Analysis of adoption of genetic modification in pork production chains

PROMOTOR

Prof. dr. ir. R.B.M. Huirne Hoogleraar Agrarische Bedrijfseconomie Wageningen Universiteit

CO-PROMOTOREN

Dr. ir. M.P.M. Meuwissen Senior onderzoeker bij Institute for Risk Management in Agriculture (IRMA) en Universitair docent bij de leerstoelgroep Bedrijfseconomie Wageningen Universiteit

Dr. I.A.C.M. van der Lans Universitair docent bij de leerstoelgroep Marktkunde en Consumentengedrag Wageningen Universiteit

SAMENSTELLING PROMOTIECOMMISSIE Prof. dr. ir. W. Verbeke Universiteit Gent, Belgi
Dr.ir. G.B.C. Backus Landbouw Economisch Instituut, Den Haag Prof. dr. ir. L.A. den Hartog Wageningen Universiteit Prof.dr.ir. J.A.M. van Arendonk Wageningen Universiteit

Dit onderzoek is uitgevoerd binnen de Mansholt Graduate School of Social Sciences

Tatiana A. Novoselova

Analysis of adoption of genetic modification in pork production chains

Proefschrift ter verkrijging van de graad van doctor op gezag van de rector magnificus van Wageningen Universiteit, Prof. dr. M.J. Kropff in het openbaar te verdedigen op vrijdag 8 juni 2007 des namiddags te vier uur in de Aula

Analysis of adoption of genetic modification in pork production chains PhD-thesis Wageningen University – with references – with summaries in English and Dutch Novoselova T.A., 2007 ISBN: 978-90-8504-666-0

Abstract During the past decades the agro-food sector has changed and developed due to the impact of new emerging technologies. Genetic modification (GM) presents one of the recently widely- discussed new technologies. The main objective of this study was to gain insight into adoption of genetic modification in the pork production chain. The project covered three important stages of the chain: the farrowing and fattening stages and the consumers. First, an integrating framework with important aspects for analysing the adoption of new technologies was proposed. Following this framework, consumers and chain participants and their concerns and benefits were analysed in an integrated way. First, consumer preferences for GM pork in general and in specific consumer segments were analysed. An attribute-based stated choice modelling approach (choice experiment) was used to investigate how consumers value certain GM applications in pork production chains, and whether and how consumers trade them off against the potential benefits. On average, the results indicate that conventional pork is preferred over GM pork. Among the four GM applications investigated, GM feed is preferred most. Furthermore, findings reveal that the attractiveness of different benefits offered does not influence consumer choices towards choosing GM pork over regular pork. It is only when consumers have already decided to choose GM pork that their further decisions regarding which GM pork to choose depend on the benefits offered. Next, five distinct consumer segments were obtained: “Anti-GM” (37.2%), “GM proponents” (29.3%), “Occasional eaters” (12.3%), “Application sensitive” (12.3%) and “Benefit seekers” (9.1%). Greater insight into consumer acceptance of GM pork in each segment was provided by profiling consumers regarding attitudes and socio-economic characteristics. Furthermore, the economic impact of GM applications for farmers was investigated in a farm-level model. The results showed that although GM applications improved output results in both stages, the GM applications considered have a limited economic impact. Overall, pork producers can expect reductions in their cost prices of 0.3% to 3.7% in the farrowing stage and of 0.2% to 2.4% in the fattening stage by using GM applications. The results obtained suggest that the current possible range of improvements in economic results due to the GM applications was rather limited and most likely would not be able to motivate farmers to introduce GM into their production system.

Keywords: genetic modification, technology adoption, pork production chain, consumer acceptance, consumer preferences, stated choice modelling approach, segmentation, economic impact

Preface Now my PhD thesis is done. It has been a long and edifying way for me, which I could not have completed without the help of many people who supported me during all these PhD years and directly or indirectly contributed to the success of this work. First, I would like to thank my supervisors Ruud Huirne, Miranda Meuwissen and Ivo van der Lans. I am very grateful to these people for their daily support and assistance that they provided during all these years and their criticism and encouragements. Ruud, without your “always active” approach I would not have began a PhD at Business Economics. Thanks a lot for your supervision, insightful comments and showing that there are no problems, at least in PhD, without an appropriate solution. Also, I would like to thank Miranda for always supporting me, for teaching me being an independent researcher and showing how to apply the scientific results to the real-life. Miranda, thanks a lot for your “geen dank” that I will always remember. To you Ivo, I am indebted not only for crucial comments and suggestions regarding methodology and multivariate data techniques but also for always having time for me, even without an official appointment. I am very grateful to all colleagues that were always with me in Farm Management Group and later at Business Economics. Dear colleagues, thank you so much for the very friendly atmosphere which supported me a lot during my PhD. In particular, I would like to thank Marian for all the assistance I got (also from the beginning of my MSc study here in Wageningen), for very prompt actions, advises and personal care. Martin thanks a lot for being not only an ICT manager, but a caring person. Lan, Paul, Lusine and Argirys I enjoyed very much our coffee breaks and pleasant chats. Special thanks also go to Monique, Vitya and Svetlana for being very nice officemates. Anne and Karin thanks for arranging all the practical matters. Specific gratitude goes to Egbert Kanis for always having time for me to discuss my work and willingness to explain difficult technical things in simple words. Also, I would like to thank the members of my scientific committee, Karel de Greef and Ge Backus, for their critical questions and comments and useful suggestions and discussions. I thank the consumers who were willing to participate in my survey. Also, I am very grateful to all the experts Age Jongbloed, Loek Jonker, Egbert Kanis, Aart Mul and Johan Turkstra who were willing to spend their time discussing peculiarities of the topic studied. Age thanks a lot for the calculations you made for me. For the funding of this project I would like to thank the Institute for Risk Management in Agriculture and Mansholt Graduate School of Social Sciences. I am very grateful to all my friends who supported me a lot from a distance (from Moscow) and here (in Wageningen). Natasha and Ira thank you for being always there for me. Alena and Anya thanks for all your encouraging letters and support I got. Trudi, hartelijk bedankt voor je ondersteuning. My special thanks go to Igor for his great help. Finally, I would like to thank my family. My dearest mama and papa, and another mama and papa, Masha and Maksim thank you very much for your help, encouragement and support through all these years.
    ,                . Last but not least I would like to thank my husband Zhenya. Zhenya, thanks a lot for sharing ups and downs with me, for learning to see things in a perspective, for always supporting me. Without you I would not have done so much.

