Livestock to 2020 the Next Food Revolution

Food, Ag ri cul ture, and the En vi ron ment Discus sion Paper 28

Live stock to 2020 The Next Food Revolution

Chris to pher Del gado Mark Rosegrant Henning Ste in feld Simeon Ehui Claude Cour bois

IN TERNATIONAL FOOD AND IN TER NA TIONAL FOOD AG RI CUL TURE LIVE STOCK POL ICY OR GANI ZA TION RE SEARCH RESEARCH OF THE IN STI TUTE INSTI TUTE UNITED NA TIONS “A 2020 Vision for Food, Ag ri cul ture, and the Envi ron ment” is an ini tia tive of the In ter na tional Food Policy Re search Insti tute (IFPRI) to de velop a shared vision and a con sen sus for action on how to meet fu ture world food needs while reduc ing pov - erty and pro tect ing the en vi ron ment. It grew out of a con cern that the in ter na tional com mu nity is setting pri ori ties for ad dress ing these problems based on incom plete in for ma tion. Through the 2020 Vision initia tive, IFPRI is bring ing together diver - gent schools of thought on these is sues, gener at ing re search, and iden ti fy ing recom - menda tions.

This dis cus sion pa per series pres ents techni cal research results that en com pass a wide range of sub jects drawn from re search on policy-relevant as pects of ag ri cul - ture, pov erty, nu tri tion, and the envi ron ment. The dis cus sion papers con tain ma te rial that IFPRI be lieves is of key in ter est to those involved in ad dress ing emerg ing food and de vel op ment prob lems. The views ex pressed in the papers are those of the authors, and not nec es sar ily en dorsed by IFPRI. These dis cus sion papers underg o re view but typically do not present final research results and should be con sid ered as works in progress . Food, Agri cul ture, and the En vi ron ment Discus sion Paper 28

Livestock to 2020 The Next Food Revolution

Chris to pher Del gado Mark Rosegrant Hen ning Ste in feld Simeon Ehui Claude Cour bois

In ter na tional Food Policy Research Insti tut e 2033 K Street, N.W., Washing ton, D.C. 20006- 1002 U.S.A.

Food and Agriculture Organization of the United Nations Viale delle Terme di Caracalla, 00100 Rome, Italy

International Livestock Research Institute P.O. Box 30709, Nairobi, Kenya

All rights re ser ved. Sections of this report may be repro du ced without the ex press per mis sion of but with ackno wledg ment to the In ter na tio nal Food Po licy Research Insti tu te .

ISBN 0- 89629-632-6 Content s

Forewor d vii Acknowledgments viii 1. The Live stock Revolu tion 1 2. Recent Trans for ma tion of Live stock Food Deman d 5 3. Accom pa ny ing Trans for ma tion of Live stock Sup ply 13 4. Pro jec tions of Future De mand and Sup ply to 2020 21 5. Impli ca tions of the Live stock Revo lu tion for World Trade and Food Prices 31 6. Nutri tion, Food Secu rity, and Pov erty Al le via tion 37 7. En vi ron men tal Sustain abil ity 44 8. Public Health 49 9. Tech no logy Needs and Prospects 52 10. Tak ing Stock and Moving Forwar d 59 Appen dix: Regional Clas si fi ca tion of Coun tries Used in this Paper 66 Refer ence s 67

iii Ta bles

1. Increase in food consump tion of meat, milk, fish, and major ce re als, 1971–95 2 2. Per cap ita meat and milk consump tion by region, 1983 and 1993 6 3. Percent of calories and pro tein from ani mal products, 1983 and 1993 6 4. Past popula tion, urban popu la tion, and GNP per cap ita growth rates 7 5. An nual per cap ita consump tion of se lected live stock foor prod ucts and percen t of total calories con sumed from each prod uct, 1973 and 1993 9 6. Consump tion of meat and milk by region, 1982–94 9 7. Trends in the food consump tion of vari ous live stock products, 1982–94 10 8. De mand elastici ties for major food prod ucts of ani mal ori gin from a cross- country, systems es ti ma tion, 1970–95, devel op ing re gions 11 9. Pro duc tion trends of vari ous live stock products, 1982–94 13 10. Trends in the pro duc tion of meat and milk, by region, 1982–94 14 11. Shares of total world pro duc tion of meat and milk, by region, 1983 and 1993 14 12. Dis tri bu tion of the world’s live stock animals, 1983 and 1993 15 13. Growth rates of live stock out put and number of animals slaughtered or milked, 1982–94 16 14. Produc tiv ity by re gion and ani mal type, 1992–94, and produc tiv ity growth rate, 1982–94 17 15. Trends in the use of ce real as feed, 1982–94 20 16. Pro jected trends in the food consump tion of various live stock prod ucts, 1993–2020 24 17. Pro jected trends in meat and milk con sump tion, 1993–2020 24 18. Pro jected trends in meat and milk pro duc tion, 1993–2020 25 19. Pro jected trends in pro duc tion of various live stock products, 1993–2020 26 20. Pro jected trends in use of cere als as feed, 1993–2020 26 21. Changes in IM PACT’s base line assump tions: a severe Asian crisis and high Indian meat consump tion 27 22. Differ ence be tween baseline pro jec tions and pro jec tions of aggre gat e consump tion in 2020 due to changes in as sump tions about the Asian cri sis and In dian consump tion 28 23. Differ ence be tween baseline pro jec tions and pro jec tions of aggre gat e consump tion in 2020 due to changes in as sump tions about feed conver sion 29 24. Net ex ports of various livestock prod ucts in 1993 and 2020, baseline scenario 32

iv 25. Dif fer ence be tween baseline pro jec tions of net ex ports in 2020 and projec tion s with changes in assump tions about the Asian crisis and In dian consump tion 32 26. Dif fer ence be tween baseline pro jec tions of net ex ports in 2020 and projec tion s with changes in assump tions about feed conver sion effi cienc y 34 27. Past trends in real prices of se lected crop, feed, and live stock products 34 28. Real prices of se lected crop and live stock prod ucts as pro jected by the IM PACT model 35 29. The place of live stock in the income of the rich and poor 40

Illus tra tion s

1. Per cap ita consump tion of meat and milk, de vel op ing and devel ope d countries, 1983 and 1993 5 2. The re la tion ship be tween meat consump tion and income 6 3. Trends in the prices of major ce real and meat commodi ties, 1970–72 to 1994–96 8

v Fore word

The com bined per cap ita con sump tion of meat, eggs, and milk in devel op ing coun tries grew by about 50 per cent from the early 1970s to the early 1990s. As incomes rise and cit ies swell, peo - ple in the devel op ing world are diversi fy ing their diets to include a variety of meats, eggs, and dairy products. This trend toward diversi fied eating habits is likely to con tinue for some ti me to come and it has led to con sid er able contro versy about the risks and opportu ni ties involved. Some observers fear that greatly increased demand for feed grains will raise the price of cere als to the poor. Others are concerned that higher con cen tra tion of live stock produc tion near ci t ies adds to pollu tion. Still others worry about the pub lic health effects of increased con sump tio n of animal fats and the rapidly increasing incidence of diseases passing from animals to humans. On the other hand, many analysts point to the nutritional bene fits of increased con sump tion of an i - mal prod ucts for popula tions that are still largely defi cient in intake of protein and micro n u tri - ents. Further more, live stock tra di tion ally have been an important source of income for the r ural poor in devel op ing countries. Finally, increased demand for live stock prod ucts may provide an engine for sus tain able intensi fi ca tion of small holder food and feed produc tion systems. A team of research ers from the Interna tional Food Policy Research Insti tute (IFPRI), the Food and Agricul tural Organi za tion of the United Nations (FAO), and the Interna tional Live stock Research Insti tute (ILRI) col labo rated to pro duce this com pre hen sive and even- handed attemp t at defin ing the nature, extent, scope, and impli ca tions of what they term the “Live stock Revo lu t ion” in develop ing countries. Look ing for ward to 2020, they argue convinc ingly that the struc tu ral shifts in world agricul ture being brought about by shifts in developing-country demand for foods of ani mal ori gin will con tinue and that increas ingly global markets have the ability to sup p ly both cereal and ani mal products in desired quan ti ties with out undue price rises. They empha size, h ow- ever, that policy deci sions taken for the live stock sec tor of devel op ing coun tries will dete rmin e whether the Livestock Revo lu tion helps or harms the world’s poor and malnour ished. The report empha sizes the importance of con tin ued invest ment in both research on and devel op ment of ani - mal and feed grain pro duc tion and process ing, and the need for policy action to help small, po or live stock produc ers become better integrated with commer cial live stock mar ket ing and proc ess - ing. It details a host of require ments in the area of technol ogy devel op ment for produc tion and proc ess ing of live stock prod ucts, potential benefits from new technolo gies, and criti cal p olicy issues for environ mental conser va tion and pro tec tion of pub lic health.

Per Pinstrup-Andersen, Di rec tor Gen eral Inter na tional Food Pol icy Research Insti tut e

Ab dou laye Sa wa dogo, Assis tant Director-General, Agri cul ture De part ment Food and Agri cul tural Or gani za tion of the United Nation s

Hank Fitzhugh, Di rec tor Gen eral International Livestock Research Institute

vii Acknowl edg ment s

The ambitious and multi dis ci plin ary topic of this paper hints at the extent to which the auth ors had to rely on help from colleagues with a wide variety of dis ci pli nary and geographic exper - tise. They are too numer ous to men tion indi vidu ally, but several col leagues stand out because of the degree of their sup port for this col labo ra tive project and the depth of their insights on pre vi ous drafts. Per Pinstrup-Andersen and Rajul Pandya- Lorch, direc tor gen eral and head of the 2020 Vision ini tia tive, respec tively, of the Inter na tional Food Policy Research Institu t e (IFPRI), pro vided an institu tional frame work for the project, constant encour age ment, and detailed and insight ful com ments through out the process. Abdoulaye Sawadogo, assis tant director-general of the Food and Agricul ture Organiza tion of the United Nations (FAO), quickly rec og nized the value of this col labo ra tion, discussed issues with the team, and facil i - tated the con di tions for effec tive FAO par tici pa tion. Hank Fitzhugh, direc tor gen eral of t he Interna tional Live stock Research Institute (ILRI), con sis tently sought to integrate ILRI’s broad strengths with respect to live stock issues in develop ing coun tries with our activities a nd facilitated ILRI’s effective partici pa tion . A number of formal and infor mal external reviewers of earlier drafts of the report greatly improved the final prod uct. Particu lar men tion should be made of the very detailed and help ful com ments of Cees de Haan of the World Bank and Maggie Gill of Natural Resources Interna - tional (U.K.) in this regard. While the tech ni cal live stock pro duc tion aspects of this report, authored as it is by economists, probably still falls short of their high standards, there is no doubt that they sub stan tially improved it over what it would have been oth er wise. Cath er ine Geissler of King’s Col lege, London, helped improve the nutri tional insights of the report; the remain ing defi cien cies are entirely the respon si bil ity of the authors. The latter would also like to express their thanks for very help ful written comments to Mercy Agcaoili- Sombilla of the Interna tional Rice Research Insti tute (IRRI), Jan Slingen bergh of FAO, Steve Staal of ILRI, Claudia Ringler of IFPRI, Bob Havener of the World Food Prize Office, and Tjaart Schillhor n Van Veen of the World Bank. Finally, they would like to commend Uday Mohan, an IFPRI editor, on his cheer ful and success ful struggle under time pressure to turn our col lec tivel y authored, carefully hedged prose into a readable final report.

viii 1. The Live stock Revolu tio n

A revolu tion is taking place in global agricul ture The 23 percent of the world’s popu la tion liv ing that has pro found impli ca tions for our health, live - in developed coun tries pres ently con sume three to li hoods, and environ ment. Popula tion growth, four times the meat and fish and five to six times the urbani za tion, and income growth in develop in g milk per cap ita as those in devel op ing countrie s 1 coun tries are fueling a massive global increase in (Delgado, Cour bois, and Rosegrant 1998). But mas - demand for food of ani mal origin. The resulting sive annual increases in the aggre gate con sump tion demand comes from changes in the diets of billions of ani mal prod ucts are occur ring in devel op ing coun - of peo ple and could provide income growth oppor - tries. From the early 1970s to the mid 1990s, con - tuni ties for many rural poor. It is not inap pro pri ate sump tion of meat in devel op ing coun tries grew by to use the term “Livestock Revo lu tion” to describe 70 million met ric tons, whereas con sump tion in the course of these events in world agricul ture over devel oped countries grew by only 26 million metric the next 20 years. Like the well- known Green tons (Table 1).2 In value and caloric terms, meat con- Revo lu tion, the label is a sim ple and con ven ient sump tion in devel op ing countries increased by more expression that sum ma rizes a complex series of than three times the increases in devel oped countries. interre lated processes and outcomes in produc tion, Milk con sump tion in the devel op ing world increased consump tion, and economic growth. As in the case by more than twice as much as milk consumption in of cereals, the stakes for the poor in develop in g the devel oped world in terms of quantity, money coun tries are enormous. And not unlike the Green value, and calo ries. Revo lu tion, the “revolu tion ary” aspect comes Even more reveal ing is the compari son be - from the partici pa tion of develop ing countries on a tween developing-country increases in meat, milk, large scale in trans for ma tions that had previ ously and fish consump tion dur ing the 1971–95 pe riod occurred mostly in the tem per ate zones of devel - with the increase in ce real consump tion (Table 1). oped countries. But the two revolu tions differ in The pe riod spans the well known Green Revolu tion, one fun da men tal respect: the Green Revolu tio n when seed-fertilizer in no va tions in ce real pro duc - was supply-driven, whereas the Livestock Revolu - tion dramati cally increased wheat, rice, and maize tion is driven by demand. out put in devel op ing coun tries, mak ing more food The Live stock Revolu tion will stretch the ca pac - avail able and increas ing farm incomes. But dur ing ity of ex ist ing produc tion and distri bu tion systems the same pe riod there was also a dra matic, if often and ex ac er bate envi ron mental and pub lic health over looked, rise in consump tion of animal-origin prob lems. Gov ern ments and indus try must pre pare food prod ucts in devel op ing countries. On a quan - for this con tinu ing transfor ma tion with long-ru n tity basis, the addi tional meat, milk, and fish con - policies and in vest ments that will satisfy consumer sumed be tween 1971 and 1995 in de vel op ing coun - de mand, improve nu tri tion, direct in come growth tries was two- thirds as im por tant as the increase in oppor tu ni ties to those who need them most, and al le - wheat, rice, and maize con sumed (Table 1). The vi ate envi ron mental and pub lic health stress. Green Revo lu tion pro vided many more calo ries

1 Fish con sump tion and produc tion are also un der go ing revolu tion ary changes, but this is o utside the scope of the cur rent paper. In - ter ested read ers are re ferred to Wil liams (1996) and Westlund (1995). 2 All tons are met ric tons in this report .

1 2

Table 1— In crease in food consump tion of meat, milk, fish, and major cere als, 1971–95

Caloric value of Con sump tion in crease Value of consump tion in crease a con sump tion increas e Com mod ity De vel oped De vel op ing De vel oped Devel op ing Devel ope d Devel op ing

(mil lion met ric tons) (bil lion 1990 US$) (tril lion kilo calo ries) Meatb 26 70 37 124 38 172 Milk 50 105 14 29 22 64 Fish c 5 34 27 68 4 20 Major cere al sd 25 335 3 65 82 1,064

Sources: The changes in quan ti ties and in calories are from FAO 1998. Money values are com puted for the disag gre gated com modi ties (shown in the notes below) us ing the 1990–92 av er age price. The com mod ity prices used for beef, sheep an d goat meat, pork, poultry, wheat, rice, and maize are de tailed in Table 28. Disag gre gated fish prices are 1990–92 av er age im port un it val ues calcu lated from FAO (1998) im port data. Developed-country im port unit val ues are used for developed- country con sump tion and dev eloping- country im port unit val ues are used for developing- country con sump tion. Notes: Cal cu la tions repre sent aggre gate changes be tween three-year av er ages centered on 1971 a nd 1995. aCalcu lated us ing 1990 world prices ex pressed in con stant, av er age 1990–92 US$. bBeef, sheep and goat meat, pork, and poul try. cMarine and freshwa ter fin fish, cephalo pods, crusta ceans, mol luscs, and other marine fish. dWheat, rice, and maize used di rectly as hu man food. than the coin cid ing increase in meat consump tion , tures, and grasses indigesti ble by humans) to one but the addi tional meat con sumed was worth almost that actively seeks new resources for the produc - three times the increase in cereal consump tion at tion of animal food products (Ste in feld, de Haan, constant world prices. and Blackburn 1997). Further more, if the consump tion pat terns in The recent rapid expansion of live stock food devel oped coun tries are an indica tion of where pro duc tion in develop ing countries resulted pri - develop ing coun tries are going, future growth in marily from increased numbers of ani mals rather cereal consump tion as food is likely to be much than higher car cass weight per animal. In develop - smaller than that in meat. Dur ing 1971–95 addi- ing countries this has contrib uted to large concen - tional consump tion of meat, milk, and fish in the trations of ani mals in urban envi ron ments where devel oped coun tries was much larger than that of the regu la tory framework govern ing live stock cereal in terms of weight and value. In develop ing pro duc tion is weak (for example, Addis Ababa, countries many people will soon be reaching satia - Beijing, Lima, and Mumbai). Larger concen tra - tion in their consump tion of cere als, while meat and tions of ani mals have also led to degra da tion of milk consump tion is likely to con tinue to grow even rural grazing areas and the clearing of for est. more robustly into the next cen tury. Growing con cen tra tions of ani mals and people in Not sur pris ingly, major trans for ma tions of the the major cit ies of develop ing countries also lead mag ni tude of the Live stock Revolu tion are not to rapid increases in the inci dence of zoonotic dis - without prob lems. Although domesti cated ani - eases, such as salmo nella, E-coli, and avian flu, mals have been a source of human food, clothing , which can only be dealt with through enforcement tools, transpor ta tion, and farm power since pre- of zoning and health regu la tions. historic times, the current rapid changes in Other pub lic health is sues raised by the Live - demand for animal foods in develop ing countries stock Revo lu tion are also of major im por tance. The are putting unprecedented stress on the resources inten si fi ca tion of live stock pro duc tion is lead ing in used in livestock pro duc tion. The com bi na tion of many parts of the world to a build- up of pes ti cides higher demand, more people, and less space is and an ti bi ot ics in the food chain. Further more, as rap idly leading to a global trans for ma tion of the the scale of output increases, espe cially in the trop - livestock sec tor, from one that mobilizes surplu s ics, food safety risks from micro bial contami na tion and waste resources (back yard slops, remote pas - are becom ing more preva lent. 3

Over con sump tion of animal food prod ucts raises almost en tirely made up of starchy staples. The an other con cern. A growing conscious ness of the addition of milk and meat pro vides protein, cal - dangers of large amounts of saturated ani mal fats in cium, vi ta mins, and other nutri ents that go lacking diets exists in most de vel oped countries. Some ex - in diets that are exclu sively made up of sta ples perts have con cluded that policies should prevent a such as cere als . similar over- consumption in devel op ing countries by Besides pro vid ing food, the driv ing force be - discour ag ing pub lic invest ment in livestock produc - hind increased live stock pro duc tion, live stock have tion (Brown and Kane 1994; Geissler forthcom ing ; other valuable uses. Live stock re main the most im - Goodland 1997; Pimen tel 1997). por tant if not the sole form of nonhu man power Increas ing live stock consump tion may also af - avail able to poor farm ers in much of the de vel op ing fect cereal prices. Because ru mi nant live stock such world. The poor, in par ticu lar, use fer til izer from as cattle, sheep, and goats con sume grain, and livestock opera tions, espe cially when rising pe tro - monogastric live stock such as pigs and poul try de - leum prices make chemical fer til iz ers unaf ford able. pend on grain in the in dus trial pro duc tion sys tems Live stock also store value and pro vide in sur ance of de vel oped countries, some ana lysts argue that the for people who have no other fi nan cial market s high de mand for live stock prod ucts in devel ope d avail able to them. Skins, wool, oil, and other re- countries and rap idly increas ing de mand and pro - sources are used as in puts in other in dus tries. duc tion in devel op ing coun tries de plete the grain This report will ex am ine in de tail the in ter- avail able for di rect consump tion by people . relationships over time between supply and de - Live stock pro duc tion and consump tion have mand for live stock and feed grain, using IFPRI’s pro po nents as well. Live stock pro duc tion is an IMPACT model. 3 It will in ves ti gate the plausi bil - especially im por tant source of income for the rural ity of the projected demand increases for livestock poor in de vel op ing coun tries. It enables poor and products and the impli ca tions of these in creases landless farmers to earn income using pub lic, for world markets in feed, milk, and meat. The common-property resources such as open range - paper will argue that world grain markets currently land. Live stock con sume many crop by-product s have suffi cient ca pac ity to han dle the addi tional that would other wise become waste, they often can de mand for feed coming from in creas ing livestock be raised on land that has no other sustain able ag ri - produc tion, even under a va ri ety of dif fer ent sce - cultural use, and they can em ploy labor dur ing pe ri - nar ios for techno logi cal devel op ment and global ods of slack in other ag ri cul tural ac tivi ties. Poor eco nomic per form ance. women in particu lar often rely on the cash in come The paper will argue fur ther that the structural from a dairy cow or a few chickens kept in the shift in developing-country diets toward ani mal household. As live stock consump tion increase s pro teins is a given that must be dealt with. It will re - there is con sid er able inter est in how the poor can view the evidence on the impact of live stock prod- retain their market share of live stock production. ucts on nutri tion in devel op ing coun tries and on the Livestock products are an ap peal ing and con - food de mand and income growth of the poor. The ven ient nu tri ent source. Protein and mi cro nu tri ent in dus tri ali za tion of live stock pro duc tion in devel - defi cien cies remain wide spread in devel op in g op ing coun tries can harm the wel fare of the poor if coun tries because peo ple sub sist on diets that are other poli cies arti fi cially re duce the cost of in dus -

3 The model was de vel oped by Rosegrant and col leagues (Rosegrant, Agcaoili-Sombilla, and Perez 1 995; Rosegrant et al. 1997; Rosegrant, Leach, and Ger pa cio 1998; and Rosegrant and Ringler 1998). It is global in na ture and bal ances sup ply and demand within ag ri cul ture with market- clearing prices for ma jor ag ri cul tural commodi ties, incl u d ing livestock prod ucts and feed. Start ing with exoge nously speci fied trends in na tional in comes for 37 country groups, the model traces food de mand, feed de mand, and sup - ply levels for 18 commodi ties, iter at ing to market- clearing prices for ma jor com modi ties an nu ally through 2020. The re sults are based on a large number of pa rame ter as sump tions taken from the litera ture, includ ing as su mp tions about the open ness to trade. The model is use ful for il lus trat ing how demand-led shocks in Asia, for exam ple, work them selves out in markets around the world. It also illus trates that in systems of in ter linked global mar kets for livestock prod ucts and feed, the net ef fect of price- mediated pol icy inter ven tions can be quite dif fer ent from what was en vis aged. 4 trial turn- key opera tions and other wise frustrate the tici pa tory insti tu tions of collec tive ac tion for small- partici pa tion of small farmers. The paper will sug - scale farmers that allow them to be ver ti cally inte - gest that un der stand ing the oppor tu ni ties and dan - grated with live stock proces sors and input sup- gers of the Live stock Revo lu tion is critical to de - pliers; (3) cre at ing the envi ron ment in which farm - sign ing poli cies that promote the in cor po ra tion of ers will increase in vest ment in ways to improve the rural poor into eco nomi cally and en vi ron men - produc tiv ity in the live stock sector; and (4) promot - tally sustain able growth pat terns. ing effec tive regu la tory insti tu tions to deal with the The rapid increase in de mand for live stock threat of en vi ron mental and health crises stemmin g prod ucts in devel op ing coun tries pres ents cru cially from live stock. impor tant policy dilem mas that must be resolved Techno logi cal prog ress in the pro duc tion, proc- for the well- being of both rural and urban people in ess ing, and dis tri bu tion of live stock prod ucts will devel op ing coun tries. These dilem mas involve be central to the posi tive outcome of the Live stock complex en vi ron mental and public health is sues in Revolu tion. Rapid advances in feed im prove ment the con text of weak regu la tory en vi ron ments. and ge netic and re pro duc tive technolo gies offer Taken together, the many oppor tu ni ties and dangers scope for overcom ing many of the tech ni cal prob - of the Live stock Revo lu tion sug gest that it would be lems posed by increased live stock pro duc tion. In sti - foolish for devel op ing coun tries to adopt a lais sez tutional and regu la tory devel op ment will also be faire policy for live stock de vel op ment. Many spe - criti cal to secur ing de sir able en vi ron mental and cific recom men da tions for con crete ac tion are given pub lic health outcomes. In sum, the demand- driven in chapters ahead. The over all focus of the paper, Live stock Revo lu tion is one of the larg est struc tural however, is on the four broad pillars on which to shifts to ever af fect food mar kets in de vel op ing base a de sir able live stock devel op ment strat egy for countries and how it is han dled is crucial for fu ture devel op ing coun tries. These are (1) re mov ing pol - growth prospects in devel op ing coun try agri cul ture, icy distor tions that arti fi cially mag nify econo mies for food se cu rity and the liveli hoods of the rural of scale in live stock produc tion; (2) build ing par - poor, and for en vi ron mental sustain abil ity . 2. Re cent Transfor ma tion of Live stock Food Deman d

Per Cap ita Consump tio n Fig ure 1— Per capita consumption of meat and milk, developing and developed Progres sive economic dif fer en tia tion betwee n countries, 1983 and 1993 coun tries over the last few cen tu ries, has led to a situa tion where people in de vel oped countries typi cally consume three to four times the meat and five to six times the milk as do those in de vel op ing coun tries (Figure 1). But this pattern is chang ing. Peo ple in devel op ing countries have increase d their con sump tion of ani mal food products over the past 20 years, and the factors driving those in - creases are ro bust and unlikely to subside in the near future .4 Between 1983 (av er age of 1982–84) and 1993 (av er age of 1992–94) per cap ita an nual meat con sump tion rose from 14 to 21 kilogram s and milk con sump tion grew from 35 to 40 kilo - grams. Dur ing the same period per cap ita consump tion of meat in de vel oped coun tries rose only 2 kilo grams and per cap ita milk con sump tion fell. At the regional level, Asia witnessed the most dra matic increases in per capita consump tion of ani mal food prod ucts. In China per cap ita con sump - tion of meat and milk dou bled between 1983 and 1993 (Table 2). Per cap ita meat consump tion also increased in Other East Asia, South east Asia, and The rela tive im por tance of animal food prod - Latin America. Per cap ita milk con sump tion ucts in the diets of people in de vel op ing countrie s increased in India, Other South Asia, and Latin rose as well. Con sum ers obtained a greater share of America. In Sub-Saharan Africa and West Asia and calo ries and protein from ani mal food prod ucts in North Africa (WANA) per cap ita con sump tion 1993 than in 1983 (Table 3). Through out Asia the of meat and milk stagnated or declined (see the share of calories and pro tein coming from ani mal Appen dix for the regional clas si fi ca tion of coun- food prod ucts in creased, almost dou bling in China, tries used in this paper). indi cat ing that many con sum ers are increas ing con -

4 Throughout this pa per, “food” is used to distin guish di rect food con sump tion by humans fro m uses of ani mal products as feed, fuel, cos met ics, or cover ings. Sta tis ti cal data in this paper are taken from the FAO Sta tis ti c al Da ta base (FAO 1997, 1998), un less oth er - wise identi fied, and the world is clas si fied into nine countries or coun try ag gre gates, wit h de tails given in the Appen dix. The years 1983 and 1993 in all tables and figures re fer to three-year moving av er ages cen tered on the years shown.

5 6

Table 2— Per cap ita meat and milk consump tion Fig ure 2— The relationship between meat by re gion, 1983 and 1993 consumption and income

Meat Milk Region 1983 1993 1983 1993 (kilograms) China 16 33 3 7 Other East Asia 22 44 15 16 In dia 4 4 46 58 Other South Asia 6 7 47 58 South east Asia 11 15 10 11 Latin Americ a 40 46 93 100 WANA 20 20 86 62 Sub-Saharan Af rica 10 9 32 23 De vel op ing world 14 21 35 40 Devel oped world 74 76 195 192 United States 107 118 237 253 World 30 34 76 75

Source: FAO 1997. Notes: Con sump tion refers to di rect use as food, measured as un - Note:Each dot is an ob ser va tion for 1 of 78 de vel op ing and devel oped cooked weight, bone in. Meat includes beef, pork, mut ton, coun tries exam ined. The solid line is a sta tis ti cally signifi can t goat, and poultry. Each number is a three-year mov ing av er - trend. age centered on the two years listed. Milk is cow and buffalo milk and milk prod ucts in liq uid milk equiva lents.WANA is West Asia and North Af rica. cap ita income and per cap ita meat consump tion . Within this trend certain coun tries dif fer en ti ate sumption of ani mal food prod ucts more rapidl y them selves for cul tural or other reasons. China, for than they are of other foods such as ce re als. ex am ple, lies above the trend, reflect ing the impor - But there remains a great dis par ity be tween the tance of pork in Chinese diets, and India lies below per cap ita animal food con sumed in de vel oped and the trend because of re lig ious prefer ences against devel op ing countries. Na tional income is a critical meat. At higher incomes, per cap ita consump tion of de ter mi nant of this dispar ity. Figure 2 dis plays the meat lev els off because people reach satu ra tion. posi tive, curved re la tion ship between na tional per This explains why de vel oped coun tries have had much smaller increases in per capita meat and milk Table 3— Per cent of calo ries and pro tein from consump tion over the past 20 years compared to ani mal prod ucts, 1983 and 1993 developing coun tries. Countries at lower income lev els are far from Calo ries from Protein from ani mal prod ucts ani mal prod ucts reach ing the meat consump tion sa tia tion point de- Regio n 1983 1993 1983 1993 spite recent increases. In the first half of the 1990s,

(per cent) people in de vel oped coun tries con sumed 76 kilo- China 8 15 14 28 grams of meat per cap ita per year as food, with Other East Asia 11 15 29 38 higher amounts in the United States and lower In dia 6 7 14 15 amounts in some of the European countrie s Other South Asia 7 9 19 22 South east Asia 6 8 23 25 (Table 2). Milk consump tion in de vel oped coun - Latin Americ a 17 18 42 46 tries was 192 kilo grams per capita. People in devel - WANA 11 9 25 22 op ing coun tries con sumed on av er age 21 kilograms Sub-Saharan Af rica 7 7 23 20 De vel op ing world 9 11 21 26 of meat and 40 kilo grams of milk. Devel oped world 28 27 57 56 In Latin America, people con sume 46 kilo - World 15 16 34 36 grams of meat and 100 kilo grams of milk per capita, Source: FAO 1997. lev els that are much higher than else where in the Notes: Each number is a three-year mov ing av er age centered on the devel op ing world, though still about half the two years listed. Ani mal prod ucts, us ing the FAO defi ni tion, developed-country aver age. Per capita meat con- in clude meat, dairy, egg, and freshwa ter and marine ani mal prod ucts. WANA is West Asia and North Africa . sumption in Other East Asia (44 kilo grams per cap- 7 ita) does come close to the Latin American aver age Table 4— Past popula tion, urban popu la tion, and exceeds the Chi nese average. Sub-Saharan and GNP per capita growth rates Africa has some of the low est per cap ita con sump - Urban GNP tion lev els: 9 kilo grams of meat and 23 kilo grams of Popula tio n popu lation per cap ita milk per capita per year. Regio n 1970–95 1970–95 1980–95 The share of calories and protein com ing from (per cent change per year) meat is also much lower in devel op ing countrie s China 1.6 3.8 8.6 than in de vel oped ones (Table 3). In de vel oped In dia 2.1 3.3 3.2 countries people ob tain an av er age of 27 percent of Other East Asia 1.6 3.0 n.a. their calories and 56 percent of their pro tein from Southeast Asia 2.1 4.0 4.3 Latin Americ a 2.1 3.0 – 0.4 ani mal food products. The aver ages for devel op ing Sub-Saharan Af rica 2.9 5.0 –1.3 countries are 11 percent and 26 per cent, respec - Devel op ing world 2.1 3.8 2.1 tively. People in Sub-Saharan Af rica, WANA, Devel oped world 0.7 1.1 1.7 South east Asia, Other South Asia, and India get a World 1.7 2.6 0.9 third or less than a third as many calories and half Source: UNDP 1998. as much protein from ani mal prod ucts as people in Note: n.a. in di cates not available. De vel oped world is the UNDP de vel oped coun tries. industrial coun tries. Data for WANA were unavail able . These low consump tion lev els give an indi ca - tion of how far ani mal food product consump tion in Prices of major meat and ce real food com - de vel op ing countries could grow. The Live stock modities have trended down ward over the past Revo lu tion of the past 20 years will begin to look 20 years, making food more af ford able to con sum - mod est in compari son to the one to come if the fac - ers of all incomes (Figure 3). Real cereal prices fell tors that promote meat and milk consump tion exert 38– 46 per cent (depend ing on the grain in ques - their full in flu ence. tion) be tween the early 1980s and early 1990s, while deflated liquid milk prices fell 37 percen t and real meat prices fell 23–35 per cent. Although De ter mi nants of Changes in ce real prices fell faster than meat and milk prices, Per Cap ita Consump tio n many consum ers have begun to diver sify their diets into meat and milk because they are nearly The growth rate of per cap ita con sump tion of anima l sati ated with cere als. Some have even reduced food prod ucts is de ter mined by economic fac tors their con sump tion of cere als . such as incomes and prices and lifestyle changes that The most im por tant lifestyle change oc cur ring cause people’s dietary pat terns to evolve in qualita - in recent years is ur bani za tion. Con sum ers in urban tive ways. Per capita con sump tion in creased in the areas are more likely to diver sify their diets into re gions where incomes grew rapidly during the meat and milk (Huang and Bouis 1996; An der son 1980–95 period. For devel op ing countries as a et al. 1997). Urban con sum ers have greater food whole, GNP per cap ita grew at 2.1 per cent per year choices and more diverse die tary and cul tural in flu - (Ta ble 4). In China, which had the most dramatic in - ences than those typically found in rural areas. creases in per cap ita meat and milk con sump tion, Urban con sum ers also often pre fer foods that offer GNP per cap ita grew at the ex traor di nary rate of vari ety and conven ience rather than maxi mum 8.6 per cent per year. India and Southeast Asia caloric con tent. also had high in come growth rates, fu el ing increases Urban popula tion growth has been substan tial in per capita animal food product con sump tion. throughout the de vel op ing world in re cent years Latin American in come growth was about zero (Table 4). Between 1970 and 1995 cities in Asia

(– 0.4 per cent), but the re gion still managed a slight grew 3 per cent per year and higher. The high est increase in per capita meat and milk con sump tion. rate of urban growth, 5 per cent, occurred in Af rica. Sub-Saharan Af ri ca’s per capita GNP fell sig nifi - The aver age for all de vel op ing coun tries was cantly, ex plain ing the region’s drop in per cap ita 3.8 percent, more than three times the developed- con sump tion of meat and milk during the pe riod. country rate. 8

Fig ure 3— Trends in the prices of major cereal Near East, Asia, and Sub-Saharan Af rica. South and meat commodities, 1970–72 to Asia has lower lev els of meat consump tion than low 1994–96 income alone would sug gest because of cul tural and re lig ious reasons. Growing health conscious ness in de vel oped coun tries has increased consump tion of lean meats such as poul try and limited growth in the of consump tion of red meat. These prefer ences are reflected in the ag gre gate changes in per cap ita con - sumption be tween 1973 and 1993 (Table 5).

