AN ECONOMIC ANALYSIS OF STRUCTURAL CHANGES IN SRI LANKAN
TEA INDUSTRY
A Thesis
Presented to
The Faculty of Graduate Studies
of
University of Guelph
by DEEPANANDA P. HERATH
In partial fulfillment of requirements
for the degree of
Doctor of Philosophy
December, 2001
O DEEPANANDA HERATH, 2001 National Libtary Bibiiithèque nationaie 1+1 ,-da du Canada Acquisitions and Acquisitiorrs et B'iographic Services services bibliiraphiques
The author has granted a non- L'auteur a accordé une licence non exclusive licence dowing the exclusive permettant à la National Lihqof Canaôa to Bibliothèque nationale du Cariada de reproduce' loan, distribute or sen reproduireyprêter, disûibuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/film, de reproduction sur papier ou sur format électronique.
The author retauis ownership of the L'auteur consene la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fiom it Ni la îhèse ni des extraits substantiels may be printed or otherwise de celîe-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation. ABSTRACT
Deepananda P. Herath Advisor: University of Guelph, 200 1 Professor Alfons Weersink
This study is an empirical investigation into the structural changes of the Sn
Lankan tea industry. The industry has changed fkom one dominated by large, vertically integrated plantations to one where small famiers grow the green leaf tea and sel1 it to independent processors. The main contribution of this thesis is to develop the conceptual and ernpirical models to explain these structural changes as the cumulative outcomes of differential responses to changes in the economic incentives and constraints by the different economic agents in the tea industry.
The findings of the research indicate that relatively lower labor wages facing the small fmsector compared to the unionized labor wages paid by large famis has been an important deteminant of the expansion in small fms relative to large plantations.
Support services exclusively for srnaIl tea fmshave also contributed to the proliferation of small fanns. Despite the widely held notion that land reform policies, which took land fiom large plantations, have contributed to the reduction in large farms, these policies were found to have no effect on changes in the distribution of farm size. Taxing the exports of black tea has hurt al1 domestic growers of green leaf tea but particularly small farms. Results indicated that pncing intervention in the green leaf market promoted the growth in the nwnber of srna11 fams as the intervention may have served as a remedy for the hold up problem in the green leaf market facing these growers. The study also found that differences in labor-related production costs between
vertically integrated plantations and independent producers were important drivers of the
changes in vertical coordination arrangements. Green leaf pticing intervention has also
increased market mediated green leaf trade between vertically independent growers and
processors as the potentiai hold up threat in the arms length transactions appears to be
remedied by the pricing intervention. The nature of ownership regime and the
subsequent changes in the management incentives seerns to have had little effect on the
movement away fiom vertical integration.
The changes on the distribution of fm size and vertical coordination
arrangements in the Sri Lankan tea industry have been influenced both by various
government policies and other economic forces. in general, government progrmi and policies appears to be more influential on these changes relative to other economic forces.
However, among these policies there are some contradictory effects such as between the effect of small farm support services and export taxation. A major driver affecting both dimensions of structural change has been the relative oppominity cost of labor as large vertically integrated plantations have had to pay higher wages for their unionized employees. ACKNOWLEDGEMENTS
1 would like to dedicate this thesis to my Appachchi and Arnma who encourage
and appreciate my educational achievements from the early days.
Many thanks are extended to my advisor, Professor Alfons Weersink, for his
patience and encouragement. For the time he gave to reading the draft of this dissertation
1 am in his debt. 1 have been extremely privileged to have Professors Michael Hoy,
Glenn Fox, and Rhakal Sarker in my supe~sorycornmittee. They help me to travel
along the treacherous interface between the theory and fact. This dissertation owes a
great deal to their insightfùl comments and criticisms. 1 am gratefbl for my fnends and
teachers in the Department of Agricultural Economics and Business. They provided
excellent intellectually stimulating and exciting environment for higher leaming. I
extend my gratitude to the Canadian Commonwealth Scholarship Program, which provided the financial support for my doctoral studies at Geulph.
This dissertation would never have seen the light of the day without caring support and lirnitless endurance of Aruni, my best niend, best cntic, and wife. She has provided love, fiiendship, and a purpose for my endeavoa. Along with Amni, my two daughters, Isun and Thilini, and rny son Sachin waited (some time irnpatiently) to see my
Ph.D. through to the end. They brought me the joy and pleasure in the darkest hours of my thesis work, 1 am in their debt. TABLE OF CONTENTS
Page ACKNOWLEDGEMENTS 1 TABLE OF CONTENTS
LIST OF TABLES
LIST OF FiGURES
CHAPTER 1. INTRODUCTION
2.1. Background 1.2. Motivation: Why this study is needed 1-3. Economic Problern 1-4. Economic Research Problern 1S. Purpose and Objectives 1.6. Outline of the Thesis
CHAPTER 2. SRI LANKAN TEA INDUSTRY: DEVELOPMENT AND CHANGES
2.1. Economic Importance of Sri Lankan Tea lndustry 2.1 .A. The Importance of the Tea industry in the Sri Lankan Economy 2.1 .B. Importance of Sri Lankan Tea Production in Global Tea Market 2.2. How Tea is Grown, Processed and Marketed in Sri Lanka 2.3. Evolvement of Industrial Organization in Sri Lankan Tea Sector 2.3.A. Origin of Plantation Production Systern in Sn Lanka 2.3 .B.Origin of Larger Holdings in Sri Lanka with Co ffee Plantations 2.3.C. Establishment of Tea Plantations in British Colonial Penod: 1865 to 1920s 2.3 .D. Continuation of Plantation Companies and Agency Houses: 1920 to 1948 2.3.E. British Interests in Tea Plantations after Independence: 1948 to 1975 2.3.F. Nationalization & Land Refonn: Governent in Tea Industry: 1975 to 1992 2.3.G. Privatization and end of the State-Managed Tea Industry: 1992 to Present 2.4. Structural Changes In Sri Lankan Tea Industry 2.4.A. Changes in Farm size Distribution of the Sri Lankan Tea Industry 2.4.B. Changes in Vertical Coordination arrangements in Sri Lankan Tea Industry 2.5. Sumrnary
CHAPTER 3. REVIEW OF ECONOMIC ANALYSES OF STRUCTURAL CHANGES IN AGRICULTURE
3.O. Introduction 60 3.1. Conceptual Exposition of FmStructure: Dimensions and Measurements 61 3.1 .A. Importance of FmSize Distribution as a Measure of Fann Structure 62 3.1 .B. Inadequacies of Farrn Size Distribution as a Measure of Structural Changes 63 3.1 .C. Farm Size Distribution and Changes in other Structutal Dimensions 65 3.1 .D. Importance of Coordination Arrangements as a Measure of Farm Structure 66 3.2. Models to Explain the Forces Affecting Farm Size Distribution Changes 68 3.2.A. Technology Models of Farm Size Distribution Changes 3.2.A. 1. Economies of size mode1 3.2.A.2. Technology Adoption Model 3.î.A.3. Labor-Capital Factor Prices Model 3.2.B. Human Capital Model 3 -2.C. Financial Model 3.2.D. Public Programs and Policies as a Causative Factor of Structural Changes 3.2.D. 1. Commodity Programs as a Causal Factor of Structural Changes 3.2.D.2. Credit Policies as a Causal Factor of Structural Changes 3.2.D.3. Tax Policies as a Causal Factor of Stmc tural Changes 3.2.D.4. Research and Extension Policies as a Causal Factor of Structural Changes 3.2.E. National Economic Conditions as a Causative Factor of Structural Changes 3.3. Models to Explain the Forces Affecting Vertical Coordination Changes 3.3 .A. Structure- Conduct- Performance Model 3.3 .B.Principal and Agent Model 3.3.C. Transaction Cost Economics Approach 3.4. Summary
CHAPTER 4. CONCEPTUAL MODEL OF CHANGES IN FARM SIZE DISTRIBUTION IN SRI LANKAN TEA INDUSTRY
4.0. introduction 4.1. Technological Factors: Nature of Labor Capital Substitution 4.1 .A. Continuation of Labor Intensive Tea Production 4.1 .B-Implication of Labor Intensive Production in Fann Size Distribution Changes 4.1 .B. 1. Factor Shares in Green leaf Tea and Black Tea Production 4.1 .B.2. Simif ar Labor Requirements in Large and Small Tea Farms 4.1 .B.3. Opportunity Cost of Labor among Different Size Tea Fms 4.1 .C. Relationship between Changes in Labor Costs and Farm Size Distribution 4.2. Public Programs and Policies of the Changes in Farm Size Distribution 4.2.A. Land Reform Policies and Tea Fann Size Distribution Changes in Sri Lanka 4.2.A. 1. Land Redistribution and Changes in Tea Farm Size Distribution 4.2.A.2. Bureaucratic Management in Large Farms in Post-Re form Period 4.2.A.3. Relationship between Land Reform Policies and Farm Size Distribution 4.2.B. Small Tea Farm Support Services and Size Distribution Changes 4.2.B.1. Relationship between Srnall FmSupports and Farm Size Distribution 4.2.C. Tea Industry Taxation Policy and FmSize Distribution Changes 4.2.C. 1. Nature and Evolvement of Tea industry Taxation in Sri Lanka 4.2.C.2. The Effect of Tea Export Taxes on the Price received by the Tea Producers 4.2.C.3. How would Export Duties AfTect Different FmSizes Differently? 4.2.D. Govertunent Intervention in Green Leaf Tea Market 4.2.D.1. Evolvement of Green Leaf Tea Price Intervention in Sri Lanka 4.2.D.2. Inefficiently low Green Leaf pnces before 1968 period 4.3. Export Market Price Instability and Changes in Tea Farm Size Distribution 4.4. National Economic Conditions and Changes in Size distribution in Tea Farms 4.5. Sumrnary
CHAPTER S. CONCEPTUAL MODEL OF CHANGES IN VFRTICAL COORDINATION IN TEA INDUSTRY
5.0. introduction S. 1. Overview 5.2. Changes in Production Costs and "Malce" versus "Buy" decisions 5.3. Changes in Management Costs and "Make" versus "Buy" Decisions 5.4. Changes in Transaction Costs and "Make" versus "Buy" decisions 5.4.A. Green leaf tea Pnces in Arms length Transactions 5.4.A. 1.Cornpetitive behavior Among Green leaf tea Producers and Processors 5.4.A.2. Green Leaf Prices and Arms length Transaction in Other Countries 5.4.B. Hold Up Threats in the Arms Length Transaction 5.4.C. Conceptua1 Mode1 of Green Leaf Market in Sri Lankan Tea Lndustry 5.4.D. Optimal (Joint Profit Maximizing) Green Leaf Pricina Arrangement 5.4.E.Green Leaf Prices with the Possibility of Hold up Problem 5.4.F. Green Leaf Pnces under Hold up Threats & "Buy" Decisions 5.4.G. The effect of Pncing intervention & Changes of "Make" vs. "Buy" decisions
5.5. Strengths and Weaknesses of Hold-Up Arguments in Green Leaf Market
5.6. Summary
CHAPTER 6. EMPIRICAL MODELS OF STRUCTURAL CHANGES IN SRI LANKAN TEA INDUSTRY
6.1. Empirical Model for the Changes in Size Distribution of Sri Lankan Tea Farrns 6.1 .A. Dependent Variable for Sri Lankan
Tea Farm Size Distribution (Yi) 6.1 .B.Explanatory Variables for Farm Size Distribution Changes 6.2. Empirical Model for the Changes in Vertical Coordination in Tea Industry 6.2.A. Dependent Variable for the Vertical Coordination
Arrangements (Y2) 6.2.B. Explanatory Variables for Changes in Vertical Coordination Arrangements 6.3. Summary
CHAPTER 7. ESTIMATION OF THE EMPIRICAL MODELS
7.0. introduction
7.1. Functional Form Specification of the Econometric Models 227 7.2. Estimation of the Empirical Model of Farm Size Distribution changes 7.2.A. Testing for Competing Functional Forms
7.2.B. interpreting the Results of Econometric Estimation
7.3. Estimation of the Model for the Changes in Vertical Coordination Arrangements 248 7.3.A. Testing for Competing Functional Forms 249 7.3.B. Interpretation of the Results of Econometric Estimation 250 7.4. Summary 253
CHAPTER 8. SUMMARY AND CONCLUSIONS 8.1. Summary 8.2. Conclusions and Policy Implications 8.3. Limitations of the Study
REFERENCES ANNEX 6A ANNEX 6B ANNEX 6C ANNEX 6D ANNEX 6E ANNEX 6 F ANNEX 6G ANNEX 6H LIST OF TABLES Table Page 2.1. Composition of Gross Domestic Product of Sri Lanka 15 2.2. Export Earnings from Sri Lankan Tea sector 16 2.3. World Major Tea Producers and Exporters 17 2.4. Regional Distribution of Tea Plantations in 1920s (acres) 37 2.5. Net Immigration of South indian laborers 1879 to 1920 (in Thousands) 38 2.6. Ownership Changes in Sri Lankan Tea Land 1959 to 1972 (in acres) 44 2.7. Land expropriated by Land Reform Commission: by crop type (acres) 46 2.8. Post-Nationalization Tea Production in Sri Lanka (Kg Mn) 48 2.9. Regional Level Tea FmSize Distribution Changes: 1978 to 1994 52 2.10. Size Distribution of Sri Lankan Tea Fms(Number & Area) 54 2.1 1. Changes in Green Leaf Purchase by Processors (no. of processors) 57 2.12. Black Tea Production from Purchased & Self-grown Green Leaf (Mn kg) 58 3.1. Structural Dimensions of Farming Industry: a Synthesis 62 4.1. Changes in Labor use for Tea Production 1954- 1993: Large Tea Farms 102 4.2. Labor per acre in Sri Lankan Tea lndustry (large farms) 103 4.3. Labor Cost in Tea Production (Cents/lbs) 107 4.4. Labor Use for Green Leaf Production: Large and Small Tea Farms 108 4.5. Land acquisition by Land Reform Commission: by crop type (acres) 116 4.6. Recipients of Lands under the Land Refonn Program ('000 acres) 117 4.7. Redistribution of Tea Land acquired with 1975 Amendment 118 4.8. Cooperative Involvement in Management of Tea Land (acres) 123 4.9. Chronological Information of Support Services for Small Tea Fann 128 4.10. Total Costs, DutiesTTaxes, and Profits arnong World Tea Producers 136 4.1 1. Concentration Ratio of Tea Sales by Major Firms in Retail Markets 139 4.12. Four-Buyers Concentration Ratio in three Major Tea Auctions 139 4.13. Export Taxation in Tea: Kenya and Sri Lanka, 1985 140 4.14. Price Elasticity of Supply among different Countries in Tea Production 143 4.1 5. Decline in Real Black Tea Price index in 1960s (1960=100) 148 4.16. Changes in Colombo Consumer Price index 149 4.17. Changes in Guaranteed Minimum Price for Green Leaf Tea 4.18. Price Instability index for Tropical Beverages and al1 Commodities 5.1. Large Fann (Estate) & Independent Green Leaf Processors in Sn Lanka 5.2. Estimated Cost of Developing and Bringing into Bearing Tea and some Alternative Crops in Sn Lanka (nominal prices in 1968) 6.1. Factors Affecting Relative Share of Farms sizes: Variable Description and the Expected Signs of the Coefficients
6.2. Factors effecting on the Share of Tea Production in Vertically Independent Sector: Variable Description and the Expected Signs of the Co-efficient
7.1. Test for Competing Functional Foms for Fann size Distribution Changes
7.2. Maximum Likelihood Estimates for the FmSize Distribution Changes
7.3. Test for Competing Functional Forms for Vertical Coordination Changes
7.4. Maximum Likelihood Estimates for the Vertical Coordination Changes LIST OF FIGURES Figure
2.1. Major Activities and the Participants of the Tea industry 2.2. Tea Land Area in Sri Lanka from 1868 to 1998 2.3. Farm Size Distribution Changes in the Sri Lankan Tea industry 3.1. Causal Factors of Changes in Structural Dimensions: A Synthesis 4.1. Dnvers of Farm Size Distribution changes in Sri Lankan Tea Industry 4.2. Relative Price changes in labor and capital in Sri Lankan Tea Industry 4.3. Minimum Labor Wages for Agriculture and Non-Agriculture Sectors 4.4. Marginal Value Product and Cost of Labor in Small and Large Farms 4.5. Changes in Management Input: 1980 to 1998 4.6. Support services for small tea fmsin Sri Lanka 4.7. Government Revenue and Expenditure in Sri Lanka 4.8. Tea Tax as a Percentage of Tea Export Earnings 4.9. Relative changes in Kenyan Tea Exports and Sn Lankan Tea Taxes 4.10 Nature of Export Taxes Burden on Primary Producers 5.1. Changes in Black Tea Production among VerticaIly Independent and Vertically Integrated Sectors in Sri Lanka 5.2. Changes in Joint Profit with the Change in Green Leaf Price 5.3. Variation in Joint Profit under Di fferent Green Leaf Tea Pices 5.4. Changes in Green Leaf Production under State Price Intervention 5.5. Tea industry with no Price intervention in Green Leaf Tea (1 960-1 968) 5.6. Tea Industry with Guaranteed Pnce for Green Leaf Tea (19694985) 5.7. Tea Industry with Formula Priced Green Leaf Tea (1986-1998) 6.1. Black Tea Production: Low Grown and fiom Purchased Green leaf CHAPTER 1
INTRODUCTION
1.1.Background
The structure of the agriculture sector has motivated the fmpolicy debate in many countries throughout the past century. Structure, as defined by elements such as average farm size and vertical linkages, has been the evidence provided to support policy efforts ranging fiom those that increase efficiency to those that protect a certain way of life. Changes in the structure of agriculture are attributable to changes in the choices made by fanners and other agents in the sector as they respond to underlying changes in economic constraints and incentives such as changes in relative prices of inputs and outputs, changes in technological parameters, and changes in costs of exchange arrangements. The cumulative outcome of their choices over time is manifested as structural changes in agriculture. Economic policies attempt to shape constraints and incentives with the aim of attaining a desired structure of agriculture. Thus, it is vital to undentand the association between farmers' responses and changing economic constraints and incentives, if one were to evaluate the potential of economic policies to affect farm structure.
In economic, social, and political realms, concerns are expressed about the causes as well as the consequences of structural changes in agriculture. These concerns do influence the various economic policies intended to change the structure of agriculture.
Economic concems are related to the concentration of larger fmsand higher production efficiency in such fms(Rhodes, 1998, Zering, 1998). Differences in efficiency under different farm size distribution influence the cost of fmproducts and the prices of farm products (Sumner, 1985). On the other hand, it is argued that concentration of production
within fewer, larger fmswould lead to anti-cornpetitive outcomes, and prices of farm
products may increase.
Social concerns of structure are based on society's desires to have a particular
structure in agriculture. The long-term growth in fmsize has led people to believe that
large fms are more efficient than small hsand consequently, farnily fhrms are
endangered (Tweeten, 1984). Many argue that family hsare essential to social well-
being through more equitable wealth distribution and to competitive production of adequate food supplies (Tweeten, 1993, Stanton 1993, Raup 1978). A relative incorne
level between fmsizes/types and non-fm sectors is also a social concem. Other equity issues dealing with structural change include the reliance on non-fam incorne b y the farm sector and the lack of comrnunity seMces in rural areas dominated by a few
firms (Bal1 and Hçady, 1972).
Political concerns of farm structure arise from inequitable effects of fmpolicies on various pressure groups (Knuston, Penn, and Boehm 1995). Generally, fmsupport prograrns are geared to protect fmproducen imposing welfare losses on taxpayers. It has been argued that these programs could be the outcome of differential lobbying activities by vanous pressure groups (Gardner, 1989, 1992). Thus, it is argued that farm support programs are a means of achieving income redistribution given the unequal lobbying activities of these pressure groups. Collectively, agriculture structure affects the wellbeing of the farm population, consurners, and the environment, so different structures will bnng about different impacts for these domains. This is a study about structural changes in the Sri Lankan tea industry over the
last four decades and about the causes of such changes. The tea industry is one of the
most important perennial crop sectors in the Sri Lankan economy and its structure, most
significantly the farm size distribution and vertical coordination arrangements, have
substantially changed. This study attempts to explain the structural changes in the light
of choices made by tea fmers and processors with the changing economic incentives
and constraints faced by them. This knowledge would allow one to assess the scope and
capabilities of the economic policies that strive to change the structure of Sri Lankan tea
industry.
1.2. Motivation: Wby this study is needed
The structural changes in the Sn Lankan tea industry contrast the structural
changes witnessed in developed world agriculture. in North American and European agriculture, the components of the supply chain are increasingly dominated by fewer and
larger fanns whose efforts are coordinated through the verticatly related stages by negotiated contracts rather than the traditional market mediated coordination (Hallberg,
2001, Bohelje, 1999, Hennessy, 1999, Monke et al, 1993). However, in the Sn Lankan tea industry the relative share of small fmsas a proportion of the total number of tea
fmsand as a proportion of the total tea area has been increasing since the 1970s. The proportion of black tea processors who procure their green leaf through market has also increased while the share of vertically integrated green leaf production and black tea processing firms has declined since the 1970s. in sum, the Sri Lankan tea sector has moved from an industry structure dominated by plantations which produce the green leaf and process it into black tea, one where the green leaf is now predominantly grown by small fmers and subsequently sold to independent processors.
The changes fiom a vertically integrated plantation sector to one with separate production and processing units is not unique to the Sri Lankan tea sector and such changes have been noted for many other developing countries in Asia and Afnca such as in the Kenyan tea and coffee industries (Lele and Agrawal, 1989), the indian rubber industry (Tharian, Haridasan and Sreekumar, 1988), and Guyana sugar industry (Thomas,
1979). Sajhau and Von Muralt (1987), IL0 (1989), Tiffen and Mortimore (1990), and
IL0 (1994), discuss such structural changes for many other developing countries. Most of these countries show similar histoncal episodes carkaturing the different stages of the changes in the plantation industry structure.
British or other European entrepreneurs, in their world-conquering era of the eighteenth and nineteenth centuries, established commercial agricultural enterprises. The production organization of such commercial enterprises came to be known as plantations.
The main characteristics of plantations are the captive labor force whose wages were artificially kept low, and the hierarchy of supervisory and technical personnel, with large- scale cultivation of crops (Beckford, 1972). Plantations were viewed as a technologically superior sector producing for export while there CO-existeda technoiogically backward peasant sector producing at a subsistence level. The dichotomy between the technologically advanced "export crops" sector and the backward "subsistence crops" sector was known as the dualistic mode1 of the developing economies (Bocke, 1953).
In the latter half of the twentieth century, many countries obtained independence fiom their European colonial rulers and nationalization regimes followed. Since independence, governrnent involvement haç been extensive in agriculture with a major development goal being the improvement and integration of the peasant sector (Barlow and Jayasooriya, 1986, Hayami, 1994). Sweeping state interventions in these plantation industries were noted in most of the above countrîes. Policies such as small fmsupport programs, export taxation, land refonn, pricing interventions of the export crops, over- valued foreign exchange control regimes and in some countries nationalized plantation management were noted (Sajhau and Von Muralt, 1987, Tiffen and Mortimore, 1990,
ILO, 1994). These plantation industries, during the last couple of decades, bave evolved to a different industry structure. This structure is characterized by the predominance of small fanns together with the non-market methods of vertical coordination arrangements such as contract production, or parastatal organizations for processing and marketing
(Barlow and Jayasuriya, 1978, ILO, 1989, Stall, Delgado and Nicholson, 1997).
Interestingly, these changes are opposite to the changes in agriculture within the developed countries, where the enlargement of farm sizes and tighter vertical coordination arrangements have been witnessed.
The literature that fucuses on the changes in plantation agriculture in developing countries has been in the broad sociological tradition with the main contributions fiom the sociologists, political scientists and histonans' (Pryor, 1982). Even the political economic analysis of plantations such as Beckford (1972), Courtney (1980), De Silva
(1982), Bandarage (1982), and nurnerous others2 are focused on explaining the factors
- I See the collection of articles in "Plantation Systems of the New World. Pan American Union Washington D.C. 1959.
see Bieber, 1997 for a collection of such studies that have given rise to the establishment of plantations as a production organization in the
time of western colonial expansion.
Political economic analysis specifically focused on the Sri Lankan tea industry
such as Dawood (1984) and Rote (1986) have described the trends but have not looked at
the underlying causes specifically changing economic environment facing individual
firms. These studies do not evaluate the relative importance of economic fundamentals
such as changes in relative prices, technological parameters, or the cost of transacting as
the causes of structural changes witnessed in the Sri Lankan tea industry. There is a lack
of knowledge as to how the responses of tea fmers to the changes in economic
incentives and constraints have led to the structural changes observed in the Sri Lankan
tea industry. How could these structural changes be explained as the cumulative outcome of the changing choices of fmers as they respond to the changes in economic incentives and constraints?
The literature that focuses on structural changes in developed world agriculture is substantially different fiom the historical and sociological description of the changes in the structure of plantation agriculture in developing countries. The literature about the structural changes in North American and European agriculture attempts to empirically analyze the fanning industry changes as a response of economic agents to the unfolding economic incentives and constraints in the fanning industry. For instance, changes in relative factor pnces and factor ratios of labor and capital in the farming industry have been able to explain almost al1 of the variation in farms size distribution in United States agriculture (Kislev and Peterson, 1982). The knowledge gained through such analysis is critical in evaluating the effectiveness of the policies proposed to change the farming industry structure to harmonize with various economic, political and social objectives.
The analytical approach that focus on more fundamental motives of economic agents in the famiing industry could be miitfûlly expnded to explain the structural changes taking place in the Sn Lankan tea industry.
1.3. Economic Problem
Govermnent policies are one of the important vehicles to bring about structural changes in agriculture to meet the economic and political objectives of the economic agents in the agriculture sector. These policies have efficiency and distribution implications. For instance, it is argued that fmprice subsidies impose welfare losses on consumers and can lower international competitiveness (Gardner, 1992). However, these programs are implemented with the objective of supporting the agricultwal sector and especially small and medium size fms(Tweeten, 1984). Whether the desired structural changes result fkom the policy depends on the responses of individual fmers. Thus, in evaluating the effectiveness of these policies, knowledge about the relationships between the observed structural changes and the responses of farmers to the changing incentives and constraints fiom the policies and other economic forces is important. Without the knowledge of the relationships that determine the responses of bersto the changes in the incentives and constraints, evaluation of the policies designed to influence these incentives and constraint would be impossible. The general dissatisfaction about fm programs is partly attributable to lack of this knowledge (Rodgers, 1988).
As in many other developing countries, governent involvement is observed in the Sri Lankan tea industry with policies such as the nationalization of large tea farms, export taxation, mal1 farm subsidies, and pricing intervention in the intermediate product
(green leaf) market. Some of these policies are implemented with a policy objective of achieving a particular structure in the tea industry (Peiris, 1984). For instance, a small farm subsidy is a policy with an objective of influencing fmsize distribution.
However, the response to this incentive among small tea fms, which would finally determine the possibility of this policy to affect the structure, is an inadequately understood process.
There are many theories about what causes structural changes and data indicating that structural change has occurred. However, the knowledge of how farmers respond to changing economic incentives and constraints, which give rise to the structural changes over time, is lacking in the context of Sri Lankan tea indusûy.
1.4. Economic Research Problem
This study examines the factors leading to structural changes in the Sri Lankan tea industry. This investigation is a departure fkom the existing literature about the structural changes of the Sn Lankan tea industry, which has been confined to political economic assessrnents and to stating hypotheses about the impacts of certain structural change drivers. These studies tend to describe the types of govemment intervention in the tea industry with qualitative discussions on the possible associations between these govemment intervention and some dimensions of the structure (Dawood, 1980, Peiris,
1984, Rote, 1985, Sajhau and Von Muralt, 1987).
However, the most important departure from this literature in this thesis is that it acknowledges that structural changes corne about due to differential responses of the different categories of farmers (eg., small versus large, or vertically integrated versus
vertically independent) to the changes in economic incentives and constraints. If famiers
in al1 the categories responded proportionately to the changing incentives and constraints,
the structure of the fming industry would be unchanged. Therefore, it is important to
use economic reasoning to capture the differential responses of the various categories of
farmers to the changes in economic incentives and constraint. The multifaceted nature of structural changes are rarely acknowledged and modeled in the existing literature.
Economic reasoning can provide a conceptual mode1 to explain why different categories of fmsizes or vertical coordination arrangements respond differently to economic incentives and constraints. The different responses of farms in different categories of a given structural dimension (i.e, size distribution or vertical coordination arrangements) to these economic forces must be empirically demonstrated. How one can explain these different responses by appealing to the economic reasoning is the economic research problem of this study.
The economic forces affecting the changes in these structural dimensions could be many and among such variables, policies that are argued to shape the economic incentives and constraints face by farmers must be included. Data on the variables affecting farm structure and the data on the changes in structural dimensions has to be gathered in order to empirical verify these claims. The changes in structural dimensions such as size distribution and vertical linkages have to be represented as quantifiable variables. The thesis develops a conceptual framework to explain the structural changes witnessed in the Sri Lankan tea sector and empirically test the resulting hypotheses. 1S. Purpose and Objectives
The purpose of thîs study is to investigate the forces that have caused the Sri
Lankan tea industry to move fkom an industry structure dominated by large, vertically integrated plantations to an industry structure consisting primarily of small fmers producing the green leaf tea and selling it to independent processors who convert it into black tea. The knowledge on the association between the economic forces (incentives and constraints) and structural changes is necessary to evaluate the effectiveness of govemrnent policies and their impact on the farrn size distribution and vertical linkages.
The specific objectives are to;
(1). Describe the structural changes in the Sn Lankan tea industry during the last four decades and identify the important structural dimensions that have undergone changes.
(2). Identify the potential drivers of structural changes by reviewing the literature on the economic analysis of factors affecting the different dimensions of industry structure.
(3). Develop a conceptual fkamework to explain the effects of the economic forces on the structural dimensions of farm size distribution and vertical linkages and to generate refutable hypothesis fiom this fiamework.
(4). Empirically specify the variables to represent the structural dimensions and the structural change drivers to test the above hypotheses.
(5). Extend the implications of the empincally verified associations between the economic incentives and constraints and the structural changes to evaluate the impact of policies that are argued to be effective in influencing structural changes in Sri Lankan tea industry and comparable situations in other countries. 1.6. Outline of the thesis
Chapter 2 describes the major structural changes that have occurred in the Sri
Lankan tea industry over the last four decades. The chapter begins with a discussion of
the Sri Lankan tea industry within its domestic economy and within the global tea
market. The second section provides a synopsis of how tea is grown, processed and
marketed in Sri Lanka. The third section describes the historical evolvement of the
current industrial organization in the Sri Lankan tea sector including the inception of
plantation agriculture with coffee cultivation by the British in the middle decades of the
19" century to the establishment of tea plantation. The final section presents a detailed
description of the changes in farms size distribution and vertical coordination
arrangements in the Sri Lankan tea industry.
Chapter 3 provides a review of the literature on the economic analysis of
structural changes in agriculture. A conceptual exposition of the structure of agriculture
is provided, major structural dimensions of fming industry are identified, and their
measurements are outlined. This chapter is organized as follows; Section 3.1 synthesizes
and describes the important dimensions of fmindustry structure and the measurements
of the changes in these dimensions. Section 3.2 synthesizes the literature about the
important economic forces or drivers that shape the different dimensions of farming
industry structure. Subsections under Section 3.2 review the major theoretical models of
fann size distribution such as technology models, the human capital model, and the
financial model. Section 3.3 reviews the major theoretical models of vertical coordination arrangements such as the structure-conduct-performance model, the
principal-agent model, and the transaction cost approach. Chapter 4 develops a conceptual fiarnework to explain the changes in fam size distribution that occurred in the Sri Lankan tea industry over the last four decades. The specific details of major factors or dnvers that affect the changes in farm size distribution in the Sn Lankan tea industry are presented in this chapter. The economic reasoning for the effect of specific drivers on the changes in tea fami size distribution is presented.
Refütable hypotheses about the effects of specific drivers of the changes in tea farrn size distribution are developed.
Chapter 5 develops a conceptual model to explain the changes in vertical linkages within the Sri Lankan tea industry. Section 5.1 provides an overview of the important determinants of the choice of vertical coordination arrangements. Section 5.2 discusses the production costs differences due to the scale economies in the intermediate input production and its effect on the "make" vs. "buy" decisions. Then the green leaf production cost differences in the "make" versus "buy" decisions for the Sn Lankan tea industry is discussed. Section 5.3 discusses the changes in management incentives and agency costs that consequently affect the changes in "make" vs. "buy" decisions. The changes in agency costs and management incentives in the Sri Lankan tea industry and the hypothesized effect on vertical coordination changes are outlined. Section 5.4 develops a theoretical model of the impact of transaction costs on the vertical coordination arrangement. The model is focused on the green Ieaf tea market and incorporates the effect of hold up problems in the green leaf market equilibriurn. The mode1 is extended to incorporate the effect of state pricing intervention in green leaf market and the consequent effect on the observed changes in vertical coordination arrangements. Testable hypotheses are developed at the end of the sections 5.2 to 5.4.
Section 5.6 summarizes the main points of the chapter.
Chapter 6 specifies the empincal models to test the hypothesized effects of the
dnvers of the changes in these smictural dimensions and to describe the variables
representing the structural dimensions and the dnvers of their changes. The first section
specifies the empirical mode1 for the changes in farm size distribution, describes the
variables representing size distribution change and the drivers. The second section does
the same for the changes in vertical coordination arrangements.
Chapter 7 explains the results of the econometric estimation of empincal models.
Section 7.1 reviews the empirical model of the tea fmsize distribution changes and
then discusses the results of econometric estimations. Section 7.2 reviews the empincal
model of the vertical coordination changes and summarizes the results of the econometric estimations. Section 7.3 surnrnarizes the main findings of the analysis of the structural changes in the Sri Lankan tea industry.
Chapter 8 summarizes the thesis and discusses the conclusions and policy implications. The limitations of the theoretical approach and the rnethods along with the data are also discussed in chapter 8. CHAPTER 2
SRI LANKAN TEA INDUSTRY: DEVELOPMENT AND CHANGES
The purpose of this chapter is to describe the major structural changes that have occurred in the Sri Lankan tea industry over the last four decades. Subsequent chapters will develop the conceptual and empirical models for detemining the factors driving these changes. The chapter begins with a discussion of the Sri Lankan tea industry within its domestic economy and within the global tea market. The second section provides a synopsis of how tea is grown, processed and marketed in Sri Lanka. The third section descnbes the establishment of plantation system by the British in the 19th century with the coffee cultivation and the subsequent evolvement of the industrial organization of the
Sri Lankan tea sector. The final section describes the stnictural dimensions of the Sn
Lankan tea industry that have substantially changed during the last four decades.
2.l.Economic Importance of Sri Lankan Tea Industry
2.l.A. The Importance of the Tea Industry in the Sri Lankan Economy
Tea is one of the most important perennial crops in Sri Lanka in tenns of land use, employment, government revenue generations, and foreign exchange earnings.
Presently, tea covers about 195,000 hectares or about 9% of the cultivated land of the country (Annual Report, Central Bank Sn Lanka, 2000). The tea industry employs about
500, 000 and with direct and indirect linkages with the other industries, about a million people are dependent on the tea industry for their livelihood (Mendis, 1992). Presently, tea has become an important cash crop for the rural peasants in Sri Lanka. The relative importance of the tea industry to the general economy has been declining during the last few decades as its share in GDP fell from 17% in the 1950s to 2% in 1990s with the industrial and service sector gaining in dominance (see Table 2.1).
Table 2.1. Composition of Gross Dornestic Product of Sri Lanka
1950 1960 1970 1982 1990 1996 1997 1998 Rs mil. % Rs mil. % Rs mil. % Rs mil, % Rs mil. % Rs mil. % Rs mil. % Rs mil %
...... Other 1303 33 1899 31 2365 24 22546 24 27007 2 1 145775 21 163089 20 178217 20
Services 1521 39 2417 40 4438 45 44828 47 62463 48 355770 51 41 1747 51 468818 51
GDP 3942 100 6072 100 9888 100 94679 100 129167 100 695934 IO0 803698 IO0 912839 100 Sources: 1950 and 1960, fiorn Snodgrass, 1966 (Table A-8) and al1 the other years are ftom various issues of Annual Reports of Central Bank of Sri Lanka.
Notes: Values for 1950 and 1960 are cwrent factor costs, 1970 is with 1959 factor costs, 1982 and 1990 are with 1982 factor costs, and 1996,1997 and 1998 are with 1996 factor costs
Industry: Mining and Quarrying, Manufacturing, Construction, Electricity, Gas, Water and Sanitary Services
Services: Transport, Storage and Communication, Wholesale and Retail Trade, banking, Insurance and Real Estate, Ownership of Dweltings, Public Administration, and Defense
Yet, tea significantly contributes to export earnings. Tea exports accounts for about 70% of the total agricultural export earnings and about 15% of the total export earnings of the country (see Table 2.2). Since the industrial sector utilizes a larger volume of imported inputs, its net export earning is lower than that of tea.
Tea was an important source of governent revenue before 1992, however, since then the export and ad-valorem duties on tea were abolished. For example, in the period of 198 1-1985 annual average tax revenue from tea exports was about US. $ 108 million. This was about 11% of the government tax revenue durhg this period (Annual Report of
Central Bank of Sri Lanka 1990 Table 67).
Table 2.2. Export Earniags frorn Sri Lankan Tea sector
Export Eamings (million US $) 1996 1997 1998 Tea 616.37 720.97 778 -11 Total agricultural exports 962.64 1062.28 1087.19 Total industrial exports 3013.15 3442.97 3552.04 Total ex~orts 4 103.36 4647.83 4742.44
Tea share from aericulturalY ex~ortx eamines . 64% 68% 71% Tea share fiom total export eamhgs 15% 15% 16% Source: Annual Report 1998, Central Bank of Sri Lanka (Table 77).
2.1.8. Importance of Sri Lankan Tea Production in Global Tea Market
Developing countries in south Asia and east Afiica account for more than 85% of world tea production and exports (see Table 2.3). india is the biggest tea producer in the world and China, where tea drinking originated, cornes in second place. Sri Lanka is the third largest producer contributing about 10% of the world tea production, Since the two largest producers are also the largest consumers, their tea exports are small (only about
20% of their production) compared to Sri Lanka (as much as 95% of her production is exported). Thus, Sri Lanka has been the world's largest tea exporter contributing about a quarter of the total world exports. Despite some setbacks in the 1970s and 80s, which made the Sri Lankan share of world exports decline, it continues to be a major tea exporter. In 1999, Sri Lankan tea production reached its highest level ever (284 million- kg) and in 1998, the volume of tea exports peaked at 272 million-kg (Annual Report
Central Bank of Sri Lanka, 1999). Table 2-3.World Major Tea Producers and Exporters
Production (000 tones) Exports (000 tones) 1993-95 1996 1997 1993-95 1996 1997 (average) (average) World Total 2575 2648 2687 1088 1114 1167
China 613 617 637 187 173 205 Sri Lanka 24 1 259 277 223 234 258 indonesia 139 166 149 96 102 67 Japan 88 89 91 - - - Bangladesh 50 53 54 27 26 25 Africa Kenya 222 257 22 1 203 244 209 Malawi 36 38 44 36 37 49 Tanzania 24 20 22 19 18 19 Others Argentina 40 43 55 43 41 56 Source: Food and Agricutture Organization, Commodity Market Review 1998- 1999: Commodity Trade Division, Rome, 1999: p.39.
2.2. How Tea is Grown, Processed and Marketed in Sri Lanka
Tea industry activities are spread over three distinct phases: (i) green leaf production (ii) black tea processing and (iii) black tea marketing. The links among the economic agents involved in the three phases are shown in Figure 2.1. A tea bush has an economic life span of about thirty to fifty years but it takes about five to six years fiom initial land preparation to the first harvesting. Thus, extensive soil rehabilitation is done before the field establishment of tea plants to ensure long lasting desirable soil physical properties. The soil rehabilitation is accomplished by growing a deep rooting, fibrous gras variety (Tripsaczrm Lmmor Cymbopogan confertiflorus) for about two years and lopping and spreading the grass in the field. Field Cultivation of Tea
f Planting Fertilizing Small holders Larger Estates Pest and Disease Control Shade Management Soi1 Conservation Weeding Pruning /\ /
w Black Tea Processing by Black Tea Processing 1 Independent Processors by Large Estates
F W ithering Rolling Fermentation Firing Shi fting and grading Packing Tea Chest
\ 1-4 1-4 Transportation to Auction Houses in Colombo +
Warehouse Facilities Tea Tasting Brokerage Activities 1
Tea Auctioning by the Colombo Tea Auction World Market Prices
Figure 2.1. Major Activities and the Participaats of the Tea Industry Young tea plants for the field establishment cornes from commercial nurseries.
Leaf cuttings (a single node cutting consisting of a nodal leaf with an auxiliary bud and inter-node) are taken from a stock of mother bushes and are grown in the nursery for about a year to eighteen months. About 12,500 plants are planted per hectare at a spacing of 4 feet between rows and 2 feet between plants in the row. Growth of the young tea plants is modified into a iow spreading bush during the first three years in the field. This bush orientation provides a convenient height for manual harvesting and rnany harvesting points. After about three years in the field, the young tea plant is ready for continual harvesting.
Harvesting of tea by selective hand picking is known as plucking. The harvest is the tender shoots containing two Ieaves and a bud fkom the growing apical points. The shoots, popularly known as 'green leaf, are plucked at regular intervals of 4 - 10 days throughout the year. Therefore, about fifty to sixty (depending on the interval) plucking rounds are carried out during a year. The shorter the plucking interval the more plucking rounds per year. The plucking interval is dependent on the climatic conditions, and vigor of the tea bush. Generally, a plucking round provides about 175 kg of green leaf from a hectare of tea. Pruning is one of the most important field cultivation practices which enables year round harvest by maintaining the vegetative phase of the tea bush.
The two producer categories, which are involved in green leaf production (see
Figure 2.1), are arbitrarily classified on the basis size. Yet, these two size classes have distinct production orientations. Small holders are considered those farms with tea gardens less than 50 acres and without their own processing facilities. This definition was operationally established with the 1972 and 1976 land refonn laws when al1 the tea farms larger than 50 acres were acquired by the state. Presently, there are about 206,000
small holder tea producers in Sn Lanka who grow about 259,000 acres of tea. Average
size of a smallholder fmis about 1 acre, and more than 98% of the smallholdings are
less than 5 acres in size. Their fming approach is labor intensive with low fixed costs
(Sriyaratna, 1994). However, the fixed cost requirement in tea growing is higher
compared than alternative cropping activities. This is due to specific requirements in
land preparations such as soil rehabilitation, soi1 erosion control measures, and shade
trees establishment. The small holders are O fien characterized as owner-operators, with a
small amount of hired labor. Family labor playa a major role compared to hired or
permanently employed labor. They are often part time tea growers as 62% of them have
off farm income. The main sources of non-fm income are fiom other small enterprises,
govemment employment, and private sector employment (Report of the Baseline Survey of Tea Small Holdings, 1994, p. 10).
Presently, about 400 large estates grow tea in Sri Lanka, which are owned and operated by plantation management companies. Agarapatana Plantation Company manages the Iargest tea area of about 7,400 hectares covering 21 estates. The smallest
Company, Kegalle Plantation Company, manages a tea land area of 1480 hectares covering 12 estates (Plantation Sector Statistical Pocket Book, 1997 p. 109). These large estates tend to use a plantation mode of production that can be characterized by a large residential labor force, a hierarchical labor management system, own processing facilities, a large volume of working capital, higher fixed costs, and an intensive fming approach (De Silva, 1982). The green leaf produced by large estates and small holders enter as an intermediate product in black tea processing. Tea processing starts with the withering of the harvested green leaf by circulating heated air through them. The withered leaves are then rolled with heavy rollers. Rolling crushes the green leaves into small granules. This exposes the ce11 contents for an enzyrnatic reaction, which is known as fermentation.
Ferrnented tea granules (dhool) are then placed in a drier. Once dried, the tea is cleaned and classified with a variety of equipment into different grades.
Black tea processing is a capital-intensive undertaking carried out either by the large estates themselves or by independent processors. The basic operational aspects such as capacity, working capital and labor requirements are not different between the two types of processors. An annual average processing capacity of a processor is about
250,000 kg of processed tea. There is no small-scale cottage processing technology available to small holders.
In 1996, there were about 381 large estate processors and 215 independent processors in the black tea processing sector. The independent processors are also known as "bought leaf processors" since they purchase their green leaf mainly from small holders but also occasionally from the larger estates. The large estate processors, on the other hand, get their green leaf fiom their own plantations with some supplemented by the small holders' green leaf. For instance, in 1996, 111 of the 381 large estate processors used only their own green leaf. The remaining 270 large estate processors supplemented their own green leaf with the small holders green leaf (Sriyaratna, 1994).
After the green leaf is processed into black tea, it is graded and packed in wooden cases or paper bags each weighing about 50 kilograms. The market channels to sel1 the black tea include the Colombo tea auctions, private sales, forward contracts, direct sales, and ex-factory sales. Except for the Colombo tea auctions, al1 the other market channels are seller (processor)-initiated and sales are directly to the packers. Tea processors, by law, are required to sel1 at least 50% of processed tea through the Colombo auctions.
However, about 97% of the country's tea production is sold through the Colombo tea auction and so it is the most important marketing channel.
The Colombo Tea Traders Association conducts the Colombo tea auction. It is the largest tea auction in the world and handles about 220 million kilos of black tea annually. Teas at the auction are sold by one of the seven licensed brokers. The processors dispatch black tea 'lots' to the brokerage firms and each 'lot' is a specific grade of tea. The brokerage firms store the tea in their warehouses, draw samples fiom the tea 'lots', taste and value the tea, catalogue it, send samples to prospective buyers, and put up tea for sale at the auctions. In the auction, these 'lots' are offered for bidding among the buyers and after the sale the brokers pay the processors afier coltecting payrnent fiom the buyers.
There are presently about 150 registered buyers that can be divided into two groups. One category of buyers is the agents of multinational companies that dominate the world tea market by blending and packing for retail level distribution. Before export, some of the bulk teas purchased in the Colombo auction by such agents go through value addition such as tea bagging and packaging. The remaining bulk teas are exported for blending and retaihg by the multinational firms. This category dominates the Colombo tea auction (the principal buyer is Uni-lever, which owns both Brook Bond and Lipton).
In 1993, the largest 10 buyers bought about 62% of al1 the teas auctioned in Colombo. Nine of those buyers were in the 10 largest tea exporters accounting for about 55% of total tea exports (Report of the Presidential Commission on the Tea Industry and Trade
1995 p.33 and 57). The other category consists of medium or small size buyers se~cing for local wholesalers and sometimes for export orders.
The above basic components of the tea industry as depicted in Figure 2.1 and the production and marketing organization has been there throughout the tea industry history.
However, the relative importance of these components as well as their relationships has markedly changed during these historical episodes. Changes that occurred in the organization of green leaf production among srna11 farms and large fmstogether with vertical linkages between the green leaf production and black tea processing are significant. However, the black tea marketing arrangement through an auction system has not been significantly changed during this penod.
The structural changes in Sri Lankan tea industry are unique due to the contrasting nature of these changes compared to what has been observed in North
American and European agriculture. The observed pattern in these countries is the increase in fmsizes and decrease in number of farms. in addition, an increasing trend in non-market coordination arrangements (vertical integration, contract production) has been observed in these countnes. Yet, in the Sn Lankan tea industry the number of famis has increased and average fmsize has been declined along with a move toward more market-based vertical coordination. In the next section, the major episodes of the history of the Sri Lankan tea industry is outlined paying attention to the nature of the changes that took place in the Sri Lankan tea industry. 2.3. Evolvement of Industrial Organization in Sri Lankan Tea Sector
The previous section provided a brief synopsis of the current organizational
structure of the Sn Lankan tea industry. A histcricai evolvement of the tea industry into
its current organizational structure is surnmarized in this section. The objective of this
surnmary is to discuss the historical process that determined the dimensions of farm size
distribution and vertical coordination arrangements at the beginning of the tea industry.
A vertically integrated plantation mode of industrial organization in the Sri
Lankan tea industry was established by the last decades of the 19" century and it
remained largely unchanged until about 1970s. The larger production units or plantations
in the tea industry appear to have been inherited from the coffee plantations, which
preceded the tea plantations in Sri Lanka. Larger production unit (plantations) came first,
foilowed by onsite processing rather than having the nature of processing requiring larger
production units of green leaf tea production. Since the 1970s, the plantation mode of
industrial organization has been replaced by the small individual producers in green leaf
production and independent tea processors. In chapter 4 and 5, the causes of the change
fiom this initial industry structure is explained and hypotheses are developed that are
later tested in chapter 6.
2.3.A. Origin of Plantation Production System in Sri Lanka
Tea is referred to as a plantation crop and was established in Sn Lanka as a plantation production system. The plantation entrenchment and the European colonial expansion in the tropical Asian region during the 18'h and 19Ih century were tightly linked occurrences during which the Sri Lankan tea industry was established. The term plantation agriculture connotes large-scale cultivation of crops for commercial purposes.
Though there are definitional difficulties (Higman, 1969, Courtenay, 1980, Proyr, 1982,
Tiffen and Mortimore, 1990) the term generally imply a set of characteristics.
International Encyclopedia of the Social Sciences (vol. 12) defines plantation as:
"A plantation is an economic unit producing agricultural comrnodities (field crop or horticultural
products but not livestock) for sale and ernploying a retatively large number of unskilled laborers
whose activities are closely supervised. Plantations usually employ a year round Iabor crew of
some size, and they usually specialize in the production of only one or two marketable products.
They differ fiom other kinds of fanns in the way in which the factors of production, primarily
management and labor, are combiner (p. t 54).
Why was Sri Lankan tea production organized as plantations fiom its inception in the 19Ih century? Tea plantation was the successor to the coffee plantation in Sri Lankan plantation agriculture. The production structure intmduced by coffee plantations was later taiIored to the tea plantations (Snodgrass, 1966, Bandarage, 1982). Forrest (1 965) wrote that "tea estates of today are, in fact, the coffee estates of 200 years ago, re- deployed, extended and growing a different crop" (p.29). Tt is acknowledged that al1 the characteristics, which remained as the basis to the plantation system, first appeared in the
1gth century coffee industry (Snodgrass, 1966, Jogaratnam, 1964, Bandarage 1982). Tea, rubber or coconut that were introduced later as plantation crops "did not fundamentally alter the structure of this system"(Snodgrass, 1966, p.2 1).
However, plantation coffee was introduced to Sri Lanka in the 1830s amidst a highly successful small holder coffee cultivation sector @nesen, 1953). Small holder coffee growing was a successful commercial agricultural enterprise not only in early decades of the 19~century but also in the middle and last decades of the lgthcentury, when the plantation coffee dorninated coffee production in Sri Lanka. What had caused the initiation of coffee growing in plantations, which eventually had dominated the small holding coffee production in Sn Lanka? Without explaining the factors that gave rise to plantation coffee production, one cannot explain the initial organizational structure of the tea industry, which persisted until about 1970s. in the following discussion, it is established that the choice of vertically integrated production of tea in plantations was an historical outcome rather than a choice due to economic and technical superiority of plantation green leaf production or vertically integrated processing.
2.3.B. Origin of larger holdings in Sri Lanka with Coffee plantation
Coffee was cultivated as a small holder crop on gardens surrounding the peasant's homes and along the roadsides before the plantation coffee sector initiated in the later part of 1820s (Dreisen, 1953). The peasant coffee exports were about 125,000 pounds per year in the 1800s and this grew to about 216,000 pounds per year within next 15 years (Bandarage, 1982, p.70). Within the ten years from 1815, peasant coffee exports doubled fiom about 540,000 to 1,214,000 pounds (Dreisen, 1953). The success of the homestead coffee was due to steadily improving demand in Europe and the improved prices throughout the lgthcentury @riesen, 1953). It is noted that the improved real incomes with the wealth generated by lgthcentury industrialization in Britain and north-
West Europe increased the demand for tropical food products such as coffee, tea, and sugarl, which were earlier considered as semi luxunes (Courtenay, 1980). Annual per capita demand for coffee in Britain grew fiom 0.6 pounds in 1801 to 11 pounds by 1841 giving a "big boost to coffee production in the colonies" mandarage, 1982, p. 72). The increasing demand for Sn Lankan coffee is reflected in the upward trend of pnces in
London Market. From 1834 to 1842 coffee prices increased by 300 per cent (about 15 sheiiing per bushel to 45 shelling per bushel) for Sri Lankan coffee @nesen, 1953, p.