Tatiana Novoselova, 2007 Contents

Page Abstract

Preface

Chapter 1 General Introduction 1

Chapter 2 Adoption of GM technology in livestock production chains: an 7 integrating framework

Chapter 3 Analysis of consumer acceptance of genetically modified applications in 41 the pork production chain

Chapter 4 Heterogeneity in consumer acceptance of GM applications in pork 71 production: attitudinal and socio-economic determinants

Chapter 5 A feasibility study of GM pork production in the Netherlands 97

Chapter 6 General Discussion 115

Summary 129

Samenvatting 137

List of Publications 145

Curriculum Vitae 147

Training and Supervision Plan 149

Chapter 1

General Introduction

General Introduction

1. Introduction During the past decades the agro-food sector has changed and developed very drastically and rapidly (Gray, Boehlje, Amanor-Boadu & Fulton, 2004; Harwood, Day-Rubenstein, Dunahay, Heisey, Hoffman, Klotz-Ingram et al., 2001). The undergone major changes involved the creation of new and expanded international markets (FAO, 2001a), applying “sustainable” farming systems (Van Calker, Berentsen, Romero, Giesen & Huirne, 2006), improving food safety (Valeeva, Meuwissen, Lansink & Huirne, 2005) and adopting new technological developments (FAO, 2001b). The nature of agro-production has changed, and will change further, due to the impact of new emerging technologies (Brinkhorst, 2000). A considerable number of new technological developments are related to food production. The emerging technologies, aiming at improving current food production systems, bring special values to either the producers or consumers or both. Along with the benefits provided, such technologies have raised controversy and “arguments for and against their implementation” (FAO, 2001b). Genetic modification (GM) presents one of the recently widely-discussed new technologies. The commercial application of GM began in the early 1990s. Advances in GM technology varied among different products as to which technology was applied. The majority of the technical developments refers to crop production, in particular, i.e. to insect-resistant (e.g. Bacillus thuriengensis (Bt) cotton and corn) and herbicide-tolerant (e.g. Roundup ReadyTM soybeans) crops. Up to now the GM crops have been the most successful application of genetic modification in agro-production (Moschini, 2000). Another example of a commercial application of genetic modification is the production of food ingredients, including enzymes, vitamins and other food additives. Over 60% of all industrial enzymes, including enzymes for food production, are produced in genetically modified organisms (Gilissen & Nap, 2000). One of the first commercial examples in the industrial production of enzymes was enzyme chymosine, which is used in the fermentation technology for cheese production (Engel, Frenzel & Miller, 2002). In livestock production there are some significant developments in animal breeding (genetic modification of the animal itself), animal growing (GM feed, GM additives and medicines) and processing (using GM bacteria, in meat processing after slaughtering of animal). With a few (famous) exceptions, applications of genetic modification in breeding only exist at laboratory level. Some GM feed (soy, corn) and GM feed additives (GM enzymes) are already widely used. The speed of adoption and progress of new developments depend on decisions of many parties such as consumers, producers, different organisations and the government. The

2 Chapter 1 adoption of genetic modification in particular seems to be a very complex issue as many parties are involved (Haniotis, 2001). Low consumers’ acceptance of foods derived from genetically modified organisms and continuing public debates have influenced the adoption of GM foods. Previous research mainly concentrated on technology adoption by farmers (see for example, Barham, Foltz, Moon & Jackson-Smith (2004) or on acceptance of GM food by consumers (for example, Fortin & Renton (2003)). However, the complexity of the adoption issue is influenced by many actors in the food production chain, not only by the opinion of producers or consumers. Therefore, more investigation is needed to study technology adoption from a complex chain perspective, where consumers’ and producers’ options and preferences are taken into account. This thesis analyses the GM technology adoption from a chain perspective. The focus of this research was on the adoption of GM technology in the pork production chain. In particular, three important stages of the chain: the farrowing and fattening stages and the consumers were analysed. GM technology in this thesis is presented, in line with description of the developments provided above, by four GM applications, i.e. GM animal, GM feed, GM additives & medicines and GM bacteria, which are possible in the pork production chain.

2. Objectives of the thesis The main objective of this study was to gain insight into adoption of genetic modification in the pork production chains. The following research questions were formulated to address the main objective in this study: 1. What aspects are relevant in the adoption process of genetic modification in pork production chains? 2. How do consumers value and make trade-offs between specific GM applications in pork production and benefits GM can offer them? 3. How heterogeneous are consumers’ preferences regarding different GM applications and how can heterogeneity be explained by attitudinal and socio-economic determinants? 4. What is the farm-level economic impact of selected GM applications in pork production chains?