Total Consump tio n The im por tance of even small increases in per cap - ita consump tion is compounded by rap idly in creas - ing popula tions in many de vel op ing regions. On average, popula tion in devel op ing coun tries grew by 2.1 percent per year between 1970 and 1995 (Table 4). The popula tion in Sub-Saharan Af rica grew the most—almost 3 percent per year dur ing the pe riod. Rapid popula tion growth cou pled with increased per cap ita consump tion resulted in dra- Sources: Past data are from ERS 1997, IMF 1997, USDA 1997, and matic increases in the total consump tion of anima l World Bank 1993. World Bank projec tions and the Manufac - food products through out the de vel op ing world turing Unit Value index used for ex press ing values in con - stant 1990 US dollars are from World Bank 1997. (Table 6). For de vel op ing coun tries as a whole, total Notes: Wheat is U.S. no. 1, hard red win ter, or di nary pro tein, ex port meat consump tion grew 5.4 percent per year and price deliv ered at Gulf ports for ship ment within 30 days. total milk consump tion grew 3.1 per cent. The com - Rice is Thai 5 percent broken, WR, milled in dica tive survey price, gov ern ment standard, f.o.b. Bangkok. Maize is U.S. para ble fig ures for de vel oped countries were no. 2, yel low, f.o.b. U.S. Gulf ports. Soy beans are U.S. c.i.f. 1.0 percent for meat and 0.5 percent for milk. China Rot ter dam. Soymeal is any ori gin, Ar gen tina 45–46 per cent ex pe ri enced an extremely high meat con sump tion ex trac tion, c.i.f. Rot ter dam, prior to 1990, U.S. 44 per cent. Fishmeal is any origin, 64–65 percent, c.i.f. Ham burg, n.f.s. growth rate of 8.6 per cent, a value that is disputed. Beef is Aus tra lian/New Zea land, cow fore quar ters, froze n China’s role as the fast est growing market for live - bone less, 85 per cent chemical lean, c.i.f. U.S. port (East stock prod ucts in the world is not in dis pute, but its Coast), ex dock. Pork is European Comu nity pork, slaugh ter growth rate may not have been so far ahead of the whole sale price. Poul try is broil ers, twelve- city com po site whole sale price, read to cook, de liv ered. Lamb is New Zea - next fast est growing region, Other East Asia. The land, fro zen whole car casses, wholesale price, Smithfiel d issue is contro ver sial because China con sti tutes a mar ket, London. Milk is U.S. whole milk sold to plants and large com po nent of world de mand. dealers, U.S. Depart ment of Ag ri cul ture. The food consump tion figures used in this re - port are from the FAO statis ti cal data base (FAO 1997, 1998). For China, as for most coun tries, the In ad di tion to income growth, price changes, numbers are taken from food balance sheets pre - and ur bani za tion, cul tural differ ences have played pared from national sources and are based primar ily an im por tant role in consump tion patterns. Poul try on esti mates of pro duc tion and net trade to de rive meat and eggs are the most accept able live stock esti mates of con sump tion. Re cently, the use of this com modi ties throughout the world. Lactose- method ol ogy for es ti mat ing live stock produc tion intolerance, found par ticu larly in East Asia, has figures in China in the 1990s—but not in the limited milk con sump tion. Pork, while par ticu larly 1980s—has been challenged (Ke 1997). valued by East Asians and people of Europea n Al though there is some uncer tainty here, inde - descent, is excluded from the diet of a large share of pend ent esti mates of consump tion based on house- the world’s popu la tion, espe cially Moslems in the hold surveys and feed use sug gest that meat con - 9

Table 5— Annual per cap ita consump tion of selected livestock food products and percent of total calo ries consumed from each product, 1973 and 1993

De vel oped countrie s Devel op ing coun tries Commod it y 1973 1993 1973 1993 (kilograms) (percent) (kilograms) (percent) (kilograms) (per cent ) (kilo grams) (per cent )

Beef 26 3 25 3 4 1 5 1 Mut ton and goat 3 1 3 1 1 0 1 0 Pork 26 4 29 5 4 2 9 3 Poul try 11 1 20 2 2 0 5 1 Eggs 13 2 13 2 2 0 5 1 Milk and product s exclud ing butter 188 9 195 9 29 2 40 3 Four meats 67 10 78 11 11 3 21 6 Four meats, eggs, and milk 268 20 285 21 42 6 65 9

Source: FAO 1997. Notes: Four meats includes beef, pork, mut ton and goat, and poultry. Values are three-year mov ing av er ages centered on the two years shown; percent ages are calcu lated from three-year mov ing aver ages. Milk is cow and buffalo milk a nd milk prod ucts in liq uid milk equiva lents. Food is used to dis tin guish direct food con sump tion by hu mans from uses of ani mal prod ucts as feed, fuel, cos met ics, or cov er ings. sumption in the early 1990s in China proba bly ran early 1980s to the early 1990s would be 6.3 per cent, closer to 30 mil lion metric tons (25 kilo grams per closer to the 5.4 percent per year observed in the capita) than the 38 mil lion metric tons (33 kilo- rest of Asia. grams per capita) given in the ta bles in this re port Whether the true growth rate of meat consump - (Ke 1997). If the lower figure is cor rect, the ac tual tion in China was exceed ingly high (6.3 percent per growth rate of meat consump tion in China from the year) or astro nomi cally high (8.3 percent per year),

Table 6— Consump tion of meat and milk by region, 1982–94

Annual growth rate of to tal meat con sump tion To tal meat con sump tion To tal milk con sump tion Re gion 1982–94 1983 1993 1983 1993

(per cent) (mil lion met ric tons) (mil lion met ric tons) China a 8.6 16 38 3 7 Other East Asia 5.8 1 3 1 2 In dia 3.6 3 4 34 52 Other South Asia 4.8 1 2 11 17 Southeast Asia 5.6 4 7 4 5 Latin Americ a 3.3 15 21 35 46 WANA 2.4 5 6 21 23 Sub- Saharan Af rica 2.2 4 5 12 14 Devel op ing world 5.4 50 88 122 168 De vel oped world 1.0 88 97 233 245 World 2.9 139 184 355 412

Sources: Annual growth rate of to tal meat con sump tion for 1982–94 is the growth rate from regres sions fit ted to FAO an nual data (FAO 1998). To tal milk and meat con sump tion for 1983 and 1993 are three-year mov ing av er ages cal cu late d from FAO 1998. Notes: Con sump tion refers to direct use as food, meas ured as uncooked weight, bone in. Meat in clude s beef, pork, mutton, goat, and poultry. Milk is milk and milk prod ucts in liq uid milk equiva lents. Met ric tons are three-year mov ing av er ages cen tered on the two years shown. Milk is cow and buffalo milk and milk prod ucts in liq uid milk equiva lents. WANA is Western Asia and No rth Africa . aSee text for qualifi ca tion on China. A lower es ti mate of 6.3 per cent per year growth, closer to the 5.4 percent observed in the rest of Asia, may be more ac cu rate. This would mean a 1993 to tal meat con sump tion of 30 mil lion met ric tons. 10 the total amount in conten tion is less than 5 per cent Quan ti fy ing the Effects of of es ti mated an nual world meat consump tion in the early 1990s. It should also be noted that the con tro - Growth Factor s versy does not include the dis tri bu tion of con sump - Quan ti fy ing the effects of indi vid ual forces that are tion among meats in China, nor does it involv e driv ing real con sump tion requires a mod el ing international trade in meat, because the downwar d approach to statis ti cal esti ma tion that is capable of re vi sion in the pro duc tion figures is matched by a sorting out the simulta ne ous influences of a host of corre spond ing down ward re vi sion in the con sump - deter mi nants in order to isolate the con tri bu tion of tion numbers. each element. Research ers typi cally use a mul ti ple- Ac cord ing to FAO, the total quan tity of meat regres sion econometric approach, although the con sumed worldwide rose by 45 mil lion metric tons degree of com plex ity in their mod els var ies greatly. be tween 1983 and 1993 (Table 6). Total milk con - The end objective is the esti ma tion of robust elas - sumption rose by 57 mil lion metric tons in liq uid ticities that measure the effect on consump tion of a milk equiva lents. In 1983 de vel op ing countrie s 1 percent increase in the determi nant in ques tion. con sumed 36 percent of all meat and 34 percent These esti mates often are obtained from a cross- of all milk con sumed worldwide. By 1993 those section of households in a particu lar region at a par - percent ages had risen to 48 percent and 41 percent , ticular time, yielding elastici ties that are usu ally respec tively . quite satis fac tory for the time period and region The breakdown of the growth rates of consump - con cerned, but which are too specific to use across tion of par ticu lar commodi ties (Table 7) shows that countries or over long time periods. in de vel oped countries total con sump tion grew Elastici ties from national data over long time slowly for all commodi ties ex cept poul try. In the peri ods are rarely es ti mated because of the dif fi - developing coun tries poul try led the field as well culty in gather ing data sets that sat isfy the eco nomic with 7.6 per cent growth in con sump tion per year. and econometric as sump tions of the un der ly ing de - Beef and milk grew at about 3 per cent, and pork con - mand model. Such es ti ma tion can, how ever, pro - sump tion grew at 6.2 per cent. vide a better fore cast of the evo lu tion of na tional

Table 7— Trends in the food consump tion of vari ous live stock prod ucts, 1982–94

Annual growth rate of total To tal con sump tion Per cap ita con sump tion consump tion Re gion/product 1982–94 1983 1993 1983 1993

(per cent) (mil lion met ric tons) (kilograms) Devel oped world

Beef – 0 .0* 32 32 27 25 Pork 0.6 34 36 29 28 Poul try 3.1 19 26 16 20 Meat 1.0 88 97 74 76 Milk 0.5 233 245 195 192 De vel op ing world Beef 3.2 16 22 5 5 Pork 6.2 20 38 6 9 Poul try 7.6 10 21 3 5 Meat 5.4 50 88 14 21 Milk 3.1 122 168 35 40

Sources: Annual growth rate of to tal con sump tion 1982–94 is the growth rate from regres sions fitted t o FAO an nual data (FAO 1998). To tal and per cap ita con sump tion for 1983 and 1993 are calcu lated from FAO 1998. Notes: Con sump tion refers to direct use as food, meas ured as uncooked weight, bone in. Meat in clude s beef, pork, mutton, goat, and poultry. Milk is cow and buffalo milk and milk prod ucts in liq uid milk equiva lents. Met ric tons and kilogr a ms are three-year mov ing av er ages centere d on the two years shown. WANA is Western Asia and North Africa. *Not sig nifi cantly dif fer ent from zero at the 10 percent level. 11 consump tion patterns over time. De spite the dif fi - Schroe der, Bar kley, and Schroe der did not re - culties, Schroeder, Bar kley, and Schroe der (1995) port price elastici ties or the ef fects of other struc - es ti mated the effects of na tional per cap ita incom e tural changes on per capita con sump tion. Del gado growth on national per cap ita con sump tion, using and Courbois (1998) es ti mated ex pen di ture, price, annual data from 32 coun tries for 1975–90. The and urbani za tion elastici ties based on data from authors found that the larg est ef fect of a US$1 in - 64 devel op ing coun tries for 1970–95. They used a crease in income on meat consump tion occurred in sys tem of equations that sorted out rela tive price countries with the low est lev els of na tional incom e effects among ani mal prod ucts and that con trolled and meat con sump tion. As countries got richer the for many cul tural, geographic, physical, and eco- im pact of an increase in income on meat con sump - nomic differ ences be tween coun tries. tion got weaker. The result ing expen di ture elastici ties (Table 8) Schroe der, Barkley, and Schroe der (1995) found es ti mate the percent age increase in the weight of that for coun tries with annual per cap ita incomes in beef, pork and mutton, poul try, or milk con sumed the neighbor hood of US$1,000 (at 1985 prices), due to a 1 percent increase in total expen di ture on each 1 per cent in crease in per cap ita in come would all animal food prod ucts in the de vel op ing coun tries increase con sump tion of pork by 1 per cent, poul try in the sam ple. Thus the whole- sample ex pen di ture by nearly 2 per cent, beef by more than 2 percent, elas tic ity of 1.36 for milk sug gests that the rela tive and lamb by more than 3 per cent. At per cap ita in - share of milk in total ani mal product consump tion come lev els above US$10,000, a 1 per cent in - in creases as real expen di ture on ani mal food prod - crease in in come would in crease per cap ita con - ucts increases across coun tries and over time, once sumption of any of the com modi ties by approx i - the ef fects of rela tive prices, ur bani za tion, and other mately 1 per cent or less. These results in di cate factors are taken into account. The 0.27 coef fi cient first, that an in crease in in come in a richer country for poul try sug gests that its share decreases with a will have a substan tially smaller im pact on meat 1 percent increase in total expen di ture on all anima l con sump tion than the same in crease will have in food prod ucts, poorer countries. Sec ond, the results indi cate that A com pari son of the poorest third of coun tries in coun tries with low but ris ing per capita in comes, to the richest third at subsam ple means suggest s per cap ita con sump tion of most meat commodi tie s that the prefer ence for addi tional milk and beef de - is likely to grow faster than growth of per cap ita creases mar gin ally when moving from poorer to in come. richer de vel op ing countries. Prefer ence for poultry

Table 8— Demand elas tici ties for major food prod ucts of ani mal origin from a cross- country, systems es ti ma tion, 1970–95, de vel op ing re gions

a Expen di ture elas tic ity Urban Poores t b third Whole Riches t b third Own price population share Commod it y of countries sample of countries elas tic ity elasticity

Beef 0.72 0.65 0.57 –0.14 –0.20 Pork and mutton 0.96 1.10 1.30 –0.39 0.46 Poultry 0.28 0.27 0.26 –0.17 0.38 Milk 1.43 1.36 1.26 –0.86 –0.17

Source: Del gado and Cour bois 1998. Notes: These pa rame ters were esti mated as a sys tem (with other explana tory vari ables and ex clu si ons not shown) for 64 countries us ing annua l data. N=1,143 and McElroy’s multie qua tion R 2 (Judge et al. 1985, 477) was 0.86. All coef fi cients were sta tis ti cally signifi cant at a 10 per - cent level or bet ter. aEx pen di ture is the to tal expen di ture on ani mal food prod ucts in cluded in the study. Ex pe n di ture elastici ties are calcu lated at the subsample mean, to proxy for income elastici ties for spe cific sub groups. bMean per cap ita gross do mes tic product dur ing the 1970–95 pe riod was used to clas sify all co un tries into one of three groups: poor est (<$800), mid - dle ($800–$3,000), and rich est (>$3,000). The sam ple was di vided into thirds, with each third h av ing the same number of coun tries. 12 is remarka bly sta ble across wealth groups of coun - Most impor tant among these was the measure of tries. Prefer ence for pork and mutton rises with in- urbanization. The ur bani za tion elas tici ties sug gest creas ing in come.5 The last re sult may hide changes that as the percent age of the popula tion living in in the quality of meat products con sumed when cities increased, so did the impor tance of pork, mov ing to richer countries, es pe cially given the mut ton, and poultry in ani mal food prod uct con - large variation in quality in pork/mutton meats. sump tion, while the impor tance of beef and milk Fur ther more, while these elas tici ties are useful for de creased. indi cat ing the rela tive re spon sive ness of con sump - The two key messages of the data and analy sis tion of differ ent products to in come, demand for sum ma rized in this chapter are that ani mal food indi vid ual products in indi vid ual coun tries may be product demand has increased dramati cally in the more or less re spon sive than these multi coun tr y past and that it is very likely to increase in the esti mates in di cate. Demand would de pend in large future. The same fac tors that drove the enor mous part on whether the coun try in question ex hib ited increases in total meat consump tion are expected to more or less income-responsiveness for ani mal exert their influence into the next cen tury. Popula - products as a group. tion is pro jected to grow more modestly but still at The own-price elastici ties in Table 8 meas ure an average of 1.5 percent per year in develop in g the change in consump tion of various ani mal prod - countries (UNDP 1998). ucts in re sponse to rela tive price changes within the With that rate of popu la tion growth, even with - group of ani mal products. As expected, price rises out a change in per capita consump tion, de mand for for a given com mod ity are as so ci ated with de - ani mal foods will grow enormously. But per capit a creased consump tion of that commod ity, other con sump tion is also ex pected to in crease. In the next things being equal. The es ti mated price re spon sive - 15 years, urban popula tions are ex pected to grow ness for both beef and poul try is rather mod est. The 2.9 percent per year on aver age for all devel op ing price respon sive ness of pork and mutton is higher, coun tries (UNDP 1998). Per capita income will also but still ine las tic. Only milk among the major live - grow. Provided the poor bene fit from these trends, stock food items is somewhat price re spon sive in they will signifi cantly increase their de mand for the cross-country regres sions. animal food prod ucts with that new in come. Other It would be an error, however, to infer that there fac tors may fur ther boost de mand. Greater trade and is little scope for price poli cies to slow down the commu ni ca tions, for exam ple, will expose people growth in de mand for meat. Delgado and Courbois even in remote areas to other cultures and foods. (1998) briefly surveyed elastici ties from several rig - Whether the world has the capacity to meet orous econometric de mand analyses that in cluded this surging new demand with increased animal dis ag gre gated animal product de mand using multi - food production will be a major question for the year sam ples and na tional level data for in di vid ual rest of this report. The next chap ter will look at the countries. They found a consump tion re sponse to evolution of supply systems over the past two own prices for meats of – 0.5 to –1.0, suggest ing that decades leading up to the Live stock Revo lu tion. price respon sive ness within countries is much Subse quent chap ters will exam ine whether future higher than across countries . demand trends coin cide with future resource avail - The cross-country regres sion sys tem re ported ability and at what cost. Finally, the paper will in Table 8 involved doz ens of vari ables to contro l address simi lar questions more quali ta tively and for the many dif fer ences across time and coun tries look at envi ron mental, health, nutri tional, food that af fect con sump tion of ani mal food products . security, and techno logi cal issues.

5 Pork and mutton are com bined since most coun tries consume a large amount of one or the other, but not both. De pend ing on the coun try, ei ther pork or mutton is the main sub sti tute for beef. 3. Accompanying Transformation of Livestock Supply

Produc tion of ani mal food prod ucts grew most rap - Saharan Af rica and WANA (where milk pro duc tion idly in the same re gions where consump tion did. fell mar gin ally), indi cat ing that do mes tic sup ply kept Total meat pro duc tion in devel op ing coun tries grew up with popula tion growth in most areas (Table 10). at 5.4 percent per year be tween 1982 and 1994, al - In 1993 both Other East Asia and WANA had sub - most five times the developed-country rate stantial discrep an cies be tween per cap ita meat con- (Table 9). The high est pro duc tion growth rates for sump tion and produc tion, indi cat ing that those re- meat occurred in Asia, espe cially in China where gions imported large amounts of meat to keep up with total meat pro duc tion increased by at least 6.3 per - growing de mand (Tables 2 and 10). South east Asia cent and possi bly as much as 8.4 percent an nu ally and Sub- Saharan Af rica had substan tially higher per (Table 10).6 cap ita milk con sump tion than produc tion in 1993. Per cap ita meat and milk produc tion rose be - Poultry had the fast est total pro duc tion growth tween 1983 and 1993 in all regions except Sub- rate in both devel op ing and devel oped countrie s

Table 9— Produc tion trends of vari ous live stock products, 1982–94

Annual growth of To tal produc tion Per cap ita pro duc tion to tal pro duc tion Region/produc t 1982–94 1983 1993 1983 1993

(per cent) (mil lion met ric tons) (kilograms) De vel oped world Beef 0.1* 36 35 27 26 Pork 0.7* 35 37 29 29 Poul try 3.2 19 27 16 21 Meat 1.1 90 100 76 78

Milk – 0.4* 365 348 305 272 Devel op ing world Beef 3.1 16 22 5 5 Pork 6.1 21 39 6 9 Poul try 7.8 9 21 3 5 Meat 5.4 51 88 15 21 Milk 3.7 113 164 32 39

Sources: Annual growth of to tal produc tion 1982–94 is the growth rate from regres sions fitted to FAO an nual data (FAO 1998). To tal and per cap ita produc tion 1983 and 1993 are calcu lated from FAO 1998. Notes: Beef includes meat from cat tle and buffalo. Poul try includes all fowl listed in FAO 1998. Meat includes beef, pork, mutton, goat, and poultry car cass weights. Milk is cow and buffalo milk and milk prod ucts in liq uid milk equiva lents . Met ric tons and kilo grams are three-year mov ing av er ages cen tered on the two years shown. *Not sig nifi cantly dif fer ent from zero at the 10 percent level.

6 See discus sion of China data in Chap ter 2. The more likely figure of 6.3 per cent per year is still the highest in the world. Meat pro - duction grew at a mod est 2.9 per cent in Latin Amer ica and just 2.1 percent in Sub- Saharan Af rica. In China, Other East Asia, Latin Amer ica, and Sub- Saharan Af rica to tal meat pro duc tion grew less than to tal meat con sump tio n, though for devel op ing countries as a whole produc tion grew at the same rate as con sump tion.

13 14

Table 10— Trends in the pro duc tion of meat and milk, by re gion, 1982–94

Annual growth of to tal meat pro duc tion Per cap ita meat pro duc tion Per cap ita milk pro duc tion Re gion 1982–94 1983 1993 1983 1993 (per cent) (kilo grams) (kilo grams) China 8.4 16 33 3 6 Other East Asia 5.0 16 24 15 30 In dia 3.7 4 5 51 66 Other South Asia 4.8 6 8 50 62 South east Asia 5.7 11 16 2 3 Latin Americ a 2.9 43 48 94 101 WANA 3.9 14 16 58 57 Sub-Saharan Af rica 2.1 10 9 19 19 De vel op ing world 5.4 15 21 32 39 Devel oped world 1.1 76 78 305 272 World 2.9 30 34 102 93

Sources: Annual growth of to tal meat produc tion 1982–94 is the growth rate from regres sions fitted to FAO an nual data (FAO 1998). Per capita meat and milk produc tion for 1983 and 1993 are calcu lated from FAO 1998. Notes: Meat includes beef, pork, mutton, goat, and poul try car cass weights. Milk is cow and buffalo m ilk and milk prod ucts in liq uid milk equiva - lents. Kilograms are three-year mov ing av er ages cen tered on the two years shown. WANA is West Asia and North Africa .

between 1982 and 1994 (Table 9). Pro duc tion grew poul try, and milk increased for devel op ing coun - slowly in de vel oped coun tries for all other live stock tries as a group. prod ucts, with total and per capita out put of beef The large dis crep an cies between devel op ing - and milk falling. In devel op ing coun tries total meat and developed- country total produc tion growth rates and milk pro duc tion grew rapidly, espe cially pork are shifting world ani mal pro duc tion, with all its and poultry. Even the growth of beef pro duc tion, bene fits and costs, from de vel oped to devel op in g which was near zero in de vel oped coun tries, coun tries (Table 11). In one decade the devel op ing - amounted to a robust 3.1 percent per year in devel - coun try share of world meat and milk produc tion op ing countries. Per capita pro duc tion of pork, rose from 36 to 47 per cent and from 24 to 32 percent,

Table 11— Shares of total world pro duc tion of meat and milk, by region, 1983 and 1993

Beef Pork Poul try Meat Milk Re gion 1983 1993 1983 1993 1983 1993 1983 1993 1983 1993 (per cent)

China 1 4 25 38 5 12 12 20 1 1 Other East Asia 0 1 1 1 1 1 1 1 0 0 In dia 4 5 1 1 0 1 2 2 8 12 Other South Asia 2 2 0 0 1 1 1 1 2 4 South east Asia 2 2 3 4 5 6 3 4 0 0 Latin Americ a 19 20 6 4 13 15 12 12 8 9 WANA 2 2 0 0 5 5 3 3 3 4 Sub-Saharan Af rica 5 4 0 1 2 2 3 3 2 2 De vel op ing world 34 41 37 51 32 44 36 47 24 32 Devel oped world 66 59 63 49 68 56 64 53 76 68 (mil lion met ric tons) World 48.8 55.0 55.7 75.5 28.8 47.3 141.4 188.0 477.4 511.8

Source: FAO 1998. Notes: Meat includes beef, pork, mutton, goat, and poul try car cass weights. Milk is cow and buffalo m ilk and milk prod ucts in liq uid milk equiva - lents. Val ues are calcu lated from three-year mov ing av er ages centered on the two years show n. WANA is West Asia and North Africa . 15 re spec tively. Chi na’s share of global meat supply 1983 and 1993, espe cially in Asia where 40 percen t rose from 12 to 20 per cent. Milk produc tion is of all chickens and other fowl were located in 1993. mainly con cen trated in the de vel oped world, but China had the larg est increase in numbers of India in creased its share of world produc tion from chicken and other fowl as well. Sheep and goats 8 to 12 per cent during the period. At the rate that pro - were most highly repre sented in WANA and Sub- duc tion has been shifting to developing coun tries, it Saharan Af rica. The re gional dis tri bu tion of sheep is likely that more than 50 per cent of the world’s and goat numbers shifted lit tle from 1983 to 1993. meat is now produced in the devel op ing world and In de vel op ing coun tries rap idly increas ing meat that the same will be true for milk by 2020. and milk pro duc tion coin cided with rap idly in creas - ing numbers of animals. The developing-countr y share of the world’s stock of ani mals rose to two- Sources of Growth in the Output thirds of all pigs, fowl, sheep, and goats, and three- of Livestock Food Products quarters of all cat tle and buf falo in 1993. By con - trast, numbers of cat tle and pigs fell be tween 1983 The geographic dis tri bu tion of the world’s live - and 1993 in the devel oped coun tries, despite in- stock animals reflects dif fer ent consump tion pref er - creased beef and pork output. In de vel oped coun - ences and trends (Table 12). Cattle and buf falo are tries growth in numbers of animals was impor tan t found where beef and milk are con sumed in large only for poul try output. Increased output in devel - amounts, primar ily the developed world and Latin oped coun tries was made possi ble primar ily by in - America, which are high beef-consuming regions, creased produc tiv ity per animal, defined as greater and South Asia, which is a high milk-consuming amounts of meat or milk out put per ani mal and per region. Growth in numbers of cat tle and buffal o unit of inputs. occurred through out the develop ing world, with Com par ing the loca tion of the world’s live stock most develop ing regions increas ing their share. in Table 12 to out put shares in Table 11 pro vides an Pigs are con cen trated in the key pork- indi ca tion of the rela tive produc tiv ity lev els of the consuming coun tries of East and South east Asia. regions. Although three- quarters of the world’s cat- China’s share of the world’s pigs rose from 38 to tle and two- thirds of the world’s pigs, poul try, 44 percent be tween 1983 and 1993, and the ma jor - sheep, and goats lived in devel op ing coun tries in ity of the world’s pigs re side now in Asia. Num bers 1993, those coun tries pro duced less than half of the of chicken and other fowl grew rap idly betwee n world’s meat and a third of the world’s milk.

Table 12— Distri bu tion of the world’s live stock animals, 1983 and 1993

Cattle and buf falo Pigs Chickens/fowl Sheep and goats Re gion 1983 1993 1983 1993 1983 1993 1983 1993 (percent)

China 5 7 38 44 15 24 11 12 Other East Asia 0 0 1 1 1 1 1 1 In dia 19 20 1 2 3 2 9 10 Other South Asia 5 5 0 0 2 2 6 7 Southeast Asia 3 4 5 5 7 11 1 2 Latin Americ a 22 23 10 9 12 12 9 8 WANA 3 2 0 0 6 7 14 13 Sub- Saharan Africa 11 12 1 2 5 5 15 16 Devel op ing world 69 74 57 64 52 65 65 69 De vel oped world 31 26 43 36 48 35 35 31 (mil lion head) World 1,378 1,457 776 878 8,680 12,936 1,607 1,722

Source: FAO 1998. Note: WANA is West Asia and North Africa . 16

Table 13 compares growth rates of ani mals cent per year between 1982 and 1994, while both slaughtered or milked with growth rates in meat and milk and beef output grew by more than 3 percent . milk output, giving an in di ca tion of the ex tent to Also in Asia with out China, pig numbers grew at which produc tiv ity and in creased numbers of ani - about four-fifths the rate of pork output, in di cat ing mals con trib uted to output growth. Coun tries in the a small amount of produc tiv ity growth. Less pro - rela tively land-abundant Latin America and Sub- ductiv ity growth occurred in poul try pro duc tion, as Saharan Af rica re lied mostly on growth in num bers chicken numbers grew at ap proxi mately the same of animals for their increased live stock pro duc tion. rate as out put. In Latin Amer ica and Sub-Saharan Af rica, the Produc tiv ity is much higher in de vel oped coun- number of cat tle that were slaughtered or milked tries than that typi cally found in de vel op ing coun - grew at rates nearly equal to, or above, the growth tries. Table 14 pres ents the number of kilo grams of rate of beef and milk output, in di cat ing that number meat or milk pro duced per animal. Produc tiv ity by of animals was more im por tant than produc tiv ity in this yardstick was clearly higher in de vel oped coun- provid ing the addi tional meat. Pig numbers grew at tries, espe cially for beef and milk. Pork and poultry about the same rate as pork output in both regions, produc tiv ity lev els showed greater similar ity across indi cat ing lit tle produc tiv ity growth. The number of re gions. chickens grew at about the same rate as poul try out - Certain devel op ing re gions appear to be catch - put in Af rica. In Latin America the number of ing up with developed-country rates for per ani mal chickens grew more slowly than poul try out put, produc tiv ity. Produc tiv ity growth rates in some de - suggest ing the exis tence of produc tiv ity growth in velop ing coun tries exceed those in de vel oped coun - poultry output. tries for some com modi ties (Table 14). Beef pro - In Asia, where land is scarce, growth in num - ductiv ity through out Asia has been grow ing at rates bers of animals made up a smaller propor tion of higher than the 0.9 percent growth rate in the devel - out put growth for beef and pork. Produc tiv ity oped world. Milk produc tiv ity growth rates in Asia, growth was relatively more impor tant. Not in clud - with the excep tion of China, also exceeded those in ing China, where reported produc tiv ity growth was de vel oped coun tries. Beef and milk produc tiv ity in even greater, cat tle numbers grew less than 2 per - Latin America and Sub-Saharan Af rica fell in creas-

Table 13— Growth rates of live stock out put and number of animals slaughtered or milked, 1982–94

Cat tle Milk Pigs Chickens Re gion Output Slaughtered Output Milked Output Slaughtered Output Slaughtered

(percent growth per year) China 20.0 15.5 10.0 11.6 7 .2 5 .8 13.1 10.3 Other East Asia 3.3 2 .0* 8 .1 3 .0 5 .6 3 .5 8 .9 8 .4 In dia 3 .6 2 .2 6 .4 1 .6 2 .8 2 .8 11.9 11.9 Other South Asia 2.4 0 .7 2 .9 1 .7 4 .9 3 .8 8 .2 6 .3 South east Asia 4.2 3 .4 4 .4 2 .1 5 .7 4 .8 7 .1 7 .5 Asia, exclud ing China 3.4 1 .8 4 .2 1 .8 5 .7 4 .6 7 .5 7 .4

Latin Americ a 2.1 1 .8 2 .5 1 .8 0 .1* – 0 .4* 6 .6 5 .5 WANA 3.0 0 .3* 2 .7 1 .2 5 .9 5 .8 5 .9 5 .6 Sub-Saharan Af rica 0.3 0 .8 2 .9 2 .3 7 .8 7 .7 4 .0 4 .1 De vel op ing world 3.0 2 .5 3 .8 2 .0 6 .1 4 .8 7 .7 6 .9 De vel op ing world, exclud ing China 2.1 1 .6 3 .6 1 .8 3 .3 2 .9 6 .6 6 .2

Devel oped world 0.1 – 0 .8 – 0 .4 –1 .7 0 .7 0 .3 2 .7 1 .9 World 1.1 0 .6 0 .5 0 .3 3 .1 2 .5 4 .7 4 .0

Source: Es ti mated from FAO 1998 data. Notes: As dis cussed later in the re port, the offi cial China produc tion figures for the mid- 1990s a re currently under re vi sion and may be reduced . China has been ex cluded from the Asia fig ures to avoid any bias this may in tro duce. *Not sig nifi cantly dif fer ent from zero at the 10 percent level. 17

Table 14— Pro duc tiv ity by re gion and ani mal type, 1992–94, and productivity growth rate, 1982–94

Beef from cat tle Milk Pork Poultry Pro duc- Growth Pro duc- Growth Pro duc- Growth Pro duc- Growth Re gion tivity rate tiv ity rate tivity rate tiv ity rate

(kilograms / (per cent/ (kilograms / (per cent/ (kilograms / (per cent/ (kilograms / (per cent/ head) year) head) year) head) year) head) year)