42). A further 100 per cent pnce increased is noted for Sn Lankan coffee between 1870 and 1880 (Dreisen, 1953, p.164). The increased demand had increased the profits fiom coffee cultivation and it is asserted that "profits were about 25 per cent of sales"
(Snodgrass, 1966 p. 27).
The British who had overthrown the Dutch in the maritime area of Sn Lanka by
1796, gradually advanced inside of the country and by 18 15 the British ruled the entire country. The increasing pnces for coffee and increasing demand in European countnes made it clear for the Bntish that coffee cultivation was lucrative. Moreover, prospects of expansion in small holder coffee production to meet the increasing demand were bleak.
The resources for a large increase in coffee production were not available to the peasants.
Their labor supply was constrained by the subsistence type of agriculture, which was main1y rice based (Bandarage, 1982). More important1 y, the communal property right system of land, which usually obviates land grabbing by any individual, was not conducive tor peasant to expand their coffee cultivation area (Meyer, 1992).
1 In 1850s, average annual per capita consumption of tea in United Kingdom was 1.86 pounds. This grew up by three times (4.43 pounds) by 1870. Similarly, annual per capita sugar consumption increased by about 135 per cent in the same period (Courtenay, 1980, p.66). In responding to the profitable opportunities of coffee production rendered by the
above circumstances, British entrepreneurs faced two obstacles. First was the land for
coffee cultivation. Second was the adequate and continuous supply of labor. As in other
b'colonies of exploitation" the British did not encounter open resource situation for land in
Sri Lanka (De Silva, 1982). However, there were "communally owned" large tracks of
unused land available in the early 19" century in Sri Lanka. Large tracks of land for the coffee cultivation were acquired with the help of Crown Land Encroachment Ordinance,
No 12 of 1840. This bill allowed the colonial state to lay claims to al1 uncultivated and unoccupied land unless the natives were able to prove ownership by the stringent cnteria the state laid down (Bandarage, 1982). Since the pre-colonial land tenure is one of communal village property, particularly in Kandyan kingdom, proofs of ownership were not in existence. Therefore, large volurnes of land ended up being "Crown" land under the jurisdiction of colonial state (Bandarage, 1982). These lands were not pennanently occupied and predominantly forest land (Jogaratnam, 1964). However, as the coffee industry expanded, partially used grazing lands, lands traditionally claimed to be assigned to scared temples and other religious establishments were expropriated. Meyer (1992) noted that before 1850 the planters had a marked preference for virgin forest land which they considered more fertile, the expanding market for coffee resulting fkom upward economic trends in Europe "led them to look for any accessible land" (p.326).
It was reported that about 260,000 acres of crown land were sold during the penod of 1833-43. This period of the early 1840s, coined as " coffee mania"
(Jogaratnarn, 1964 p.98), "great rash of CO ffee planting" (Courtenay, 1980 p.67), "sudden rise of the coffee industry" (Snodgrass, 1966 p. 17). One historia noted that land sales between 1833 to 1889 were about ten per cent of the surface area of the island or about
1.1 million acres (Bandarage, 1982, p. 7, and Jogaratnarn, 1964 p.98). These lands were mainly sold to British planters for nominal prices (50 cents per acre) and almost al1 of these lands ended up as coffee plantations (Snodgrass, 1966, Bandarage, 1982).
Snodgrass (1966) noted that the average sale to a European between 1833 and 1860 was
97 acres and this cm be taken as a first approximation of "the size of the average coffee plantation" (p.23). Another researcher (McCune, 1949 p.229) noted that by 1857 there were 404 plantations with an area of 86,950 acres indicating an average size of 200 acres for a coffee plantation.
The land development with initial jungle clearing involved substantial investments (Fay, 1936). The jungle wood was Felled, dried, and bumt. The road networks inside the estates were developed, the lands were drained, and then coffee was planted. Al1 this required a penod of about six-months with a substantial labor force. It has been reported that in 1844 about 3000 sterling pounds were invested in setting up an average sized coffee estate (Snodgrass, 1966 p.26). In the boom of plantation coffee, during the period of 1841-47, it is reported that the annual investment of British capital in coffee cultivation exceeded 100,000 sterling pounds (Driesen, 1953). The majority of such investrnents were fiom individuals rather than corporations (Snodgrass, 1966). A considerable amount of these investments was allocated to convert the virgin forest and uncultivated lands to cultivable standard (Forrest, 1967 p. 30). The "larger holdings" characteristic of the plantation industry in Sri Lanka came into being with this expropriation of communally owned land by British planters. In India, Brazil, West
Indies, and many other sirnilar tropical colonies the availability of land was the impetus behind the initiation of Iarger operation units in the colonial penod (Beckford, 1972,
Courtenay, 1980, De Silva, 1982). Hayami (1994) summed up the conditions that lead to plantation establishment as follows;
Given these historical conditions, that is, the continuing rise in the dernand for tropical
cornrnodities by industrialized nations and the urgency of opening new lands to reap the gains
fr-orn the production and marketing of high-valued crops, the establishment and the development
of plantations in areas where labor and social capital were deficient (or nonexistent) were therefore
inevitable. Viewed from this perspective, it follows that the plantation system evolved not
because it was generally a more efficient mode of productive organization than the peasant mode
but because it was the rnost effective type of agricultural organization for extracting the econornic
benefits accruing from the exploitation of sparsety populated vùgin areas" (p. 128).
The considerable investments on land to bring up it to cultivable status had locked the British capital in land. The returns to the large-scale, land-based investment were dependent on the returns fiom the land-based activities such as cultivation of coffee. The capital investments that brought the land to cultivable status necessitated the early entrepreneurs to continue on land-based production to recoup this investment. This historical fact is important in analyzing the initial industrial organization of coffee (as well as tea) production in Sri Lanka and also important to understand why a plantation vertically integrate backwards into production of the raw material. The processing requirement in coffee was not a constraint that determined the size of the production unit.
The simple nature of primary processing had been one reason for the coexistence of small holder coffee production with plantation coffee production. Wickizer (1960) noted this situation about Brazil coffee production; "Historically, the modem coffee industry developed on a plantation basis in Brazil, where
originally there was ample land and labor for large scale exploitation, and a systern of land holding
that favored Iarge-scale operatiom. Small holders and plantations are equally equipped to care for
the crop under present conditions of technology, and processing of the harvested chemes does not
present insurmountable obstacle" (p. 86).
The primary processing requirement for coffee berry was insignificant. At the producer level, coffee processing is confined to the removal of skin or cherry to take out the coffee beans (Fay, 1936). When the skin has been removed the raw beans were sent in parchment fonn to curers and finally it is roasted and ground overseas. The limited amount of pnm~processing was not the main value generation activity in the coffee plantations. Therefore, the decisions about whether to produce raw material by themselves (vertically integrate) or to purchase raw material fiom others (vertically de- integrated) for this limited primary processing was not an economically important driver of selecting plantation mode of production. These plantations were pnmady large-scale, land-based raw material producers (whether it is coffee beans or black tea in bulk fonn).
To recoup the investment on land, the wealth generation had to be on iand-based raw material production, not the primary processing. The value addition by processing has been always with the overseas secondary processors. The processing requirement was not a major determinant of plantation coffee production. The primary processing was done only up to the level which allowed the buik produce to be shipped to the major processing activities overseas (Wickizer, 1960).
Therefore, the arguments of expensive processing equipment that confer scale economies such that only the larger producers can profitably use such equipment to achieve lower average costs was not a important determinant of plantation coffee production in Sri Lanka. The size determination of raw material producing wit in coffee was not based on the requirement of larger quantities of raw material for running a processing facility at full capacity to recoup the large fixed cost of the processing equipment. The relationship between the size of the unit of raw material production and the nature of vertical coordination between raw material production and processing was irrelevant in plantation establishment in coffee production.
In contrast, Shlomowitz (1984) argued that the development of large, sophisticated sugar milling was able to confer substantial economies of scale in sugar milling leading to lower processing costs relative to traditional smaller mills. The raw matenal requirement for ru~ingsuch Iarge mills necessitated large production areas of sugar cane. The investment in large sugar mills were always had been conditional on procuring adequate sugar canes to carryout the processing at kll capacity of the mil1
(Shlomowitz, 1984 p.8). However, large operations of coffee at the begiming in Sn
Lanka were not driven by such considerations. Primary processing of coffee was unlikely to have economies of scale, which could have favored the plantation coffee production. The heyday of the coffee industry is noted during the 185 1 to 1870s during which Sn Lankan coffee fetched remunerative pnces and shipped to almost al1 the important countries (Perera, 1951). By the 1870s plantation coffee production was well entrenched in Sn Lanka covering about 275,000 acres, which had production of three times more than the small holder production. 2.3.C. Establishment of Tea Plantations in British Colonial Period: 1865 to 1920s
The nse in the coffee plantations in the middle and latter decades of 19thcentury came to a sudden end and the country's coffee cultivation was wiped out by coffee rut fùngus, which first appeared in 1868 (Pim, 1946, Jogaratnam, 1964). The coffee crop was devastated by the "CO ffee mst fiingus" (Hemileia vastatrix) which ravaged through the country's coffee cultivation in the 1880s and at the turn of the 2othcentury only about
4000 acres of coffee was left in Sri Lanka (Peebles, 1982 p. 126). The British coffee planters expenmented with alternative crops such as Liberian coffee and cinchona and after a chaotic period of trial and error, tea proved to be the best choice with its steady prices in the European markets and its success in adapting to a plantation mode of production (Forrest, 1967 and Wickremeratne, 1972).
The production structure introduced by CO ffee plantations was tailored to the tea plantations (Snodgrass, 1966, Bandarage, 1982). Tea cultivation was introduced to the dying coffee plantations. The pnmary concem of the coffee planters was the additional cost of introducing a new crop and to recoup the investments they already made on their coffee plantations. For instance, descnbing the establishment of the tea industry in Sri
Lanka, Wickremeratne (1972) noted;
More germane to cost considerations was the fact that abandoned coffee lands, or lands on which
coffee was being indifferently grown, saved the planter the expenses of clearing the jungle and
gave him an income from the existing shmbs. Moreover such lands were invarïably well roaded
and drained and possessed buildings-and in some instances rnachinery too, which could be used
for the requirement of tea" (p. 135). The new crop benefited from the infiastructure of old coffee plantations, including the transportation system, the presence of plantation labor and the agricultural expertise of
Eutopean planters and the commercial and marketing arrangements built up over fifty years. Thus, tea area reached the peak of its cultivation by about 1915, just about 30 years after its initiation (see Figure 2.2).
The larger production unit for tea cultivation at its inception was more of a historical accident than a decision of vertically integrated production of green leaf in a
Iarger production unit to feed a large processing facility. The land-based production came first in tea production in the Sri Lankan tea plantation history. in order to successfully operate the large raw material producing unit, for which the foundation had been laid by the coffee cultivation, tea plantations needed to have its own processing facility. Thus, the vertical coordination arrangement of green leaf production and black tea processing in the initial period of tea production was historically detemined. In the initial period of the tea industry, establishing a processing facility without having a raw material producing unit was not an economic proposition; Fay (1936) noted "normally the position still is one estate, one tea factory but not always" (p.634). It is important to appraise this historical bias towards the vertically integrated tea production when one evaluates the causes of changes in the vertical coordination arrangement and fmsize distribution changes. Farm size for the tea industry at its inception was historically detemined and vertical integrated production was determined largely due to this large farm size.
Tea processing facilities were available in varying capacities and except for relatively smaller plantations, the size of the plantation was not insurmountable constraint to profitable operation of processing facility (De Silva, 1982). In answering the question
of what is the optimum size of tea plantation, Fay (1936) noted that in India and Sn
Lanka it was determined historically by the capacity of the individual planter in pre-
motor days to finance and supervise the development of the estate, its cultivation and
working and the treatment of its "product in the factory on the estate" (p.633). However,
there are increasing returns at the level of processing and in house green leaf production,
which were not significant in coffee cultivation. The need for closer coordination
between green leaf production and processing ideally suited the larger production unit,
which has been already introduced by coffee cultivation. Due to the capital intensive-
processing requirement in tea, the small holder tea producers were worse off relative to
small holder coffee producers.
The majority of tea plantations were established in the central hilly region where
the coffee had been growing. The central hilly region is categorized into two regions
based on altitude; the High Grown region is 1220 meter above mean sea level and the
Mid Grown region is between 610 and 1220 meters in altitude. The third region
involving tea production is the Low grown, which is from the mean sea level to 610 meters in altitude. The earliest data available about the regional distribution of tea plantations in these three elevation groups were in 1920 (see Table 2.4). About 86% of the tea land area was established in High Grown and Mid Grown regions or the central hilly region. Except for the Ratnapura district, none of the other disticts in the Low
Grown region was important during the initial stage of tea plantations in Sn Lanka. Figure 2.2. Tea Land Area in Sri Lanka from 1868 to 1998
Sources: (1) From 1867 to 1949, Marby (1972 p.28) and Snodgrass (1966 Table A-37) (for the years 1923, 1924 and 193 1, 1932 are estimations). (2) From 1950 to 1980, Athukorala (1 984 p.84). (3) Frorn 1 98 1 to 1998, Annual Reports of Central Bank of Sn Lanka ( 1990/98). Table 2.4. Regional Distribution of Tea Plantations in 1920s (acres)
Elevation Group Districts Tea Area Percentage High Grown Nuwara Eliya 106,000 24.1 1 Badulla 77,000 17.5 1 High Grown Sub Total 183,000 4 1-62 Mid Grown Kandy 160,000 36.00 Matale 20,000 4.50 Kegalle 19,000 4.32 Mid Grown Sub Total 199,000 44.82 Low Grown Colombo Kalutara Galle Matara Ratnapura 35,000 7.96 Low Grown Sub Total 57,500 13.00 Total 439500 100.00
Source: Peebles, 1982 (Sri Lanka: A Handbook of Historical Statistics, p. 1 18)
Another important feature associated with the rapid expansion in tea was the influx of south Indian laboren to work in the tea plantations. The net flow of immigrant south Indian labon was smaller during the era of coifee plantation (see Table 2.5), since labor requirements for coifee cultivation was seasonal (Bandarage, 1982). However, since the 1890s, when tea was becoming the mainstay of the plantation economy with a much higher and year round labor requirement than coifee, a large labor force imrnigrated to the country (see Table 2.5). The average number of Indian laborers (men women and children) reported as working on plantations dunng the last quarter of 1929 was about 742,300 (Sundararn, 1931). This ethnic group called the "indian Tamils" settled in the country and the majority of them (62%) still live in High Grown and Mid
Grown regions working on tea plantations. They composed about 5.55% of the country's population of 14.85 million in 1981 (Department of Census and Statistics, 198 1). Table 2.5. Net Immigration of South Indiao laborers 1879 to 1920 (in Thousands)
Year Net Immigration 1871-1875 88.6
Source: Fernando, 1980 page 103.
Plantations had to invest heavily on providing basic arnenities such as food, housing,
water, and health facilities for this residential labor force and even today such provisions
are made by the plantations at substantial costs (De Silva, 1982).
The adoption of tea into the plantations required significant capital investment.
During the coffee era fiom 1820 to 1886, private proprietonhip had been the principal
form of business organization (Ramachandran, 1963). The first investors were those
most informed about the potential profits from coffee; the civil and military officiais of
the British colonial govemment who were stationed in the country. They generated the
required capital through private arrangements such as their own savings, partnerships,
mortgaging the coffee land or fiom advances on future crops (Snodgrass, 1966, De Silva
1982, Rote, 1986).
Capital requirements were considerably higher for tea cultivation due to its need
for sophisticated processing equipment relative to the processing equipment for coffee
(Forrest, 1967). Private proprietors were no longer capable of sustaining such capital for tea cultivation. Rote (1986) noted "the acquisition of capital for investment in the more capital intensive tea industry required some fom of business organization that was less risky than an enterprising individual planter" (p.35). Consequently, joint stock companies were incorporated to raise the required capital. In 1899, there was one such joint stock company incorporated in the United Kingdom and necessary capital was raised though issuing shares (Rarnachandran, 1963 p. 186). Their number increased to
12 by 1920 and by 1959, there were about 80 such companies, where the majonty of shareholders were British. The joint stock companies incorporated in the United
Kingdom called the "Sterling Companies" and those incorporated in Sri Lanka were called the "Rupee Companies". By the 1950s, there were about 170 plantation companies whose main interest was tea. About 350 plantation companies grew tea with other plantation crops (mainly rubber).
The present giants of world tea trade such as Lipton, and Brook Bond, entered the
Sri Lankan tea industry during this penod. in 1890, Thomas Lipton, the owner of one of the largest retailing grocery chains in Britain, came to Sn Lanka to invest in tea production. In 1922, Brook Bonds bought both the tea shipping company and some estates, establishing their permanent interest in Sri Lankan tea industry. This marked the beginning of corporate ownership and management of the tea plantation sector, which was quite different fiom the coffee era when private proprietorship was the principal business organization.
Along with the British corporate ownership, management of plantations were carried out by some intermediaries popularly known as "Agency Houses"(Rote, 1986).
These Agency Houses managed the cultivation, production, and marketing of estate produce on behalf of the Companies. Agency Houses also provided the services of exporting of tea, importing of inputs, insurance, storage etc. Cost of management services
through the Agency Houses is argued to increase to the costs of production in these large
tea fanns compared to that of the owner-operated srnall tea farms. Such higher costs
were mainly in the form of management fees, which were based on the size of tea output
rather than on profits (De Silva, 1982).
The devastation of coffee cultivation by coffee rust hngus also affected small
coffee cultivators who were known as peasant coffee growers (Dreisen, 1953).
Smallholder coffee production covered approximately 50,000 acres near its peak period
(Roberts and Wickremeratne, 1973). However, peasants did not switch to tea cultivation
immediately. Despite the encouragement given by some govemment offcials in the fom of demonstration plots, free tea seed distribution etc., "native interest in tea cultivation was significantly difised" (Wickremeratne, 1972 p.139). He noted that peasants knew
little about the cultivation of tea or about the preparation of its leaf. Therefore, to peasants, this was not a "safe investrnent" (p. 141). By 1899, only about 3000 acres were grown by srnaIl tea producers (tea gardens smaller than 10 acres). This was less than
0.8% of the total tea area of 378,000 acres in 1899 (Ramachandran, 1963 p. 1). Compared to the comparative importance of the peasant coffee sector, peasant involvement in tea was insignificant dunng the initial stages of the tea industry.
2.3.D. Continuation of Plantation Cornpanier and Agency Houses: 1920 to 1948
Dunng 1920 to 1948, the tea plantation system established at the tum of the century become entrenched in the country and the Agency Houses, as the intermediaries, tightened their grip over the plantation companies in al1 spheres of the tea industry. It is said that by the time of the tea centenary (1967) the management of tea estates was largely in the hands of Agency Houses, whose services were used by most of the companies that grew and manufactured tea (Rote, 1986). Other important developments during this time that affected the tea industry were largely political.
Growing antagonism for British ownership and management of the tea industry was one of the important features in this period. This wai a result of gains in political power by the national elite movement, which agitated the public for independence from
British (De Silva, 1973). When partial autonomy in the legislature was granted under the political reform in 1931, the idea that the tea industry should be under the Sri Lankan control gained momentum. There were political demands to stop expatriation of profits from tea and to reinvest these in local development schemes (Rote, 1986).
Another important political milestone during this period was the initiation of labor unionization. The south indian plantation workers were drawn towards a political organization called Ceylon indian Congress which was established in 1939 to "defend interests of Indian Tamil community" (Kearney, 2971 p. 122). The labor union of the
Ceylon Indian Congress, which is called, Ceylon Workers Congress, was established in
1940 and today it is the largest plantation labor union in the country with about 350,000 members. The Political importance of the vote block of the plantation labor force was changed with sweeping constitutional changes that took place in 1978. These constitutional changes, especially two of the changes, made these ethnically oriented votes important in gaining political power. First was the introduction of an executive presidential system making the executive president the head of state, who is directly elected through a presidential election. Second was the changing of the parliament electoral system from a "fiat-past-the-post" system (a candidate who gets a plurality of the votes cast wins the parliament seat) to a "proportional representation system" (a party gets nurnber of parliarnent seats according to the proportion of the votes they get in a district). Both these electoral changes necessitated gaining the plantation labor votes if one were to win in the presidential election and the parliamentary election2. in general, other labor unions are not based on the ethnicity and they are divided arnong many political parties in the country. There are no other union-based political representations in the parliarnent, except for the plantation labor unions.
The peasant farmer involvement in tea cultivation gradually increased during this period. Wickremeratne (1972) noted that in the Low Grown districts, peasant cultivation increased considerably. These peasant producers usually were in physical proximity to a larger plantation to sel1 their green leaf tea, which was processed by the large plantation
(Snodgrass, 1966 p.38). In the 1!ZOs, the srnallholding tea sector ( farm less than 10 acres in size) was about 62,000 acres or about 1 1% of the total tea area of the country.
However, during this period, small tea fanris contribution to the country's total tea production was insignificant (Snodgrass, 1966).
2.3.E. British Interests in Tea Plantations after Independence: 1948 to 1975
Sri Lanka gained independence fiom the British in 1948. By then two-thirds of the tea sector was owned by sterling companies and managed by British-controlled
' A telling example of the importance of the plantation labor votes in gaining the political power of the country is that the leader of the Ceylon Workers Congress has been the Cabinet Minister since 1977 till his death in 1995, then his heir (one of his grand children) assurned this portfolio and continues to-date, despite the changes in ruling party and the president in 1994. Ceylon Workers Congress usually smkes a deal with the party, which can become the parliament rnajority with their support. This has made thern influential in resolving labor grievances, particularly in increasing the labor wages. Agency Houses in Colombo (Caspersz, 1975). The British were involved in not only in the tea sector, but in almost al1 facets of the economy such as banking, insurance, and shipping (Snodgrass, 1966). The major change in the post-independence era was the outflow of sterling companies and increased Sn Lankan ownership of the tea lands.
Uncertainty surrounded the foreign ownership of sterling companies after independence. From the conferral of independence until 1952, the legislature was not elected from a general election. The first general election of independent Sn Lanka was held in 1952 and the second in 1956. The 1956 election was a landmark in putting a strain on British business interests in the country. Nationalization of foreign-owned tea land became a major issue in this election and the govenunent came to power in 1956 accepted nationalization as a policy (Fernando, 1980).
The uncertainty for British (sterling company) ownership in this period was reflected by the reduction of British ownership in the tea sector since independence.
From 1952-53, British owned about 365,000 acres or about 64% of the total tea area
(Caspersz, 1975 p. 42). By 1959, this dropped to about 202,000 acres or 35% of total tea area and by 1972 British ownership dropped fiirther to 158, 000 acres (Fernando, 1980 p.
132). A large share of British (sterling company) owned tea estates was sold to rupee companies and the Sri Lankan ownership of plantation companies increased considerably during this period. The other important development during this penod was the perceptible increase in small tea fanners (see Table 2.6). Table 2.6. Ownership Changes in Sri Lankan Tea Land 1959 to 1972 (in acres)
Ownership Category 1959 % 1967 YO 1972 % Sterling Companies 201,905 35.4 177,676 29.2 158,147 26.5 Rupee Companies 82,742 14.5 153,651 25.7 150,889 25.2 Non-SriLankanhdividuals 55,067 9.7 15,062 2.5 10,859 1.8 Sri Lankan hdividuals 148,539 26.0 144,233 24.0 152,468 25.5 Sri Lankan and Non-Sri 8,723 1.5 4,741 0.8 4,085 0.7 Lankan Individuals State - - 2,277 0.4 13,530 2.3 Sri Lankan Small holders 73,594 12.9 101,174 16.9 107,687 18.0 Total 570,570 100.0 598,8 14 100.0 597,645 100.0 Source: Fernando, 1980 p. 132 (Table 4.3)
Although nationalization was not camed out until almost twenty years later, in
1975, the confidence for the British to continue in the Sri Lanka's tea industry was shattered afier independence. It is argued that due to the uncertainty in ownership and property rights, maintenance and care was compromised in sterling tea estates and conditions deteriorated rapidly dunng this period (Rote, 1986). The aggravating labor disputes or "emergence of a new militancy in the plantation trade union movement"
(Sinnatharnby and Devaraj, 1987 p. 223) also encouraged the British owners to sel1 their estates to Sri Lankan companies.
The basic management orientation did not change with the ownership change from sterling companies to mpee companies. Larger plantations were still managed by
Agency Houses. in 1970, the Agency Houses managed 47% of the total area planted to tea and accounted for 62% of the total tea production of the country (Fernando, 1980 p.135). These Agency Houses experienced a process of "Sri Lankanization" of management personnel. Yet, these Sri Lankans followed their foreign predecessors as far as management and persona1 life style were concemed.
Sri Lankan individuals bought a sizable extent of the sterling tea estates since independence (Rarnachandran, 1963). These were owner-manager style mal1 properties and there was a considerable fragmentation of large estates when these were sold to such individuals. Ramachandran (1963, p. 173-175) noted that many of the British sterling companies who left Sri Lanka during this period started tea plantations in Kenya at about the sarne tirne. This has been recognized as the major reason for the rapid expansion in
Kenyan tea production in the1950s and 1960s.
2.3.F. Nationalization and Land Reform: Government in Tea Industry: 1975 to 1992
The nationalization of the tea industry finally took place in the early 1970s with the land refonn legislation which was one of the most influential agricultural policies that affected the ownership of plantations since its inception (Peins, 1984). The Land Refonn
Law, No 1 of 1972 imposed a ceiling on the private (excluding public companies) ownership of land (25 acres for paddy and 50 acres for other categories of land). Three years later, the scope of the reform was extended through the Land Reform (Amendment)
Law, No 39 of 1975 and with this land and related capital assets held by public companies engaged in agriculture were nationalized.
The Land Reform Commission expropriated a total of 98 1,357~acres with the
1972 and 1975 enactment (see Table 2.7). This was about 23% of the total land area under permanent agriculture in the country. The tea area alone constituted about 38% of
' . Detail account of land reform activities of Sri Lanka is given in Peins, 1978 and Sadaratma, 1972. the total land acquisition. About 63% of the country's total tea axa ended up in the jurisdiction of the Land Reform Commission by 1975.
Table 2.7. Land expropriated by Land Reform Commission: by crop type (acres)
Crop 1972 Law % 1975 amendment % Total % Tea 139,354 24.7 237,592 56.8 376,946 38.4 Coconut 1 12,523 20.0 6,406 1.5 118,929 12.1 Rubber 82,563 14.7 94,835 22.7 177,398 18.1 Sub Total 334,440 59.4 338,833 81.0 673,273 68.6 Jungle & uncultivated 176,505 3 1.3 Other 52,455 9.3 79,124 19.0 Sub total 228,960 40.6 79,124 100 308,804 31.4 Grand Total 563,400 100 417,957 981,357 100 Source: modified fiom Fernando, 1980 p.208
About 95% of land acquired fiom company-owned large tea plantations were vested to two statutory bodies: People's (Janatha) Estate Development Board and Sn
Lanka State Plantations Corporation. Both of these state corporations followed the same plantation management system of the former owners (companies), keeping the larger estate character intact in the post-reform period (Fernando, 1980). Yet, a few of the tea estates were put into different management experiments such as collective fming, and
CO-operative faming which later reverted back to the purview of the above two corporations (Simathamby and Devaraj, 1987). The state involvement in tea plantations was considered as "substitution of public ownership for the different types of ownership that had previously existed" (Si~athambyand Devaraj, 1987 p.228). The management of these large state corporations was plagued with contiision fiom its initial stage. The central coordination and management of 45,000 hectares of plantation land
(both rubber and tea) which covered 650 estates dong with 440,000 residential labor force was not a simple feat. There was a lack of planning and understanding as to who would be entrusted to manage these properties once these were acquired (Fernando,
1980). The inunediate post-land reform institutions in the plantation sector were introduced in haste without adequate thought to their workability (Simatharnby and
Devaraj, 1987). The complications in establishment of post-reform institutional arrangements confounded with the management difficulties deteriorated both the output and the profitability in the tea estates managed by the state corporations. Dunng the 15 years of the post-reform period, tea production in state-owned estates declined by about
19%~,while the small (private) tea fams registered an irnpressive growth (see Table 2.8).
Profits for the state-owned estates declined and dunng 1980 to 1991 there were five years in which the profit margins were negative (Plantation Sector Statistical Pocket
Book 1999 p. 136, 137). The cost of black tea production before the nationalization period (average is calculated for the five years 1972 to 1976) was about Rs/kg 3.00 (1952 constant). This had increased to Rs. 5.86 (1952 constant) during the last five years of state ownership (average is calculated for 1986 to 1991); an almost two fold increase
(Plantation sector statistical Pocket Book 1999 p. 136).
The crisis in central coordination of the management of large tea farms led to bringing the two corporations under two separate Ministries, which were under the direct control of the President of Sn Lanka.
'The average annual production of state-owned estates during the five years of 1976-1980 was about 18 1 million kilos. This dropped to 146 million kilos ( 19%) last five years (1 986-1 99 1) of state ownership. Table 2.8 Post-Nationalization Tea Production in Sri Lanka (Kg Mn)
Year State-Owned Estates Regional private. Total Other Plantation Sector EDB' SLSPC~~~~C~Total cornPanies& (small fms) 1976 188.30 O 7 -64
Sources: Annual Report Central Bank of Sri Lanka, 1983, Plantation Sector Statistical Pocket Book, 2000 (p.23)
'JEDB:Janatha (People's) Estate Development Board, # SLSPC: Sn Lanka State Plantations Corporations, w Private: Tea fms less than 50 ac in size (values for 1976 to 1979 is an estimate) 4 Regional Plantation Companies are the re-privatized state-owned tea fmsfrom 1992 The management difficulties, political interference, shear sizes of these corporations, lack of incentives for the sound management decisions were noted as the main causes for the failures of state ownership and management of the nationalized plantations (Rote, 1986,
Simatharnby, 1993, Bandaranaiyake, 1993, Ranasighe, 1995). World Bank (1 997) surnrned up this situation "Between 1972 to 1992, misdirected sectoral planning and policies, combined with total governrnent control over production and marketing of tree crops, left the sub-sector completely without direction for firture growth and international competitiveness" (p.52). The budgetary allocations to the two ministries of plantation sector increased Fom Rs 87.67 million in 1981 to Rs 1683.7 million in 1991 (nominal) and there was increasing displeasure that these large ministries put an unbearable strain on state coffers (Kelegama, 1991). It was increasingly becoming clear that state management could not be prolonged and privatization of the plantation sector gained momentum in the 1990s.
2.3.G. Privatization and end of the State-Managed Tea Industry: 1992 to Present
The Plantations Restructuring Unit was assigned to study how to revitalize the Sri
Lankan plantation industry in late the 1980s. The recomrnendation given by the
Plantation Restmctunng Unit was that management of the state-owned plantation had to be entmsted to the private sector (Ranasinghe, 1995). This proposa1 was finally put into action in two phases: first handing over the management of the state-owned plantation to private management companies and then transferring the hl1 ownership of these large tea fanns to the private sector (Liyanage, 1993). The first phase started in 1992 by setting up 22 state-owned regional joint stock companies, each acquinng about 20-25 estates (majority of the companies had both tea and mbber) fiom the 450 state-owned estates. Then the state-owned companies were contracted out to 22 local private companies (many of them were former agency houses) to manage the estates on behalf of the state under a profit sharing arrangement. The agreement between the govemment and these management companies was a 5-year management lease. The 22 regional companies were fully owned by the state and the
managing companies were merely their operating agents.
There were some accusations that private management would resort to short-terrn profit increasing practices such as exploitative harvesting, restricting fertilizer application, at the cost of long terni viability of these tea estates with the shorter period of the management lease (People's Bank of Sri Lanka, 1995). The full divestiture of ownenhip to the private sector was needed to provide incentive for private management to undertake the long tem capital development and investments (World Bank, 1997). in
1995, the govemment started privatizing the ownership of state-owned regional companies by selling 51 per cent of the shares though open bids. By May 1997, ownership of 14 companies was sold and the stocks of six of these were trading on the
Colombo stock exchange. The shares of the other regional companies were sold during
1st few years and presently most of these state-owned plantation companies have become public liability companies. There were improvements in the plantation management with the pnvatization and particularly in lowering the cost of production.
Under privatized management, the average cost of production of black tea came down
from Rs 6.47 (1952 constant) in 1992 to Rs 4.84 in 1998 (Plantation Sector Statistical Pocket Book, 1999). Black tea production of the previously state-owned tea fmsunder
private management improved fiom about 131.7 million kilos in 1993 to 140.7 by 1998
(see Table 2.8).
During the chaotic period of state ownership and management of the large tea
fms, privately held small tea fmsmade impressive progress. Their numbers as well
as the area of cultivation went up year aAer year, particularly in the 1980s and 1990s. in
1998, small fmsproduced close to a half of the national output (see Table 2.8) and they
are enjoying better yields compared to the large fams sector (Central Bank of Sri Lanka
Annual Report 1998 p. 35). The area under the tea small fanns increased by 52% from
1978 to 1994 while the area planted to tea under large farrns dropped by 46% (see Table
2.9). The Low Grown region, where the tea industry had not traditionally played an
important role has become the most important tea producing region since the 1980s. in
1994, this region occupied about 37% of total tea area of the country (see Table 2.9). The
principal contributor in this development is the small fmerand the tea land area of the
small fmsin the Low Grown region increased by 156%.
2.4. Structural Changes In Sri Lankan Tea Industry
During the historical episodes discussed in section 2.3, the relative importance of
the components shown in the Figure 2.2 has changed reflecting the fundamental changes
in the industry structure. These are identified as structural changes and in this section, the two most important structural changes are described. Table 2.9. Regional Level Tea Farm Size Distribution Changes: 1978 to 1994 (Figures are in hectares) Elevation Class Small Farms Large Farms 1978 1994 % Change 1978 1994 % Change High Grown Nuwara Eliya 3,276 6,303 + 92 39,671 44,153 +Il Badulla 4,408 6,244 + 41 33,686 23,507 - 30 Sub Total 7684 12547 + 63 73357 67660 - 8 Mid Grown Kandy 18,127 9,733 - 46 60,600 13,658 - 77 Matale 847 1,235 + 45 7,008 2,970 - 57 Kegalle 3,689 2,128 - 42 8,789 3,333 - 62 Sub Total 22,663 13,096 - 42 76397 19,961 - 74 Low Grown Ratnapura 5,750 15,667 + 172 23,556 10,453 - 56 Galle 7,O 11 17,080 + 143 7,304 2,205 - 70 Matara 5,692 15,868 + 178 9,604 2,694 - 72 Kalutara 1301 1962 + 50 3,825 1470 - 61 Sub Total 19,754 50,577 + 156 44,289 16,822 - 62 Grand Total 50,101 76,220 +52 194,043 104,443 -46
Sources: 1978: Report of the Cornmittee Appointed to Study the Constraint and Problems of the Tea Small Holder Sector, 1980. 1994: Plantation Sector Statistical Pocket Book, 1999. Notes: in 1978 small farrns were defined as farms less than 10 acres in size. In 1994 small farms were defined as fanns less than 50 acres in size.
2.4.A. Changes in Farm size Distribution of the Sri Lankan Tea lndustry
The tea censuses have classified tea fanns into four different size classes: larger than 500 acres, between 100 and 500 acres, between 10 and 100 acres, and smaller than
10 acres. The first three size classes are traditionally called estates or large tea farms and the fourth size class is referred to as small holders. However, since 1978, with the introduction of the land reform law, small holders have been categorized as tea farms less than 50 acres in size.
The changes in farm size distribution are in sharp contrast to the North American and European pattern (see Table 2.10 and Figure 2.4). First, the total number of tea fms in Sri Lanka has increased by about 138% (86,842 in 195 1 to 207,056 in 1995). The number of small fms(< 10 acres) has gone up by 143% while the land area under this size class has gone up by 136% (67,414 acres in 1951 to 159, 470 acres in 1995). The number of medium size hs(10-400 acres) has also increased by about 95% from
1951 to 198 1, while the land area under them has doubled from about 43,000 acres to
86,000 acres. Secondly, the tea fms in the two largest size classes have declined in number by about 10% from 195 1 to 198 1, and in terms of land area planted to tea in these size classes have declined by about 15%. There is no information about these four size classes since 198 1. Thus, relative changes of these four size classes during 1st 2 decades are not known. Moreover, in 1978 the definition for small fmswas changed to farms less than 50 acres in size. This was the ceiling for the pnvate land ownership, which was enacted with the land reform laws in 1972. Under this definition, by 1995, small fams have represented about 44% of the country's total tea area (see Table 2.10). ignoring the definitional changes of small fams, the area under small tea fmshave increased from
12% of total tea area in 1951 to 44% of total tea area in 1995 (in absolute tenns fiom 67,
414 acres in 1959 to 204,900 acres in 1995). Yet again, ignoring the definitional changes of large tea farms, the area under large tea fmshas declined from about 500,000 acres in 1951 to about 262,000 acres by 1995 (see Table 2.10). The conceptual mode1 to explain the farm size distribution changes are discussed in chapter 4. Table 2.10. Sue Distribution of Sri Lankan Tea Farms (Number & Area)
(Percentages are in parentheses)
Year Nurnber of Tea Farrns Arca of the Tea Fms (acres) c10ac 10-cl00 100-~500~500 Total c10ac 10-cl00 100-~500 >500 Total 1951 84363 1540 593 346 86842 67414 43128 168127 288619 567288
< 10 acres > 10 acres Total < 10 acres > 10 acres Total
(99.02) (0.98) ( 100.00) (34.1 5) (65.85) ( 100.00) < 50 acres > 50 acres Total c 50 acres > 50 acres 1995 206652 404 207056 204893 262042 466935
Sources: 2959: Snodgrass, 1966, Table A 38 1968: Ceylon YearBooks Caspersz, 1975 p.4 1, Table II Report of the Cornmittee Appointed to Study the Constraints and Problems of the Tea Small Holding Sector, 1980 Fernando, 1984 Report of the Presidential Commission on the Tea Industry and Trade, 1995 p.69 Sriyaratne, 1994 p. 1 15. Figure 2.3. Farm Size Distribution Changes in the Sri Lankan Tea Industry 2A.B. Changes in Vertical Coordination arrangements in Sri Lankan Tea Industry
Vertical coordination in green leaf production and black tea processing is another important sîructural dimension that has significantly changed during he above historical episodes. In the 1960s and 1970s, when the largest farm sizes dominated the tea production, green leaf production and black tea processing were vertically integrated under the same ownership and management. As discussed in the section 2.3, vertically integrated coordination in the early penod of tea industry was largely historically detemined mode of production. However, in the 1980s and 1990s, when green leaf production of the small famis expanded, the market activity becarne an important coordination process between the black tea processors and small green leaf tea producers.
The changes in vertical coordination of green leaf producers and the black tea processors are summarized in Table 2.1 1. Black tea processots are classified into four groups depending on the relative size of their green leaf purchase. The first group buys no green leaf for their processing needs. These are tùlly integrated estates or larger tea fams. The second consists of large estate processors who buy less than 50% of their need (the published data categorized processors in this manner, see the source in Table
2.1 1). The third group is comprised of large estate processors whose bought leaf share exceeds 50% of their need. The fourth and final group is made up of independent processors who buy their entire green leaf supply. The proportion of large estate processors who purchase green leaf has gone up fiom 39% in 1973 (324 of 816) to 48% in 1993 (286 of 592). The number of processors who use only their own green leaf has dropped fiom 429 in 1973 to 1 17 in 1993. The processors who completely outsource their green leaf requirements has gone up €rom 63 to 184 or more than three times during this penod. in 1973, about 19.3% of al1 processed black tea was based on purchased green leaf. By 1993, this percentage had risen to 52% (Report of the Presidential
Commission on the Tea hdustry and Trade, 1995).
Table 2.1 1. Changes in Green Leaf Purchase by Processors (no. of processoa)
% of Green leaf High Grown Mid Grown Low Grown Total Purchased 1973 1993 %A 1973 1993 %A 1973 1993 %A 1973 1993 %A
100% 1 3 +200 29 33 +13 33 148 +348 63 184 +192 % of Processors buying green leaf 15.7 43 -6 54.1 83.7 68.0 98.0 47.4 79.3
- -
(Source: Report of the Presidential Commission on the tea industry and trade 1995 p.80)
Along with the increase in number of processors who have increased the outsourcing of green leaf requirement, black tea production from purchased green leaf has substantially gone up. Table 2.12 shows the amount of black tea processed from vertically integrated green leaf production and Erom market-procured green leaf. From the total black tea production of 195.4 million kg in 1980, 73% was produced in vertically integrated units. By 1998, this share has dropped to 45%. During this period, the amount of black tea processed from purchased green leaf by the large farms increased slightly fiom 20.45 million kilos in 1980 to 33.17 million kilos in 1998. However, the independent processors or the processors who fully outsource their green leaf requirernents have tripled their production fkom 52.65 million kilos in 1980 to 157.17 million kilos in 1998 (see Table 2.12). The rnove f?om vertically integrated tea plantation to separate green leaf growers
(that are increasingly small holders) and independent processors in general is opposite to the trends in vertical coordination arrangements across the developed world agriculture.
The causative factors for the vertical coordination changes in the Sri Lankan tea industry is analyzed in the Chapter 5.
Table 2.12 Black Tea Production from Purcbased & Self-grown Green Leaf (Mn kg)
From vertically integrated green leai 1 From market-procured green leaf Total (10+11) Year JEDB" SLSPC. RPC* Other Total JEDB SLSPC RPC Total Independent S tate Processors Agencies 1 2 3 4 5 6 7 8 9 10 11 12 142.731 7.00 13.45 O 20.45 32.20 52.65
I Source: Plantation Sector Statistical Pocket Book 1999; Columns 2,3,4,7,8, and 9 f'iom Table 5.23 (p.139), and column 5 and 11 from Table 2.5 (p.24).
'JEDB (Janatha (People's) Estates Development Board), 'SLSPC (Sri Lanka State Plantation Corporation), 'RPC (Regional Plantation Companies). 2.5. Summary
The structure of the Sri Lankan tea industry has changed considerably over tirne.
From its inception in the mid 1800s to the early 1970s, large estates that grew the green
leaf tea and processed it into black tea dominated the sector. Since then, farm size
distribution has changed. Small farm share in total tea land area has considerably
increased and the green leaf production and black tea processing sectors have become
increasingly separated. independent and small fmproducers now grow more of the
green leaf tea and independent processors purchase it for processing into black tea. The
break of the vertically integrated plantation structure within the Sri Lankan tea sector is
particularly evident at the iowest elevation region.
The structura1 changes in Sri Lankan tea industry pose an interesting puzzle, for
these changes are counter to the increasing vertical coordination witnessed in the agri-
food sector elsewhere (H~M~ssY,1996). Why does the large tea farm loss its importance relative to small tea fms? Why does the vertically integrated organizational structure replace by market-based vertical coordination? The next chapter reviews the literature and summarizes the general economic forces found to affect the farm size distribution and vertical coordination changes. The following chapters apply these general concepts to the Sn Lankan tea sector and develop a conceptual mode1 and associated hypotheses to explain changes in tea farm size distribution (Chapter 4) and in vertical coordination relationships (Chapter 5). CHAPTER 3
REVIEW OF ECONOMIC ANALYSES OF STRUCTURAL CHANGES IN
AGRICULTURE
3.0. Introduction
The purpose of this chapter is to assess the literature related to the economic analysis of structural changes in agriculture. The objective of this assessment is to identim and characterize the important drivers (or factors) that shape the structural changes in agriculture. These dnvers will be used to develop a conceptual fiamework to explain structural changes that have occurred in the Sri Lankan tea industry. The assessment is focused on the farm size distribution and vertical coordination arrangements, since these have been the most important structural dimensions that substantially changed in the Sn Lankan tea industry.
This chapter is organized as follows; Section 3.1 synthesizes and describes the important dimensions of fann industry structure and the measurements of the changes in these dimensions. The comparative strength and weakness of the various structural dimensions in characterizing structural changes in agriculture are noted. Section 3.2 synthesizes the literature about the important economic forces or drivers that shape the different dimensions of farming industry structure. Subsections under Section 3.2 review the major theoretical models of farm size distribution such as the technology models, the human capital model, and the financial model. Section 3.3 reviews the major theoretical models of vertical coordination arrangements such as the structure-conduct-performance model, the principal-agent model, and the transaction cost approach. Section 3.4 summarizes the main points discussed in the chapter. 3.1 Conceptual Exposition of Farm Structure: Dimensions and Measurements
How do agrïcultural economists characterize the structure of the fming industry? Structure has been measured by dimensions such as the total number of farms, the average farm size, the population living on fms, and tenure arrangements. Table 3.1 presents a synthesis of fmstructural dimensions fkom previous studies. These dimensions can be categorized as either focusing on the component or the relationship characteristics of the fming industry structure.
The "cornponent" dimensions of fami structure describe the composition of farm industry structure. Such dimensions include size distribution of fms, level of specialization, tenure/ownership distribution, and types of business organization. The
"relationship" dimensions describe the coordination and exchange arrangements of the inputs and outputs of the fming industry. These are inter- and intra-sector linkages such as contract production, vertical or horizontal integntion, and off-farm involvement of farm people.
At a given point in time, the number of farms under each category of these dimensions would characterize farm structure. in this sense, farm structure is a static concept. Structural change is the change in the farm numbers within the categories of each dimension. For instance, under the size distribution dimension, the number of small farms is decreasing while number of the large farms is increasing over time in developed world agriculture. In the vertical linkage dimension, the number of fmsdepending on the market for input procurement and output marketing is decreasing while nurnber of fmsentering contract production and complete integration is increasing in agriculture of the developed world. Table 3.1. Structural Dimensions of Farming Industry: a Synthesis
Dimension Categorùes of farms usually noted in the Iiterature
~om~onentDl'mensions Farm Size Distribution Small Medium Large Very large
Level of specialization/ No of enterprises No of enterprises No of enterprises (< 10) No of enterprises Diversification (< 2) (< 5) ('10)
Type of Business Family or Parmership Corporations Cooperatives and Organization Individual (fdlyheld or non- uts, family held) rnunicipalities Relationshrj, Dimensiorts Tenure (Ownership) Full owner Part owner Operator Tenants Share Farming distribution Operator
Vertical Iinkages with Completely Production contracts Ownership contract Completely inter and intra sectors market mediated integrated
Extent of off-farm Full time farming 1-99 days per year 100- 199 days per year 2 200 days per Emplo yment off-fm employment off-farm employment year
Sources: Boehlje 1999, 1992, Goddard et al 1993, Reining 1990, Rodgers 1988, Tweeten 1984, Sumner, 1985, Brinkman and Warley 1983, Tweeten and Huffman 1980, Babb 1979
3.l.A. Importance of Farm Sue Distribution as a Measure of Farm Structure
Usually farm size is measured by gross sales but it can also measured by acres,
assets, total input value, value added, and equity (Tweeten, 1984). Throughout
agriculture in developed countries, the trend has been toward increasing average fm
size (Monke, Langwothy and Avillez 1993, Reining 1990, and Bnnkman and Warley
1983). increasing fmsize was an important phenornenon related to the changes in
fming industry structure of the United States in the 1950s to 1970s (Stanton, 1989).
This has been the bench mark measure of structural change as evidenced by the quote
that ''structural change in US. agriculture refers to change in the distribution of fms
sizes" (as measured by cash receipts by sales class) (Reining 1989, p.39). Thus, examining changes in size distribution remains the "most important evidence to
evaluate9'(Stanton 1989, p.16). It is also related to the larger question of the demise of
fmily farms. Concentration of a larger share of farm production in increasingly fewer
fmshas also raised the concem of these large firms possibly exerting market power and
influencing fami prices. Together, these issues have made farm size distribution the most
important dimension in economic analysis of the structural changes.
3.1.B. Inadequacies of Farm Size Distribution as a Measure of Structural Changes
There are questions over whether the change in fmsize distribution is the most
appropriate or representative dimension of structural change. Some of the concerns
surround the impreciseness of the measurement of farm size. Gardner and Pope (1978)
argue that when the yardstick is imprecise the measurements of farm size distribution and
its changes are erroneous. Others (Le., Boehlje, 1999) concerned about the inability of
the size distribution to capture the changes in other important structural dimensions.
The argument dealing with the measurement precision of fami size begins with
the definition of a fm. The U.S. Bureau of Census defined a fm; "any place fiom
which $ 1,000 or more of agricultural products were produced and sold or nomally
would have been sold dunng the census year" (1992)'. Thus, some activities, which
generate marketable agricultural output are too small to be regarded as legitimate
commercial enterprises.
I Statistics Canada defines census farrn as follows: "Census farm refers to a farm, ranch or other agricultural holdings which produces at least one of the following products intended for sale: crops, livestock, poultry, animal products, green house or nursery products, mushroorns, sod, honey, or maple symp products. Census farms are also commonly referred to as farm operations or agricultural holdings" (p. 37 1 ), Census Overview of Canadian Agriculture: 197 1 - 199 1, Statistics Canada. On the other hand, at the upper end of the scale, enormous "hi-tech" enterprises might better fit into other Special Industrial Categorization (SIC). The empirical basis of deterrnining the fmsize distribution with gross receipts would be flawed if pnce changes of different agriculture products were taken into account (Gardner and Pope
1978). Price changes between census years could move a large nurnber of farms between sales classes. Therefore, the increase in fann sizes, when measured with fann sales alone may be due to increases in pure scale, defined in terms of total inputs committed per h,and/or to increases in product prices. Gardner and Pope (1978) showed that acreage as a size measurement gives a substantially different picture of farm size enlargement compared to the nominal gross sales as a size measurernent.
In general, if the gross sales data are not adjusted for inflation, farm sizes can get bigger with time without substantial changes in the farming sector. Even with the inflation- adjusted gross sales, vastly different farms could end up in the same gross sales class. For instance, fiom hivo hog finishers, if one buys al1 his feeds while the other rnakes feeds by himself, their inflation adjusted gross sales may be the same, yet these are two different farms From their size of activities.
Furthermore, gross sale value cannot capture differential changes of size distribution arnong fming activities since different fming activities are aggregated across a given sales class. Cater and Johnston (1978) noted that the extent of transition in
U.S. fming vax-ied significantly by commodity group and by region. Gmss sales could not capture such variability among commodity groups.
Some of the above shortcornings in farm size measurements using gross sales could be corrected with value addition as a measure of farm size. Value addition of a fmis the monetary difference between the value of inputs purchased and the value of outputs sold. However, with the use of value added as a size measure, vastly different primary input users could be classified as the same size fmers. For instance, the sarne value added is possible fkom a hydroponics tomato grower and a land-based tomato grower (when one takes monetary value of the inputs and outputs). Yet, these fmsare not "similar" to be in the same size class. Rodgers (1988) suggests that a general measure of size would be a weighted-sum of al1 prirnary inputs owned and controlled by the fm. The weights can be either opportunity cost of the primary inputs, or marginal value products of the inputs.
3.1.C. Farm Size Distribution and Changes in other Structural Dimensions
Fmsize distribution is incapable of capturing the changes in vertical coordination Iinkages, ownership arrangements, and level of specialization/diversification. The most drarnatic structural changes presently occurring within the North American farming sector deals with the relationship dimensions of structure (Bohelje 1999). Agricultural markets depend less on spot pricing and central markets and more on forward contracting (largely through fiitures markets, and less formal shorter-term cash contracting), production and marketing contracts and vertical integration (O'Brien 1994). Henderson (1994) discussed the changes in the use of productiodmarketing contracts and vertical integration in 17 commodities from 1970 to
1990. Except for fed cattle, al1 the other commodities showed an increase in either contract production or vertical integration, with the largest in hogs, fiesh vegetables, eggs, and sheep and lamb (p.553). Fmsize does not capture the changes in such vertical linkage arrangements.
In addition to vertical linkage dimensions, size as a measure of €am structure does not capture the changes in fam~business organization or fmtenure arrangements.