3. Outline of the thesis Chapter 2 presents definitions of the main terms used throughout this research i.e. genetic modification and genetically modified organisms. Also, a conceptual framework with

3 General Introduction important aspects for analysing the adoption of new technologies is presented here. This framework is used to review the literature on the adoption of genetic modification in livestock production chains. Following this framework, consumers and chain participants and their concerns and benefits should be analysed in an integrated way. Therefore, the subsequent analyses in thesis are organised according to this framework. First, consumer preferences for GM pork in general and in consumer segments are analysed (Chapters 3 and 4). Furthermore, the economic impact of GM applications for farmers is investigated in a farm-level model (Chapter 5). Chapter 3 provides definitions of the four major GM applications that capture the whole production chain and potential benefits (price reduction, improvements in quality, environment, animal welfare and reduced residues in meat) that can be expected from the use of these applications. In this chapter the consumers’ acceptance (of) and preferences for GM pork produced using these four GM applications are investigated by means of a choice experiment. The choice experiment research design explicitly allows for a trade-off between different alternative applications of genetic modification and specific benefits offered due to the use of GM technology. In this chapter consumers are analysed as a single group. However, an examination of existing literature suggests that there are different consumer segments with different attitudes (Hobbs, 2001). Therefore, the main issue of Chapter 4 deals with heterogeneity of the consumers’ preferences regarding different GM applications used in the pork production chain. Specifically, this chapter identifies distinct consumer segments and further profiles consumers in segments, using attitudinal and socio-economic characteristics. In this chapter finite-mixture regression analysis is used to obtain the consumer segments. Factor analysis is applied to assess consumers’ attitudes in different segments towards consumption and production of GM pork by using different GM applications. Chapter 5 quantifies technical and economic performance of GM applications in the pork production chain. In order to capture these relationships, an economic farm model with farrowing and fattening stages is used. Model outputs present the cost price of a 25 kg piglet in the farrowing stage and 1 kg of pork produced in the fattening stage respectively, for various GM applications Lastly, Chapter 6 discusses different issues related to the analysis of genetic modification in food chains using an integrated approach, results obtained and further implications of the project.

4 Chapter 1

References Barham, B. L., Foltz, J. D., Moon, S. & Jackson-Smith, D. (2004). "A Comparative Analysis of Recombinant Bovine Somatotropin Adoption across Major U.S. Dairy Regions." Review of Agricultural Economics, 26(1), 32-44. Brinkhorst, L. J. (2000). Biotechnology - Implications for Agriculture and Society. In Biotechnology: the Science and the Impact, the Hague, The Netherlands. Engel, K.-H., Frenzel, T. & Miller, A. (2002). "Current and future benefits from the use of GM technology in food production." Toxicology Letters, 127(1-3), 329-336. FAO (2001a). "Ethical issues in food agriculture." Series: FAO Ethics Series -1, ISBN: 9251045593, available on-line at http://www.fao.org/docrep/003/X9601E/X9601E00.HTM. FAO (2001b). "Genetically modified organisms, consumers, food safety and the environment." Series title: FAO Ethics Series - 2, ISBN: 9251045607, available on- line at http://www.fao.org/DOCREP/003/X9602E/x9602e00.htm. Fortin, D. R. & Renton, M. S. (2003). "Consumer acceptance of genetically modified foods in New Zealand." British Food Journal, 105(1/2), 42-58. Gilissen, L. J. W. J. & Nap, J.-P. (2000). A contribution to the discussion how a broad choice of GMO-free food products can be guaranteed and organized for consumers. Wageningen, Plant Research International: 34. Gray, A., Boehlje, M., Amanor-Boadu, V. & Fulton, J. (2004). "Agricultural Innovations and New Ventures: Assessing the Commercial Potential." American Journal of Agricultural Economics, 86(5), 1322-1329. Haniotis, T. (2001). The Economics of Agricultural Biotechnology: Differences and Similarities in the US and the EU. In G. C. Nelson, Genetically Modified Organisms in Agriculture: Economics and Politics Academic Press. Harwood, J., Day-Rubenstein, K., Dunahay, T., Heisey, P., Hoffman, L., Klotz-Ingram, C., Lin, W., Mitchell, L., McBride, W. & Fernandez-Cornejo, J. (2001). Economic Issues in Agricultural Biotechnology. Washington, DC, USDA. AgricultureInformation Bulletin No. 762. Hobbs, J. (2001). Consumer Responses to Food Quality, Food Safety, and Health Concerns. In M. Fulton, H. Furtan, D. Gosnellet al, Transforming Agriculture: The Benefits and Costs of Genetically Modified Crops. Prepared for The Canadian Biotechnology

5 General Introduction

Advisory Committee, Project Steering Committee on the Regulation of Genetically Modified Foods. Moschini, G. (2000). Economic Benefits and Costs of Biotechnology Innovations in Agriculture.In Agricultural Trade Liberalization: Can We Make Progress? Quebec, Canada, October 27-28. Valeeva, N. I., Meuwissen, M. P. M., Lansink, A. G. J. M. O. & Huirne, R. B. M. (2005). "Improving Food Safety Within the Dairy Chain: An Application of Conjoint Analysis." Journal of Dairy Science, 88(4). Van Calker, K. J., Berentsen, P. B. M., Romero, C., Giesen, G. W. J. & Huirne, R. B. M. (2006). "Development and application of a multi-attribute sustainability function for Dutch dairy farming systems." Ecological Economics, 57(4), 640.