China 149 4.5 1,530 –1.6 76 1.4 1.3 2.8 Other East Asia 207 1.3H 1,983 5.1 72 2.1 1.1 0.5*H In dia 103 1.4 973 4.8 35 . . . 0.9 0.0 Other South Asia 111 1.7 538 1.2 H 37 1.1 1.0 1.9 Southeast Asia 170 0.9H 628 2.4 58 0.8 1.1 –0.3 Asia, de vel op ing countries, exclud ing China 147 1.5 672 2.3 61 1.1 1.1 0.0* Latin Americ a 194 0.2 1,137 0.7 71 0.5 H 1.4 1.1 H WANA 135 2.7 1,236 1.5 H 69 0.1*H 1.1 0.2 Sub- Saharan Af rica 132 –0.5 340 0.6 45 0.1 0.9 –0.1* Devel op ing world 162 0.5 896 1.9 72 1.2 1.2 0.8 H Devel op ing world, exclud ing China 164 0.5 879 1.8 63 0.5 H 1.2 0.5 De vel oped world 242 0.9 3,739 1.3 82 0.4 1.4 0.8 World 204 0.5 2,073 0.2 76 0.6 1.3 0.7

Source: FAO (1998) figures on to tal food com mod ity produc tion per region and item are di vided by th e cor re spond ing number of ani mals slaughtered (or milked) from FAO 1998. Growth rates of produc tiv ity are from regres sions fit ted to annual es ti mates of produc tiv ity ob tained through this means. WANA is West Asia and North Africa . *Not sig nifi cantly dif fer ent from zero at the 10 percent level. HNot sig nifi cantly differ ent from devel oped world at the 10 per cent level. ingly be hind the de vel oped coun tries. Pork pro duc - ering prod ucts for an in creas ingly quality- conscious tiv ity growth rates exceeded developed- country urban popu la tion. Indus trial live stock pro duc tion lev els in Asia. Poul try produc tiv ity growth was on maxi mizes the use of scarce re sources, no ta bly av er age more rapid in devel oped coun tries than in land, labor, and feed, and it involves the de vel op - de vel op ing coun tries ex clud ing China. ment of genotypes, ap pli ca tion of bio tech nol ogy, gen eral improve ment in ani mal hus bandry and veterinary care, and advances in the back ward and Sources of Productivity Growth for ward linkages of live stock output (such as meat Produc tiv ity growth in de vel oped regions mainly mar ket ing systems, feed mills). Pro duc tion costs occurs through fur ther tech no logi cal progress. of monogas trics, such as pigs and poultry, tend to Farmers can raise many more animals per unit of fall faster than those for ru mi nants in land- scarce land by using capital- intensive mecha ni za tion that situations because monogas trics require less space reduces labor re quire ments, by increas ing per ani- and are more effi cient at con vert ing feed con cen - mal feed use and feed qual ity, and by invest ing in trates to meat. im proved ani mal ge net ics and health. Nearly 37 Live stock pro duc tion in de vel op ing coun tries per cent of the world’s meat supply comes from in- relies much more on tradi tional opera tions. A quar - dus tri al ized live stock pro duc tion (FAO 1995b). In ter of the world’s land is used for grazing, which re cent years, indus trial live stock pro duc tion grew sustains about 10 percent of world meat pro duc tion globally at twice the rate (4.3 percent) of more tra - (FAO 1995b). Grazing systems typi cally in crease di tional, mixed- farming sys tems (2.2 percent), and pro duc tion by increas ing the number of heads and more than six times the rate of grazing sys tem (0.7 the land area used. As land becomes scarce, graz ing per cent) (Sere and Ste in feld 1996). sys tems lead to ei ther land deg ra da tion and eco- In dus trial live stock pro duc tion is knowl edge- nomic de cline or mixed or indus trial live stock pro- and management-intensive, espe cially when deliv - duc tion systems. 18

Mixed live stock and crop pro duc tion is the velop ment led to the adop tion of a number of tech - most common form of live stock opera tion in devel - nologies that dealt first with land scar city and then op ing coun tries, pro vid ing more than 50 percent of re placed labor with capi tal. the meat produced in the world (FAO 1995b). The The first set of in no va tions intro duced into crop com po nent of the farm pro vides residues for Asian mixed- farming sys tems in cluded basic ani mal rough age, while the live stock com po nent pro vides health care, such as con trol of in fec tious diseases and ani mal traction, fertil izer, ani mal fi bers, a form of parasites. Next came the provi sion of ad di tional feed, sav ings or collat eral, and a role in social func tions. first from crop byprod ucts, but then in creas ingly Live stock kept in mixed systems are primar ily large from cere als and other con cen trates. In In do ne sia and small ru mi nants because they are effi cient at and elsewhere, cut and carry systems, es sen tially a con vert ing pastures, crop residues, and other rough - combi na tion of for age cul ti va tion and stall feeding , ages into meat. Such fi brous mate ri als and grasses de vel oped for milk produc tion. Breeding con tin ued have lit tle or no alter na tive use. Large ru mi nants to be based on local se lec tion for pre ferred traits; can also pro vide farm power. crossbreed ing was prac ticed to some extent . Currently an esti mated 250 million working As demand for live stock foods expanded rap- ani mals provide draft power for mixed farms that idly, mixed- farming sys tems could not keep up. cover about 28 per cent of the world’s ar able land. Feed requirements could no longer be met from Approxi mately 52 per cent of avail able cropland in domes tic supplies of cere als and other con cen trates. devel op ing coun tries is farmed using ani mal draft Asia began to import large amounts of feedgrains , power. The use of ani mal draft power increased in mainly from the developed coun tries. At this point the 1970s and 1980s in those parts of West Afric a the indus trial pro duc tion of pork, poultry, and eggs where the technol ogy was relatively new, dis ease that emerged was more effi cient at using imported was being controlled, and intro duc tion of new crops feeds. Indus trial sys tems make use of imported live - such as cotton and maize re quired added farm stock genetic material and sophis ti cated feed ing power (Pingali, Bigot, and Binswanger 1987). prac tices, such as phased feeding, the use of feed Eastern and Southern Africa and South Asia in additives, and, at an even more advanced stage, particu lar have longer tradi tions in the use of draft synthetic amino-acids. As of this writing, in early ani mals and are likely to con tinue to use them for 1999, these sys tems appear to have been dispro por - some time to come. tion ately disad van taged by the Asian eco nomic cri - Mechani za tion is rap idly occur ring in other sis, which raised the cost of imported feeds and areas, such as East and South east Asia (Ste in feld depressed urban demand. 1998). In these areas mixed farm ing is evolv ing as WANA ex hib its differ ent trends. Land is not a the food value of ani mals increases and the value of limit ing factor per se but agri cul tural poten tial is, their other uses declines. The dif fu sion of ma chin - given the scarcity of water. Tra di tional pastor al is m ery, fertil izer, syn thetic fi bers, and finan cial serv - and, to a limited extent, mixed farming con tinue to ices re duces the value of live stock’s other roles. exist, but oil reve nues and the re sult ing eco nomic ex - Mechani za tion improves the produc tiv ity of ani mal pan sion since 1973 have in tro duced im ported indus - food pro duc tion because animals no longer need to trial pro duc tion units, for poultry and dairy in par- be kept into adult hood for draft power, ena bling ticular. These units have state- of the- art technol ogy more rapid slaughter rates. Mecha ni za tion also but require many im ported inputs and the do mes tic makes possi ble a shift from large and small ru mi - produc tion of others (for ex am ple, forage produc tion nants to animals such as pigs and chickens that re - for dairy cows). For the most part they cannot com - quire less time and space for pro duc tion. pete with world mar kets but are maintained through In past decades live stock devel op ment in Asia protec tion as a mat ter of politi cal choice. suf fered from an ex treme short age of land and qual - Because of the emergence of these sub si dized ity feed but en joyed a rela tive abundance of labor indus trial pro duc tion systems, as well as other and water. Capital, first in short supply, be came less major distor tions in the food mar ket, lit tle tech no - of a limit ing fac tor as rapid indus trial devel op ment logical change has occurred in the graz ing and took place and incomes in creased. Progress in de - small mixed- farming sec tors in WANA. Feed lots 19 for small ru mi nants have de vel oped to some ex tent, cen ters, and the tech nol ogy used in ex ten sive ranch - mainly in re sponse to the market re quire ments of ing only involves such basic measures as fenc ing, the region. These require ments include the sea sonal disease pre ven tion and treatment, and some ge netic demand for whole ani mals dur ing re lig ious holi - de vel op ment. days. In more ecol ogi cally favor able envi ron ments, Recently, in ten sive poul try pro duc tion and, to no ta bly the Nile valley in Egypt, com peti tive dairy some extent, dairying have de vel oped in Latin systems have emerged that use a mix ture of do mes - America. These live stock systems took advan tage tic and imported feed re sources and inter me di at e of the tradi tion ally high ur bani za tion rate and a labor- intensive technol ogy . resurgence of eco nomic growth in the 1990s. Inten - The relative abundance of land and extrem e sive in na ture, these sys tems use many of the same scar city of capital resulted in lit tle produc tiv ity in - basic technolo gies found in de vel oped coun tries but crease per ani mal in Sub-Saharan Af rica over the use them at lower lev els of in ten si fi ca tion. past two dec ades. Without signifi cant per cap ita income growth the re gion lacked the stimuli and the means for adopt ing meat pro duc tion tech nol ogy Feed Use outside the poul try sec tor in large coastal mar kets. Dump ing of live stock prod ucts from de vel oped Recent rapid increases in meat pro duc tion caused countries dur ing the late 1970s and throughout the global use of cereal as feed to rise at 0.7 percen t 1980s also discour aged pro duc tion inno va tion, as annu ally between 1982 and 1994. This growth rate did overval ued ex change rates that favored im ports. reflects neg li gi ble growth in use of cereal as feed in Sub- Saharan Afri ca’s live stock sec tor contin - the developed coun tries and a more than 4 percen t ues to be largely made up of ru mi nants located in per year growth rate in the develop ing countrie s tsetse- free areas and fed lo cally available feed. (Table 15). Despite the higher growth rate, devel op - By and large health care is ru di men tary and only ing coun tries still use less than half as much cereal simple feed sup ple ments, such as min er als, are pro - for feed as do developed coun tries. In 1990–92 con - vided. Close to urban centers, and where agro- cen trated cereal feed pro vided between 59 and ecological con di tions per mit, semi- intensive and 80 percent of the nutri tion pro vided to animals in inten sive dairy ing has de vel oped using culti vated the developed world. By con trast, cere als accounted fodder and agro-industrial by- products. Poul try for only 45 percent of total concen trate feed in pro duc tion has begun to in dus tri al ize. Apart from Southeast Asia, the devel op ing region with the most low income and a short age of quality feed, live stock intensive use of feedgrains . pro duc tion tech nol ogy in Africa contin ues to be As output of live stock prod ucts grows in devel - severely constrained by dis ease, espe cially trypano - oping regions, ani mal pro duc tion meth ods and feed- somia sis, which is spread by the tsetse fly (Al ex an - ing patterns are shifting rapidly. Grazing system s dratos 1995). Cattle pro duc tion is diffi cult in areas are rapidly diminish ing in importance through out infested with the tsetse fly. Other diseases must be the world. Land avail able for grazing is caught in a con trolled as well before widespread inten si fi ca tion squeeze. Urbaniza tion and crop produc tion are of pork and poul try pro duc tion can occur. encroach ing on tra di tional grazing areas. Preser va - Latin America is char ac ter ized by an abun - tion efforts are lim it ing expansion of grazing opera- dance of land. The region expe ri enced consid er abl e tions into virgin areas. livestock devel op ment prior to the 1960s. Ur bani - Mixed-farming systems also face limits. In no - zation had already ad vanced by that point, un like in vations in crop pro duc tion have reduced crop resi - the rest of the devel op ing world. Income stag na tion dues and non grain bio mass available for feed ing. in the 1970s and early 1980s and low world marke t Crop research has largely ignored the feed value of prices for meat slowed produc tiv ity growth and crop residues. Un im proved varie ties in low ex ter nal tech nol ogy adoption. It was gen er ally cheaper to input systems typi cally pro duce three to four times ex pand live stock pro duc tion into new areas than as much nongrain biomass as grain, whereas mod - invest in new tech nol ogy. Pastures con tinue to be ern hy brids often produce an equal share or less un im proved, except for areas close to con sumer nongrain bio mass as grain. 20

Table 15— Trends in the use of cereal as feed, pro duc tion in par ticu lar. But small-scale backyard 1982–94 op era tions are disap pear ing because of low re turns to labor and in creased compe ti tion from large-scale Annual growth rates To tal ce real pro duc ers. Although each backyard opera tion is To tal cerea l To tal ce real use as feed pro duc tion, use as feed, small, at the ag gre gate level such sys tems act as Regio n 1982–94 1982–94 1983 1993 major transform ers of waste into meat and milk. (per cent) (mil lion Because large opera tions are unlikely to find it metric tons) cost-effective to collect small amounts of waste China 2.1 5 .8 40a 84 from many households, this source of ani mal feed Other East Asia –2 .3 6 .8 3 7 may be under used in indus trial systems . In dia 3 .0 3 .5 2 3 The use of ce re als as feed has been fast est in Other South Asia 2.0 0 .9 1 1 South east Asia 2.4 7 .2 6 12 Asia, where out put growth has risen the most and Latin Americ a 0.9 2 .6 40 55 land is scarce (Table 15). In Other East Asia, South - WANA 3.7 1 .9 24 29 east Asia, and Sub-Saharan Af rica, ce real use as Sub-Saharan Africa 4.0 5 .2 2 3 De vel op ing world 2.3 4 .2 128 194 feed grew faster than meat pro duc tion, in di cat ing Devel oped world 0.2* –0 .5* 465 442 that those re gions are inten si fy ing their use of feed World 1.2 0 .7 592 636 per unit of meat output (Ta bles 10 and 15). Most of Asia, WANA, and Sub-Saharan Afric a Sources: Pro duc tion and use growth rates for 1982–94 are from regres - lack the capac ity to pro duce substan tial amounts of sions fitted to FAO an nual data (FAO 1998). To tal use for 1983 and 1993 is calcu lated from FAO 1998. feedgrain at com peti tive prices. The grow ing Notes: Ce re als in clude wheat, maize, rice, bar ley, sorghum, mil let, amounts of feedgrains imported into these region s rye, and oats. Met ric tons are three-year mov ing av er ages attest to this de fi ciency. Given that many de vel op - centered on the two years shown. WANA is West Asia and North Af rica. ing coun tries can not expand crop area, two possi - aSimp son, Cheng, and Miyazaki (1994) report 40 mil lion met ric tons bilities re main: in ten si fi ca tion of exist ing land re - from U.S. Depart ment of Ag ri cul ture data. That figure is used here be - sources and impor ta tion of feed. Because much of cause it is more consis tent with the feed quan ti ties and feed/meat con - ver sion ratios in Rosegrant et al. 1997. FAO (1998) reports 49 mil lion the gain from in ten si fi ca tion will probably go to- metric tons. ward meet ing the increas ing de mand for food crops, *Not sig nifi cantly dif fer ent from zero at the 10 percent level. sub stan tially more feedgrains will have to be im - ported by de vel op ing coun tries in the fu ture. Will Household food waste, such as tuber skins, feedgrain avail abil ity, the infra struc ture for moving stems, and leaf tops, has tra di tion ally been an other large amounts of grain, and other access-relate d impor tant feed resource, for backyard monogastric factors keep up with pro jected surges in demand? 4. Projections of Future Demand and Supply to 2020

Trends dur ing the past 15 years sug gest that the IFPRI (Rosegrant, Agcaoili- Sombilla, and Perez developing coun tries are in the midst of a demand- 1995; Rosegrant et al. 1997; Rosegrant, Leach, and driven Live stock Revo lu tion that already has in - Gerpa cio 1998; and Rosegrant and Ringler 1998) creased the shares of world live stock prod ucts con - that is particu larly ap pro pri ate for looking at these sumed in those coun tries by a substan tial mar gin. issues and assess ing the sen si tiv ity of results to Even more im por tant ef fects are expected in the different assump tions. The tool in ques tion is the foresee able future. These fu ture changes are of such Inter na tional Model for Policy Analysis of Ag ri cul - magni tude that they cannot occur in a vac uum. tural Com modi ties and Trade (IM PACT), June They will af fect the global economy and in turn be 1998 version. 7 (For fur ther in for ma tion on the affected by it. model, see the box.) This chapter and the next will ex am ine the likely mag ni tude of the changes, the fea si bil ity of the changes given the world’s capac ity to pro duce Projected Consumption livestock and feed, the like li hood of the require d Trends to 2020 in the pro duc tion in creases oc cur ring in devel op ing or Baseline Scenario developed coun tries, and the impli ca tions for world prices of meat, milk, and ce re als. This chapter pres - The baseline sce nario rep re sents the most real is tic ents both an approach to ad dress ing these questions set of as sump tions govern ing inter na tional and and some an swers in the form of base line result s na tional econo mies. With the sole excep tions of from a global eco nomic model. It also looks at what beef in the devel oped countries and milk in the some plau si ble changes in fu ture scenar ios might developing coun tries, con sump tion of food prod - en tail for world milk and meat. Chapter 5 ex am ines ucts of animal origin is projected to grow at a sub - the pro jected impact of the Live stock Revo lu tion on stantially lower annual rate over the 1993–2020 world trade in meat and milk and world food prices pe riod than it did in the 1982–94 pe riod (Table s under dif fer ent scenar ios . 16 and 7). The projected rates of growth to 2020 Global eco nomic mod els are subject to many are ex pected to be about half those ob served in the un cer tain ties and are typi cally speci fied at such a last 15 years in most cases. Three factors pro duce high level of product aggre ga tion (meat and milk these lower growth rates to 2020. First, re cent are usu ally lumped in with “food”), that they are not rapid growth in consump tion means that the base very use ful for the purpose at hand. For tu nately, the for pro ject ing growth beyond 1993 is larger than authors have at their dis posal a model de vel oped at that in 1983. Thus a given abso lute an nual in cre -

7 The June 1998 version of IMPACT in cor po rates pa rame ters and as sump tions that re flect lowe r ex pected overall eco nomic growth in Asia relative to the as sump tions in prior ver sions, in ad di tion to updated pa rame ter es ti mates for Af rica (Rosegrant and Ringler 1998). De mand and sup ply parame ters for milk have also been ex ten sively re viewed and re vised in this ver sion; milk was in cluded previ ously, but not reported on sepa rately. Read ers inter ested in the str uc ture of model equations, which have not changed, are referred to the appen dix of Rosegrant, Agcaoili- Sombilla, and Perez (1995 ), pub lished in the same serie s as the pres ent pa per.

21 22

The Baseline IMPACT Model

IM PACT is a global food model that di vides the world are too de tailed to re port here, but at ten tion is given to into 37 countries or coun try groups. Its base line ver sion those parame ters that mat ter the most, such as the de - rep re sents the most re al is tic set of as sump tions ac cord - tailed struc tural parame ters of shift ing meat de mand in ing to the model team. The coun tries and groups can be China, taken from Huang and Bouis (1996). con ven iently aggre gated into re gions that are com pati - Na tional in come, popula tion, and ur ban growth ble with FAO defi ni tions (see the Ap pen dix). Fan and rates are also as sumed for each coun try group, along Agcaoili- Sombilla (1998) com pared the results of early with antici pated changes in these rates over time. The ver sions of the model and other in ter na tional mod els for model uses the re vised United Na tions medium- ce real pro duc tion and con sump tion in China in 2010 variant pro jec tions for 1996 for demo graphic as sump - and 2020. They found that the pro jec tions from the tions. Na tional income pro jec tions are es ti mated based IMPACT model are mid dle of the road in out look and on a re view of pro jec tions drawn from sources such as nei ther pessi mis tic nor bull ish with re spect to the is sues the World Bank. The model is solved on an annual raised at the be gin ning of this chap ter. basis by link ing each coun try model to the rest of the The IM PACT model cov ers 18 com modi ties, in- world through com mod ity trade. The market-clearin g clud ing beef, pork, poul try, sheep meat, goat meat, bo - con di tion solves for the set of world prices that clears vine milk, and eggs. The base data used in the cur rent in ter na tional com mod ity mar kets, so that the to tal im- ver sion are av er ages of the 1992–94 an nual data from ports of each com mod ity equals to tal ex ports. World the FAO Sta tis ti cal Da ta base (the same source used for prices of com modi ties thus act as the equili brat ing the ta bles in pre vi ous chapters, with 1993 data ex actly mecha nism and main tain the model in equi lib rium. equiva lent to the IM PACT base year). Since each of the When an ex oge nous shock is in tro duced in the model, 37 coun try groups pro duces and/or con sumes at least such as an increase in crop yields from higher in vest - some of each com mod ity, lit er ally thou sands of sup ply ment in crop re search, the world price will ad just and and de mand parame ters had to be speci fied (in come, each ad just ment is passed back to the effec tive pro- price, and cross- price elas tici ties of de mand; pro duc tion ducer and con sumer prices. Changes in do mes tic pa rame ters in clud ing crop area, yield growth trends, prices sub se quently af fect the sup ply and demand of and herd size and pro duc tiv ity; price re sponse pa rame - the com modi ties, ne ces si tat ing their itera tive read just - ters; ini tial lev els and trends in feed con ver sion; trade ments un til world sup ply and de mand are in balance dis tor tion pa rame ters; and so on). Pa rame ter es ti mates and world net trade again equals zero. were drawn from economet ric analy sis, as sess ment of The out come of this an nu al ized it era tive pro cess is past and chang ing trends, ex pert judge ment, and syn the - an es ti mated an nual se ries of pro jected market- clearing sis of the ex ist ing lit era ture. The myr iad assump tions prices, con sump tion lev els by com mod ity and coun try

ment accounts for an increas ingly smaller per - developed coun tries in 2020, up from 28 percent in cent age incre ment. Second, slowdowns in the rate the early 1990s. Yet in ag gre gate terms 62 percen t of overall in come and ur bani za tion growth will of the world’s meat and 60 percent of milk con - occur for the same reason. Third, con sum ers sumption will take place in the de vel op ing coun - begin to get sa ti ated as the impor tance of meat in tries in 2020. This is a major change from the early their diets increases. 1990s, when 52 percent of the world’s meat and Even by 2020, per capita consump tion of milk 59 percent of the world’s milk were con sumed by prod ucts in devel op ing coun tries is expected to be the de vel oped world. on av er age only one- third that of de vel oped coun - The consump tion trends by re gion and the tries (up from less than one- fifth in the early 1990s). annual growth rates to 2020 are given in Table 17. Per cap ita consump tion of meat in devel op ing The pro jected consump tion growth rates for countries is pro jected to be 36 percent of that in China of 3 percent for meat and 2.8 percent for 23

group, feed- use lev els, pro duc tion area, yield and pro - period is modi fied by the rela tive price move ments of duc tion lev els by com mod ity and re gion, and net trade feeds, which change in the model depend ing on both across coun try groups by com mod ity. The analy sis ex - human and ani mal demand and by the speci fied rates of tends to the year 2020. The in creas ing globali za tion of change in feed ing effi ciency due to tech ni cal prog ress. ag ri cul tural mar kets is rep re sented in the model by its Browse, back yard slops, and other non ce real, non- set of en doge nous world prices es ti mated through an - oilcake feeds are implic itly costed as free, and are thus nual it era tive so lu tions. Events af fect ing net ex ports not explic itly mod eled. Report ing focuses on cere als from one coun try group will af fect prices in oth ers, in used as feed, since this is the aspect of feed use that is turn chang ing sup ply, de mand, and net trade, even tu ally con tro ver sial and it is also in all like li hood the bind ing con verg ing to a so lu tion. con straint in pro duc ing ade quate mixed feed rations. An im por tant ad van tage of a mod el ing ap proach The approach has the drawback that where pro- such as IM PACT is that it can be used to in ves ti gate duc tion sys tems are shift ing from low cereal use sys- the fea si bil ity of pro duc tion in creases simul ta ne ousl y tems (as in back yard poul try or range- fed rumi nants) with the de ter mi na tion of pro duc tion costs and out put to indus trial feedlots, cur rent ratios of cereal feed to prices. Ac tual pro duc tion of commodi ties in the model spe cific meat prod ucts may under es ti mate future re quires the use of suf fi cient quan ti ties of all needed ratios. The exoge nous feed ratio effi ciency paramete r in puts in the coun try and year con cerned. Par ticu lar in the feed demand equa tion is speci fied to com pen sate attention is de voted to cereal- based feeds. Ce real feed for this in some coun tries. It also allows fur ther sen si - de mand in IM PACT is driven by live stock pro duc tion, tiv ity test ing in this area. ce real feed ratios, own- and cross- price re la tion ships The ra tio of to tal ce real feed use to to tal meat pro - among feed crops, and an ex oge nously speci fied “ef fi - duc tion (with feed use for milk and eggs net ted out) in ciency pa rame ter” that can be used to model exoge - the early 1990s was 1.40 to 1.00 in China and 3.64 to nous tech ni cal prog ress or other secu lar changes af - 1.00 in the United States. Between 1983 and 1993, fect ing feed de mand. feed con ver sion ef fi ciency in the United States in - The mix of feeds used for a spe cific live stock prod - creased by about 15 percent across all meats, with the uct starts from actual aver ages in the 1992–94 base in creases for poul try out strip ping those for beef period. For each live stock prod uct, the feed ing ratio for (CAST forthcom ing). The pro por tion ately greater use each main feed com mod ity in the model—maize, other of natu ral pas tures, house hold wastes, roots, tubers , coarse grains, wheat, oil cakes, cas sava, sweet pota toes, and by prod ucts for feed ing in de vel op ing coun tries ex - and pota toes—is speci fied for the base period. What hap - plains why their ce real feed ra tios are lower than those pens to the amount of each com mod ity fed after the base in de vel oped coun tries.

milk are on the low side of the litera ture (Ke above the 0.6 and 0.2 percent av er ages projected 1997).8 They are fully compa ra ble to those pro - for de vel oped coun tries as a whole. Per capit a jected for other parts of Asia, but are signifi cantly meat consump tion in 2020 is pro jected to remain

8 As noted in Chapter 2, the FAO consump tion and produc tion figures for meat in China in the ea rly 1990s, used here for the 1992–94 base, may over state ac tual consump tion and pro duc tion by up to 30 percent (Ke 1997). Ke stud ied t he discrep ancy be tween food- balance- sheet and sur vey esti ma tions of meat consump tion in China in de tail. He con cludes that ac tual to tal de mand for meat is only likely to grow at 3 to 5 per cent per year in China in the foresee able fu ture, a lower rate than that pre dicted by “so me” (un speci fied) per sons. While the IMPACT-estimated level of meat de mand in China in 2020 might be seen as somewhat high be cause o f a base pe riod es ti mate of meat consump tion that is “too high,” the IMPACT es ti ma tion is also the re sult of pro jected an nua l growth over 27 years that is “low” be cause of the conser va tive as sump tions built into the model. There is also lit tle concern about the ef fect of an over es ti mate of Chi nese livestock production in the base period (1992–94) on im ports in IMPACT be cause the over es ti mate of pro du c tion, if there is one, is bal anced by an over es ti mate of consump tion. There does not ap pear to be a com pel ling rea son to revise th e es ti mates for China given here. 24

Table 16— Projected trends in the food consumption of various livestock products, 1993–2020

Projected an nual To tal An nual per cap ita growth of tota l consump tion con sump tion consump tion , Re gion/Product 1993–2020 1993 2020 1993 2020 (per cent) (mil lion met ric tons) (kilograms) Devel oped world Beef 0.4 32 36 25 26 Pork 0.3 36 41 28 29 Poul try 1.0 26 34 20 25 Meat 0.6 97 115 76 83 Milk 0.2 245 263 192 189 De vel op ing world Beef 2.8 22 47 5 7 Pork 2.8 38 81 9 13 Poul try 3.1 21 49 5 8 Meat 2.8 88 188 21 30 Milk 3.3 168 391 40 62

Sources: To tal and per cap ita con sump tion for 1993 are calcu lated from FAO 1998. Projec tions are up dated fig ures follow ing the same for mat as that reported in Rosegrant et al. 1997. Notes: Con sump tion refers to direct use as food, meas ured as uncooked weight, bone in. Meat in clude s beef, pork, mutton, goat, and poultry. Milk is cow and buffalo milk and milk prod ucts in liq uid milk equiva lents. Met ric tons and kilogr a ms are three-year mov ing av er ages centere d on the two years shown. low in Sub-Saharan Af rica, partially vegetar ia n Projected Production India, and other coun tries in South Asia. The Trends to 2020 assumptions con cern ing tastes for meat in India will be tested below as part of the sensi tiv it y Pro jected produc tion trends for meat to 2020 analy sis, as will assump tions about pro duc tiv ity closely fol low those pro jected for consump tion , growth and the se ver ity of the Asian finan cia l with the excep tion of WANA (Ta bles 17 and 18). crisis . Gener ally IM PACT proj ects that in 2020 defici t

Table 17— Pro jected trends in meat and milk con sump tion, 1993–2020

Projected annual growth To tal Per cap ita of to tal con sump tion, con sump tion con sump tion 1993–2020 in 2020 in 2020 Regio n Meat Milk Meat Milk Meat Milk

(percent) (mil lion met ric tons) (kilograms)

China 3.0 2.8 85 17 60 12 Other East Asia 2.4 1.7 8 2 67 20 In dia 2.9 4.3 8 160 6 125 Other South Asia 3.2 3.4 5 41 10 82 South east Asia 3.0 2.7 16 11 24 16 Latin Americ a 2.3 1.9 39 77 59 117 WANA 2.8 3.0 15 51 24 80 Sub-Saharan Af rica 3.5 3.8 12 31 11 30 De vel op ing world 2.8 3.3 188 391 30 62 Devel oped world 0.6 0.2 115 263 83 189 World 1.8 1.7 303 654 39 85

Sources: Projec tions are updated fig ures follow ing the same for mat as that reported in Rosegrant et al. 1997. Notes: Con sump tion refers to direct use as food, meas ured as uncooked weight, bone in. Meat in clude s beef, pork, mutton, goat, and poultry. Milk is cow and buffalo milk and milk prod ucts in liq uid milk equiva lents. Met ric tons and kilogr a ms are three-year mov ing av er ages centere d on the two years shown. WANA is West Asia and North Africa. 25

Table 18— Pro jected trends in meat and milk production, 1993–2020

Projected annual growth of to tal produc tion , To tal pro duc tion Per cap ita pro duc tion 1993–2020 in 2020 in 2020 Re gion Meat Milk Meat Milk Meat Milk

(per cent per year) (mil lion met ric tons) (kilograms)

China 2.9 3.2 86 19 60 13 Other East Asia 2.4 3.9 7 3 55 29 In dia 2.8 1.6 8 172 6 135 Other South Asia 2.6 3.1 4 46 9 92 Southeast Asia 3.1 2.9 16 3 25 5 Latin Americ a 2.2 2.0 39 80 59 121 WANA 2.5 2.6 11 46 18 72 Sub- Saharan Africa 3.4 4.0 11 31 10 30 Devel op ing world 2.7 3.2 183 401 29 63 De vel oped world 0.7 0.4 121 371 87 267 World 1.8 1.6 303 772 39 100

Sources: Projec tions are updated fig ures follow ing the same for mat as that reported in Rosegrant et al. 1997. Notes: Meat in cludes beef, pork, mut ton, goat, and poultry. Milk is cow and buf falo milk and milk prod ucts in liquid milk equiva lents. Met ric tons and kilo grams are three-year moving aver ages cen tered on the two years shown. WANA is We st Asia and North Africa .

countries will im port feed rather than meat. This per cent or less for each of the major live stock food con tra dicts the conven tional wisdom that it is commodi ties other than poul try (Table 19). In the cheaper to trade the com mod ity with the higher de vel op ing coun tries the pro jected an nual rates of value-added (meat as opposed to bulk cere als). But growth are 2.7 percent for meat and 3.2 percent for it is con sis tent with cur rent trade pat terns and with milk. The major ity share of live stock pro duc tion the ex pe ri ence of rap idly devel op ing coun tries such will take place in devel op ing coun tries, even though as Taiwan. These live stock trade pat terns may re - per cap ita pro duc tion lev els will be much higher in flect the reali ties of rela tive over all costs and bene - the de vel oped coun tries. fits of pro duc ing and trading strate gies better than a simple compari son of value to bulk. Compari son of pro jected milk pro duc tion and Projected Cereal Feed consump tion in 2020 shows a somewhat dif fer ent Use to 2020 story. Unlike meat, milk is itself a major input into do mes tic live stock pro duc tion. The con sump tion The trends in ce real feed use since the early 1980s figures in Table 17 in clude only what is con sumed (Table 15) highlighted the fact that the growth rates by humans, but milk pro duc tion figures include of ce real feed use ex ceeded cereal produc tio n milk used as feed. Milk pro duc tion in 2020 in the growth rates in all regions of the world excep t de vel oped coun tries is pro jected to exceed milk WANA. Compari son of pro jected and his tori cal consump tion by 108 mil lion metric tons (or 29 per - growth rates of cereal feed use shows a slowdow n cent of pro jected pro duc tion). Seventy-seven mil - through 2020 in China, Other East Asia, Southeast lion metric tons of milk will be used as feed and the Asia, and Sub-Saharan Af rica (Ta bles 15 and 20). re main ing surplus will be exported to de vel op ing But feed use accel er ates in India, Other South Asia, countries. Only 41 mil lion metric tons (10 per cent and WANA, largely due to fast- growing milk pro - of pro jected produc tion) are likely to be used as duction. feed in de vel op ing countries, de spite a herd size Per cap ita cereal use as feed increases by only larger than that in the de vel oped coun tries. 8.1 percent in ag gre gate from the early 1990s to In de vel oped countries the an nual growth rate 2020 in de vel oped coun tries, but rises by 44 percent of pro duc tion through 2020 is pro jected to be 0.7 in de vel op ing countries. Given this stagna tion in 26

Table 19— Projected trends in production of various livestock products, 1993–2020

Projected annual growth of to tal produc tion , To tal pro duc tion Per cap ita pro duc tion Re gion/product 1993–2020 1993 2020 1993 2020

(per cent) (million met ric tons) (kilo grams) Devel oped world Beef 0.6 35 38 26 28 Pork 0.4 37 41 29 29 Poul try 1.2 27 36 21 26 Meat 0.7 100 121 78 87 Milk 0.4 348 371 272 267 De vel op ing world Beef 2.6 22 44 5 7 Pork 2.7 39 81 9 13 Poul try 3.0 21 47 5 7 Meat 2.7 88 183 21 29 Milk 3.2 164 401 39 63

Sources: To tal and per cap ita produc tion for 1993 are calcu lated from FAO 1998. Projec tions are up d ated fig ures, fol low ing the same for mat as that reported in Rosegrant et al. 1997. Notes: Meat in cludes beef, pork, mut ton, goat, and poultry. Milk is cow and buf falo milk and milk prod ucts in liquid milk equiva lents. Met ric tons and kilo grams are three-year moving aver ages cen tered on the two years shown. WANA is We st Asia and North Africa . de vel oped coun tries, which pres ently have high ce- An addi tional 292 mil lion metric tons of cere als real feed use, global cereal feed use is pro jected to will be used as feed in 2020 compared to the early in crease by only 46 percent by 2020. This trans lates 1990s. For fur ther compari son, a nor mal U.S. maize to a 1.4 percent compounded an nual rate of growth, crop in the early 1990s to taled around 200 millio n noticea bly higher than the 0.7 percent an nual metric tons of grain. The increase in grain use as growth rate observed from the early 1980s to the feed will be met primar ily through expan sion of early 1990s. Cereal pro duc tion is pro jected to grow yields in the tradi tional export ing coun tries and ex - by 1.3 percent per year through 2020. pansion of culti vated area in South and East Asia.