Although U.S. fmsare getting bigger, the majority (about 85%) remained as farnily- owned business enterprises. Excluding small farnily-held corporations, farm corporations made up only 0.4% of al1 farms, 1.3% of al1 farm acreage, and generated only 6% of al1 receipts (US Bureau of Census, Census of Agriculture 1992). An important research question is why agriculture resisted the transition to large corporate ownership (Allen and
Lueck 1998, Johnson and Ruttan 1994). Similarly, the nature of tenancy and share fârming arrangements remains an important research topic particularly in developing countnes (Cheung 1969). Changes in the distribution of farm sizes are unable to shed light on changes on either ownership or tenancy dimensions of farm structure.
3.1.D. Importance of Coordination Arrangements as a Measure of Farm Structure
Recent literature stresses the importance of changes in the coordination arrangements in understanding fming industry structural changes. For example,
Boehlje (1999) emphasized the importance of quantifying the determinants of the coordination efforts arnong increasingly interdependent business partners. Similar views are expressed by Barkema, 1994, Hennessy and Lawrence, 1999 and Cozzarin and
Westgren, 2000.
if one asserts that farm structure is changing in the coordination or vertical linkage dimension, one has to show that the number of farrns under different categories in this dimension has changed. How do agricultural economists categonze farms in the vertical linkage or coordination dimension? Compared to the categorization of fmsizes, coordination categorizes are relatively hazy. in characterizhg fmsize distribution, one has to select a size measurement (Le., gross sales or land area), and then it is a matter of putting the nurnber of farms under each size category. Yet, such a continuum is not clear
(at least outwardly) with the vertical linkage dimension. What are the different categories of farms within this dimension and what is the basis of that categorization?
To define vertical linkage categories, a useful concept is the supply chain. Supply chain or value chain is known as the process that begins with the purchase of raw materials and ends with the distribution and sales of finished goods and services
(Besanko et a2 2000). The categories of farms in the vertical linkage dimension are determined by the methods that fms use to transact with the other vertically related economic agents in the value chain. Vertical transaction is a transfer of an interrnediate good or service as an input in the production of another good or service in the value chain
(McFetridge 1994). There are many ways to carry out a vertical transaction. A vertical transaction is intemal when the buying and selling farms are jointly owned. In such a transaction, market pnces do not play a role and administrative decisions determine the transfer of goods. This is an exarnple of the two stages of the value chain being vertically integrated. A vertical transaction is market-coordinated when the two parties to the transaction carry it out in the market.
There are an endless variety of vertical relationships in between the market transaction and vertically integrated transaction (Mahoney 1992). These different varieties could provide the basis to categorize fmsacross coordination dimensions. Al1 the following categories noted in the industrial organization literature could be a coordination arrangement in a value chain; relational contracting (transactions between bvo parties on a continuing basis involving reciprocal obligations or private ordering), hybrid (contractual arrangements inside the fimi), franchise arrangements (extraction of brand name rents), strategic alliances (cooperative agreement between firms for some strategic purpose), and joint ventures (jointly owned new enterprise of two or more independent firms) (Henderson 1994). Moreover, the possible combinations of these coordination arrangements would be countless when backward linkages (input procurement), forward linkages (output marketing) or both of the directions from the fmare considered. Categorizing farms in such countless coordination alternatives may be impossible but may not even be necessary to comprehend the changes in coordination aspects.
3.2. Models to Explain the Forces Affecting Farm Size Distribution Changes
Ln the previous section, ways to characterize farm structure and ways to measure the farm structural changes have been discussed. What are the essential forces that shape fmstructure across time? How do agricultural economists explain or mode1 the effect of such forces on different structural dimensions? These aspects will be considered in this section and are sumrnarized in Figure 3.1.
Figure 3.1 shows the structural dimensions (second row) and the drivers (grouped into six categories) of changes in these dimensions. The categorization is neither exhaustive nor mutually independent. It provides a synthesis of generaily agreed groupings of the drivers of structural changes that have appeared in the literature. Aggregate Mix of Farm Industry Structure 1 Dimensions 1 Structural Farm Sue Vertical Types of Tenurd Level of Off- f. Distribution Coordination Business Ownership SpcciaIization/ Employment Dimensions Arrangements Organization Distribution Diversification
Technological Factors
economies of size: technical, pecuniary, and external, labor capital substitution, differences in technology adoption, specialization with technological breakthrouj$i Drivers of Human Capital Development managerial improvements, changes in consumption patterns, off-fam employment, opporhmity cost of labor, Structural Prices and Financial Sector Factors asset pricing/ land value appreciation, price instabilities in export markets and risk management, substitution of inputs, substitutions of consumption goods, price-cost margins Changes Public Policies and Programs comrnodity programs, farm credit, taxation, research and extension, socialization of risk, monetary and fiscal policies, unemployment policies, environmental regulations and ceilings In National Economic Crowth changes in consumption and demand pattern, non-farm income, rural urban migration
Agriculture Agribusiness Orientation
tighter coordination requirement., inadequacies of coordination through prices, concentrations in input and output marketing, strategic motivations, asymmetry of information, increased asset specificity
Figure 3.1. Causal Factors of Changes in Structural Dimensions: A Synthesis
One important question is what factors or drivers do affect the changes in what structural
dimension. There is a countless web of possible influences. As Brinkman and Warley
(1982) noted structural change is a two-word vocabulary, that encapsules a "multifaceted
topic of great complexity (p.8)". The structural change drivers include virtually every
external force acting on the many dimensions of the structure of farming industry, and a
variety of influences that exist within the sector itself. The focus here is on the dnvers of the changes in fam size distribution and vertical coordination arrangements since these dimensions are the ones that are important in the
Sn Lankan tea industry. The important analytical models for explaining farm size distribution changes could be categorïzed into the technology model, the human capital model, and financial model (Bohelje 1992 and Rodgers 1988). In addition, there is a less fomalized approach in the literature about the effect of the drivers on the fming industry structure. Instead of analyzing the effect of one economic variable on one structural dimension, researchers discuss the possible simultaneous effects of a driver on many dimensions. For instance, taxes in agriculture have been discussed, in the aspect of its effects on size distribution changes, ownership changes, and changes in off-fm employment. The analysis by Tweeten (1984), Brinkman and Warley (1983), Tweeten and Hufhan (1980), Babb (1979) are in this nature. Both the formal models and the informal discussions about the changes in farm size distribution are reviewed here.
3.2.A.Technology Models of Farm Sue Distribution Changes
Technological changes are important dnvers of structural change. Technology generally refers to the input-output relationship faced by fmers in transforming resources into products in the fming industry. There are three technology-related models in the literature; economies of size model, adoption of technology model, and labor-capital factor price model. 3.2.A.1. Economies of sue model
The economy of size2 model is based on the relationship berneen the size of the operation and the long-mn cost curve in farm production. The implication for farm structure is that if economies of size are present, larger fmshave a lower cost per unit output compared to smaller farms. Thus, in the long run, fmswant to become larger given everything else constant. The historical association between advances in mechanical technology, growth in labor productivity, and increases in farm size have been taken as evidence of size economies in farming (OTA 1986). Therefore, according to this argument, the increase in farm size is a naturai consequence of economies of size or lowering of long run average cost with the increase in size of fmoperations.
Other than exogenous technological advancements, Boehlje (1992) noted that the following aspects may also affect the long run cost curve of fmsdetermining its shape and shifi; govemrnent policy including taxation, commodity programs (direct subsidies, pnce and income support), risk reduction, marketing, and pecuniary economies in input market (volume dependent discounts) and output (volume dependent premiums). These factors could affect the shape of the long-run average cost hnction and consequently the optimal fmsize.
Economies of size models have a mixed record on explaining size distribution changes, according to the results of empirical research. Hallam (1989) reviewed many of the studies about the impact of economies of size on farm structural change and
'The expression "econornies of size" and "econornies of scale" are not interchangeable. Technical scale relationships imply constant factor proportion, which are almost impossible to achieve in the real world and dystùnctional in the context of structural changes analysis. This is because apart fiom the changes in factor proportions, the entrance of new factors is inevitable with stmctural changes. Hence, an economy of size (the relationship between size of out put and average costs) is the relevant and widely use term. concluded "some economies of size or scale may exist for livestock fms, but significant economies, at lest conventionally defined, do not exist for most crop production activities" (p. 62). Tweeten (1984) estirnated the relationship between the total resource costs of al1 inputs of different farm sizes and their gross receipts for 1981 census data (in parity ratio of 1910-19 14 input prices). He found that smaller fms(gross receipts less than $100,000) required 200 percent of 1910-1 914 parity to cover al1 costs in 198 1.
However, for the rest of the fmsizes (fiom gross receipts of $100,000 to $700,000) al1 input costs were covered at just around 54 per cent of 1910-1914 parity. Tweeten (1984) along with many others (Stanton, 1978, Miller, 1979) have found the long run average cost curve to be "L" shaped, indicating most economies of size can be realized on units operated by one family with minimal hired help.
An unanswered issue in the economies of size mode1 is the source of size economies. Although economies of size models are fiequently discussed in the literature, the sources of economies of size are implicitly assumed. Kislev and Peterson (1996) evaluate the sources of economies of size. They dispute the assertion by Hayami and
Ruttan (1985) that the size economies usually stem fiom the Iumpiness or indivisibility of fixed capital. Kislev and Peterson argue that size economies are a long run concept, and the long-run size distributions of the assets such as machinery, land, structures, are continuous rather than "lumpy". In the few cases, where large machines are the most efficient (combine harvesters and cotton pickers) rental markets develop. They conclude that "arnong the conventional inputs of land, labor and capital, long-mn indivisibility should not be given senous consideration as a source of economies of scale" (p.158).
They also disputed the "managerial" or farmer's human capital indivisibility as a source of economies of size. The growth in off-farm employment provides a long nui divisibility
of this "fixed" input. The economy of size approach has not been very successful in
explaining fmsize distribution changes.
3.2.A.2. Technology Adoption Model
An important theoretical model that analyzes the impact of technology on changes
in fmsize distribution is the 'Treadmill model" by Cochrane (1979). This model
assumes a constant returns to scale technology, exogenous technological change, and a perfectly competitive farm sector consisting of a large number of fmswhich have a very similar cost structure. it also assumes that some fmers consistently adopt new technologies more rapidly than others do. These early innovators achieve lower unit costs and hence increased profits. They expand their output levels and consequently output prices start to faIl with the increased supply. Both the higher unit cost and lower output prices would adversely affect the laggards or late adopters of the technology. Yet, under the perfectly competitive assumptions this is a dis-equilibnum situation. The equilibrium restores when the early innovators absorb the productive resources of the laggards until a no-profitho-loss situation result. In the process, the farming sector ends up with fewer and larger fms (Rodgers 1988). The next round of exogenous technological advance will tt-igger the same cycle and equilibrium will restore itself with still bigger fmsyet smaller in number.
The main weakness of this model, as Rogers (1988) discusses, is its inconsistency in the assumptions. He noted that the main assurnptions of constant retums to scale in agriculture and the possibility of "early adopters" eaming positive profits in short-run are not consistent. Rogers argued "faced with constant retums to scale in a competitive environment, the logical response by "early birds" or "rent earning" innovators is to expand indefinitely" (p.270). If a no profitho loss stable equilibrium is to be resorted after a wave of exogenous technological advancement, there must be diseconornies of scale to prevent indefinite expansion of "early birds".
3.2.A.3. Labor-Capital Factor Prices Mode1
Kislev and Peterson (1982) show that fmsize can change without any change in technology. The driving force in their model is a change in relative prices between machines and labor, which originates fiom changes in the non-farrn sector. For farming, the pnce of labor is the alternative earning opportunities outside agriculture. Machinery is the substitute for farm labor. Relative changes in the opportunity cost of farm labor
(rneasured by manufacturing sector wages) against machinery cost (measured by the real value of rental custom rates in combine harvesting of wheat) is an important determinant in farm input use. The index of wage to rental charges increased (constant price of 1987) fiom 30 in 1929 to 100 in 1987. These relative price changes provided fmers with an incentive to increase the arnount of machinery on fms, enabling the family farm to cultivate a larger amount of land. Kislev and Peterson (1996) note that coinciding with the energy cnsis in the mid- 1970s, the index of wage-rental ratio exhibit a sharp decrease
(from 104 in 1969 to 74 in 1982) and average farm size (measured in acres) remained relatively constant during this period.
Rodgers (1988) has criticized this model on its inability to shed light on changes of the distribution dispersion or higher order movements of farm size distribution. The rate at which large fms get large is much higher than the growth in average fmsize.
Although the model explains the "growth in average" farm size, it could not explain the differential changes among différent fmsizes.
3.2.B. Human Capital Model
The human capital model assumes the impacts of increased investment in human capital and the improved managenal capacity in the farm sector favor larger farm sizes.
Managerial input is cntical to the cost and production relationships of any Brm (Rodgen
1988). If managerial capacity is a fixed factor, then costs will eventually rise with increased size since higher levels of' output are receiving less and less managerial input, which is essential for efficient production. However, there have been increased investments in human capital in agriculture over time which has allowed farmers to more cost effectively manage larger scale specialized units (Boehlje 1992). There are inherent variations of managenal input due to factors such as experience, education, and ski11 levels. With this inherent variation, the optimal fami size will Vary by individual but in general it increases in size with improved human capital.
The main weakness of the hurnan capital model is the argument of managerial indivisibility. The increasing availability of off-fm employment is implies managerial input or the famer's time is divisible (Tweeten, 1984). The majority of fmers with small operations are engaged in off-farm employment. The relationship between increasing fmsize and improvements in managerial abilities through human capital development depends on the status of off-fann employment. When the off-farm employment opportunities increases, the increase in hurnan capitaI does not necessarily increase fann size, since off-farm employment could accommodate the increase in hurnan capital.
3.2.C. Financial Mode1
Lowenberg-DeBoer and Boehlje (1986) develop a financial model to show that fa.size is influenced by the balancing of annual incomes and capital gains. The optimal mix of durable inputs such as real estate and other capital and non-durable inputs such as seed, fertilizer, and chernicals depends on the relative contribution of these to farm profits. The finance charges and capital gain parameters in holding the durable inputs (especially real estate) play an important role in deciding the optimal mix of the durable and non-durable inputs. in equilibrium, retums and costs of durables
(contribution to production, appreciation or depreciation) and non-durable inputs (only contribute to production) determine their optimal mix and consequently the optimal farm size. Lowenberg-DeBoer and Boehlje (1986) noted that fmland capital gains might go prevent the adoption of intensive technologies since appreciation in the value of fannland may induce bringing more land into the farm operations, and thereby increase the net worth of the €m.In turn, this may help explain the increase in acreage per fann and relative low growth in land productivity. In general, the model suggests that in an environment of large capital gains, farmers will tend to enlarge fann acreage and incur higher debt loads in their attempts to take advantage of the value appreciation of fannland. 3.2.D. Public Programs and Policies as a Causative Factor of Structural Changes
The following fmpolicies have been noted by many studies as having important effects on farm structure: commodity programs, credit programs, tax policies, and research and extension programs.
3.2.D.1. Commodity Programs as a Causal Factor of Structural Changes
Commodity prograrns include a variety of policy instruments such as income support (deficiency payments, disaster payments, and crop insurance) price stabilization, commodity credit, and/or supply management (grain reserves, marketing orders, acreage and quota allotment) (Meyers and Westho ff 1989). Commodity programs invariabl y help marginal fmers to remain or attract them into fming (Penn 1979, Tweeten 1984).
Gardner (1978b) proposed the impact of cornmodity programs on structure is known best by knowing who would leave fming without such programs. Richardson, Smith, and
Knuston, 1989 tested this fmsurvival this hypothesis with farm data and analyzed what farm size would benefit rnostly from commodity programs. They argued that the farm survival rate would be higher with higher off-farm incarnes and that the lower and moderate debt producers are better able to survive in a cash flow cnsis compared to higher debt fmers. Tenant operators are less able to survive in a crisis than part or full owners because tenant farmers could not borrow against equity in land to meet cash flow deficits. Economies of size correspond to lower cost of production, and therefore they argue that farms with size economies are better able to survive in a crisis. Then they looked at the association between farm size and the above four charactenstics of farm survival. They found that mid-size fms in 1986 (annual gross sales $ 40,000 to 250,000) have lower levels of off-fm incomes, higher debt-to-asset ratios (as much as
bigger famis) but relatively fewer means to manage nsk, a higher share of tenant fmers,
and a lack of economies of scale. Therefore, they conclude that mid-size farms would be
the category hurt most without commodity prograrns.
Another significant impact of commodity programs on fmstructure is through
its impact on nsk in fming. Penn (1979) noted that with less nsk, farmers tend to
specialize in specific commodities to achieve technical, pecuniary, and extemal
economies that are reflected in reduced unit production costs. Lowered risk tends to
induce the desire to expand farm size 6om a given size of equity. Therefore, cornmodity
programs may tend to increase the relative number of specialized and larger farms
through the reduction in nsk.
The income transfers and reduced risk of commodity programs have provided
incentives to acquire more fixed-assets. Therefore, commodity program benefits tend to
become capitalized in asset prices, mainly in farmland (Penn 1979, Weersink et ai 1999).
Pasour, 1990 (p.139) argued that the pnce of land hinges on the discounted Stream of
expected future returns. If there is an increase in product price so that the expected returns to land increases, land prices will increase. It is argued that increased land pnces act as an entry barrier for the new entrant (Penn, 1979). However, it has been also argued that the existing large farms are in a relatively advantageous position to debt finance and acquire such land through their large equity base (Penn, 1979).
The above account shows that the effect of commodity programs on structural changes in the farming industry is mixed. Tweeten (1984) and Sumner (1985) concluded that fmcommodity programs in the United States have been neutral in their overall impact on fmsize and nurnbers.
3.2.D.2. Credit Policies as a Causal Factor of Structural Changes
The general intent of credit policies in agriculture has been to ensure adequate capital availability for agriculture (Penn 1979). Readily available capital at reasonable rates and terms plus technologies that aid pro fitability provides a suitable environment for technology adoption (OTA 1986). Various govemrnent and govemment-supported lending agencies provide low-cost credit for needy farm operaton in many countries.
Credit arrangements and policies of the past are believed to have contributed to maintaining a structure of the farm production sector that, compared with many other sectors is largely a small-scale, pluralistic, non-corporate, cornpetitive market organization of ownership, management, and control (OTA 1986).
In general, the impacts of farm credit on structural changes are considered to be exerted through factors such as asset prices (land) and adoption of technology, nsk taking and slower mobility of some resources (OTA 1986). Readily available credit may have increased the cornpetition for land and further increasing land pnces (Penn 1979).
Together with entry barriers at higher land prices, improved credit facilities tend ta encourage growth in farm size and consolidation of hing units. Lins (1979) has shown that fiom 1950 through 1979 the average size of fmsin US. increased from 213 acres to 443 acres, while farm mortgages increased during the same period fiom $ 5.6 billion to $ 63.6 billion. This indicates a substantial increase in debt financing in acquinng new lands for farming. It is argued that larger famis are. in more favorable position to finance asset acquisition compared to smaller famis.
in a similar vein, improved credit facilities hastens the adoption of new technologies, which generally favor larger size fms (Penn 1979 and Lins 1979).
Increasing labor costs prompt €mers to use more labor saving technologies, which usually require larger and lumpy financial commitments. Credit facilities are important in financing such technologies. The higher debt to asset ratios in non-real estate assets of the largest fmclass (42% in 1970 and 46% in 1982) suggests to the smallest Fmclass
(8.7% in 1970 and 12% in 1982) reflect that larger fms uses their equity base in financing non-real estate assets such as machinery (Tweeten 1984). Pasour (1990) noted that in view of' widespread public concems about fann size and capital requirements in commercial agriculture, it is ironic that government operated and govemment-sanctioned credit programs may have contributed to the trends towards larger and more highly rnechanized farms.
While some argue that an impact of credit policies for agriculture has kept many marginal fmers operating who wouid othetwise have left the farming (Penn 1979), others argue that the abundant credit facilities have attnbuted to the expansion of farm size (Lins 1979, Tweeten 1984). Empincal evidence suggests subsidized and readily availabie credit has tended to expand fmsize.
3.2.D.3. Tax Policies as a Causal Factor of Structural Changes
Preferential tax treatments for agriculture through capital related taxation and more flexible provisions of overall tax calculation have had an impact on hm size distribution agriculture (ESCS Staff 1979, Brinkman and Warley 1983, and OTS 1986).
Tax policies of the capital assets in farming are different f?om other sectors and this has encouraged labor substitution by capital and hence growth in fmsize (Tweeten 1984,
Pasour 1990). In most sectors, investments in capital assets are not allowed to be deducted as current expenditures fiorn taxable income. However, in farming, investrnents in certain capital assets are allowed to be written-off as current expenditures. In addition, fming sector tax rules tend to allow for larger tax write-offs for depreciation of capital assets (Tweeten 1984). These are tax benefits if the costs of depreciable assets are written off before the property stops contributing to fmincome. Moreover, investment tax credits for agriculture have been commonly used over time. Such preferential treatment in capital assets has encouraged substitution of capital for labor by lowering the real cost of capital relative to labor (OTA 1986). The adoption of capital-intensive innovations such as large-four wheel drive tractors, circle irrigation systems, minimum tillage system, and large-scale improved harvesters are considered to be influenced by the tax provisions
(OTA 1986).
There are some flexible provisions in tax calculations for fming, which have enhanced the attractiveness of investing in fming (Pasour 1990). Accrual accounting calculates net income by matching income and expenses over the ordinary course of the business regardless of whether such income and expenses have been received or spent by the time of calculation. Income taxes in agriculture are based on the cash accounting system, which calculates net income as the income received and expenses incurred. inventories of unsold goods and unused inputs are ignored. When compared with producen operating under the accrual system, the cash accounting system permits taxes to be deferred. This reduces the tax liabilities of farrners since it is always beneficial to receive a benefit sooner rather than later (in terms of present value) and sometime this provision allows a reduction in the total tax liability (Penn 1979). Since the cash accounting method is allowed to be used by "family" farms, this favors certain ownership classes and influences the fming industry ownership structure.
While the marginal tax rate usually increases as income increases, progressive taxation is not as effectively operated in the farming industry as in non-fann sector
(Tweeten 1984). The marginal tax rate in farming is somewhat proportional across al1 the fmincome groups (Tweeten 1984). The emerging picture is that overall, taxes not only were lower for farmers than for othen, but also were less progressive, suggesting tax policy has promoted larger and fewer farms.
3.2.D.4. Research and Extension Policies as a Causal Factor of Structural Changes
The public sector research and extension programs in agriculture are considered another factor that can influence the fming industry structure. The primary focus of agricultural research and extension activities continues to be agricultural productivity.
Tweeten (1984) argued that most publicly supported agricultural production research has been for biological technologies, such as improved crop varieties or output increasing technologies, which are scale neutral. He also noted that according to an estimate for
1979,85% of publicly supported agricultural production research was scale neutral.
Tweeten (1984) noted that private sector research place more emphasizes on mechanical and chernical technologies which is usually laborsaving and tends to be scale biased. He argued that the most important single source of structural change in U.S. agriculture has been the tractor and its accessories, which are developed by private firms.
These research payoffs have increased the labor productivity immensely and caused farm size to increase. Dramatic improvements in agriculture technologies are mainly responsible for increasing the supply of farm products. Consequently, real fmpnces are declining al1 over the developed world. However, demand for farm products, which increases mainly due to increases in population and increases in consumer income has not increased to the same extent as the supply of farm products. The resultant depressed prices lowers margins and generally leads to large fanns who are able to generate sufficient income from the smaller margins.
Agriculture extension also has an impact on fmsize distribution. Larger farms and agibusiness fims have sufficient resources to obtain information such as pnce forecasts and new technological developments on their own giving them an edge over their cornpetition. However, many moderate and smaller fanners could not afford private information services without publicly supported agriculture extension services that provide for al1 end-users. Given the availability of prices and rapid technological developments, small fmsurvival would have been harder without a publicly supported extension service.
3.2.E. National Economic Conditions as a Causative Factor of Structural Changes
Pasour (1990) noted that problems in the farm sector could be attributable to destabilizing effects of economic growth. National economic conditions and growth are tied pnmarily to eamings in the non-farm sector. Ruttan (1988) noted that in developing countnes the speed of adopting labor replacing technologies is slower than that for biological and chemical technologies. He attributed this to an insignificant increase in the relative prices of labor in such economies due to slower growth in the non-fm sector.
However, in the developed world, the increase in non-farrn labor earnings relative to fhrm labor eamings has caused two developments. First, as discussed previously, the change in relative prices between labor and capital induced mechanization and hence the increase in land area cultivated per labor (Kislev and Peterson 1982). Second, income fiom fming must also be increased on par with the non-fm sector. Tweeten (1984) estimated that, given everything else stays the sarne, fmsize (measured in sales) had to scale up by about 1.2% in the 1950s and by 3% in the 1960s to match the increase in real per capita disposable income. Thus, the increasing non-farm labor earnings caused farm size to increase in these two ways.
A related aspect is the incidence of off-farm employment and its impact on the requirement for scaling up fmsize. The availability of off-farm employrnent tends to counter the trend towards fewer and larger fmsas it provides another income source for particularly small farms with less of a tirne constraint, that can support their fârm operation (Goddard et ai 1993). Off-fm opportunities are related to the economic growth of the country and the growth of the non-farm sector. Therefore, the net effect of the changes in fmsize distribution fiom growth in the non-farm sector is not clear.
3.3.Models to Explain the Forces Affecting Vertical Coordination Changes
Drivers of changes in vertical linkages are also shown in the Figure 3.1. Exchange mechanisms coordinate buyers and sellerç with one another to trade products and senices. At the producer and first handler level, exchange arrangements include
bargaining, formula pricing, administered pricing, private treaty, and central or auction
markets (Tomek and Robinson 1981). These mechanisms differ by the amount of
information, transaction costs, productive efficiency, risk and market power allocated to
buyers and sellers (Klimer 1987). Do these qualities of an exchange arrangement change
with time for a given transaction? If such changes are happening, what causes these
changes? The consensus arnong analysts is that vertical linkages in agricultural
cornrnodity marketing channels have been evolving to permit tighter coordination
arrangements so as to better face the challenges of various exogenous changes (Boon
1999, Barkema, 1994b).
The substitutions among alternative exchange mechanisms are pt-imarily driven by the efficiency concems over resources used in the exchange process (Kilmer 1987). hefficiencies are created due to the conflicting interests among parties to exchange and theones to analyze such conflicts have been embraced by the analyst in the structural change analysis of agriculture. Mahoney (1992), Sporlder (1 9W), Hennessy and
Lawrence (1999), and Boehlje (1 999) noted that agency theory (Jensen and Meckling,
1976), and transaction cost economics theory (Williamson 1979; 1985) are the important theoretical approaches in analyzing vertical linkages dimensions. However, familiar industrial organization models that explain the relationship between structure
(competitive nature of the fims in the industry) and performances (cost-price margin and output decisions of the firms) are used to analyze concentration issues of the agi-food sector. Therefore, the traditional structure-conduct-performance and market power explanation, principal-agent models, and transaction cost economics approaches are reviewed in the following section. These approaches in the literature attempt to explain the changes in vertical linkages in agriculture and related industries. The reviews are brief and are not exhaustive given that their applications in various sectors are rapidly increasing.
3.3.A. Structure- Conduct- Performance Mode1
Industrial organization modeis were initially used to explain the reasons for selechng one coordination arrangement over another (Trifon 1959, Mighell and Jones
1963, North Central Regional Project 1 17, 1976). The main concern in the structure- conduct-performance approach is the welfare losses due to deviations from the nom of competitive markets. The motive of various coordination arrangements according to this tradition is the possibility of firms gaining an advantage over rivals as a consequence of vertical integration across either factor or output markets (Sporlder, 1992).
The important dimensions of industry structure are the concentration of buyers and sellers (distribution of sizes of firms), degree of product differentiation, and conditions for entry and exit. These determine the conduct of firms in terrns of their pricing and output policies, product development, promotion, and reactions to their rivals.
The conduct thereby determines the market performance measured by price-cost margins, production efficiency, advertising, etc. The functional relationship between these three dimensions of industry structure has been identified as the stmcture-conduct-performance paradigm (Henderson 1988).
In structure-conduct-performance models, market power is the driving force explaining firms' decisions to change the way they vertically transact. Market power is not an easy concept to pin down. Mason, (1956)) noted "It is not possible, nor will it
ever be possible, by calculating market shares, dividing price minus marginal cost by
price, or other hocus pocus, to present an unambiguous measure of the degree of
monopoly. Market power has many dimensions". The specific motives of changes in
coordination arrangements with a vertically related hsare elimination of factor price
distortions, elimination of successive mark-ups in the presence of imperfect competition,
creation of entry barriers, and price discrimination (Henderson, 1988).
Assessing the extent of concentration and its effect on competition has been a
primary focus of this paradigm. For instance, Marion has concluded that a 10% increase
in the four firm concentration ratio (CR 4) for the cattle slaughtering industry results in a
1O to 23 cents per hundredweight decline in live cattle prices to producers. However,
Sumner (1994) concluded that while there is evidence of some concentration in the food
industry, there is not enough for the firms involved to feel secure or an active anti- concentration policy is likely to raise prices for farm sellers or lower pï-ices for consumers. Sexton (2000) concludes in a similar tone that market power created in the industrialization and consolidation process appears, based on extant literature, to be rather modest, and probably 'worth it' from an efficiency calculus" (p. 1101). Therefore, market power appears to be an incomplete explanation for the changes in vertical linkages.
3.3.B. Principal and Agent Model
Recently agriculture economic tesearchers such as Hemessy (1996), Lajili, Barry,
Sonka and Mahoney (1 997), and Goodhue (2000) have the used principal-agent approach
' quoted in Shepherd, W. D. Market Power and Economic Welfre. Random House, New York, 1970: p.24. to explain exchange arrangements in various comrnodity sectors. Principal-agent theory
recognizes conflicts of interest between different economic actors and formatizes these
conflicts through the inclusion of observability problems and asymmetric information.
Arrow (1985) used an example of stockholders and management to describe the
importance of unobservability of the managers' action on which the stockholders' profits
are dependent. The stockholders' profits are dependent on both the exogenous factors
and the actions of the managers. Unobservability is due to the inability to distinguish the
impact of managers' action on profits net of the impact of exogenous factors.
This approach recognizes that the interests of parties to an agreement can diverge
and that it might be impossible to constnict an incentive mechanism that maximizes total potential surplus when the inputs and/or outputs of agents are not observable by the principal. Principal and agent methods seek to identify and optimize the trade-offs concerning the nature of the incentives. Comrnon incentive structure decisions facing a contract designer include whether payrnent should consist of a low base fee and high incentives, a fixed payrnent, or something in between.
A recent application of this mode1 to the broiler industry seeks to explain the choice of compensation method by integrators to the growers (Goodhue 2000). She addresses the issue of how the integrator can maximize his profit given the different degree of risk aversion and capabilities of growers so that the growers are rewarded according to their relative performance. She shows that input provision in broiler production contracts (resource providing contract production) is a measure of controlling informational rent4. Othenvise, the principal would have to pay this rent for private
returns the superior growers could earn from their unique abilities as the growers, but this information known only to these growers information for highly productive growers (who can achieve maximum input use efficiency).
The payment method of toumarnents (tournarnents are a method of payment, which depend on the relative performance of a given grower among a set of growers whose performances could be treated comparably) can remove the cornrnon uncertainty in production to which al1 the growers will face equally.
The main weakness of the principal agent heworkis that it cannot explain why a particular coordination method is more suitable than the another in a given context.
Hart (1996) noted that pnncipal-agent theory could explain why it might make sense for two parties to vwite a profit sharing agreement, whereby rewards to agents are based on receipts to the principal. This encourages the agent to supply his best effort. Yet, the theory does not tell whether it matters if this profit-sharing agreement is accomplished through vertical integration, or whether both should remain as separate firms. The theory says linle about the evolving coordination arrangements. Specifically, it is not clear, why the task was given to the agent or why two parties do not interact through the market place (Hennessy and Lawerance 1999). Principal-agent theory tells us about optimal incentive schemes but not about organizational form (Hart 1996).
3.3-CmTransaction Cost Economics Approach
As early as 1937, Coase noted " the main reason why it is profitable to establish a firm would seem to be that there is a cost of using the pnce mechanism" (p.390). Based on this premise Coase argued that a firm will tend to expand until the costs of organizing an extra transaction within the firm become equal to the costs of canying out the same transaction by means of exchange on the open market or the costs of organizing in
another firm. The main cost of transacting in a "pricing system" or in the market is the
cost of learning about and haggiing over to discover the relevant prices. When market
transaction is more costly, transacting parties will select another exchange arrangement to
minimize transaction costs. Developments in this tradition over the last few decades,
broadly categorized as transaction costs economics (Williamson 1985), has become one
of the main theoretical means to explain the choice of alternative exchange arrangements.
The characteristics of transactions and the characteristics of parties to the
transaction play important roles in determining the size of transaction costs in a given
transaction (Kreps, 1990, p.749). Williamson (1985) shows that transactions can be
characterized according to its 1) complexity, which includes the amount of uncertainty
that the transaction bears (especially about future contingencies); 2) thinness of the
transaction (number of alternative trading partners involved in it); and 3) the extent of
impacted information in it (information only some parties to the transaction possesses).
The parties to the transaction cm be characterized: 1) as opportunistic (pursue selfish
interest in guilehl manner), and 2) as limitedly rational (it is costly and sometime
impossible for them to carry out al1 the elaborate thinking required to find the tnily
optimal course of action or think through al1 contingencies that might bear on the
transaction). Relatively greater complexity andor thi~essandor impacted information, joined with relatively greater guile andor relatively more limited rationality will raise
transaction costs (Kreps 1996).
The parties to a transaction can select an exchange arrangement to carry out the
transaction that will minimize the costs of exchange. Exchange arrangements can span fiom complex organizational choices (brand name rents could be collected through
&anchking or a fixed leasing arrangement, a Company product could be distributed through an independent network or commissioned employees) to a simple exchange arrangement (spot market purchase). Transaction costs economics identifies al1 such alternatives as governance structures, which Williarnson (1979) defines as the
"institutional matrix within which transactions are negotiated and executed" (p.239).
Why could not only one exchange arrangement or "govemance structure" facili tate al1 transactions? Why are there di fferent exchange arrangements depending on the transaction? Each transaction is a surplus generating activity for the parties to the transaction and this is mutually beneficial, and yet mutually dependent. Moreover, the surplus could be changed with changes in the exogenous conditions arguably leading to a conflict in surplus sharing, unless the arrangements under such contingencies are specified ex ante. In such a conflict, when the contingent plan is absent each party will wish to arrange the terms of trade in such a way as to appropriate for him as large a share of that surplus as possible (Williarnson 1998). Different exchange arrangements resolve such conflicts with different adaptability while attempting to keep the total surplus as large as possible. More adaptable exchange arrangements in surplus sharing are not necessarily the total surplus maximizing arrangements and these opposite pulls are different for different exchange arrangements. The size of these opposite pulls
(adaptability to changing economic conditions and capability to maximize the surplus) and attributes of the transactions (uncertainty that the transaction bears, number of alternative trading partners involve, and extent of impacted information) decide the most appropriate exchange arrangement for a given transaction, Riordon and Williamson (1985) argued the principal factor that is responsible for difference in the cost of exchange arnong transactions is variations in asset specificity.
Assets cm be specialized in many contexts to support transactions. Alchian (1984) delineated the specialized assets in most generalizable terms;
in forming a coalition, some members will make investments that value of which
elsewhere will be less than the value in the coalition (and also Iess than its costs). If its
value in the coalition is higher than elsewhere, it is defined to be specific to the coalition.
Some of the resources have values independent of the coalition. They will be able to eam
just as much elsewhere. They are nonspecific to the coalition. If the coalition fails, they
lose nothing. But the resources that are specific to the coaIition are those that will lose
the excess of the investment cost over the salvage value (p.36).
Specificity of an asset is measued as the percentage of investment value that is lost when the asset is used outside the specific setting or relationship (Milgrom and
Roberts 1992). It is an exchange or interaction related (behavioral) concept rather than a physical concept (Alchian 1984). The essence of asset specificity is about, "the ability of parties to tum to alternative partners should one seek to gain at the expenses of the other"
(Masten 1991, p.8). This ability could be limited by factors such as availability of alternative partners, cost of reaching them, and the storability of the item for trade. These aspects are highly variable for different enterprises, regions and across tirne. Thus, these aspects are important in explaining how an investment could become a specific asset in a given context.
When assets are specialized, parties to the transaction lose the ability to regulate the transaction by tuming to alternative partners should one seek to gain at the expense of the other. This possibility is identified as the hold up problem (Klein, Crawford, and
Alchian 1978). The hold up problem is defined as: "The general business problem in which each party to a contract womes about being forced to accept disadvantageous terms later, after it has sunk an investment, or womes that its investrnents rnay be devalued by the actions of others" (Milgrom and Roberts 1992, p. 136).
Can asset specificity and hold up problems explain the changes of vertical linkage dimensions in agri-food industries? The applications of this tradition in agriculture are few (Masten 199 2). However, there are early contributions by Knoebar, (1983), (1989), and Dye, (1994). The early contributions do not specifically address the vertical linkage dimensions of structural changes but use the hold up issue to explain the observed exchange arrangements. In these studies, the hold up problems are related to the difficulties in finding alternative trading partners with perishable agricultural products.
This has been identified as temporal specificity in agricultural vertical relationships
(Sporleder 1992). Due to the perishable nature of the raw materials (milk, hits and vegetable, sugar cane, green leaf tea), high weight to value ratio and immediate processing to reduce this catio, site specificity is likely with the processors. The transacting parties resort to alternative coordinating arrangements such as long term contracts by passing the spot market (Binswanger and Rosenweig 1986, Masten 1991,
Grosh 1994, and Delgado 1999).
A representative work in this tradition is Knoebar (1983) who explores the alternatives to litigation when fniit and vegetable growers and processors breach forward contracts. These contracts are agreements negotiated before planting for annuals and before harvest for perennials for a particular grower to supply his crop to a particular processor For a specified price. Since spot prices may diverge fiom contract pnces and so provide an incentive for one party to default, the unreliability of the forward contract is noted. Even with no divergence between spot and contract prices, both parties have an incentive to threaten default on the contract and renegotiate price. The hold up problem looms for the grower at harvest time with a very penshable crop and if default by the processor would mean either loss of the crop or a quick (costly) sale in the thin spot market (and likely some crop deterioration as well). Similarly, for the processor, default fiom the grower would interrupt the processing or require quick replacement fiom a spot market.
Knoebar (1983) then showed the role of bargaining cooperatives in deiüsing the hold up threat. Their only hction is to contract with processors for the sales of member's crops. Member growers have agreed to sel1 their crop through the cooperatives. Yet, they also have to agree to pay damages (worth 25% of the crop sold by passing the cooperative) if the agreement for the cooperative is breached. The cooperative keeps a deposit fiom the members, which can be liquidated in case of renege on a contract. Knoebar explained this as a "default bond" with a third party, which is a check against the hold up threat.
Recent applications in other hold up potentials are by Martinez, Smith, and Zering
(1998) in the pork industry with investments in quality improvements, and McLaren,
(1999) in the coffee industry where raw bean growers hold up threats are prevented by increasing unprocessed coffee prices. Perishability of raw products does not necessarily cause hold up threats in an agriculture transaction. Even if the raw agricultural product is perishable (non-storable), if there are closely scattered alternative buyers, if transportation facilities are inexpensive, then the investment in assets producing that product is unlikely to create a hold up problem. Therefore, generalization is highly unlikely from situation to situation.
The main weakness of transaction cost economics is that its prediction and empirical testing is confined to the matching of observed exchange arrangement outcomes to the nature of the transaction. There is no explicit attempt to measure the size of transaction costs with a given transaction and different exchange arrangements. There are attempts to put measurement proxies on various attributes of transactions and to predict the exchange arrangement outcomes (Lyons 1994). The measurement difficulties of various attributes that give rise to transaction costs is noted as one reason why that approach is difficult to use empirically (McFetndge 1994).
Reputation is yet another aspect that makes the validity of the hold up problem questionable in explaining the choice of vertical coordination arrangements. Reputation is a worthwhile capital good, which has to be expended in contract breaching with hold up threats. Therefore, reputation is an in-built check against the hold up problem in vertical coordination arrangements. Klein (1994) introduces the self-enforcing range of contracts or private sanctioning to explain the reputation impact on the hold up problem.
He showed that the potential capital loss of contract breaching with hold up threats is the discounted future values of the quasi rent from the investment to stay in the business. He argued that each transacting party compares the potential hold-up gain from breaching the contractual understanding with this capital loss. When the hold up gain is less than the capital loss then reputation (by means of capital loss) is a credible loss for the breaching the contract. However, if the reverse is the case, reputation cannot stop the hold up problem and contract breaching.
3.4 Summary
This chapter reviews the economic analyses of farming industry structure and its changes. Having identified the different dimensions in characterizhg farming industry structure, it discusses the inadequacies of farm size distribution, the most commonly used structural dimension to represent the farming industry structural change. The importance of the other structural dimensions such as vertical coordination arrangements and ownership distribution is highlighted. The major theoretical models used to explain the changes in farm size distribution and vertical coordination arrangements are discussed and the main weaknesses of such models are highlighted. The structural change drivers identified fiom this review will be discussed in relation to the Sn Lankan tea industry in the following chapters. Chapter 4 develops a conceptual mode1 on the basis of this review to explain the changes in the size distribution of Sri Lankan green leaf tea growers. Chapter 5 does the same but in tems of'explaining the change from a vertically integrated sector to one of independent growers and processors. CHAPTER 4
CONCEPTUAL MODEL OF CHANGES IN FARM SIZE DISTRIBUTION
IN SRI LANKAN TEA INDUSTRY
4.0. Introduction
The purpose of this chapter is to develop a conceptual hework to explain the changes in fmsize distribution that occurred in the Sri Lankan tea industry over the last
four decades. As discussed in Chapter 2, fmsize distribution is one of the important structural dimensions that have undergone a substantial change during the last four decades. The relative share of small fmsas a proportion of the total nurnber of tea
fmsand as a proportion of the total tea axa has been increasing. What has been driving the changes in these structural dimensions? How does one go about analyzing such changes?
The previous chapter reviewed the major factors or drivers that affect the changes in farm size distribution. On the basis of that review there are four major structural change driven used in the conceptual mode1 of fmsize distribution changes in the Sri
Lankan tea industry (see Figure 4.1 ) : 1) tec hno logical factors, 2) govemrnent programs and policies, 3) prices and financial sector factors, 4) national economic condition. These factors are related to the major policies and other exogenous economic factors specific to the Sn Lankan tea industry. Economic reasoning is used to develop a conceptual mode1 that explain the effects of these drivers on the changes in the farm size distribution of Sn
Lankan tea industry. Sri Lanka Tea Industry Structure Structural
( ~tmctural 1 1 Farm Size Distribution 1 Dimension 1
Drivers of Tecbnological Factors labor capital substitution, economies of size: technical, pecuniary, and external, Ï / Structural Government Policies and Programs Land reform programs, small farm support services, taxation, ~t&e&terventionin geen leaf market Changes Prices and Financial Sector Factors pnce instabilities in export markets and risk management
In National Economic Growth Alternative opportunities in other agriculhual production Agriculture
Figure 4.1. Drivers of Farm Sue Distribution changes in Sri Lankan Tea Industry
The refutable propositions1 of these effects are denved from the conceptual model. The
hypothesized effects of the drivers on the changes in fami size distribution of the Sri
Lankan tea industry are tested in the empirical section of chapter 7.
This chapter is organized as follows: Section 4.1 develops the arguments about
the nature of labor-capital substitution in the Sri Lankan tea industry and its impact on
labor costs changes for small fmsand large fms. It then develops the hypothesis on
the effect of relative labor cost changes for small and large farms sector on the observed
changes in tea fmsize distribution. Section 4.2 discusses the impacts of public programs and policies on the changes in tea fmsize distribution. Land reform policies,
small farm support services, tea export taxation, state pncing intervention, are discussed
' . Refutable propositions or hypotheses "is a set of unequivocal statements about situations that would not be observed if the theory were correct. Failure to observe these conditions is a failure to falsify the theory...... Such.. .hypothesis must take the form that event E will occur and it mut be possible to observe event E not occuning." (Fox, 1997 p.56). and hypotheses are developed about the effect of these policies on relative number of
small and large tea fms. Section 4.3 discusses the impact of tea price instability on
fmsize distribution. Section 4.4 discusses the ef5ects of relative changes in the
profitability of alternative agricultural enterprises on farm size distribution changes.
Section 4.5 surnrnarizes the main points of the chapter.
4.1. Technological Factors: Nature of Labor Capital Substitution
Input substitutions in production take piace due in part to changes in relative input prices. Profit for a firm is rnaximized when inputs are used such that the ratios of marginal value products (MVP) are equated to their pnce (P)(pnnciple of equi-marginal retum of input use). For instance, the profit maximizing labor (L) and capital (C) usage is where;
MVP , - MVP (4.1). PL Pc
The ratio of MVP to input pnce will be equal to one in the unconstrained profit maximization case. If PL (the opportunity cost of labor) increased relative to Pc (the opportunity cost of capital), MVPL would have to rise relative to MVPc other things being equal. A different interpretation would be that the marginal value product of labor has increased by using more capital per labor. Thus, increasing the relative cost of labor causes a more intensive use of capital and thereby increases the capital-labor ratio.
For instance, Kislev and Peterson (1996) showed that fiom 1949 to 1987 in the
United States agriculture, charges for rental machinery (per acre custom rate of wheat harvesting) fell by about 25%, and price of diesel fiel stayed the sarne, while labor wages increased by about two folds. The result of these changes in the relative factor prices was that labor input per fmacre (hired plus fmily, measured in labor-years) decreased Eom
0.010 to 0.003 or more than 300% during this period. While the machinery input has increased bom $ 32.48 per farm acre in 1949 to $ 167.46 per farm acre in 1987 (1987 prices) or more than 400% increase during the same period. Thus, substitutions in inputs are influenced by corresponding changes in their prices.
In contrast to the U.S fann sector, Mesubstitution has occurred between labor and capital in tea production. Tea remains a labor-intensive crop, despite increase in labor wages relative to mechanization costs2 in Sri Lanka (see Figure 4.2). As long as existing technology allows for input substitution, a change in the input prices will lead to substitution of one input for the other. However, labor has not been substituted by capital to any considerably degree in Sn Lankan tea industry.
4.1.A. Continuation of Labor Intensive Tea Production
The total labor use (man-days per acre per year) for the production of green leaf tea has remained almost identical between 1950 and 1990 (see Table 4.1). During this penod, labor replacing technological innovation did not play an important role in Sri
Lankan tea production. The major technological changes were ori yield enhancing and pest and disease control (Snodgrass, 1966).
' Since large rnachinery (similar to combine harvester or Cotton picker) is not used in tea industry. it is impossible to fmd a suitable approximation to depict the changes in relative capital cost in tea production. However, the variable cost in machinery is Iargely made out of fuel costs. T'us,diesel prices are used to approximate the relative prices of capital. Since factor ratio (which decide the etasticity of input substitution) for Iabor and fuel is not known the physical uni& are used in the Figure 4.2. This does nui indicate the input substitutability. Figure 4.2. Relative Price changes in labor and capital in Sri Lankan Tea Industry
Sources: Labor wages (Daily wage rates of RsDay) fiom Annual Reports of Central Bank of Sri Lanka Diesel price fiom Annual Reports of Sn Lanka Petroleum Corporation (both deflated by CPI lW2= 100)
Note: Oil pnces are subsidized in Sri Lanka, in 1978, with the liberalizing policies, the oil subsidies were partly removed. Due to such biologically-oriented technological developments, the composition of labor usage has changed among the activities of in green leaf tea production. For instance, labor use for harvesting has increased due to the increased output per acre that must be picked but labor use for Pest and diseases has decreased due to the increased use of pesticides (see Table 4.1). Together, the total labor requirement has not changed significantly and tea continues to be a labor-intensive crop.
Table 4.1. Changes in Labor use for Tea Production 19541993: Large Tea Farms (Figures are for annual requirement per acre) Activity Labor days '?40 Labor days YO 1954 1993 Harvesting (green Ieaf) 127.6 44.4 15s3 62 Weeding 53.7 18.7 45 19 Fertilizer Application/Forking 12.3 4.3 14 6 Pruning 5.5 1.9 6 3 Pest & Disease Control 19.2 6.7 8" 3 Others 25 -2 8 -8 15 7 Total Field Production (Green LeaB 243.3 84.8 243 ZOO Processing 24.4 8.5 N/A - Services 19.3 6.7 N/A - Grand Total 287. O 200.0 - - Sources: 1954 fiom Lamb, 1954, p. 193 (Table 1) 1993 from Liyanage, 1993, p. 82 (Table 4.1)
.' nie increase in labor use in harvesting fiom 1950s to 1990s is amibutable to increase in yield. Black tea production (kgha) increased fiom 1950- 1954 average of 653 to 1990- 1994 average of 1087 (about 66%) (Annual Reports, Central Bank of Sri Lanka 1998 Table 11, Athukorala 1984, p.84).
* The reduction in labor requirement for pest and disease control and weed control is attniutable to development of tea clones with disease resistance (Handbook on Tea, Tea Research Institute Sri Lanka 1990). Snodgrass, 1966 (p. 139- 142) gives a general account of technological improvements in tea production by Sn Lanka Tea Research Institute. Another hdicator of the total labor use is labor-land ratio in tea production. It has rernained stagnant around 1.00 labor-year per acre for the 1st 60 years (see Table 4.2).
Despite there are fluctuations in the labor-land ratio there is no trend in the changes of the ratio. It ranged fiom 1 to 1.12 man-years per acre during the last four decades.
Table 4.2. Labor per acre in Sri Lankan Tea Industry (large farms)
Pe riod Acreage Workers WorkerdAcre
1942-1945 447,078 447,348 1 .O00 1946-1950 448,557 479,160 1 .O68 1951-1955 452,522 494,427 1 .O92 1956-1959 446,340 503,622 1.128 1977 473,626 431,185 1 .O98 1995. 57,878 63,640 1.108 1996 42,8 19 48,06 1 1.122 1997 42,785 46,489 1 .O86 1998 42,115 43,407 1 .O30 1999 4 1,453 46,334 1.1 17 Sources: (1) 1942 to 1959 fiom Lim 1968, p.462 (averages were taken for each 5 years pet-iod)
(2) 1977 acreage From iTC Bulletin 1980, labor force figure Sinnathamby and Wicramasekara, 1984, p.9 1.
(3) 1995, 1996, 1997, 1998 and 1999 tiom Plantation sector statistical pocket books, 1997: p. 108 and 1999: p. 1 17 and 2000: p. 1 14 (the values are for the summation of three Plantations Management Companies that grow onIy tea (extent in bearing); Maskeliya, Agarapatane, and Madulsima)
The figures From 1995 to 1999 are not representative enough to make inferences about the entire country. There are two arguments in the literature about the continuation of labor intensive
tea production.
(1). Disincentives in the large fams to invest in labor replacing technologies, and
(2). Agro-botonical (technical) reasons involving tea production
(1). Disincentives in the large tea fmsto invest in labor replacing technologies
Beckford (1972) argued that the essence of the plantation production systern is the
large labor force, fiequently brought fiom another country, and kept in captivity by
physical isolation fiom the rest of the native population. He argued that use of unfree
labor with an authoritarian structure of labor relations was the main spnng of the
plantation system. With the establishment of large tea farms in the eighteenth and
nineteenth centuries, about a million south Indian laborers migrated to Sn Lanka (see
Table 2.5). The resultant resident labor force was controlled and utilized with duress
preventing them fiom becoming free labor (Little, 1999, Bandarage, 1982, De Silva,
2982). The plantation system "perpetuated the use of a captive labor force long afier the
fonnal demise of slavery and the indenture system" (Little, 1999 p.34). The alienated
nature of plantation laborers fkom the surrounding peasant economy made provisions
such as housing, healthcare and food. a responsibility of the plantation management
(Bandarage, 1982). Once the plantations were established with their private, resident
labor force, which reproduced itself over time, future labor supply was guaranteed. The
investment in a captive, low wage labor was liable to be wasted by labor-displacing technology, thus, it is argued that during these low labor wage regirnes for the large fm laborers, labor-saving innovation was unnecessary (De Silva, 1982). However, the main weakness of this argument is that it does not explain the lack of labor replacing technological innovations even in the period of relatively higher labor wages with the labor unionization. The resident labor force was paid low wages relative to labor market wages until these expatriate laborers organized and started to bargain as a strong labor union. Since the late 1970s (with the increased political clout of the plantation labor unions), large fmlabor wages increased relative to diesel prices (see figure 4.2) and industrial sector labor wages (see figure 4.3). Yet, labor displacing technological improvements have not been observed in the large tea farms (see Table 4.1 and Table 4.2). Labor wages have increased during the governent ownership of the large tea fm. It could be argued that incentives for searching labor-replacing technologies were absent under the govemment ownership, compared to the private ownership. Moreover, employment generation was an important policy goal of the nationalization of large tea fms (Fernando, 1984), thus labor replacing technological innovations are inconsistent with such policy goals.