6

Chapter 2

Adoption of GM technology in livestock production chains: an integrating framework

Tatiana A. Novoselova, Miranda P.M. Meuwissen, Ruud B.M. Huirne Business Economics, Wageningen University and Institute for Risk Management in Agriculture, Wageningen University

Trends in Food Science and Technology 18(2007): 175-188

Abstract: This paper presents an integrating framework for analysing the adoption of new technologies in food chains. We review the literature on the adoption of genetic modification in livestock production chains and conclude that an integrated chain approach is currently lacking. Such an approach is, however, essential to analysing the adoption of technologies, such as genetic modification in livestock production. The proposed framework integrates consumers and chain participants and their concerns and benefits.

Chapter 2

1. Introduction Over the past few decades a number of new food technologies have appeared. These include, for instance, irradiation, nutriogenomics and genetic modification, all of which aiming at improving food safety, animal welfare and the environment. Other technologies reduce production costs and the prices of final products. Although new food technologies hold the promise of such benefits to producers, consumers and society, not all of them are readily adopted and accepted (Frewer, 2003). Technology adoption presents a complex process which has been widely studied from different perspectives. It is acknowledged that many different factors play an important role in the adoption of existing and emerging food technologies. Rogers (2003) investigated technology adoption processes by categories of adopters and adoption rates. Many studies have concentrated on producers’ adoption and specifically on farmers’ adoption of emerging technologies. Research has suggested that adoption of food technologies by producers depends on the perceived and actual profitability, costs of transition, uncertainty of outcomes, as well as on consumer concerns with respect to the specific technology (Hillyer, 1999). Much research has been done on public opinion regarding the acceptability of new technologies and their applications (ABE, 2003; Braun, 2002; Bruhn, 2003; Cook, Kerr & Moore, 2002; Eurobarometer, 2005; Miles, Ueland & Frewer, 2005). In addition, many studies have been done on technology adoption at individual farmer level or on acceptance at consumer level, usually with a focus on a single application of the technology or a single concern ( Barham, Foltz, Moon & Jackson-Smith, 2004; Fortin & Renton, 2003; Grunert, Bech-Larsen, Lähteenmäki, Ueland & Åström, 2004; Lockie, Lawrence, Lyons & Grice, 2005; Marra, Pannell & Abadi Ghadim, 2003; Noussair, Robin & Ruffieux, 2004; O'Connor, Cowan, Williams, O'Connell & Boland, 2006; Scully, 2003; Tauer, 2002; Vilella-Vila, Costa- Font & Mossialos, 2005). However, there is a lack of insight into the integrated effects of technology applications at chain level. This is important, because for some technology applications, costs occur at one stage of the chain while, the benefits are enjoyed in another. Although much work has been done to date, more investigation is needed to study technology adoption from a complex chain perspective. The purposes of this paper are to review the available literature on technology adoption from a chain perspective and to provide a framework for the analysis of new technology adoption within the whole-chain context. This paper focuses more specifically on the adoption of genetic modification technology in livestock production chains including feed suppliers, producers, processors and consumers.

9 Adoption of GM technology in livestock production chains: an integrating framework

The paper proceeds as follows. In the second section we outline findings from the literature on recent developments as to genetic modification in plant (feed) and livestock sectors of agriculture. Possible applications of genetic modification are discussed by stage of the chain. The third section presents the overview of previous research on the analysis of new technologies (Non-GM and GM) from the chain perspective. The fourth section introduces the integrating framework of genetic modification technology adoption in food chains with four key elements: consumer acceptance, producer acceptance, legislative aspects and chain characteristics. The last section presents a discussion and the conclusions.

2. Genetic modification in livestock production chains 2.1 Genetically modified crops and animals There are many definitions of genetically modified organisms (GMO) and genetic modification technology. The term “genetic modification” covers a range of ways of manipulating genetic material and means a process of moving genes into a new species and getting them to function. Although each plant and animal variety is a “genetically modified organism”, the term GMO refers to the product or technique which recombines specific sections of the genetic code from one organism to produce a new plant, animal, or microbe (Gilissen & Nap, 2000). Genetic modification is widespread in crop farming. There are a number of successfully introduced GM crops such as corn, soy or rape. These crops can be used as ingredients for animal feed. Animals fed with such feed are called GM animals. However, the term “GM animals” is rather broad, as it also applies to animals fed with feed containing GM additives or enzymes, those given GM vaccines and hormones (Bonneau & Laarveld, 1999), and those that are genetically modified themselves, such as cloned animals (Visscher, Pong-Wong, Whittemore & Haley, 2000).

2.2 Overview of the livestock chain and applications of genetic modification The chain presents a mechanism of many activities, from input through processing to delivery of the final product to the consumer (Samaranayake, 2005). The livestock production chain consists of animal breeding, growing, slaughtering, processing, retail and the consumer (Figure 1). At the breeding and growing stages animals receive medication (vaccinations),

10 Chapter 2

feed, which is mostly supplied from other farms (in EU and non-EU countries), and feed supplements (hormones, additives and microorganisms).