Table 20— Pro jected trends in use of cereals as feed, 1993–2020

Projected annual To tal ce real Per cap ita ce real growth rate of total use as feed use as feed ce real use as feed Re gion 1993–2020 1993 2020 1993 2020

(per cent) (mil lion met ric tons) (kilograms)

China 3.4 84 178 62 125 Other East Asia 2.2 7 20 116 167 In dia 5.0 3 14 4 11 Other South Asia 2.9 1 4 6 7 South east Asia 2.7 12 30 32 45 Latin Americ a 2.0 55 92 116 140 WANA 2.5 29 66 93 104 Sub-Saharan Af rica 3.5 3 5 4 5 De vel op ing world 2.8 194 409 45 65 Devel oped world 0.6 442 519 346 374 World 1.4 636 928 115 120

Sources: To tal and per cap ita use for 1993 is calcu lated from FAO 1998. Projec tions are up dated fig u res fol low ing the same for mat as that in Rosegrant et al. 1997. Notes: Ce re als includes wheat, maize, rice, bar ley, sorghum, millet, rye, and oats. Met ric tons and kilo grams are three-year mov ing av er ages cen - tered on the two years shown. WANA is West Asia and North Af rica. 27

Table 21— Changes in IMPACT’s baseline as sump tions: A severe Asian crisis and high Indian meat consump tio n

Nature, mag ni tude, Phenome no n and du ra tion of Sce nario being modele d Mechanis m changed parame ters

Severe Asian crisi s En dur ing depre cia tion of Asian In ser tion of ad di tional wedge Last ing one- shot in crease in exchange rates be tween do mes tic and world wedge of 5 to 13 percent of commod ity prices for Asian world prices de pend ing on coun tries coun try

Lower nonagricultural in come De crease in ex oge nously Last ing de crease of 30 to 45 growth in Asia specified growth rates of GDP percent in GDP growth rates, for Asia depend ing on coun try

High In dian meat Shift of In dian tastes to ward Income elas tic ity for ani mal Last ing rise in elasticity to 1.5 con sump tion ani mal foods includ ing beef foods raised for India to 2.0, de pend ing on the com mod ity

Pro duc tion re sponse of India n Per ma nent in crease in the trend Fixed in ter cepts shift upward herds of In dian herd growth rates by 0.3 to 0.7 percent by com mod ity

In crease in cere al feed use in Progres sive upward shift in feed By 2020, India has ratios higher In dia con ver sion ratios (kilogram of than most other devel op ing ce re al/kilogram of ani mal food) countries, but lower than de vel oped coun tries

Source: The sce nario for a severe Asian cri sis is de scribed in Rosegrant and Ringler 1998.

Feed grain culti va tion has been of relatively minor version effi ciency over time (due perhaps to in - impor tance in South and East Asia compared to creasing use of cereal feeds as live stock pro duc tion foodgrain cul ti va tion, but maize pro duc tion is in - inten si fies). creasing rap idly while rice pro duc tion is slow ing. What Happens in the Event of a Sensitivity of Projections to Long-Run Economic Crisis in Asia? Changes in Assumptions Because de mand from devel op ing coun tries, pri - marily in East Asia, drives the Live stock Revolu - Four ex peri ments that test some of the assump tion s tion, a pro longed Asian eco nomic crisis could have in IM PACT and take into account some major a major effect on live stock trends through 2020.9 changes that might occur in the world are discussed The scenario for a more severe outcome in Asia is here: (1) a pro longed de cline in eco nomic growth in laid out in Table 21. The scenario includes an en - Asia; (2) a struc tural change in tastes in India to - during depre cia tion of Asian ex change rates and a ward in creased consump tion of milk and meat; (3) a last ing decrease in income growth rates. Depend ing broad, secular increase over time in feed conver sion on the country, price wedges be tween domes tic and effi ciency (due to techno logi cal progress, perhaps) ; world prices are increased by 5 to 13 per cent. An - and (4) a broad secular decrease in cereal feed con - nual GDP growth rates are cut by 30 to 45 percent .

9 The June 1998 base line used here al ready incor po rates less op ti mis tic as sump tions about Asian growth than pre vi ous ver sions of IMPACT. See Rosegrant and Ringler (1998) for a de tailed discus sion. The cur rent base line as su mes that long- term nonag ri cul tura l GDP will grow at 5.5 percent per year in China, 4.5 percent in Indo ne sia, 5 percent in Mala y sia, 3.5 per cent in Ko rea, 5 per cent in In - dia, and 5 per cent in Thai land. 28

This scenario, presented in de tail in Rosegrant and relative to the baseline, a neg li gi ble change con sid - Ringler (1998), is a “worst- case” pes si mis tic sce- ering it is spread over 27 years. nario, espe cially because the ef fects are as sumed to last until 2020. The results are presented here to il- What Happens if People in India lus trate the robust ness of the Live stock Revo lu tion, Dra mati cally In crease Their Meat even in the face of a last ing slow down in Asia. Con sump tion? A compari son of the results of the severe Asian scenario to those of the base line model are given in Another major struc tural change that could affec t Table 22. A severe Asian crisis would cause meat, the global live stock econ omy is a shift in Indian milk, and feed consump tion to fall, most no ta bly in tastes that accel er ates meat consump tion (Bhalla, China and India. But even where consump tion de - Ha zell, and Kerr 1998). Ur bani za tion, in come clines by 20 percent rela tive to the baseline, the growth, and secu lari za tion could cause people in base line pro jec tions of 200–300 percent growth for India to increase their meat consump tion at a rate meat and milk in India and China by 2020 mean that simi lar to high milk consump tion growth rates. con sump tion will grow 160–240 percent in those Total milk consump tion in India grew by 53 percent countries, still a large increase. between the early 1980s and 1990s (Table 6). Even with a severe Asian cri sis, the prospects This sce nario as sumes that In dian income elas - for long-term live stock expan sion in Asia and the ticities increase to be tween 1.5 and 2.0, de pend ing world are quite robust. In Latin America and Sub- on the product (Table 21). It also as sumes that In - Saharan Africa the severe Asian scenario increase s dian pro duc tion sys tems change in tan dem, with meat and milk consump tion slightly because of per ma nent in creases of 0.3–0.7 percent in the trend price effects (cheaper world prices for meat and of In dian herd- size growth rates. Finally, the sce - feed). For the world as a whole, consump tion of nario as sumes that live stock pro duc tion in land- live stock com modi ties de clines by 8 percent or less scarce India in creases primar ily through in ten si fi -

Table 22— Difference between baseline projections and pro jec tions of ag gre gate consump tion in 2020 due to changes in assump tions about the Asian crisis and Indian consumption

Severe Asian cri sis High Indian meat con sump tion Beef and Pork and Ce real Beef and Pork and Ce real Re gion mut ton poul try Milk used as feed mut ton poul try Milk used as feed

(per cent differ ence be tween new sce nario and base line)

China –20 –19 –18 –13 – 3 – 2 – 7 4

In dia –14 –17 – 27 –16 343 505 154 250

Other Asia (includ ing WANA) –8 –13 – 5 –1 – 4 – 3 – 9 7

Latin Amer ica 1 4 2 – 5 – 3 – 4 – 4 8

Sub-Saharan Af rica 2 4 4 – 4 – 4 – 2 – 8 5

De vel op ing world – 6 –13 –13 – 8 33 6 59 14

Devel oped world – 2 2 0 – 2 0 0 – 2 3

World – 4 – 8 – 8 – 5 19 4 34 8

Notes: The severe Asian cri sis in cor po rates lower projected income growth and de pre cia tion in ex change rates for coun tries in Asia. De pend ing on

the country, income growth projec tions are 30 – 45 percent lower than in the base line scenario, and exchange rate depre cia tion re sults in

5 –13 percent higher do mes tic prices. The high In dian meat con sump tion scenario in cor po rates a shift in tastes to ward con sump tion of ani mal foods and in creases in produc - tion and feed use ratios by In dian live stock produc ers. Income elastici ties for ani mal pro d ucts rise to be tween 1.5 and 2.0, de pend ing on the

commod ity. The trend growth rate of In dian herd size rises by 0.3 – 0.7 per cent (de pend ing on the com mod ity), and feed con ver sion ratio s rise above those typical in de vel op ing countries . Mut ton includes sheep and goat meat. Milk includes all dairy prod ucts in liq uid milk equiva lents. The large percent age changes for meat and feed in In dia re flect a low ini tial base. Projec t ed beef con sump tion in 2020, for exam ple, is still only 31 kilo grams per cap ita under the high meat scenario, less than in China today. WAN A is West Asia and North Africa . 29 cation and that the amount of ce real feed require d assumed conver sion ratios in ei ther direc tion. The per kilo gram of meat rises as a conse quence . base line scenario basi cally as sumes that these two The high In dian meat consump tion sce nario ef fects can cel each other out. pulls in the oppo site di rec tion from the severe Asian In the first, “op ti mis tic” scenario, feed con ver - crisis scenario. The effect on the growth of India n sion ratios improve pro gres sively. About 1 percen t con sump tion is huge in both abso lute and percent - more meat per year is pro duced per kilo gram of age terms, but still rela tively mod est in per cap ita feed in devel op ing coun tries, with a compounded terms (Table 22). Milk consump tion and cereal feed ef fect over time. This mir rors the rapid tech no logi - use rise the most in abso lute terms. For the world as cal prog ress observed in the de vel oped countrie s a whole, milk consump tion in 2020 would exceed from the early 1980s to the early 1990s. Under the the baseline by 34 percent and feed use by 8 per cent. opti mis tic scenario, feed conver sion effi ciency in de vel oped coun tries improves by only 0.5 percen t Scenarios with Different Rates of Feed per year. This lower rate reflects less scope for Conversion Efficiency adopting tech nol ogy from other regions. The re sult Feed can make up to three-quarters of the variable is that in 2020 devel op ing coun tries pro duce ap- cost of live stock pro duc tion in in ten sive sys tems proxi mately 60 percent more meat per kilo gram of (Sere and Stein feld 1996). Changes in the feed feed than they do in the baseline pro jec tions. The requirements of meat and milk pro duc tion play an ef fect is half as large in de vel oped coun tries. im por tant role in de ter min ing the final cost of live - In the second, “pessi mis tic” scenario, feed con - stock pro duc tion. Techno logi cal inno va tions that version ratios worsen pro gres sively. About 1 per - in crease the amount of meat obtained per unit of cent less meat per year is produced per kilo gram of high energy feed and inten si fi ca tion of pro duc tion feed in de vel op ing coun tries, with the effect com - (substi tu tion of ce re als for browse or slops) both pound ing over time. This scenario reflects in creas - can change the amount of ce real used per unit of ing shifts from backyard pro duc tion to more inten - meat and milk pro duced. Two op po site scenar ios sive feed ing in lots. Because in ten si fied feed ing are pre sented: an increase in the rate that feed use practices have already taken hold in devel oped effi ciency improves and a decrease in that rate. countries for the most part, the wors en ing con ver - These are both modeled in somewhat extreme form sion ratio for de vel oped coun tries is as sumed to be to test the sen si tiv ity of re sults to changes in the only 0.5 percent per year. The result is that in 2020

Table 23— Difference between baseline projections and pro jec tions of ag gre gate consump tion in 2020 due to changes in assump tions about feed conversion

Increas ing feed con ver sion ef fi ciency De creas ing feed conver sion effi cienc y Beef and Pork and Ce real Beef and Pork and Ce real Region mut ton poultry Milk used as feed mut ton poul try Milk used as feed

(per cent differ ence be tween new sce nario and baseline) China 0 0 0 –15 0 0 0 17

In dia 0 1 0 – 9 0 0 –1 11 Other Asia (includ ing WANA) 0 1 1 –10 –1 0 –1 12 Latin Amer ica 1 1 0 –15 0 –1 –1 19 Sub- Saharan Af rica 0 0 0 –12 0 0 0 14 Devel op ing world 0 1 0 –13 0 0 –1 16

De vel oped world 0 0 0 – 2 0 0 0 2

World 0 0 0 – 7 0 0 0 8

Notes: The increas ing feed conver sion ef fi ciency sce nario as sumes an in crease in the ef fi ciency of con vert ing maize to meat. This re sults in a decreased rate of in crease for con ver sion ratios in re gions where they are in creas ing, and an increased rate of de crease for con ver sion ratios in re gions where they are decreas ing . The decreas ing feed con ver sion ef fi ciency sce nario as sumes a de crease in the ef fi cienc y of con vert ing maize to meat. This re sults in an increased rate of in crease for con ver sion ratios in re gions where they are in creas ing, and a de creased rate of de crease for conver sion ratios in re gions where they are decreas ing . Mut ton includes sheep and goat meat. Milk includes all dairy prod ucts in liq uid milk equiva lents. WANA is West Asia and North Af rica. 30 devel op ing coun tries pro duce approxi mately 60 Pro gres sively lower feed effi ciency raises the percent less meat per kilo gram of feed than they do world price of ce real feed to the point where substi - in the base line pro jec tions. The ef fect is half as tutes become cost- effective. Produc tion shifts from large in de vel oped coun tries. regions and commodi ties that cannot under take these Table 23 shows the effects of the opti mis tic and sub sti tu tions at low enough cost to those that can. pes si mis tic feed con ver sion scenar ios on consump - Coun tries such as Ar gen tina, with large produc tion tion in 2020. The most striking result is that large capac ity on ranges, are fa vored in the “pes si mis tic” changes in ei ther di rec tion have vir tu ally no impact scenario because they can compete more eas ily, with on live stock product con sump tion. Large changes, more expen sive lot-fed beef, for exam ple. The reverse however, are asso ci ated with differ ences in the is true under the “opti mis tic” sce nario. Higher feed amount of cereal used to produce live stock prod - effi ciency (lower conver sion ratios) tends to encour - ucts. Al though not shown, changes in feed con ver - age in creased use of ce re als as feed and favor those sion ra tios also cause changes in the loca tion of live- countries where ce real supply is relatively cheap and stock pro duc tion in a competi tive market system. ce real feed ing practices are well- established. A crucial point to note in Table 23 is that the amount A striking re sult is that as long as ce re als are of cereal used for feed in 2020 changes by only avail able world wide in rela tively elas tic sup ply, feed 2 percent in devel oped countries and 13 to 16 per - con ver sion does not play a critical element in de ter - cent in devel op ing coun tries, relative to the base line. mining human con sump tion of livestock prod ucts, The change is small because increased feed ef fi - although it does affect the com peti tive ness of indi - ciency leads to falling cereal feed prices, which vidual pro duc ers of live stock product. Feed effi - encourages substi tu tion back into cereal feeds ciency and relative cereal prices, there fore, are likely among all produc ers, lessen ing the overall effect of to be quite im por tant in de ter min ing the geo graphic changes in ef fi ciency. di rec tion of trade. 5. Implications of the Livestock Revolution for World Trade and Food Prices

An emerging, central fea ture of world mar kets for trade was minimal in the early 1990s. Roughly 0.4 meat, milk, and feed is that they are increas ingly in- mil lion metric tons flowed on a net basis from the ter linked. IM PACT im plic itly takes this inter link - de vel oped to the devel op ing world. WANA im - age into ac count with its an nual market-clearin g ported the most, while Latin America ac counted for feature, which al lows prices to move until sup ply 18 percent of net world ex ports of beef. Devel ope d and de mand are in balance. Domes tic mar kets are countries imported a net total of 0.7 mil lion metric cleared in the model through back ward and forward tons of pork from de vel op ing coun tries in the early it era tions between sets of domes tic and world 1990s, much of it from China. De vel oped countries prices. World prices dif fer from domes tic ones by had net ex ports of 0.5 mil lion metric tons of poul - means of fixed price wedges speci fied for each try, almost 19 mil lion metric tons of milk, and a lit - coun try group that catch the ef fect of protec tion is t tle more than 93 mil lion metric tons of ce re als. poli cies or major trans port costs to remote markets. Beef is pro jected to become the most sig nifi - Do mes tic prices in equilib rium may al ways be cant meat im port of devel op ing coun tries in 2020, above or below world prices for meat and feed, de- at 2.7 mil lion metric tons net. Pork will remain a net pending on the coun try group in question, but they export of de vel op ing coun tries, though only mar- are al ways af fected in the model by movements in ginally. Poul try im ports will rise to 1.8 mil lion met - world prices. This in ter linked move ment of prices ric tons. Milk will almost double in size as a net ex - is a better ap proxi ma tion of re al ity than delinked port of the devel oped world. Cere als will con tinue mar kets. In IMPACT, changes in de mand for live - to be the most signifi cant net agricultural expor t stock prod ucts in Asia, for exam ple, affect live stock from the de vel oped to the devel op ing world. Com - and ce real feed prices every where. No part of the pared to 1993, net ex ports from de vel oped to devel - world will be unaf fected by the events in Asia or op ing coun tries are pro jected to rise by 133 million India or changes in feed use ef fi ciency. metric tons, an amount equiva lent to ap proxi mately Trade flows are mod eled in IMPACT as the net 60 percent of the en tire U.S. aver age corn crop in an nual resid ual between produc tion and consump - the early 1990s (FAO 1998). tion at world market- clearing prices for each coun try group and com mod ity, leading to either net ex ports or net imports (negative net exports) for that com - Changes in Net Exports under modity and country group in the year in question . Changed Asian Assumptions The model does not de ter mine trading partners, but the world wide sum of net exports for a given com - The live stock and feed trade picture in 2020 modity is zero for each year in equilib rium. changes sig nifi cantly under the severe Asian cri sis and high Indian meat con sump tion scenar ios Projected Trade in Livestock Food (Table 25). The extreme as sump tions in the Asian Products and Feed to 2020 crisis scenario do not change ag gre gate live stock consump tion by much. However, they do promot e Net ex ports of specific com modi ties by region are large changes in pro jected world cereal feed flows. given in Table 24, for 1993 and 2020. World beef The changes in feed flows are in fact the most sig -

31 32

Table 24— Net ex ports of vari ous live stock prod ucts in 1993 and 2020, baseline scenario

Beef Pork Poul try Milk Cere al s Region 1993 2020 1993 2020 1993 2020 1993 2020 1993 2020

(million met ric tons)

China 0.1 0 .1 0 .7 0 .3 0 .1 0 .3 – 0 .9 0 .5 –0.9 –46 .2

Other East Asia – 0.4 – 0 .6 0 .0 – 0 .1 – 0 .3 – 0 .6 – 0 .5 – 0 .4 –20.0 –31.9

In dia 0 .1 0 .1 0 .0 – 0 .1 0 .0 0 .0 0 .0 0 .1 1 .3 –7.1

Other South Asia 0.0 – 0 .3 0 .0 0 .0 0 .0 – 0 .3 – 0 .6 – 2 .8 –4 .5 –21.9

South east Asia – 0.2 – 0 .6 0 .2 0 .6 0 .1 0 .5 – 4 .0 –7.8 –3.2 –5.4

Latin Americ a 0.7 0 .6 – 0 .2 – 0 .2 0 .1 – 0 .1 – 5 .7 – 5 .6 –16 .0 –13 .0

WANA – 0.7 –1.7 0 .0 0 .0 – 0 .4 –1.3 – 5 .0 –12.2 –37.7 –74.6

Sub-Saharan Af rica 0.0 – 0 .2 0 .0 – 0 .1 – 0 .1 – 0 .2 –1.5 –1.8 –11.6 –24.9

De vel op ing world – 0.4 – 2 .7 0 .7 0 .2 – 0 .5 –1.8 –18 .7 – 30.9 –93 .3 –226.1

Devel oped world 0.4 2 .7 – 0 .7 – 0 .2 0 .5 1 .8 18.7 30.9 93.3 226 .1

Source: Projec tions are from IMPACT. Notes: Net trade in 2020 is projected produc tion mi nus con sump tion for the com mod ity and region s hown. Met ric tons are three-year mov ing av er - ages cen tered on the two years shown and, for meat, re fer to carcass weight. Milk is cow and bu ffalo milk and milk prod ucts in liq uid milk equiva lents. Cere als in clude wheat, maize, rice, bar ley, sorghum, millet, rye, and oats. Net ex port fig ures may not sum to zero over all be - cause of rounding. WANA is West Asia and North Af rica. Minus signs indi cate im ports.

Table 25— Difference between baseline pro jec tions of net ex ports in 2020 and projections with changes in assump tions about the Asian crisis and Indian consumption

Severe Asian cri sis High Indian meat con sump tion Beef and Pork and Beef and Pork and Re gion mut ton poul try Milk Cere al s mut ton poul try Milk Cere al s

(per cent differ ence between new scenario and base line)

China 211 556 62 27 169 612 321 – 2

In dia 86 53 18,857 19 – 4,038 – 2,719 – 55,464 – 480 Other Asia (includ ing WANA) 21 293 13 –1 28 66 30 4

Latin Amer ica –172 – 527 –55 4 168 176 95 – 2

Sub-Saharan Af rica –112 –119 – 90 –15 141 24 157 33

De vel op ing world 2.9 261 43 4 – 47 4 – 83 –10

Devel oped world – 2.9 – 261 – 43 – 4 47 – 4 83 10

Notes: The severe Asian cri sis in cor po rates lower projected income growth and de pre cia tion in ex change rates for coun tries in Asia. De pend ing on

the country, income growth projec tions are 30 – 45 percent lower than in the base line scenario, and exchange rate depre cia tion re sults in

commod ity. The trend growth rate of In dian herd size rises by 0.3 – 0.7 per cent (de pend ing on the com mod ity), and feed con ver sion ratio s rise above those typical in de vel op ing countries . Mut ton includes sheep and goat meat and edible prod ucts. Milk includes all dairy prod ucts i n liq uid milk equivalents . The large percent age changes for meat and feed in In dia re flect a low ini tial base. WANA is West Asia and North Africa . 33 nificant result of the severe Asian crisis as sump - of the unre al is tic as sump tions of the high In dian tions, espe cially because the abso lute amounts of sce nario, the world sys tem is flexible enough to ad - feed traded in the baseline scenario are alread y just with out major dislo ca tions in consump tion . large. Baseline live stock trade is rela tively mod est, which should be borne in mind when inter pret in g the large percent age changes shown for live stock Changes in Net Exports under items in most cases. Changed Assumptions about Geographi cally the severe Asian crisis scenario Feed Efficiency sharply decreases pro jected net im ports of live stock foods and decreases pro jected net ex ports from The previ ous chapter assessed the sensi tiv ity of Latin America and Sub-Saharan Af rica.10 Pro jec - base line results to assump tions about feed con ver - tions of net im ports of feedgrains by Asia also de - sion ratios and found that consump tion pat terns of cline ap pre cia bly, by as much as 27 percent in livestock prod ucts did not change much, but cerea l China. Net ce real feed im ports by de vel op ing coun - use as feed rose or fell depend ing on whether ef fi - tries de cline by 4 per cent. Net Chinese meat ex ports ciency decreased or in creased. The results are dif- rise by 3.4 mil lion metric tons and milk ex ports rise fer ent for trade, as shown by the wide geographic by 0.3 mil lion metric tons. varia tion in trade pro jec tions rela tive to the base line For Asia other than China and India, the severe (Table 26). Asian crisis sce nario changes the region from a net An equal percent age increase in feed con ver sion meat im porter in 2020 to a net ex porter of 3.2 mil - ef fi ciency ap plied to all devel op ing countries leads lion metric tons. The re gion’s net milk im ports de- to differ ing results in differ ent countries, includ ing a crease by 5.3 mil lion metric tons. The ef fects of the 25 per cent in crease in net exports of beef from India Asian crisis decrease domes tic de mand and in - and a 9 per cent de cline in net beef exports from crease competi tive ness in world markets. India re - China. The price changes that re sult from de creas ing mains a minor par tici pant in world trade of live - feed conver sion ef fi ciency work to in crease net ex - stock prod ucts in 2020 under the base line and Asian ports of cereal feed from the de vel oped coun tries, crisis scenar ios. Coun tries out side Asia tend to in - India, and Sub- Saharan Af rica, and in crease net im - crease their im ports of live stock prod ucts under se- ports of cereal feed by China and Latin Amer ica. The vere Asian crisis assump tions because of lower small percent age of trade in total baseline produc tion world prices. and con sump tion causes the large percentage trade The high In dian meat (and milk) con sump tion responses seen in some cases. scenario turns India into a major world im porter of meat, milk, and ce real feed in 2020, a substan tia l shift from its minor trad ing role in the base line pro - Past Trends in Real World Prices jec tions (Table 25). Other re gions of the world in - crease their net ex ports of meat and milk, al though for Crop and Livestock Products China and Latin America ac tu ally margin ally in- and Projections to 2020 under crease their im ports of ce real feed in re sponse to the Different Scenarios expanded meat and milk ex port oppor tu ni ties. The very large percent age changes for India are caused Debates over the use of cere als as feed need to be by the very low pro jected trade lev els in the base - cast in the context of steadily lower re turns to feed- line projec tion. Base line net ex ports of 77,000 met - ing cat tle for meat pro duc tion over the past 25 ric tons of liquid milk equivalents in 2020 turns into years. Real world ag ri cul tural prices have slipped 42.6 mil lion metric tons of net im ports in the high stead ily relative to manu fac tured goods prices since Indian meat consump tion scenario. Even in the face the early 1970s (Table 27). Clearly, live stock

10 The con cept of “net ex ports” could equally well be ex pressed in this case as increas ing net e x ports from Asia or decreas ing net imports by Latin Amer ica and Sub- Saharan Africa . 34

Table 26— Difference between baseline pro jec tions of net ex ports in 2020 and projections with changes in assump tions about feed conversion efficiency

Increas ing feed conver sion ef fi ciency Decreas ing feed con ver sion ef fi ciency Beef and Pork and Beef and Pork and Region mut tonc poultry Milk Cere al s mut ton poul try Milk Cere al s

(per cent differ ence between new scenario and base line)

China – 9 16 – 6 46 7 – 21 5 – 53

In dia 25 2 –1,075 – 32 – 26 – 2 1,075 36

Other Asia (includ ing WANA) –1 – 3 8 5 1 3 – 9 – 6

Latin Amer ica – 5 –14 1 26 6 14 –1 – 40

Sub-Saharan Af rica – 3 8 46 – 31 1 –11 – 50 34

De vel op ing world – 2 2 6 9 2 – 3 –7 –12

Devel oped world 2 – 2 – 6 – 9 – 2 3 7 12

Table 27— Past trends in real prices of selected crop, feed, and live stock products

Year Wheat Rice Maize Soy beans Soymea l Fish meal Beef Pork Poul try Lamb Milk

(con stant 1990 US$/met ric ton)

1970–72 232 524 215 476 415 750 5,144 n.a. n.a. 3,248 485 1980–82 236 534 169 384 338 615 3,536 2,344 1,474 3,730 413 1990–92 135 288 104 234 195 444 2,585 1,781 1,139 2,440 280 1994–96 156 270 116 238 192 424 1,761 n.a. 1,113 2,474 261 World Bank pro jec tions 2000 135 279 102 230 189 n.a. 1,773 n.a. n.a. n.a. n.a. 2010 118 262 92 236 196 n.a. 1,629 n.a. n.a. n.a. n.a.

Sources:Past data are from USDA 1997a and b, IMF 1997, and World Bank 1993. World Bank projec tions and th e Manufac tur ing Unit Value in dex used for ex press ing val ues in constant 1990 U.S. dollars are from World Bank 1997. Notes: Wheat is U.S. no. 1, hard red win ter, ordi nary pro tein, ex port price deliv ered at Gulf ports for ship ment within 30 days. Rice is Thai 5 percent broken, WR, milled, in dica tive survey price, govern ment standard, f.o.b. Bang kok. Maiz e is U.S. no. 2, yel low, f.o.b. U.S. Gulf ports.