(2). Agro-botonical (technical) reasons involving tea production
Many argue that the elasticity of substitution between labor and capital is relatively small due to technical reasons in tea production. Sajhau and Muralt (1987) notes that tea along with rubber and coffee is not yet suitable for advanced mechanization especially at the harvesting stage. Selective hand picking of young shoots is hard to mechanize. This is the most labor-intensive activity irrespective of fmsize as it accounts for about 60% of total labor used for green leaf production on large farms and about 54% on small farm (see Table 4.4). It is argued that the present method of tea field establishment with the objective of complete ground cover from the canopy of tea bush is an inherent drawback to the adoption of large-scale mechanization. This method is recommended to prevent the rainfall-induced soil erosion in the steep, erosion-prone terrain and to manage weeds, by blocking the sunlight reaching the soil (Handbook of Tea 1986). In Sn Lanka, the undulating terrain of tea farms is yet another difficulty in adopting large-scale mechanization (Lamb 1954). Thus, other field practices such as pruning and fertilizer/pesticide application are unlikely to be mechanized.
However, main weakness of this argument is that it does not accommodate the role of relative input pnces in the labor capital substitution. It is unlikely that the agronomic limitations would govern the factor proportion irrespective of relative price changes of the factors. It is more likely that agronomic limitations would be over come with technological innovation if changes in relative factor price are large enough. It is not clear that why tea remains labor intensive crop. Yet, the continuation of labor intensive tea production is a fact.
4.1.B. Implication of Labor Intensive Production in Farm Size Distribution Changes
What is the implication of a labor-intensive production process on the changes in tea fam size distribution? In this section it is argued that 1) labor is the costliest factor in tea production, 2) labor requirements per unit land area for small and large fmsin green leaf tea production is sarne, 3) the oppomuiity costs of labor for large and small tea fms have been changing at different rates, 4) changes in relative labor wages cause differences in the profitability among fmcategories leading to size distribution changes. 4.1.B.l. Factor Shares in Green leaf Tea and Black Tea Production
The labor cost shares in the green leaf tea production and black tea processing are
presented in Table 4.3. The labor cost is about 50% (140.82/283.75) of the total costs of
green leaf production and black tea processing. The costliest activity is producing green
leaf tea. It is about 50% of the total cost (137.63/283.75). More importantly, the share of
labor cost is about 72% for the cost of green leaf production (98.94/137.63). The general
charges are usually for large fm overhead costs that include quasi-fixed costs
(supervisory staff salaries, various other benefits to the labors,) and sunk costs (rental
charges, taxes, insurance, depreciation). Therefore, in input cost shares, labor cost is the
single most important input cost in green leaf production.
Table 4.3. Labor Cost in Tea Production (Centdlbs)
Item TotalCost %ofTotalCost Laborcost % of (Centsnb) (Cen Wb) Labor Cost Green leaf production 137.63 48.5 98.94 70.25 Processing 68.29 24.0 12.77 9 .O General Charges 70.06 24.7 27.09 19.23 Transport 7.77 2.7 2.02 1 -4 Marketing 21.37 7.5 O O Total 283.75 ZOO 140.82 100 Source: Sarker 1978, p.4
4.1.B.2. Similar Labor Requirements in Large and Small Tea Farms
Labor requirement for green leaf production is same between large and small tea
fms(see Table 4.4). Green leaf harvesting uses the same amount of labor both in small and large tea farms and it uses up more than half of the total labor. Table 4.4. Labor Use for Green Leaf Production: Large and Small Tea Farms
(Figures are for annual requirement per acre) Activity Labor days YO Labor days % (large fms) (small farms) Harvesting (green leaf) 155 62 139 54 Weeding 45 19 72 28 Fertilizer Application/Forking 14 6 06 2 Pruning 6 3 5 2 Pest & Disease Control 8 3 - - Lopping Shades - - 04 O1 Green Leaf Transport - - 12 O5 Others 15 7 20 08 Total 243 100 258 100 Sources: for large fms, Liyanage, 1993, p. 82 (Table 4.1) For small fms, Femando1982, p.7 (Table 5)
4.1.B.3. Opportunity Cost of Labor among Different Size Tea Farms
The opportunity costs of labor differ for the two size classes. The opportunity
cost of labor for the owner-operator of a small farm depends on his forgone eamings and
this is proxied by the industnal worker's daily wage rate. Labor on large tea fms
(plantations) is unionized and wage rates are deterrnined through tripartite agreements
between the unions, govemment labor officials and management of the large tea farms
(ILO, 1966). Figure 4.3 shows the opportunity cost of labor for these two farm
categones. The difference is explained by the power of unionized labor on the plantations. The majority of plantation workers are south Indian tamils who migrated in the wake of plantation establishment. Figure 4.3. Minimum Labor Wages for Agriculture and Non-Agriculture Sectors
Workers in Agriculture: Tea Growing and Manufacturing, Rubber Growing and Manufacturing, Coconut Growing.
Workers in Industry: Baking, Tile Manufacturing, Coconut Manufacturing, Printing, Textile, Tire and Tube Manufacturing, Coir Mattresses, and Bristle export, Hoisely Manu fachiring, Engineering, Garment Manu facturing, Biscuit and Confectionery , Tea Export and Rubber Export trade only.
Sources: (1) Central Bank of Sri Lanka, Monthly Bulletin 1979 June, Annual Reports, 1990, 1998 (2) Peebles, 1982 (Sri Lanka: A Handbook of Historical Statistics p. 107) For instance, during the 1960s annual average plantation employment was about 574,500
workers of which around 432,300 (about 75%) were Lndian tamils (Peebles, 1982, p. 102).
They have organized into several trade unions and the largest is the Ceylon Workers
Congress. The political importance of the plantation Iabor force have been increased with
the constitutional changes took place in 1978. The govemrnent involvement in deciding
the labor wages of large hm has been criticized due to it detrimental impact on large
fmprofitability (Betz, 1989, Report of the Presidential Commission on Tea Industry
and Trade, 1995 p. 148).
The net result is that green leaf production remains a labor-intensive activity and
the increase in labor costs for large plantations relative to small fams has been a shift in
the relative profitability between the two types of fanns. The relative changes in labor
wages are argued to be an important economic factor that influences the farm size
distribution changes.
4.1.C. Relationship between Changes in Labor Costs and Farm Size Distribution
Suppose labor productivity (tea output per unit of labor input) is same for the
small and large tea farms and so is the output pnce they face. Suppose diminishing
marginal product of labor is applicable in tea production. Thus, unconstrained profit
maximizing labor use for both large and srna11 farms would be as follows;
MVP, = ,and MVP, = plSF
Where MVP, = marginal value product of labor ,P," = opportunity cost of labor for large fms,and cSF= opportunity cost of labor for small fms. The above, unconstrained profit maximizing condition, is depicted in figure 4.4. The marginal value product of labor is shown as MITl and it is assumed to be same for both fmsizes. The small farrn labor opportunity cost (per unit of labor) is shown as and the large fam labor opportunity cost is shown as (perfectly elastic labor supply is assumed in both sectors). The labor opportunïty cost (marginal factor cost) and marginal value product of labor detemine the size of tea output, keeping everything else constant.
The changes in relative labor wages lead to changes the relative amount of tea production in small and large fanns. Thus, the area under tea production for them should also change along with the changes in production. As the relative labor wages of the small fmsector decreases (compared to the large farm sector), their share of total tea production increases and so would tea area under the srnall fms (see Figure 4.4).
Therefore, as the relative labor wages in the small fmsector decreases, it is expected that the share of small tea fanns in total tea area increases.
Hypothesis 1: The ratio of opportzrnity cost of labor for small and large tea farms is negatively related to the shore of tea area cultivoted by the small farms. Large Farm Labor wages before Strong Labor Unions
-- - - -+ Smali Large Farm Fm Tea Production Output Output pF= Labor opportunity cost in large fmsector pF= Labor opportunity cost in small farms sector MV& = Marginal Value Product of labor in tea production
Large FmLabor wages with Strong Labor Unions
Large Fm Small Fm Output Output Tea Production
Figure 4.4. Marginal Value Product, Cost of Labor in SrnaIl& Large Farms 4.2 Effect of Public Programs and Policies on the Changes in Farm Sue Distribution
As discussed in Chapter 3, commodity programs, credit programs, tax policies,
and research and extension programs are public policies that have had an ambiguous
effect on farm size distribution changes in industrialized countries (see section 3.2.D).
This section argues that the important public policies for the Sn Lankan tea industry such
as land reform policies, small fami support services, export taxation polices, and
intervention in green leaf pricing have favored smaller, independent fams and have
contributed to the shift in production away fiom large plantations.
4.2.A. Land Reform Policies and Tea Farm Sue Distribution Changes in Sri Lanka
As discussed in detail in chapter 2, the Sri Lankan govenunent acquired about
377,000 acres of tea land (about 63% of totai tea area) in 1972 and 1975. Govemrnent
continued to manage the majority of these as large tea farms. What is the impact of this
land reform legislation on the changes in size distribution of tea fms? Land reform
policies would have affected the relative shares of tea area under small fmsand large
farm due to: 1) the redistribution of lands; and 2) bureaucratic inefficiencies in large fann
management. Each of these will now be discussed.
(1). Redistribution of land for the small farming sector
The relative availability of capital for acquiring land is an important determinant
of fmsize distribution as reviewed in chapter 3. It has been argued that large fams can
use their larger equity to leverage the financing of more land than a smaller farm could purchase (Lin, 1979). In the face of increasing land prices, only the large fmscould purchase more land, while small fms and prospective newcomers could not compete with large fms in financing such purchases (Tweeten, 1984). These size dependent, non-technological economies have been identified as pecuniary scale economies. It has been argued by development economists that land reform programs may extend the same benefits of pecuniary economies of scale (enjoyed by large famis) to small fms
(Barlow, 1978). For instance, if government redistributes adequate land to small fmers, this will overcome the barriers of capital requirements facing small fmers and will offset the pecuniary scale economies of large fms (Barlow and Jayasunya, 1986).
Many argue that government provision of land is one of the most important reasons for small fmexpansion in the developing world (Hayami 1994, Barlow and Jayasuriya
1986). This intent of such public policies and programs in countries like Sn Lanka to prornote srna11 fms is well-documented (Hayami 1994, Lelle and Agrawal 1989,
Tharian, Haridasan and Sreekumar 1988). Therefore, the provision of land under redistribution policies increases the relative share of small fms directly. If the redistribution has been achieved by fkagmenting the large fms, then there is also an indirect effect.
(2). Bureaucratic inefficiencies in large fmmanagement
While some of the land taken by the govenunent was distributed to small fmers, the majority of the large tea farms were maintained but operated by govemment officials.
The lack of managerial incentives under state controlled fmproduction and salaried managers is noted as an obstacle for efficient decision making and optimal resource use
Fernando, 1980, Betz, 1989). On the other hand, for small fms, whose resources are usually owned by the fmer, there is the motivation and incentive for efficient resource
use. Consequently, the relative share of fmsues could change with differences in the
relative incentives for productive efficiency. These two aspects wiil be examined in the
following section in context of land reform policies of Sri Lanka.
4.2.A.1. Land Redistribution and Changes in Tea Farm Size Distribution
The effects of land reform policies depend on how much land was redistributed
arnong small farms and how many large fms were fragrnented. The Land Reform
Commission acquired a total of 981,357 acres or about 23% of the total land arca under permanent agriculture in the country with the 1972 and 1975 enactment (see Table 4.5).
Of this total acquired land, about 38.4% was fiom tea plantations with approximately
12% fiom coconut plantations and 18% fiom rubber plantations. About 63% of the country's total tea area ended up in the administration of the Land Reform Commission by 1975.
The total of 981,000 acres acquired by Land Reform Commission was redistributed to public sector agencies, govemment-sponsored cooperatives, and to peasants for farnily fms or for residential purposes (Peiris, 1978, Fernando, 1980 and
Moore, 1985). By 1976, of the total of 981,000 acres of redistributed lands, more than two thirds (656,000 acres) were allocated to public sector agencies, a little more than a fourth (240,000 acres) was allocated to cooperatives. Only about 5% of land (52,000 acres) was distributed to peasants (see Table 4.6). Table 4.5. Land acquisition by Land Reform Commission: by crop type (acres)
Crop 1972 Law % 1975 amendment % Total YO Tea 139,354 24.7 237,592 56.8 376,946 38.4 Coconut 112,523 20.0 6,406 1.5 118,929 12.1 Rubber 82,563 14.7 94,835 22.7 177,398 18.1 Sub Total 334,440 59.4 338,833 81.0 673,273 68.6 Jungle & uncultivated 176,505 3 1-3 Other 52,455 9.3 79,124 19.0 Sub total 228,960 40.6 79,124 100 308,804 31.4 Grand Total 563,400 100 417,957 981,357 100 Source: modified fiom Fernando, 1980 p.208
The cooperatives failed as a post-reform management system, and as a result some of
these lands were given back to the public agencies, and some were redistributed among peasants confining the cooperative control to a mere 34,000 acres in 1978 (Fernando
1980, Peiris 1984). With this, by 1978, total allocation of the redistributed lands to the public sector agencies increased to about 83% and peasant allocation increased to about
11 % (1 09,000 acres).
It is not known how much of the 109,000 acres ended up with srnall tea fms
(Peins, 1984). But what is known is that out of the 83% of total land area allocated to public agencies, 69% or about 680,000 acres were in the hands of two state corporations
(State Ptantation Corporation and Janatha Estate Development Board). They managed al1 the plantation crops (tea, rubber, and coconut) as large farms with a resident labor force and bureaucratic management without "any noteworthy intemal changes in the system of plantation management" (Sirmatharnby and Devaraj 1987 p.288). Table 4.6. Recipients of Lands under the Land Reform Program ('000 acres) Recipient 1976 1978 Extent % Extent % Public Secfur Agencies State Plantation Corporation People's (Janatha) Estate Development Board Land Cornmissioner's Department Up Country Coop. Estate Development Board Estate Management Services Scheme Revenue Department District Land Reform Authority Tea and Rubber and Coconut Research Institute Live Stock Development Board Other Agencies Sub Total
Electoral cooperatives 185 18.8 - - Land Refom Cooperatives (Janawasa) 50 5 .O 3 1 3.1 Multipurpose Cooperative Societies 3 0.3 3 0.3 Special Cooperatives and others -7 0.2 - - Sub Total 240 24.3 34 3.4 Peasant Small Holdings 52 5.3 109 11.1 Others 31 3.2 20 2.0 Grand Total 981 100 981 100 (source: modified fiom Moore, 1985, p. 80)
In a recent review of the land reform program, Ranasinghe (1995) noted no important changes have taken place fiom the above redistribution pattern since 1978. About 95.5% of the large tea fmswere given to the state corporations (see Table 4.6). Both People's
Estate Development Board and State Plantations Corporation followed the same large tea farm management system in the post-reform period (Femando, 1980). Thus, the larger tea estates were not hgrnented and distributed in a widespread marner.
Table 4.7. Redistribution of Tea Land acquired with 1975 Amendment
Administering Institution Acres ?40 People's Estate Development Board 145,287 61.0 State Plantations Corporation 8 1,968 34.5 Land Reform Co-operatives 3,695 1.6 Tea Research institute 3,027 1.3 Up Country Cooperative Estate Development Board 2,3 10 1.O Special Co-operatives 835 0.4 District Land Reform Authonties 470 0.2 Total 237,592 100.0 Source: Femando, 1980 p.229 (this is as of mid 1976)
Fernando (1980, pp. 248-254) discusses a sarnple survey on the allocation of tea plantation land to villagers. in his 29 large tea farm sample, seven estates had to redistnbute cultivated tea land to villagers. The seven estates allocated about 11% (553 acres out of the total of 5 1 17 acres) of their area to peasants. He also reported another 17 estates (not included in his sarnple) with a total area of 15,80 1 acres had about 1093 acres or 7% allocated arnong villagers. Thus, the land refom policies redistributed only a limited amount of land from plantations.
The fragmentation of large tea farms and the associated redistribution arnong small holders was not encouraged due to the "prevailing notion that the fkagmentation of the estates would have adverse effects on their productivity" (Peins, t 984 p.39). Peins (1978) cited a report5, which claimed "estates when acquired by government and
divided up among peasants in smallholding resulted in rapid deterioration of agricultural
conditions". The International Labor Organization ~ission~to Sri Lanka claimed in
1970 that "for land holdings above 50 acres, the techno-economic case for a distributive
reform is weak [since] it is the bigger, not the smaller units that are higher yielding, more
intensive and better manager. The highest authonties in the tea industry too
discouraged the large fmfragmentation on account of the degradation the soils through
the lack of conservation practices used by small fmers (~amb', 1954).
Peiris (1 W8), summarized the prevailing view against fragmentation, by noting
"the claim made repeatedly during the past few decades that in plantation agriculture the conversion of estates to small holding is counterproductive and leads inevitably to a detenoration of agricultural standards has had an impact upon alienation policy relating to the acquired tea and rubber plantations, especially those located in the physically well- endowed areas" (p.621). Accordingly, the Land Reform Commission avoided both the hgmentation of plantations and the disruption of their former managerial and supervisory infrastructure. Therefore, the distributioîî of land among land-less peasants has hitherto been relatively unimportant in the overall pattern of land redistribution under the reform program.
What is the likelihood of the 11.5% of total redistributed land that was given to villagers ending up as small tea fms? There are references suggesting that peasants received lands that were more suitable for residential purposes than cultivation. In this account, Betz (1989) claimed that the "village expansion program", which redistributed
' Cited in Peiris, 1978 p.622: nie Six Year Program of Investment (Planning Secretariat: Colombo, 1955) 6 Cited in Peiris, 1978 p.622: Matching Employment Opportunities and Expectations, Geneva 1970. about 113,600 acres of nationalized estates was "entirely politically motivated, created
very srnall plots and was not accompanied by a thorough selection of beneficiaries nor by
credit programs or marketing supports" (p.53). Others such as Moore (1985), Peins
(1984), Gooneratna and Wesumperuma (1984) and Gold (1977) also noted that the
Iimited amount of land given to villagers under the land reform programs was Iow value,
marginal agricultural lands. Peiris (1984) argued that many of the marginal holdings
given to the villagers ended up in residential use rather than as tea gardens.
The above discussion highlights that a relatively small arnount of land was
redistributed to small holders and this land was a limited use for agricuttural purposes.
The acquired larger estates were given to public agencies to be managed as large units
rather than fiagmenting the land for redistribution. Hence, the relative shares of small
fmsand large fmswas likely not affected to any considerable degree through the land redistribution process.
4*2*A.2.Bureaucratic Management in Large Farms in Post-Reform Period
The effect of reallocating large estates fiom pnvate to public management on the size distribution depends on the success or failure of the bureaucratie management.
Fernando (1%O*), Sandaratne (1972), Peins (1 984), Rote (1 986), Betz (1989),
Sinnatharnby and Devaraj (1993), and Ranasinghe (1995) discussed the post-reform management and retated issues.
7 Director, Tea Research Institute of Sn Lanka 8 See Fernando, 1980 Chapter 6 "Impact of the Land Reform on Tea Production" p. 207-254. These discussions could be synthesized in two aspects; (1) initial post-reform
management institutions, including a) statutory trustees and b) cooperative (collectivist)
fmsand (2) state corporations' management of acquired large tea hms.
(1). Weaknesses in the initial post-reform management: statutory trustees
Soon after the Land Reform Commission acquired the lands, the previous
landowners and the agency houses were appointed as the statutory trustees of these lands
and were held responsible for the good and proper management of these lands. Despite
the various threats of litigation (Fernando, 1980 p.2 14), trustees neglected the proper
management of assets they did not personally own. While the statutory trustee
arrangement was designed to be in effect until new institutions were set up to take over
management fùnctions, the govenunent was compelled to cancel tnisteeship of all the
management agents as of 30 March 1976 (Fernando, 1980). The Land Reform
Commission was not adequately equipped, in terms of work force and technical expertise, to manage the lands that were acquired and the result was that large tea Farrns were severely neglected within this statutory trustee penod of about a year.
(II). Weaknesses in the initial post-reform management: cooperative (collectivist) farm
The cooperative (collectivist) style of management was attempted soon after the
land acquisition on some of the tea estates. Table 4.7 indicates that by 1975, about 23% of the tea land acquired in 1972 was managed by various cooperatives9 such as, land reform cooperatives and electoral cooperatives. This share increased to 32% by the end
Peins, 1978 p.620-621 discuss the establishment and huictioning of various cooperatives in estate management. of 1976. Large farms were converted into collective fmsand youths from various parts of the country were brought together and given land to cultivate on a cooperative basis.
The governent provided the inputs and facilities for marketing their products.
Fernando (1980) noted the cooperatives were shrouded with difficulties inherent in and complicated by outside political interference. The rigid rules of the CO-operativesystem1' when applied to estate management were not sufficiently flexible for quick decision making, which is an integral part of successful estate management. The Board of
Management of these cooperatives were political appointees" who oflen did not have estate management experience. Fernando (1980) (p.226) reported that in some instances tea yield of the cooperative tea fmswere 4540% lower than adjacent large tea farms of the State Plantation Corporation. The next elected government in mid 1977 dissolved the cooperatives and by 1978, they accounted for a mere 3.8% of the lands acquired in 1972
(Peiris 1984). AAer these cooperatives were dissolved, most of the properties rnanaged by them were handed over to the State Plantations Corporation and Peoples Estate
Developrnent Board (Peins 1984). This cm be seen From Table 4.8, as land area under
State Corporations increased to 87% in 1978 with the shuffling of allocations.
10 These rules designed to guide, control and administer consumer CO-ops,which distnbute basic consumer goods but were not suitable for the Iand management CO-ops(Fernando, 1980 p.224). 1 I Members of Parliament who represented the ruling party served as a chairman of the Board (Fernando, 1980 p.223-224). Table 4.8. Cooperative Involvement in Management of Tea Land (acres)
Administrative Institution 1972 940 1976 VO 1978 YO State Corporations 73870 53 N/A NIA 121154 87 Various Cooperatives Land Reform Cooperatives 8,152 5.8 6,399 4.6 5,3 15 3.8 Electoral Cooperatives 24,107 17.3 37,531 29.9 - - Subtotal 32,259 23. 43,930 34.5 5,315 3.8 District Land Reform Authorities 23,678 17.0 10,821 7.7 896 0.6 National Agriculhiral Diversification and - 11,927 8.6 Settlernent Authority Others 9,547 6.9 62 -- Total 139,354 100.0 139,354 100.0 139,354 100.0 Source: Fernando (1980) Table 6.4 (p.218) slightly modified.
(2). State corporations' management of acquired large tea farms
As of 1978, about 96% of large tea fârms acquired by the Land Reform
Commission (fiom both the 1972 and 1975 legislation) were managed by two public corporations; Peoples' Estate Development Board (about 61%) and State Plantations
Corporations (about 35%). They managed these fmsuntil 1992 when the management was assigned to private companies on a profit sharing basis with the ownership retained by the state. The management by these two public corporations had a considerable bearing on the size distribution changes in Sn Lankan tea farms.
Between 1978 and 1992 tea production by these two corporations declined by
19% and there were five years with negative profit margins (see section 2.3.D). The cost of producing black tea had doubled fiorn 3.00 (1952 constant Rs/kg) in 1972-76 (5 year average in privately-owned large tea fms, before land refonn) to 5.86 in 1986-91 (5 year average in state-owned large tea fms). in the re-privatization process in 1992, 52 large tea farrns were identified as non-viable or uneconornical to continue because they had losses in each of the previous 5 years and the yield of tea was below ZOO0 kg per hectare (World Bank 1997). These 52 large tea fmscovered about 47,640 acres or about 13% of the total tea land acquired with the land reform laws of 1972 and 1975.
Apart fiorn poor management, other factors affected the performance of large tea farms, including fluctuations in world market prices, labor unionization and the consequent increase in real wages, labor shortages, heavy taxation for plantation products, and severe weather conditions (Thorbecke and Svejnar, 1985). Rather than attempt to determine how much each of these was responsible for the decIine of large tea fms during the 15-year period since 1978, one can look into simple measures of management performance in following manner. The amount of administrative staff used by the two state corporations to mange the large tea farms increased from 4 staff per 100 labor in 1980 to 5.5 staff per 100 labor in 1990 (see Figure 4.5). However, with re- privatization in 1992, this dropped to about 4.5 in 1995. Similarly, staff per 100 hectares increased to about eight dunng 1980s but dropped sharply with re-privatization to about
5.5 per 100 ha. These two measures approximate the efficiency in management of the public corporations that controlled the large tea farms in 1976 to 1992.
Bandaranayake (1993) was critical about the management deficiencies in the two public corporations. Motivating the employees by rewarding success and punishing failure was absent. The deterioration in work incentives for management under state ownership aggravated the decline in the productivity of the large tea fmsand thereby affected the size distribution changes. Figure 4.5. Changes in Management Input: 1980 to 1998
Source: Plantation sector Statistical Pocket Book, Colombo: 1999 (p. 122, 123, 126) 4.2.A.3. Relationsbip between Land Reform Policies and Farm Sue Distribution
What would be the inferences that one can draw about the impact of land reform policies on the change in tea farm size distribution? The land refom policy seems to have increased the share of tea area under small tea fms. First, about 1 13,442 acres (about
11% of total land area acquired) were distributed among peasants and some of these would have entered into tea production as small tea farms. This would have increased the share of tea are under the small fmsdirectly. Second, during the post-land reform period, some of large tea farms managed by the state were poorly run and eventually were forced out of production due to continued losses. This would have increased the share of tea area under the small tea fms indirectly. To test the above assertion about the impact of land reform policies on the changes in tea fmsize distribution following hypothesis is developed;
Hypothesis 2: The share of tea area under the small farms is Iarger in the post-land reform period compared to the pre-land reform period. 4.2.B. Small Tea Farm Support Services and Size Distribution Changes
Large farms generally enjoy pecuniary and technological economies of size (Lh,
1968, Babb, 1979, Hayami, 1994). Pecuniary size economies include quantity discounts for larger levels of input procurement, premiurns for larger output marketing @ulk handling), and better terms fkom credit suppliers (lower credit administration costs).
Technological economies of scale exist in processing and transportation (Lim, 1968,
Etherington, 1971, Sarker, 1978, Hayarni, 1994).
However, the state provision of services such as credit, inputs (fertilizer), centralized processing, transportation have been the ways in which small tea growers can offset the advantages of scale economies for large farms (Etherington, 1971 ). Moreover, with the exception of fruits of publicly fûnded research, the state provisions of these services to large fams are uncornmon (Barlow and Jayasuriya, 1986). Even the benefits of research fiom the public sector tend to be scale neutral since these are largely agro- botanical and yield enhancing (De Silva, 1982, Tweeten, 1984). Thus, the govenunent provision for small fmers of service generally perceived to have economies of size fmers would improve their competitiveness relative to large farms and consequently increase the relative share of small fanns.
The chronological provision of support services for srna11 tea fmsby the state in
Sri Lanka is provided in Table 4.9. The next section examines whether the level of support intensity has changed for these services. In the initial period of the development of the tea industry before the 1930s, small tea farms received little attention fiom the state. In 1932, a small holder advisory service was started with the appointment of an officer by the Tea Research Institute (Gadd, 1932). In 1952, a scheme was started to provide small fàrmers with fkee seeds. Then in 1956, under the new settlement prograrn,
expenses involved in land clearing and soil conservation work were paid by the state.
This scheme financially assisted the small holders through Tea Producer Co-operat ive
Societies to purchase fertilizer and construct contour drains and terraces. The
cooperatives were supported with loans for organizing the collection and transportation of their green leaf to the factories.
Table 4.9. Chronological Information of Support Services for Small Tea Farm
Small tea farm support Functions Source of Information prograrn 1932: Small tea fmadvisocy Advising on conditions and Gadd, 1932 officer (appointed by Tea methods of tea cultivation Research Institute) 1952: Free seed provision Facilitating access to planting Report of the cornmittee on small materials holder problem, 1980 1954/55 to 1959160: Rehabilitation, Pest control and Six Year Program of Investment Loan grants, through Tea soi1 erosion control and fertilizer 1954/55 to 1959160, Planning Producers Cooperative Societies application Secretariat, p. 62 1956: High Land Senlement Land clearing expenses, soi1 Report of the cornmittee on small Program conservation work expenses hoider problem, 1980 covered by the state 1958:Rehabilitation Scheme Improving agricultural conditions Ceylon National Development continued till 1966 by fertilizer application, adoption Plan, p.7 of soil conservation measures, infilling 1966: Fertilizer Subsidy scheme Subsidized fertilizer Ceylon Year Book, 1970 p.71 continued till 1970 1977:Subsidy scheme for new To start new small tea farms Report of the cornmittee on srnall planting holder problem, 1980 1979:Subsidy for infilling To cover the bush vacancies in Report of the cornmittee on small srnall tea farms holder problem, 1980 in 1958, with the Tea Rehabilitation Scheme, assistance was provided to improve agicultural conditions of the small fms, by fertilizer application, and support for the establishment of shade trees and green manure. The support was extended as a subsidy in
1966 under the Tea Fertilizer Subsidy Scheme. The Tea Small Holding Development
Authority was established in 1975 with the increasing importance of the tea small holding sector. The adrninistering support services for the srnall tea farrns were totally assigned to the Tea Small Holding Development Authority. By 1977 and 1979 the subsidy was extended to new planting and the vacant areas with new tea plants (Report of the cornmittee on tea smallholding problems, 1980).
Presently, small tea fmsare supported by a subsidy for new planting, replanting and filling tea bush vacancies. The new planting subsidy is Rs 36,000 per hectare.
Replanting subsidies Vary according to the elevation class with High and Mid Grown areas receiving Rs 110, 000 and the Low Grown areas receiving Rs 100,000 per hectare
(Plantation Sector Statistical Pocket Book).
4.t.B. 1. Relationship between Small Farm Supports and Farm Size Distribution
The total subsidies and unit land area subsidies (sum of different type subsidies) for small tea farrns in Sn Lanka from 1960 to 1998 shown in Figure 4.6. Although the reliability of the data may be questioned (see Annex 6C in Chapter 6) the close association between total subsidies provided and subsidies received per unit land area indicates that whoever requested the subsidies received it. The importance of this fact is that, unless constrained by eligibility criteria (such as soi1 condition of the land, 1 +Subsidies pa unit bnd area (Corrstarit 1952) +Total Subsidg (Constant 1952 Rs-Mn) /
Figure 4.6. Support services for small tea farms in Sri Lanka
Source: see Annex 6C in Chapter 6. gradient of the land), the maIl famis were assured of the subsidies. Thus, the availability of subsidies would have improved the access to the resources for small farms offsetting the pecuniary scale economies of the large fms. Since, the subsidies were made available only for srnall famis, the subsidies would have been expected to increase the share of tea area under small fms. It is expected that the larger the subsidies for small fitrms (ceteris paribus) the larger would be the share of tea area under small tea fms.
This expected relationship is developed into the following hypothesis;
Hypothesis 3: The srnail farm subsidy per unit fand area is positive& related to the share of tea area under srnail fanns. 132
4.2.C. Tea Industry Taxation Policy and Farm Size Distribution Changes
The effects of taxation policy on structural changes generally have been due to the differential tax rates and the availability of tax credits (see chapter 3). The important tax policy in the Sri Lankan tea industry was tea export taxes. This section begins with a brief description of how export duties on tea evolved. The impact of these duties on exporters and producers is then discussed specificaIly in relation to the derived demand for black tea from the exporters and the short run and long run supply responses of the tea producers. Finally, the hypothesized effects of the export duties on the changes in farm size distribution are proposed.
4.2.C.1. Nature and Evolvement of Tea Industry Taxation in Sri Lanka
The primary reason for export crop taxes is to raise govemment revenue (Rote,
1986). The colonial govement introduced export duties in 1915 as a means to collect additional revenue to fund the World War 1 effort. Although the tax continued throughout the colonial period, the tax burden was insignificant (Rarnachandran 1963).
Changes occurred in export taxation policies when Sri Lanka becarne an independent state in 1948. The profits generated in British-owned large farms or plantations were identified as a possible source of govemment revenue, which would enable government to maintain and expand services, including social and welfare services such as health and education (Tiffien and Mortimore, 1990). Since the first elected govemment took office in independent Sri Lanka in 1952, the budget deficit has grown as illustrated in Figure
4.9. The post 1956 penod has been characterized by "rapidly increasing social expenditures and later the state-centered industrialization program" (Betz 1989, p.55). Figure 4.7. Covernment Revenue and Expenditure in Sri Lanka
Source: Peebles 1982 (Sri Lanka: A Handbook of Historical Statistics, p.233 & 234) deflated by CPI (WS2=lOO).
Notes: The deficits were financed by resort to extemal assistance and partly by increased reliance on deficit finance (see, Singh, 1972) The ever-increasing govemment expenditures on health, education, food subsidies and infiastructure had to be financed through taxes on its major export crops, since the tax base of the country was small in this era12 (Rote, 1986, Thorbecke and Svejnar, 1987).
The wealth obtained from tree crop exports was seen as a justification for, and the basis of, food subsidies (Bhalla, 1991).
The average export duty of 35 cents per pound of black tea exported in 1947 was increased to 69 cents per pound by the end of 1958 while the CoIombo auction price of tea only increased from 1.60 (RsAb) in 1947 to 1.73 in 1958 (ITC Bulletin, 1960). The export tax earnings fiom tea sometimes contributed up to 28% of total govermnent revenue" and generally was about 8% of total government revenue. in the period of the
1960s, about 20% of the export earnings kom tea were collected as export taxes and by the latter part of the 1970s and 1!%Os, these increased considerably (see Figure 4.10).
The export duties on Sn Lankan tea were considerably higher than the taxes in other major tea producing countries (see Table 4.10). Consequently, the after tax profit margin of Sn Lankan tea producers was the second lowest among the major exporters. The export duty alone snatched away about 40% of the pre-tax profit margin. Even after the
British ownenhip of the large estates was terminated in the mid 1970s, export tax policies were unchanged. Tea taxes went up to more than 50% of the tea export earnings with the currency depreciation in 1977 (see Figure 4.8).
" Snodgrass 1966, Table A-65 shows the composition of government revenue in this period. Corporate plus individual incorne tax were only about 17% of the governrnent revenue while export duties contributed about 28% of the government revenue. Details are given in Snodgrass 1966 p. 187- 192 13 In 1978, total governrnent revenue of about Rs. million 1 16,000 and tea export taxes was about Rs. million 3,300 (Annual Report , 1980 Central Bank of Sri Lanka). This was the ever-highest export tax earning fiom tea. Figure 4.8. Tea Tax as a Percentage of Tea Export Earnings
Source: Annual Report Central Bank of Sri Lanka (Various Issues) Table 4.10. Total Costs, DutiedTaxes, and Profits among World Tea Producers
(in 196û-70, U.S cents per pound)
N. India S. India Sri Lanka Uganda Kenya Malawi Tanzania Total Cost 35.37 30.62 32.55 35.91 32.61 31.00 36.16 London Price 48.07 42.26 49.30 50.98 52.98 43.41 52.80 Pre-Tax Margin 12.70 11.64 16.75 15.07 20.37 12.41 16.64 Duties and Taxes cesse 0.24 0.24 1.56 0.28 0.28 0.33 0.10 Excise Duty 3 .O9 2.35 - - - - - Export Duty 4.75 3.70 6.72 - - - 1 -40 Sales Tax - - 3.34 - - - - Total Duties/taxes 8.08 6.29 11.62 0.28 0.28 0.33 1.50 Post Tax Margin 4.62 5.32 5.13 14.79 20.09 12.08 15.14 Source: Report of the World Tea Economy, Report No. EC-178 (IBRD 197 1, p. 38) Notes ( 1). The authors indicated that the information must be used as a guide since data were not gathered fiom representative sample. (2). Total cost included al1 the costs fiom selling point plus Freight, shipping, brokerage, commissions and insurance. In UK,landing charges, warehousing, insurance and brokerage. (3). cesse is a fwid that is collected for tea industry related financing (tea promotion, research and subsidies etc.,) (4). Information about Uganda, Kenya, Malawi, and Tanzania is for 1970 while India and Sri Lanka for 1968.
Export taxation came to end in 1992 with the hand over of state-owned large tea fms hand over to private management companies. The export duties depressed the profits accruable to the tea industry thereby constraining long-term investments in the industry
(Perera 1976, Thorbecke and Svejnar, 1987, Sinnathamby and Devaraj 1987, Betz 1989).
However, it is important to isolate the impact of the export duties on tea producers, compared to the tea exporters since these two groups are distinctly different in the Sn
Lankan tea industry. The impact of export duties on the producers' price (hence profits) depends on the price elasticity of demand for their product by the tea exporters. The
more inelastic the demand by the exporters, the less will be the impact of the export duty
on producer prices.
4.2.C.2. The Effect of Tea Export Taxes on the Price received by the Tea Producers
World tea trade is organized in such a manner that tea producers face an elastic
denved demand for their bulk teas (primary produce). The countries in which tea is grown and primarily processed (green leaf to black tea) for bulk export are distinct Çom the countries in which these teas are imported, fûrther processed (blending and packing of black tea), and sold to consumers in the retail market. The former consists of mostly the south Asian and east Afiican countries while the latter is made up of developed western countries (United States, United Kingdom, and Canada). Primary tea producers (green leaf to black tea in bulk form) are not organized as a cartel and each country independently decides upon its own tea production14. in contrast, the blenders, and packers, who process the bulk tea for consumable fom, is made up of a few multinational corporations with considerable market concentration (UNCTAD, 1984 p.21).
The role played by these secondary processors (blenders and packers) in the world tea market is vital. Tea as a primary product (in bulk form) is a heterogeneous commodity. Its quality varies by tea farm, region, elevation, climate, and season as well as by field practices and processing methods (Majumdar, 1977).
14 The first International Tea Agreement was in effect from 1933 to 1955. Suice then, there were unsuccessful atternpts to re-established the Agreement. See Goonewardene, 1976, Etherington and Jones, 1976, and Edwards, 1976. However, the tea consumer requires a product uniform in quality throughout the year.
The blenders convert the heterogeneous primary product to the homogeneous consurnable product. The general pattern conceming blenders and packers is for a hi& and increasing concentration of their activities in the hands of a small number of transnational
companies (UNCTAD, 1982). Close to 65% of ready to consume tea trade in major tea- consuming countries is controlled by Unilivers (formerly Brooke Bond and Lipton, 35%),
Nestle/Twining (6.8%)' Lyons Tetley (1 3%), and Typhoo/Schweppes (10%) (Dawood,
1980 p.83). in 11 of the 20 major tea consurning countries, these four firms alone
accounted for about 80% of the market for tea while, in the other nine countries the
degree of concentration was about 60% (see Table 4.1 1).
In each of the major producing countries, these multinational corporations have
agents to buy bulk tea in auctions. Tea auctions in Colombo (Sri Lanka), Mombassa
(Kenya), and Cochin and Calcutta (India) provide about 56% of the world's total tea
exports of about 1556 million kilos (ITC Bulletin, 1996). Table 4.12 shows the concentration ratios of the four buyers in each of these three (Culcutta and Cochin
together) auctions.
The concentration in both the retail market (ready to consume tea) and primary product markets (bulk tea exports) do not prove collusive behavior but do indicate the possibility in controlling the prices received by the primary producers. The standard
practice by blenders and packers is to provide price instructions to local buying agents at
the auction centers (Presidential Report on the Tea industry and Trade, 1995). The
possibility of controlling the purchase of bulk tea in different auction centers of producing countries enables the exporters to switch between their suppliers. Table 4.11. Concentration Ratio of Tea Sales by Major Firms in Retail Markets
Country Year % of sales No, of Firms Austria Australia Canada Denmark France Germany (west) India (a) Ital y ireland Japan Mexico Netherlands Nigeria (a) Norway Pakistan (a) Philippines Portugal Sweden United Kingdom United States 1973 74 3 Source:UNCTAD, 1984, p.21 {(a) packet tea only), Total of 4 fims as discussed above
Table 4.12. Four-Buyers Concentration Ratio in three Major Tea Auctions
Buyer Sn Lanka (1993) Kenya (1981) india (1979/80)
Mn Kilos YO Mn Kilos % Mn Kilos '%O Buyer 1 22.6 9.94 28.1 1 15.8 14.87 8.4 Buyer 2 19.0 8.38 21.89 12.3 13.52 7.6 Buyer 3 18.5 8.16 2 1.O8 12.2 12.78 7.2 Buyer 4 16.6 7.30 20.54 11.6 7.77 4.4 Total 76.7 33.78 92.22 51.9 48.94 27.6 Sources: Sri Lanka; Report of Presidential Commission on Tea industry Trade, 1995 p.33 Kenya and India; Rote 1986, p. 187 The substitutability of teas fiom different tea producing regionsl' in the blending and other secondary processing is also helphil for the exporten to shuffle between pnmary producers (Majumdar, 1977). More importantly, if these exporters were subjected to larger export duties in any country they would switch to another country with lower export taxes keeping everything else constant. Thus, if prices in the Colombo auction for comparable teas are consistently far more (due to larger export taxes) than those of other auctions, exporters will tend to move into other auctions.
The tea export taxation in Sn Lanka and Kenya show that both the average tax rate (percentage of export tax from the F.0.B pnce) and marginal tax rate (increment in tax rate per unit increase in tea prices) have been considerably higher in Sn Lanka relative to Kenya (see Table 4.13).
Table 4.13. Export Taxation in Tea: Kenya and Sri Lanka, 1985
FOB price($/kg) Kenyan Tax Rate(% fiom price) Sri Lankan Tax Rate (% from orice). Average Marginal Average Marginal 1.20 0.00 O 22.4 O
Source: World Bank, 1986 (World Development Report, p.76)
'' For instance, darker brewing quality reputed to be with Kenyan tea and with Sri Lankan Low Grown tea. (See Forrest 1985, (p.20-97). It has been argued that this moderate export taxation in Kenya is a possible reason for its rapid expansion in world export share. The Kenyan share of world exports increased fiom 3% in 1960-63 to 9% by 1980-82, while the Sn Lankan export share dropped fkom
33% to about 20% during the sarne period (World Bank, 1986 p.76). The positive association between the Sri Lankan export tax rate in tea (1952 Rslkg) and the relative increase in Kenyan (Mombassa) tea auction sales against Sri Lankan (Colombo) tea auction sales strenghens the above argument (see Figure 4.9).
The centralized operations of large multinationals in shipping, warehousing, blending, packaging and retailing allows for the possibility of resorting to the strategies to buy the bulk tea fiom the primary producers at the lowest possible prie (Rote, 1986).
Therefore, given the above structure of the world tea market, producers of bulk tea face an elastic derived demand by exporters.
4.2.C.3. How would Export Duties Affect Different Farm Sues Differently?
The extent to which an export tax is borne by the primary producers mainly depends upon the price elasticity of derived demand, which is arguably elastic as discussed in the previous section, and the price elasticity of supply by primary producers.
The elasticity of tea supply to own price is small as evidenced by the estimated short- tenn elasticities fiom previous studies (see Table 4.14). In view of the gestation pet-iod of about 6 years fiom newly planted tea bushes to first commercial harvest, tea production tends to be price-inelastic in the short nui. Therefore, it is argued that taxation does not affect the short-run output of either large or small farms (Bhalla, 1991). Thus, export taxes could be conveniently passed to primary producers in the short-mn. Sri bnkan Tea Export Tax Rate (1952 Rs/kg)
Figure 4.9. Relative changes in Kenyan Tea Exports and Sri Lankan Tea Taxes
Sources: Sri Lankan tea export tax rate: see Annex 6D Colombo and Mombassa auction sales are fi-orn International Tea Committee, Annual Bulletin of Statistics, various issues Tabte 4.14. Price Elasticity of Supply among different Countries in Tea Production
Source Response Short rua Price Long run Elasticity E las ticity
Akiyama and Trivedi ( 1987) Black Tea 0.04 (price lagged- - for 3 N/A Sri Lnkan tea industry years)
Akiyama and Trivedi ( 1987) B lac k Tea 0.1 2 (price lagged- - for 3 N/A indian tea industry years) Akiyarna and Trivedi ( 1987) Black Tea 0.1 7-0.40 (Adjusted N/A Kenyan tea industry (srna11 payment for green leaf) holders) Akiyama and Trivedi (1987) Black Tea 0.26-0.86 Kenyan tea industry (large farms) Askari and Cumrnings ( 1976) Black Tea 0.264 (for last years price) N/A (~.259)Indian Tea industrv Askari and Cummings (1976) Black Tea 0.2 12 (for last years price) N/A (p.259) Indian tea industry Ramanujam ( 1986) Sri Lanka Black Tea 0.0408 India Black Tea 0.1536 Kenya Black Tea O. 163 1 Malawi Black Tea 0.1058 Indonesia Black Tea 0.0724 Mozambique Black Tea 0.073 1 Tanzania Black Tea 0.0479 0.443 1
Adan and Rajagopalan,--- ( 1989) Acreage 0.0707 (Tamil Nadu) 0.2664 (Tamil Nadu) SOU~~Indian tea indus~ ~creaie 0.0653 (Kerala) 3.9~76'~(Kerala) N/A= not available
Thus, in the short-run, the elastic denved demand for bulk tea from tea exporters and the inelastic supply of the primary producen means a disproportionately large share of the export tax burden is passed on to primary producers (see Figure 4.10). Producer profits decline.
l6Ajjan and Rajgopalan noted " compared to Tamil Nadu, the long-run elasticity of acreage response to price changes was significantly very large in Kerala, This was not totally unexpected because in Kerala the area under tea in recent years had shown erratic variation in response to policy of growving different plantation crops" (p.71). The authors noted that in previous studies the estimates for long-nui supply elasticities for tea were in the range of 0.02 to 0.72 (p.71). Price of bulk tea / Supply of bulk tea
Derived-demand for bulk tea w Q' Q* Quantity of bulk tea
Figure 4.10 Nature of Export Taxes Burden on Primary Producers
P' = Market equilibriurn bulk tea price before export taxation Q' = Market equilibriurn bulk tea quantity before export taxation E' = Specific Export Tax PL= Market equilibrium bulk tea price afier export taxation Q' = Market equilibrium bulk tea quantity after export taxation P'+E, = Exporter's price with the export tax of Et (per kg of bulk tea) P'-E, = Producer's pnce with the export tax of Et (per kg of bulk tea)
Exporter's price increased by {(Pt+Et)- P'} Producer's price decreased by (P' - (Pt-Et)) Over time, long run supply elasticities corne into play, and the decline in profits leads to a decline in output and acreage. How would the long run elasticities Vary for the mal1 and large f'arm? It is argued below that the long run supply elasticities of small farms are greater than those for large fms. Thus, with larger export taxes and resulting lower profits, there would be fewer small fanns.
The supply elasticities from the previous studies are summarized in Table 4.14.
The values confinn the general notion of larger long run elasticities relative to the short run elasticities (Le Chatelier Principle). However, the distinction between small farms and large fmlong-run supply elasticities is not apparent fiom Table 4.14. It is assumed here that large fmshave a relatively less elastic long run supply due to the production organization in such famis. Saker (1972) noted that large tea farms require a large arnount of skilled, permanent labor and a large land and capital asset base which are
'sunk costs' which do not have much alternative economic use. The possibility of diversion of these resources to alternative enterprises is limited. The size of 'sunk costs' for the large fmsare larger due to investments in labor housing, and other infiastructure
(roads and other structures in the large fms)provided in such fms. For example, there is no other crop that cm be prûcessed with the tea processing equipment. The production changes requiring the changes in these sizable sunk costs is relatively slow compared to the flexibility of smaller farmer.
in contrast, small famis do not invest in processing facilities nor do they provide housing and other facilities (medical, educational and social) that large tea farms must provide for their labor (Etherington, 1971). in commenting on the differences of supply responses between the Kenyan small and large tea farms, Akiyama and Tnvedi (1987) showed that the small fmers responded to relative price increases in the long nin by
more entry and expansion of their production area. Large tea farms responded with a
greater use of yield-increasing agricultural practices since their expansion in the
production area is limited. Akiyama and Tnvedi noted "given high adjustment costs of
fixed inputs (for large fms) this may be an optimal response" (p. 156). Since the sunk
costs for small farmers are much less relative to large farms, a larger export taxes is more
likely to cause small fams to "pull up the tea bushes", and convert into other enterprises
compared to large farms whose shut down costs are generally much higher (Sarker,
1972).
Export taxes depress a tea producer's profits regardless of size. It is argued that
the downward adjustment of supply for small tea farms is relatively easier than for large
farms. Thus, small farms tend to exit fiom tea cultivation relatively easily in the instances of lowered profits since they incur less costs in shutting down tea production compared to
large fms. Thus, it is hypothesized that the larger the export duties on tea ceteris paribus the smaller will be the share of smalI tea fanns in the total tea area.
Hypothesis 4: The size of the tea expon tares is negatively relared to the share of small tes farms in total tea area. 4.2.D. Government Price Intervention in Green Leaf Tea Market
Govemment intervention in agricultural markets has been justified on the basis of
stabilizing producer prices, supporting fmincome, protecting consumers from high
prices, reducing the dependency on imports, and conserving foreign exchange (Tomek
and Robison, 1990). The effects of govertunent intervention on fmsize distribution
depend on the type of policy (see chapter 3). To assess the role of price intervention
policy in Sn Lanka, a brief review on the evolution of pricing intervention in the green
leaf tea market is provided.
4.2.D.1. Evolvement of Green Leaf Tea Price Intervention in Sri Lanka
The pricing intervention in green leaf tea market started in 1968 (Report of the
Presidential Commission on the Tea industry and Trade, 1995). Price information on green leaf tea before 1968 is not available in published sources. However, researchers have shared a similar view about green leaf prices during this period (see section 5.4.
Cl). The view was that green leaf prices were inefficiently low to the level that small,
independent green leaf producers shied away from green leaf tea production.
Gadd argued "There can be very little profit for the grower (small fm) when good quality village leaf sells for 3 cents per pound, but, when the price falls to 2 cents or less (.) there can be no profit at al1 after his pluckers and others charges on his holding have been paid" (p. 89). Assuming a 4.5 conversion ratio (4.5 pounds green leaf to produce 1 pounds of black tea), the green leaf price, as reported by Gadd would translate into price of black tea of 9 cents (2~4.5)to 13.5 cents (3x43 per pound. The black tea price in Colombo auction in this tirne was about 66 cents per pound (ITC Bulletin). Thus, the share of payments for green leaf would be 13% (9+66) and 20% (1 3.5+66) of the price of black tea. The question is whether this share was adequate to compensate the green leaf producers for the opportunity cost of their resources in green leaf production.
According to Gadd, when the share for the green leaf is 13% of the black tea price, the small farmer receive no profits. The green leafprices received by small fmers dropped due to a large drop in black tea prices during the 1960s (see Table 4.15). The real pnce of black tea fell about 50% in the Low Grown region, where the highest concentration of small tea hswas found, and about 30% in Mid Grown and High Grown regions.
Table 4.15. Decline in Real Black Tea Price Index in 1960s (1960=100)
Year High Mid Low 1960 1O0 1O0 1O0 1961 97 96 92 1962 98 89 81 1963 89 82 77 1964 88 83 73 1965 86 86 80 1966 83 76 66 1967 80 75 57 1968 80 81 73 1969 70 63 57 1970 71 67 55 Source: ITC Bulletin various issues in an effort to improve the low relative prices for green leaf tea, the Sri Lankan
govemment intervened in 1968 and implemented the following price formula (al1 the
components are expressed in unit weight basis);
1 Black tea price - {cost of processing + processor' s profit) Pnce of green leaf (Rskg) = 4.5
The Tea Controller Department determined and declared the relevant values for the cost of processing (Rs/kg of black tea) and the processorTsprofits (Rs/kg of black tea). The divisor of (4.5) is the weight of green leaf required to make a unit weight of black tea.