Medication Animal breeding Feed from outside of the farm Animal growing Supplements to feed

Slaughtering

Processing Bacteria

Retail

Consumer

Figure 1. Livestock production chain

Applications of genetic modification in livestock can be categorised into three main areas: animal production, human nutrition and human health care (Sang, 2003). This paper focuses on applications relevant to animal production. Figure 1 shows the applications of genetic modification in the following stages of the chain: animal breeding (genetic modification of the animal itself), animal growing (GM medication, feed and feed supplements, such as additives, hormones and microorganisms) and processing (using GM bacteria, in meat processing after slaughtering of animal). To date, only for these stages of the chain have genetic modification experiments been done, or do GM products already exist. However, with a few (famous) exceptions, applications of genetic modification in breeding programmes only exist at laboratory level. Some GM feed (soy, corn) and GM feed supplements (GM enzymes) are already widely used. Table 1 shows various areas in which these applications of genetic modification can be advantageous, such as costs of production, animal welfare, food safety, sensory quality of the meat and the environment.

11 Adoption of GM technology in livestock production chains: an integrating framework

Table 1. Potential applications of genetic modification per chain stage and type of benefit Costs of Animal Sensory Food Environment production welfare quality safety

Breeding + + + + + Growing Animal Feed + + + ? + Feed additives + + + ? + Hormones + + + ? + Microorganisms + + + ? + Vaccines + + n.a. + ? Processing Bacteria ? n.a. ? ? n.a. + possible application(s) ? application(s) not yet available n.a. not applicable

Animal breeding Currently there is no GM livestock or meat on the market and it will probably not be for some years to come. However, there are some interesting developments at laboratory level. New breeding programmes may bring various benefits, such as leaner meat, improvement of feed efficiency and increased growth rate. Japanese scientists have created pigs with an implanted spinach gene. These animals produce less fat and therefore their meat is healthier. Pork from such ‘green pigs’ contains 20% less saturated fat than normal pork (Saek et al., 2004). Another application also aimed at improving meat fat content is the incorporation of a growth factor gene (IGF-1). In a study by Pursel, Coleman, Wall, Elsasser, Haden, DeMayo et al. (1996), it was found that the IGF-1 transgene helped reduce carcass fat and boost lean body mass, making each hog worth $6 more on the market. Another application of GM technology in animal breeding causes sows to produce more milk, leading to increased growth in transgenic-reared piglets, which results in lower feed costs, less use of antibiotics and less environmental pollution (Noble, Rodriguez-Zas, Cook, Bleck, Hurley & Wheeler, 2002). Phosphorous pollution from manure of monogastric animals, including pigs and poultry, is a major environmental issue. Canadian scientists have developed GM pigs that produce manure containing up to 75% less phosphorus (so-called “Enviropigs”) (Golovan, Meidinger, Ajakaiye, Cottrill, Wiederkehr, Barney et al., 2001). Genetic modification technology can help to improve resistance of animals to diseases such as BSE and scrapie (Paterson, DeSousa, Ritchie, King & Wilmut, 2003). In addition, the use of genetic modification to increase animals’ disease resistance may be beneficial from a food safety point of view.

12 Chapter 2

Animal growth At the growing stage, there are many possibilities for genetic modification in medication, feed, feed additives, hormones and vaccines. As for GM feed, the most-used crops are those of the so-called first generation, meaning that they are herbicide tolerant and disease resistant, but have a composition, digestion and feeding value for the animals similar to those of conventional crops (Gilissen et al., 2000; Phipps & Beever, 2000). However, there are also some new crop varieties with improved feeding value through incorporated phytase. This is not only economically profitable, but also benefits the animal by helping it in digesting phytate. Moreover, this application reduces phosphorus and thus benefits the environment (Zhang, Kornegay, Radcliffe, Wilson, & Veit, 2000). Some current studies focus on incorporation of edible vaccines, medicines, antibodies, enzymes and hormones into plants (Giddings, Allison, Brooks & Carter, 2000; Streatfield, 2005; Streatfield, Jilka, Hood, Turner, Bailey, Mayor et al., 2001). These applications will not only have economic implications, such as reduced costs and an increased growth rate of animals, but also affect aspects of animal welfare and the environment. Feed additives such as nutrients, enzymes and immune product supplements have a significant role in enhancing livestock nutrition (Forano & Flint, 2000). With better nutrition, animals digest feed more efficiently and, therefore, become healthier and more productive.

Processing of livestock products At the processing stage, applications of GM bacteria (e.g. lactic acid-producing bacteria) ferment meat products and help make the preservation process more effective. The currently used GM bacteria are similar to those used for fermenting a range of raw agricultural products (FAO, 2004).

3. Previous research on technology adoption in food chains In the literature, the adoption of different new technologies in food chains has been mainly analysed quantitatively by using modelling techniques. This section presents an overview of previous studies on the adoption of technologies from the chain perspective. Therefore, three types of chain models are distinguished: (1) “one-stage” models, including only one stage of the chain, (2) “various stages” models, covering more than one stage of the chain, and (3) “integrated” chain models, analysing the adoption of new technologies in the entire chain, including all components/stages.