Soybeans are U.S. c.i.f. Rotter dam. Soy meal is any origin, Ar gen tine 45 – 46 percent ex trac tion, c.i.f. Rot ter dam, prior to 1990, U.S. 44 per-

cent. Fish meal is any origin, 64 – 65 percent, c.i.f. Ham burg, n.f.s. Beef is Aus tra lian/New Zea land, cow fore quar ters, fro zen boneless, 85 percent chemical lean, c.i.f. U.S. port (East Coast), ex- dock. Pork is European Com mu nity pork, slaugh ter wholesale price. Poul try is broil ers, twelve- city com pos ite wholesale price, ready- to- cook, de liv ered. Lamb is New Zea land, fro zen whole car casses, wholesale price, Smith field mar ket, London. Milk is U.S. whole milk sold to plants and deal ers, U.S. Depart me n t of Agri cul ture. n.a. in di cates that com pa - ra ble prices for those years are not available .

producers have felt the pain in recent years. Beef in 1980s and 1990s (Table 15) occurred at a time 1994–96 was priced at only 34 percent of its when maize and soy beans declined in price by more inflation- adjusted level in 1970–72, while maize than one- third. This lends sup port to the ar gu ment was 54 percent of its inflation- adjusted price in the that addi tional concen trate feeds will be sup plied same pe riod. with lit tle real increase in price. This ar gu ment is Further more, the 10 percent global expan sio n tested by the IM PACT model, which produces a se - of cereal feed use by weight be tween the early ries of market- clearing world prices in equilib riu m 35

(Table 28). For compari son, single- equation price a severe eco nomic crisis in Asia, but in the op po site projec tions to 2010 by the World Bank are reporte d di rec tion. at the bottom of Table 27. Increas ing feed conver sion effi ciency low ers Com pared to the World Bank projec tions to real world maize prices by 17 percent rela tive to 2010, the IMPACT baseline projec tions show 38 base line pro jec tions for 2020. Decreas ing feed con- percent higher maize prices and 3.5 per cent higher version effi ciency increases maize prices by 21 per- beef prices. The global food supply and demand ap - cent. Live stock prices, on the other hand, are hardly proach of IMPACT cap tures the effect of the Live - affected. Ce real produc ers and con sum ers clearly stock Revolu tion on cere als prices, whereas single- have a lot at stake in feed conver sion effi ciency . equation method ol ogy does not. Even so, IMPAC T Live stock produc ers in in di vid ual coun tries have a proj ects that real world maize prices in 2010 will be lot at stake as well, even if technol ogy is shared by only about 10 per cent above actual 1994–96 prices, all pro duc ers, because increas ing feed effi ciency fa - which were histori cally low. Thus, the Live stock vors some produc ers while decreased feed ef fi - Revo lu tion pre vents cereal prices in the baseline pro - ciency fa vors others. Live stock con sum ers have jections from falling further from their currently low sur pris ingly lit tle at stake, at least as far as the cost lev els, and per haps even increases them slightly, but of live stock food prod ucts is con cerned. nowhere near their level in the early 1980s. The severe Asian crisis assump tions de crease real world prices in 2020 by 7 percent for maize and Why Is the Livestock Revolution 5 percent for beef. These are signifi cant ef fects, but Not Likely to Raise World Cereal the assump tions for this scenario are extreme be - Prices Significantly? cause they posit that the crisis lasts though 2020. The high In dian meat scenario increases world Com pari son of the IM PACT price pro jec tions maize prices by 13 percent and world beef prices by under vari ous scenar ios with the histori cal rate of 9 per cent. Chang ing tastes in India have even decrease of real food prices in the last quarter cen - greater ef fects on world live stock mar kets than does tury sug gests that the rate of de cline of food prices

Table 28— Real prices of selected crop and live stock prod ucts as projected by the IM PACT model

Year Wheat Rice Maize Soybean s Beef Pork Poultry Lamb Milk

(con stant 1990 US$/met ric ton) Base line prices 1992–94 148 275 126 263 2,023 1,366 1,300 2,032 234 Base line pro jec tions 2010 146 293 127 244 1,835 1,260 1,175 1,915 217 2020 133 252 123 234 1,768 1,209 1,157 1,842 199 Se vere Asian cri sis scenario pro jec tions 2020 124 248 114 221 1,676 1,104 1,074 1,807 187 High In dian meat con sump tion sce nario pro jec tions 2020 148 268 139 267 1,927 1,287 1,259 2,203 219 In creas ing feed con ver sion ef fi ciency scenario pro jec tions 2020 126 243 102 228 1,738 1,188 1,134 1,817 196 De creas ing feed conver sion ef fi ciency scenario pro jec tions 2020 141 262 149 242 1,802 1,233 1,183 1,870 202

Sources:The up dated IM PACT baseline pro jec tions and the se vere Asian cri sis sce nario are from Rose grant et al. 1997 and Rosegrant and Ringler 1998. Notes: The sce nar ios are inde pend ent of each other. The IM PACT baseline prices in 1992–94 are fully com pa ra ble to the ac tual series given in Ta ble 27. 36 during the next quarter cen tury is likely to be much come about to a large extent in response to the higher weaker than in the past. This is un doubt edly a result cereal prices that made in vest ing in produc tion inno - of the Live stock Revolu tion. Real ce real prices, va tions eco nomic (Hayami and Ruttan 1985). however, are not likely to rise very much by 2020, Another expla na tion of the high sup ply re spon - contrary to the fears of some reported in Chapter 1. sive ness of ce re als is that in a sys tem of global mar - The pri mary expla na tion for IM PACT’s pro- kets, where ac tors all over the world respond to jection that cereal prices will con tinue to be low changing price in cen tives, in di vid ual shocks are even as de mand for both food and feed in creases is smoothed out over time through myriad adjust - that the world is thought to have con sid er able re - ments through out the system. In other words, world serve capac ity for addi tional ce real produc tion. supply will be more price- responsive than in di vid - This assump tion not only fits the evidence of the ual coun try supply, a phenome non well- captured in past 25 years, during which world pro duc tion out - IM PACT. put changed in re sponse to price, but also accord s A final phenome non of di rect relevance to ce- with the dozens of price re sponse pa rame ters for cereal sup ply is that in areas contain ing a signifi can t reals (for 37 coun tries and re gions times a half share of the world’s poor, the rise in the consump - dozen crops) built into the model and mostly taken tion of calories from ani mal prod ucts is matched by from in de pend ent sources (Rosegrant, Agcaoili- a decrease in calories from starchy sta ples. People Sombilla, and Perez 1995; Pinstrup- Andersen, in China’s rural areas, for exam ple, con sumed Pandya-Lorch, and Rosegrant 1997). nearly 200 kilo grams per cap ita of grain di rectly as The tra di tional expla na tion for the high suppl y food in 1991, while people in the richer urban areas respon sive ness of cere als found in the em piri cal lit - consumed an av er age of 130 kilo grams per capita, a erature is that the large grain-exporting coun tries, fig ure much closer to av er age consump tion level s such as Aus tra lia, Canada, and the United States, elsewhere in the de vel op ing world (Huang and have the abil ity to bring large amounts of land typi - Bouis 1996). Rice and wheat are the grains typi - cally not sown to grain into cereal produc tion as cally con sumed in China. These grains have sig nifi - prices rise. Further more, cereal cul ti va tion by the cantly lower av er age yields per hectare than maize, large grain export ers usually is less input- intensive a feed grain increas ingly culti vated by Chinese than in many im port ing coun tries. Av er age wheat farm ers. Thus, sub sti tu tion of meat and milk for yields in Britain, for ex am ple, are much higher than grain in the diet liber ates some grain from direc t in the United States. This leaves con sid er able scope consump tion as food, and the conse quent increase for ex port ing countries to in crease produc tion in ag gre gate grain supply is ampli fied by the rela - through higher fer til izer use, which becomes profit - tively higher yields of feedgrains per hec tare. All of able when cereal prices are high. The tremen dou s this sug gests that the Live stock Revo lu tion is inti - produc tiv ity increases in ce re als in re cent years also mately wrapped up with nutri tional and food secu - at test to the spur of prices. These in creases have rity, the subject of the next chapter . 6. Nutri tion, Food Se cu rity, and Pov erty Alle via tio n

The trends noted in the pre vi ous chapters strongly “anti-livestock” po si tions in policy mak ing. But sug gest that live stock consump tion pat terns in de - whether the tra di tional livestock en ter prises of the velop ing countries are rapidly con verg ing with rural poor can co ex ist with in creased indus tri ali za - those in de vel oped coun tries, put ting severe pres - tion of live stock produc tion is an open question. Fi - sure on pro duc tion systems in devel op ing countrie s nally, in creas ing inten si fi ca tion of produc tion raises to do the same. The con sid er able con tro versy in other ethical issues, in clud ing con cerns about ani mal both the popular and scien tific lit era ture about the wel fare and the un pleas ant ness suffered by those de sir abil ity of live stock consump tion and pro duc - who live downwind from major indus trial hog farms. tion pat terns in the richer coun tries raises concern s As with other struc tural changes in the food about what the Live stock Revo lu tion por tends for sec tor that are im por tant to human wel fare, such as human wel fare in the poorer countries. It seems the Green Revo lu tion in ce re als, popular dis cus sion prob able, for ex am ple, that mas sively larger of the effect of the Live stock Revo lu tion on devel - amounts of cere als will be used as feed to pro duce op ing coun tries is oc ca sion ally highly emo tional items con sumed primar ily by better- off urban peo - and often imper fectly grounded in facts. The objec - ple in countries where outright lack of food is still tive of this chapter is to show how the research pre- com mon (Brown and Kane 1994; Good land 1997; sented in the rest of this paper re lates to the discus - Pimen tel 1997). sion of wel fare issues. These concerns raise com plex nutri tional, food secu rity, and pov erty al le via tion issues. Nutri tiona l is sues in this con text include the ef fect of con sump - The Livestock Revolution and tion of specific ani mal food prod ucts on health and Nutrition in Developing Countries well-being in specific situa tions. Food se cu rity is - sues encom pass the ability of people to secur e The dis cus sion of nutri tional is sues in this section enough food on a regular basis for healthy and pro- does not ad dress the health ef fects of consum ing ductive lives. Pov erty al le via tion is sues in clude the contami nated food products. Those ef fects are dis - extent to which the pro duc tion and sale of animal cussed in Chapter 8. Even if contami na tion can be prod ucts might lead to widespread net im prove - dealt with, is increased consump tion of meat and ment in the live li hood of the rural poor. milk a good thing? Concerns about medi cal prob - The IM PACT pro jec tions shed light on likely lems clearly as so ci ated with very high meat and trends in ce real and ani mal product prices under dif - milk consump tion in socie ties such as the United fer ent scenar ios for the Live stock Revo lu tion. The States must be taken seri ously (Bar nard, Nichol son, im pact of live stock on the purchas ing power of the and How ard 1995). But sometimes these concern s poor, the other side of the cereal af forda bil ity issue, are extrapo lated di rectly to de vel op ing coun tries in must be assessed with highly dis ag gre gated data. fairly blanket fashion (Good land 1997; Pi men tel As will be seen below, evidence from house hold 1997) and sometimes se lec tively, as in the case of stud ies around the de vel op ing world sug gests that China’s cit ies (Geissler forthcom ing). in many cases the rural poor, espe cially women, get Sci en tific con clu sions about the ef fects animal a larger share of their income from livestock prod ucts can have on nutri tion in devel op ing coun - sources than do the relatively wealthy. These find - tries not only depend on the demograph ics and in - ings raise questions about the moral ity of blanket come lev els of target groups and the commodi ties

37 38 under study (CAST 1997a), but they vary also be - Authori ta tive litera ture on human nutri tion in cause of the method olo gies em ployed (Hu and Wil - the de vel op ing world stresses the widespread lett 1998). And random ized clini cal end point tri als prevalence of mild to mod er ate protein-energy mal - or other forms of in ves ti ga tion designed to con trol nu tri tion affect ing up to one- third of all childre n for spu ri ous cor re la tions when compar ing long- and perhaps a higher share of preg nant or lactat in g term “be fore” and “after” con di tions may be dif fi - women. The litera ture empha sizes the critical role cult to con duct. adequate balance and ade quate lev els of bioa vail - One often- cited ap proach to study ing the link able protein, calories, and key mi cro nu tri ents (such be tween meat consump tion and health relies on cor- as vi ta min A, iodine, and iron) play in men tal and re la tions be tween disease rates and patterns of live- physi cal devel op ment (Cal lo way 1995; Sharma et stock product consump tion across loca tions. A al. 1996; Latham 1997; Neu mann and Har ris 1999; promi nent ex am ple is a study that looks at such cor- Geissler forthcom ing). Research also shows that re la tions across 65 coun ties of rural China (Chen et ade quate, balanced nutri tion helps pre vent morbid - al. 1990). Yet while this approach raises issue s ity. Animal prod ucts are portrayed in this lit era ture quickly and exploits large amounts of data, its re li - as ex cel lent sources of absorb able forms of iron, abil ity is open to ques tion as Chen et al. (1990) zinc, vita min B 12, and re ti nol. Cer tain meats and themselves point out. The China study could not milk prod ucts are cited as variously being good fully con trol for other variables that might also ex - sources of thia min, calcium, vita min B 6, ri bo fla vin, plain the results (Hu and Willett 1998). vi ta min A, and other min er als required in higher Hu and Willet recently (1998) con ducted a thor - amounts during growth pe ri ods (Cal lo way 1995; ough re view of the rela tion ship between consump - Latham 1997; Neu mann and Har ris 1999). tion of ani mal prod ucts and the risk of chronic dis - Al though a large combi na tion of crop- based ease for the World Bank. The goal was to sort nu tri ents could also pro vide the neces sary amino through con flict ing evi dence in order to provid e acids and trace nu tri ents to meet nutri tional needs, guidance to the World Bank’s invest ment policies in secur ing such a balance through out the year with devel op ing countries. They drew in evi dence from vegeta ble matter alone is not easy for the rural poor all over the world, assum ing that while diets and nu- in devel op ing countries. On the other hand, in - tri tional needs change across en vi ron ments and work creased consump tion of even a relatively small ad - regimes, human bi ol ogy probably does not. ditional amount of meat and milk would supply the Hu and Wil lett found that red meat (beef, pork, neces sary protein and mi cro nu tri ents and a fair lamb) consump tion be yond a low thresh old level share of needed addi tional calories, espe cially to proba bly increases the risks of coronary heart dis- chil dren, and would do so with a much less var ied ease (CHD); dairy prod ucts may also do so but to a vegetable-matter diet than using crops alone lesser ex tent; eggs are probably unre lated to CHD (Latham 1997; Hu and Willet 1998; Geissler forth - up to one egg a day; and poul try and fish proba bly coming; Neu mann and Har ris 1999). decrease CHD. Mod er ate red meat in take, how ever, For all these reasons the Live stock Revolu tion may re duce the risk of hemor rhagic stroke in cases appears to have many po ten tial benefits for nutri - where initial intake lev els are very low. High lev els tion in devel op ing coun tries. At present these po - of red meat intake may increase the risk of vari ous ten tial benefits proba bly out weigh the po ten tial nu- forms of cancer, but sub sti tu tion of poul try and fish tritional costs. Whether nutri ent balance is secured for red meat proba bly reduces the risk. Dairy prod - through a wide va ri ety of crops or a smaller vari ety ucts may be a risk fac tor in prostate can cer. Hu and of crop and animal prod ucts, good nutri tion clearly Willett con clude that health policy in de vel op ing requires the overall intake of an ade quate amount of countries should distin guish poul try and fish from food. The primary nu tri tional issue for most poor beef and pork because the former proba bly pro vide peo ple in most de vel op ing countries probably re - greater nutri tional benefit than the lat ter. Hu and mains se cur ing both an adequate amount of food and Willett also see benefits from expan sion of egg and a bal anced diet. dairy consump tion where consump tion lev els are as Once a person regu larly con sumes excess food low as they are in devel op ing countries . calories, the sorts of health prob lems asso ci ate d 39 with excess consump tion of cho les terol and satu- Econo mists tend to reduce the issue to whether rated fatty acids in the de vel oped countries will increased ani mal pro duc tion increases or decrease s probably become more prevalent in de vel op ing cereal and meat and milk prices rela tive to the in - countries, espe cially among the relatively wealthy comes of poor peo ple, other things being equal. in urban areas. How ever, av er age per capita con - This ap proach has the merit of defin ing food secu - sumption of meat and milk in devel op ing countrie s rity in terms of the ability people have to purchase in 2020 are pro jected to be less than half the present food staples. But it does not address fully whether developed-country aver age (see Table 16). At lev - all members of the household have physical ac cess els that low in 2020, the major ity of people on the to food at av er age world prices or whether peo ple planet are unlikely to face the prob lem of ex cess who are sick can usefully ab sorb that food at all meat and milk consump tion for quite some time. times. Assess ing the impact of increased livestock The possi bil ity that the people who could most pro duc tion on house hold incomes and food prices is benefit from in creased meat consump tion may not an improve ment over non en ti tle ment ap proaches share in the Live stock Revo lu tion should be a and about the best that can be done here. greater concern than overcon sump tion. Whether The previ ous chapter showed that the simple in - they do benefit will depend princi pally on the evo - ference that increased live stock consump tion will lu tion of food prices and the incomes of the needy, sig nifi cantly raise cereal prices by 2020 does not top ics discussed in the fol low ing sec tion. mesh with the rather dif fer ent and com plex picture drawn by the IM PACT pro jec tions. Recall that base line, inflation- adjusted rice and wheat prices in The Livestock Revolution and 2020 are pro jected to be 8 to 10 percent below lev - Food Security of the Poor els in the first half of the 1990s. Maize prices are pro jected to de cline by only 2 per cent. Further more, Both ecologists and ani mal scien tists, un like econo - av er age world cereal prices in the mid- 1990s were mists, tend to view the ef fect of in creased anima l only about two- thirds (maize) to half (rice) their av - pro duc tion on food supplies in devel op ing coun - er age inflation-adjusted prices in the early 1980s tries as a trade-off between using ce re als (and the (Table 27). land and water used to produce them) as food or In the “worst- case” sce nario in ves ti gated by feed. Cal cu la tions are then presented to argue IM PACT, rice and wheat prices exceed their whether extra pro duc tion of ani mal prod ucts adds inflation- adjusted lev els in 1992–94 by only 8–11 or subtracts from a hy po theti cal calorie balance per cent (Table 28). The price of maize goes up 21 avail able for hu mans (CAST 1994; Pi men tel et al. percent. IM PACT also proj ects mod est long-ter m 1997; Good land 1997; Brown and Kane 1997). declin ing price trends for vari ous meats and milk. Pro-livestock authors tend to stress that ru mi - The consump tion pat terns of the rich may pre vent nant animals in devel op ing coun tries mostly use livestock prod ucts from be com ing even more af- natural grasses and other feed with little use as fordable to the poor, but they do not appear to be food. They also stress the im por tant nonfood uses likely to raise the price of live stock prod ucts much of live stock (Fitzhugh 1998; CAST forthcom - above their cur rent levels. ing). “Anti- livestock” authors empha size that In sum, food prices have been decreas ing over monogastric live stock ac count for much of the the long term, de spite rapid in creases in the use of growth of produc tion under the Live stock Revo - cereal feeds. The Live stock Revo lu tion is likely to lu tion, that these ani mals require high energ y slow this decreas ing trend somewhat by 2020. But feed such as ce re als, and that they do not fully re - the trend reverses only under extreme condi tions, place the ce real calories used to produce them. and even if it does, the price increases are expecte d These authors also argue that because live stock to be small rela tive to the real price de clines seen food product consump tion increases rap idly with since the early 1980s. The Live stock Revolu tion’s in come, the rich proba bly will bid away the food- ef fect on the food se cu rity of poor peo ple, through stuffs of the poor in the market place (Pi men tel cereal prices, is likely to be far less im por tant than 1997; Brown and Kane 1997). its ef fect on the income of the poor. 40

The Livestock Revolution and adequacy from all over the world are pre sented in Table 29. In Paki stan in the late 1980s, for ex am ple, Incomes of the Poor Adams and He (1995) found that about 25 per cent of Live stock are central to the live li hood of the rural the in come of the poor est 20 per cent of rural house - poor in devel op ing coun tries in at least six ways holds in their sample came from livestock. The rich - (Live stock in Devel op ment 1998). First, they are est 20 per cent re ceived only 9 per cent of their in- an im por tant source of cash in come. Second, they come from live stock. In Egypt in the mid-1970s, are one of the few as sets available to the poor, es pe - Fitch and So liman (1983) found that an av er age of cially poor women. Third, live stock manure and 63 per cent of the in come of landless or near landless draft power are vital to the pres er va tion of soil fer - households came from live stock. Only 14 per cent of til ity and the sustain able inten si fi ca tion of farm ing the in come of large landown ers came from live stock. sys tems in many de vel op ing areas facing increas ing Von Braun and Pandya-Lorch (1991) iden tify four popula tion den sity. Fourth, live stock allow the poor coun tries where the mal nour ished get more of their to exploit com mon prop erty resources, such as open in comes from live stock than those who are not mal- graz ing areas, in order to earn in come. Fifth, live - nour ished. Vosti, Wit cover, and Carpen tier (1998) stock prod ucts enable farmers to diver sify in comes, show an excep tion to this trend in their study of Bra - help ing to reduce income vari abil ity, espe cially in zil. In parts of Latin Amer ica, such as the Ama zon semi arid systems charac ter ized by one cropping and Argen tine Pam pas, success ful animal grazing re- season per year. Sixth, live stock pro vide a vital and quires control of large amounts of land. often the only source of income for the poor est and Poor people have few oppor tu ni ties to in crease most mar ginal of the rural poor, such as pas tor al ists, their incomes because of lim ited access to land and share crop pers, and wid ows. capi tal. Small-scale and backyard live stock pro duc - A broad vari ety of an ec do tal evidence from case tion enables the poor to earn income from ani mals studies in Af rica, Asia, and parts of Latin Amer ica grazed on com mon prop erty pas tures or fed house- shows that the poor and landless derive a higher hold waste. Live stock pro duc tion of fers one of the share of their house hold in come from live stock few rap idly growing mar kets that poor, rural peo ple sources than do the rela tively better-off in the same can join even if they lack substan tial amounts of rural commu ni ties. Es ti mates of the share of house - land, training, and capi tal. hold in come coming from livestock for house holds The im por tance of live stock for women’s in - with differ ent in come lev els, farm sizes, and die tary comes in devel op ing coun tries has been widely em -

Table 29— The place of live stock in the income of the rich and poor

Per cent of house hold in- Wealth/pov erty come from Pe riod/size Country in di ca tor Stra tum livestock of sample Source

Bra zil (Western Amazon ) Household in come Lowest 1/5 37 a 1994, 154 ru ral Vosti, Wit cover, and a stra tum High est 1/5 64 households Car pen tier 1998

Ethio pia Household income Very poor 6 1988– 89, 550 rural Webb and von Braun stra tum Poor 24 households 1994 b Kenya Household income Lowest 1/5 61 1998, 310 dairy Staal and Balten wec k b stra tum High est 1/5 38 farm ers 1998

Paki stan Household income Lowest 1/5 25 1986– 89, 727 ru ral Ad ams and He 1995 stra tum High est 1/5 9 households c Phil ip pines Household income Lowest 1/5 23 1984–85, 500 ru ral Bouis 1991 stra tum High est 1/5 10 c household s (contin ued) 41

Table 29—Continued

Percent of house hold in- Wealth/pov erty come from Pe riod/size Country indi ca tor Stratum live stock of sample Source

Sene gal Kelly et al. 1993 Sahe lian (driest) zone Household income Lowest 1/3 24 1988– 90, 29 ru ral stra tum High est 1/3 14 household s

Su da nian (tran si tion) zone Household income Lowest 1/3 10 1988– 90, 58– 67 ru ral stra tum High est 1/3 8 household s

Guinean (forested) zone Household income Lowest 1/3 6 1988– 90, 92– 102 ru ral stra tum High est 1/3 6 household s

Su dan Household income Lowest 1/5 14 1989, 240 rural Teklu, von Braun, and stra tum High est 1/5 13 household s Zaki 1991

Egypt Land hold ings Landless or near 63 1976–77, 165 Fitch and Soliman landless households 1983 Largest landholders 14

India (Andhra Pradesh Land hold ings Lowest 1/5 of land 5 1997, 699 households Kerr 1998 and Maharash tra) distribution Highest 1/5 of land 6 distribution India (Andhra Pradesh Land hold ings Landless 7 1975–78, 240 Singh, Asokan, and and Maharash tra) Largest landholders 15 households Walker 1982

Mo zam bique Mozambique Monapo Landholdings Lowest 1/4 of land 2 1991, 343 smallholder MOA/MSU/UA distribution households Research Team 1992 Highest 1/4 of land 5 distribution

Ribaue Landholdings Lowest 1/4 of land 6 distribution Highest 1/4 of land 5 distribution

Angoche Landholdings Lowest 1/4 of land 3 distribution Highest 1/4 of land 2 distribution Pakistan Landholdings Landless 14 1986–89, 727 Adams and He 1995 Largest landholders 11 households

Brazil Dietary adequacy Malnourished 32 1984, 384 rural von Braun and Not malnourished 27 households Pandya-Lorch 1991

Pakistan Dietary adequacy Malnourished 16 1986–87, 1,082 rural von Braun and Not malnourished 14 households Pandya-Lorch 1991 Philippines Dietary adequacy Malnourished 10 1983–84, 792 rural von Braun and Not malnourished 9 households Pandya-Lorch 1991

Sri Lanka Dietary adequacy Malnourished 4 1984, 480 rural von Braun and Not malnourished 1 households Pandya-Lorch 1991 aPercent of income from cattle. bPer cent of income from dairy ing. cPer cent of income from live stock, fruit, and vegeta bles . 42 phasized (Quisumbing et al. 1995; Valdi via, Dunn, investment that does not cur rently exist in most and Sher bourne 1995). Dairy coop era tives have in de vel op ing countries. fact been a major means of bringing women in poor Oppor tu ni ties for rapid growth are rare in poor areas suc cess fully into the cash econ omy in East rural areas. Live stock produc tion may provide one Africa (Brok ken and Sey oum 1992), India (Schnei- such op por tu nity. The poor around the world have der 1995), and Bolivia (Valdi via, Dunn, and Sher- shown their abil ity to produce livestock, and the fu - bourne 1995). ture for the sec tor as a whole looks good. How ever, A pat tern that shows that the poor earn a higher poli cy mak ers and research ers urgently need to find share of their income from live stock than do the the best market-oriented means for ensur ing that wealthy raises the possi bil ity that the Live stock small op era tors bene fit from growth in the livestock Revo lu tion will be good for the poor. The revo lu - sector. To ac com plish this, policies will have to tion of fers two main rea sons for opti mism. First, the focus on rural organi za tion. Live stock produc tion poor can more easily improve their incomes when may provide one of the major op era tional themes in they have a major stake in a sec tor that is grow ing. effec tive rural poverty al le via tion during the next 20 Second, the cur rent rapid inten si fi ca tion of animal years, but things could also go the other way. Fail ure pro duc tion comes at a time when the rural poor des - to address how policies have tended to skew live - per ately need higher re turns to their shrink ing land stock devel op ment in favor of overly large produc - than field crops alone can offer. tion units, and fail ure to promote the verti cal co or di - The main rea son for concern is that increase d nation of small op era tors with proc es sors, will lead inten si fi ca tion might make small opera tors uncom - to a major missed oppor tu nity. Worse, those who petitive compared to large pro duc ers. But true need the ac tiv ity most could be driven out of live - economies of scale in live stock pro duc tion may not stock produc tion . be great once ex plicit and im plicit govern ment sub - si dies to larger produc ers are taken into ac count. Large produc ers in many areas enjoy capital sub si - Other Ethical Issues Raised by dies, tax holi days, free govern ment services, ex on - Trends in Livestock Production eration from certain pol lu tion and health require - ments, and subsi dies on pub lic graz ing land rented Some non food ethical issues asso ci ated with the in large units. Elimina tion of sub si dies or their re di - Live stock Revolu tion cannot be neglected. To date rection toward small-scale produc ers could shift the these issues have manifested themselves pri mar ily in mar ket in favor of the poor. devel oped coun tries, where inten si fi ca tion has pro - A less tracta ble prob lem is that large produc er s ceeded the most. But there is no rea son to think that often find it easier to con tract or ver ti cally inte grate these concerns will not show up elsewhere soon. with proces sors than do small produc ers in de vel op - Ani mal wel fare is a grow ing ethi cal concern in ing coun tries, and there are major econo mies of the era of mas sive hog fac to ries and indus tria l scale in the proc ess ing of per ish able prod ucts. chicken houses (CAST 1997b). As pro duc tion in - Trade in per ish ables benefits greatly from process - ten si fies in de vel op ing countries, concerns over ing arrange ments that stan dard ize quality and make ani mal wel fare will surely surface. Evidence for the quan tity sup plied plenti ful and reli able. Large this is suggested by cul tural at ti tudes toward cows pro duc ers also benefit greatly from mar ket ing in fra - in India, by the strong re lig ious views on ritual structure that alle vi ates the need to sell perish ables slaughter held in many parts of the world, and by the same day to the end user, as is tra di tion ally the the esteem that tradi tional stockrais ers eve ry where case in the devel op ing trop ics. hold for their cat tle. Small op era tors can overcome these bar ri ers The air and water pol lu tion that in ten sive ani- by joining insti tu tions of collec tive action, such mal in dus tries oc ca sion ally create for their neigh - as out grower schemes or par tici pa tory pro ducer bors, espe cially in peri ur ban areas with little par tici - co op era tives (Staal, Del gado, and Nicholson patory govern ment, is also likely to bring increased 1997; Delgado 1998). But the estab lish ment of conflict. Pure market forces in the absence of effec - such insti tu tions requires or gani za tion and tive po liti cal back lash could create hog la goons 43 along side residen tial neigh bor hoods. Ad vance next on humans. Yet the in ser tion of human genes plan ning and ap pro pri ate regula tion by ac count able into transgenic live stock in order to mass produc e gov ern ment authorities can help overcome the fact human proteins in ani mal milk may be es sen tial for that mar kets do not al ways reflect the full costs low er ing the produc tion costs of nu mer ous phar ma - borne by all par ties, in clud ing neighbors. ceu ti cals that remain out of reach for all but the rela- Fi nally, ge netic engi neer ing could push the lim- tively rich. Such ex peri ments may also lead to the its of ethical debates in the next 20 years (see Chap - devel op ment of new drugs, for ex am ple for the treat- ter 9 for a dis cus sion of the tech nol ogy). Cloning of ment of cys tic fi bro sis (Gil lis 1999). Re gard less of the animals raises under stand able concerns about possi - view point adopted, few people anywhere who are ble abuses, particu larly given the temp ta tion to work aware of the is sues will re main mor ally indif fer ent. 7. Environmental Sustainability

Rapidly in creas ing livestock pro duc tion can cause condi tion by increas ing its water- holding capac ity , se ri ous damage to the envi ron ment, but it can also and improves soil structure. Ani mal manure also be har mo ni ous with or even bene fi cial to the envi - helps maintain or create a better climate for micro ron ment when appro pri ate types and levels of pro - flora and fauna in soils. Graz ing ani mals improv e duc tion are in place.11 Techno logi cal progress can soil cover by dispers ing seeds, con trol ling shrub further re duce dam ag ing effects while in creas ing growth, breaking up soil crusts, and re mov ing bio - output. This chapter will iden tify the en vi ron- mass that other wise might be fuel for brush fires. mental problems that are now or likely to become These activi ties stimulate grass til ler ing and im - the most se vere and the policies that exac er bat e prove seed ger mi na tion, and thus improve land these problems. It will also identify tech nolo gies quality and vege ta tion growth. and policies that can enhance the envi ron mental Live stock enable farm ers in the resource- poor sus tain abil ity of live stock pro duc tion as the Live - areas of de vel op ing coun tries to al lo cate plant nutri - stock Revolu tion unfolds. ents across time and space. Land that can not sus tain crop pro duc tion can be used for grazing to produce manure that can make other land more produc tive . Livestock-Environment Grazing live stock can ac cel er ate trans for ma tion of Compatibility nu tri ents in crop by prod ucts to fertil izer, speed ing the process of land recov ery be tween crops. His tori cally, live stock have played a critical role in Large parts of the devel op ing world, espe - the process of agri cul tural in ten si fi ca tion. Live - cially in subhu mid Af rica, are only now begin - stock recy cle nu tri ents on the farm, pro duce valu - ning to gain the benefits of farming that mixes able output from land that is not suitable for sus - crops and live stock. As disease constraints are re - tained crop pro duc tion, and pro vide energy and moved, large live stock animals can be in te grated capi tal for success ful farm opera tions. The inte gra - into crop opera tions, provid ing farm power and tion of live stock and crop opera tions is still the manure. In other parts of the world, such as Asia, main ave nue for sustain able inten si fi ca tion of ag ri - high crop ping and land-use inten si ties have been culture in many re gions of the devel op ing world. sus tained over cen tu ries by in te grat ing crop and This is espe cially true in semi arid and sub hu mid sa - livestock activi ties. Live stock continue to be kept vanna areas receiv ing 600 to 1,200 mil li me ters of for manure and power in these areas, but food annual rainfall. Much of the inte rior of West, East, pro duc tion is becom ing more promi nent with in - and Southern Af rica, north east Bra zil, and much of creasing commer ciali za tion. South Asia be long to this agro cli matic cate gory. Mixed live stock and crop farm ing takes on a Live stock can help main tain soil fer til ity in differ ent form in the more in ten si fied pro duc tion soils lack ing ade quate organic con tent or nutri ent s sys tems in both de vel oped and devel op ing coun - (Ehui et al. 1998). Adding manure to the soil in - tries. Crop and live stock opera tions are inte grate d creases the nutri ent reten tion capac ity (or cation- into a local sys tem where waste from each opera - exchange capac ity), improves the soil’s physical tion is proc essed and transported to become a cheap