(Report of the Cornmittee on Tea Small Holdings Sector, 1980 p. 17). The above formula ensured profits to processors but did not necessarily correct for the low green leaf prices to fmers. Small holders objected to the above formula and it was modified in 1978 to include a guaranteed minimum price for green leaf. However? the inflation levels Sri
Lanka experienced afler 1978 negated the effectiveness of the minimum price (see Table
4.16).
Table 4.16. Changes in Colombo Consumer Price Index
Year 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 CCPI' 198.3 200.7 203.7 227.3 257.2 318.2 375.5 416.1 474.2 553.1 %A - 1.2 1.5 11.6 13.1 23.2 18.0 10.8 14.0 16.6 Source: Annual Reports Central Bank of Sri Lanka Colombo consumer Price Index
The consumer price index almost tripled within 7 years and consequently there were several revisions to the guaranteed minimum price since 1978 as shown in Table 4.17. Table 4-17. Changes in Guaranteed Minimum Price for Green Leaf Ten
Effective Date Green Leaf Price (Rskg) 1'' October 1978 1.54 ln Apnl 1979 1.87
1" November 1979 2.20 1'' March 1981 2.53 lstJanuary 1952 2.86 ln May 1982 3.10 1983 3.10 Source: Annual Report Central Bank of Sn Lanka 1983 p.3 1, Fernando 1982, p. 12
During this period the government also revised the allowable processing costs and profits (Annual Report, Central Bank of Sn Lankan 1983). The net result is that the establishment of an effective minimum price beginning in 1978 allowed small producers to cover their costs. The minimum pnces were proportionately increased tallying with the inflation level. In 1984, the guaranteed minimum pnce was eliminated by the
"Reasonable Pncing Formula" which split the black tea price with 25% going to the processor and 75% to the green leaf supplier. in 1985, the shares changed to 30% and
70% and in 1987 to 32% and 68%. This formula along with the green leaf to black tea conversion at 4.5 (or 22.22%) has been in effect to date in detemining the green leaf price in Sri Lankan tea industry. For example, if a processor received Rs 100 per kg of black tea from the auction he has to pay for his green leaf suppliers Rs 15.1 1 per kg of green leaf (100~0.68~0.2222).The green leaf pnce is crucially dependent on the conversion factor and instead of 22.22% if the conversion factor is taken at 2 1% (1 00 kg of green leaf to produce 2 1 kg of black tea) the green leaf price would be Rs 14.28
(1OOxO.68xO.î 1) per kg of green leaf.
4.2.D.2. Inefficiently low Green Leaf prices before 1968 period
If the green leaf prices were inadequate and returns from green leaf tea production
were lower relative to the returns fiom other opportunities, small farmers would not
invest in producing green leaf tea. The explanations for such inefficiently low prices,
which would prevent mutual beneficial exchange for small fmers and processors, is
explored in a model of a hold-up problem in Chapter 5 (section 5.4). It is argued there
that pncing intervention restricts the processors fkom engaging in inefficient hold up
behavior, thereby green leaf production and processing by vertically independent small
fmsand processors move to an efficient equilibrium (see Figure 5.5 to 5.7).
in supporting the claims of inefficiently low green leaf prices, three points are
used in the hold up problem model discussed in Chapter 5. First, in the 1960s, about 96%
of the processors had their own green leaf supply. There were only a few independent
processors who were purchasing green leaf fkom small fanners during this period (see
Table 5.2). Thus, the majority of the processors were not dependent on small farm green
leaf supply for their processing needs. in 1951, about 96% of processors were large fm processors who grew their own green leaf. This percentage had slightly decreased to
92% by 1972. These processors did not depend on small farrners their for green leaf tea but the small farms were dependent on such processors since independent processors represented only 4% to 8% of the total nurnber of processors. Second, even with an independent processor, the small fmer was in a weak bargaining position given the difficulty in hauling his produce to another independent processor. One can calculate the area of a given elevation region and the number of independent processors that was available for that area dunng the 1960s. in 1951, for the
High Grown elevation class", there were only three independent processors within 1563 square miles. In the region of Mid Grown, there were 17 independent processors for an area of 2326 square miles and in the Low Grown region, there were 22 independent processors in an area of 2384 square miles.
Third, the bargaining power of the small hersdepends on the condition of the transportation infrastructure within these areas. A survey on small tea fms in 1994 found that about 48% of the small fmswere located in a locality without road access and about 27% of these were half a mile away fiom a road (Report of the Baseline
Survey of Tea Small Holdings, 1994). Chaudhuri (1978) referring to indian tea industry argued that the reason to have processing facilities within the plantations is "(green) tea leaves are weight-loosing and perishable in nature and cannot be transported to long distance for processing" (p.29). The combination of a large majonty of processors who were not dependent on green leaf from small farms and the difficulties in moving their product between available processors, lead to a weak bargaining position for small fmers and low green leaf pnces.
The intervention by the Sri Lankan government beginning in late 1960s has removed the hold up threats in the investments in green leaf production and raised the
17 Land area of elevation classes; High Grown = Sq miles 1563 (NuwaraEliya 474 + Badulla 1089), Mid Grown = 2326 (Kandy 914tMatale 770.2+ Kegalle 642), Low Grown= 2384 (Galle 652+Matara 48 1 +Ratnapura 1250) price of green leaf tea received by producers. The increase has benefited largely
independent suppliers who tend to be small famiers. The rernoval of the hold up problem
in the green leaf market through pnce intervention does not affect the large tea farms as
much since they do not usually sel1 their green leaf tea but process it with their own processing facilities (see Chapter 2 section 2.2). Thus, pricing intervention is likely to
increase small fams relative to large farms and thereby change the tea farm size distribution. Therefore, it is hypothesized that the;
Hypothesis 5: The share of srnaff tea farms in the total tea area is Iarger during the period of price intervention in green leaf market, relative to the period without the price
intervention in green leaf markel. 4.3. Export Market Price Instability and Changes in Tea Farm Sue Distribution
The short-term fluctuations in the pnces of primary cornmodities have been noted as an important problem in export-based economies (Maizels, Bacon, and Mavrotas,
1997, and Athukorala, 1984). The fluctuations in earnings of producing countries fiom their exports can have a major impact on incornes, employment, and living standards in these countries. The analyses of price movements in primary commodities tend to examine changes in the averages (trend) pnce and the nature of the short- term (year to year or within year) oscillation of prices around that average (trend) price (Athukorala,
1984). The focus here is about the impact of price instability on the changes in fmsize distribution. The impact of changes in price and hence the profitability of tea production relative to other enterprises on the farm size distribution changes is discussed in the
Section 4.4.
The degree of price instability experienced by an export comrnodity is determined by two major factors: the degree of price elasticity of export demand and supply, and the fiequency and the magnitude of the shiAs in supply and demand. The more inelastic the demand and supply curves, and the more fiequent their periodic shift, the higher the degree of pnce instability (Athukorala, 1984).
The price instability index, which is calculated as the mean annual percentage deviation fiom an exponential trend of the pice series, for tropicai beverages and al1 commodities is shown in Table 4.18. The variation in prices for tropical beverages is greater than that of the other cornmodities. While price instability in the world tea market has varied substantially during the last few decades, it has declined generally (Athukorala and Huynh, 1987). The oscillation of tea prices, on average, has been about 20% higher or lower fiom the exponential price trend during the period of 1962 to 1986 (see Table
4.18). This oscillation has been reduced by about half during the penod of 1986 to 199 1.
Table 4.18. Price Instability Index for Tropical Beverages and al1 Commodities
Price lnstabiGty Index Penod Tropical Beverages Al1 Commodities Cocoa Coffee Tea Total 1962- 1980 27.7 28.4 20.9 25.5 15.2 1980-1986 11.1 11.8 19.3 10.4 8.2 1986-1991 9.1 16.8 9.9 10.0 8.5 Source: Maizels, Bacon, and Mavrotas, 1997 p.23
Such fluctuation in prices does not necessarily mean the relative profitability of
tea has gone down compared to alternative enterprises, and a need for switching to such
alternative enterprises. Yet, the fluctuations have an impact through the risk preferences of fmers in the selection of their crop mix. For instance, when farmers are risk averse they would rather select an enterprise with less variation in return (less nsky) among the enterprises that would generate the same expected retums, given everything else constant.
The greater the short-term fluctuation of tea prices the riskier is tea growing as an enterprise to risk-averse farmers. This would influence the entry decision into tea farming given the varying ability to manage the risk among the farrners. It has been argued that large fmsand small fmshave different cornpetencies in managing the risks hence to deal with short term fluctuations in incomes as a consequence of short-term fluctuations in pnces (Tweeten, 1984). Farmers who can better manage the short-term fluctuations in income are more likely to survive with price instability, and such
competencies would influence the nsk preferences of fmers.
Some argue that large fmshave more competencies to fare through such short-
term fluctuations in income relative to small fms. In an unstable price environment,
access to information and better utilization of infonnation is crucial. The increase in overhead costs of such infonnation retrieval systems is often feasible only if such costs can be spread over a large amount of outputs. As discussed in chapter 3, to the extent
that large fmshave the resources to obtain better market information and use other
measures to improve forecasts, they have an advantage over small fmsin an unstable market environment (Tweeten, 1984, Key and Runsten, 1999). Moreover, large farms have an additional advantage over small fmsin holding liquid financial reserves, or have better access to short tenn credit or overdraft facilities in cash flow crises (Lins,
1979).
On the other hand, small farmers are more able and likely to engage in off-fm employment which can buffer the income fluctuations fiom riskier cropping enterprises
(Tweeten, 1984). The possibility of engaging in a diversified enterprise of crops and livestock to stabilize average income is also noted as a relative advantage for small farmers. Cornmenting on the Sri Lankan srna11 fmand large farms sectors, Peiris
(1972) noted that owners of small farms can transfer their labor fiom one avenue of production to another when market conditions warrant or demand such a transfer, thereby providing the small fmsector with a degree of flexibility which the large fmsector does not possess. Such diversification would be relatively difficult for large farms whose specialization in a limited number of enterprises with related vertically integrated activities such as processing and distribution would be disturbed with diverse enterpnses
(Tweeten, 1984).
ïhese competing perspectives about the relative cornpetencies of different fam sizes to fare with pnce instabilities could be empincally tested by examining the association between a variable representing the changes in farm size distribution and a variable representing price instability. The outcome of such empirical testing would indicate whether price instability has been an important driver in fmsize distributional changes within the Sn Lankan tea industry. The following hypothesis is tested in
Chapter 7:
Hypothesis 5: Tea price instability index is negatively related with the share of the area of srnaZZ tea fams of the total tea land area. 4.4. National Ecoaomic Conditions and Changes in Size distribution in Tea Farms
Farmers, regardless of their size, respond to changing economic conditions by
altering the optimal mix of enterprises. These decisions are necessarily long run
adjustments to the relative pnces of the economy. However, the flexibility to change
may Vary with size, as the finn will need to compare the welfare iosses of exit Fiom one
enterprise to the increased profit with another enterprise. The welfare loss of shutting
down a production activity depends on the losses of rent fkom the fixed factors (Just,
Hueth and Schmitz, 1982, p.53). The firm may also need to invest up front in specific
investrnents that has no value to other firms (and therefore has no value on a second-hand
market) and cannot be allocated to another use within the firm (Tirole, 1994 p.307).
These costs known as the "sunk" costs cannot be avoided even if the firm goes out of
business. Whyte (1994) argued that if these assets still possess substantial economic
value or have not fully depreciated yet, the decision makers in the firrn will tend to
consider these sunk costs in their deliberations about fùture courses of action.
The important differences between large and small tea farms are the nature of labor supply and the managerial input (Peiris, 1972). In large tea farms, a residential
labor force and the hierarchy of supervisory and technical personnel require some infrastructure and other overhead services, the provision of which implies a substantial sunk cost is incurred (De Silva, 1982). Such investments are mainly on labor housing, and other amenities of health and water supplies, housing for supervisory staff and roads and the electricity distribution inside the farms (see Section 4.2.3.C). However, except for the substantial sunk costs involved in establishing and bringing up the tea field to an income generating stage, which is the same for al1 fms, these infrastructure related sunk costs of small farms are much less relative to large farms.
These differences in sunk costs between large fmsand smaller fmswould play a role in the ability to respond to changes in relative prices. Large fams would not cover the sizable sunk costs if it shut down the tea production to switch to other cropping activities unless these other crops are very profitable. Akiyama and Trivedi (1987) showed that small farmers has responded to relative price increases by more entry and expansion of the production area, while large tea farms responded improving yield through greater use of yield-increasing agrïcultural practices. Their expansion in the production area was limited.
Thus, small fmsseems to be relatively more flexible to respond to economic incentives unfolding in a changing economy. Relative to large fms, the land share of the small hsthat has been devoted to commercial crops cm be more flexibly allocated to other commercial crops if such crops were to become relatively more profitable.
Hence, it is hypothesized that the larger the profitability of alternative crops compared to tea ceteris paribus the smaller will be the relative share of the small tea farms.
Hypothesis 6: Changes in relative profiability of alternative crops is negatively related to the share of tea area under the small tea farms in the total tea area.
4.6. Summary
This chapter developed the conceptual fiarnework to derive the hypotheses about the effect of various drivers on the farm size distribution changes. The six broader categones of structural change drivers discussed in chapter 3 is used to identify the potential drivers of the farm size distribution changes in Sri Lankan tea industry. The major focus in developing these hypotheses was to capture the differential responses of small and large fanns to changes in a given driver. The hypotheses are developed about the following potential drivers on the changes in fm size distribution; relative opportunity cost of labor between small and large farm sectors, land reform policies, srnaIl fami support senices, tea export taxation, state pricing intervention, tea price instability and profitability of alternative agricultural enterprises. The specific hypotheses are;
Hypothesis Ir me ratio of opportuniîy cost of labor for small and large tea farms is negatively related to the share of tea area cultivated by the small farms.
Hypothesis 2: The share of tea area under the small fams is larger in the post-land refonn period contpared to the pre-land reform period.
Hypothesis 3: The small farm subsidy per unit land area is positiveiy related to the share of tea area under small farms.
Hypothesis 4: nie size of the tea export taxes is negotively related to the share of srnail tea farms in total tea area.
Hypothesis 5: Tea price instability inda is negativeiy related with the share of the area of small tea farms of the total tea land area.
Hypothesis 6: Changes in relative proftabiliîy of alternative crops is negatively related to the share of tea area under the small tea farms in the total tea area. The empincal specifications of the variables and the data for such variables representing the fmsize distribution and various dnvers of its change will be discussed in chapter 6. in chapter 7, these empincal specifications are estimated and these hypotheses are tested. CHAPTER 5
CONCEPTUAL MODEL OF CHANCES IN VERTICAL COORDINATION IN
TEA INDUSTRY
5.0. Introduction
The purpose of this chapter is to develop a conceptual model to explain the changes in vertical linkages within the Sri Lankan tea industry. As descnbed in chapter
2, black tea production is organized either as vertically integrated large fms or as vertically independent green leaf producers and black tea processors. Until about the late
1970s, the share of the vertically independent sector had been stagnant at around 20% of total black tea production. Since then, its share has rapidly increased and by 1998, it had reached about 55% of total black tea production (see Figure 5.1). While the share indicates the increase relative to the vertically integrated sector, the absolute increase is also substantial (see Figure 5.1). Thus, vertically integrated tea production has been substantially replaced by independent green leaf tea production and processing dunng this period. What has caused these changes in the vertical linkages? This chapter develops a conceptual model to analyze the effect of factors that have caused the changes of vertical linkages in the tea industry.
The chapter is organized as follows. Section 5.1 provides an overview of the determinants of vertical coordination arrangements in relation to "make" venus "buy" decisions. Section 5.2 discusses the production costs differences due to the technological reasons (size economies) in the intennediate input production and its effect on the
"make" vs. "buy" decision. Then the green leaf production cost differences in the
"make" versus "buy" decision for the Sri Lankan tea industry are discussed. Figure 5.1. Changes in Black Tea Production among Vertically Independent and Vertically Integrated Sectors ia Sri Lanka
Source: see Annex 6G Section 5.3 discusses the changes in management incentives and agency costs that
consequently affect the changes in "make" vs. "buy" decisions. The changes in agency
costs and management incentives in the Sri Lankan tea industry and the hypothesized
effect on vertical coordination changes are outlined. Section 5.4 develops a theoretical
model of the impact of asset specificity and hold up threats in the vertical coordination
arrangement. The model is focused on the green leaf tea market and incorporates the
effect of hold up problems in the green leaf market equilibrium. The model is extended
to incorporate the effect of state pncing intervention in the green leaf market and the
consequent effect on the observed changes in vertical coordination arrangements.
Testable hypotheses are developed at the end of the sections 5.2 to 5.4. Section 5.5 discusses the main strengths and weaknesses of the hold-up argument in the Sn Lankan green leaf market and section 5.6 summarizes the main points of the chapter.
5.1 Overview
The changes in the vertical coordination arrangements are tinked to changes in the relative profitability of the methods in obtaining intermediate inputs. If a firm has two production processes in which the entire output of one of the production processes is employed as (part or al1 of the requirernent of) an intermediate input of the other production process, the finn is said to be vertically integrated (Peny, 1989). An alternative arrangement of obtaining the intermediate input is outsourcing or purchasing it korn other firms. Which of these two methods for obtaining the intemediate input is chosen is lcnown as the "make vs. buy" decision of the fmn (Besanko et al, 2000). These decisions determine the vertical boundaries of the firms. The vertical integration of green leaf production and black tea processing could be seen as the 'make" decision while the
vertically independent green leaf production and black tea processing could be seen as
the "buy" decision. Thus, changes in the method of obtaining green leaf tea for the
processing of black tea can be viewed as a movement fiom the "make" decision to the
"buy" decision. What are the important determinants for such changes in the vertical
coordination arrangements?
The choice of coordination mechanisms for the activities in the vertical chain,
such as vertical integration (make) or reIying on market forces (buy) is dependent on a
variety of factors. These factors compare the benefits and costs of using the market as opposed to organizing the activity in-house. They capture the differences in both the
management costs as well as transaction costs pertinent to these alternatives (Coase,
1993, Demsetz, 1988, and North and Wallis, 1994). Coase (1960) defined transaction costs as the costs of searching for a partner with whom to exchange, bargaining with potential trading partners to reach an agreement, and verification of the agreed conditions are fiilfilled by the parties. Transaction costs is the cost of organizing resources across markets while management costs are costs of organizing resources within fim. Coase
(1937) distinguished between the management costs and transaction costs and then explained that activities carried out by firms are those that are less costly to handle within an organization compared to with market transactions. Profit maximization requires the substitution of firms for markets if the costs of using markets become larger than the cost of managing.
However, Demsetz (1988) argued the relative costs of transaction and management in organizing resources are not easy to distinguish. The "make" decision does not necessarily eliminate transaction costs since these decisions require the purchase of inputs across a market to "make" the good. Similady, purchasing goods fiom another fim does not necessady eliminate management costs, since it involves an implicit purchase of the management services of this other fim. Therefore, Demsetz argued that the choice among alternative production organizations depends on whether the sum of management costs and transaction costs incurred through in-house production is more or less than the surn of management and transaction costs incurred through purchase across markets.
Demsetz also argued that the cost of transacting is only one element of the difference between the cost of purchasing fiom others and the cost of making something oneself. That crucial difference also depends on technology and operating-cost differences between firms. The relative cost differences due to technologically detemined reasons between firms are also important in the "make" versus ''buy" decision. Thus, Demsetz argued even if transaction costs are zero and management costs are positive the "make" decision may still be profitable. This would be the case when the firm specific technology give nse to more efficient in-house production compared to the production in another firm. Thus, contrary to a simple comparison of transaction costs and management costs assurning the sarne efficiency in production in al1 firms, "make" versus "buy" decision is also dependent on the differences of production efficiencies among firms. Therefore, realistically, management, transaction and production costs are al1 assumed to be positive, the correct decision is reached by assessing whether merger or independent production yields the lowest unit cost, taking al1 the costs into account, over the relevant range of output (Demsetz, 1988). The relative changes in the transaction costs, management costs, and production
costs would influence the choice of the most profitable vertical coordination arrangement.
In this chapter, the economic reasoning for the changes in these costs is discussed and
their impact on vertical coordination changes in the Sri Lankan tea industry during the
last four decades is investigated.
5.2. Changes in Production Costs and "make" versus "buy" decisions
One significant advantage of the "buy" decision arises when size economies are
important in producing the intermediate input. An individual firm's level of use of the
input may not allow the firm to produce this input in-house at the point of minimum total
average cost in the cost function or at minimum efficient scale'. If this is the case, one
large firm can satisfy the intermediate input needs of many firms situated downstrearn,
thereby enjoying the economy of size. In addition, the single large firm can exploit its
experience producing for many firms to obtain leaming economies (Milgrom and
Roberts, 1992). In such circurnstances, market procurement of the intermediate input
would be the lower cost option compared to in-house production, ignoring transaction
costs.
Stigler (1 95 1) used the above argument of economy of scale through pooled
production to outline a theory of vertical integration and dis-integration, which is known
as the life cycle theory (Mahoney, 1992). Stigler defined a firm as a collection of
I One can argue that the fim still can reach the minimum efficient scale and sel1 the excess of the intermediate input. This phenornenon is known as "taper integration" (McFertridge, 1994 p.526). Given the option of taper integration, there is nothing to deter cornrnon ownership of successive stages of production even where intemal vertical transactions constitute a small fiaction of total activity. However, this would bring about other management diseconomies due to the expansion of the fin size (see Williamson, 1985 Chapter 4 and McFertridge and Smith 1988 p. 15-16). interrelated activities or fhctions each with a separate cost function with different output levels at the minimum of the average cost (efficient size). A completely vertically integrated firm undertaking necessary activities in-house is burdened by inefficiencies that arise fiom the fact that these activities may not achieve minimum average cost at the implied rate of output. As the market grows in size, it becomes easier to avoid these inefficiencies by separating the vertical stages of production into separate industries populated with specialized firms. The essence of Stigler's argument was that if economies of size are achieved at levels of intermediate input production larger than a single firm's requirement, and if another firm amalgamates such requirements for many downstream firms and reaps the size economies, the "buy" decision will be more economical than the "make" decision ignoring transaction costs. How are the size economies through "buy" decisions relevant to the change in vertical coordination arrangements for the Sri Lankan tea industry?
In the Sn Lankan tea industry, the "buy" scenario occurs when processors elect to procure green leaf tea fiom independent, small production units. According to the economies of size argument alone, the costs of green leaf tea production should be higher for small size fms than for large farms. However, there is no empirical evidence to support the existence of size economies in green leaf tea production. Mendis (1992) and
Roberts (2989) showed that Sri Lankan tea production exhibits constant returns to scale technology? Ramachandran (1963), Etherington (1 971 ), Casperze (1979, Fernando
(1981), and De Silva (1982) have also rejected scale economies in green leaf production
' However, Roberts estirnated using restricted profit fûnction and his estimation of size econornies shows "in the field (green leaf production) they declined as area increases until about 300 hectares, above which they are either unitary or even decreasing (p. 193-194). Meanwhile, Mendis examined green leaf production using a Cob-Douglas production function and showed technology is constant returns to scale. and the subsequent cost advantage associated with larger production units in green leaf production.
instead of cost disadvantages there are sources of cost advantages in producing
green leaf tea in smaller production units compared to vertically integrated larger farms.
The main sources of the differences in cost of green leaf tea production between smaller
scale farms and larger farms is labor and overhead costs. The cost of green leaf tea
production is largely determined by labor costs (see Chapter 4 section 4.1 .A). The labor cost component is as much as 72% of the cost of green leaf tea production (see Table
4.2). Moreover, the labor requirement for green leaf tea production is the same for the vertically integrated large famis and vertically independent small fms(see Table 4.1 ).
Thus, changes in labor cost differences would be an important determinant of whether to procure green leaf from small fms or grow in-house ignoring the transaction costs.
Wage rates have increased in the vertically integrated sector due to a unionized labor
force relative to the opportunity cost of labor for the independent growers (see Figure
4.3). Thus, the opportunity cost of labor in the vertically independent sector has been relatively lower giving them a cost advantage in producing green leaf tea compared to the verticaliy integrated sector.
The relatively lower costs in green leaf tea production of the vertically
independent sector gives them a cost advantage in the "buy" decision. The changes towards market procurement of green leaf tea ('buy" decision) avoiding in-house production ("make" decision) could be seen as an attempt to reap the cost advantages of green leaf tea production nom vertically independent small fms. Since the relative opportunity cost of labor has been decreasing for the vertically independent, small-scale producer, the relative share of vertically independent tea production should increase as the relative production costs in this sector decreases. The hypothesis about the association between the changes in relative opportunity costs of labor and the changes in vertical coordination arrangements of black tea production is as follows:
Hypothesis 1: Changes in relative production costs (as driven &y the opportunity cost of labor) wiii affect the share of green leaf grown by verticaliy integrated processors as cornpared to the share of green leaf purchased by bdependent processors.
5.3. Changes in Management Costs and "Maken versus "Buy" Decisions
Demsetz (1988) argued that management costs differences between in-house production and independent firm production should also play a role in "make" versus
"buy" decisions. Are there any reasons to argue that the management cost systematically differ between vertically integrated firms and vertically independent firms in the Sri
Lankan tea industry? If so it could be expected that ceteris paribus vertical coordination arrangement that has a greater management cost per unit of output would be less adopted compared to the vertical coordination arrangement with lower management cost per unit of output. It is argued below that management costs have increased in the vertically integrated plantations during the state ownership relative to vertically independent sector and vertically integrated production becarne less profitable.
There are two strearns of argument in the literature about the general factors that give rise to differences in management costs between "make" and "buy" scenarios. The first argument is about the "diminishing return to the management" and the second argument is about the "benefit of competitive pressure in the market". The argument of diminishing returns to the management is related to the difficulty in coordinating the
number of production activities of the firm get bigger. Bringing what was formerly a
market transaction in-house increases the size of an organization, which often results in
additional hierarchical levels. The increased size implies greater coordination dificulties
between most subordinates and their superiors. Both subordinates and superiors make
decisions that contribute to the profitability of a firm. When subordinates and superiors
knowingly do not act in the best interest of their firm, they are shown to be slacking
(Farna and Jensen, 1978). Slacking is a result of sub-goal pursuit and opportunisrn by
lower level employees and sub-units in complex organizations. Agency costs are the
costs associated with the slack effort and with the administrative controls to deter slack
efforts (Besanko, et al 1999). As firms get bigger, with additional hierarchical levels and
greater distances between superiors and subordinates, the agency costs likely increase.
On the other hand, Williamson (1 985) argues that markets promote high-powered
incentives and restrain bureaucratic distortions more effectively than interna1
organizations. in the strategic management literature (Boone and Verbeke, 199 l), bureaucratic costs have been widely used as an explanation for increased vertical dis-
integration witnessed through subcontracting or "greater recouse to market and quasi market transaction" in the electronics, automobile, and steel industries (p.186). When a
firm decides to "buy" across markets instead of "make", it enjoys an efficiency gain of the supplier (market firms) through the high powered incentives that prevail in the market
firm compared to in-house production (ignoring the transaction costs). Therefore, the incentives in a market favor 'bbuy" decisions over "make" decisions ceteris paribus.
Mahoney (1992 p.569), and Blois, (1972 p. 321) note that a lack of direct cornpetitive pressure in controlling the cost of production of intermediate inputs might allow increasing levels of slack. For instance, an in-house division performing an activity solely for the firm of which it is a part (e.g., data processing department in a bank) is known as a cost center (Besanko et al 2000). It is difficult to evaluate the efficiency or the profitability of cost centers if there is no comparable market based provision or other contractual arrangements of such a service. The costs related to slacks would be especially problematic in such situations. McFertridge and Smith (1988) noted that when divisional profits are an arbitrary component of total corporate profits, divisional shortcomings are less readily detectable, and their consequences are borne in part by other divisions.
Have there been changes of incentives and agency costs in the Sri Lankan tea industry rendering vertically integrated coordination less profitable? It is argued below that the role of incentives and agency costs have been changed under the state ownership of vertically integrated tea fmscompared to the period of their private ownership. It is fùrther argued that during the period of state ownership (1976 to 1992), it is likeIy that state-owned vertically integrated plantations were burdened by larger management costs relative to vertically independent sectors. Thus, vertically independent production became relatively less expensive coordination arrangement compared to vertically integrated plantations during the regime of state ownership of plantations.
The role of agency costs and incentives in deciding the arrangement of vertical coordination in the Sn Lankan tea industry substantially changed after the nationalization policies of the mid 1970s. First, during the period of state ownership, management incentives radically changed and arguably, this would have escalated the slacking and poor decision-making. The state acquired about 377,000 acres of pnvately owned and managed tea lands (about 63% of the total tea area of the country, fiorn tea fmslarger than 50 acres in size) with the nationalization and land reform poiicies in 1972 and 1975
(see section 4.2.A in Chapter 4). About 96% of this state-owned tea Land was assigned to two state corporations (State Plantation Corporation and Jantha (Peoples) Estate
Deveiopment Board). The incentives for managers that had existed under private ownership, such as profit-sharing bonuses and the prospect of climbing higher in the estate managing Agency houses (see Chapter 2 section 2.3.B), were absent under the nationalized-management (Rote, 1986 p.2 15). Government salary scales were applied to the managers in the state corporations and these salaries were less than the manager's remuneration under private ownership (Rote, 1986). in addition, the salaries and other remuneration of the managers/bureaucrats were independent of the profits or losses of the tea fmsthey managed (Bandaranaike, 1993).
Critics have noted the management failures in maintaining the hge verticatly integrated tea farms dunng the period of state ownership and management. Management failures in the vertically integrated tea fmsduring the state ownership were noted for various aspects such as general agricultural conditions including land and soi1 management practices, tea bush maintenance (Tea Master Plan, 1978), adequate fertilizer application (Simathamby and Deveraj, 1987), and replanting (Peiris, 1984, Si~atharnby,
1993). As shown in Figure 4.6 in Chapter 4, management performance of the state- owned tea farms during the 1980s as measured by administrative staff per 100 laborers and per 100 hectares, changed in a manner indicating a deterioration of management efficiency. The change in the cost of production of black tea dunng the penods of state ownership and private management is relevant here. During the five years preceding
nationalization (1972 to 1976), the annual average cost of production of black tea was
about Rs 3.00 (1952 Rskg) (Peebles, 1982 p.322). During the last five years of state
ownership (1986 to 1991) the annual average cost of production increased up to Rs. 5.86
(1952 Rskg); an almost two-fold increase (Plantation sector statistical Pocket Book 1999
p. 136). Under re-privatized management since 1992, the average cost of production of
black tea came down fiom 6.47 (1952 Rskg) to 4.47 (1952 Rskg) in 1999; a 30%
reduction (Plantation Sector Statistical Pocket Book, 2000, p. 133).
Management cost differences of vertically integrated large fmsunder the state
ownership and private ownership are linked to the viability of the coordination
arrangements dunng different ownership penod. The burden of larger management costs
or lower retums for the management input of vertically integrated production under state
ownership is expected to reduce the relative profitability of vertically integrated
production compared to vertically independent production.
Hypothesis 2: Given the fack of incentives to managers of verticaliy integrated firms under the slate ownership of tea land and processing plants. the share of tea production from the verticafiy independent units is expected to be higher during the period of state ownership of vertically integrated jinns (1976 to 1992) relative to the periods ofprivate ownership vertically integratedfims (1960 to 1975 and 1993 to 1998).
with everything else constant. 5.4. Changes in Transaction Costs and uMaken versus '(Buy" decisioos
One of the important determinants in the choice of vertical linkages is the value of the resources used up in a market transaction. Coase (2937) showed searching for trading partners, negotiation and bargaining for terms of trade and the verifkation of that the agreed conditions are hlfilled by the parties gives rise to resource used up in market transaction. Williamson (1985) argued that the above costs are ex ante transaction costs, which are associated with contract formation. He argued these as "the cost of drafting, negotiating, and safeguarding an agreement" (p.20). He introduced et post transaction costs as those that arise due to failures in on going contracts with changing economic conditions, when the contracts are incomplete (boundedly rational individual cannot think of al1 the possible contingencies to include into the contract) and parties to the contract are opportunistic (seeking self-interest with guile). Williamson argued that when bounded rationality and opportunism are taken as behavioral postulates of human behavior, the magnitude of ex post transaction costs are an important determinant of the vertical coordination arrangements with a transaction-specific investment. The combination of durable, relationship-specific investrnents, uncertainty about future demand and costs, and contractual incompleteness arising from the costs of writing, monitoring, and enforcing full contingent claims contracts increase the likelihood of a firm vertically integrating rather than relying on the market to buy intermediate inputs keeping everything else constant (Joskow, 1988, Williamson, 2 979).
Transaction specific investments or asset specificity is argued to raises the prospect for opportunism. Thus, when the parties to a transaction have invested in relationship specific assets, they are more likely to expose to hazards of opportunistic behavior relative to the parties in a transaction who have not invested in reiaiionship
specific assets. Demsetz (1988) distinguished the difference between costs of market transaction including search, bargaining and verification costs and cost of opportunistic behavior involving transaction specific investments. He claimed that transaction specific
investment increases the losses due to failure of agreement rather than these investrnents increase the cost of transacting involving search, negotiation, and verification costs.
Losses due to failures in agreement (contract breaching) involving relationship specific assets are large relative to the losses due to failures in agreement involving non-specific assets. This is because retum from the next best alternative trading opportwiity is far too small for relationship specific assets while returns from the next best alternative trading opportunity for the investrnent in non-relationship asset is much closer to the retums from the best trading opportunity (Klein, Crawford and Alchian, 1978). The deciding factor of the success of contracts between two parties with transaction-specific investrnents is how to minimize the losses due to failures of agreement in the future. It is argued that institutional arrangements evolve to allay such failures and to minimize the losses due to transactional nictions (North, 1990). However, other forces such as govenunent regulations could also shape the institutionai arrangements goveming the exchange process and thereby allay transaction Enctions (Mulherin, 1986).
The institutional arrangements that govem the green leaf trading activities in the
Sri Lankan tea industry have substantially changed during the last four decades. As discussed in Chapter 4 (section 4.2.D), green leaf tea prices were unregulated before
1968. From 1968 to 1978, the government decided the amount receivable to the processors fiom auction prices of black tea without guaranteeing minimum prices to green leaf producers. A guaranteed minimum price for the growers was introduced in
1978 and it was effective till 1984 with several amendments.
In 1984, government introduced the reasonable price formula under which the green leaf price is directly attached to the Colombo tea auction prices. It is possible that these changes in the institutional arrangements are important in the observed changes in
"make" vs. "buy" decisions of green leaf procurement. How can one conceptualize the impact of pricing intervention in the vertical coordination arrangements in Sn Lankan tea industry?
It is argued that a large number of market participants as evidenced in the Sn
Lankan tea industry, would result competitive prices for green leaf tea. In the 1950s, there were a large number of processors (about 950) and green leaf producers (about
86,840). Based on this reality, some argued that government pricing intervention in green leaf market should have been prevented and market forces should have been allowed to determine green leaf pnces (Report of the Presidential Commission on Tea
Trade and hdustry, 1995). However, literature indicates arms length transaction between green leaf producers and processors in Sn Lanka and other tea producing countries such as Kenya have been unsatisfactory.
Green leaf processors would be indifferent to "make" vs. "buy" decision for procunng green leaf tea if green leaf tea prices are competitive, given that the management and production costs differences between these arrangements are unimportant. However, if green leaf prices with the arms length transaction are not competitive and if such prices lead to losses either to green leaf producers or processors, vertically integrated green leaf tea production would be encouraged and the arms length transactions would be discouraged. Has the intervention in green leaf pricing been able to promote the arms length transaction between green leaf producers and processors? If so the pricing intervention has been infiuential in changing the vertical coordination arrangement.
S.4.A. Green leaf tea Prices in Arms length Transactions
One must carefully assess the evidence whether competitive green leaf tea prices would Iikely have prevailed in the arms length transactions between green leaf producers and processors in tea industries in Sn Lanka and in other countries. If this was the case, the green Ieaf pricing intervention may not have influenced the changes in vertical coordination arrangements in the Sri Lankan tea industry. Particularly, one must also recall that in Sn Lanka tea plantations were established as vertically integrated larger production unit. How likely are the chances that the activities of independent green leaf tea producers and processors would lead to the competitive green leaf price in the arms length transactions? There are number of economic forces such as competitive behavior among green leaf tea producers and processors, and reputation of these economic agents that would result cornpetitive green leaf tea prices. The large number of green leaf tea producers and processors in the Sri Lankan tea industry suggests a competitive green leaf tea market. The reptation for being an honest and non-opportunistic partner in business is necessary if one were to continue in business. On the other hand, green leaf tea prices could be non-competitive with monopsony power of the processors, monopoly power of the green leaf tea producers, andor hold up potentials of the returns fiom investment in green leaf tea production or in green leaf processing. The relative dominance of economic forces that tend to gravitate the green leaf prices towards the competitive green leaf price would facilitate the "buy" decision in tea production given everything else (management costs and production costs) constant. On the other hand, the relative dominance of the forces rnaking green leaf pnce unfavorable to one party would promote "make" decisions in tea production, given everything else constant. Thus, the relative importance of these opposite economic forces must be carefiiily discussed.
5.4.A.1 Competitive behavior Among Green leaf tea Producers and Processors
The discussions about green leaf pricing in Sri Lanka are confined to an uncontested notion of exploitative green leaf pricing by the processors. These discussions might be echoes of the lobbying activities of green leaf tea producers, who are generally, small farmers and lobby for higher prices for their produce regardless of prevailing prices. Despite the unanimity among researchers about the exploitative green leaf prices offered by processors, the researchers were not able to demonstrate how such
"exploitative" green leaf prices prevailed simultaneously among a large number of green leaf tea processors. Green leaf tea processing is a profitable enterprise, and there are no barrien to entry, thus exptoitative green leaf pnces are not consistent with a competitive processing sector. Despite this inconsistency, the researchers emphasized the
"exploitative" nature of green leaf tea prices by the green leaf tea processors. For example, Ramachandran (1963) cornrnenting about the green leaf tea market before price intervention noted;
Even greater were the difficulties that smallholders had to contend with in the processing
and sale of their green leaf tea. Lack of finance, skill, and knowledge meant that srnallholders could not undertake these operations for thernselves and had to se11 their
tea-leaf to estate and their factories; and the cornplaint was oflen heard that the factories
paid the smallholders unduly low prices (p.44).
Yet, the author was unable to show how the pnces were "unduly low" or to provide economic reasoning as to why this situation prevailed. Forrest (1965) presented a similar comment;
The average smallholder still sen& his produce to one of the commercial 'bought leaf
factones but except where special factories have been started for the exclusive use of
Smallholders' Co-operatives, his best chance of a fair deal is probably to make contract
with one of the larger estates which has spare manufacturing capacity and which is
prepared to take an interest in his output and apply fairly rigorous standards to it. With
an ever-rising intake of estate leaf putîing pressure on factories, such arrangements are
not easily come by today. The bought leaf factones themselves are working beyond their
rated capacities, with adverse effects on the quality of their tea" (p.241).
Again, the author was unable to provide any data or economic reasoning as to why the smallholders could not get a "fair deal" fiom the "bought leaf factories". De Silva
(1982), on the issue of lower prices for green leaf, has also noted;
The price paid for green leaf by the tea factories was barely remunerative. The financial
condition of smallholders (with their constant need for cash), and deterioration in the
quality of green leaf within about six hours of being plucked, weaken their bargaining
position in the market for leaf (p.282). De Silva was also unable to explain why the factories were in a bbstrong**bargaining position (despite about 950 processors indicating a competitive processing sector). He argued that "the profit differential between estates and smallholders thus contains a rent element (fiorn monopoly of processing) and, to this extent, reflects the depressed pnces for unprocessed produce" (p.287). Yet, De Silva was unable to show how the "monopoly of processing facilities" was created, particularly when a large nurnber of processors were in the processing sector. Fernando (1982) has also noted the presence of a monopoly in the processing sector. He argued;
Since tea is a perishable commodity, it has to be sold soon afier plucking. This reduces
bargaining power of producers to a significant extent. in certain areas factory capacity is
not adequate to cope with supply of green leaf and there exists transport problems too.
These factors make small tea farmers more vulnerable to monopolistic elements in the
market (p. 1 1).
Again, Fernando was also unable to substantiate the daims about monopoly in the processing sector. The monopoly argument was also put forward by the Report of the
Cornmittee Appointed to Study the Constraint and Problems of the Tea Small Holding
Sector, 1980 (p. 12). Despite the shared view arnong the researchers that the green leaf pnce was unlikely to be competitive, there were inconsistencies in their views. The monopsony (rather than monopoly) power of the green leaf processors was unlikely wiîh about 950 processors.
Forrest (1965) in his comment noted that green leaf producers may have dificulties when they try to se11 their green leaf to the processors who have their own green leaf production but have some excess capacity to accommodate the small producers' green leaf. Yet, it is not clear why this is the case. 1s it because processors who have their own green leaf tea production are not dependent on the green leaf tea
supply fiom small producers?
Few of the authors noted the pecishability of the green leaf tea and the
requirement for quick transportation to the processor, which have nothing to do with
monopsony power of the processors. Monopsony is a situation where the processing
sector is confined to a single buyer of the green leaf tea from the producers, and there are
barriers to entry in to the processing. Thus, it is unclear whether the absence of competitive forces, as implied by the monopsony argument, is a valid explanation of
lower green leaf prices. There were no systematic analyses to prove the contrary.
Moreover, an important ~e~ort~about the Sn Lankan tea indushy, argued that market forces should be allowed to detemine the green leaf tea price as in many other small holder crops such as rubber, coconut, spices, vegetables and paddy. The Report argued that the green leaf tea producer and the processor are in symbiotic relationship and that the two parties should be aware of their interdependence with the green leaf price determined by market forces reflecting this interdependency. This recommendation is an acknowledgement of a competitive green leaf market in Sri Lanka. However, the arguments for a competitive nature of the green leaf tea market have been rejected by many and supported by few. An objective assessrnent has not been done about the competitive nature of the green leaf market and its impact on the green leaf tea prices.
Thus, the possible impact of green leaf pricing intervention on the changes in vertical coordination arrangements in Sri Lankan tea industry is not cIear.
3 Report of the Presidential Commission on the Tea Indusrry and Trade I995, Sri Lanka: Sessional Paper N0.11- 1995. 5.4.A.2. Green Leaf Prices and Arms length Transaction in Other Countries
Researchers in many other countries in addition to Sn Lanka have also noted the unsatisfactory performance of market forces to price green leaf tea. Grosh (1994) gave a detailed account about the avoidance of arms length transactions between producers and processors and prevalence of contract farming for green leaf tea in Afica. She claimed that although plantation green leaf tea production is expensive relative to owner-operator
(small fmer) due to costly labor supe~sionin the former category, the spot market activity in green leaf was not successfül and green leaf continued to be grown in plantations. When small farmers were involved in green leaf production, the contracts were used to govem the transaction between them and the processon. She claimed "Tea is contracted because a spot market in green leaf never works" (p.254). However, the only explanation she provided for this condition was that contracting would often be the best form of organization for those crops for which "penshability requires careful co- ordination between growing and processing.*'(p.256). However, it is not clear why adequate green leaf pices in arms length transaction could not suppon "careful co- ordination*' between green leaf tea producers and processors.
Delgado (1999) noted that green leaf tea production and processing is one of the most difficult products to coordinate through a spot market and claimed that tea has to be grown through contract farming by small holders or on large farms (p. 183). Supporting this argument he claimed "The difficulties smallholders face in finding reliable markets for perishables is one source of transaction costs, due to the low bargaining power of a famer whose product is spoiling9'(p.183). Does this imply the searching component of the transaction cost is large with perishable products? If so, provision of information about the potential buyers would substantially lower the search cost. Yet, he also has not provided any explanation in what respect the spot market coordination is not satisfactory.
Haymi (1994) noted the need for close coordination between fmproduction and processing underlies the pervasive use of the plantation system for black tea manufacturing since "low transaction costs is associated with the supply of farm- produced materials to central processing'*(p.13 1). However, this cornparison of transaction costs between the "farm-produced" (make) and 'outside produced" (buy) options are not adequate to explain the avoidance of a green leaf market. As noted in the section 5.3, the "farm-produced" material should have to be cheaper net of the transaction costs for it to become a relatively more profitable method of getting green leaf tea compared to procuring it fiom outside.
The Kenyan Tea Development Authonty is a para-statal organization which contracted with small holder tea growers for their green leaf tea and these growen are paid according to a agreed upon formula based on the world prices realized. There is an initial monthly payment .for the green leaf producers without considering the world market black tea pnces and then the annual payment will make the link between the world market pnces and the growers' final compensation. This formula-like green leaf pricing has been identified as an important element in the incentives for the green leaf producers (Lamb and Muller, 1982 p.4). Evidently, Kenyan small growers too do not depend on the spot market to pnce their green lear tea.
The above account shows that despite many claims about the avoidance of anns length transaction between green leaf tea producers and processors, there is no satisfactory explanation provided. It could be argued that exploitative green leaf pnces by the processors could eventually hurt the processors themselves. This is because reputation related losses are possible with such exploitative green leaf prices by the processors. The empirical verification of the impact of green leaf pricing intervention on the changes in vertical coordination arrangement in the Sri Lankan tea indus6 requires an economic explanation about the avoidance of a arms length transaction for green leaf tea.
A number of arguments about the failures of arms length transactions in green leaf tea market in Sri Lanka and in other countries have been discussed, yet none of them alone provided a satisfactory explanation. These reasons include monopsony power of the processors, perishability in green leaf tea, rapid transportation requirements to deliver to the processors, lack of transportation facilities, and difficulties in selling green leaf tea to processors who have their own green leaf production. Do such factors act as obstacle in reaching competitive green leaf tea prices? While the impacts of these factors have not been substantiated, the researchers have failed to explore another possible explanation for the failures in anns length transactions in green leaf market. This is the possibility of hold up threats involved with the arms length transaction in the green leaf market.
5.4.B. Hold Up Threats in the Arms Length Transaction
Transactions that involve specific investments require arrangements to protect the investor against early termination or opportunistic renegotiations of the terms of the production relationship. This aspect is established in the literature as the hold-up problem. The hold-up problem is defined as "the general business problem in which each party to a contract womes about being forced to accept disadvantageous terms later, after it has sunk an investment, or worries that its investment may be devalued by the actions
of others" (Milgrom and Roberts, 1992 p. 136). The hold up problem and the institutional
arrangements governing the exchange process are related phenomena. The changes in the
institutional arrangements governing the arms length transaction in green leaf market are
important in changing the vertical coordination linkages. in general, institutional
arrangements evolve to allay transactional fictions, such as hold up problems.
Transaction cost economics tries to explain under what circumstances incentives are more Iikely to exist that lead to opportunistic behavior by the parties to the transaction
(Williarnson, 1985). Opportunistic behavior has been related with the existence of significant transaction-specific sunk costs, or specific assets, which are of little economic value outside the buyer-supplier relationship. The concept of asset specificity is related to transaction specific sunk costs. Asset specificity anses when the ability to generate value fkom a non-salvageable investment is conditional on another party's behavior. It is an exchange related (behavioral) concept rather than a physical concept (Alchian, 1984).
The essence of asset specificity is that, when the investment is made, there is a loss in
"the ability of parties to turn to alternative partners should one seek to gain at the expenses of the other" (Masten, 199 1 p.8).
One can ask how does this apply to green leaf tea production and processing where the green leaf tea production assets cm be used to sel1 to a large number of processors. The number of processors and green leaf producers alone is misleading parameters to assess the potential for hold up problem. It is argued below that some of the factors suggested by the authors who noted the unsatisfactory green leaf pricing with the amis length transactions rnight fit into the asset specificity argument. Factors such as perishability of green leaf tea, the cost of searching for available alternative processors, additional transportation cost to reach them, whether such processors use their own green leaf, whether the green leaf can be stored till one finds another processor (in case if one processor offer low prices) may influence the green leaf prices in arms length transaction.
Masten, 1991 argued temporaI specificity or the perishability related limitations to trade with alternative trading partnen would make investments in producing agricultural intermediate products relationship specific investments. The arms length transactions might fail if the penshability related limitation forces the green leaf producers to accept unsatisfactory terms of trade from the processors. The possibility of hold up problems in the arms length transactions in green leaf market and the possibility of allaying such hold up problems by the green leaf pricing intervention has not yet been explored. A conceptual model is developed in the following section to accommodate the hold up threats in the arms length transaction and possibility of explaining the failures in arms length transaction in green leaf market. Then the model derives plausible explanations of the impact green leaf pnce intervention in the vertical coordination changes in the Sn
Lankan tea industry. The refùtable hypothesis are derived fiom these explanations and these are tested in chapter 7.
5.4.C. Conceptual Model of Green Leaf Market in Sri Lankan Tea Industry
The model is similar to Blair and Kesserman (1987) who focus on the possibility of introducing formula pncing in an intermediate product market showing the indiîference between "make" versus "buy" decisions. The present model extends the
Blair and Kesserman (1987) model by introducing post-negotiated incentives leading to oppottunistic behavior by one party when the other party has transaction specific investrnents.
initially, it is assumed that green leaf producen and processon are vertically independent agents. Green leaf producers and processon as vertically independent firms would respond to green leaf price by adjusting their production and processing decisions.
Assume there is a homogeneous group of vertically independent green leaf tea pmducen
(denoted by the subscript G for grower). The green leaf tea producers are assumed to have the sarne fmsize, cost functions, yields, and investment characteristics in green leaf tea production.
Quantity of green leaf tea is denoted by X and the green leaf tea production by a representative green leaf producer, G, is denoted by XG (kilos per fmper penod). The average total cost of producing XG kilos of green leaf tea is a hnction of XG and the input price vector faced by the growers (PzG).The short run average total cost function is the surn of average variable cost, Cm(X~,&), and average fixed cost (I&) with IG denoting per period amortized investment cost of establishing a green leaf tea farm.
Changes in the technology of green leaf production are captured by an additive exogenous shifi parameter in the average total cost function denoted by a. Increase in efficiency of green leaf tea production would be represented by a reduction in the value of a. The average total cost fùnction is assumed to be a convex function of green leaf tea production (i.e.,Xv.r(Xd//dXc >O and ~c,(x~)/~x~>o).A representative green leaf tea producer is a pnce taker in green leaf market and he would receive pnce (P,) per kilo of green leaf tea. Profit for a representative green leaf producer is DG is: A representative green leaf tea producer maximizes equation (5.1) through the choice of green leaf tea production (XG).
Pr -C:,(X,.P,)X, -C,(X,,P,)-a =O (5-2)
Solving the resulting first order condition in terms of XG,generates the green leaf tea producers output supply hinction x*~;
Given the optimal supply for a representative green leaf producer (XeG)and the nurnber of green leaf producers (fi(assumed) maximized total profit of the green leaf tea production sector ( n, ) is
The green leaf tea processors (denoted by the subscript P for processor) process green leaf tea into black tea. They are assumed homogeneous in their cost of processing, technology of processing, and processing capacities. The green leaf tea to black tea conversion in processing is assumed to be a general transformation function, Q = Q(Xp), where Q'>O and Q"e O representing the decreasing marginal product from the green leaf input in the transformation function. The short run average total cost of the processing is the sum of average variable costs Cv~(Qfx~)),PZp), and average fixed cost Id,where Q(Xp)is black tea output, PZpis input prices faced by processors, Ip is per
period amortized investment cost of establishing processing facility. The short run
average total cost fhction for the processors is assumed to be a convex fùnction of the
level of black tea output (Le., ZYQ(Q,)/@>O and ~CQ(Q.)/~Q>O). The green leaf
processors are price takers both in the final output (black tea) market and inputs (green
leaf tea and other inputs) market. A representative green leaf tea processor would receive
a price of PQ per kilo of black tea. Thus, the per period profit for a representative green
leaf tea processor, np,can be summarized as;
A representative processor maximizes his profit (5.5) through the choice of the amount of
green leaf purchased (Xp);
Solving the resulting first order condition, the processor's input (green leaf) demand
function would be;
The aggregate profits for the processors where NPis the total number of processors is Under the market clearing assurnption the total amount of green leaf produced must equal the arnount demanded by processors or;
N"X; = N~X; (5 -9).
in summary, individual members within each group are pnce-takers in both the intermediate product market and output market. The two groups' profits and hence the joint profits depend on the supply and demand functions of green leaf. Supply and demand are dnven by the green leaf price along with a given set of exogenous parameters. The joint profits for the green leaf producers and processors will be;
The green leaf tea price that maxirnizes the joint profit in equation (5.10) is argued to achieve in a cornpetitive market equilibrium, where both green leaf tea producers and green leaf tea processors act as pnce takers, as shown in Figure 5.2 (al1 the exogenous parameters such as a, PZp.PZG. and Pp are assurned fixed). When the green leaf pnce is at P', joint profits are maximized as shown by the triangle abc, with the net returns to growers of D+E+F and to processors of A+B+C.