13 Adoption of GM technology in livestock production chains: an integrating framework

Given the lack of studies on the adoption of GM in food chains, first general chain models with the adoption of non-GM technologies have been reviewed (Table 2). Several distinctions are made in the table regarding the type of chain model, the kind of technology analysed, modelling methods used, objective functions applied and the static/dynamic nature of the model. General chain models (non-GM) Several authors have presented different frameworks analysing the adoption of new technologies in one of the chain stages. These were mostly models of farmer’s adoption of new crop varieties or consumer’s adoption (willingness to pay) of foods produced with the help of new technologies (Alfnes, 2004; Fisher, Norvell, Sonka & Nelson, 2000; Flett, Alpass, Humphries, Massey, Morriss & Long, 2004; Marra et al., 2003; Nayga Jr., Poghosyan & Nichols, 2004). An individual farmer’s adoption decision is generally modelled as a dynamic, multi-year and multi-stage decision process, accounting for risk and uncertainty (Abadi Ghadim & Pannell, 1999; Marra et al., 2003). Abadi Ghadim et al. (1999) have proposed in their adoption framework to include “farmer’s personal perceptions, managerial abilities and risk preferences”. The decision to adopt a new crop, or not, is represented here as a decision problem with the objective of profit maximisation. In contrast, Flett et al. (2004) tried to explain a farmer’s adoption of new technologies not by focusing on the economic benefits, but by modelling the general psychological process of farmer’s decision making by focussing on personality type, intelligence, attitudes and objectives. In general, although the contribution of such factor as “changing consumer preferences” to the complexity of agricultural innovations has been acknowledged (see for example Fisher et al. (2000)), this component is still missing in farmer’s adoption studies. As to one-stage models, analysing only consumer acceptance, these are mainly models concentrating on the consumer willingness to pay for foods produced with the help of new technologies (Alfnes, 2004; Nayga Jr. et al., 2004). Besides obtaining willingness to pay estimates, these studies characterise and cluster consumers based on their preferences with respect to the food investigated.

14

Table 2. An overview of general (non-GM) chain models Chain stagesa Reference Technologyb Methodc Objective function Static/ Comments (Year) Dynamicd One-stage Abadi Ghadim New crop variety Static portfolio and Profit maximisation Dynamic Adoption framework also (farm) and Pannell. dynamic adoption model (several years) includes farmers’ perceptions, (1999) with uncertainty managerial abilities and risk preferences One-stage Flett et al., Dairy farming Factor and discriminant Technology adoption n.a. Model focuses on the (farm) (2004) technologies analysis personality type, intelligence, attitudes and objectives of the farmer One-stage Nayga et al., Irradiation Ordered probit model Willingness to pay for n.a.e Model with sample selection (consumer) 2004 irradiated beef One-stage Alfnes, 2004 Hormone-treated beef Mixed logit model Consumer preferences for n.a. The analysis of consumer (consumer) beef choices considers effect of price variable and socio- economic variables Various Griffith et al., New management Budgeting approach and Within and between Dynamic Model considers farm and stages (1995) technology optimisation enterprise resource market components; benefits adjustments for consumers only in terms of increased supply Various Valeeva et al., Food safety Optimisation Cost-effective food safety Static Cost-effectiveness of food stages (2003) technology measures safety measures in dairy chain Various Van der Gaag Food safety Simulation Cost-effective preventive Static Cost-effectiveness of food stages et al.,, (2004) technology measures safety measures in pork chain Integrated Den Ouden, Animal welfare, Simulation and Multi-criteria Dynamic Multi-optimisation criteria: (entire chain) (1996) environment optimisation economics, environmental improvements pollution, hog welfare a The chain stages that were considered in the studies b The different kinds of technologies in food chains that were analyzed in the studies c The method column refers to the methods that were used to analyse the adoption. Mainly these are the econometric models. For a detailed description, please see following textbooks Churchill and Iacobucci (2002), Greene (2003), Hair (2006), and Winston (2003) d Static/Dynamic indicates whether models account explicitly for the factor of time (like several years of adoption) e n.a. means not applicable

Adoption of GM technology in livestock production chains: an integrating framework

Another group of the models, covering various stages of the chain, such as the work by Griffith, Vere and Bootle (1995), Valeeva, Meuwissen and Huirne (2003) and Van der Gaag, Saatkamp, Backus, van Beek and Huirne (2004), considers multiple stages of production chains, but does not establish connections between them and the consumer component. For example, the model of Griffith et al. (1995) integrates farm and market components and analyses the impact of new lamb management technology for the production of large, lean, lamb. The consumer component is only presented in terms of increased lamb supplies on the market. Valeeva et al. (2003) and Van der Gaag et al. (2004) analyse the effect of food safety improvements in dairy and pork production chains, respectively. Valeeva et al. (2003) account for the fact that food safety and traceability investments at feed level have cost and safety assurance implications for the farm and processing level. Van der Gaag et al. (2004) integrates the results of epidemiological and economic models for finishing, transportation, lairage and slaughtering chain stages. Nevertheless, none of these studies investigates the consumer reaction and willingness to pay for food safety improvements. The last group, “integrated” chain models, establishes links between the various stages of the chain, starting from the producer and finishing with the consumer stage. Because these models are just emerging, a very limited number of studies is available. For instance, Den Ouden (1996), first evaluates animal welfare perceptions of consumer-related respondents and then incorporates these results into an optimisation model that takes animal concerns into account. Moreover, in an economic pork chain simulation model, the effects of different measures aimed at improving societal concerns, such as animal welfare or environmental pollution, are quantified. GM chain models With a difference to the general chain models, the analysis of adoption of genetic modification has been mainly done on a one-stage basis, focusing on farm or consumer level. Table 3 lists a number of examples.