11 This chapter draws on Ste in feld, de Haan, and Black burn (1997) and de Haan, Stein feld, and B lack burn (1997).

44 45 input for the other. Re gion wide in te gra tion of live - Scoones, and Ker ven 1993). Where irri ga tion or stock and crop opera tions emerges fairly soon in crop pro duc tion have inter fered with live stock mixed farming. Small-scale rural infra struc ture, resources and move ment patterns, ecol ogi cally such as feeder roads and light motor ized or ani mal sound systems often have been disrupted. Popula - trans port, fa cili tate this inte gra tion. In more inten - tions have become seden tary around water points in sive indus trial farming systems, the po ten tial for many cases. The result ing over graz ing and land over load ing fields with nu tri ents from farm ma nure degra da tion have threat ened the live li hood of pas - becomes a prob lem. toral com mu ni ties. Flexibil ity and mobil ity are essen tial for achieving sustain able range land use in arid areas. Environmental Problems in the Unfor tu nately, both were se ri ously impaired in the Low-Intensity Livestock Systems past by poli cies that set tled pastor al ists and at - of Developing Countries tempted to regulate stocking rates from above rather than in con sul ta tion with stock raisers. Well- Tradi tional, low-intensity live stock pro duc tion intentioned changes often dis rupt tradi tional sys- methods remain in many re gions of the world. Pro - tems and some times make them un sus tain able. For duc tion lev els in these systems are deter mined by ex am ple, new water points in duce set tle ment of locally available re sources. In creased de mand pres - large ani mal and human popu la tions in dry areas. sure can push these systems to pro duce beyond their They also make pos si ble year- round graz ing instead ca pac ity. This can bring them into con flict with the of al low ing stock ing lev els to fluctu ate so that they en vi ron ment, ne ces si tat ing changes in tra di tional are low dur ing the dry sea son. The re sult ing over - practices to reduce dam age. load of animals on the range severely de grades Live stock graze on about 26 per cent of the graz ing re sources around the new source of water. world’s land area. Grazing sys tems in de vel op ing re - Semi arid zones can sus tain more in ten sive ag ri - gions mainly rely on native grassland and are only culture than arid zones, mak ing possi ble greater partially mixed with crops. These sys tems usually do human and live stock popu la tions. Crop en croach - not involve inputs from out side the sys tem. Over - ment on pastures, de for es ta tion through fu el wood graz ing can cause soil compac tion and ero sion, and col lec tion, and overgraz ing of remain ing pas tures can de crease soil fertil ity, organic mat ter con tent, can then ensue. Crop encroach ment on mar ginal and water infil tra tion and storage. Over graz ing in land not only ex poses the soil di rectly to the erosive hilly envi ron ments can accel er ate erosion . ef fects of winds and heavy rains, but it hampers the The United Nations En vi ron ment Pro - flexibil ity of ani mal move ment by obliter at ing pas - gramme (UNEP) esti mates that since 1945 about sages between wet- and dry-season graz ing areas. 680 mil lion hec tares, or 20 percent of the world’s Drought emer gency pro grams that hand out subsi - grazing lands, have been sig nifi cantly degraded dized concen trate feed fur ther contrib ute to range (Oldeman, Hakkeling, and Sombroek 1991). Re- degra da tion by ena bling too many animals to be cent evidence suggests, however, that graz ing main tained on the range and pre vent ing natural re - systems are more resil ient than once thought, genera tion of the range vegeta tion after drought. even in the worst cases, where drought has ex - These feed subsi dies increas ingly have become an tended desert margins. Satel lite imagery shows en ti tle ment for pas tor al ists, espe cially in North Af - re cent vege ta tion growing in the West Af ri can rica and the Middle East. Sahel at the same northern lim its where it once Human popula tion pressure is a key con tribu tor grew before the big droughts of the 1970s and to en vi ron mental deg ra da tion in many areas where 1980s (Tucker, Dregne, and New comb 1991). livestock are kept. When insti tu tional fail ure allows Arid grazing systems have extremely lim ited people to extract private goods (live stock pro duc - poten tial for increasing produc tiv ity and great tion) from public goods (com mon rangeland) with - poten tial for suf fer ing last ing dam age. Pro duc tion out limit, the deg ra da tion is even greater. The situa - in these areas has adapted con tinu ously to highly tion is ag gra vated fur ther when stock numbers can variable rainfall and feed avail abil ity (Behnke, be kept high at all times because of water de vel op - 46 ment and local crop ping in previ ously pastoral pended (Faminow and Vosti 1998). How these areas. These well- intentioned but ecol ogi cally farmers will sur vive on land that is increas ingly de - harm ful poli cies often ex plain why deg ra da tion graded is uncer tain. No clear al ter na tives exist. tends to be more severe in semi arid zones than in In many re gions farm sizes may shrink because arid range lands. of popu la tion pres sure, urban encroach ment, and Tropical rain for ests cover about 720 mil lion subdi vi sion among heirs. Expand ing farm opera - hec tares and con tain ap proxi mately 50 percent of tions into com mu nal or new land under these condi - the world’s bio di ver sity. More than 200 millio n tions often leads to de for es ta tion or overgraz ing. As hec tares of tropical rain for est have been lost since com mu nal grazing and new land become in creas - 1950. Con trib ut ing factors include ranching, crop ingly scarce, whole farm ing sys tems may lose their culti va tion, and for est exploi ta tion. Ranching- livestock compo nent. First to go are large stock associated de for es ta tion has been linked to the loss such as cattle, jeop ard iz ing the nutri ent balance in of some unique plant and ani mal spe cies in South farms along with farm ers’ liveli hoods. Re ported and Central America, the world’s rich est source of nu tri ent deficits range from about 15 kilo grams of biodi ver sity. In Central Amer ica the area under pas - ni tro gen per hec tare per year in Mali, to more than ture has increased from 3.5 mil lion to 9.5 millio n 100 kilo grams of nitro gen per hectare per year in hec tares since 1950, and cat tle popu la tions have the high lands of Ethiopia (de Wit, Wes tra, and Nell more than dou bled from 4.2 mil lion to 9.6 mil lion 1996). head (Kai mo nitz 1995). In vir tu ally all tropi cal high land areas (for ex - Clear ing for est and sa vanna to estab lish pas - ample, the Hima la yan hills, Afri can highlands, An- tures in humid areas causes soil nu tri ents to leach dean countries, and Java) rela tively high human out rap idly under high rainfall and high tempera - popula tion den si ties are tra di tion ally sustained by tures. Weeds soon displace grasses and arti fi cial complex mixed farm ing systems. Fur ther popula - pas tures can only be sustained for a pe riod of up to tion growth in many of these re gions has made it 10 years. More than 50 percent of the pas ture areas impos si ble for the tradi tional mixed sys tems to sur - in the Ama zon are now un grazed fallow, with a sig- vive. Modi fied mixed smallholder farming is possi - nificant por tion being abandoned because of deg ra - ble when market devel op ment per mits in vest ment da tion. Natural regen era tion of for ests is diffi cult, in in ten sive technolo gies and in puts. es pe cially when the cleared areas are large. Policies that make titling of land easy and pro - vide finan cial in cen tives for large ranches have Environmental Problems of been the main reasons for ranch-induced defor es ta - High-Intensity Industrial tion (Kai mo nitz 1995). In the late 1960s and 1970s, Livestock Production Brazil sub si dized ag ri cul tural loans and beef ex - ports, play ing an im por tant role in ranch expan sion . As livestock pro duc tion in ten si fies, pro duc ers These poli cies have now been phased out and in - adopt technolo gies that minimize overt di rect costs vest ment in large ranches by absen tee own ers has asso ci ated with land and labor and take maxi mum declined, and so has the rate of de for es ta tion. In advan tage of free access to en vi ron mental pub lic Central America in the 1980s rain for ests dis ap - goods and capi tal subsi dies. These highly in ten sive peared at an an nual rate of 430,000 hectares per indus trial pro duc tion methods are the rule in devel - year. In 1990–94 that rate was 320,000 hectares an- oped coun tries and grow ing rap idly in im por tance nu ally. The lesson of the 1960s and 1970s is that in the devel op ing world. subsi diz ing horizon tal expan sion of live stock pro - Es ca lat ing de mand for ani mal prod ucts leads to duc tion can cause great en vi ron mental dam age and ani mal con cen tra tions that are out of balance with create dis in cen tives for in ten si fi ca tion. the waste absorp tion and feed sup ply capac ity of In creas ingly evi dent in the Ama zon today are avail able land. High con cen tra tions of animals small crop farm ers who switch their cleared-fores t close to human agglom era tions often cause enor - fields over to live stock grazing only after the soil mous pol lu tion problems. Large areas of West ern nu tri ents neces sary to raise crops have been ex- Europe (Neth er lands, northwest ern Ger many, Brit- 47 tany [France], the Italian Po val ley), the north east - highest when feed quality is low, which is common, ern United States, and, increas ingly, coastal South - es pe cially in the low land trop ics and subtrop ics of east Asia and large plain areas in China now show de vel op ing countries. enor mous nutri ent surpluses that range from 200 to Twenty percent of methane ema nat ing from more than 1,000 kilo grams of nitro gen per hectar e ani mal pro duc tion comes from manure stored under per year (Stein feld, de Haan, and Blackburn 1997). anaero bic con di tions, such as in holding ponds Glob ally pig and poul try in dus tries produce 6.9 (USEPA 1995). The high lev els of methane ema - mil lion tons of nitro gen per year, equiva lent to 7 nat ing from an aero bic con di tions are usu ally asso - per cent of total inor ganic nitro gen fer til izer pro- ci ated with high lev els of produc tiv ity and inten sit y duced in the world. Excess nitro gen and phos pho rus and large pro duc tion units. leach or run off the land, affect ing ground wa ter Ani mal manure also produces ni trous oxide, the qual ity and dam ag ing aquatic and wet land eco sys - most dam ag ing green house gas (320 times more so tems. Tests in Penn syl va nia have shown that about than carbon di ox ide). Ani mal waste contrib ute s 40 percent of the soil sam ples from mixed dairy and about 0.4 mil lion tons of nitro gen per year, or 7 per - crop farms exhib ited exces sive phos pho rus and po - cent of the total global emis sions (Bouwman, Bat- tas sium levels. Sur plus nu tri ents from saturated jes, and Bridges 1992). soils leach into surface water and pol lute the en vi - The feed re quire ments of expand ing meat and ron ment (Narrod, Reyn nells, and Wells 1994). A milk pro duc tion exceed what can be sustained by simi lar scenario is unfold ing in Brittany, where one tra di tional feed resources such as graz ing land and in eight dis tricts had soils with nitrate lev els of crop by- products. Increas ingly the world live stock more than 40 mil li grams per liter in the 1980s. Now sec tor re sorts to ex ter nal inputs, no ta bly high- all eight dis tricts re port such nitrate levels, which energy feed such as cere als and oilcakes. Rough age can cause ex ten sive damage to the re gion’s aquatic is decreas ing in im por tance as feed and being re- systems (Brand jes et al. 1995). placed by ce real and agro-industrial by- products. A Ex cess con cen tra tions of live stock and live - corre spond ing shift to ward pro duc ing monogastric stock waste also pro duce gases. Some, such as am - animals, mainly poul try and pigs, is tak ing place. mo nia, re main local. Others, such as car bon di ox ide While rumi nant meat ac counted for 54 percent of (CO 2), meth ane (CH 4), and ni trous oxide (N2O) af- total meat pro duc tion in the devel op ing coun tries in fect the world’s atmos phere by trapping the sun’s 1970 (FAO 1995a), it ac counted for 35 percent in energy and con trib ut ing to global warming. the early 1990s. This spe cies shift is driven in part There are three main sources of livestock- by the better conver sion rates for concen trate feeds related carbon diox ide emis sions. First, domes ti - offered by monogastric ani mals. cated animals emit carbon di ox ide as part of basic Feed crops dif fer in their pes ti cide needs and meta bolic func tion ing or res pi ra tion, at an es ti - pro pen sity to de plete soil mois ture, water, and nu - mated level of 2.8 bil lion metric tons an nu ally. Sec - tri ents. In gen eral, cereal crops (maize in particu lar) ond, car bon di ox ide emis sions result from biomass have the po ten tial to cause greater en vi ron mental burning, part of which can be at trib uted to land damage than other crops. Ce re als require heavy fer - clearing and bush fires on pas ture used to enhance til izer and pes ti cide use and a great deal of water pas ture growth. Third, car bon diox ide is release d and offer poor ground cover in the early stages of when fos sil fuels are con sumed for livestock- plant de vel op ment. Legumes, such as soybeans and related manu fac tur ing and transpor ta tion . pulses, gen er ally have the least po ten tial for dam - Live stock and manure man age ment contrib ut e age. Maize and wheat de plete nitrates and phos- about 16 percent of the global an nual pro duc tion of phates the most, while cassava and sweet po ta toes 550 mil lion tons of meth ane. Ru mi nants mainly deplete the most soil nu tri ents. produce the methane as a by-product of digest in g large amounts of grasses and other fibrous feeds. Pigs and poul try emit rela tively low amounts of The Environmental Challenge meth ane because they cannot di gest these fibrou s In dus trial live stock pro duc tion is rap idly spring ing feeds. Methane emis sions per unit of product are up close to urban centers in devel op ing coun tries, 48 be cause of weak in fra struc ture, high trans port costs, ucts and continu ous sup ply of high- value food com- and weak regu la tion. Like many cit ies in East and modi ties to urban popula tions are often the over rid - South ern Af rica, Dar- es- Salaam, the capital of Tan- ing policy objec tives, par ticu larly in de vel op ing zania, now has 20,000 dairy cows kept within the countries . city limits. Urban pig ger ies are increas ingly com - In the de vel oped world the pol lu tion of land, mon in Asia, espe cially coastal China. Large- scale water, and air has raised acute aware ness of the en - poultry opera tions are found in pe ri ur ban areas vi ron mental problems as so ci ated with in dus trial throughout the devel op ing world. livestock pro duc tion. In many cases this has trig - In dus trial and in ten sive mixed farm ing systems gered the estab lish ment of poli cies and regu la tory present the most severe en vi ron mental challenge in measures that address these prob lems. De vel op ing the live stock sec tor. These systems have often countries, on the other hand, typi cally exhibit an ab - bene fited from policy distor tions and the absence of sence of ap pro pri ate and en force able regu la tions, regu la tions or their lack of enforce ment. The regu - to gether with a surge in de mand and a lack of effec - la tory vacuum has often given this type of pro duc - tive politi cal ex pres sion of concern about grow ing tion a competi tive edge over land- based systems . en vi ron mental and health hazards . Further more, some poli cies have misdi rected re - Con cerns about the long-term produc tiv ity of source use and encour aged the devel op ment of natural resources, in clud ing land, water, air, and tech nolo gies that are inef fi cient out side the dis- biodi ver sity, will not be reflected in market prices torted context. Many devel op ing coun tries not only un less govern ments and in ter na tional or gani za tions provide di rect sub si dies for feed but also for en ergy define and es tab lish mecha nisms to reflect the pres - and capi tal, which consti tute some of the major di - ent and fu ture value of natural resources. Insti tu - rect costs of in dus trial pro duc tion. Economy wid e tions that pro vide regu la tory frame works need to be poli cies that offer subsi dies for energy and credit de vel oped, local groups need to be empow ered, and often end up favor ing indus trial pro duc tion over a legal author ity that imple ments envi ron menta l less-intensive mixed farm ing and grazing . policies needs to be estab lished or re in forced. Prices Tech nol ogy can offer solu tions to many en vi - should be adjusted through taxation to correct for un - ron mental problems, espe cially under indus tria l charged envi ron mental costs and en cour age ef fi cient condi tions. But most policy frame works enable the resource use. Appro pri ate techno logi cal change, the cheap sup ply of ani mal prod ucts at the ex pense of key to solving envi ron mental problems, must be fa- the envi ron ment. Self-sufficiency in ani mal prod - cilitated by govern ment sup port. 8. Public Health

Live stock pro duc tion is a source of risk to human ingly diffi cult. The clas sic worry in de vel op ing health in both low- and high- in ten sity pro duc tion countries is the micro bial contami na tion of food. sys tems. Risks exist not only from the un con trolled Ba cilli or bac te ria such as Sal mo nella , Escheri chia en demic diseases found in devel op ing regions, but coli, Clostrid ium botulinum, and Staphylo coc cu s also from those that appear in highly de vel oped pro- are the most fre quently observed causes of food- duc tion systems when ani mal con cen tra tions are borne diseases. They are usu ally re lated to improper high, feeds con tain con tami nants, or meat and milk food prepa ra tion or inade quate refrig era tion . are improp erly handled. Hu mans are ex posed to The World Health Or gani za tion (WHO) re - these risks through several pathways. Zoono tic dis - ports that hun dreds of millions of people world - eases, those shared by hu mans and animals, can wide suf fer from diseases caused by con tami - mu tate and spread in ani mal hosts be fore passing to nated food and that prod ucts of animal ori gin hu mans. Ani mal wastes can carry disease or chemi - rank at the top of the list of causes (WHO 1997). cal tox ins into the envi ron ment. And milk and meat More than 3 mil lion chil dren under five years of can expose hu mans to disease and tox ins con tained age die each year because of diar rhea. Con tami - in the ani mals or produced by improper han dling nated water and food-borne pathogens cause and process ing . much of this diar rhea. Salmo nella infec tions are Although great prog ress has been made in con - also on the rise and threaten to become major trol ling en demic diseases of animals, inten si fi ca - public health concerns, espe cially in big Asian tion of both tradi tional and modern live stock pro - cities where large bird flocks are kept for food duc tion systems has brought about new risks. High and fecal contami na tion is hard to control . concen tra tions of animals, often moving between Risks from the in tense indus tri ali za tion of live - loca tions for dif fer ent stages of pro duc tion, can be - stock pro duc tion are also appear ing in devel oped come breed ing grounds for dis ease. Proxim ity of countries. Move ments of ani mals for birthing, these facili ties to high human popula tions can mag - wean ing, and fatten ing have ag gra vated health risks nify the po ten tial damage of an outbreak. be cause ani mals from dif fer ent lo ca tions are ex - Live stock pose a particu lar risk in de vel op ing posed to each other and to humans. Risks of trans - regions, where animal con cen tra tions are often lo - mission are fur ther in creased by greater inter na - cated near and around cit ies because of limited tional trade of live stock and live stock prod ucts. transpor ta tion facili ties and infra struc ture. The Inten sive live stock pro duc tion units can also risks are compounded by inade quate or non ex ist ent breed new and more viru lent strains of zoono tic dis - health in fra struc ture, regu la tions, monitor ing, or eases. An ti bi ot ics used in in ten sive ani mal pro duc - enforce ment. Zoono tic diseases such as tuber cu lo - tion of poul try have led to the emer gence of Sal mo - sis and brucel lo sis, which are nearly or entirely con - nella, Listeria , and E. coli re sis tant to an ti bi ot ics. trolled in de vel oped re gions, con tinue to pose major There is increas ing concern about new strains of in - prob lems in these regions. flu enza devel op ing in pigs and chickens that could Health risks in devel op ing coun tries are still be trans mit ted to humans, such as the avian flu often monitored only at the house hold level. As strain that forced eradica tion of chick ens in Hong livestock pro duc tion lev els rise, the ability of peo - Kong in 1997. Some re search ers go so far as to say ple to gauge the quality of the ani mal food product s that the next lethal flu pan demic might emerge from they purchase in market places becomes in creas - Eu ro pe’s crowded pig barns (MacKen zie 1998).

49 50

Despite long-established food qual ity assur ance Beef that was con tami nated in the slaughter - sys tems in de vel oped coun tries, new food con tami - house is the prin ci pal cause of E. coli infec tion in na tion risks have emerged. Ac cord ing to WHO, humans. Bacte ria are trans mit ted when infected seven food-borne pathogens (Campy lo bac ter jejun i, meat is con sumed raw or under cooked. Ground Clos trid ium per frin gens, E. coli O157:H7, Listeri a meats, such as hamburg ers, are par ticu larly as so ci - mono cy to genes, Sal mo nella, Staphylo coc cus au - ated with infec tions because the in fected mate rial is reus, and Toxo plasma gondi i) are re spon si ble for an mixed through out the product dur ing the grind ing esti mated 3.3 to 12.3 million in fec tions and 3,900 process. deaths annu ally in the United States (WHO 1997). Hygi enic slaugh ter ing prac tices will re duce Fur ther more, global sur veys by WHO indi cate that contami na tion of carcasses but will not guaran tee food- borne dis eases may occur 300- 350 times more the absence of micro bial contami na tion from prod - fre quently than re ported. ucts. Simi larly, the contami na tion of raw milk on Sal mo nel lae are blamed for more than 50,000 the farm can not be completely prevented. The only cases of bacte rial food poison ing in the United effec tive method of eliminat ing danger ous mi - States every year (WHO 1997). Transmis sion of crobes is to heat (thoroughly cook or pasteur ize) or this microbe usu ally occurs through in suf fi ciently irra di ate food. cooked meats and eggs. Chickens are a major res er - Lis teria monocy to gene s (Lm ) has been rec og - voir of salmo nella. Ingest ing foods con tami nated nized as a risk only recently. It in fects cheese and with signifi cant amounts of salmo nella can cause meat that has been re frig er ated for a long time. Lis- in tes ti nal infec tion. teria monocy to gene s infec tions can cause miscar - The increased size of pro duc tion units and fre - riage in women and septi ce mia (blood poison ing ) quency of contact be tween ani mals greatly in - and men in gi tis in infants and per sons with weak - creases the over all impact of outbreaks of illness. ened im mune systems. Six mil lion pigs were destroyed in Hol land to eradi- Non mi cro bial contami na tion of food also poses cate clas si cal swine fever in 1997. As this repor t a risk. Food-borne chemi cal tox ins increas ingly was being writ ten, Ma lay sians were in the pro cess preva lent in rap idly de vel op ing coun tries in clude of de stroy ing a mil lion pigs in their larg est pork pes ti cide residues (often contain ing mercury or ar - pro duc ing region to stem an outbreak of a new form senic), met als (zinc, cadmium, and cop per), and of viral en cepha li tis that has killed more than 100 heavy met als (espe cially mercury from pigs eat ing people in six months (Washing ton Post 1999; fungicide-treated corn). ProMed 1999). Live stock feeding is partly to blame as well for Use of new feed resources has also brought new the accu mu la tion of toxic lev els of heavy met als in risks. In suf fi cient tem pera tures used in render ing the envi ron ment. Trace elements are often added to ani mal tissue into feed has been clearly linked to the ani mal feeds to sup ply mi cro nu tri ents and enhanc e appear ance of BSE (bo vine spongiforme encepha - feed conver sion ef fi ciency. Copper and zinc are de- lo pa thy), or mad-cow dis ease, in cattle. The infec - lib er ately added to a va ri ety of animal feed con cen - tious prions in volved are similar to those that cause trates, while heavy met als like cad mium are in tro - CJD (Creutzfeldt-Jakob dis ease). More than 20 duced invol un tar ily via feed phos phates. Dur ing human vic tims in the United King dom are thought diges tion the elements are concen trated, re sult ing in to have contracted the disease through inges tion of manure and slurries that can con tain high lev els of ani mal nerve tissue . the toxic ele ments. Soils on which pig and poul try In fec tion with Escheri chia coli O157:H7 (E. manure are regu larly ap plied at high rates can accu - coli) was first described in 1982. E. coli has mu late large amounts of heavy met als, which can in emerged rap idly as a major cause of diar rhea and turn con tami nate crops and pose a health risk to hu - acute renal failure. The infec tion is some times fatal, mans. High con cen tra tions of toxic elements can particu larly in chil dren. Outbreaks, gen er ally asso - also accu mu late in the meat of animals fed feeds ci ated with beef, have been reported in Austra lia , with trace met als. Can ada, Japan, the United States, vari ous European Resi dues of growth hor mones, anti bi ot ics, and countries, and southern Africa (WHO 1997). in sec ti cides are increas ingly found in the tis sues of 51 animals raised in indus trial pro duc tion systems . come rec og ni tion that many risks cannot be con - The presence of an ti bi ot ics in ani mal food product s trolled at the household and farm lev els alone. can cause al ler gies. Overdoses of an ti bi ot ics are Consum ers cannot adequately judge the safety of often used ille gally in feed to pro mote growth food that was produced and proc essed far from where enforce ment systems are weak. Con sid er able the household. Farmers act ing indi vidu ally can - debate (and trade dispute) also ex ists about the not fully protect their animals from diseases that safety of growth hor mones such as BST (bo vine so - spread from farm to farm. In many cases insti tu - mato tropin). The il le gal use and misuse of growth tions are required to de velop, monitor, and en - pro mot ers such as DES (diethyl-stilbetrol) and force quality stan dards across the market ing clen buterol can also create problems . chain. Unfor tu nately, govern ment serv ices are Tra di tion ally the health risks from live stock being cur tailed in this area in many poor coun - and live stock prod ucts have been borne by con - tries as the size of the overall public sec tor is sum ers and farmers. With in ten si fi ca tion has being re duced. 9. Technology Needs and Prospects

Increases in the produc tiv ity of animal food produc - most preva lent in pro duc tion systems in which ani- tion come from the devel op ment and trans fer of ani - mals move un con trolled and un moni tored over mal pro duc tion technolo gies, particu larly for animal large distances. Vac ci na tion and sur veil lance pro - health; improved feed and feed use; ge netic en - grams are needed to keep these diseases in check. hancement; and bet ter posthar vest handling. Live- The global eradica tion of rinderpest by 2010, for stock tech nol ogy policy faces two chal lenges in its ex am ple, re mains an achievable and im por tant goal efforts to raise produc tiv ity. First, appro pri ate exist - for de vel op ing countries. ing and new tech nolo gies and produc tion system s Live stock disease control has under gone a have to be adapted and dis semi nated to the devel op - para digm shift in recent years. More flexible con - ing world to elimi nate low pro duc tiv ity. Sec ond, the trol strate gies that focus on re gions of high est re - limits of livestock produc tion tech nol ogy and sys - turns within a coun try are replac ing country wid e tems have to be ex tended to in crease ef fi ciency fur - eradica tion programs. Risk analysis and ani mal ther and the envi ron mental and public health prob - health eco nom ics help deter mine where disease lems that have ap peared in high-intensity livestock con trol in vest ment will have its great est benefit . produc tion have to be solved. This chap ter survey s The accep tance of disease- free or low-risk status the tech nolo gies and policies avail able or in devel op - for re gions (rather than for entire countries) in inter - ment to meet these chal lenges. na tional agreements such as the sani tary and phyto - sanitary (SPS) agreemen t of the World Trade Or - gani za tion (WTO) il lus trates this trend. The treaty Addressing Animal Health makes possi ble ex port of meat and other livestock Constraints prod ucts from a coun try where a disease such as FMD ex ists, as long as the disease is ei ther not pres- In fec tious and para sitic diseases affect ing live stock ent in the region where the meat is produced or the re main im por tant constraints to prof it able live stock risk of transmis sion through mul ti ple con trols is ex - opera tions in many de vel op ing regions. Disease s tremely low. re duce incomes di rectly by causing consid er abl e Biotech nol ogy offers prom ise for the im - livestock losses and in di rectly by neces si tat in g proved diag no sis and treatment of animal dis - health re stric tions on exports. In some areas (in - ease. Even as the inci dence of zoonotic disease s clud ing the former social ist countries) the proble m rises because of increased con cen tra tions of ani- is wors en ing because of weak vet eri nary and ad - mals near peo ple, live stock health research bene - minis tra tive services, the absence of ac count able fits from the greater resources available to human local govern ment, and civil strife. health research (Fitzhugh 1998). For ex am ple, In fec tious diseases such as rinderpest, foot and ge nom ics is a new sci ence appli ca ble to human s mouth disease (FMD), conta gious bo vine pleuro - and live stock that per mits sequenc ing and map - pneu monia (CBPP), and peste des pe tits ru mi nants ping of the ge nome (a ge netic map of a liv ing or - (PPR) still pose major threats to live stock pro duc - ganism). Genom ics takes advan tage of the work tion in de vel op ing coun tries. These epizo otic dis - on the ge nomes of disease or gan isms and per mits eases, so- called because they often manifest them - the de vel op ment of new genera tions of vac cines, selves as epi dem ics affect ing large numbers of includ ing those that use re com bi nant an ti gens to animals of the same spe cies in a given area, are patho logi cal agents. Live stock disease organ isms