The changes in green leaf prices would change each party's profit and the joint profits. These changes of profits depend on the shape of the green leaf supply
(X'G=X'~(P.,))and demand (x*~=x*P(P,)) curves. Note the effect of changing green leaf price on joint profits is ambiguous. Extremely high or low prices would make both parties worse off relative to P'. This is because extremely low or high prices would lead to extremely low level of operations by one party preventing the other party making any Price of Green leaf
leaf
Figure 5.2. Changes in Joint Profit with the Cbange in Green Leaf Price
P' = Equilibrium green leaf price X*= Equilibrium green leaf quantity
Area abc = Joint profit for the green leaf producer and processor P' = Green leaf price when the producers are successful in increasing the pnce Po = Green leaf price when the processors are successfùl in lowering the price XO = Green leaf production with prices Po or P'
C+D = Net loss in joint profits (moving fkom price P* to either to PO or PI) profits. In such a situation, the joint profits shown in the equation (5.10) would be small.
Generally, up to a point, price increases (fkom extremely low green leaf prices) or decreases (fiom extremely high green leaf prices) possibly make both parties better off and joint profits would increase.
One unique green leaf price would make the equation (5.10) maxirnized. This green leaf price is generally assumed to be reached in a competitive equilibrium.
Deviation Eiom this unique green leaf price to either higher or lower green leaf prices would result lower joint profits and inefficient outcome relative to the joint profit maximizing green leaf price. For instance, if the green leaf processors are able to depress4 the prices to PO, or if the green leaf producers are able to increase5 the prices to
P', the joint profits would be reduced by the area of C+D. The literature indicates that green leaf prices in Sn Lanka and some other countnes with the arms length transaction are more likely to be at PO rather than at P'. However, there were no systernatic analyses to substantiate the claims of low green leaf prices, high green leaf pnces, or whether the prices are truly competitive with the amis length transactions. Green leaf pnces of P> and
Piare likely to discourage arms length transaction since at PO green leaf producers might make losses and at P' processors might make losses together both parties jointly make lesser profits relative to the joint profits at P*. Thus, outcome of the green leaf market with PO and P' are inefficient relative to the joint profit maximizing green leaf price.
'' When the green leaf tea processing sector is a rnonopsony, such a condition would occur. However, hold up problems would aIso give rise to such possibilities (see, Klein, Crawford and Alchian, 1978 p.299)
The possibility of green leaf producers to estabIished bargaining cooperatives and influence the green prices have not been ruled out. However, there are no evidences in Sri Lankan tea industry to this effect. The general nature of the changes in joint profits could be depicted with a profit possibility fiontier. It shows the locus ofjoint profits (sum of the profits to the green leaf tea producers and the green leaf tea processors) with different green leaf ptices, when al1 the other exogenous parameters (PQ..PzG.. PzG,, a etc.,) held constant. The profit possibility fkontier is the profit transformation curve, where profit is transformed fiom one party to the other through changes in green leaf pnce. The joint profits are maximized somewhere along the frontier. The maximum can be found by moving out the iso-profit lines until it is tangent with the possibilities fkontier. Assuming profits from production and processing are valued the same, the slope of the iso-profits line is -1 (45 degree line). Joint profits in Figure 5.3 are maximized when the green leaf price is at P'.
With P*, green leaf producer profits are D+E+F and processor profits are A+B+C.
Anywhere along the joint iso-profit curve I profits will total (A+B+C+D+E+F). With the low green leaf pnce such as PO, both green leaf producers and processors might end up in the inefficient equilibrium (E?) with lower joint profits (A+B+E+F) shown by joint iso profit curve II. This outcome is also possible with the higher green leaf prices such as
P'. As the green leaf price goes up fiorn PO or goes down fiom P', both parties7 profits and the joint profits would increase up to a point and afier reaching this point, one party's profit would increases only at the expenses of the other parties profit. However, it is possible that both green leaf producers and processors attempt to enlarge their share of the profits by manipulating the green leaf prices and end up with inefficient outcome such as E2 with green leaf pnce of 9 (see Figure 5.3). Do arms length transactions in green leaf market lead to such inefficient outcornes? To answer this question one must assess the likelihood of reaching to P' with the arms length transaction in green leaf market. Joint Iso - Profit Curve I (nt)
leaf price P.)
Profit Possibility Frontier
1 leaf price Po)
FIGUER 5.3. Variation in Joint Profit under Different Green Leaf Tea Prices A unique joint profit maximizing equilibnum (El) is achievable with a green leaf pnce of P.. The mode1 could derive the green leaf pnce that achieves this maximum of joint profit with a given set of exogenous variables. The next section derives this joint profit maximizing green leaf pncing and shows how incentives might exist with arms length transaction to deviate from this joint profit maxirnizing green leaf pncing.
S.4.D. Optimal (Joint Profit Maximizing) Green Leaf Pricing Arrangement
Suppose a social planner carries out both green leaf production and processing activities in such a way as to maximize the joint profits from both stages. In other words. the social planner mns the two activities as a vertically integrated operation (denoted by subscnpt 1 for the integrated operation). The profits of the social plamer, ff,, would be;
Notice that since it is an integrated production the subscnpt for independent green leaf tea producer (c) and green leaf tea processor (p) is irrelevant as is the green leaf tea price
(Px).Assume that input pnces in green leaf production and green leaf tea processing are fixed. The joint profit now becomes a function of size of green leaf production along with the exogenous pararneters6 of PQ, and a. The optimal green leaf tea production (x') that maximizes the equation (5.1 1) is found by solving the following.
These two parameters (Pa. a) can illustrate the conflict between the green leaf tea producers and greenleaf tea processors when either of these two parameters are held as private information. The implicit function theorem states that when the equation (5.12) holds at the optimal
level of green leaf production, there exists a function; X'=X*(PQ, 4) relating the
exogenous variables to optimal green leaf production. Solving for the inverse of the
optimal level of industry or individual green leaf supply fùnction (5.3), the equation
X'=X*(PQ,.~)could be substituted to determine the joint profit maximizing green leaf
price relationship with the above exogenous parameters as shown in (5.13).
The green leaf production that maximizes joint profit varies with the exogenous parameters of PQ, and cr. Suppose the social planner detennined the values of the exogenous parameters and used (5.13) to set the joint profit maximizing green leaf pnce
in (P'). If the independent green leaf producea and processors adhere to this price, their joint profit will be maximized as in a vertically integrated production, if so parties may be indifferent to vertical coordination arrangements given everything else stays them same among the different vertical arrangements.
It is usually assumed that with a large number of market participants as evidenced in the Sri Lankan tea industry, the resultant competitive forces would lead to an optimal price for green leaf (P.). Even in the 1950s, there were a large number of processors
(about 950) and green leaf producers (about 86,840). Based on this reality, some argued that govemrnent pricing intervention in green leaf market should have been prevented and market forces should have been allowed to determine green leaf prices (Report of the
Presidential Commission on Tea Trade and Industry, 1995). However, possibility of hold up threats with the amis length transactions might prevent the parties achieving P'. 5.4.E. Green Leaf Prices with the Possibility of Hold up Problem
This section explores the potential of hold up threats in the arms length transaction between green leaf' producers and processors. Despite lack of concrete evidences for hold up problems in the Sri Lankan tea industry, it is a possible explanation of failures in arrns length transactions in green leaf market. The hold up problems in the green leaf market is possible in two ways. Green leaf producers can hold up the profits of the processors or processors can hold up the profits of green leaf producers. In either way, green leaf prices in arms length transactions would deviate from the joint profit maximizing price (P*). Such lower or higher green leafprices discourage the arms length transaction between independent green leaf producers and processors since such prices bring lower joint profits compared to the joint profits in vertically integrated operations.
If the pricing intervention prevent the potential for hold up threats, this would encourage the anns length transaction behiveen green leaf producers and processors. It is possible that pricing intervention could have prevented both types of hold up problems. However, one could not definitely Say whether green leaf prices has been increased (by lowering the hold up threats by processors) or decreased (by lowering the hold up threats by growers) without having green leaf prices before and after the pncing intervention.
While acknowledging the possibility of higher green leaf prices prior to the pricing intervention (green leaf producers holding up the processors), here the focus is about the prevalence of lower green leaf prices or the processors holding up green leaf producers profits.
Suppose the green leaf tea producers and processors resort to axms length transaction adhering to joint profit maximizing green leaf price (P.) as denved in the equation (5.13). Assume the investment cost, IG, in equation (5.1) is per penod amortized investment cost (Rupees per fm)for a representative green leaf tea producer.
A representative tea fmis assumed to have a salvage value S, (Rupees per hm), which is the rentai revenue from the tea fm. It is assumed that the rental revenue is less than the arnortized investment cost. The retm to the investment of a representative green leaf producer is the net revenue he could generate producing green leaf tea. The total revenue
(P'X x,) minus total variable costs (C, (XG))&) equals net revenue. The opportunity cost of holding the asset is its salvage value and should thus be deducted fiom net revenue.
The resultant value of net revenue minus salvage value is known as the quasi-rent of the asset (Klein, Crawford and Alchian, 1978). Quasi-rent is the retums that the investor will be deprived of by quitting his production and this is a more general producer surplus measure than profits (Just, Hueth and Schrnitz, 1982 p. 54). Thus, the quasi rent for a representative green leaf tea producer is:
Quasi Rent = {P' - C,(XG) )) XG - S (5.14).
C", is average variable costs and is a function of the green leaf production XG, as previously shown in equation (5.1). Since there are no barriers to entry in green Ieaf tea production such as patent or quotas that prevent the small fmers from becoming engaged in green leaf production, it could be assumed that in the long run quasi-rent would be zero. If the joint profit maximizing green leaf tea pnces were reached in arms length transaction, green leaf producers would breakeven the investment in a tea garden.
The basic premise of hold up threat is that retums fiom the next best alternative trading arrangement is lower than the returns from best trading arrangement (Klein, Crawford and Alchian, 1978). Knowing this, the trading partner in the best arrangement might try to exploit part of the quasi rent hence leading to lower pnces and joint profits.
Suppose the next best alternative buyer is another processor (denoted as processor B) with whom the green leaf producer can arrange to sell his green leaf for the joint profit maximizing price of P.. The green leaf producer retums per kilo of green leaf from the processor B would be P.- T(D), where captures the additional cost that a green leaf producer has to incur doing the business with the next alternative processor. It is reasonable to assume that perishability related costs would increase with the distance (D) to travel to the next alternative processor. When the green leaf is transported to the processor B, value of green leaf decreases by T(D) where D is the additional distance to travel. Therefore, quasi rent fkom next best alternative to the green leaf producer would be [PX'- T@) - CV~(XG)]XG.
The hold up argument shows that differences between quasi rents fiom the best trading arrangement and next best trading arrangement is specialized quasi rent (Klein,
Crawford and Alchian, 1978). They argue that specialized quasi rent is accmabie only by trading with the best trading partner. Thus, the specialized quasi rent (SQR) for a representative green leaf tea producer is;
SQR = {(P*- CVr(Xd)XG- S }-( (P.- TP)- CdXd)XG- S 1 (5-15)
According to the hold up argument the best trading partner might try to extract the SQR knowing that this portion of quasi rent is available to the other party only through him.
7 In the hold up literature the returns fiom the next best alternative trading parmer is argued to be closer to zero. However, in green leaf tea grower cm sell his green leaf to another processor at P.. However, what the researchers have noted is that due to perishability of green leaf tea hauling to another processor is less likely. Thus, the specialized quasi rent for a green leaf producer is vulnerable to hold up threats
by the best trading partner (processor A). It is possible that processor A might act
opportunistically8 and misrepresent information relevant to the joint profit maxirnizing
price (P') and offer prices, P, lower than P' such that P' -T(D). This is a possible
explanation of how lower green leaf prices is resulted in the arms length transaction and
possible explanation of green leaf producers to avoid the arms length transaction in green
leaf market.
5.4.F. Green Leaf Prices under Hold up Threats & "Buy" Decisions
The above is an possible economic explanation for green leaf prices less than joint
profit maximizing pnce in the Sri Lankan tea industry with the arms length transactions.
Any price, P. that is paid by the processor A, such that P2 P* - T(D) and P2 Cm, is
adequate for the green leaf producer to sell his tea to processor A. Once the investment is
committed to planting tea, the green leaf producer's best choice is to sell tea at this price,
P2 P' - T(Dl rather than switch and sel1 to processor B provided that T(D1 2 0. There
may be extreme situations when T(D) is prohibitively large, and P can be effectively
pushed to its limit such that P = C,.leaving green leaf producers able to just cover their
variable costs.
What would be the impact of the potential hold up problem on the investment decisions in green leaf production by independent green leaf producers? With the joint profit maximizing price, P', a green leaf producer will break-even on the investment in green leaf tea production IG if (P' - Cm(XG))XG - S = O (see equation (5.14)).
8 Hold up argument of Klein, Crawford and Alchian, 1978 do not explain the reputation related losses with the opportunistic behavior. Reputation as a limitation of hold up argument will be discussed in section 5.- Green leaf prkes lower than P' can lead to long run losses on the grower's invesiment
(fixed) cost and eventually exit from green leaf production or expecting the hold up threats, green leaf producen may have avoided in investing in green leaf production.
The impact of the potential hold up threats on the size of green leaf tea production by independent farms is show in Figure 5.4. The unregulated market pnce for green leaf tea could be in the range of C& PS P' depending on the size of Tm.When TfD) is larger, processor A will move towards to the lower bound of CAPS P', and when T(D) is smaller he will move towards its upper bound (see Figure 5.4). Green leaf tea pnces that satisG the condition of C,tr< P5 P' could result in an inefficient equilibrium (&) for vertically independent secton. This is an ineficient outcorne, as shown in the joint profit frontier (see EZ in Figure 5.3). The sum of profits for both parties (green leaf tea producers and processors) in the vertically independent sector could be increased by moving towards to El. However, hold up threats might discourage the entry of independent green leaf producers into the production. The inefficient hold up behavior might have prevented the mutually beneficial exchange opportunities for both parties.
This is a possible explanation of avoidance of amis length transaction in the green leaf market in Sri Lanka and other countries. The 'buy" decisions with prices less than P' lead to lower joint profits, and lower level of green leaf production, relative to the
"make" decisions in which hold up threats are absent. Thus, it is possible that "make" decisions were encouraged and "buy" decisions were discouraged due to the possibility of hold up problem in the arms length transactions in the Sri Lankan green leaf market. 5.4.G. The effect of Pricing Intewention & Changes of "Maken vs. "Buy" decisions
Green leaf pricing intervention was introduced to the Sri Lankan tea industry in
1968. First, it was stipulated a base price regardless of green leaf growers' cost of production (1968 to 1978), then a minimum guaranteed price was imposed and revised several times (1978 to 1984), and since 1985 the pricing formula has been in place, which links green leaf prices directly to the Colombo auction tea prices. The price intervention might impose a lower bound on the green leaf price range of C,I P< P'. instead of Cm, the lower bound would be PI which is the green leaf price after the intervention. The price intervention in green leaf market possibiy restricted the severity of hoId up threats by narrowing down the possible range of prices. The possible effect of state intervention to increase the green leaf tea price to P' (in Figure 5.4) would be an expansion in independent sector green leaf production (2in Figure 5.4). The new equilibrium (EG)is hypothesized by a reduction in hold-up threats and more conducive institutional arrangements for exchange in the green leaf market relative to the inefficiently low level of green leaf market activity at Er.
These outcornes have been witnessed in the Sri Lankan tea industry with the green leaf price intervention. Independent green Ieaf production (measured in black tea production by bought green leaf) expanded fiom about 35 million kg in 1960 to 56.68 by
1978 (about 58%) and increased to 157 million kg by 1998 (about 171%, see Figure 5.1).
These assertions about the impact of pricing intervention on the changes in vertical coordination arrangements could be explained by focusing at the changes of equilibrium output in the vertically integrated and independent sectors along with the total tea production of the country (see Figure 5.5 to 5.7). Pnce of Green leaf
xo X[ X' Quantity of green leaf
Figure 5.4. Changes in Green Leaf Production under State Price Intervention
P' = Joint Profit Maximizing Green Leaf Price X' = Joint Profit Maximizing Green Leaf Quantity P* - T(D) = Green Leaf Price under the Hold up Threats Xo = Green Leaf Production under the Hold up Threats P' = Government Intervened Green Leaf price X' = Green Leaf Production under Governrnent intervention The relative changes in the equilibrium output in the vertically integrated sector and independent sector during the period of 1960 to 1998 are sumrnarized in Figure 5.5 to
5.7. Figure 5.5 shows the Sri Lankan tea industry without the price intervention, Le, fiom
1960 to 1968. Figure 5.6 is related to the regulated market but prior to the formula pricing Le., 1969 to 1984, and Figure 5.7 is related to the green leaf market with formula pricing i.e., 1985 to 1998. Total black tea production as the sum of the equilibrium output in the two sectors is given in panel (c) in each Figure. It is assumed that demand for the Sri Lankan black tea, (D(world)), is unchanged during the penod of analysis which is reasonable given that total Sn Lankan production was stagnant around 230 to
250 million kilos during this period (see Figure 5.1 ). The total industry supply of black tea (S(total)l and world market demand for Sri Lankan tea, (D(worM))).establish the price of black tea (Pp).
The black tea pnce (PQ) determines the amount of green leaf produced by the vertically integrated fartns. The black tea price (Pp) could be transformed to the optimal green leaf price through the relationship shown in the equation (5.13), PX = P*,(PQ.~).
Where P*.~intersects the vertically integrated sector's supply function &(Green).
Determine their green leaf production. In Figure 5.5, optimal green leaf production in the vertically integrated sector is at gb,. In the absence of hold up problems, ceteris paribus, optimal green leaf pnce or the joint profit maximizing green leaf price, P*.~would be observed in the vertically independent sector. However, the vertically independent sector in Figure 5.5, panel (a), is hypothesized to be constrained by hold up threats during the no price intervention regime (1 960 to 1968). Figure 5.5. Tea industry with no Price intervention in Green Leaf Tea (19604968)
(a) Vertically Independent Sector (b) Vertically Integrated Sector (c) Total Production
Figure 5.6. Tea Industry with Guaranteed Price for Green Leaf Tea (1969-1985) Green Leaf Price Black Tea Price Green Leaf Price (mg) S (Total) /
. . a a 1 b 1 I 9 Green ~e&'ea b ~lackea q~geai 2- (Mn kg) Green Leaf Tea Q : 2 green (Mn kg)
(a) Vertically Independent Sector (b) Vertically Integrated Sector (c)Total Production
Figure 5.7. Tea Industry with Formula Priced Green Leaf Tea (1986-1998)
Green Leaf Price Green Leaf Price Black Tea Pnce
b S (Green) 3 ------P
b Green Leaf Tea Q : Black Tea (Mn kg)
q b + q Id' 1 (Mn kg) (a) Vertically Independent Sector (b) Vertically Integrated Sector (c) Total Production An inefficient equilibriurn of green leaf tea production at Po and qat is the result. The
total arnount of green leaf grown is qal and qbl, which is transformed into Q'/ of black tea. In Figure 5.6, the pricing intervention in the green leaf market raises the green leaf
prices to independent producers fiom Po to and green leaf tea production expands to
9%. At the same time, the vertically integrated sector faced with higher labor wages resulting in a shift in their green leaf supply towards the lefi to (green). Thus, the arnount of green leaf produced by vertically integrated sector falls to qb2. The total
arnount of black tea remains the same (Q: = Q'?) but the share of green leaf gomby
independent growen has increased ( Q: < qa6 Q:. These assertions on the impact of pricing intervention in lowering the hold up threats and the consequent effects on the changes in vertical coordination arrangements have to be empirically tested. The following two hypotheses are developed to test these assertions. Hypothesis 3: Ifhold iip threats do lead to lower levels of black tea production a~dgreen leaf output from a vertically independent sector relative to a vertically integrated sector. ceteris paribus, the share of both outprtts from independent sectors should increase with the ment of pricing intervention in the green leaf market which reduces the poten tial for hold rtp threats. Hypothesis: If 60th the verticaiiy independent sector and integrated sector respond to black tea prices identically, ceteris paribus, there shotild not be any systematic rehtionship between black tea prices and the share of black tea production by either sector. However, in the presence of hold up threats. ceteris paribus, the price response by the verticaliy integrated sector wotdd be larger thari that for the verticaily iridependent sector. 5.5. Strengths and Weaknesses of Hold-up Arguments in Green Leaf Market The main weakness of the hold up argument is that it does not take into account the possibility of self-inflicting economic damages to the processors through their opportunistic behavior in holding up green leaf producers' profits. If a processor acquires a reputation for offering low prices, then potential green leaf producers will hear about this. If this is the only processor in the region, those potential green le& growers will grow something else. Then the processor will not have access to supply of green leaf in future. If the processor realizes this then he will offer prices sufficient to attract the supply he needs. Thus with the hold up threats to capture the profits of green leaf producer, processors possibly inflict long run economic damages on themselves. On the other hand, the economic darnages to the processors by having deprived the outside supply of green leaf are related to their requirement of such supplies to run the processing in full capacity. Arguably, economic damages by depriving outside green leaf supply would be larger for a processor who procures his entire green leaf requirement fiom outside suppliers, relative to a processor who does not buy green leaf at al1 and have his own green leaf supply, given everything else is same between these two types of processors. At the time of the introduction of green leaf price intervention in Sri Lanka about 96% of the processors (908 out of 950) were vertically integrated (estate) processos who had their own green le& production. The rest of the processon are independent processors who procure their entire green leaf requirement from the outside green leaf suppliers and they were only 4% (4 out of 950 processors) from the total number of processors (see Table 5.1 ). Table 5.1. Large Farm (Estate) & Independent Green Leaf Processors in Sri Lanka Elevation Class Number of Processors Estate Independent Estate Independent Estate Independent High Grown 316 3 234 1 152 5 Mid Grown 357 17 296 29 99 27 Low Grown 235 22 223 33 130 183 Total 908 42 753 63 381 215 (Sources: International Bank for Reconstruction and Development, 1953 p.231 and Report of the Presidential Commission on Tea trade and industry, 1995 p. 80) While the number of processon in these two categories do not show the nature of economic losses of not having access to the outside green leaf supplies, such economic damages generally would be lower to the vertically integrated processors. In the absence of outside green leaf supply, they might adjust in-house green leaf production to fil1 the gap for the requirement in full processing capacity. Arguably, the reputation related damages by engaging in hold up threats would be less for these processon given everything else is same. Another weakness of hold up argument in the green leaf market is the debatable nature of the switching costs of the investment in green leaf production. The hold up problern arises with the transaction specific investments, which have lqe witching costs and difficulty in finding alternative trading partners. Facing hold up threats, if the grower decided to "pull" out the tea bushes he incurs substantial losses. The cost of estabIishing green le& tea fields is comparatively large among the peremial crops, yet the revenue fiom tea is sarne with other perennial crops in Sri Lanka (see Table 5.2). Tea takes approximately seven years from its initial field establishment to reach harvestable age and this is the second longest waiting period to reach the harvestable age among the perennial crops. A tea garden can last about 50 years providing econornic retums (Etherington, 1971). As noted in Chapter 2, prior to field establishment, about 2 years of soil-rehabilitation is recomrnended (Handbook of Tea, 1986). in addition, there are specific field related investments, such as shade trees, drains, and terraces, which are not useful in the production of any other crops. Together, these aspects make production related sunk costs comparatively larger for green leaf tea production among the peremials. in embarking on such a long-term investment, the green leaf producer should make sure that his crop would generate adequate retums, which are dependent on the pnce offered by the black tea processor. The relatively large investment requirement in green leaf production would make the green leaf growers to be vigilant about the possibility of opportunistic behavior by the processors. Thus, it is likely that if a tea grower expects any hold up threats frorn the processors, he would select some other cropping activity. It is possible that green leaf tea garden could be sold to another tea farmers in the instance of imminent threats of hold- up. How the buyer going to avoid the hold up threat is an important issue which determine the possibility of selling. Table 5.2. Estirnated Cost of Developing and Bringing into Bearing Tea and some Alternative Crops in Sri Lanka (nominal prices in 1968) Crop Cost Annual Ready to Economic Life (RsperAcre) Revenue Harvest Span @/Acre) At Years (Years) Tea 4000-6000 326 6-7 40-50 Coconut Oil Palm Coffee Cocoa Cimamon Cardamom Pepper Cloves Nutmeg and Mace Source: IBRD, Review of World Tea Economy, 1970 (Annex 1, Table 18 & 19) Tilakaratna, 1984 (Table 4, p.22 1) Notes: Coffee, Cocoa, Pepper, Cloves cost up to 5" year including establishment Cinnamon, Cardamom costs up to 3rdyear including establishment Nutmeg and Mace cos& up to 6" year including establishment One another weakness of the argument for hold up threat in the green leaf market in the Sri Lankan tea industry is about the difficulty in finding alternative trading partners. One party can successfully hold up other party's profit only in the instance that the other party could not find alternative trading partner easily. How difficult is a green leaf producer to find another processor in the instance of one processor attempt to hold up the profit? Hold up problem is more likely as the difficulty to find another processor (who likely to buy the green leaf) increases. This difficulty depends on the penshability of green leaf, geographical distribution of the processors, nature of transportation facilities available. Yet, there are no concIusive evidences to show that the importance of these factors in hold up threats. However, some of the following information would be helpfÙ1 in deciding the degree of difficulties of green leaf producers to find alternative trading partners prior to the pricing intervention. in 1951 for the High Grown elevation class, there were only three independent processors within 1563 square miles9, 17 processors within 2326 square miles in the Mid Grown region and 22 processors within 2384 square miles in the Low Grown region. The significance of these distances on the bargaining power of the green leaf producers depends on the condition of the transportation infiastructure. A survey of small tea fmsin 1994 found that about 48% of the green leaf tea producers were located in a non-transportable locality and they have to haul their produce to the near by road. About 27% of these were half a mile away from a transportable road. Only about 12% had their own transportation facilities to transport green leaf to the chosen green leaf producers (Report of the Baseline Survey of Tea Small Holdings, 1994). Some argue that perishability impose a time limitation on finding alternative buyers (Masten, 1991). Related to the green leaf market in Sn Lanka Roberts (1 989) argued that the difficulty in finding alternative trading partners is aggravated by the penshability of green leaf tea. He noted that "since withering starts at plucking and darnage in transport cm prematurely start fermentation, green leaf is rarely transported more than 30 kilometers" (p. 186). The reference of perishability in green leaf tea is related to the chemical fermentationio. The bruises to the green leaf in transportation exposes the enzymes to the substrate and starts the chemical fermentation prematurely, - - - 9 Land area of elevation classes; High Grown = Sq miles 1563 (NuwaraEliya 474 + Badulla 1089), Mid Grown = 2326 (Kandy 9 14tMatale 770.2+ Kegalle 642), Low Grown= 2384 (Galle 652tMatara 48 1 +Ratnapua 1250) 10 Chernical fermentation is carefully handled in tea processing first by maceration of green tea leaves to facilitate the enzymatic reaction between ce11 content and the enzymes (which brings the charactenstic liquor color and quality) along with controlling the temperature and moisture content. adversely affecting the quality of the processed product (Handbook of Tea, 1986 p. 158). Delays and excessive crushing with longer distance transportation trigger premahire and uncontrolled fermentation of the green leaf and result in poor quality at the time of delivery. The Handbook notes "prompt transport of leaf to the factory (processor) is, therefore, essential to minirnize those effects."(p.l58). Moreover, except for green leaf tea, none of the outputs of other peremials requires immediate processing. Perishability is not as important as in tea for al1 the other crops, since these do not need imrnediate processing at the raw produce stage. Thus, perishability would not be a constraint to find alternative buyers for the growen of these other peremial crops. From the green leaf producers' perspective, when the difficulties in finding alternative buyers for his produce is imminent, the investment in green leaf production may become a site-specific asset relative to the investments in other peremials. There is a no test one cm apply to ascertain whether hold up problem (either green leaf producers holding up processors or processors holding up green leaf producers) was a reality. Green leaf and black tea prices before and after pricing intervention could provide a comparison to see whether green leaf prices have increased or decreased. Even then, one has to make allowance for other factors that bear upon green leaf prices. For instance, processors will pay higher green leaf prices for outside green leaf supply, irrespective of the pncing intervention, if in-house green leaf production is costlier relative to procuring fiom outside. Hold up problem is therefore only a possibility, which might explain the avoidance of arms length transaction in green leaf market. The above account of the strengths and weaknesses of hold-up problerns in the green leaf market in Sri Lanka assess this possibility. 5.6. Summary The structural changes which have occurred in the Sn Lankan tea industry are characterized by vertical de-integration of the sector with a switch to independent green leaf production and processing sectors particularly since the 1970s. There is no explanation provided in the literature for what may have caused these structural changes and theoretical reasoning for such causal effects. This chapter has developed several hypotheses to explain the structural shift. The hypotheses are based on the relative incentives associated with the "make" (vertically integrated) versus "buy" (vertically independent) decisions. First, it is expected that there would be a movement toward purchasing green leaf fiom small, independent producers since their relative costs of production has fallen over time. Labor costs, the largest expense item, has increased for vertically integrated units due to the pressures of labor unions. Second, profits are generally greater fur independent firms due to market incentives and lower agency costs. However, the movement towards using the market to purchase green leaf (the "buy" decision) was stymied during the period of state ownership of the large plantations. Thus, it is expected that there would be a movement towards independent production before (before 1976) and after (after 1992) the penod of state ownership of the plantations. Finally, the make vs. buy decisions is influenced by differences in transaction and govemance costs associated with the alternative coordination arrangements. A theoretical mode1 was developed to explain how pricing intervention rnay affect the vertical coordination in the Sri Lankan tea industry. First, it develops a formula for the socially optimal green leaf price that maiiimizes profits for both the green leaf producer and the processor and those profits are equal to the profits in a vertically integrated production. Then it demonstrates that given the perishability of green leaf tea and the difficulty in locating alternative buyers, an investment in green leaf tea production creates location specific assets. When the small holder tea garden becomes a specialized asset to a particular processor, there are specialized quasi rents created. The subsequent potential for opportunistic behavior to capture these rents could lead to (unregulated) spot market prices being less than the socially optimal green leaf price. These lower pnces lead to long mn losses for the small holder producers and may have discouraged them fiom investing in green leaf production. The legal enforcement of the green leaf prices is hypothesized to be a restraint on the hold up threats. This restraint on hold up threats has improved the spot market coordination of green leaf market activity leading to the observed structural changes in the Sri Lankan tea industry. To test hypotheses developed in chapter 4 and chapter 5, the empirical specification of the conceptual relationships is developed in chapter 6. The empirical variables representing both farm size distribution changes and vertical coordination changes and explanatory variable of the drivers of these changes is discussed in chapter 6. The empirical specifications are estimated in chapter 7. CHAPTER 6 EMPIRICAL MODELS OF STRUCTURAL CHANGES IN SRI LANKAN TEA INDUSTRY in the previous two chapters, conceptual rnodels were developed to explain the changes in fmsize distribution (chapter 4) and changes in vertical coordination arrangements (chapter 5) in the Sri Lankan tea industry. The purpose of this chapter is to speciS the empirical models to test the hypothesized effects of the drivers of the changes in these structural dimensions and to describe the variables representing the structural dimensions and the dnvers of their changes. The first section specifies the empirical model for the changes in farm size distribution, describes the variables representing size distribution change and the drknof its change, and reviews the hypotheses in Chapter 4. The second section does the same for the changes in vertical coordination arrangements. The final section sumrnarizes the main points of the chapter. 6.1. Empirical Model for the Changes in Sue Distribution of Sri Lankan Tea Farms In this section, the measurement of farm size distribution variables, and its eight explanatory variables is presented. In addition, the descriptive statistics for al1 eight variables are briefly discussed while annual data values for each variable from 1960 to 1998 are given in Annex 6A to 6E along with the data sources. In Chapter 7, the specified model is estimated and econometric results are discussed. 6.1.A. Dependent Variable for Sri Lankan Tea Farm Size Distribution (Y1) Fann size could be measured using a variety of indicators as discussed in Chapter 3. Land area, as a measure of fmsize, is common in almost al1 tea producing countries (IL0 1966). The categorization of farm size groups in the Sri Lankan tea industry is based on the official legal definition of the terni for "plantation" or "estate" which, according to the Estate Labor (Indian) Ordinance of 1889, was a tea fmnot less than 10 acres in size (IL0 1966). Tea farms less than IO acres in size were viewed as small holders or small farms. Ever since, these arbitrary size groups have been entrenched in the literature as the measurement of tea fann sizes. Censuses in Sn Lanka have used the 10 acre size as delimiting small and large tea farms and has also Mersub-grouped large tea fms(10-100 acres, 100-500 acres, and larger than 500 acres, see Figure 2.10). However, when the land reform program was introduced in the mid 1970s, the legal ceiling of 50 acres becarne important in categonzing smallholdings. Thus, since 1978, small holders were redefined as tea farms less than 50 acres in size without their own prxessing facilities. The tea farms between 10 acres and 50 acres were reported to be about 10% of total tea area of the country in 1978 or about 60,000 acres (Report of the Cornmittee on problems of the tea small holdings sector, 1980 p.11). Since the definition of srna11 holders and estates changed in 1978, it is not possible to use the tea census to obtain a consistent measure of farm size over time. The other source of data on size distribution is aggregate acreage for the tea ims smaller than 10 acres in size and larger than 10 acres in size as provided in the Bulletin of Statistics of International Tea Council (see Annex 6A). These two acreage aggregates are taken in this study to represents large and small tea fms. The dependent variable assessing changes in the size distribution of tea fms is approximated by the share of area planted by small farms (40 acres) of the total tea area of the country. The data fiom 1960 to 1998 on tea area planted by these fmsize categories and the relative share planted by small fmsis listed in Annex 6A. For exarnple, in 1960 tea fmssmaller than 10 acres in size covered 79,466 acres of the total tea area of 58 1,821 acres so the relative share of small tea fms was 0.1366 (Le, 79466/58 182 1). The relative share of total tea area grown by small farms has grown from 13 -7% in 1960 to 3 5% in 1 998 and the average small farm share is about 22% dunng this period. 6.1 .B. Explanatory Variables for Farm Size Distribution Changes There are four major drivers used in the conceptual mode1 of fmsize distribution changes (see Figure 4.1): 1) technological factors, 2) public prograrns and policies, 3) prices and financial sector factors, 4) national economic conditions. In Chapter 4, specific factors under each of the above four groups were discussed and the hypothesized effects of these factors on the changes in tea farm size distribution were derived. These specific factors (drivers) are: (1 ) Technological factors (1). Nature of labor capital substitution and opportunity cost of labor (2) Public Programs and Policies (0. Land reform polices (m.Small farm support services (III). Tea sector taxation (IV). State intervention in green leaf market (3) Prices and financial sector factors (0. Price instability in export markets (4) National Economic Conditions (0. Expansion in non-traditional export crops. The notations and definition of these variables and their hypothesized effects are summarized in Table 6.1. Table 6.1. Factors Affecting Relative Share of Farms sues: Variable Description and the Expected Signs of the Coefficients Variable Notation Expected Sign Dependent Relative share of small tea farms y1 Explartatory Ratio of opportunity costs of labor for small and large WR - tea farms Dummy for the land reform policies Land Refonn + Small tea fam subsidies Support + Rate of export taxation on tea Export Tax - Dummy for the pricing intervention (1969-1984) Price Inter 1 + Dummy for the pricing intervention (1985- 1998) Price Inter II + Price instability in Colombo tea auction Price Var - FOB price index for alternative crops Alt Ret - Ratio of Smnll/arrn and Large Fam Labor Opportunity Costs (WR) Changes in labor oppominity costs for large fatms and small fmsis discussed in Chapter 4 (4.1.B). Large fmlabor wages has been influenced by the labor unions and these were appeared to be kept above the labor market clearing wages. The opporhinity cost of labor For the large tea fmsis taken as the plantation sector daity (real) wage rate while that of small tea farms is taken as the industrial sector daily (real) wage rate obtained fiom annual reports of the Sri Lankan Central Bank. An ancillary assumption is that the owner-operator of the small farms is able to obtain the industrial sector employment without additional costs (commuting expenses for the industrial centers from rural area, search costs etc., are zero). The ratio between industrial sector wages and plantation sector wages is the proxy variable that represents the change in the opportunity cost of labor between small and large tea farms. As the relative labor costs change, it is hypothesized that the relative profitability shifts towards to the tea fms with the less expensive labor. Annual data collected hm1960 to 1998 is presented in Annex 6B. The average of the wage rate ratio is 1.3 but it has dropped fiom the highest value of 1.85 in 1971 to a low of 0.91 in 1997. lmplernentation of Land Reform Policies (Land Reform) A dummy variable (Land Reform) is introduced to capture the effects of land refonn policies on the tea fmsize distribution changes. Land reform policies are reputed to favor expansion in small farm through redistributing land of the large hs. In addition, state management of the large fms, which was arguably a period of mismanagement of large farms, begun with the land reform policies. Both these aspects are hypothesized to favor the small fams while these are negatively affected on the share of large fms. The period fiom 1960 to 1975 is before the land refom policy (Land Refom = O) while Land Reform=l for the years from 1976 onwards when the policy was in effect. Small Tea Fann Support Services (Support) Support services for small tea farms range fkom extension services to fertilizer subsidies. However, the main plank of support services has been monetary subsidies provided for small fmrehabilitation, for fertilizer application, and for replanting. Small tea fmsubsidies per hectare (1952 constant) is taken as a representative measure of the state support extended to small tea fanns. Total subsidies allocated for the small tea fmswas divided by the small tea farm area (tea acreage of -40 acres size farms) to calculate the subsidy per unit land area. Data fiom 1960 to 1980 had to be gathered fiom various sources and some of the values are crude approximations of the small fann subsidies. Al1 the data sources and the approximations are explained in Annex 6C. Data fiom 1980 to 1998 are more reliable as these are recorded yearly in Plantations Sector Statistical Pocket Book. Small farm support services could increase the relative share of small farrns and these supports are not extended to large tea fms. The support services variable has shown substantial variation and averaged at 101.38 (1952 Rs/ha) and recent values are the lowest around 40 while the highest value was for 1986 at 191. Rate of Wot-tTaxes on Tea (Export Tax) Taxes on exported tea include fixed export taxes (specific tax), ad-valorem taxes (depend on the tea pnces in the auctions) and an export cess (cess is a levy on tea exports, which is used to finance production subsidies plus, tea research and marketing efforts of the Sri Lanka Tea Board). The swnmation of these three different taxes is assumed to represent total export taxes. The tax per kilo of tea exports or rate of export taxes was calcuiated by dividing this total tax by total tea exports. The rea1 value of taxes per kilo of tea exports was obtained by deflating the nominal tax rate by the Colombo Consumer Price Index. This tax rate was calculated ftom 1960 to 1998 and al1 the data sources and the approximations are explained in Annex 6D. The export taxes are hypothesized to be disproportionately affect the production sector due to the nature of the export demand elasticity and supply elasticity of pnmary producers. Export taxes depress the producer pnces and small farms are more flexible in responding to such sustained price depressions compared to large farms. The export tax rate averaged at 1.507 (1952 Rskg of exported black tea). However, the lowest is at 0.099 in 1967 and highest being 7.822 (1952 Rskg) in 1978. State interventions in Green Leaf Market (Price hter i and Price hfer1. Dummy variables (Price Inter 1 and Price Inter II) are introduced to capture the effects of green leaf tea price interventions on the farm size distribution changes. As discussed in Chapter 4 (4.2.D.l), the period from 1960 to 1968 there was no intervention in the green leaf market. From 1969 to 1978, there was a formula for sharing black tea pnce, which did not guarantee a minimum pnce for the growers, since 1978 a minimum price was granted, yet due to the inflationary pressure the minimum price was adjusted upward time to time. The Reasonable Price Formula was put into effect in 1985 and this was most influential in sharing the auction prices for black tea between the growers and the processors. The price intervention is hypothesized to reduce the hold up threats in the green leaf market. This in tum has facilitated the market exchange process and has expanded the equiiibrium output of green leaf tea fiom inefficiently low levels of green leaf production by small, independent green leaf producers to a larger level (see Figure 5.5 to 5.7). It is also asserted that, Reasonable Price Formula is the most effective regirne of price intervention in allaying the hold up threats, since it directly links the green leaf prices to black tea prices. Two dummy variable are introduced to differentiate the three periods of price interventions i.e., without pricing intervention (1960 to 1968), pricing intervention other than Reasonable Pricing Formula (1969 to 1984), and pricing intervention with Reasonable Price Formula (1985 to 1998). Thus, Pnce Inter 1 equaI to one for the period of 1969 to 1984 and equal to zero otherwise. Price Inter II equal to one for the period 1985 to 1998 and zero otherwise. It is hypothesized that mean value of the relative share of small farrns is largest in the period of Pnce Inter II and smallest in the base (without price intervention) period. Tea Price Imtabifity in Colombo feu Auction (Price Var) Short-term price instability is usually measured by some method of capturing the deviation of the prices fkom the price trend. Following Athukorala and Huynh, 1987, a five year moving average is calculated as the trend in the Colombo auction tea prices. The deviations are calculated as the percentage difference fkom the actual price and the five-year moving average. A two years lagged of the price instability index is taken assuming instability in the past two years captures the size of risk to which small and large fmsface with different competencies. It is hypothesized that large fmare better equipped to face short-term liquidation crises so small fmsare adversely affected with larger pnce fluctuations. Annual data are collected from 1960 to 1998 and the data along with the data sources are explained in Annex 6E. Percentage deviation from the price trend is sometimes as large as 79 and sometime as Meas 0.073. There is no perceptible trend yet, in 1970s there were large fluctuation (both the largest and srnallest values are in 1970s). Rettîrns from Alternative Agric~îlturalEnterprises (Ait Ref) Tea growers respond to economic returns fiom alternative enterprises. Tea growers would reallocate their pnmary inputs if these inputs could be employed more gainfully in alternative enterprises. The most important pnmary inputs in growing tea are land and labor and the alternative enterprises in which these primary inputs could be employed are assumed cropping and livestock enterprises. The alternative cropping enterprises were selected on the basis OFwhether they represented reasonable substitute to tea growing. Therefore, cropping activities with higher imgation requirement (paddy and some other field crops), with larger capital or labor intensities (horticultural crops), and with more centralized post-harvest processing requirements (sugar cane, palm oil) were not considered. Given the availability of pi-ice data for the pe60d 1960 -1 998, six non- traditional peremial export crops (pepper, cimamon, cloves, nutmeg, cardamom, and cocoa) are selected as the alternative cropping enterprises. These crops have been proposed as alternatives for unprofitable tea growers (Jacob, 1984 and National Agriculture, Food and Nutrition Strategy, 1984 p.69-73). The profitability of these crops is captured by their FOB prices. Since the export volumes of these crops are comparable (see Annex 6F), average of FOB prices (1952 constant) are taken to represent the profitability of these alternative crops. To capture the relative pro fitabiiity changes the average FOB price for these alternative crops is divided by black tea price. Annual data and the data sources are explained in Annex 6F. Responses to the profitability changes in alternative crops are hypothesized to be different between small and large hs. Small farms are more flexible (less fixed costs) to respond to such changes compared to the large hs. These alternative crops has show a wide variability in their FOB prices (minimum is 4.2 (1952 Rskg) in 1960 and maximum is 37.08 in 1979 and standard deviation is 8.56) however, it has shown a increasing trend until 1980s and since a gradua1 decline. 6.2. Empirical Model for the Changes in Vertical Coordination in Tea Industry In this section, measurements for the changes in vertical coordination arrangement and its five explanatory variables are presented. In addition, the descriptive statistics for al1 five variables are briefly discussed with annual data values for each variable from 1960 to 1998 given in Annex 6F to 6E along with the data sources. in Chapter 7, the specified mode1 is estimated and econometric results are discussed. 6.2.A. Dependent Variable for the Vertical Coordination Arrangements (YI) The changes in vertical coordination arrangements within the Sri Lankan tea industry are proxied by the amount of black tea produced from market-procured green leaf (vertically independent production) as compared to the amount of black tea processed fiom self-grown green leaf (vertically integrated production). The amount of black tea processed fkom market-procured green leaf and self-grown green leaf is available for years after 2980 fiom Plantation Sector Statistical Pocket Books. As discussed in Chapter 2 (see Table 2-13), black tea production fiom market procured green leaf has increased fiom 53 million kilos in 1980 to 157 million kilos in 1998. However, the data on the source of green leaf for processing are not directly available for years pnor to 1980. Estimates of black tea production fiorn market procured and self-grown green leaf for the period of 1960 to 1980 are based on elevation production levels. The main supplier of market-procured green leaf is the small tea fm and these farms tend to dominate the Low Grown elevation class (see Table 2.10 and 2.12 in Chapter 2). Thus, it is assumed that the amount of black tea processed with market-procured green leaf has to be associated with tea production in the Low Grown Region. Figure 6.1 depicts this association for the period fkom 1980 to 1998. There is a significant correlation between black tea production fiom market- procured green leaf and tea production in the Low Grown region (Pearson correlation coefficient is 0.9977). Figure 6.1. Black Tea Production: Low Grown and from Purchased Green Ieaf Source: Plantation Sector Statistical Pocket Book, 1999 (p. 24,25, and 139) Thus, the amount of black tea processed from bought green leaf is estimated as dependent variable in a linear regression with the amount of black tea produced in the Low Grown region as the explanatory variable. The estimated equation using the available data fiom 1980 to 1998 is; Black Tea from Boright Green Leaf = - 17 .O 12 + 1-1 45 87 (Lotv Gro wn BIack tea Prodric fion) (-4.6797) (3 1.74) Adj R' = 0.982 (t statistics are given in the parenthesis). Data on Low Grown black tea production from 1960-to 1980 were then plugged into the linear regression equation to extrapolate the amount of the black tea produced from bought green leaf. The extrapolated data fiorn 1960 to 1979 and actual data fiom 1980 to 1998 of black tea production fiom market-procured green leaf are given in Annex 6G. The increase in black tea production fiom bought green leaf fiom 1960 to 1983 was about 54% while the increase fiorn 1983 to 1998 is about 2 85%. Average production of black tea fiom purchased green leaf is about 69.80 million kilos of black tea with the maximum is observed in 1998 at 15 7.17 mi Ilion kilos. 