16

Table 3. An overview of GM chain models Chain stagesa Reference Technologyb Methodc Objective function Static/ Comments (Year) Dynamicd One-stage Darr & Chern, GM crops (Soybeans, Regression (Tobit ) Adoption probability and Dynamic GM adoption as a function of (farm) (2002) Bt Maize, HR Maize) number and share of acres (3years) attitudes, perceptions and behaviour planted with GMO One-stage Chung et al., GM Feed (GM Optimisation Feed costs minimisation Static Model includes different growth (farm) (1998) soybean) stages of animals One-stage Henriques & GM hormone (bST) Binary choice probit Importance of feed Static Models evaluate interrelation (farm) Butler, and feed management and multivariate probit management practices in between feed management (2002) models the adoption decision technology and bST and interrelation between technologies One-stage Tauer, GM hormone (bST) Profit function model Impact of bST on profit Dynamic Studies different types of farmers (farm) ( 2002) One-stage Barham, Foltz, GM hormone (bST) Multinomial logit Key determinants of bST Static Model shows to which degree (farm) Moon and model (dis) adoption technology affects performance of Jackson-Smith dairy sector (2004) One-stage Loureiro & GM tomato with Multiple bounded Willingness to pay for GM n.a.e GM benefits: enhanced nutrition (consumer) Bugbee, 2005 different consumer probit model tomato value, pesticide reduction, increased benefits shelf food life, increased profit for farmers and enhanced flavour One-stage Kaye-Blake et GM apple with Multinomial logit Willingness to pay for GM n.a. GM benefits: reduced level of (consumer) al., 2005 different consumer model apple chemical insecticides, improved level benefits of antioxidants and improved flavour Various Hayenga et al., GM hormone (pST) Partial budgeting; Initial changes in costs and Dynamic Four possible scenarios are analysed: stages (1989) quarterly econometric returns after pST product efficiency, demand, carcass supply and demand application and impacts on merit, and composite models and simulation pork industry a The chain stages that were considered in the studies b The different kinds of technologies in food chains that were analysed in the studies c The method column refers to the methods that were used to analyse the adoption. Mainly these are the econometric models d Static/Dynamic indicates whether models account explicitly for the factor of time (like several years of adoption) e n.a. means not applicable

Adoption of GM technology in livestock production chains: an integrating framework

The majority of the one-stage GM models at farmer level are aimed at either the GM crop adoption, from growing crops (Darr & Chern, 2002) to feeding them to the animals (Chung & Pettigrew, 1998), or bST hormone adoption (Barham et al., 2004; Henriques & Butler, 2002; Tauer, 2002). In these studies, the adoption process is analysed from different perspectives. For example, Chung and Pettigrew (1998) evaluate the potential producers’ benefits such as costs reduction and changes in feed from new soybean varieties in each livestock growth stage. In addition, authors perform a sensitivity analysis to see the robustness of potential benefits from GM technology under alternative ingredients’ prices and availability of new ingredients. Also, the adoption of GM crops by farmers is analysed by estimating the probability of adoption, and number and share of acres planted with GM crops (Darr et al., 2002). Finding the explanatory factors affecting adoption of GM technology has been a main objective of some studies as well (see for example, Barham et al. (2004); Darr et al. (2002); Henriques et al. (2002)). Furthermore, in some GM studies, as in (Tauer, 2002), farmers are differentiated and analysed as to the size and productivity characteristics of their farms and different adoption phases of GM. The consumer one-stage GM models focus mainly on the consumer acceptance and, as a result, on their willingness to pay for foods produced with the help of genetic modification (Kaye-Blake, Bicknell & Saunders, 2005; Loureiro & Bugbee, 2005). GM products in these studies are presented to the consumers with different benefits related to the nutrition value of the product, pesticide or chemical insecticides reduction, improved flavour and shelf life of the product and profit for farmers. Besides these one-stage GM models, there is a study by Hayenga, Buhr, Skold, Johnson and Grundmeier (1989) who analyse the adoption of a GM growth hormone in various stages of the pork production chain. This study presents several scenarios with respect to demand and supply of GM pork. However, it does not differentiate between consequences for individual chain stages. In spite of extensive literature on the new technology adoption in food chains, there seem to be just a few studies on the modelling of new technology adoption process in the entire food chain. Even more, in case of GM technology adoption in food chains, to the author's knowledge, there have been no such studies at all. Furthermore, although the importance of risk and uncertainty factors have been emphasised in many studies about the adoption process (see for example, Griffith et al. (1995) and Marra, Pannell & Abadi Ghadim (2003)), they have not been addressed in the GM adoption studies.

18 Chapter 2

4. An integrating framework for technology adoption The possible applications of genetic modification along the chain presented in section 2.2, and the potential numerous benefits and concerns they bring, complicate the understanding of the adoption of new technologies. As previous research showed, there are several omissions in the current studies on the adoption of new technologies in food chains and a general integrated chain approach is missing. Our integrating framework is aimed at structuring all information related to the adoption of new technologies in food chains for further analysis (Figure 2).

General attitude to the technology

Concerns Consumers Benefits

Regulations Chain

Concerns Producers Benefits

New technology adoption feasible

Fig. 2. An integrating framework for new technology adoption in food chains.