52 53 can also pro vide use ful mod els for study ing The po ten tial to gen er ate lo cally available fod- human health (Ole Moi-Yoi 1995). der and cereal feed resources is great. But gov ern - Afri can swine fever is a major con straint to ex- ments may need to create the scien tific and trans - panding pork pro duc tion in Af rica. In a series of re - porta tion in fra struc ture neces sary to reap the cent outbreaks, pig numbers have been reduced by benefits of world feed re sources and new feed tech - between 30 and 70 percent in a wide range of nologies as they become avail able. The projec tions coastal Afri can countries. Movement and sani tary in Chapter 4 in di cate that the neces sary sup plies of con trol can limit fu ture outbreaks but require more feed concen trates will be available on world mar - effec tive veteri nary in sti tu tions. kets without undue price rises. These projec tion s Farmers in de vel op ing regions typi cally lack de pend on maintain ing trends in tech no logi cal low-cost, easy-to- use di ag nos tics, vaccines, and progress that raise yields of major feedgrains such con trol strate gies for disease or gan isms and vec- as maize. tors. Among the para sitic diseases, trypano somia sis But even if ag gre gate quan ti ties are available , (sleeping sick ness) trans mit ted by tsetse flies, poses en sur ing that hundreds of mil lions of small pro duc - an enormous con straint to cat tle pro duc tion in most ers have access to feed mar kets under developing- of the humid and sub hu mid zones of Af rica. country con di tions will be no small challenge. Chi - Com bi na tions of aer ial insec ti cide sprays, ad - na’s port and grain-importation infra struc ture indi - hesive py re the roid insec ti cides, im preg nated cates the kinds of prob lems requir ing remedy. In screens and traps, ster ile in sect mating, and try - 1995 China imported an addi tional 15 mil lion met - pano cide drugs hold the prom ise of gradu ally re - ric tons of grain, put ting con sid er able stress on han- cov er ing in fested areas for mixed farming and in - dling and dis tri bu tion facili ties (Pinstrup- Andersen, creased live stock output. These strate gies also are Pandya-Lorch, and Rosegrant 1997). Signifi cant likely to improve crop out put. bot tle necks occurred. Yet the base line pro jec tion Other im por tant para sitic disease groups in - for 2020 indi cates that China will im port 45 millio n clude hel minthia sis and tick-borne diseases. Hel - metric tons more cere als than in 1993 (Table 24). minths are rarely fatal, but they limit produc tiv ity in Both de vel oped and devel op ing re gions must many pro duc tion systems. They become a lim it ing con tinue to use rations more effi ciently. The re - fac tor in the inten si fi ca tion stage but can be con - quire ments and the quality and quan tity of feed re- trolled. Ticks have the capac ity to transmit diseases, sources used for ra tions dif fer across spe cies, no ta bly east coast fever in Eastern and South ern Af - breeds, systems, and regions. Research to re duce rican coun tries. An effec tive vaccine for this dis- costs and improve effi ciency will have to be highly ease may soon be avail able, with a poten tially large tar geted, but even so it will have spillo ver effects . im pact on ru mi nant produc tiv ity in these countries Research that de fines chemical compo si tion and di - (ILRI 1998). gestibil ity charac ter is tics will contrib ute to crop- breeding strategies, par ticu larly the need to change the phy to chem is try of pri mary crop prod ucts and Improving Feed Quantity resi dues fed to live stock (Fitz hugh 1998). and Quality New technolo gies that en hance the quan tity and quality of available tropical feed re sources are A large number of the world’s livestock, par ticu - being assessed through nutri tion research. The larly ru mi nants in pas toral and low- input mixed- identi fi ca tion of suitable traits and their mo lecu lar farming systems, suf fer ei ther perma nent or sea - markers help improve the quality of tropi cal feeds sonal nutri tional stress. In many regions these pres - de rived from foodcrops. Breeders use the mark ers sures have been alle vi ated through better stor age of to de velop dual- purpose crops with im proved grain locally available resources. Stor age and conser va - yields and pro tein con tent for hu mans and non ru mi - tion of for age, use of high- protein le gu mi nous fod - nants and higher-quality crop residues for rumi - ders and fodder trees in ra tions, treatment of crop nants. resi dues, and ad di tion of min eral nu tri ents to feed Plant genom ics and phy to chem is try will tackle all offer ways to im prove for age in some areas. an tinu tri tional factors (ANFs) in plants. Some of 54 these ANFs can be poison ous to ru mi nants. Micro - the nu tri ents in their feed. Ge netic improve ment biologi cal techniques will help en rich ru mi nant and better balanc ing of feed will enable this trend to eco sys tems with microbes that can better de tox ify con tinue. Pre ci sion in ani mal feeding is foresee able ANFs. Maize with reduced phytic acid content has in de vel oped countries. Nu tri ents ex creted by ani - re cently been reported to im prove feed conver sio n mals will be greatly re duced, so that nutri ent s for chicks by up to 11 percent (CAST forth com ing). needed and sup plied will be fairly equal (CAST The ability to increase starch con tent in feed - forth com ing). grains already evi dent in the de vel oped world may Finally, using growth hormones, such as bo vine become more attrac tive to devel op ing coun tries by so ma to tropin (BST), with high-energy feeds has the 2020. But these technolo gies for maize and sor - po ten tial to increase milk yields. But the tech nol ogy ghum are unlikely to ex tend feeding value per unit is likely to bene fit farms in de vel oped countrie s by more than 15 per cent, even in de vel oped coun - more than those in de vel op ing coun tries. Es ti mated tries. Pul ver iz ing grain with steam before feed ing yield responses for BST vary widely, fa vor ing proba bly will offer a more feasi ble and economi ca l farms with feeds, breeds, and man age ment prac - means for improv ing conver sion effi ciency (CAST tices that are already extremely produc tive (Jar vis forth com ing). 1996).The harsh con di tions in devel op ing coun tries In de vel oped coun tries, and increas ingly also are un likely to pro vide the en vi ron ment neces sar y in in ten sive produc tion sys tems in devel op ing for large-scale benefits from BST. This will change countries, a wide range of probi otic and an ti bi - in the long run, how ever, once live stock produc tiv - otic feed addi tives are part of the ra tion fed to ity in devel op ing regions has bene fited from other, livestock. Micro bial ac tion can help break down less so phis ti cated technolo gies (Jar vis 1996). other wise hard-to- digest roughage so that nutri - ents are better absorbed . Finding better ways to use the vast abun dance Improved Reproductive and of fi brous bio mass available in the world of fers a Genetic Technologies particu larly excit ing area of research. Rumen micro bi ol ogy research fo cuses on the isola tion of Ar ti fi cial insemi na tion has been used for more than fiber-degrading enzymes. Better use of fi brous feed 50 years in de vel oped coun tries, primar ily on com - mate ri als will increase the avail abil ity of feed re - mer cial dairy herds. An es tab lished technol ogy, its sources inedi ble to humans. Chang ing the ca pac ity further spread is likely to occur primar ily through of the rumen to di gest high-fiber fodders would dra - mar ket processes. In the early 1990s, no more than matically improve the prospects of ru mi nant pro - 17 percent of the 50 mil lion first insemi na tion s duc tion in the sub hu mid sa van nas of Af rica and given annu ally took place in de vel op ing coun tries. Latin America, where extremely large quan ti ties of But usage is ad vanc ing rap idly in Asia, espe ciall y biomass of low feed quality are pro duced. In sert ing India, where grow ing milk de mand has made it eco- en hanced cellulase- producing capac ity into rumen nomical (Chu pin and Thi bier 1995; Chupin and bacte ria should be possi ble in the not too distant fu - Schuh 1993). Ar ti fi cial insemi na tion has con sid era - ture (de Haan, Stein feld, and Blackburn 1997; Cun - bly spurred ge netic upgrad ing as large-scale test ing ning ham 1997). Micro bial genom ics will increase of the prog eny of bulls and the subse quent use of the pace of prog ress in this area of research (Wal - valuable breeders has be come possi ble. Wide - lace and Lahlou-Kasi 1995; Odenyo, Osuji, and Ne - spread adop tion of ar ti fi cial insemi na tion is likely gassa 1999). to occur in the more favored pro duc tion en vi ron - With the great advances tak ing place in genet - ments of devel op ing countries, such as temper at e ics, more prog ress should also be made in the feed highlands and peri-urban com mer cial pro duc tion con ver sion of monogas trics. Dur ing the past dec ade areas. The de mand for milk will pro vide re turns for feed conver sion rates for pigs and poul try have im - its intro duc tion and the neces sary technol ogy and proved by 30-50 per cent, in part through breed ing in fra struc ture are be com ing avail able. and in part through ad di tion of enzymes to feeds. The use of embryo transfer, al low ing cows of Still, monogas trics cap ture only 25 to 35 percent of high ge netic po ten tial to pro duce a much larger 55 number of calves than with normal re pro duc tion, is traits and the ability to adapt to harsh climate s cur rently lim ited to only a small part of the com - and re sist diseases. mercial herds in some de vel oped countries. This Genetic research in de vel oped re gions focuse s form of re pro duc tion probably will not becom e less on pro duc ing the hearty animals that are neces - wide spread in the de vel op ing coun tries within the sary in stress ful developing-country envi ron ments next 20 years (Cun ning ham 1997). and more on making animals that pro duce higher Cross ing of local breeds in de vel op ing countrie s quality prod ucts at minimum cost. Un der stand ing with highly produc tive varie ties from the de vel oped the ge netic make- up of animals has al lowed par - coun tries has be come com mon place for dairy cat tle ticular product traits, such as low cho les terol level s in the trop ics. Consid er able gains in produc tiv ity per or the ability to produce high con cen tra tions of a ani mal (25 percent) have been ob tained. Those gains pharma ceu ti cal, to be added to animals. Recent ad - can be main tained with ju di cious inter breed ing or ro - vances in cloning of em bryos could poten tially tational breed ing. How ever, the gains are typi cally have a large impact on live stock pro duc tion, par - lost in sub se quent gen era tions (Cunning ham 1997). ticu larly of dairy cat tle in de vel oped coun tries. But Other authors re port gains of up to 50 per cent, but this is still an area where a number of com plex ethi- stress the one-shot nature of the trans fer and the re- cal and sci en tific is sues have yet to be re solved stricted number of breeds that can be drawn upon (Cun ning ham 1997). (CAST forth com ing). Selec tion from local breeds, es pe cially for mutton and goat meat, may hold con - sid era bly more promise . Postharvest Technology: The charac teri za tion, conser va tion, and use of tropical ani mal breeds are vital to the ability to re- Protecting Public Health and spond to in evi ta bly changing pro duc tion en vi ron - Increasing the Value of Livestock ments. Adapted live stock are more re sis tant to dis - Output ease and en vi ron mental chal lenges. They can main tain produc tiv ity without the need for high- In the next 20 years the trans fer of meat and milk value inputs, increase farm in come, and con trib ute proc ess ing technol ogy to de vel op ing coun tries by to pov erty al le via tion (Rege 1997; Ham mond and the private sec tor under pub lic regula tion is likely to Leitch 1995). be es pe cially impor tant for food pro duc tion. Ad vances in ge net ics also offer new means to Growth rates for the out put of processed product s im prove live stock. Marker- assisted selec tion and will probably be even higher than the growth rates detec tion of quan ti ta tive trait loci, for ex am ple, for meat and milk pro duc tion in de vel op ing coun - combine results from molecu lar and quanti ta tive tries. An increas ing share of total food pro duc tion is ge netic research. In ter ac tions between ge netic traits expected to pass through market ing and process ing and the envi ron ment need to be ad dressed in order chan nels. to em ploy ad ap ta tion traits, such as ge netic resis - The estab lish ment of dairy plants and slaugh- tance to parasites, as pro duc tion traits. As with dis - terhouses in pro duc ing areas, together with mar ket ease resis tance, in sights from human ge net ics re - de vel op ment, will play an im por tant role in stimu - search can be brought to bear on the ge netic lat ing market- oriented pro duc tion. The increas ing im prove ment of live stock (Fitzhugh 1998). im por tance of trading meat and milk over long dis- It became possi ble during the past decade to tances in tropical cli mates will also en cour age tech- pro duce maps of ge netic linkages in order to nol ogy devel op ment and trans fer for food com- iden tify the gene lo ca tions of eco nomi cally im - modi ties such as ultra- pasteurized dairy product s portant traits (disease resis tance, per form ance). and vacuum- packed meat. Food safety is likely to This tech nol ogy, which is being de vel oped for a provide the major im pe tus for tech nol ogy de vel op - large range of traits by a number of research in sti - ment over time. Food safety concerns will occa - tutes worldwide, carries the prom ise of a shortcut sionally con flict with the ob jec tive of small opera - to ge netic improve ment in devel op ing countries. tors to remain competi tive. Cur rent debates over Live stock can be bred for spe cific produc tiv e milk pas teuri za tion in East Af rica bear wit ness to 56 this conflict. East Afri can con sum ers usu ally boil tech no logi cal devel op ment can of fer some hope in un pas teur ized fresh milk col lected earlier in the day devel oped coun tries, and will be increas ingly use ful be cause pas teuri za tion makes prod ucts much more in this re gard in devel op ing countries . expen sive (Staal, Delgado, and Nichol son 1997). Live stock are cited as culprits in global cli mate In the de vel oped coun tries risk analysis is typi - change because they emit green house gases. In de- cally used to help evalu ate exist ing pro grams for vel oped coun tries live stock meth ane emis sions can food safety. The trend to ward globaliza tion of trade be re duced through the use of treated feeds and a ra - makes this kind of analysis increas ingly im por tant. tion closer to nutri tional re quire ments. While emis- Risk analysis will have to be widely intro duced in sions will con tinue to grow in abso lute terms in de - devel op ing coun tries over the next two dec ades. It velop ing countries, inten si fi ca tion of pro duc tion involves an itera tive process with three sets of ele - will enhance the digesti bil ity of feed, reduc ing ments. First is risk as sess ment, es ti mat ing the prob - emis sions per unit of out put. abil ity and po ten tial se ver ity of damage re sult ing Live stock also affect the en vi ron ment by com - from food haz ards. Sec ond is risk man age ment, the pacting the soil structure, which results in ac cel er - de vel op ment of policy for food safety and food ated run-off and soil erosion. High stock ing rates safety programs. Third is risk com mu ni ca tion, pro - and un con trolled grazing cre ate this problem, espe - ductive in ter ac tions between policy mak ers and cially in the hilly areas of devel op ing coun tries. In stake hold ers. order to man age the biomass needs of ani mals at Hazard Analysis Criti cal Control Points varying grazing inten si ties and main tain a criti cal (HACCP) analysis is pres ently the tool of choice for vegeta tive cover to minimize soil erosion, graz ing handling the risks noted above. It identi fies critica l options and pas ture spe cies have to be de fined areas in the food chain that must be monitored to (Mwendera and Mohamed Saleem 1997; Mwen- ensure food safety. Four steps are in volved: as sess - dera, Mohamed Saleem, and Woldu 1997). Graz ing ment of risks in the food chain, de ter mi na tion of the sys tems will remain an im por tant source of anima l criti cal con trol points and criti cal limits for ensur - prod ucts for the fore see able future. To some exten t ing food safety, devel op ment of moni tor ing sys - these sys tems can sus taina bly inten sify produc tion tems, and imple men ta tion of proce dures for veri fi - with stronger insti tu tions, local em pow er ment, and cation. Posthar vest methods that a short time ago regu la tion of access to resources . seemed feasi ble only in the context of developed- Mixed farming sys tems around the world will country indus trial proc ess ing are now be com ing continue to in ten sify and grow in size. Grazing sys - common place in the shrimp and high-value fish er - tems may evolve into mixed farming systems where ies ex port sec tor of devel op ing countries. Simila r there is poten tial for mixed farm ing, as there is in the pro ce dures for high-value meat and milk product s semiarid and sub hu mid trop ics. Im por tant produc tiv - should fol low suit in the next two decades. ity gains can be achieved in mixed farming by fur - ther en hanc ing nu tri ent and en ergy flows be tween crops and live stock. In mixed sys tems livestock sub - Technology to Improve the stitute for natu ral and pur chased inputs, in addi tion to Environmental Effects of pro duc ing meat and milk. The envi ron mental and Livestock Production eco nomic sta bil ity of mixed sys tems make this form of farming in de vel op ing coun tries a prime can di date Live stock are of ten cited as the source of envi ron - for tech nol ogy transfer and devel op ment . mental woes (Chap ter 7). Not all of this nega tive Novel concepts are being devel oped to inte - public ity is merited, es pe cially un der developing- grate crop and live stock pro duc tion in a farm ing country condi tions. Where the prob lem does exist, it area rather than on indi vid ual farms. Are awide is of ten a mat ter of fun da men tal economic structure s crop-livestock inte gra tion allows individ ual, spe - and insti tu tions that must be ad dressed through pol - cialized enterprises to operate separately but icy change rather than tech nol ogy per se. Envi ron - energy for farming and flows of organic and min - mental degra da tion from defor es ta tion and over - eral matter to be linked by markets and regu la - stock ing on the commons are cases in point. Yet tions. This produces the highest efficien cies at the 57 enter prise level while maximiz ing social bene fits. for in creased pro duc tion of pork and poul try will be A form of areawide mixed farming is already fairly largely trans fer able. com mon in develop ing countries, where manure is Issues re lat ing to live stock and the en vi ron ment bartered for feed. cannot be solved with tech ni cal inno va tions alone. In dus trial systems usu ally have a com peti tive A com pre hen sive policy frame work is needed to fa - edge over land-based systems. But in areas with cili tate the adop tion of effec tive technolo gies . high ani mal den si ties, indus trial sys tems will have Tech nol ogy remains the key com po nent because to ab sorb increased pro duc tion costs as a result of fu ture de vel op ment, in clud ing that of the live stock more strin gent regu la tions and pol lu tion levies. In sector, will depend upon land- and water- saving some cases, this will re move the competi tive edge tech nol ogy to substi tute for use of natural re - of indus trial pro duc tion. sources. This trend to ward knowledge-intensiv e Tech nol ogy will play an im por tant role in the sys tems can be widely ob served. Smart tech nolo - process ing of ani mal wastes into use ful prod ucts. gies, supported by astute policies, can help to meet Such technolo gies, for the pro duc tion of dry-pellet fu ture de mand while main tain ing the integ rity of fertil izer from chicken waste, for exam ple, are be - the natural resource base. Better in for ma tion on com ing fairly com mon in de vel oped countries which to base deci sion mak ing is, there fore, ur - (CAST 1996). The eco nom ics of waste dis posal in gently required . crowded devel op ing coun tries are likely to evolve New forms of com mer cial and spe cial ized pro- in this direc tion as well, though probably not in the duc tion that are based upon the resource endow - next 15 years. ments of a region and that main tain nutri ent bal - ances will have to be estab lished. In ten sive sys tems need to be inte grated into a wider frame work for Likely Pathways for the Transfer land use in order to blend resource-saving tech nolo - of Livestock Technology to gies with the ab sorp tive capaci ties of the sur round - Developing Countries ing land. This is par ticu larly neces sary for pig and poultry pro duc tion. New organ iza tional arrange - A number of im por tant techno logi cal devel op ments ments will have to be found to allow special ized are tak ing shape, par ticu larly in ge net ics and repro - units to capital ize on economies of scale. The fu ture duc tion, feeding, and ani mal health. By 2020 these is likely to bring a trans for ma tion of family- based de vel op ments proba bly will be widely in use. mixed farms into spe cial ized and commer cial enter - Demand-driven pro duc tion sys tems in de vel op ing prises in rural areas. countries will likely adopt these technolo gies fairly Policies will need to pro mote areawide inte gra - rapidly. Most of these systems will be in East Asia, tion (as op posed to in di vid ual farm in te gra tion). peri ur ban India, and Latin Amer ica out side the An- This kind of pro duc tion or gani za tion is a long-ter m dean areas. Where demand is grow ing less objec tive of en vi ron men tally sustain able ag ri cul - quickly—most of South Asia, Sub-Saharan Af rica, ture in high- potential zones. In the devel oped and the An dean coun tries—tech nol ogy uptake will world, exces sive ani mal con cen tra tion is being con - be slower and im por tant pock ets of tech no logi cal trolled through quanti ta tive limits on ani mal num - stag na tion will remain. Public- sector research and bers and re lo ca tion of ani mal pro duc tion to areas of ex ten sion for live stock will have a high payoff in lower den sity. Regu la tions for in dus trial pro duc tion the fast- growing areas if they comple ment private- sys tems in urban and pe ri ur ban en vi ron ments need sector activ ity and fa cili tate access to small farmers. to en force vir tu ally zero green house gas emis sions In the slow- growing areas public- sector re search and restric tive licens ing . and exten sion will pro vide the main tech no logi cal To main tain the nutri ent balance in nutrient- ve hi cle for ad dress ing these is sues. deficit mixed farming and to en hance crop- In demand-driven set tings, adop tion usu ally oc - livestock in te gra tion in devel op ing coun tries, poli- curs through market forces, as long as input prices cies need to pro vide in cen tives and serv ices for re flect the rela tive scar city of inputs and create the tech nol ogy up take. To re duce nutri ent surplus in ap pro pri ate incen tives. In par ticu lar, tech nolo gies mixed farming systems, regu la tions to control ani- 58 mal den si ties and waste dis charge and incen tive s incen tives. Similarly, taxation for pasture and crop - for waste re duc tion are required. Often this im plies land in rainfor est areas can dis cour age for est con ver - that sub si dies on concen trate feed and inputs used sion. This needs to be accom pa nied by regu la tions in the pro duc tion of feed need to be removed . protect ing the most valu able areas in terms of the To address envi ron mental degra da tion in the environment or bio di ver sity. graz ing areas of devel op ing coun tries, policies need Finally, as policy mak ers and de vel op ment to fa cili tate local em pow er ment and create prop erty- part ners become increas ingly aware of the key rights instru ments, while al low ing for the flexi bil ity poverty alle via tion role of live stock in de vel op - of movement of pastor al ists. Land- tenure ar range - ing coun tries, agen cies charged with facili tat ing ments can also help limit the ex pan sion of ranching increased live stock produc tiv ity are likely to put into the rainfor est frontier in Latin Amer ica. Lack poverty alle via tion concerns higher on their of market ac cess also de grades the land and needs agenda. This in turn is likely to bind technol ogy to be ad dressed. Other incen tives may help to reduc e de vel op ment more tightly with coop era tives and graz ing pres sure in the semiarid zones, for exam ple, other small-producer insti tu tions that help small the intro duc tion of full cost re cov ery for water and op era tors overcome the transac tion costs of entry animal health serv ices. Grazing fees would provid e into a major commer cial activ ity. 10. Taking Stock and Moving Forward

Truly it is not in ap pro pri ate to use the term “Live - glob ally inte grated markets; (4) in creased substi tu - stock Revolu tion” to de scribe events in world ag ri - tion of meat and milk for grain in the human diet; culture in the next 20 years. Like the well- known (5) rapid rise in the use of cereal- based feeds; Green Revolu tion, the label is a simple and con ven - (6) greater stress put on grazing resources along ient expres sion that summa rizes a com plex series of with more land-intensive pro duc tion closer to cit- in ter re lated processes and outcomes. As in the case ies; and (7) the emergence of rapid tech no logi cal of ce re als, the stakes for the poor in devel op ing change in live stock pro duc tion and process ing in coun tries are enormous. Not unlike the Green indus trial systems . Revolu tion, the “revo lu tion ary” aspect comes from The dangers and oppor tu ni ties raised by each of the partici pa tion of devel op ing coun tries on a large the seven charac ter is tics of the Live stock Revolu - scale in trans for ma tions that had previ ously oc - tion will be exam ined in turn. Then four major pil - curred mostly in the tem per ate zones of devel oped lars on which to base forward- looking policy mak - countries. And like the gradu ally but steadily ris ing ing for the live stock sec tors of devel op ing coun tries ce real yields in the 1970s and 1980s that typifie d will be sug gested. Many of the specific actions are the Green Revolu tion, the Live stock Revolu tion dis cussed in more de tail in the con clu sions to Chap - started off gradually and in creased its rate of ters 6 through 9 of the re port. growth. But the simi lari ties end there. The Green Revo lu tion for ce re als was a supply- side phe nome non; it rested on funda men ta l Charac ter is tics of the Livestoc k tech no logi cal change and the ad ap ta tion and exten - Revolution sion of seed-fertilizer in no va tions in devel op ing countries. The Live stock Revo lu tion is demand- Rapid Increases in Demand That Affect driven. With no ta ble excep tions for milk and poul - Production Patterns and Trade try in the devel oped coun tries, where techno logi cal Rapid de mand growth in devel op ing coun tries pro - progress ar gua bly preceded and pre cipi tated pels the global Live stock Revolu tion. Expand ing changes in de mand through lower prices, the sup ply demand is the result of a com bi na tion of high real side of the Live stock Revo lu tion until now has income growth, swel ling popu la tions, rapid ur bani - mostly re sponded—often under dis torted incen - za tion, and the on go ing diver si fi ca tion of tives—to rapid in creases in demand. developing-country diets away from very high lev - This paper shows that the Revo lu tion has seven els of starchy staples. Milk consump tion has grown spe cific charac ter is tics, each of which of fers both more than 3 percent per year in devel op ing coun - dangers and posi tive oppor tu ni ties for human wel - tries since the early 1980s and is pro jected to grow fare and en vi ron mental sustain abil ity. The seven even faster through 2020. Meat consump tion has are: (1) rapid worldwide increases in consump tion been grow ing about 5 percent per year and is ex - and pro duc tion of live stock products; (2) a major pected to grow 2.8 percent per year through 2020. in crease in the share of de vel op ing coun tries in total These developing-country growth rates can livestock pro duc tion and consump tion; (3) on go ing be compared to the slow fore cast for consump - change in the status of live stock pro duc tion from a tion growth in the de vel oped world: 0.7 percen t mul ti pur pose activ ity with mostly nontrad able out- per year for meat and 0.4 percent per year for put to food and feed pro duc tion in the con text of milk through 2020. This low growth in de vel oped

59 60 countries is largely explained by slow popu la tion standing the stakes in volved helps mo ti vate those growth, slow ing ur bani za tion, sa tia tion of diets, who seek to meet the emerg ing challenges . and grow ing health concerns about high in takes of cho les terol and satu rated fatty acids from World Market Share Steadily Shifts to some animal prod ucts. Developing Countries The rapid growth in consump tion of food from animals has not been and is not likely to be evenly A steady increase is under way in the share of de - distrib uted across regions or even within coun tries. vel op ing countries in the world’s pro duc tion of Since the early 1980s, most growth in con sump tion meat and milk. Shares for the latter com modi ties of meat and milk has occurred in the rap idly devel - amounted to only 31 and 25 per cent, re spec tively, op ing coun tries of East and South east Asia and to a in the early 1980s. The base line pro jec tions esti - lesser ex tent in Latin America. Af rica has lagged mate that developing-country shares in 2020 will be behind, as has India in the case of red meat, al- 60 percent and 52 per cent. Clearly the brunt of the though India has recently witnessed rapid growth in bene fits and costs expected from the Livestoc k milk and poul try consump tion . Revo lu tion will accrue to the devel op ing coun tries. Produc tion trends in the de vel op ing countrie s Sensi tiv ity analysis sug gests that these pro - are follow ing consump tion trends. By 2020 the de - jected trends are robust. Even under the as sump tion velop ing coun tries are expected to pro duce 95 mil - of pro longed eco nomic crisis in Asia, the growth of lion metric tons more meat per year and the devel - aggre gate consump tion of live stock prod ucts re- oped coun tries 20 mil lion metric tons more mains strong in de vel op ing coun tries. Fur ther more, com pared to pro duc tion lev els observed in the early if tastes in India shift toward increased red meat 1990s. This pro duc tion level equals an addi tiona l con sump tion, as they already appear to have done 15 kilo grams per cap ita of meat in the de vel op ing for chicken, the nega tive ef fects of a severe Asian world, given expected popula tion in 2020. The economic crisis on world live stock mar kets are value of the an nual increase in animal food pro duc - wiped out. tion in de vel op ing coun tries cur rently far exceed s that of the growth in pro duc tion of all the major ce - Change in Status from a Local Multipurpose reals com bined. The ca loric value of animal live- Activity to a Global Food Activity stock food product increases will exceed the caloric value of ce re als some time in the next 10 years. The Tradi tion ally, livestock in both the de vel oped de vel oped countries are also expected to add an ad- and de vel op ing coun tries were kept on- farm for ditional 15 kilo grams per cap ita of an nual meat pro - a va ri ety of purposes, in clud ing food, sav ings, duc tion by 2020, much of which will be sold to the animal draft power, fiber, hides, and so forth. In devel op ing coun tries. devel oped countries the use of livestock has be - Such rapid change is cre at ing new oppor tu ni - come spe cial ized. The same trend is being ob - ties for live stock produc ers in de vel op ing coun tries, served in de vel op ing coun tries be cause the op- where some of the world’s poor est people live. The por tu nity for sell ing meat and milk has increased in crease in live stock food prod ucts also holds and the share in pro duc tion of spe cial ized food promise for re liev ing widespread mi cro nu tri ent and animals, such as pigs and poultry, has risen in re - pro tein malnu tri tion, while mak ing posi tive contri - sponse to food de mand. bu tions to the sustain able in ten si fi ca tion of small - Live stock food sectors—tra di tion ally a major holder agri cul ture. Yet, signifi cant new dangers are forum for nontar iff barri ers—are now under go in g also arising. Some forms of live stock pro duc tion rapid changes in na tional policies, in clud ing trade en cour aged by policy distor tions are lead ing to se ri - lib er ali za tion; in vest ment in trade infra struc tur e ous en vi ron mental and health risks. Fur ther more, (roads, ports, refrig era tion facili ties); priva ti za tion increas ingly global live stock and feed mar kets are of pro duc tion and market ing; de regu la tion of inter - sharing eco nomic pains faster and more di rectly, nal markets; and reduced govern ment spend ing on even as they share eco nomic prosper ity. Under - research, ex ten sion, and in spec tion. These changes 61 have spurred consump tion of live stock product s level. As income growth and ur bani za tion proceeds, and in creased trade. the diseases of afflu ence as so ci ated with exces s The expan sion and globaliza tion of world mar- cho les terol intake could well become more of a kets for feed and live stock prod ucts increases the problem, as they already are in much of the devel - extent to which de mand and supply shocks, and the oped world. overall effects of eco nomic boom or bust, are spread to feed and meat prices through out the Rapidly Rising World Demand for world. Major export ing coun tries such as Ar gen - Purchased Feed Concentrates tina, Austra lia, and United States need to pay close Available evi dence, both from histori cal trends and at ten tion to Asia. The 50 percent fall in prices for from in- depth econometric stud ies, sug gests that feedgrains and meat during 1998 in the midwester n price respon sive ness is con sid er able in world cere - United States was largely at trib ut able to mac - als pro duc tion. Simu la tions under various scenar ios roeconomic events in Asia and the former Sovie t indi cate that the net ef fect of the Live stock Revolu - Union. Even per sons solely in ter ested in the course tion on cereal prices in 2020 would be to pre vent of fu ture live stock prices in East Afri can coun tries, them from fal ling fur ther from their cur rently low for exam ple, can not ne glect events affect ing live - levels. Under pes si mis tic as sump tions prices may stock in Asia at the present time, al though doing so in crease by 10 to 20 per cent, but no where near the would not have been so unthink able only a short high lev els of the 1980s. time ago. This view of cereal prices is confirmed fur ther by events in world mar kets during the past 25 years. Substitution of Meat and Milk for Grain in De mand in creases for meat and milk have largely the Human Diet been met through expan sion of feed pro duc tion or Wide spread evidence ex ists that as socie ties grow im ports at world prices that are declin ing in real wealthier they substi tute higher-priced livestock terms. Histori cally, the live stock sec tor has helped calo ries for lower-priced starch calories, at a de - sta bi lize world ce real supply. Evidence from cerea l creasing rate as the sub sti tu tion proceeds. In rural price shocks in the 1970s and 1980s sug gests that China, for exam ple, where meat and milk con sump - re duc tions in cereal sup ply were largely ab sorbed tion is still very low, people may well re duce their by reduc tions in feed for live stock. di rect an nual per cap ita consump tion of ce re als In any event, IM PACT proj ects that an addi - over the next few decades by 20 to 30 kilo grams. tional 292 mil lion metric tons of cere als will be Their cereal consump tion lev els will approach used an nu ally as feed by 2020, with mini mal impact those in the wealthier segments of devel op ing so - on cereal prices (a 2 percent increase in 2020 rela - cie ties today. On the other hand, the an nual per cap - tive to the early 1990s in the base line simula tion) . ita use of cere als as feed is pro jected to rise by about In the worst case scenario, maize prices rise 21 per - 60 to 70 kilo grams in China by 2020. This widely cent above the baseline pro jec tions for 2020. repro duced trend in de vel op ing coun tries will be Simula tions testing the effect of changes in the hard to stop if income and urban popula tion growth ef fi ciency of grain conver sion to meat or milk show continue. that effi ciency and cost matter greatly to the rela tive Contrary to the situa tion in de vel oped coun- competi tive ness of in di vid ual coun tries. Live stock tries—es pe cially the United States, where live stock exports in coun tries with ample graz ing re sources, foods are heavily con sumed—even small increases such as Argen tina, did well when feed grain prices in the consump tion of meat and milk would be were high because they use less feed to pro duce a benefi cial for most women and children in de vel op - unit of meat. Coun tries with ample feedgrain pro - ing coun tries, at least out side the rich est urban duc tion and feed-intensive pro duc tion prac tices, areas. Protein and mi cro nu tri ent defi cien cies re - such as the United States, exported more live stock main widespread in devel op ing coun tries, as Chap - when feedgrain prices were low. Feed conver sion ter 6 points out. Some evidence ex ists that increased ef fi ciency also mattered to the compo si tion of poultry and fish consump tion would be pref er able specific feeds used, ce real de mand gen er ally, and to in creased red meat consump tion above a certain world trade patterns, but it barely af fected the level 62 and dis tri bu tion of live stock consump tion across meat and milk per unit of land to avoid furthe r countries. In a price- responsive global system of degra da tion of exten sive resources. mar kets, in creased feed de mand will be met with Thirty-seven percent of global live stock pro- increased supply, with mini mal changes in the final duc tion already occurs under in dus trial farming supply of meat. condi tions. Increas ing the level of in dus trial pro - Yet the soar ing feed de mand pro jected under duc tion in many cases is lead ing to diffi cult prob- the Live stock Revo lu tion also raises a major ca - lems of manure dis posal and nutri ent over loads in veat. Just because world maize and other feed - adjoin ing soils and bod ies of water. While these grain yields ex panded steadily under signifi can t prob lems have raised concerns primar ily in devel - public and private in vest ment in the past, does oped coun tries, they will soon become se ri ous prob - not guaran tee that this will continue in the fu ture, lems in a number of fast- growing devel op ing coun - es pe cially if produc tiv ity research stops or public tries as well. policy discards the yield poten tial of biotech nol - Growing concen tra tion of animals and people ogy. IM PACT pro jec tions show that an nual in the devel op ing world’s major cit ies (Addis growth in maize yields for most coun tries av er - Ababa, Beijing, Lima, and Mum bai, for ex am ple) ages just above 1 percent through 2020, a rate are lead ing to rapid increases in food safety prob - sig nifi cantly below histori cal trends. If future lems. Infec tions from salmo nella and E. coli, and yield growth rates fall below those baseline pro - emerg ing diseases such as the avian flu observed in jec tions, ce real prices would be higher than pre - Asia, give much cause for con cern. Pesti cides and dicted. Further more, an un fore seen melt down in anti bi ot ics are also build ing up in the food chain be- global maize pro duc tion, perhaps due to dis eases cause of live stock pro duc tion prac tices in the devel - spread because of decreased ge netic diver sity in oped coun tries and in coun tries where moni tor ing maize varie ties, would increase grain and meat and enforce ment are lax. Further more, as the con - prices in the im me di ate term because of the Live - sumption of live stock prod ucts increases in tropi cal stock Revo lu tion. cli mates, espe cially among those popu la tions for whom every day use of such prod ucts is relatively Greater Stress on Extensive Resources new, food safety risks from micro bial infec tion be - and More Intensive Livestock Production come more prevalent. While the list of dangers above may seem Closer to Cities uniformly nega tive, this is also the ap pro pri ate Demand-led increases in live stock pro duc tion have place to note that the prospects for sustain able in - led to the inten si fi ca tion of pro duc tion near major ten si fi ca tion of smallholder agri cul ture under urban mar kets and areas with plen ti ful sup plies of rainfed con di tions would be much more dif fi cult high energy feed. These increases have de graded with out a dy namic live stock sector. Fami lies liv - pro duc tion resources, such as pasture. To date, the ing on a hec tare or two can not sur vive eco nomi - highest lev els of in ten si fi ca tion have occurred pri - cally with crops alone, espe cially in pe ri ur ban marily in the devel oped coun tries. Developed- areas. More in ten sive live stock pro duc tion on country expe ri ences may pro vide ex am ples of pit - these farms pro vides both a higher re turn to farm- falls that de vel op ing coun tries can avoid. ers’ labor and land and a source of or ganic mate - The past expan sion of live stock food pro duc - rial and soil nutri ents gen er ally lack ing in such tion in devel op ing coun tries has come primar ily sys tems. Women’s smallholder dairy de vel op - from increased herd sizes. Range deg ra da tion has ment in East Af rica il lus trates the prom ise that a been observed in exten sive produc tion areas, new live stock activ ity can offer to a farm ing sys - such as graz ing land in the Sahel. In for est areas tem under economic stress. such as the Amazon, perverse incentives led to irre versi ble defor es ta tion by large cattle ranches An Era of Rapid Technological Progress, in the 1970s. As the Live stock Revo lu tion pro - Especially for Industrial Livestock ceeds, incre ments to produc tion will have to As set out in de tail in Chapter 9, the last two or three come increas ingly from higher produc tiv ity of dec ades have seen rapid techno logi cal prog ress in 63 indus trial live stock pro duc tion, mostly in devel - Removing Policy Distortions that Promote oped countries. Progress has come in the form of Artificial Economies of Scale in Livestock im proved feed conver sion, higher out put per ani - Production mal, and higher quality of final product. In the de vel oped countries, techno logi cal prog - The politi cal economy in devel op ing countrie s ress for both rumi nants and monogas tric animals in - often fa vors systems that sig nifi cantly benefit a few volved repro duc tive and ge netic tech nol ogy, includ - targeted indi vidu als or insti tu tions rather than pro- ing ad vances in biotech nol ogy; feed improve men t vid ing small in cre ments for large numbers of peo - through blend ing, process ing, ge netic means, and ple. Inap pro pri ate live stock devel op ment pat terns, chemical treat ment; use of growth hormones; and such as high- cost and highly capi tal ized in dus trial im prove ments in animal health mainte nance. Some pig, milk, and poul try pro duc tion in the pe ri ur ban of these indus trial technolo gies, es pe cially for pigs areas of de vel op ing countries, are often the result of and poultry, have been fairly eas ily transferred to de - delib er ate pol icy choices. Arti fi cially create d velop ing coun tries. economies of scale in pro duc tion add to tech no logi - The private sec tor has played an im por tant and cal economies of scale that may exist for poul try often domi nant role in boost ing live stock pro duc - and pigs (though they proba bly do not exist for tivity and solving en vi ron mental problems in indus - dairy or ru mi nant meat). The result may be a live - trial systems. While live stock systems and pro duc - stock indus try dominated by a few large produc ers, tivity lev els in some de vel op ing coun tries have with few oppor tu ni ties for poor farm ers and littl e begun to converge with those in de vel oped coun - con trol of en vi ron mental or health risks. tries, some whole regions, such as Sub-Saharan Af - Even in the absence of specific tax breaks or rica, have fallen behind. This raises the ques tion of sub si dies, distor tions in domes tic capi tal market s how produc tiv ity increases in those areas are best that pro vide cheap capi tal to large enter prises and de vel oped and propa gated. lim ited or expen sive loans to small-scale opera - tions, favor the large over the small and the sub sti - tu tion of capi tal for labor. Poli cies dealing with in - Strategic Elements for Livestock frastruc ture and access to natural re sources can un in ten tion ally benefit large-scale produc ers at the Policy in Developing Countries expense of smaller ones. Taken to gether, the many op por tu ni ties and dan gers Urban pig ger ies and dair ies that do not ade- of the Livestock Revolu tion dis cussed above sug gest quately dispose of waste mate ri als often op er ate in that it would be foolish for devel op ing countries to poor regu la tory en vi ron ments, under distor tions in adopt a lais sez faire policy for livestock devel op - the mar ket ing chain that pre vent compe ti tion from ment. Many specific recom men da tions for concrete rural areas, and without legal ac count abil ity for pol- action are given in the in di vid ual con clu sions to lution. Over graz ing often results from inade quate Chap ters 6–9. The fo cus here, how ever, is on the property rights devel op ment and en force ment four broad pil lars on which to base a de sir able live - mecha nisms as well as politi cally moti vated subsi - stock devel op ment strategy for devel op ing coun - dies to large pro duc ers. The point is that there is a tries. These are (1) remov ing policy dis tor tions that de gree of choice in how govern ments pro mote live - arti fi cially mag nify econo mies of scale in livestock stock de vel op ment, and the deck is often stacked in produc tion; (2) building partici pa tory insti tu tions of favor of sys tems that harm the en vi ron ment and col lec tive action for small- scale farmers that allo w provide lesser benefits to large numbers of poor them to be verti cally in te grated with livestock proc - rural people . essors and input sup pli ers; (3) cre at ing the envi ron - Alter na tives to large, indus trial pro duc tion ment in which farmers will in crease invest ment in sys tems might be de vel oped through ver ti cal co - ways to improve produc tiv ity in the livestock sec tor; or di na tion of spe cial ized crop and live stock ac - and (4) promot ing ef fec tive regula tory insti tu tions to tivities in high-potential areas. This would permit deal with the threat of envi ron mental and health cri- special ized enter prises to perform ef fi ciently ses stemming from live stock. while maintain ing the bio physi cal links be tween 64 crops and live stock. Such contracts between spe - terprises and in proc ess ing and dis tri bu tion of per - cial ized feed and live stock produc ers are more ish able com modi ties gener ally, sug gests ways likely to emerge if verti cally in te grated com pa - should be found to inte grate small produc ers verti - nies in the indus trial system have to bear the full cally with live stock food proces sors who are also in costs of their use of the envi ron ment. a posi tion to manage input supply. In devel ope d The main alter na tive to an in dus trial pro duc tion countries this has been accom plished through con - sys tem in many devel op ing countries is one that tract farm ing or par tici pa tory producer coop era - uses the labor and quality- control of many small tives, espe cially in the case of milk and poul try. farmers in pro duc tion, but also benefits from the ex - Gov ern ments and de vel op ment partner s per tise, technol ogy and assets of large-scale com pa - seeking to in vest in eco nomic capacity- building, nies that are under contract for input pro vi sion, while facili tat ing the partici pa tion of the poor in proc ess ing, and dis tri bu tion of output. Such ar - commer cially viable activi ties, need to follow the rangements in fact char ac ter ize the poul try in dus try Live stock Revo lu tion closely. The stakes are in the United States and cor re spond to con tract high and growing, and the rap idly rising de mand farm ing of high-value crops in many devel op ing for output increases the prob abil ity of success. countries. As discussed above, many appar en t The worst thing that well- motivated agen cies can economies of scale are in fact located in input sup - do is to pre vent public invest ments that could fa- ply and output proc ess ing and distri bu tion. This cili tate sustain able and market- oriented live stock suggests that insti tu tional inno va tions can allow the pro duc tion by small pro duc ers. This will not stop poor to enjoy a greater share of the bene fits of the the Live stock Revo lu tion, but it will help en sure Live stock Revo lu tion. that the form it takes is less favor able for pov erty al le via tion and sus tain abil ity. Building Institutions for Incorporating Poor Producers into the Benefits of the Livestock Creating Urgently Needed Public Goods for Revolution Livestock Production Creation in Developing Countries Con sid er able evidence from field stud ies around the world shows that the rural poor and landless The benefits pro vided by tech nol ogy de vel op ment presently earn a higher share of their income from and exten sion in the in dus trial live stock sec tors of live stock than better-off rural peo ple. Pov erty alle - de vel oped coun tries largely accrue in the market - viation poli cies must find a way to help the rural place. The private sec tor, there fore, will con tinue to poor par tici pate in the growth made pos si ble by the play the lead ing role in fur ther live stock technol ogy Live stock Revolu tion. The al ter na tive might be that de vel op ment and diffu sion in developed- country the poor are driven out by in dus trial live stock pro - indus trial systems. In dus trial technolo gies for pigs duc ers, and the one grow ing market they pres ently and poul try are largely trans fer able to de vel op ing compete in will be closed to them. countries, suggest ing that the need for pub lic goods Small produc ers face many hid den costs under provi sion for these items is modest. The prob lem is developing-country con di tions. They find it dif fi - that technol ogy devel op ment and exten sion are also cult to gain access to produc tive as sets such as required for cat tle and other types of live stock pro - credit and refrig era tion facili ties and to in for ma tion duc tion. The role of the pub lic sec tor becomes an such as knowledge of micro bial contami na tion pre - issue here, espe cially in devel op ing coun tries, ven tion pro ce dures. In addi tion, the small pro ducer where large private compa nies rarely op er ate out - of per ish ables in the trop ics typi cally is at a bar gain - side the in dus trial live stock sec tor. ing disad van tage with mar ket ing agents, be cause As the stakes rise with the Live stock Revolu - the product must be moved im me di ately or lose its tion, policy regard ing the costs of live stock pro duc - value. Contract insti tu tions re store the balance and tion in de vel op ing countries become even more also benefit dis tribu tors by assur ing quality and re- criti cal than be fore. Educa tional, vet eri nary, re - li able supply. The probable exis tence of genuine search, ex ten sion, and input pro vi sion are not yet economies of scale in input sup ply to live stock en - fully pri vat ized and in many cases can not yet be pri - 65 vat ized at pre vail ing stages of de vel op ment. The duc tion need to be defined. These steps will re quire pub lic goods aspect of live stock devel op ment in de - the compi la tion and dissemi na tion of exten siv e vel op ing coun tries has al ways ex isted. The dif fer - ecologi cal and eco nomic data pertain ing to live- ence now is that because the market is growing, cre - stock pro duc tion to comple ment the ex ten sive tech - at ing oppor tu ni ties and risks, the pub lic goods ni cal data avail able. aspect really matters, espe cially in disease-contro l for smallholder- produced ani mals in ver ti cally inte - Regulating Environmental and grated in dus trial systems. Public Health Concerns Meat and milk pres ently contrib ute more than 40 percent of the value of food and ag ri cul tural pro - In the present phase of live stock devel op ment in the duc tion in the world, but receive a dis pro por tion - more dy namic de vel op ing coun tries, the size, com - ately small al lo ca tion of pub lic in vest ments for fa - po si tion, and end desti na tion of live stock pro duc - cilitat ing pro duc tion. Greatly in creased atten tion tion ar gua bly have outgrown the limited insti tu - must be given to live stock produc tiv ity is sues in de - tional capac ity of minis tries to cope with result ing vel op ing coun tries, in clud ing post har vest process - en vi ron mental and pub lic health is sues. Regu la tory ing and mar ket ing. Policy not only needs to facili - agen cies need to func tion in a way that is commen - tate the shift from increas ing herd size to increas ing su rate with the kinds of problems arising. For ex - produc tiv ity, but it needs to steer this de vel op ment ample, meat hy giene needs to be better enforced in away from over in ten si fi ca tion and en vi ron mental urban China and graz ing better managed in West degra da tion . Africa. Gener ally, technolo gies that deal with envi - Above all, pub licly funded re search and exten - ron mental and pub lic health dan gers stem ming sion should focus on agri cul tural re source man age - from the Live stock Revo lu tion will not work un less ment that com pre hen sively fur thers policy goals that regula tory enforce ment backs them up. Such in sti - re late to human needs. Rather than empha siz ing out - tutional de vel op ments will likely occur only when put maximi za tion above all else, re search and exten - the politi cal de mands for better regu la tion becom e sion should find ways to use a dy namic live stock strong. Such was the case in an earlier era in the de - sector to im prove food secu rity and al le vi ate poverty veloped coun tries. In devel op ing coun tries, an and at the same time mini mize adverse effects on ounce of pre ven tion may eliminate the need for a public health and the en vi ron ment. The design of pound of cure. public invest ment must there fore go beyond a strict In sum, it is unwise to think that the Live stock techni cal ori en ta tion and con sider the so cial, eco - Revo lu tion will somehow go away in re sponse to nomic, and eco logi cal di men sions of the in ter ac tion moral sua sion by well-meaning devel op ment part - of livestock with the bet ter ment of liveli hoods . ners. It is a struc tural phenome non that is here to Enhanc ing the design of public in vest ment in stay. How bad or how good it will be for the popula - the live stock sec tor of de vel op ing coun tries re - tions of de vel op ing coun tries is intri cately bound up quires substan tial improve ments in the crea tion, with how coun tries choose to approach the Live - dis semi na tion, analy sis, and use of policy- relevant stock Revo lu tion. Poli cies can sig nifi cantly im - in for ma tion con cern ing livestock. Spe cifi cally, an prove pov erty al le via tion, en vi ron mental sus tain - im proved inven tory and monitor ing sys tem of the abil ity, and pub lic health, but only if new ac tions changes in the availabil ity, use, and manage ment of are un der taken. Failing to act risks throw ing away the ag ri cul tural resource base worldwide has to be one of the few dy namic eco nomic trends that can be created. Differ ences and over laps between the eco - used to improve the lives of poor rural people in de- logical and eco nomic ef fi ciency of live stock pro - velop ing countries. Appendix: Regional Clas si fi ca tion of Countrie s Used in this Paper