6.2.B. Explanatory Variables for Changes in Vertical Coordination Arrangements The conceptual model in Chapter 5 derived a hypothesis about the effect of the dnvers (factors) that influence the changes in relative production shares between vertically integrated sector (make decision) and vertically independent sector (buy decision). There are three major drivers used in the conceptual model of Chapter 5: 1) Changes in labor-related production costs between vertically integrated and independent sector, 2) Changes in incentives and agency costs in the vertically integrated sector, and 3) Changes in transaction costs and hold up probIems in the green leaf market. In Chapter 5, specific factors under each of the above dnvers were discussed and the hypothesized effects of these factors on the changes in vertical coordination arrangements were derived. These specific factors are: (1). Changes in labor- related production costs (I). Production costs advantages for vertically independent tea production through lower labor costs. (Ratio of labor wages small and kuge fms) (2). Changes in incentives and agency costs in the vertically integrated sector (1). During the state ownership in the tea industry, vertically integrated production was declined. (Durnmy variable for the period of state ownership) (3). Changes in the transaction costs and hold up problems in the green leaf tea market (1). Decrease in hold up potentials in the green leaf tea market with the state pricing intervention (Dummy variables for pncing interventions) (II). Differential price responses due to hold up problems in the vertically integrated and independent sectors (Black tea price). The notations and definitions of these variables and their hypothesized effects are sumrnarized in Table 6.2. Each of these is discussed in fùrther detail below. Table 6.2. Factors effecting on the Share of Tea Production in Vertically Independent Sector: Variable Description and the Expected Signs of the Co-efficient Description Notation Expected Sign Dependent Share of black tea production in the vertically independent sector (Mn Kg) Explunatory Ratio of opportunity costs of labor for vertically WR - independent and integrated tea fms Dumrny variable for the Period of State Ownership State + Dummy variable for the Pricing intervention 1 Price Int 1 + Dummy variable for the Pricing Intervention II Price Int II + Black Tea Prices Price - Ratio of Opportzrnity Costs of Labor For Vertically Independent and integrated Farms Changes in labor opportunity costs for vertically integrated fms and vertically independent small fmswas discussed in detail in Chapter 4 (4.1.B). The opportunity cost of labor for the vertically integrated sector is taken as the plantation sector daily (real) wage rate while that of vertically independent sector is taken as the industrial sector daily (real) wage rate. The data were obtained from the annual reports of the Sri Lankan Central Bank. The ratio between industrial sector wages and plantation sector wages is the proxy variable that represents the change in the opportunity cost of labor between small and large tea fms. As the relative labor cost of vertically integrated sector increases, i t is hypothesized that the vertically independent tea production would be relatively more profitable, given everything else stays constant. Annual data collected nom 1960 to 1998 is presented in Annex 68. The average of the wage rate ratio is 1.3 but it has dropped fiom the highest value of 1-85 in 1971 to a low of 0.91 in 1997. State Oronershïp of the Verticaiiy integrated Sector (State) A dummy variable (State) is introduced to capture the effects of distorted incentives under the state ownership in the verticalIy integrated sector. The management incentives did not play an important role under the state ownership of the large plantations. Thus, it is argued that under the state ownership of vertically integrated plantation sector the relative pro fitability and the viability of the vertically integrated production declined relative to privately managed vertically independent sector. Thus, it is hypothesized that under the period of state ownership the share of black tea production by the vertically independent sector has increased relative to the vertically integrated sector. The penod fiom 1976 to 1992 is the period under state ownership and management, thus State = 1, while private management existed fkom 1960 to 1975 and 1993 to 1998 so State = O. It is expected to see positive relationship between the share of vertically independent sector production and the State dummy variable. State Interventions in Green Leaf Market (Price Inter I and Price inter 10 Durnrny variables (Price Inter 1 and Price inter II) are introduced to capture the effects of green leaf tea price interventions in allaying the hold up problems and expanding the share of black tea production in the vertically independent sector. During the penod fiom 1960 to 1968, there was no intervention in the green leaf market. From 1969 to 1978, there was a formula for sharing black tea price, which did not guarantee a minimum price for the growen. A minimum price was granted in 1978 but it did not account for the inflationary pressure so the minimum price was adjusted upward time to time. The Reasonable Price Formula was put into effect in 1985, which specified a sharing rule of the auction prices for black tea between the growers and the processors. The price interventions are hypothesized to reduce the hold up threats in the green leaf market. Two dummy variables are introduced to differentiate the three periods of government involvement in the green leaf market; Price Inter I is set equal to one for the period of 1969 to 1984 and equal to zero otherwise. While Price Irzter II equal to one for the period 1985 to 1998 and zero otherwise. Tt is hypothesized that during the period of unabated hold up threats, the share of black tea production by the vertically independent sector would be smallest. In contrast, the pricing intervention is expected to increase the amount of black tea processed with purchased green leaf. The reduction in the hold up potential is hypothesized to be the greatest during the period of Reasonable Price Formula so that positive coefficient of Priceinter II is expected to be larger than PriceIn ter I. Black tea Prices (Price) Higher black tea prices should increase green leaf production for both vertically independent growers and vertically integrated growers. However, this response is dependent on the share received by the green leaf producer for his green leaf. The share is influenced by the potential for hold up problems yet this is not relevant to the vertically integrated sector. The response by the vertically independent sector could be constrained by hold-up potentials. If independent green leaf growers do face the threat of hold-up, it is hypothesized that the own-price response of the vertically integrated sector is larger than that for the vertically independent sector. Thus, the share of green leaf production by the vertically independent sector is negatively related to the black tea pnces, given everything else stays constant. A three-year lag of black tea price is used to capture the effect of black tea pricess on the relative production shares. The average Colombo auction black tea prices (deflated by Consumer price Index, 1952 Rskg) were obtained from the annual report of Central Bank of Sri Lanka and are listed in Annex 6H. These pnces ranged from a Iowest in 1966 of Rskg 2.62 to a highest of 1 1-63 (1952 constant prices) in 1984. The average was about 4.83 Rs/kg. 6.3.Summary This chapter explained the empirical specification of the dependent variable and the explanatory variable for the models of farm size distribution changes and changes in vertical coordination arrangements. The variable measurements were explained and general trends were discussed. The hypothesized relationships between the dependent variables representing the two dimensions of the structural change and the explanatory variables were briefly explained and summarized in Table 6.1 and 6.2 respectively. These specifications will be tested and econometnc results will be discussed in Chapter 7. CHAPTER 7 ESTIMATION OF THE EMPIRICAL MODELS 7.0 Introduction This chapter explains the results of the econometric estimation of empirical models specified in the previous chapters, the changes in fmsize distribution (chapter 4), and the changes in vertical coordination arrangements (chapter 5). The data and empirical specifications for these conceptual models were developed in chapter 6. The primary objective of the econometric estimation is to test the hypotheses about the effects of structural change drivers discussed in the conceptual models. The chapter is organized as follows. Section 7.1 discusses the functional form specification, Box-Cox transformation of the variables, and the flexibility in functional specification. Special econometric properties of the estimated parameters with the Box- Cox regression are noted. Section 7.2 reviews the empirical model of the tea farrn size distribution changes and summarizes the results of econometric estimations. Estimation results are contrasted with the expected theoretical relationships. The competing functional forms for the empirical specification are tested with the Box-Cox regression. Sections 7.3 reviews the empincal model of the vertical coordination changes and summarizes and discusses the results of the econometric estimations as in Section 7.2. in Section 7.4, the main findings of the analysis are summarized. 7.1 .Functional Form Specification of the Econometric Models Economic theory does not guide the selection of appropriate functional form for economic relationships (Zarembka, 1974). Mis-specified functional foms violate some of the assurnptions of estimation procedures and give inconsistent and biased estimators. inferences with such mis-specified iùnctional forms are erroneous. The Box-Cox transformation is an econometric application for functional form selection (Davidson and MacKimon, 1993). It enables the estimation of hctional form parameters that best fit the data. Transformation of variables introduced by Box-Cox (1964) is as follows; There are three types of different transformations within the general model of 7.1 (Davidson and MacKinnon, 1993 p. 483). In the first type, only the dependent variable is transformed. In the second type, variables in both sides of the regression hnction are transformed. in the third type, the dependent variable and some of the independent variables are transformed. The transformation could be done by using the same A. for al1 the variables (dependent and al1 the independent) or with different As for each variable. Thus, the Box-Cox transformation could estimate a linear model (when al1 the variables are transfonned by A= 1), a logarithmic model (when al1 the variables are transformed by k O), a semi-log model (when the dependent variable is transformed by A= 1 and al1 the independent variables by A= O), and a log-inverse model (when the dependent variable is transformed by A= O and independent variables by k1). More importantly, the transformation parameter (A)could be estimated with the Box-Cox regression so it best fits the data. Imposing various restrictions on the values of A. then allows for a variety of fiinctional forms to be tested. Consider estimating the parameters in a relationship with transformed variables; The standard practice in Box-Cox regression is to assume the errors (et's) are normally distributed and to apply maximum likelihood procedures (Davidson and MacKimon, 1993). However, the nonnality assurnption of the errors is incompatible with the transformation. This is because A (when R # 1 or O) would bound the possible upper and lower values of ((#-'-l )l A)). For instance, when &O the value of ((yA'-1)l A)) cannot be greater than -l/ A (see Davidson and MacKinnon, 1993 p. 484). Thus, the errors are distributed as a truncated nonnal rather than as a normal distribution in the range of -oc and + a. To proceed with maximum likelihood estimation, it is assumed that the truncation effects are negligib le and e,*s are independent, identicall y distributed normal random variables with zero mean and d variance. The log -1ikelihood function of equation 7.2 would be; The last term in the right hand side of equation 7.3 is the log of the Jacobian tenn (dei/&) when Yi is transformed to ((yA-1)/ A)). T is number of observations. The first step of estimation procedure is to find an estimate of A to maximize the log likelihood iünction. Then the estimated R is used as a "known" fixed value to obtain maximum likelihood estimators of fi and 2; Thus, maximum likelihood estimators of A and fi and d are not jointly determined and @ and b are conditional on the value of A. The covariance matnx of the P coefficients estimated kom &AKX '"x '"1-' is conditional on the value of A. These conditional standard errors are underestimated since the estimation procedure ignores the variability caused by A (being an estimate rather than known fixed value). The conditional standard errors are lower than the standard errors computed using the actual second derivatives of the log-likelihood function (inverse of the information matrix) (Green, 1993 p. 33 1). However, Spitzer, 1982 argued that in practice, the expected value of the second derivative matrix could not be evaluated for functions, which contain the Box-Cox transformation (p. 308). The econometric implication is that the conditional standard errors would overstate the reliability of the estimators of P with larger t- ratios and reject too kw nul1 hypotheses. The inferences would be erroneous with such estimators. Spitzer (1984) suggested a leeway fiom the above difficulty of estimating the valid covariance matrix for hypothesis testing. He showed that transforming the mode1 to nonlinear least squares would allow the estimation of covariance matrix using the second derivative of the log likelihood fùnction much more easily. in order to convert to nonlinear least squares approach, Spitzer showed that one has to re-scale the data by dividing each variable by its geometric mean. The Box-Cox regression of the re-scaled data would get nd of the problem of invalid standard errors and the problem of overstated reliability of the estimators of P coefficients. Such a re-scaling leaves the parameter A unaffected. Yet, P coefficients will now have a different interpretation. These linear coefficients cannot be interpreted as the partial denvatives of the independent variables with respect to the dependent variable. These linear coefficients are now the point elasticities estimated at their geometric mean. Their standard errors become valid estimates with the re-scaled data compared to the underestimated conditional standard emrs with the original data (Spitzer, 1984). Inferences and hypotheses testing would be legitimate with the valid standard errors that are estimated with the re-scaled data. The re-scaling (dividing through each variable's by its geometric mean) is applied to the data set of both models, prior to the Box-Cox regression using SHAZAM version 7.0. 7.2. Estimation of the Empirical Mode1 of Farm Size Distribution changes The empirical mode1 of the fmsize distribution is specified as follows; Farm size distribution (Y,) is represented as the share of small tea farms from the total tea area of the country (area of tea fms smaller than 20 acres in size/ total tea area). WR represents the ratio of labor opportunity cost of srna11 Farms and large fms (daily wage rate of industrial sector workeddaily wage rate of plantation workers (1952 Rsiday)). Support represents the small farm subsidy (total subsidies per hectare (1952 Rslhectare)), while ExportTm represents export taxes on black tea (total export taxes per kilo of black tea (1952 Rskg)). Black tea pnce instability is approximated by Pricevar (percentage deviation of the black tea price from 5 years moving average). AltRet represents the retums for the alternative crops divided by black tea price (average FOB pricedprice of black tea (1952 Rdkg)). The durnrny variables are for the introduction of land reform policies (La~tdReform=O for 1960 to 1975 and LandReform=l for 1976 to 1998), price intervention before the reasonable pricing formula (PriecInterI=l for 1969 to 1984 and othenivise zero), and pnce intervention with the reasonable pricing formula (Priceintll=l for 1985 to 1998 and otherwise zero). The error terrn U is assumed independentl y distributed with zero mean and constant variance. The Box-Cox transformation cm be applied only to the variables with strictly positive values and dummy variables have no meaningful transformation (Davidson and MacKinnon, 1993). Thus, except for the intercept dumrnies (LandRefonn, Pricehted, and PriceIntdi), al1 the other variables are re-scaled by dividing through their geometnc means before canying out the Box-Cox regression. Annual data for the period of 1960 to 1998 comprising 38 observations are used in the econometric estimation. The following expected signs for the estimated coefficients are implied by the hypothesized effects of the drivers on the change in tea fmsize distribution; Pi, P3, Pq, and ps< O, while pz, Pa, P7, and Pg >O- The maximum likelihood estirnates of the Box-Cox regression for the above model are given in Table 7.2. The estimated model explains the variability in the fm six distribution satisfactorily as the adjusted R' is 0.96. The model's F-value is significant at a 2 % level. The Durbin-Watson statistics is 1.7823 which falls into the zone of indecision about the autocorrelation since the upper and lower bounds of the Durbin- Watson statistics (with 38 observations and 8 explanatory variables) are respectively 1.029 and 2.01 7 (Gujarati, 1995). In addition, five of the eight estimated coefficients are statistically significant and these al1 have signs consistent with the theoretical propositions. These will be explained later. 7.2.A. Testing for Competing Functional Forms Box-Cox regression is a convenient econometric tool to test and reject one or more of the competing functional forms. The asymptotic likelihood ratio test could be used to test these competing models. Likelihood Ratio (LR) Statistic = - 2 (Log of Likelihood (n)- Log of Likelihood (A)) A (haf)is for the restricted Box-Cox regression, while A (tilda) is for the unreshicted Box-Cox regression. The LR statistics corresponds to a X'r distribution where r is the number of restrictions in Box-Cox regression. Since A=1 for the three dummy variables, there are three restriction in the Box-Cox regression. The likelihood ratio test for fiinctional forms of linear (A=l), and logarithrnic @=O) is shown in Table 7.1. Table 7.1 Test for Competing Functional Forms for Farm size Distribution Changes Ho: h=l Ho: A=O Log of Likelihood Function 46.667 64.4587 LR Statistics 38.38 5-46 2 (at 0.01 probability) 9.210 9.210 3 Test Result Reject the linear function Do not reject logarithrnic function 7.2.B. Interpreting the Results of Econometric Estimation The coefficient estimates for the ratio of opportunity cost of labor for small and large tea farms (m),tea export taxes (ExportTax), and price interventions in green leaf market (Pricelnter 1 and Pricelner II) are consistent with the expected relationships and significant at the 1% level. Table 7.2. Maximum Likelihood Estimates for the Farrn Sue Distribution Changes -- - Variable Estimated Asymptotic P-Value Consistent with Coefficient T-Ratio Expected Sign Constant Po= -0.13600" -4.355 0.000 Not Applicable WR Pi= -0.51903" -6.097 0.000 Yes Support Pz = 0.00055* 2.275 0.03 O Yes ExportTax P3 = -0.07288" -7.117 0.000 Y es Price Var P4 = 0.00069 O. 1633 0.871 No AIt Ret Ps = 0.05337 1 -798 0.083 No Land Reform B,= 0.0 1 840 O. 1567 0.877 Yes Pnce inter 1 PT= 0.16000" 3.460 0.002 Yes Price inter II p8=0.1151 1" 3 .496 0.002 Yes Adj R'= 0.96, Model F value =128.82", Log of the Likelihood Function = 65.85 Box-Cox Parameter (A)= -0.32, Durbin-Watson statistics = 1.7823, and *' denote statistical significance at 5% and 1% respectively. The small fam support service (Support) is significant at the 5% level. The dummy variable for land reform policy (LandReform) also has the expected sign yet it is not significant at the 5% level. The price instability index in Colombo tea auctions (PriceVar) and returns fiom the alternative crops (AltRer) have signs opposite to that expected, yet these coefficients are not statistically significant at the 5% level. The negative sign for the coefficient for WR, (the ratio of labor opportunity cost between small fmsand large fms), indicates that as the relative labor costs for the small farm sector decreases their share in the total tea area goes up. It has been observed that since the 1980s the relative opportunity cost of labor for the small tea fms has decreased rapidly in the Sri Lankan tea industry (see Figure 4.3). Thus, smaller tea farms have become relatively less-expensive method of producing green leaf tea compared to the large (plantation) farm sector. This has been an important factor in the observed changes in the fmsize distribution. The argument about labor wages being independent of labor productivity in large tea farms has been center stage in the Sri Lankan tea industry (Betz, 1989, Si~atharnby, 1993, World Bank, 1997). It has been also argued that governent intervention in detemining plantation sector labor wages for politically motivated reasons has caused the discrepancy between labor productivity and the wages (Report of Presidential Commission on the Tea Industry and Trade, 1995 p. 148- 150). This Report noted that government intervention to control wages and other terms of employrnent prevent management from extracting higher productivity from labor and any further increases in costs without a substantial increase in productivity is the "surest way of killing the tea industry" (p. 150). While labor costs for the large farms sector have been pushed due to union pressure (with governrnent support), the differences in profitability as measured by WR are separated further by the differences in labor productivity on small and large fmfor a given wage rate. Schultz (1964) argued that the economic basis of inefficiency with absentee ownership of plantations is due to the impossibility of routinizing the decision- making involving spatial, seasonal, mechanical and biological subtleties in agriculture. The decisions to deal with these subtleties cannot as a rule be made efficiently under absentee arrangements for the simple reason that it is not possible for the absentee parties to become sufficiently informed. Labor productivity is dependent on the incentives to use this knowledge in the absence of ownership. Yet, it is argued that in general, absentee parties have not been successfül in developing necessary incentives for improving labor productivity (Schultz, 1964). Thus, Schultz argues where there is cornpetition, that part of fming under absentee arrangement decreases relatively because of inherent inefficiency (p.119). The small farm sector, which is rnainly based on their farnily labor, is argued to be efficient in their labor input since they are the residual claimants of their effort, receiving the total product of their labor rather than a fixed wage (Key and Runsten, 1999). On the other hand, labor supervision on large farm can not completely deter shirking if it is impossible to perfectly observe individual labor efforts. Thus, apart from a higher labor opportunity cost for the large farm due to unionized labor force, the inherent inefficiency in labor use put Iarge tea farms in a disadvantageous position relative to small farms. The changes in relative profitability between the large farms and small fms related to the opportunity cost of labor is therefore an important economic force explaining the changes in tea farm size distribution in Sn Lanka. The positive association between the small Farm support services and the share of small tea farms is as expected. It is argued that large farms have an advantage over small fmsdue to the pecuniax-y economies of size associated with credit availability (due to larger equity based to draw against), input procurement (quantity discounts for bulk purchase), and output marketing (quantity related premiums). Government provision services to small farms such as credit, output marketing arrangements, and dissemination of technical information can offset the advantages of the large fmsand enable small fmsto become more cornpetitive and maintain their share of production. In developing world agriculture, state provision of such services has been integrated into the goals of agricultural development murlow and Jayasooriya, 1986). A widely cited example of small fann expansion associated with the state provision of the above services is the Kenyan tea small holding sector (Lamb and Muller, 1982, Lelle and Agarawal, 1983, Leonard, 1991, and Little and Watts, 1994). The Kenyan Tea Developrnent Authority is a parastatal organization, which provides planting rnaterials, chemicals, fertilizer, and processing services for small holders. The rapid expansion of the Kenyan tea small holding sector has been attributed to the services provided by this parastatal organization. The general pattern of small fmsupport services by the state and the consequent expansion in the small farm sector relative to the large hm sector has also been noted for many other developing countries (Burlow and Jayasooriya, 1986, Sajhau and Muralt, 1987 see chapter 5). The regression results demonstrate that the provision of support services to srna11 farms is also positively related to the share of small tea hsin the total tea land area of Sn Lankan tea industry. Commodity export taxes are detrimental to primary producers as they depress the producer prices (World Bank, 1986). Sustained higher export taxes lead to sustained lower producer prices. Under such circumstances, it is hypothesized that the share of small farms in total production would decrease as they have more flexibility in switching to alternative enterprises as compared to the larger fanns whose switching cost would be much larger. The expected negative relationship between the tea export taxes and the share of small tea fanns is observed in the regression with a 1% significant level. It is argued that the tea export taxes has been exorbitant in Sri Lanka. Researchers have been cntical about the irrational export tax policy in the Sn Lankan tea industry (Perera, 1976, World Bank, 1986, Sinnatharnby and Devaraj, 1987, Betz, 1989, Bhalla, 1991). Yet, the probable impact of the higher export taxes on changes in fmsize distribution has not been previously explored. The empirical relationship between the changes in fmsize distribution and rate of export taxation shows that larger export taxes discourage small farmer tea production more than it discourages large fann tea production. The negative relationship between the pnce instability index and the share of the small fms is expected. Unstable prices give rise to fluctuations in incorne and cash flow crises and the large tea farms are assumed to be more capabte in handling such difficulties compared to the small fârms. The observed sign is opposite to the expected sign, yet the estimated coefficient is not statistically significant at 5% level. As noted in Chapter 4, small fmers may be able to stabilize their average income with the diversified enterprises such as growing subsistence crops and engaging in off fm income generating activities. In fact, about two-thirds of small tea fmers in Sri Lanka engage in off-farm employment (Report of the Baseline Survey of Tea Small Holdings, 1994 Table 9). Thus, while large farms may be able to handle price instability through having access to overdraft facilities in banks, and holding financial resewes small fms may be better positioned to self-insure through more tirne availability to engage in off- fmemployment. Thus, cornpetencies to face the risk generated through price instability may not be substantially different between the small and large farms. Another puuling regression result is the positive relationship (though statistically insignificant at the 5% level) between the share of small tea farms and retums fiom the alternative crops. It is hypothesized that as the relative returns from the alternative crops to tea (such as non-traditionai export pere~ials)increases the share of small tea fms falls. This is because the relative flexibility to switch fiom tea cultivation to other enterprises is assumed greater for the small fms. Hence, as the relative profitability of alternative crops increases more small farms would switch to those enterprises relative to large farms. However, the regression results do not support this assertion. One of the limitations in the empirical specification of the relative profitability of alternative crops is that it is approximated by FOB prices (export value divided by the export volume). Costs may differ so the FOB prices may not truly reflective profitability. However, the insignificant result most likely implies that there is little difference in the relative supply response between fmsizes. According to the regression results, land refonn policies have not been effective in increasing the share of small tea fanns. Mean values of Y, before (-0.1360) and afier (0.0184) the land refonn policies are not significantly different (at 5% confidence level). The insignificant effect was expected since land redistribution for the small fmers was not an objective of the land reform policies. This is evident by the post refonn land redistribution pattern and post refonn institutional arrangement of managing the acquired land (see Table 4. 6). There is no detailed analysis about the pattern of land redistribution (Ranasinghe, 1996). However, Sajhau and Muralt (1987, p.60) noted that by 1982 only 110,000 acres had been distributed arnong 350,000 families of land-less peasants and suggested each fâmily received about one-third of an acre. Such small blocks usually serve the purpose of building a house rather than starting a tea garden. In countries like Papua New Guinea plantations that were owned by multinational corporations were acquired by the government and the ownership of such lands were transferred to the small holders (Sajhau and Muralt, 1987 p.79). Such a policy would directly increase area under the small farms at the expense of large fms. It is clear fiom the discussion of section 4.2.A.1 in chapter 4, that small holder area has not been expanded at the expense of large plantation land in Sri Lanka. Therefore, the regression result of the land reform dummy is not surprising. Yet, the empirical verification of the hypothesis is important since this indicates that the claims of land reform policies have been effective in the expansion in small tea fms(Le., Fernando, 1982) are not tenable. On the other hand, claims about the share of large tea fmarea coming down due to inefficient management regimes aAer land reform policies is also not supported by the data. Pricing intervention in the green Leaf market is hypothesized to be favorable for the small tea farmers. It is argued that in the open market regime there is a potential for hoid up problems in the green leaf production investment. Such hold up problems are not relevant to the large farms, which tend to be vertically integrated with their own green leaf production and black tea processing. Thus, pricing intervention would be favorable for small farms ensunng higher prices for green leaf, while it has no particular impact on the large farms. Therefore, it is hypothesized that during the pricing intervention regimes, the share of small fmswould be greater compared to the period of no price intervention. During the first phase of the pricing intervention in 1969 to 1984, mean value of Y,(0.1600) greater than that of (-0.13600) the period of no price intervention from 1960 to 1968 at the 1% significant level. In addition, during the formula price regime the mean value of YIis (O. 1 15 1) is also greater than that of the period of no price intervention at a 1% significant level. The paradox of why such an intervention in the green leaf market is needed in the tea industry, which appears to be reasonably cornpetitive with large number of green leaf producers and processors, will be evaluated in terms of the vertical coordination changes in the next section. 7.3.Estimation of the Model for the Changes in Vertical Coordination Arrangements The empirical model of the changes of vertical coordination arrangements was specified in the previous chapter. The model is; The change in the vertical coordination arrangement is represented by (Y2);the share of market-procured green leaf based black tea production in the total black tea production (black tea production by market procured green leavtotal black tea production). WR represents the ratio of labor opportunity cost between small farms and large farms. The dummy variables are for the state ownership of large tea farms (State=l for 1976 to 1Wî), price intervention before the reasonable pncing formula (PriecInterI=l for 1969 to 1984), and price intervention with the reasonable pricing formula (PriceIntiI=l for 1985 to 1998). The black tea price of the Colombo tea auction is represented by price. The assumption about the error term U is as in the previous econometric model. Since the Box-Cox transformation could be applied only to the variables with strictly positive values, the three intercept dummies are not transformed. The rest of variables are re-scaled by dividing through their geometric rneans. Annual data for the period of 1960 to 1998 compnsing 38 observations are used in the econometric estimation. Signs for the estimated coefficients of the dnvers responsible for the changes in vertical coordination arrangements are expected to be; y,, yz, and y3 < O, while y4, and y5 >O. Table 7.4 provides the results of the maximum likelihood estimates. The estimated model explains the variability in the changes of vertical coordination arrangements well. The adjusted $ is 0.96 and the F-value is significant at the 1% level. The Durbin-Watson statistics is 2.0044 so that the nul1 hypothesis of positive or negative first order autocorrelation in the error term is rejected (Gujarati, 1995 p.422). Al1 the estimated coefficients except the State dummy are consistent have the expected signs and al1 the estimated coefficients with expected signs are significant at the 5% level. The individual coefficients are explained later. 7.3.A. Testing for Cornpeting Functional Forms As in section 7.2.A, the asymptotic likelihood ratio test is used to test the competing models. Since A =1 for the three dummy variables, there are three restrictions in the Box-Cox regression. The likelihood ratio test for functional forms of linear (A=l), logarithmic (A=0) are shown in Table 7.3. The linear functional form is rejected but not logarithmic. Table 7.3 Test for Cornpeting Functional Forms for Vertical Coordination Changes Ho: A=l Ho:A.=O Log of Likelihood Function 40.40 46.275 LR Statistics 1 1.77 0.03 8 X2 (at 0.01 probability) 9.2 1O 9.210 3 Test Result Reject the linear functional Do not reject logarithmic form functional form 7.3.8. Interpretation of the Results of Econometric Estimation The sign of the estimated coefficient for WR (the ratio of opportunity cost of labor for vertically independent (small) farms and vertically integrated (large) fms) is consistent with the theoretical proposition and significant at the 1% level. It is argued that the "make" versus "buy" decisions depends on the production cost differences between in-house production and market firms. Labor is the costliest factor in green leaf production for both the vertically independent small farms and vertically integrated large farms so changes in the relative labor costs play an important role in the decision regarding green leaf procurement. As Figure 4.3 indicates labor wages for independent srnaIl fms declined relative to green leaf production on vertically integrated fms, since the 1980s. As the "buy" decision became less expensive relative to the "make" decision, the industry has shifted towards vertically independent sectors. Thus, the decrease in relative production costs in the "buy" decision is an important economic force associated with the vertical coordination changes in the Sri Lankan tea industry. The mean value of Y2 is expected to be larger during state ownership of vertically integrated plantations relative to the period of private ownership of such plantations. This difference is expected since the distorted management incentives during state ownership of vertically integrated plantation would make them less viable relative to the privately owned vertically independent green leaf production and black tea processing sector. Table 7.4 Maximum Likelihood Estimates for the Vertical Coordination Changes Variable Estimated Asymptotic P-Value Consistent with Coeff~cient T-~atio Expected Sign Constant y0 = -0.14788" -4.369 0.000 Not Applicable Price y2 = -0.1 43 84' -2.36 1 0.0 18 Yes State y3 = 0.0 14467 0.4 141 0.679 Yes Price Inter 1 y4 = 0.066802' 1.987 0.047 Yes Box-Cox Parameter (A) = -0.06, Durbin-Watson statistics = 2.0044, and O* denote statisticat significance at 5% and 1% respectively. The econometric results indicate that the mean value of Y2 is larger during the period of state ownership of vertically integrated fams. However, the mean value of Y? during the state ownership (0.0144) is not significantly difference at 5% level from the mean value of Y2.(-0.1478) during the pnvately owned period. The state intervention in the green leaf pricing is expected to increase Yr or the equilibrium output of the vertically independent sector because pricing intervention is hypothesized to allay the hold up problems between the vertically independent green leaf producers and black tea processors. During the price intervention of 1969 to 1984, green leaf prices were not directly linked to the black tea prices, which is the most important determinant of green leaf prices (see section 4.2.D.2). Since 1985, with the reasonable pnce formula, green leaf prices were directly linked to the black tea prices. Thus, it is argued that the impacts would be different in these two pncing interventions. This is because during the first regime of pricing intervention green leaf pnces determination was somewhat arbitrary. Due to the nsing level of inflation, these prices for green leaf were offen revised, yet the link to the black tea price is weak. Therefore, it is argued that the responses of the vertically independent green leaf producers were relatively low during the period of Pricelnferl compared to the Pricehfer11 (formula pricing regime), where black tea price was directly link to the green leaf pnce and green ieaf price determination is more transparent. The mean values of Y? are higher for both the pricing intervention regimes relative to the period without price intervention and the differences are statistically significant at 5% level. More importantly, the mean value of Y2 (0.3691) was higher during the regime of the reasonable pricing formula (Pricelnterli) relative to the mean value of Y? (0.0668) or the regime of govenunent determined prices (Pricelnterl). Both these mean values are higher than that for the no pnce intervention regime (-0.1478). Therefore, the theoretical proposition about the impact of pncing intervention in expanding the equilibrium output in the vertically independent sector is consistent with the data. The estimated coefficient for the black tea price is negative (as expected) and significant at a 5% confidence level. It is argued that if the green leaf supply decisions as a response to black tea pnces are the sarne for the vertically integrated sector and vertically independent sector, ceteris paribus, there should not be any systematic relationship between the share of black tea production by either sector and black tea price. However, if the green leaf supply decisions are restrained by hold up threats in the vertically independent sector, green le& production response to black tea prices would be different in the two sectors. With the hold up threats, vertically independent sector would not increase the green leaf production as much as in the vertically integrated sector for a given size of black tea price increase. Thus, one should expect to see a negative relationship between Y2 and black tea price with the presence of hold up threats given everything else stays the same between the two sectors. This negative relationship is observed in the regression results and it is significant at the 5% level. 7.4 Summary Results of the econometnc estimations to test the hypotheses regarding the structural changes within the Sri Lankan tea sector are reported in this chapter. Box-Cox regression is used to estimate the two empirical specifications of the changes in tea fann size distribution and vertical coordination arrangements. The ratio of opportunity costs of labor for large farm and small fann was, as expected, negatively related with the share the area of small fmin the total tea area (Yi). The expected positive relationship between smali fmsupport services and Yi was observed. Export taxes on black tea and Yi, as expected, negatively related. The expected relationships between pricing intervention in green leaf market and Yi was observed. The expected negative relationship between tea pnce instability and YIwas not observed. Land reform policies were not influential in increasing the share of small fmarea in the total tea area (Yi). The expected relationship between Yi and the profitability in alternative cropping enterprises was also not found. The expected negative relationship between the share of black tea production fiom market-procured green leaf in total black tea production (Yz)and the ratio of labor opportunity costs between vertically independent (srnall) and vertically integrated (large) fami was found. The ratio of labor opportunity costs indicated the production cost differences between vertically independent and vertically integrated sectors. The expected relationship between (Y2) and pncing intervention in green leaf market was observed. The differences in green leaf procurement though bbbuy"decisions before and afier the pnce interventions in green leaf market indicated the expected impact of the pricing intervention in green leaf market. The expected relationship between Y2 and the black tea pnce is observed and it indicates that responses to prices in "buy" decisions were not as large as the responses in "make" decisions. The differences in 'bbuy'* decisions and "malce*' decisions as response to black tea prices indicates that some fictions in the arms length exchange in green leaf market. Holds up problems provide plausible explanation for this situation. The expected sign for the relationship between Y? and regime of ownership and management was observed. The size of Y2 was determined by the changes between private and state management and ownership of the vertically integrated plantations. CHAPTER 8 SUMMARY AND CONCLUSIONS 8.1. Summary The structure of the agricultural sector has motivated the farm policy debate in many countries throughout the past century. The perception that some of the trends in the economic organization of agriculture might adversely affect the interest of fmers, consumers, or the general public makes these changes important public policy problems. In economic, social and political realms, concems are expressed about causes and consequences of structural changes in agriculture. A wide range of public policies could be used to influence the structure of agriculture and meet the economic and political objectives of the economic agents in the sector. Structural changes in agriculture are the aggregate outcomes of changes in the choices of economic agents as they respond to unfolding economic constraints and incentives such as relative prices of inputs and outputs, technologica1 parameters, and changes in costs of exchange arrangements. Therefore, knowledge about the relationships between the observed structural changes and the responses of economic agents to the changing incentives and constraints is important in policy evaluation, since policies are just another driver that influence structural changes. The econornic fundamentals driving the trends can be tested by examining contrasting patterns of structural changes in agriculture in different parts of the world. In North Arnerican and European agriculture, the components of the supply chain are increasingly dominated by fewer and larger fmswhose efforts are coordinated through the vertically related stages by negotiated contracts rather than traditional market mediated coordination. However, a different pattern of structural change has emerged in some countries with colonial plantation industries, such as in Kenya, India, Sn Lanka, and Guyana. The large vertically coordinated plantation industries in many of the developing countries have evo lved to an industry structure charactenzed b y a predominance of small farrns that market their produced through arrangements such as contract production, producer cooperatives, and parastatal organizations. The changes in farm size distribution have been biased towards small fanns against large farms and the vertically integrated activities (raw material production and processing) of the plantations has been separated into independent producers and processon. Unlike the assessment of structural changes in Western agriculture, the analyses of these changes in plantation industries within developing countries are often confined to political economic discussions without paying attention to the relative importance of potential drivers of these changes. This thesis attempted to explain structural changes in the Sn Lankan tea industry in light of the choices of economic agents as they respond to the changes in economic incentives and constraints unfolding in the economy. The results of the analysis are summarized below in terms of the specific objectives of the thesis. (1) Describe the strrrctrtral changes in the Sri Lankan rea indrrstry and identrfj, the structural dimensions that have rrndergone sttbstantial changes. The importance of the tea industry to the Sri Lankan economy in terms of employment generation, land use, and foreign exchange earnings are substantial as described in chapter 2. Green leaf production, black tea processing, and black tea marketing are the main activities in the tea industry. Small farms and plantations grow green leaf tea and plantation themselves or independent processors do the processing of green leaf into black tea Black tea is sold through a central auction system. Tea growing in Sri Lanka was started around the 1860s and by about 1920, tea area reached to its peak and since then tea growing has become firmly entrenched in the country. British corporations dominated al1 spheres of the tea industry until about 1948. The British conferred independence status to Sri Lanka in 1948 and during the post independence era, from 1948 to 1975, British ownership and management had gradually declined and in 1975, the tea industry was nationalized. State ownership and management of large tea farms continued till 1992 and it is generally agreed that during this period large tea fms were mis-managed. Large tea farms were re-privatized in 1992 and since then pnvate firms have been managing these farms. During the post independence period, the Sn Lankan tea sector has moved fiom an industry structure dominated by plantations which produce the green leaf and process it into black tea, to one where the green leaf is now predominantly grown by small fmers and subsequently soid to independent processors. The small farm share of totai tea area has increased fiom about 12% in 1960s tu about 44% by 1998. The vertically integrated production has declined fiom 80% of the total black tea production in 1960s to about 45% of total black tea production by 1998 while vertically independent production has increased fiom about 20% of total black tea production to about 55% during the same period. Both farm size distribution and vertical coordination arrangements have substantially changed. (2) Identzfi the potentiul drivers of structural changes by reviewing the literature on the econom ic anaiysis of factors affecting diferent dimensions of the farming industry Chapter 3 assessed the literature related to the economic analysis of structural changes in agriculture, specifically about fmsize distribution and vertical coordination arrangements and identified the important drivers that influence the changes in these structural dimensions. The major theoretical models dealing with farm size distribution changes are the technology rnodel, the human capital model, and the financial model. The strengths and weaknesses of these models and unresolved issues with each model were discussed. The major theoretical approaches dealing with vertical coordination changes such as the structure-conduct-performance model, the principal-agent model, and the transaction cost approach was discussed in light of their strengths and weaknesses in explaining vertical coordination changes. Six categories of dnvers that influence the changes in farm size distribution and vertical coordination were identified: a) technoiogical factors, b) factors related to human capital, c) prices and financial sector factors, d) govemment policies and prograrns e) national economic growth conditions, and f) agri-business orientation. The impacts of structural change dnvers in each of these categories were briefly discussed related to the structural changes in developed world agriculture. (3) Develop a conceptual framework to explain the eflects of these economic forces on the changes of farm size distribution and vertical linkage dimension and to generate refutable hypothesisfrom this framework: The conceptual model to explain farrn size distribution changes was developed in chapter 4. This model denved testable hypotheses on the impacts of changes in a) relative labor costs of small and large tea hs,b) land reform policies, c) small fmsupport services, d) tea export taxation, e) state pricing intervention, f) tea price instability, and g) the profitability of alternative agricultural enterprises on the changes of fami size distribution in the Sn Lankan tea industry. The differential responses of the large and small farms for the changes in these economic factors are outlined. The conceptual model to explain the changes in vertical coordination arrangements in the context of "make" versus "buy" decisions of green leaf tea procurement in the Sri Lankan tea industry was developed in chapter 5. There were three economic factors in this conceptual model; a) the changes in green leaf production costs differences for the technological reasons b) the changes in management incentives and agency costs, c) and changes in the potential for hold-up threats in the vertical coordination arrangement. Testable h ypotheses are developed about the impact of these factors on the changes in vertical coordination arrangements. (4) Empirical spectjkation of the variables to represent the srrtrcttrral dimensions and the stnrcttrral change drivers to test the above hypotheses. The empirical specifications of the dependent variable and explanatory variables for both farm size distribution changes and vertical coordination changes were provided in chapter 6. Fmsize distribution was represented as the share of small tea fmsfiom the total tea area of the country. There were eight explanatory variables; the ratio of labor opportunity cost for small fmsand large fms, the arnount of subsidies paid to small farms, export taxes on black tea, tea price instability, the relative profitability of alternative crops, the introduction of land reform policies, the price intervention before the reasonable pncing formula, and the price intervention with the reasonable pncing formula. The change in the vertical coordination arrangement was represented by the share of total black tea production fiom the market-procured green leaf tea- There are five explanatory variables: the relative opportunity cost of labor for small fmsand large fms, the black tea price fiom the Colombo tea auction, and dummy variables for the time periods of state ownership of large tea fms, the price intervention before the reasonable pricing formula, and the pnce intervention with the reasonable pricing formula. The data for al1 the variables are collected fiom 1960 to 1998 from various published sources in Sri Lanka- The econometnc estimates of these empirical specifications are reported in Chapter 7. The Box-Cox regression was used to select the functional form for the empincal specification. In order to estimate a valid covariance matrix and valid t-ratios to avoid unreliable inferences, the Box-Cox regression was carried out with re-scaled data where re-scaling was done by dividing al1 the variables (except dummy variables) through their geometric means. The asymptotic likelihood ratio tests were carried out to test the hypotheses about the functional forms. The maximum likelihood method provided the functional form parameters that best fit with the data. Six of the eight explanatory variables were consistent with the theoretical expectation of their influence on the farm size distribution changes and statistically significant at the 5% level. The share of tea area under the small fmswas negatively associated with the ratio of opportunity cost of labor for small fmsand large fms. This indicates that as the relative labor costs for the small farm sector decreases their share in the total tea area goes up. The small tea hm share in the total tea area was positively related with the amount of subsidies paid for small farms. Such subsidies were argued to offset some pecuniary economies of size enjoyed by the large fmsand were able to make small fmscompetitive with large fams. The small tea farm share in the total tea area was negatively associated with the export taxes on tea. It was argued that as larger export taxes depress the producer prices, small famis exit from tea production at a grater rate compared to the larger fms. The small tea farm share in the total tea area, on average, was greater during the period of the pricing intervention in green leaf market relative the period of without price intervention in green leaf market. It was argued that arms length transaction in green leaf market might be prevented due to hold-up problems and pricing intervention by regulating the green leaf price might remove the fiction in the arms length transaction. Tea price instability, land reform policies, and the profitability in alternative cropping enterprises were not important in explaining the changes in tea fmsize distribution. The change in vertical coordination arrangements was also well explained by the empirical specitication. Four of the five explanatory variables of the mode1 were consistent with the theoretical expectations of their impact and statistically significant at the 5% level. The vertical coordination changes or the changes in the "make" versus "buy" decisions for the green leaf procurement were detennined by economic factors such as production cost differences due to labor wages, and the success of arms length transactions in green leaf market. However, vertical coordination changes were unaffected by the regime of ownership and management. The widely believed differences in management incentives for inducing the profit augrnenting decision making (such as outside green leaf procurement) under govemrnent ownership and management versus private ownership and management was not supported by the data. The differences in green leaf procurement through "buy" decisions with and without price interventions indicate that some fnctions were present in the arms length transaction of the green leaf market. Black tea production of the vertically integrated sector was systematically greater than that of the vertically independent sector for the given increase in black tea prices. This finding was also indicative of some fnctions in arms-length transactions in the green leaf market. There is no valid test to determine what has caused the friction but hold-up problem may provide a plausible explanation for this situation. The pricing intervention in green leaf market prevent one party holding up the profits of the other party, however, most likely in the Sri Lankan tea sector that the green leaf processors were holding up the green leaf producers profits prior to the pncing intervention. Other plausible economic factors to explain the fnctions in amis-length transaction might be, moral hazard issue invoiving the difficulties in testing for the green leaf quality or the risk of timely input procurement with outside green leaf supply. The relative importance of different economic forces in changing different structural dimensions of agriculture sector has been neglected in most of the economic analysis of structural changes in agriculture. This study was able to quantify and expIain the impact of these economic forces that are important in affecting changes of farm size distribution and vertical coordination arrangements in not only the Sri Lankan tea industry but in al1 agricultural sectors. 8.2 Conclusions and Policy Implications The thesis set out to expiain the structural changes in agriculture as the aggregate outcornes of the responses of economic agents to unfolding economic incentives and constraints. The thesis developed conceptual and empirical models incorporating structural changes as the aggregate outcomes of responses by economic agents to changes in potential drivers. It has been shown that once the underlying economic forces are identified and their influences are explained, the structural changes in agriculture are the aggregate outcome of the differential economizing behaviors between different categories of fanns. The most important implication of undentanding the systematic relationships between the structural changes and responses of individual farmers to the constraints and incentives in the economy is the potential for designing appropriate agriculture policies to affect these structural changes. The knowledge about these responses is vital in evaluating the impact of various policies on the changes in the structure of the Sri Lankan tea sector. More importantly, these empirical findings show that policies that are introduced to the sector exert their influence jointly with many other important economic forces. The impact of the policies net of the impact of other exogenous economic forces in influencing the changes in structure of the Sri Lankan tea sector are not known. These unknown terri tories are important in evaluating the effectiveness of the policies. the empirical findings on the changes in farm size distribution show some policy paradoxes in the Sri Lankan tea industry particularly for small farm support services and export tea taxes. The desirability of government provision for certain services when small fmshave difficulties to obtain them such as credit, extension services, and other inputs has been acknowledged and promoted in Sri Lankan tea industry. The results of this study show that, as in many other countries, small fmsupport prograrns were effective in expanding the number of small fmsin the Sri Lankan tea sector. However, policy makers were arguably unaware of the differential impact of the tea export taxes on small and large tea fms. Some argue that small fmen are more flexible to switch arnong different enterprises relative to larger farms. As higher export taxes depressed domestic prices, it is argued that more small fms exit fiom tea fming relative to large farms leading to size distribution changes biased towards to large farms. The empirical findings of this thesis support this argument by showing a negative relationship between export taxes and the number of small farms. Thus, the objective of small farm promotion with the srna11 fmsupport services has been negated by export tax policies in the Sn Lankan tea sector. Small holding support services were associated with lower export tea taxes in tea in Kenya and the expansion in small tea farmers in Kenya has been much &ter than in Sri Lanka. Tea export taxation was abolished in 1992 with the re-privatization of the tea plantation sector and a rapid expansion of small tea fms has been observed since then in Sri Lanka. The other important policy dilemma is the govenunent involvement in labor wage determination of the tea plantation sector. Although the issue of labor wages in plantation industries is shrouded with political controversies, the competitiveness of large fms is dependent on their labor wages. Tea production remains a labor-intensive activity and so the wage rate is a major determinant of production costs. The empirical findings of the thesis show that the relatively higher wages in the large fam sector compared to srnall farms has been an important factor in explaining both the farm size distribution and vertical coordination changes in the Sri Lankan tea industry. Any intervention in the labor wages in either the large or small fmsector will have a significant effect on their relative performance and viability. Small farrns mainly depend on family labor and hence their opportunity cost is detennined by how much they can eam elsewhere in the economy. Thus, market forces through labor supply and demand in the economy determine the opportunity cost of small fmlabor, Large farms or plantations have their own residential labor forces but market forces are not allowed to play a role in wage determination of these workers. Wage rates for workers on large fanns are politically driven with labor productivity acting in secondary importance (Betz, 1989, Report of the Presidential Commission on Tea Industry and Trade, 1995 p. 148). The intervention in labor wages without considering the contribution of the labor input to tea output has undermined the competitiveness of large tea farms. The coercion-based land-lord labor relations, which was an important feature of the plantation system in the seventeenth and eighteenth centuries has been subjected to tensions with the changing social relations of this labor force. There is a lack of innovative labor management practices developed in plantation industries in Sn Lanka. Thus, industrial relations are increasingly becoming hostile and darnaging to the tea industry. The oid system of coercion-based supervision is not capable of improving labor productivity and making the large farms competitive with small fms. Care and wisdom is important in implementing labor management approaches that can accommodate the inspirations of the plantation laborers who are no longer the "semi-serf' docile immigrant who came to Sn Lanka in the seventeenth and eighteenth centuries. The passive govemment intervention in determining the labor wages ignonng the above aspects might eventually threaten the viability of large fmtea production in Sn Lanka. Pricing intervention in green leaf market is another important policy dilemma faced by policy makers in Sn Lanka. There was a substantial pressure to do away with the formula used to price green leaf as has occurred in many of the other small holding crops in the country (Report of the Presidential Commission on lea Industry and Trade, 1995 p. 106). The difficulties in organize green leaf production and sale through arms- length transactions was noted and contract production or vertical integration between green leaf producers and black tea processors was the result noted by many researchers. Monopsonistic behavior of the black tea processors was forwarded as an explanation of the low prices for green leaf paid to growers and the subsequent failures in arms-length transactions in the green leaf market prior to the govemment pncing intervention. Yet, the monopsony argument is not substantiated by any of the researchers and the presence of a large number of processors is countenntuitive to the monopsoney argument. Empirical findings of this research were consistent with the difficulties in with an auction like market for green leaf, and suggest that the pricing intervention played a role in the observed changes in the vertical coordination arrangements. The relatively higher green leaf prices under a pnce intervention and the subsequent higher supply by green leaf producers could not explain the increased amount of green leaf bought by processors, since such higher prices for green leaf is a disincentive to buy for the processors. Thus, higher green leaf prices under the pnce intervention regime does not explain the simultaneous increase in both the supply from green leaf producers and the derived demand from black tea processors. The pncing intervention, particularly the reasonable pncing formula, has introduced a de facto contractual agreement between the green leaf producers and processon. Although its implementation is not a mutually agreed-upon outcorne, its effect is similar to the effect of a mutual agreement in contract production since its resolves some of the fictions in the amis-length transactions. With such a contractual arrangement (legally imposed), it is possible that the scope for the hold up problem is reduced and the "buy" decision is facilitated. Whether the friction in arms length transaction in green leaf market is due to hold up threat or not, the pricing intervention has been able to remove this fiction and promote arms length transactions. In the other tea producing countries, as many other researchers have noted, contract production and other pncing arrangements such as a mutually agreed formula-pricing have been used to avoid arms-length transactions. The same effect is found in the Sri Lankan tea industry with the green leaf pricing intervention. However, as more independent green leaf producers and processors enter the tea sector, the likelihood of a hold up problem is diminished. Thus, market forces might be allowed to determine green leaf prices. Whether the role played by the pncing intervention has come to an end with such a thick market is open to debate, yet, according to the empirical Findings of the thesis, it has been important in prornoting the arms length transaction and the structural changes observed in the Sri Lankan tea industry. However, it is prudent to evaluate the probable impact of the factors that may give rise to hold up threats and assess the competitiveness in the green leaf processing sector prior to revising the green leaf pricing formula. One of the important implications of the findings in the empirical investigation is the mutual dependency between the two structural dimensions; changes in farm size distribution and vertical coordination arrangements. As some of the same variables were used to explain both structural dimensions, it begs the question whether the changes in farm size distribution and changes in vertical coordination are two sides of the same coin. This is important in evaluating the desirability of agricultural policies if the policy objective is to influence one structural dimension but not the other. One can argue that if the impacts of policies are to proliferate small hs,the obvious vertical coordination arrangement would be market-mediated transactions while if these policies promoted large fms, vertical integration would be the obvious choice of vertical coordination. This assertion is incorrect. Changes in size distribution could take place independent of the changes in vertical coordination arrangements. Therefore, policies to influence the changes in one dimension do not necessarïly trigger the changes in some other structural dimensions. For instance, small fams could be proliferated with contract production (in Kenyan tea industry, see Lamb and Muller, 1982) rather than market-mediated vertical coordination arrangements. Small famis could also be promoted with their own cooperative processing centers (in Kenyan dairy industry see StaIl, Delgado and Nicholson, 1997). The factors that govern the farm size distribution do not necessanly govern the vertical coordination arrangement. However, some of the factors may be mutually reinforcing such as the pricing intervention. It affects the vertical coordination arrangement directly thereby making the exchange arrangement conducive for the srnall farms. The institutions governing the exchange arrangements evolve to improve the overall efficiency that achieved with such the exchanges (North, 1990). 8.3 Limitations of the Study One of the important theoretical limitations of the study is the lack of a single coherent theoretical approach to select trul y independent causes of the structural changes, particularly the fam size distribution changes. The literature rarely questions the independent nature of the drivers of structural changes. Until recently almost al1 the structural change analyses in North Amencan agriculture were focused on fmsize distribution changes, which is narrowly defined as the changes in number of fmsin size classes defined by gross sale receipts. These studies attempt to explain the changes in average fmsize by appealinp to a host of factors drawn arbitrarily. This thesis too ha this limitation and the selection of drivers in farm size distribution changes is arbitrary but based on economic reasoning. The drivers are not endogenously determined through a more general theory. The arbitrariness of the dnvers limits the scope for the generalization of empirical results. The study is lirnited to few important dnvers of farm size distribution and vertical coordination changes. Man y other potential dnvers were not included into the conceptual model. For instance, in 1977, there was a transition in the economy fiom a welfare- onented, inward looking, and protectionist economy to a liberalized, export-import onented, open economy. Wi th this transition, arti ficially kept over-valued exchange rates were removed and consequently, the returns fiom major exports such as tea and other primary comrnodities have increased. It is argued that the differential impact on the profits with such a devaluation of exchange rates in producing subsistence crops versus export-oriented crops is important. The impact of the introduction of such export- fnendly policies in the changes of fmsize distribution in tea sector is not incolporated in to the conceptual model. The lack of a coherent theoretical foundation to pick genuinely independent causes of structural changes is the most important limitation with the conceptual models of the study. The analyses of structural changes in agriculture could be improved considerably by developing a unifjing theory in identiîjing genuinely independent structural change drivers that are important in changing given structural dimension. Ln what important ways would the results have changed if the analysis were based on such a general theory? The generalizability of the results would have been much greater. However, the drivers of stnictural changes were selected with an in depth review of the economic analysis of structural changes in agriculture. Almost al1 the important drivers are inchded into the analysis. To this extent one can be confident about the empirical findings of the thesis. One of the important aspects ignored in the "make" versus "buy" decisions in green leaf procurement is the moral hazard problem involved with the "buy" decision. It is well known that independent green leaf producers can change some quality attributes of the green leaf (i.e., adding water and thereby increasing surface moisture of the green leaf, or incorporating mature leaf) to increase their returns. The moral hazard problem is reported as an important grievance of the black tea processors who buy green leaf fiom independent green leaf producers (Report of the fresidential Commission on The Tea Industry and Trade, 1995 p.75). Incorporation of the moral hazard problem in the quality of green leaf purchased and its impact on the changes in vertical coordination arrangements is an interesting extension of this study. Such an extension would bnng another possible explanation of failures of arms-length transactions in the green leaf market. If the moral hazard problem were important what would have been the impact of prking intervention in the green leaf tea market? Was the pricing intervention able to resolve the moral hazard problem involving green leaf quality in the arms length transaction? Without adequate further research one cannot answer these questions. There were important data limitations in the empirical models of the thesis. The best representation of the size distribution changes would have been average farm size together with another higher order movement of distribution such as variance. Such data are not available in Sri Lankan tea industry. The opportunity cost of small fmlabor is approximated with industrial sector labor wages. However, this may not be the best measure of small fmlabor opportunity cost, since that earnings for a small farmer selling his labor to the other rural enterprise may be different fiom the industrial sector labor wages. If their true opportunity cost of labor were less than the industrial sector labor wages, the empirical results would still hold. The profitability of alternative agricultural enterprises was approximated with FOB prices of these comrnodities. A better measure is gross margin per unit land area relative to tea. However, the information to estimate the gross margins for these crops was not available. The empiricaI estimation would have changed if the gross margins are substantially different fiom the FOB prices. Theoretically, the hold-up problem could have been in two directions. If the green leaf growers were exercising power on the green leaf processors, pricing intervention might have decreased green leaf prices relative to pre-intervention green leaf pnces. On the other hand, if green leaf processors were exercising power on the green leaf producers, the opposite effect on green leaf prices would have occurred with the formula pnce. There are no data to ascertain what has actually happened with the pricing intervention. However, al1 the evidence suggests that the green leaf pnce has gone up with the pricing intervention. There are important regional level variations in the structural changes. For instance, among the elevation classes of tea producing regions in Sri Lanka (High Grown, Mid Grown, and Low Grown), the Low Grown region has been transformed most dramatically while both these structural dimensions has not changed substantially in the High Grown region. The causes of these regional level differences in the structural changes have not been addressed in this thesis. More importantly, the explanations about such contrasting pattern of structural changes in regional levet definitely would improve the knowledge of the causes of the structural changes. However, the regional level break down of the data is not available. Ajjan, N and V. Rajagopalan. "An Analysis of Supply Response of the Tea industry in South India." Singapore Economic Review 34(1,1989):68-74. Akiyama, T and P.K. Trivedi. "Vintage Production Approach to Perennial Crop Supply: An Application to Tea in Major Producing Counûies." Journal of Econometrics 36(l987):133-l6l. Alchian, Amen A. "Specificity, Specidization, and Coalitions." Journal of Institutionai and Theoretical Economics. l4O( 1984):34-49. Allen, Douglas W. and Dean, Lueck- "The Nature of the Fm."Journal of Law and Economics. XLI (October 1998):343-386. Annual Reports of Central Bank of Sri Lanka, Central Bank of Sri Lanka, Colombo, various issues. Annual Reports of Sri Lanka Petroleum Corporation, Colombo. Various Issues Askari, H. and J.T. Cummings. AgncliZtural Supply Response: A Survey of the Econometric Evidence. Praeger Publishers, New York, 1976. Athukorala, P. "International Trade Movement and the Plantation Crop Sector." In Plantation Agriculture in Sri Lanka, eds. W.Gooneratne and D. Wesumperurna. International Labor Organization, ARTEP, Bangkok, 198457-85. Athukorala, P. and Frank Cong Hiep Huynh. Eiport Instabilis, and Growth: Problerns and Prospectsfor the Developing Econornies. Croom Helm, London, 198 7. Azzam, Aveddine M. and Emilo Pagoulatos. "Vertical Relationships: Economic Theory and Empirical Evidence." In Vertical Relationships and Coordination in the Food System, eds. Giovanni Galizzi and Luciano Venturini. Physica-Verlag: A Springer-Verlag Company, 1999:93- 1 13. Babb, E. M. "Some Causes of Structural Change in US. Agriculture." In Structure Issues of American Agriculture. Agricultural Economic Report 438, Economic, Statistics, and Cooperative Service, USDA, 1979. Ball, A. G. and Heady, E. O. "Trends in Fmand Enterprise Size and Scale" in AG. Ball and E.O. Heady, eds. Size, Structure. and Future of Farms, Iowa State University Press, Ames, 1972: 40-77. Bandarage, Asoka. "The Establishment and Consolidation of the Plantation Economy in Sri Lanka." The Bulletin of Concerned Asian Scholars, 14(3, 1 982):2-22. Bandaranayike, R. "Privatization of Estate Management in Sri Lanka." International Tea 1 (1993): 23-30. Barlow, Colin and S. K. Jayasuriya. "Stages of Development in Smallholder Tree Crop Agriculture." Development and Change. 17(1986):635-658. Barlow, Colin. The Natural Rubber hdtlstty. Its Development, Technology, and Economy in Malayasia. Oxford University Press, Oxford: 1978. Barkema, Alan, 'New Roles and Alliances in the US. Food System" in Food and Agricultural Markets: The Quite Revolution. Eds Lyle P. Schertz and Lynn M. Daft. Washington D.C.: The National Planning Association, 1994a: 11 8- 125. Barkema, Alan. "The Changing Structure of the U.S. Food System". Canadian Journal of Agricziltiiral Economics. 42 (1994b): 54 1-547. Beckford, George, L. Persistent Poveriy: Wnderdmelopment in Plantation Economies of the Third World. London: Oxford University Press: 1972. Besanko, David, David Dranove and Mark Shanley. Economics of Strategy, 2nd Edition. John Wiley & Sons, inc: New York. 2000. Betz, Joachim. "Tea Policy in Sri Lanka." Marga 10 (1989): 48-7 1. Bhalla, Suj i t. "Sn Lanka." In The P olitical Economy of Agricultztral P ricing Policy, Vol. 2: Asia. Eds, Anne O Krueger, Maurice Schiff and Alberto Valdes, The John Hopkins University Press, Baltimore, 199 1. Bieber, Judy. (eds) Plantation Societies in the Era of Eziropean E-rpansion. An Expanding World: The European impact on World History 1450- 1800 (vol. 18). Variorium Aldershot: 1997. Binswmger, H. P and Rosenweig, M. R. "Behavioural and material determinants of production relations in agriculture." The Journal of DeveIopment Studies. 22(1986):503-539. Blair, Roger and David L. Kaseman. "A Note on Bilateral Monopoly and Formula Price Contracts." American Economic Review. 77(June 1987):460-463. Blois, K. J. "Vertical Quasi-Integration" In Firms, Organizations and Contracts: A Reader in Industrial Organization, Eds Peter J-Buckley and Jonathan Michie, Oxford University Press, 1996. Bocke, J. H. Economics and Economic Policy of Dual Societies. International Secretariat. Institute of Paci fic Relations, NewYork: 1953. Boehlje, M. "Alternative Models of Structural Change in Agriculture and Related Industries" Agribusiness: An International Journal. 8Wo3, 1992):2 19-232. Boehlje, M. "Structural Changes in the Agricultural Industries: How do we measure, analyze, and understand them?" American Journal of Agrictrlttrral Economics. 81(1999):1028-1041. Boon, Arnold. "Capabilities, Transaction Costs, and Vertical Coordination in the Food System". In Vertical Relatiorzships and Coordination in the Food System. Eds. Giovanni Galizzi and Luciano Venturini. Physica-Verlag: A Springer-Verlag Company. 1999:2 1-37. Boone, Christophe and Alain Verbeke. "Strategic Management and vertical Disintegration: A Transaction Cost Approach." In Microeconomic Contributions to Strategic Management. Eds, Jacques Thepot and Raymond-Alain Thietart, North-Holland, New York: 199 1. Brinkman, G. and T. K. Warley. Stmctrrral Change in Canadian agricriltztre: A perspective. Ottawa: Agriculture Canada, Regional Development Branch, Development Policy Directorate, 1983. Carlin, Thomas A. and Linda M. Ghelfi. "Off-Farm Employment and the Fann Structure". in Stmctttre Issues of American Agricriltzrre. Agricultural Economic Report 438, Economic, Statistics, and Cooperative Service, USDA, 1979. Carter, Harold 0.and Warren E. Johnston. "Some Forces Affecting the Changing Structure, Organization, and Control of American Agriculture." American Journal of Agriculttrral Econornics. (December, 1978):737-748. Cazperze, Paul. "The Intemal Structure and Organization of the Tea Industry: An Overview." Marga Qrrarterly Journal. 2(4, 1975):38-74. Cetral Bank of Sri Lanka. Annual Report. Colombo, Sri Lanka. Various issues, Ceylon Year Book. Department of Census and Statistics, Colombo Sn Lanka, 1970,1969,1968, 1967. Chaudhuri, M. R. me Tea Indusiry in India: A Diagnostic Analysis of lts Geo-Economic Aspects. Indian Economic Geographic Studies, Oxford Book and Stationery Co: Calcutta: 1978. Cheung, Steven N. S. "Transaction Costs, Risk Aversion, and the Choice of Contractual Arrangements". Journal of Law and Economics (April, 1969): pp.23-42. Cloutier, Martin L. Steven T. Sonka and Randall E. Westgren. 'The Economics of Leaming-based Strategies and Structural Changes in the U.S. Pork Chain: A System Dynamics Perspective". Ln Vertical Relationships and Coordination in the Food System. eds. Giovanni Galizzi and Luciano VentUrini. Physica-Verlag: Springer-Verlag Company. 19W:2 1 1-225 Coase, R. H. "The Nature of the Fim." Economica. iV (November 1937):386-405. Coase, R. H. "The Nature of Firm: Influence" In The Nature of the Firm: Origins. Evolution. and Development, Eds Oliver Williamson and Sidney G. Winter. Oxford University Press, New York. 1993. Cochrane, Willard W. The Development of American Agriculture- A Hisîorical Analysis. Minneapolis: universi ty of Minnesota Press, 1979. Courtney, P. P. Plantation Agriculture. 2ndEd. Bell and Hyman, London: 1980. Cozzarin, Brian P. and Randall E. Westgren. "Rent Sharing in Multi-Site Hog Production". American Jotirnal of Agriarltural Economics. 82(2000): 25-37 Cozzarin, Brian P. and Peter J. Barry. 'Drganizational Structure in Agricultural Production Alliances." International Food and Agribusiness Management Review. 1(2, 1998): 149-165. Davidson, Russell and James G. Mackimon. Estimation and Inference in Econometrics, New York, Oxford University Press: 1993. Dawood, Nawaz. Tea and Poverty: Plantarions and the Political Economy of Sri Lanka. Urban Rural Mission-Christian Conference of Asia: Hong Kong, 1980. Delgado, C. L. "Sources of Growth in Smallholder Agriculture in Sub-Saharan Afnca: The Role of Vertical integration of Smallholders with Processors and Marketers of High Value-Added Items. Agrekon. 38(May 1999): 165- 189. Demsetz, Harold. "The Theory of Firm Revisited." In Ownership. Control, and the Firm: The Organization of Economic Activity. Vol. 1, Basil Blackwell, New York, 1988. Deparmient of Census and Statistics Sri Lanka Census of Population, Colombo, 198 1. Department of Census and Statistics Sn Lanka. Statistical Abstracts various issues. De Silva, S. B.D. "Political Economy of Underdevelopment". Routledge & Kegan Paul: London. 1982. De Silva, K. M. "The Refonn and Nationalist Movements in the Early Twentieth Century" in University of Ceylon: History of Ceylon Volume Three. Editor-in- Chief, K.M.De Silva. The Colombo Apothecaries Company Ltd., Colombo: 1973: 38 1-407. Driesen, Vanden 1. "Coffee Cultivation in Ceylon (l)." Ceylon Historical Journal. 3(1, 1953, July):3 1-61. Dye, Alan. "Avoiding Holdup: Asset Specificity and Technical Change in the Cuban Sugar industry, 1899- 1929. The Jotirnal of Economic History. 54(Sept. 1994):628- 653. Edwards, Anthony. "Tea: From the Potential for New Commodity Cartels." Marga 3 (4,1976): 87-9 1. ESCS (Economic, Statistics, and Cooperative Service) Staff. "U.S. Tax Policy and Agriculture Structure". In Srructure Issues of Amen'can Agriatlture. Agricultural Economic Report 438, Economic, Statistics, and Cooperative Service, USDA, 1979. Etherington, D. M. Smallholder Tea Production in Kenya: An Econometric Analysis. East Afncan Literature Bureau. Nairobi. 1973 Etherington, Dan M. "Economics of Scale and Efficiency: A Case Study in Tea Production." East African Journal of Rural Development. 4(1, 1971 ):72-87. Etherington, Dan M. and R. L. Jones. "The Economic Gains fiom an international Tea Agreement." Marga, 3 (4, l976):Z 1-33. Fay, C. R. "Plantation Economy" Economic Jourrtal46 (1936):620-644. Food and Agriculture Organization. Commodity Market Review 1998-1999, Commodity Trade Division, FAO: Rome, 1999: p.39. Forrest, D.M. A Hundred Years of CeyZon Tea. Chatto & Windus, London: 1967. Forrest, D. M. The World of Tea Trade: A Survey of the Production, Distribution and Consumption of Tea. Woodhead-Faulkner, Cambridge, 1985. Fernando, Nimal. "Continuity and Change of Plantation AgricuIture: A Study of Sri Lanka's Land Refonn on Tea Plantation", unpublished Ph.D. Thesis, University of Wisconsin-Madison. 1980. Femando, Nimal. 'Weeded Redirections in Anal ysis of Efficiency of Plantations and Smallholdings in Tea Cultivation" Staffstudies of the Central Bank of Ceyfon, 10 (1&2, April-September, 1981):70-89. Fernando, Nimal. "Tea Smallholdings Sector in Sri Lanka." Progress, 2(4, December 1982):l-12. Fernando, Nimal. "Land Reform in the Plantation Sector-The Employrnent and Income Effects." In Plantation Agriculture in Sri Lanka, eds. W. Gooneratne and D. Wesurnpenima. International Labor Organization, ARTEP, Bangkok, 1984: 11 1 - 131. Fox, Glenn C. Reason and Reality in the Methodologies of Economics: An Introduction, Edward Elgar Publishers Lyrne, United States, 1997. Gadd, C.H. "Small-Holding in Tea." Tea Research lnstitute of Sri Lanka Tea Quarte* 6(1932):87-107. Gardner, B. D. and R. D. Pope. "How is Scale and Structure Determined in Agriculture". American Journal of Agriatltural Economics. 60(1978): 295-302. Gardner, B. L. "Public Policy and the Control of Agriculture Production." American Journal of Agricultural Economics. 60(1978b): 836-843. Gardner, Bruce L. "Causes of U.S. Farm Commodity Prograrns " Journal of Political Economy. 95(2, 1987): 290-3 10. Gardner, Bruce L. "Changing Economic Perspectives on the Farm Problem." Journal of Economic Literattcre, XXX (March 1992): 62- 101. Goddard, E, Alfons Weersink, Kevin Chen and Calum G. Turvey. "Economics of Structural Change in Agriculture." Canadian Journal of Agricultuural Econornics. 41 (1993): 475-489 Gold, Martin E. Law and Social Change: A Study of Land Refonn in Sri Lanka. New York: Nellen Publishers: 1977. Goodhue, Rachael E. "Broiler Production Contracts as A Mul ti-Agent Problem :Common Risk, Incentives, and Heterogeneity". American Journal of Agricul~ural Economics. 82 (August 2000): 606-622. Goodhue, Rachael E. and Gordon C. Rausser. "Value Differentiation in Agriculture: Dnving Forces and Complementarities." In Vertical Relationships and Coordination in the Food System. Eds. Giovanni Galizzi and Luciano Venturini. Physica-Verlag: A Springer-Verlag Company. 1999:93- 1 13. Goonewardene, S. "A Bnef Survey of International Action on Tea: 1933- 1976." Morga (published by Marga Institute Sri Lanka) 3(4, 1976): 127- 136. Gooneratne, W and D. Wesumperuma "Plantation Agriculture in Sn Lanka- An OveMew of Employment and Development Prospects." In Plantation Agriculture in Sri Lanka, eds. W.Gooneratne and D. Wesumperuma. international Labor Organization, ARTEP, Bangkok, 1984:3-33. Gow, Hamish R. ,Deborah Streeter, Johan F. M. Swimen. "How Private Contract Enforcement Mechanisms Can Succeed Where Public Lnstitutions Fail: The Case of Juhocukor a.s". A paper presented at the Workshop "Agroindustrialization, Globalization, and Economic Development". Aug 6-7, Renaissance Hotel, Nashville, Tennessee. 1 999. Green, William H. Econometric Analysis, 2ndEdition, Prentice Hall, Englewood, 1993. Grosh, Barbara. "Contract Farming in Afnca: an Application of the New Institutional Economics. Jortrnul of African Economics. 3,2(0ctober, 1994):23 1-260. Gujarati, Damodar N. "Basic Econometrics" 3" Edition, McGraw-Hill, Inc: New York, 1995. Hallam, Arne. "Empirical Studies of Size and Structure in Agriculture". In Determi~iant of Farm Size and Structure. Ed, Arne Hallam, Proceedings of the program sponsored by the NC-18 1 Committee on Determinant of Farm Size and Structure in North Central Areas of the United States, Jan. 7-10, Tucson, Arizona: 1989 Hallberg, Milton C. Economic Tmdin US.Agriatlture and Food Systems since World War II. Iowa State University Press, Arnes: 200 1. Hart, Oliver. "An Economist's Perspective on the Theory of the Firm." In Firms. Organizations and Contracts: A Reader in Industrial Organization, Eds Peter J. Buckley and Jonathan Michie, Oxford University Press, 1996. Hayami, Yujiro. "Peasant and Plantation in Asia". in Frorn Classical Economics to Devefopment Economics. Ed Gerald M. Meier. St. Martin Press, New York: 1994. Hayami, Yujiro, and Vernon W. Ruttan. Agrimitural Developrnent: An International Perspective. 2ndEd. Baltimore: John Hopkins University Press. 1985. Heady, Earl0. and A. Gordon Ball. "Public Policy Means and Alternative-" Size. Structure. and Future of Farms. Eds, A. Gordon Ball and Earl O. Heady. Ames, Iowa: Iowa State University Press, 1972. Henderson, Demis R. "Application of the Structure, Conduct, and Pedormance Paradigm to Research on the Structure of Agriculture." in Determinant of Farm Size and Structure, Ed, Lindon J. Robison, Proceedings of the prograrn sponsored by the NC- 18 1 Committee on Determinant of Farm Size and Structure in North Central Areas of the United States, Jan. 16-19, San Antonio, Texas: 1988 Henderson, Dennis R "Measuring and Assessing Vertical Ties in the Agro-food System". cana dia^ Journal of Agricultural Economics. 42(1994):549-560. Hennessy, David. A. " Information Asymmetry as a Reason for Food hdustry Vertical uitegration". American Journal of Agrïcultural Economics. 78 (1996): 1034- 1043. Hennessy, David A. and John D. Lawrence. "Contractual Relations, Control, and Quality in the Hog Sector." Review of Agriarltural Econornics. 2 1,1(1999):52-67. Higrnan, B. W. "Plantations and Typological Problems in Geography" The Australian Geographer Xi, 2(1969): 192-203. International Bank for Reconstruction and Development. The Economic Development of Ceylon, The Johns Hopkins Press: Baltimore: 1953. International Labour Office (ILO). Piantation Workers: Conditions and Standard of Living. Geneva, 1966 Jackson, Jeremy C. and Angela P. Cheater. "Contract Farming in Zimbabwae: Case Studies of Sugar, Tea, and Cotton". in Living Wnder Contract: Contract Farming and Agrarian Transformation in Sub-Saharan Africa. Eds. Peter D. Little and Michael J. Watts. The University of Wisconsin Press, Madison: 1994 Jacob, V. J. "Development of Uneconomic Tea and Rubber Lands." In Plantation Agriculture in Sri Lanka, eds. W. Gooneratne and D. Wesumperuma. International Labor Organization, ARTEP, Bangkok, 1984: 183-205. Jensen, Michael C. and W. Meckling "Theory of the finn: Managerial Behavior, Agency Costs and Ownership Structure." Jotrrnal of Financial Economics 3 (1976): 323- 339 Jogaratnarn, T. "Plantation Agriculture and Economic Development in Ceylon" in Internatiortal Explorations of Agriculttrral Economics. Ed, R. N. Dixey. Iowa State University Press, Arns. 1964. Johnson, Nancy L. and Vernon W. Ruttan. "Why Are Farms So Small?" World Developrnent 22 (1994): 691-706. Joskow, Paul L. "Asset Specificity and the Strcture of Vertical Relationships: Empincal Evidence." Joirrnal of Law. Econornics. and Organization, 4 ( 1, 1988):95- 1 17. Just, Richard E. Darrell L. Hueth and Andrew Schmitz. Appiied Welfare Economics and Public Policy. Prentice-Hall, Inc., Englewood Cliffs, N.J: 1982. Keamey, Roberts N. Trade Unions and Politics in Ceyion. University of California: Berkely: 1971. Kelegama, Saman. Privatization In Sri Lanka: Problems and Prospects. Mimeo, Colombo, 1991. Key, Nigel and David Runsten. "Contract Farming, Smallholders, and Rural Development in Latin America: The Organization of Agroprocessing Firms and the Scale of Outgrower Production." World Development. 27 (1999):38 1-401. Kirk, Colin. "Contracting Out: Plantations, Smallholders and Transnational Enterprise." IDS Bulletin. 18(2, 1987):45-5 1. Kislev, Yoav, and Willis Peterson. "Economies of Scale in Agriculture: A Reexarnination of the Evidence" in nie Economics of Agriculture. Vol, 2. Papers in Honor of D. Gale Johnson. Eds, John M. Antle and Daniel A. Sumner. The University of Chicago Press. Chicago, 1996: 156- 170. Kislev, Yoav, and Willis Peterson. "Prices, Technology, and Farm Size". Journal of Political Econorny, 1982 (Vol. 90, No.3 1). 1990: P.578-595. Klein, B. Crowford, R and Alchian, A. "Vertical Integration, Appropriable Rents, and the Cornpetitive Contracting process." Journal of Law and Economics. 2 1(October 1978):297-326. Klein, Benjamin. "Why Hold-Ups Occur: The Self-enforcing Range of Contractual Relationships." Economic Inquiry, 36(1996):444-463. KLimer, Richard L. 'nie Economic Efficiency of Alternative Exchange Mechanisms." Economic Ewe~zcyin Agriatltural and Food Marketing. Eds, Richard L. Klimer and Walter J. Armbruster, Iowa State University Press, Ames: 1987. Knoebar, Charls R. "An Alternative Mechanism to Assure Contractual Reliability." Journal of Legal Studies, XII ( 1983): 333-343. Knoebar, Charls R. "A Real game of Chicken: Contracts, Tournament, and the Production of Broilers." Jortrnal of Law, Econornics and Organization. 5(2, Fa11 l989):27 1-292. Knutson, Ronald D. J. B. Penn and Willairn T. Boehm. Agrimltural and Food Policy. 3rd Ed, Prentice Hall Education, Career & Technology, Englewood Cliff, New Jersey: 1995. Kreps, David M. A Course in Microeconornic Theory, Princeton University Press, 1990. Kreps, David M. "Corporate Culture and Economic Theory." In Firm Organizations and Contracts: A Reader in Industrial Organization, Eds Peter J. Buckley and Jonathan Michie, Oxford University Press, 1996. Lajili, Kaouthar, Peter S. Barry, Steven T. Sonka, and Joseph T. Mahoney. "Farmers' Preferences for Crop Contracts". Journal of Agricuhral and Resorcrce Economics (Vo1.22, No 2). 1997: pp. 264-280. Lamb, J. 'Tea as a Crop for Small Hotders in Ceylon." Worid Crops, 6(5, May 1954):206-207. Lamb, Geomey and Linda Muller. Control, Accoicntability, and Incentives in a Successfuul Development Institution: The Kenyan Tea Development Aicthority. World Bank Staff Working Papers. No 550. The World Bank: Washington D.C (1982). Lele, Uma and Manmohan Agarwal. "Smailholder and Large-Scale Agriae hure in Africa: Are nere Tradeofls Between Growth and Equity?" MADM Discussion Paper 6, World Bank, Washington D.C: 1989. Leonard, David K. African Successes: Four Public Mangers of Kenyan Rural Development. Berkeley, University of California Press, 199 1. Lim, Young. "Impact of the Tea Industry on the Growth of the Ceylonese Economy." Social and Economic Studies, 1 7 (4, December 1 968):453-467. Lins, David S. "Credit Availability Effects on the Structure of Farming". In Structrire Issues of American Agrintlticre. Agricultural Economic Report 438, Economic, Statistics, and Cooperative Service, USDA, 1979. Little, Angela. Labouring to Learn: Towards Political Economy of Plantations, People and Education in Sri Lanka. Macmillon Press New York, 1999. Little, Peter D. and Michael I. Watts. Living Under Contract: Contract Farming and Agrarian Transformation in Sub-Snliaran Africa, The University of Wisconsin Press, Madison: 1994. Liyanage, Sumanasin. "Partial Privatization of Plantations and its Possible Impact on the Performance of the Plantation Sector in Sri Lanka" in Recent Trertds in Employment and Producriviîy in the Plantation Sector of Sri Lanka, Ed. Y.R Arnerasinghe, International Labor Organization, New Delhi: 1993. Lowenberg-DeBoer, J and Michael Boehlje. "The Impact of Farmland Pnce Changes on Farm Size and Financial Structure". American Jorrrnal of Agrimltural Economics. Vol. 68. (November 1986): 838-848. Lyons, Bruce R. "Contracts and Specific Investment: An Empirical Test of Transaction Cost Theory." Jotrrnal of Econornics and Management Strategy 3 (2, 1994): 257- 278. Maizels, Alfred, Robert Bacon and George Mavrotas. Commodity Supply Management by Producing Countrïes: A Case-Study of the Tropical Beverage Crops. Clarendon Press: Oxford, 1997. Mahoney, 3. T. 'The Choice of Organizational Fom: Vertical Financial Ownership versus Other Methods of Vertical Integration." Strategic Management Journal 13 (1992): 449-584- Majumdar, S. "Problems of Quality Variation, international Substitution, and Dynarnic Adjustments in Primary Trade: The Case of Tea. The Deveioping Econumies, XV(3, l977):293-3 19. Marby, Heidmn. Tea in Ceylon: An attempt at a Regional and Temporal D%/rentiation of the Tea Growing Areas in Ceylon. Franz Steiner Verlag GMBH. Wiesbaden. 1972. Martinez, Steve W. Kevin E. Smith and Kelly D. Zering. "Analysis of Changing Methods of Vertical Coordination in the Pork industry " Jozirnal of Agriciiitural and Applied Econonlics, 30 (1998):3O 1-3 1 1. Masten, Scott E. "Transaction-Cost Economics and The Organization of Agricultural Transactions". Paper presented at NI94 World Food System Projects Symposium: Examining the Economic Theory Base for Vertical Coordination, Chicago, IL Oct. 17-18, 1991. McCune, Shannon. "Sequence of Plantation Agriculture in Ceylon" Economic Geography. 25(1949): 226-235. McFetridge, Donald G. 'The Economics of Vertical Integration." Carradian Journal of Agricziitural Economics. 42(1994): 525-53 1. McFetridge, Donald G. and Douglas A. Smith. The Economics of Vertical DisOrtegration, The Fraser Institute, 1988. McLaren, John. "Why did big Coffee Seek Regulation? A Case Study in Transaction Costs Economics" Working Paper, Department of Economics, Columbia University, Washington: 1999. Mendis, Patrick. Human Environment and Spatial Relutionships in Agricultural Production: The Case Study of Sri Lanka and Other Tea Producing Cozinrries. Peter Lang Publishing; New York. 1992. Meyer, Eric. "From Landgrabbing to Landhunger: High Land Appropriation in the Plantation Areas of Sri Lanka During the British Period". Modern Asian Stridies, 26,2(1992):32 1-36 1. Meyers, William H. and Patrick C. Westhoff. "Commodity Program Reform and the Structure of U.S. Agriculture". In Determinant of Farm Size and Structure, Ed, Arne Hallam, Proceedings of the program sponsored by the NC- 18 1 Cornmittee on Determinant of Fami Size and Structure in North Central Areas of the United States, Jan. 7- 10, Tucson, Arizona: 1989 Mighell, Ronald L. and Lawrence A. Jones. Vertical Coordination in Agriculture. US. Department of Agriculture, Economic Research Services, FmEconomics Division, Agricultural Economics Report No. 19: 1963. Milgrom, Paul and John Roberts. Economics, Organization and Mangement. Prentice Hall, New Jersey. 1992. Miller, T. M. "Economies of Size and Other Growth Incentives." Stmct~tralIssues of American Agriarltrtre. USDA-ESCS. Agr. Econ. Report 438, 1979: 108- 1 15. Monke, Eric, Mark Langworthy, and Francisco Avillez. "Structural Change and Small Farms in the European Community" in Strzrcttrral Change and Small-Fartn Agriculrttre In Northwest Portugal. Eds, Eric Moneke, et al, Corne11 University Press, Ithaca: 1993. Moore, Mick. The State and Peasant Politics in Sri Lanka. Cambridge University Press, Cambridge: 1985. Mulherin, Harold, J. "Specialized Assets, Goverment Regulation, and Organizational Structure in the Natural Gas Industrf' Jorrrnal of Institutional and Theovetical Ecortomics. 142 (1986): 528-541. Nankumba, J. S and B. Kalua. "Contract Farming in Malawi: Smallholder Sugar and Tea Authority." Eastern Africa Economic Review. (Special Issue, August 1989):42-58. National Agriczilttire. Food and Nutrition Strategv: A Change in Perspective. National Planning Division, Ministry of Fimance and Planning, Colombo, Sri Lanka:1984. North Central Regional Research Project 11 7. Coordination and Ekchange in Agriarftural Subsectors. Studies of the Organization and Control of the U.S. Food System. Research Division, College of Agricultural and Life Sciences, University of Wisconsin-Madison: 1976. North, Douglas. Institutions. Institutional Change and Economic Performance, Cambridge: Cambridge University Press, 1990. North, Douglas and John WalIis. "Integrating Institutional Change and Technical Change in Economic History A Transaction Cost Approach." Journal of Institutional and Theoretical Economics, 150 (4, 1994): 609-624. O'Brien, Patrick M. "Implications for Public Policy". In Food and Agricultzo-al Markets: The Quite Revolution, Eds Lyle P. Schertz and Lynn M. Daft. Economic Research Services, U.S. D.A, and Food and Agricultural Cornmittee, National Planning Association, Washington D.C: 1994. OTA (Office of Technology Assessment). Technology, Public Policy. and the Changing Structure of American Agriculture. U.S. Government Printing Office, Washington: 1986. Pasour, E. C. Jr. Agriarlture and the State: Market Processes and Bttreaucracy. Holrnes & Meier, New York: 1990 Peebles, Patrick. Sri Lanka: A Hand Book of Historical Statistics. G.K Hall & Co., Boston: 1982. Penn, J. B. "The Stmcture of Agriculture: An Overview of the issue". in Strrrcture Issues of American AgricuItrtre. Agricultural Economic Report 438, Economic, Statistics, and Cooperative Service, USDA, 1979. Peins, G.H. "Structural Change in Plantation Agriculture in Sn Lanka." In Plantation Agriculture in Sri Lanka, eds. W. Gooneratne and D. Wesurnperuma. International Labor Organization, ARTEP, Bangkok, 1984:33-57. Peins, G.H. "Land Reform and Agrarian Change in Sri Lanka." Modem Asian Studies, 12(4, 1978): 6 1 1-628. Peins, G. H. "Plantation Agriculture" in Sri Lanka: A Suney. Ed. K.M. De Silva. C. Hurst & Company, London: 1972. People's Bank of Sri Lanka. "Profit-Privatization and Losses-Peopolization " Econonric Review, 2 l(6, 1995 September): 7-8. Perera, A. B. "Plantation Economy and Colonial Policy". Ceylon Historical Journal. 1 (1% 1): 46-58. Perera, A. M. S. ""TheTea Industry of Sri Lanka." Marga 3 (4, l976):93- 103. Perry, Martin K. "Vertical integration: Determinants and Effects." in Handbook of industrial Organization, Vol. 1, Eds, Richard Schmalensee and Robert Willig, North-HolIand: New York, 1989. Pim, Alan. Colonial Agricuftural Production. Oxford University Press. London: 1946. Plantation Sector Sratistical Pocket Book. Planning Division, Ministry of Plantation Industries. Sri Lanka: 1997, 1999,2000 Price, Dorothy 2. Stephen J. Hiemstra, Vicki A. McCracken, and David Pnce. "American Food-Away-Frorn-Home Consumption." In Japanese and Amerïcan Agriculture: Tradition and Progiess in Conflict. Eds. Luther Tweeten, Cynthia L. Dishon, Wen S. Chem, Naraomi Imarnura, and Masaru Morishima. Westview Press: Boulder, 1993. Pryor, Frederic L. The Plantation Economy as an Economic System. Journal of Comparative Economics. 6 ( 1982):289-3 17. Rarnachandran, N. Foreign Plantation Investment in Ceylon 1889- 1958. Central Bank of Ceylon, Colombo. 1963 Ramanujarn, P. "Supply Function of Perennial Crops Re-exarnined: The Case of Tea" Upanathi (publication of Social Science Association of Sn Lanka), l(2, 1986 July): 157- 197. Ranasinghe, Sudatta. "Land Reform in Decade of IWO." Econonzic Review. Research Division, Peoples Bank, Sn Lanka, l(6, September 1995):9- 11. Rasmussen, Wyane D. "Agricultural Structure and the Well Being of Society Revisited". In Determinant of Farm Size and Stntcture, Ed, Arne Hallam, Proceedings of the program sponsored by the NC-18 1 Committee on Determinant of Farm Size and Structure in North Central Areas of the United States, Jan. 7-10, Tucson, Arizona: 1989 Raup, Philip M. "Some Questions of Value and Scale in American Agriculture". American Journal of Agricultural economics, (May l978):303-308. Reining, C. R. "Two Methods for Estimating Real Structural Change in Agriculture." Journal of Agricultural Economics Research, 4 1 ( 1, Winter 1989): 39-43. Reining, C. R. Structrrral Change in U.S. Farmland. Agricultural Economic Report. No. 6 17. United States Department of Agriculture, Economic Research Service, 1990 Report of the Presidential Commission on the Tea Indusft-y and Tmde. Sessional Paper No II, 1995. Department of Government Printing: SR Lanka. Report of the Baseline Survey of Tea Small Holdirrgs. Ministry of Plantation Industries. Sri Lanka. October 1994. Report of the Committee Appointed to Study the Constraints and Problems of the Tea SmalZ Holdings Sector. Ministry of Plan Irnplimentation, Colombo. Sri Lanka. 1980. Report of the World Tea Economy. International Bank for Reconstruction and Development, Econornic Department Report No. EC 178. 1971. Richardson, James W., Edward G. Smith, and Ronald D. Knutson. "Who Benefits From Farm Programs: Size and Structure Issues?" In Determinant of Farm Size and Structure, Ed, Lindon J. Robison, Proceedings of the prograrn sponsored by the NC-18 1 Comrnittee on Deteminant of Farm Size and Structure in North Central Areas of the United States, Jan. 16-1 9, San Antonio, Texas: 1988 Riordon, Michael and Oliver Williamson, "Asset Specificity and Economic Organization." International Journal of Industrial Organization, 3 (1985):365- 378. Rhodes, James V. 'The industrialization of Hog Production" in The industrialization of Agricuitzrre: Verrical Coordination in the US.food system. Eds, Jeffrey S. Royer and Richard T. Rogers. Ashgate, Brookfield: 1998. Robert, Michael and L.A. Wickremeratne. "Export Agriculture in the Nineteenth Century" in University of Ceyion: History of Ceylon Volume nree. Editor-in- Chief, K.M.De Silva. The Colombo Apothecaries Company Ltd., Colombo: 1973:89-118. Roberts, James. "A Micro-Economic Analysis of Tea Production Using a Separable Restncted Profit Function." Journal of Agriculturul Economics. 40(1, January 1989): 185-197. Rodgers, John L. Structural Changes in Agriculture. Ph.D. Thesis, Department of Agricultural and Applied Economics, University of Minnesota, St. Paul, 1988. Rote, Ron. A Taste of Bitterness: The Political Economy of Tea Plantations in Sri Lanka, Free University Press: Amsterdam, 1986. Ruttan, Vernon M. "Scale, Size, Technology and Structure: A Persona1 Perspective". In Determinant of Farm Size and Structure, Ed, Lindon J. Robison, Proceedings of the prograrn sponsored by the NC-181 Cornmittee on Deteminant of Fam Size and Structure in North Central Areas of the United States, Jan. 16-19, San Antonio, Texas: 1988 Sadaratne, Nimal. "Sri Lanka's New Land Refonn." South Asian Review, 6(1,0ctober l972):7- 19. Sajhau, Jean-Paul Jurgen Von Muralt. Plantations and Plantation Workers. international Labor Office, Geneva: 1987. Sarkar, Goutam. Tea: Some Policy Issues. Discussion Paper. Institute of Development Studies at the University of Sussex. Brighton: England. 1978. Sarkar, Goutam. The World Tea Economy. Oxford University Press: Delhi. 1972. Schultz, Theodore W. Transforming Traditional Agriculture. New Haven and London, Yale University Press: 1964 Sexton, Richard J. "Industnalization and Consolidation in the US. Food Sector. Implication for Competition and Welfare." American Journal of Agricultural Economics, 82 (2000): 1087- 1 104 Shlomowitz, Ralph. " Plantations and Smallholdings: Comparative Perspectives from the World Cotton and Sugar Cane Economies, 1865-1 939". Agricultural History 58, î(l984): 1-16 Singh, Manamohan. "Development Policies in Ceylon: 1966- 1971 ." Economic Bulletin for Asia and the Far East, XXII (1, 1972): 19-36. Simathamby, M and P.Devaraj. "Crisis in the Plantation Sector; the dilemmas after 1977." In Essays on the Sri Lankan Economy 1977-1983, eds. David Dunharn and Charls Abeysekara. ISS/SSA, Colombo, l987:220-248. Sinnathamby, M. "The State Sector Plantations: Recent Trends in Productivity, Employment, and Earnings of Workers." in Recent Trends in Employment and Prodzictivity in the Plantation Secfor of Sri Lanka. Ed Y.R.Amerasinghe, International Labor Organization, Asian Regional Team for Employment Promotion (ARTEP). New Delhi: 1993: 25-46. Sinnatharnby, M and P. Wickrarnasekara. "Labor Availability and Labor Demand in the Plantation Sector" In Plantation Agriczrlftrre in Sri Lanka, eds. W.Gooneratne and D. Wesumperuma. international Labor Organization, ARTEP, Bangkok, l984:87- 109. Sivaram, B. and Deepananda P. Herath. Developrnent Planning in Tea. Monograph. Tea Research hstitute of Sri Lanka. Thalawakele, Sri Lanka. 1996. Snodgrass, Donald R. Ceylone: An Export Economy in Transition. Richard D. Irwin, Inc. Illinois. 1966. Sriyaratne, H. "The relative Importance of small and large scaie production of Tea in the tea industry of Sn Lanka." Central Bank ofSri Lanka StaflStudies. 24(1 & 2, 1994): 1 12- 154. Spitzer, J. J. "A Primer on Box-Cox estimation." The Review of Economics and Statistics, 64 (1982): 307-3 13. Spitzer, J. J. "Variance Estimates in Models with the Box-Cox Transformation: Implications for Estimation and Hypothesis Testing." The Review of Economics and Statistics, 66(l984): 645-652. Sporleder, Thomas L. "Managerial Economics of Vertically Coordinated Agricultural Firms". American Journal of Agrbltural Economics- (Decernber 1992): 1226- 1231. Stall Steven, Christopher Delgado and Charles Nicholson. "Srnallholder Dairying Under Transaction Costs in East Afnca." World Development. 25 (5, 1997): 779-794. Stanton, B. F. "Perspective on FmSize". American Journal of A@mltural Economics. (December) 1978: 727-73 7. Stanton, B. F. " Changes in Farm Size and Structure in Amencan Agriculture in the Twentieth Cenhuy". Determinant of Farm Size and Structure, Ed, Arne Hallarn, Proceedings orthe prograrn sponsored by the NC-18 1 Committee on Determinant of Farm Size and Structure in North Central Areas of the United States, Jan. 7-10, Tucson, Arizona: 1989: I 1-46. Stanton, B. F. "Structural Change in Arnencan Agriculture into the Next Century". In Japanese and American Agricultrire: Tradition atrd Progress in Conflict. Eds. Luther Tweeten, Cynthia L. Dishon, Wen S. Chern, Naraomi Imamura, and Masam Morishima. Westview Press: Boulder, 1993. Stigler, George J. "The Division of Labor is Limited by the Extent of the Market." Journal of Political Economy, 59 (3, 195 1): 185-193. Sumner, Daniel A. "Concentration and Food System: The Role of Public Policy". In Food and Agriculttiral Markets: The Qttite Revoltrtion, Eds Lyle P. Schertz and Lynn M.Dafi. Economic Research Services, USDA, and Food and Agricultural Cornmittee, National Planning Association, Washington D.C: 1994. Sumner, Danial A. "Fann Programs and Structural Issues." In US.Agrictrltrrral Policy: The 1985 Farm Legislafion. ed, Bruce L. Gardner. American Enterprise Institute for Public Policy Research, Washington, D-C: 1985. Sundararn, Lanka. "Indian Labor in Ceylon." International Labour Review, XMII(3, 193 1 March):370-387. Tea Master Plan, Position Paper No. 1 (Colombo): 1978 Tharian, George K. V. Haridasan and B. Sreekurnar. "Role of Governrnent and Structural Changes in Rubber Plantation Industry." Economic and Political Weekly (November, 26 1988): 158-166. Thomas, Clive. "Agrarian Change in Plantation Economy" in Agrarian Systems and Rural Development. Eds. Dharam Ghai, Azizur Rahman Khan, Eddy Lee and Samir Radwan. Holmes & Meier Publishers, inc: New York, 1979. Thorbecke, Erik and Jan Svejnar. Economic Policies and Agriatituraf Performance in Sri Lanka: 1960-1984. Developrnent Centre of the Organization for Economic Co- operation and Development: OECD, Park: 198% Tiffen, Mary and Michael Mortimore. neory and Practice in Plantation Agriculture: an Economic Review. Overseas Development Lnstitute, Westview Press, Boulder: 1990. Tilakaratna, S. "Potential for Agricultural Diversification of Lands under Coconut to increase Productive Employment." In in Piantation Agricuiture in Sri Lanka, eds. W. Gooneratne and D. Wesumperuma. International Labor Organization, ARTEP, Bangkok, l984:207-232. Tirol, Jean. The neory of lndustrial Organization. The MIT Press;Carnbndge, 1 994. Tomek, William G. and Kenneth L. Robinson Agriniltural Product Prices. 3" Ed, Corne11 University Press, Ithaca 1990. Trifon, Rapheal. "Guides for Speculation about the Vertical integration of Agriculture with Allied industries." Journal of Farm Economics. XLI, 4( 1959):734-746. Tweeten, L. Causes and Conseqztences of Structurai Change in the Farming Iridustty . National Planning Association, Wasington, D.C.: 1984. Tweeten, L and Wallace Hufhan. "Structural Change" in Structure of Agriculture and Information Needs Regardirig SmaZl Farms. Eds, Luther Tweeten, Wallace Hufhan, Steven T. Sonka and Richard D. Rodefeld. Washington, D.C.: The National Rural Center, 1980. Tweeten, L . "American Agriculture; Organization, Structure, Institutions, and Policy". In Japanese and American Agriarltwe: Tradition and Progress in Conflict. Eds. Luther Tweeten, Cynthia L. Dishon, Wen S. Chem, Naraomi Imamura, and Masam Morishima. Westview Press: Boulder, 1993. United Nations Conference on Trade and Development (UNCTAD).Studies in the Processing, Marketing and Distribution of Commodities: Marketing and Processing of Tea: Areas for International Co-operation. United Nations, New York, 1984 U.S. Department of Commerce, Bureau of the Census. Census of Agriculture. 1992 Weersink, A, S. Clark, C. G. Turvey, and R. Sarker. "The Effect of Agricultural Policy on Farmland Values." Land Economics 75 (3, August 1999):425-439. Wickizer, V. D. "The Smallholder in Tropical Expoa Crop Production." Food Research Institute Studies l(l96O) :49-99. Wickremaratne, L. A. " The Establishment of the Tea Industry in Ceylon." Ceylon Journal of Hrrton-cul and Social Studies. 2(2, 1972 July-Dec): 131-1 55. Williamson, Oliver. "The vertical Integration of Production: Market Failure Considerations." American Economic Review LX1 (1 971): 1 12- 123. Williamson, Oliver. "Transaction Cost Economics: The Govemance of Contractual Relations." Journal of Law and Economics 22 (1 979): 3-6 1. Williamson, Oliver. The Economic Institutions of Capitalism. The Free Press. New York. 1985 Williamson, Oliver. "Transaction Cost Economics and Organization Theory. Industrial and Corporate Change.(2, 1993): 107- 156. Williamson, Oliver. "Transaction Cost Economics: How it Works; Where it is Headed." De Economist l46(l, 1998):23-58. World Bank. Sri Lanka S Tea Industry.. Succeeding in the Global Market. Discussion Paper No. 368, The World Bank, Washington, D.C: 1997 World Bank. World Development Report, Oxford University Press, New York, 1986. Whyte, Glen. "The role of asset specificity in the vertical integration decision." Journal of Economic Behavior and Organization, 23 ( 1 974):287-302. Zarembka, P. "Transformation of variables in econometric." In Frontiers in Econometrics, Ed, P. Zarembka. New York: Academic Press, lW4:8 1-104. Zering, Kelly. "The Changing U.S. Pork Industry" in The Indrtstria[ization of Agriariture: Vertical Coordination in the U.S.food system. Eds, Jeffrey S. Royer and Richard T. Rogers. Ashgate, Brookfield: 1998. mX6A Proxy Variable for Sri Lankan Tea Farm Size Distribution (YI) Year Total Area of Total Area of Total Tea Area ReIative Share < 10 acres farrns (acres) > 10 acres farms (acres) (acres) of small farms 1960 /Y466 5m33 5817 0.1 366 Source: ïïC Bulletin of Statistics Various Issues From 1993 to 1 998 an estimation with the time and 40acre farms total area; Tea area of < 10 acre farms = 2 1 8 1.6 (Year) - 4 194544 Proxy Variable for Ratio of Labor Opportunity Costs (WR) Year Workers in Plantation Workers in Industry Ratio of Industrial to Plantation Real Wage (Rdday) Real Wage (Wday) Real Wags 1960 2.09 3.61 1.73 Workers in plantation (agriculture): Tea Growing and Manufactunng, Rubber Growing and Manufacturing, Coconut Growing Workers in Industry and Commerce: Baking, Tile Manufacturing, Coconut Manufactunng, Printing, Textile, Tyre and Tube manufacturing, Cou Mattresses and Bnstle Export, Hosiely Manufacturing, Engineering, Garment Manufacturing, Match Manufacturing, Biscut and Confectionary, Tea Export and Rubber and Rubber Export trades only. Sources: Central Bank of Sri Lankan, Monthly Bulletin 1979 June, Central Bank of Sri Lanka, Annual Reports, 1990, 1998. Sri Lanka: A Handbook of Historical Statistics, 1982 P. 107 Amex 6C Proxy Variable for Small Tea Farm Support Services (SSR) Year Small Tea Farm Small Tea Farrn Subsidies for mal1 Subsidies (Rs.Million) Area (cl 0 acres) ha tea fanns (Real Rdha) 1960 3 -47 32159 104.16 1961 4.47 33803 126.29 1962 6.45 34983 173.55 1963 5.30 36356 134.0 1 1964 4.97 38730 114.33 1965 5 .O0 39395 112.82 1966 5.00 40 198 110.76 1967 4.49 40926 95.63 1968 5.30 41539 105.0 1 1969 5 -00 42034 91 -29 1970 5.00 42690 84.75 1971 6.37 43 127 104.12 1972 5.85 43572 89.0 1 1973 6.33 44129 86.47 1974 5.39 44745 64.88 1975 6.54 4528 1 72.85 1976 7.7 1 47886 80.22 1977 11.68 49074 117.10 1978 20.23 4999 1 178.07 1979 22.50 51320 173.84 1980 10.54 52085 63.60 1981 1 1.75 52614 59.47 1982 16.37 53 13 1 74.05 1983 14.53 53916 56.83 1984 16.19 54966 53 -25 1985 34.68 56266 109.83 1986 66.2 1 57246 190.86 1987 57.07 58016 150.69 1988 40.00 58626 88.14 1989 7 1.20 59257 144.73 1990 79.58 59875 131.78 1991 89.10 60673 129.79 1992 78.41 60849 102.24 1993 79.90 63596 89.2 1 1994 82.67 63373 85 -41 1995 82.98 64180 78 .62 1996 53.15 64986 42.89 1997 59.17 65793 43 .O5 1998 6 1.44 66600 40.37 The information about the subsidies provided for mal1 tea farms are gathered fiom different sources. These are outlined as follows; 1960, 1961: actual expenditure incurred on the small tea fmssubsidies as reported in "The Short-tem Implementation Program". The Depariment of National Planning, Government Press, Colombo, Ceylon, 1962: p. 105 1962, 1963, and 1964: proposed expenditure for small tea fann subsidy as reported in "The Short-rem Implementaation Program". The Department of National Planning, Government Press, Colombo, Ceylon, 1962: p. 1OS 1965; assumed to be equal for 1966 proposed expenditure 1966: Proposed expenditure for tea small farm subsidy as reported in "Ceylon National Development Han 1968". The Department of National Planning, Govenunent Press. Colombo, Ceylon 1968: p.9 1967, 1968: actual expenditure incurred on the small tea farms subsidies as reported in "Ceylon Year Books" 1969 and 1970, The Department of Statistics, Colombo, Sn Lanka (p.64 for 1969, p.72 for 1970) 1969, 1970: proposed expenditure incurred on the small tea farms subsidies as reported in "Ceylon Year Book 1969", The Department otStatistics, Colombo, Sri Lanka (p.64) 1971 to 1979: Small tea fmssubsidy is estimated as a percentage of the Cess levy (this is a ievy on tea exports fiom Sn Lanka to be used on tea research, tea promotion, and tea subsidies). From 1960 to 1970, average of 15% fiom the Cess levy has been used on small fmsubsidy. Therefore, 15% of Cess fiom 1971 to 1979 is used as the estimator of subsidies for small tea fms(Cess levy for 1960 to 1998 is given in Annex 6D). 1980 to 1998: actual expenditures on srnall tea fmsa reported in "Plantation Sector Statistical Pocket Book 1999": Sri Lanka, Colombo (p.47) Proxy Variable for Rate of Export Taxes on Tea Year Different Taxes (Rs. Million) Total Black Tea Export Tax Cess Advalorem Export Total @@ Mn) (Real Rs/kg) .. . . -. . . .. - .. . --. - Sources: 1960- 1966: cess, advalerom tax, export tax are from "Review of the WorZd Tea Economy" International Bank or Reconstruction and Development, 1970 (Annex 16) 1967:advalerom and export taxes are fiom "Review of the World Tea Economy" International Bank or Reconstniction and Development, 1970 (Annex 16) 1967: cess is estimated with the cess rate (Rskg 0.145) of 1966 1968-1969: advalerom and export taxes are from Annual Report, Central Bank of Sri Lanka, Colombo 1970 1968: cess is fiom "Report on the World Tea Economy" International Bank of Reconstruction and Development, 197 1:p.38 1969: cess is estimated with the cess rate (Rs/kg 0.203) of 1968 1970-1974: cess, advalerom tax, export tax are fiom Perera, A.M.S. "The Tea industry of Sri Lanka" Marga Vol. 3 No. 4, 1976:p.93 1975-1978: advalerom and export taxes are from Monthly Bullelin, Central Bank of Ceylon (1979 July, 1978 December, 1978 January, and 1974 December) 1975:cess is estimated with the cess rate (Rs/kg 0.205) of 1974 1976-1978: cess are calculated fiom the cess rates given in "Plantation Sector Statistical Pocket BooK', Colombo, Sn Lanka 1999:p.39 1979: cess, advalorem, and export taxes are fiom Annual Report, Sri Lanka Tea Board, Colombo, Sri Lanka 1980 1980-1998: cess, advalorem, and export taxes are fiom "Plantation Secror Statistical Pocket Book", Colombo, Sn Lanka 1999:p.41,43 Tea Exports 1960- 1982, Thorbecke and Svejnar, 1984: p. 78 and 1982- 1998 "Ptantation Sector Statistical Pocket Book", Cofombo, Sri Lanka 1999:p.8 Colombo Consumer Price index is used to calculate the real taxes. Annex 6E Proxy Variable for Tea Price Instability in Colombo tea Auction (PIN) Y ear Real Tea Prices Moving Average % Deviation (Rs/kg of Black tea) (5 years) fiom trend (absolute value) 1952 4.320 1953 4.664 1954 6.369 1955 5.403 1956 5.419 1957 4.463 1958 4.067 1959 4. Z 74 1960 4.278 1961 4.056 1962 3.857 1963 3.566 1964 3.476 1965 3.582 1966 3.161 1967 3 .O3 1 1968 3.317 1969 2.724 1970 2.728 1971 2.9 18 1972 2.924 1973 2.660 1974 3.617 1975 3.480 1976 4.609 1977 8.027 1978 6.274 1979 4.877 1980 5.761 1981 4.820 1982 5.633 1983 9.125 1984 11.352 1985 6.95 1 1986 5 .O63 1987 6.020 1988 5 -525 1989 6.578 1990 7.036 1991 5.150 1992 4.899 5.838 16.08 1993 4.918 5.716 13.96 1994 4.263 5.253 18.84 1995 4.391 4.724 7.06 1996 5 -448 4.784 13.88 1997 5.715 4.947 15.53 1998 5.880 5.140 14.4 1 Sources: Colombo tea pnces (average al1 teas) are from; 1952- 1970, Annual Report, Central Bank of Sri Lanka 1974 (Table 54). 1970 to 1980, Report of the Presidential Commission on the Tea Industry and Trade, Colombo, Sn Lanka 1995:p. 9, 1980 to 1998, Plantation Sector Statistical Pocket Book, Colombo, Sri Lanka 1999: p. Deflated by Colombo Consumer Price index (1952=100) ANNEX 6F Export Volume and Value for the Main Alternative Crops for Tea - - Volume (Thousand kg) Value (Million Rs) Year Pcpper Cinnornon CIoves Nutmeg Cardomom Cocoa Pepper Cinnomon Cloves Nutmeg Cardornom Cacoa Contd, Proxy Variable for Returns from Alternative Agricultural Enterprises Per Kg Value Average Year Pepper Cinnomon Cloves Nutmeg Cardomom Cocoa 1960 1998 13.69 - - 3.24 32.82 ------. Sources: 1967- 1975 fkom Monthly Bulletin of Central Bank of Ceylon (1 979 July, 1978 December, 1978 January, and 1974 December) 1976 -1998 fkom Annual Report, Central Bank of Sri Lanka (various issues) Average values calculated by Export Value/Export Volume. 1960 to 1966 are estimated using the trend fiom 1970 to 1967. Proxy Variable for Changes in Vertical Coordination Arrangements (&) Year Black tea production fiom market procured green leaf (Mn kg) 1960 35.83 305 ANNEX 6H Reat Black Tea Prices Y ear Average Colombo Auction Pnce (1 952 Rskg) Source: Annual Report Central Bank of Sn Lanka 2000 (Special Statistics Appendix Table 13) Deflated by Colombo Consumer Pnce index (1952=100).