Re gion Member countrie s

China Mainland China

Other East Asia Hong Kong, Ma cau, Mon go lia, North Korea, and South Ko rea In dia In dia

Other South Asia Af ghani stan, Bangla desh, Bhu tan, Mal dives, Ne pal, Paki stan, and Sri Lanka Southeast Asia Brunei, Cam bo dia, East Timor, In do ne sia, Laos, Malay sia , Myan mar, Phil ip pines, Sin ga pore, Thai land, and Viet Nam

Latin Americ a South and Central America and Car ribean Western Asia and Al ge ria, Bahrain, Cy prus, Egypt, Gaza Strip, Iran, Iraq, Jordan , North Africa Kuwait, Leba non, Libya, Mo rocco, Oman, Qatar, Saudi Arabia , Syria, Tu ni sia, Turkey, United Arab Emirates, Western Sahara , and Yemen

Sub- Saharan Af rica Af rica south of the Sahara except for South Af rica

De vel oped Austra lia, Can ada, Eastern Europe, European Union, other West ern European coun tries, Is rael, former So viet Union, Japan , New Zea land, South Af rica, United States

Devel op in g All other coun tries in FAO Sta tis tics Da ta base World All coun tries in cluded in FAO Sta tis tics Da ta base

Sources: Re gional group ings were cho sen based on FAO 1998, which is con sis tent with clas si fi ca tion in Rosegrant et al. 1997. Note: Data from some small countries were not avail able in all series in all years. Miss ing values for very small coun tries are ig nored with out note.

66 References

Ad ams, R. H. Jr., and J. J. He. 1995. Sources of in - Bouwman, A. F., N. H. Batjes, and E. M. Bridges. come ine qual ity and pov erty in ru ral Paki stan . 1992. Meth od ol ogy and data used to esti mate Research Report 102. Wash ing ton, D.C.: Inter - natural N 2O emis sions. Paper presented at an national Food Policy Research In sti tute. in ter na tional workshop on World In ven tory of Soil Emissions Poten tials, August 24–27, Al ex an dra tos, N., ed. 1995. World ag ri cul ture: To - Wagenin gen, The Neth er lands. ward 2010 . Rome: Food and Agri cul ture Or - ganiza tion of the United Na tions. Brandjes, P., J. de Wit, H. van Keulen, and H. B. van de Meer. 1995. En vi ron men tal im pact of ma - Ander son, K., B. Di maranan, T. Hertel, and W. nure man age ment. Im pact do main study for the Mar tin. 1997. Asia-Pacific food mar kets and FAO/WB study on live stock and the envi ron - trade in 2005: A global, economy-wide per - ment. Final Draft. Wagenin gen, The Neth er - spective. Aus tra lian Journal of Agri cul tural and lands: In ter na tional Ag ri cul tural Cen tre. Re source Econom ic s 41 (March): 19–44. Brok ken, R. G., and S. Seyoum, eds. 1992. Dairy Bar nard, N. D., A. Nichol son, and J. L. Howard. market ing in Sub- Saharan Africa . Pro ceed ings of 1995. The medi cal costs attrib ut able to meat a sym po sium held at the In ter na tional Livestock consump tion. Preven tive Medi cine 24 (No vem - Centre for Africa, Ad dis Ababa, Ethiopia, 26–30 ber): 646–655. Novem ber 1990. Ad dis Ababa: In ter na tional Livestock Centre for Af rica (now known as the Behnke, R.H., I. Scoones, and C. Kerven. 1993. In ter na tional Livestock Research Insti tute) . Range ecol ogy in disequi lib rium . Lon don: Overseas Devel op ment Insti tute/In ter na tional Brown, L. R., and H. Kane. 1994. Full house: Reas - In sti tute for Envi ron ment and Devel op ment , sessing the earth’s popu la tion car ry ing ca pac - Com mon wealth Sec re tar iat. ity. New York: W. W. Nor ton. Bhalla, G. S., P. Hazell, and J. Kerr. 1998. Prospects Callo way, D. H. 1995. Human nutri tion: Food and for balanc ing food needs with sustain able re - micro nu tri ent rela tion ships. Ag ri cul tural sources man age ment in India to 2020. Inter na - Strate gies for Mi cro nu tri ents Work ing Paper 1. tional Food Pol icy Research Insti tute, Wash - Inter na tional Food Policy Research In sti tute, ing ton, D.C. Mimeo. Washing ton, D.C. Photo copy . CAST (Coun cil for Ag ri cul tural Sci ence and Tech - Bouis, H. 1991. The rela tion ship be tween nu tri tion nology). 1994. How much land can ten billio n and income sources for the rural poor in a people spare for na ture? Ames, Iowa: CAST. south ern Phil ip pine prov ince. In In come sources of mal nour ished people in ru ral ar eas: ———. 1996. In te grated ani mal waste man age - Micro level infor ma tion and pol icy im pli ca - ment. Ames, Iowa: CAST. tions, ed. Joachim von Braun and Ra jul ———. 1997a. Con tri bu tion of ani mal prod ucts to Pandya- Lorch. Work ing Papers on Commer - healthful di ets. Ames, Iowa: CAST. ciali za tion of Ag ri cul ture and Nutri tion No. 5. Wash ing ton, D.C.: Inter na tional Food Polic y ———. 1997b. The well-being of agri cul tural ani - Research Insti tute. mals . Ames, Iowa: CAST.

67 68

———. Forthcom ing. Con tri bu tion of ani mal ag ri - Del gado, C., and C. Cour bois. 1998. Trade-off s culture to meet ing global hu man food de mand. among fish, meat, and milk de mand in de vel op - Ames, Iowa: CAST. ing countries from the 1970s to the 1990s. Pa - per presented at the 9th Bien nial Confer ence of Chen, J., T. C. Camp bell, J. Li, and R. Peto. 1990. the In ter na tional In sti tute of Fish er ies Eco nom - Diet, life style and mor tal ity in China: A study of ics and Trade—IIFET 98, July 7–11, Tromsö, the charac ter is tics of 65 Chinese counties . Ith - Nor way. aca, NY: Cor nell Uni ver sity Press. Del gado, C., C. Cour bois, and M. Rosegrant. 1998. Chu pin, D., and H. Schuh. 1993. Sur vey of presen t Global food demand and the contri bu tion of status of the use of ar ti fi cial insemi na tion in de - live stock as we enter the new mil len nium. Pa - velop ing coun tries. World Ani mal Revie w 74/75: per presented at the Brit ish So ci ety of Ani mal 26–35. Science-Kenya Ag ri cul tural Research In sti tute Con fer ence on Food, Lands and Liveli hoods: Chu pin, D., and M. Thibier. 1995. Sur vey of the Setting Research Agen das for Ani mal Science, pres ent status of the use of ar ti fi cial insemi na - January 27–30, Nai robi, Kenya. tion in devel oped countries. World Ani mal Re - view 82: 58–68. Ehui, S., H. Li- Pun, V. Mares, and B. Shapiro. 1998. The role of livestock in food secu rity and en vi - Cunning ham, E. P. 1997. The ap pli ca tion of biotech - ronmental pro tec tion. Out look on Ag ri cul ture nologies to en hance ani mal produc tion in differ - 27(2): 81–87. ent farming sys tems. Paper commis sioned by the Food and Ag ri cul ture Organi za tion of the United Faminow, M. D., and S. A. Vosti. 1998. Livestock- Nations in partner ship with the As so cia tion for deforestation links: Policy issues in the west ern South ern Afri can States and the Euro pean Asso - Brazil ian Ama zon. In Livestock and the envi - cia tion for Ani mal Produc tion . ron ment inter na tional confer ence, ed. A. J. Nell. Pro ceed ings of the Inter na tional Con fer - de Haan, C., H. Ste in feld, and H. Black burn. 1997. ence on Livestock and the En vi ron ment held in Livestock and the en vi ron ment: Find ing a bal- Ede/Wagenin gen, The Neth er lands, 16–20 ance. Report of a study co or di nated by the Food June, 1997, and organ ized by the World Bank, and Ag ri cul ture Organi za tion of the United Na - Food and Agri cul ture Or gani za tion of the tions, the United States Agency for In ter na tional United Na tions, and the In ter na tional Ag ri cul - De vel op ment, and the World Bank. Brus sels: ture Cen tre. Wagenin gen: In ter na tional Ag ri - European Com mis sion Directorate-General for cultural Centre . De vel op ment. Fan, S., and M. Agcaoili-Sombilla. 1998. Why pro - de Wit, J., P. T. Wes tra, and A. J. Nell. 1996. En vi - jec tions on China’s fu ture food sup ply and de - ron men tal assess ment of landless ru mi nant mand differ. The Aus tra lian Journal of Ag ri cul - pro duc tion sys tems. Work ing document live - tural and Re source Econom ic s 41 (June): 69–90. stock and the en vi ron ment: Find ing a balance . FAO (Food and Agri cul ture Or gani za tion of the Rome: Food and Ag ri cul ture Or gani za tion of United Na tions). 1995a. Produc tion year book the United Na tions, World Bank, and United 1994. Rome: FAO. States Agency for Inter na tional Devel op ment . ———. 1995b. World ani mal re view, 84/85. Rome: Del gado, C. 1998. Sources of growth in smallholder FAO. agri cul ture in Sub-Saharan Af rica: The role of ———. 1997. FAO statis tics data base. http:faostat. ver ti cal inte gra tion with proces sors of high fao.org/de fault.htm. Accessed Sep tem ber and value-added items. Paper presented at the De cem ber 1997. Inter- Conference Sym po sium of the Inter na - tional As so cia tion of Ag ri cul tural Econo mists, ———. 1998. FAO statis tics data base. http:faostat. August 10–16, Bad plaas, South Af rica. fao.org/de fault. htm. Accessed Summer 1998. 69

Fitch, J. B., and I. A. Soliman. 1983. Live stock and Judge, G., W. Griffiths, R. Hill, H. Lutke pohl, and T. small farmer la bor supply. In Migra tion , Lee. 1985. The the ory and practice of econo met - mecha ni za tion, and agri cul tural labor mar kets rics , 2d ed. New York: John Wiley and Sons. in Egypt, ed. A. Richards and P. L. Martin . Boulder, Colo.: West view Press. Kaimo nitz, D. 1995. Livestock and defor es ta tion in Cen tral America . Tokyo: United Na tions Uni - Fitz hugh, H. A. 1998. Global agenda for livestoc k ver sity Press. research . Nai robi, Kenya: In ter na tional Live - stock Research In sti tute. Ke, B. 1997. Recent de vel op ment in the live stock sec tor of china and changes in live stock/feed Geissler, C. Forthcom ing. China: The soya pork di - rela tion ship. Un pub lished Food and Ag ri cul - lemma. Pro ceed ings of the Nutri tion Soci ety 58. ture Organi za tion of the United Na tions re port, Gillis, J. 1999. Cows and clones on a VA pharm; Ani mal and Health Di vi sion (LPT2), Rome. Where animals make drugs and gene research Mimeo. goes to the fron tier. Washing ton Post , Feb ru- ary 28. Kelly, V., T. Re ar don, A. A. Fall, B. Di agana, and L. McNeilly. 1993. Consump tion and sup ply im - Goodland, R. 1997. En vi ron men tal sus tain abil ity in pacts of agri cul tural price poli cies in the pea nut agri cul ture: Diet matters. Ecologi cal Eco nom - ba sin and Sene gal orien tal. In ter na tional Food ics 23 (Decem ber): 189–200. Pol icy Research Insti tute, Wash ing ton, D.C. Ham mond, K., and H. Leitch. 1995. To ward better Un pub lished re port. Mimeo. man age ment of ani mal ge netic resources . Kerr, J. 1998. Personal commu ni ca tion, Sep tem - World Ani mal Re view 84/85 (3–4): 48–53. ber 23. Hayami, Y., and V. W. Rut tan. 1985. Agri cul tura l de vel op ment: An inter na tional perspec tive . Latham, M. C. 1997. Hu man nu tri tion in the de vel - Balti more, Md., U.S.A.: The Johns Hopkins oping world. FAO Food and Nu tri tion Serie s Uni ver sity Press. No. 29. Rome: Food and Ag ri cul ture Organi za - tion of the United Nations. Hu, F. B., and W. C. Wil lett. 1998. The rela tion shi p between consump tion of ani mal prod ucts (beef, Live stock in Devel op ment. 1998. The in te gra tion pork, poultry, eggs, fish and dairy products) and of live stock in ter ven tions into a sustain able ru - risk of chronic dis eases: A criti cal review. Un - ral live li hoods approach. In Sustain able rura l published report for the World Bank. Harvar d live li hoods: What contri bu tion can we make?, School of Pub lic Health, Boston, Mass., U.S.A. ed. D. Car ney. London: Depart ment for Inter - national Devel op ment. Huang, J., and H. Bouis. 1996. Structural changes in the demand for food in Asia . Food, Agri cul - MacK en zie, D. 1998. This lit tle piggy fell ill... New ture, and the Envi ron ment Dis cus sion Paper 11. Sci en tist, Sep tem ber 12, 18–19. Wash ing ton, D.C.: Inter na tional Food Polic y Research In sti tute. Mozam bique MOA/MSU/UA Research Team. 1992. The deter mi nants of household in come ILRI (In ter na tional Live stock Research Insti tute). and consump tion in ru ral Nampula Prov ince: 1998. ILRI 1997: Live stock, people and the en - Im pli ca tions for food se cu rity and agri cul tura l viron ment . Nai robi, Kenya: ILRI. policy reform. Work ing Paper No. 6. Mozam - IMF (Inter na tional Mone tary Fund). 1997. Inter na - bique: Minis try of Ag ri cul ture. tional fi nan cial statis tics year book. Wash ing - Mwen dera, E. J., and M. A. Mo hamed Saleem. ton, D.C.: IMF. 1997. In fil tra tion rates, surface runoff, and soil Jarvis, L. S. 1996. The po ten tial effect of two new loss as in flu enced by graz ing pressure in the biotech nolo gies on the world dairy indus try. Ethio pian high lands. Soil Use Manage men t 13: Boulder, Colo.: West view Press. 29–35. 70

Mwendera, E. J., M. A. Mo hamed Saleem, and Z. term prospect s. Food Pol icy Re port. Wash ing - Woldu. 1997. Vegeta tion response to cat tle ton, D.C.: In ter na tional Food Pol icy Re search grazing in the Ethio pian high lands. Agri cul - Insti tute . ture, Ecosys tems and Envi ron men t 64: 43–51. ProMed. 1999. ProMed-mail post ings. health net.org/ Narrod, C., R. Reyn nells, and H. Wells. 1994. Po - programs.promed.html. Accessed April. ten tial op tions for poul try waste utili za tion: A fo cus on the Del marva pen in sula. Wash ing ton, Quisumbing, A., L. Brown, H. S. Feld stein, L. Had - D.C.: United States Depart ment of Agri cul ture dad, and C. Peña. 1995. Women: The key to food and the Envi romen tal Protec tion Agency. se cu rity. Food Pol icy State ment 21. Wash ing - ton, D.C.: In ter na tional Food Pol icy Research Neu mann, C., and D. M. Har ris. 1999. Contri bu tion Insti tute . of ani mal source foods in improv ing diet qual - ity for children in the de vel op ing world. Pa per Rege, J. E. O. 1997. The contri bu tion of in dige nous pre pared for the World Bank, Wash ing ton, D.C. live stock to the economies of Afri can coun - tries, with par ticu lar ref er ence to eastern and Odenyo, A. A., P. O. Osuji, and D. Nagassa. Forth- south ern Af rica. Pa per presented at a Semina r com ing. Micro bio logi cal evalua tion of fod der on Livestock Devel op ment Poli cies for East ern leaves as rumi nant feed. Asian-Australasian and South ern Af rica, July 28–August 1. Mba - Journal of Ani mal Scienc e . bane, Swa zi land. Olde man, L. R., R. T. A. Hak keling, and W. G. Rosegrant, M. W., and C. Ringler. 1998. Asian eco - Som broek. 1991. World map of the status of hu - nomic crisis and the long-term global food man in duced soil degra da tion: An explana tory situation. Pa per pre pared for the In ter na tional note. Wagenin gen, The Nether lands: Inter na - Agri cul tural Trade Research Con sor tium Sym- tional Soil Ref er ence In for ma tion Cen tre/ po sium on Pol icy Reform, Market Sta bil ity, United Na tions En vi ron ment Pro gramme. and Food Se cu rity, June 26–27, Alex an dria , Ole Moi Yoi, O. K. 1995. Casein Kinase II in theil - Virginia, U.S.A. eriosis. Sci ence 267 (5199): 834–836. Rosegrant, M. W., M. Agcaoili- Sombilla, and N. Pi men tel, D. 1997. Live stock produc tion: Energy Perez. 1995. Global food pro jec tions to 2020: in puts and the envi ron ment. In Proceed ings, ed. Im pli ca tions for in vest ment. Food, Ag ri cul ture, S. L. Scott and X. Zhao. Pro ceed ings of the and the Envi ron ment Dis cus sion Pa per 5. 46th An nual Meet ing of the Cana dian Soci et y Washing ton, D.C.: In ter na tional Food Polic y of Ani mal Science, July 24–26. Montreal, Can- Research In sti tute. ada: CSAS. Rosegrant, M. W., N. Leach, and R. V. Ger pa cio. Pi men tel, D., J. Houser, E. Pre iss, O. White, H. 1998. Al ter na tive futures for world ce real and Fang, L. Mesnick, T. Barsky, S. Tariche, meat consump tion. In ter na tional Food Policy J. Schreck, and S. Alp ert. 1997. Water re - Research Insti tute, Wash ing ton, D.C. Mimeo. sources: Agri cul ture, the envi ron ment, and so - Rosegrant, M. W., M. Agcaoili-Sombilla, R. V. Ger - ci ety. Bio Sci ence 47 (Feb ru ary): 97–106. pacio, and C. Ringler. 1997. Global food markets Pingali, P., Y. Bigot, and H. P. Binswan ger. 1987. and US ex ports in the twenty- first cen tury. Pa per Agri cul tural mechani za tion and the evo lu tion presented at the Illi nois World Food and Sus tain - of farm ing systems in Sub-Saharan Africa . Bal - able Ag ri cul ture Pro gram Confer ence, “Meet ing timore, Md., U.S.A.: The Johns Hop kins Uni- the Demand for Food in the 21st Century: Chal - versity Press. lenges and Oppor tu ni ties for Illi nois Agri cul - ture,” May 28, Urbana- Champaign. Pinstrup- Andersen, P., R. Pandya- Lorch, and M. Rosegrant. 1997. The world food situation: Re - Schneider, F., ed. 1995. Sec tor strat egy for pro - cent de vel op ments, emerg ing is sues, and long- gramme de vel op ment in livestock produc tion and 71

dairy ing in India. Work ing Paper. Swiss De vel - op tions. Report of a study co or di nated by the op ment Co op era tion, Zolliko fen, Switzer land . Food and Ag ri cul ture Organi za tion of the United Nations, the United States Agency for Inter na - Schroe der, T. C., A. P. Berkley, and K. C. Schroe - tional De vel op ment, and the World Bank. Brus - der. 1995. Income growth and inter na tiona l sels: Euro pean Com mis sion Directorate- General meat consump tion. Journal of Inter na tiona l for De vel op ment. Food and Agri busi ness Market in g 7(3): 15–30. Teklu, T., J. von Braun, and E. Zaki. 1991. Drought Sere, C., and H. Stein feld. 1996. World livestoc k and famine rela tion ships in Su dan: Policy im - produc tion sys tems: Cur rent status, is sues and plica tions . Research Re port 88. Wash ing ton, trends. FAO Ani mal Pro duc tion and Health Pa- D.C.: In ter na tional Food Policy Research In sti - per 127. Rome: Food and Ag ri cul ture Organi - tute. zation of the United Na tions. Tucker, C. J., H. E. Dregne, and W. W. New comb. Sharma, M., M. Garcia, A. Qure shi, and L. Brown. 1991. Expan sion and contrac tion of the Sahara 1996. Overcom ing malnu tri tion: Is there an des ert from 1980–1990. Scienc e 253: 299. ecore gional di men sion? Food, Agri cul ture, and the Envi ron ment Dis cus sion Paper 10. Wash - UNDP (United Nations De vel op ment Programme). ing ton, D.C.: In ter na tional Food Policy Re - 1998. Hu man de vel op ment report 1998. New search Insti tute. York: UNDP. Simp son, J. R., X. Cheng, and A. Miyazaki. 1994. USDA (United States Depart ment of Ag ri cul ture). Chi na’s live stock and re lated ag ri cul ture: Pro- 1997a. Agri cul tural sta tis tics . Washing ton, D.C.: jec tions to 2025. Wallin gord, U.K.: CAB Inter - United States Govern ment Print ing Of fice. national . ———, Eco nomic Research Service. 1997b. World Singh, R. P., M. Aso kan, and T. S. Walker. 1982. agri cul ture: Trends and in di ca tors: World and Size, compo si tion, and other as pects of rura l re gional data. http://mann77. mannlib.cor - in come in the semi-arid trop ics of In dia. Eco - nell.edu/data- sets/in ter na tional/89024/14/ nom ics Pro gram Progress Re port 33. Andhra World.wk1. Ac cessed Decem ber 1. Pradesh, In dia: Inter na tional Crops Research USEPA (United States En vi ron men tal Pro tec tion In sti tute for the Semi- Arid Tropics . Agency). 1995. Global impact domain—Meth - Staal, S., and I. Bal ten weck. 1998. Ratio of income ane emis sions. Consult ant report for the live - from dairy activi ties to income from all farm stock and the envi ron ment study. USEPA, activi ties. Un pub lished data from the Kenya Wash ing ton, D.C. Min is try of Ag ri cul ture/Kenya Agri cul tura l Valdi via, D., E. Dunn, and J. Sherbourne. 1996. Gen - Research Insti tute/In ter na tional Live stock Re - der, livestock and household peasant produc tion : search In sti tute Col labo ra tive Smallholde r Dairy and diver si fi ca tion in crop-livestock sys- Dairy Research Project. Inter na tional Live - tems of an An dean Com mu nity. Insti tuto Bo livi - stock Research Insti tute, Nairobi. ano de Tecnolo gia Agropecuaria (IBTA) Tech ni - Staal, S., C. Delgado, and C. Nichol son. 1997. cal Report 33. http://www.ssu.mis souri.edu/ssu/ Smallholder dairy ing un der trans ac tions costs srcrsp/pa pers/TR33. Accessed Septem ber 23. in East Africa. World De vel op ment 25 (May): von Braun, J., and R. Pandya- Lorch. 1991. In come 779–794. sources of mal nour ished people in rural ar eas: A synthe sis of case stud ies and impli ca tions for Stein feld, H. 1998. Live stock produc tion in the pol icy. In In come sources of malnour ished peo - Asia and Pa cific region: Cur rent status, issue s ple in ru ral areas: Mi cro level infor ma tion and and trends. World Ani mal Re view 90(1): 14–21. policy impli ca tions, ed. J. von Braun and R. Ste in feld, H., C. de Haan, and H. Black burn. 1997. Pandya- Lorch. Work ing Papers on Commer - Livestock-environment in ter ac tions: Issues and ciali za tion of Agri cul ture and Nutri tion, No. 5. 72

Wash ing ton, D.C.: Inter na tional Food Polic y Westlund, L. 1995. Ap par ent histori cal con sump - Research In sti tute. tion and future demand for fish and fishery prod ucts—ex plora tory calcu la tions. Pa per Vosti, S. A., J. Wit cover, and C. Car pen tier. 1998. presented at the Inter na tional Confer ence on Convenio de coopera cion tecnica re gional no Sustain able Con tri bu tion of Fisher ies to Food reem bolsable No. ATN/SF- 4827- RG- Programa Se cu rity, Decem ber 4–9, Kyoto, Ja pan. 1994 de tec nolo gia agro pecuaria regional en America Latina y El Car ibe. Final re port to the Wil liams, M. 1996. The transi tion in the contri bu - Inter-American Devel op ment Bank (IDB)/ tion of living aquatic re sources to food secu rity . International Food Policy Research In sti tute Food, Ag ri cul ture, and the Envi ron ment Dis - (IFPRI) Project No. 2475- 000. In ter na tional Food cus sion Pa per No. 13. Wash ing ton, D.C.: Inter - Pol icy Research In sti tute., Wash ing ton, D.C. national Food Policy Research In sti tute. Wallace, R. J., and A. Lahlou- Kasi. 1995. Rumen ecol ogy research plan ning. Paper presented at a World Bank. 1993. Price pros pects for ma jor work shop held at the In ter na tional Livestock primary commodi ties, 1990–2005, vol ume 2. Research Insti tute, Ad dis Ababa, Ethiopia . Washing ton, D.C.: World Bank. March 13–18. ———. 1997. Commod ity markets and the de vel op - Washing ton Post . 1999. Malay sia says deadly viru s ing countries, No. 4. Washing ton, D.C.: World out break over, April 10. Bank. Webb, P., and J. von Braun. 1994. Famine and food secu rity in Ethiopia: Lessons for Af rica . New WHO (World Health Organi za tion). 1997. World York: John Wiley. health sta tis tics quarterl y 50 (Nos. 1–2).

Chris to pher Delgado and Mark Rosegrant are sen ior research fellows and Claude Cour bois is a researc h ana lyst at the In ter na tional Food Policy Research Institute. Henning Stein feld is senior of fi cer for live stock de vel op ment planning at the Food and Agri cul ture Or gani za tion of the United Na tions. Sim eon Ehui is co or di na tor of the Live stock Policy Analysis Project at the In ter na tional Live stock Re search Insti tute. Recent Food, Ag ri cul ture, and the Envi ron ment Discus sion Pa pers

27. Soil Degra da tion: A Threat to Developing- Country Food Secu rity by 2020?, by Sara J. Scherr, 1999 26. Fostering Global Well- Being: A New Paradigm to Re vi tal ize Agricultural and Rural Development, by David D. Bathrick, 1998 25. Pest Man age ment and Food Pro duc tion: Look ing to the Future, by Monta gu e Yudelman, Annu Ratta, and David Nygaard, 1998 24. Food from Peace: Breaking the Links between Conflict and Hun ger, by Ellen Messer, Marc J. Cohen, and Jashinta D’Costa, 1998 23. Seguri dad Alimen taria y Estrate gias Sociales: Su Con tribu ción a la Segurida d Nutri cional en Areas Urbanas de América Latina, by María Inés Sánchez-Griñán , 1998 22. The Nonfarm Sector and Rural Develop ment: Review of Issues and Evidence , by Nurul Islam, 1997 21. Challenges to the 2020 Vision for Latin Amer ica: Food and Agri cul ture since 1970 , by James L. Gar rett, 1997 20. Water Resources in the Twenty- First Cenu try: Chal lenges and Impli ca tions for Action, by Mark W. Rosegrant, 1997 19. China’s Food Econ omy to the Twenty- First Century: Supply, Demand, and Trade, by Jikun Huang, Scott Rozelle, and Mark Rosegrant, 1997 18. Rus sia’s Food Economy in Tran si tion: Cur rent Policy Issues and the Long-Ter m Outloo k, by Joachim von Braun, Euge nia Serova, Harm tho Seeth, and Olga Melyukhina, 1996 17. The Role of Fer til iz ers in Sus tain ing Food Security and Protect ing the Environ - ment, by Balu L. Bumb and Car los A. Baan ante, 1996 16. Manag ing Inter ac tions between Household Food Security and Preschooler Health , by Law rence Had dad, Saroj Bhat ta rai, Maar ten Immink, and Shubh Kumar, 1996 15. Potential Impact of AIDS on Popu la tion and Economic Growth Rates , by Lynn R. Brown, 1996 14. Land Degra da tion in the Develop ing World: Impli ca tions for Food, Agri cul ture, and the Envi ron ment to the Year 2020, by Sara J. Scherr and Satya Yadav, 1996 13. The Tran si tion in the Con tri bu tion of Living Aquatic Resources to Food Security , by Meryl Williams, 1996 12. Middle East Water Con flicts and Direc tions for Conflict Reso lu tion, by Aaron T. Wolf, 1996 11. Structural Changes in the Demand for Food in Asia, by Jikun Huang and Howart h Bouis, 1996 10. Overcom ing Mal nu tri tion: Is There an Ecore gional Dimen sion? by Mano har Sharma, Marito Garcia, Aamir Qure shi, and Lynn Brown, 1996 9. Agri cul ture, Techno logi cal Change, and the Envi ron ment in Latin America: A 2020 Per spec tive, by Eduardo J. Trigo, 1995 8. Major Natu ral Resource Man age ment Concerns in South Asia, by Gerard J. Gill, 1995

Note: For titles of ear lier papers within this series, please contact IFPRI. INTER NA TIONAL FOOD POL ICY RESEARCH INSTITUTE 2033 K Street, N.W. · Washing ton, D.C. 20006-1002 · U.S.A. · 1-202/862- 560 0 Fax: 1- 202/467- 4439 · E- Mail: [email protected] · Web: www.ifpri.org

FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NA TIONS Viale delle Terme di Caracalla, 00100 Rome, It aly Tel: 39-06/57051 · Fax: 39-06/57053152 · Telex: 625852/625853/610181 FAO I Tele grams: FOODAGRI ROME · Web: www.fao.org

INTERNATIONAL LIVESTOCK RESEARCH INSTITUTE P.O. Box 30709, Nairobi, Kenya Tel.: 254-2/630743 · Telex: 22040 · Cable: ILRI · Fax: 254-2/631499 E-Mail: [email protected] · Web: www.cgiar.org/ilri