Analysis of new technology adoption is not an easy task, especially in food chains, where consumers are generally very sensitive to change and evaluate new developments with special attention. Genetic modification is no exception. Previous research on genetic modification mostly concentrated on the adoption of technology by farmers and did not take consumers into account (Chung et al., 1998; Desquilbet, Lemarie & Levert, 2002; Nadolnyak & Sheldon, 2002; Qaim & De Janvry, 2003). However, the increasing concerns about healthy and safe food, as well as growing environmental and ethical concerns, give consumers’ opinion an

19 Adoption of GM technology in livestock production chains: an integrating framework

important role in the process of a successful technology adoption. The framework suggested for technology adoption, therefore, takes the consumer as a starting point. In understanding consumers’ acceptance of genetic modification, a few important things are worth noticing. First, the formation of consumers’ attitudes towards genetic modification was found to be more a top-down process rather than bottom-up process (Scholderer & Frewer, 2003). This means that consumers form their attitudes to the technology based on their general socio-political attitudes and values rather than on the attitudes to the characteristics of the specific product and the information about the potential risks and benefits of the technology. In spite of the fact that both of these processes are usually present, their influence and importance on the consumer attitude may differ to a great extent (Grunert et al., 2004; Scholderer et al., 2003; Sondergaard, Grunert & Scholderer, 2005). Second, although it was found by Bredahl (2001) that consumers “reject the technology overall rather than evaluate products on a case-by-case basis”, the experience of a “strong benefit” of the “real” GM product may positively change consumer attitudes and disregard concerns with respect to the technology (Grunert et al., 2004). In spite of consumer perceptions that are considered central in the adoption process, producers must also adopt a new technology, and for them concerns and benefits are relevant. Furthermore, their actual adoption decisions will depend at least in part on the strict regulations concerning genetic modification. Moreover, in the chain adoption of genetic modification, the specific chain issues should not be neglected; these are also discussed further.

4.1 Consumer acceptance Consumers are of paramount importance in the adoption process of a new technology (Schmidt, 2000). This seems to be especially true for the genetic modification (see for example studies by D’Souza and Quazi (2005); Miles et al. (2005)). Keeping in mind the above-mentioned peculiarities of the consumer attitudes to genetic modification, we would like to discuss in more detail general consumer attitudes, concerns that technology might cause to the consumers, and technological consumer benefits.

General consumer attitude Levels of acceptance of genetic modification vary among countries (Cantley, Hoban & Sasson, 1999; Gaskell, Allum, Bauer, Durant, Allansdottir, Bonfadelli et al., 2000). American

20 Chapter 2

and Canadian consumers seem to have a relatively positive attitude towards genetic technology (Hoban, 1999). The same applies to Japanese consumers (Macer & Chen Ng, 2000). By contrast the European public has a much more negative attitude (Gaskell et al., 2000). This fact was demonstrated in many studies examining consumer attitudes towards GM food in the UK (Burton, Rigby, Young & James, 2001), the Netherlands (Hamstra & Smink, 1996), Sweden (Koivisto Hursti, Magnusson & Algers, 2002; Lahteenmaki, Grunert, Ueland, Astrom, Arvola & Bech-Larsen, 2002), Croatia (Renko, Brcic´-Stipcevic´ & Renko, 2003), Ireland (O'Connor, Cowan, Williams, O'Connell & Boland, 2006, 2005), Denmark, Finland, Norway ( Lahteenmaki et al., 2002) and in many other European countries. Still, European consumers seem to make some differentiation according to the area and type of application (Grunert, Lahteenmaki, Asger Nielsen, Poulsen, Ueland & Astrom, 2001). Consumers have been shown to have more positive attitudes towards the use of genetic modification for medical applications than for food production purposes (Gaskell et al., 2000; Koivisto Hursti et al., 2002). In general, medical applications are considered more beneficial, less risky and more ethically correct than applications to food production (Enriquez, 2001). Frewer, Howard and Aaron (1998) found that GM microorganisms and plants were perceived to pose less risk to human health than did GM animals. Townsend and Campbell (2004) demonstrated that in their study consumer concerns about future risks of using GM animals in food were found to be one of the key predictors of willingness to purchase GM food.

Consumer concerns Among various new food technologies studied (e.g. irradiation), genetic modification elicits the highest level of consumer concerns (Cardello, 2003) that are “very complex and difficult to study” (Koivisto Hursti et al., 2002) (Table 4, part A: Concerns). The main source of concern to consumers with respect to GM organisms is health (Malarkey, 2003). The unknown long-term health effects scare away consumers from accepting and consuming GM food. This negative attitude towards the new technology is maintained by uncertainty. In most cases fear of the unknown is an inseparable part of the new technology adoption process. Worry about allergic reactions to new products is one example (Uzogara, 2000). Closely related to the health issue is the food safety concern. Past food scandals and recent attempts at introduction of new food technologies have decreased consumer confidence in food safety (Grunert, 2002). Consumers have become increasingly critical of new technologies and methods of production (Wilcock, Pun, Khanona & Aung, 2004). They fear,

21 Adoption of GM technology in livestock production chains: an integrating framework

for instance, that some microorganisms, plants or animals might escape from the test-field trials or from biotech laboratories, or that accidents might happen, at which toxic agents will be released. However, potential risk to the environment from GM animals is considerably less than that from GM plants or fish (Sang, 2003). Nevertheless, in comparison to other GM organisms, there are major public concerns about the acceptability of GM animals (Sang, 2003). Consumers motivate their fears by the fact that genetic modification of animals is relatively new and all possible “drawbacks” have not had a chance to be assessed properly. Another public concern, the so-called unnaturalness of GM food, also plays an important role in consumer attitudes towards genetic modification (Moses, 2002; Townsend, Clarke & Travis, 2004; Verhoog, 2003). Some consumers perceive modification of plants, animals or microorganisms to be an unnatural process. Unnaturalness is closely related to such ethical objections as “playing with nature”, “mixing genes among species” and “incorporation of animal genes into plants and vice versa”. Such concerns have been explicitly discussed by Cantley et al. (1999), Gaskell et al. (2000) and Moses (2002). The ethical concern embraces a complex mixed group of ethical, religious and philosophical issues. Some consumers doubt whether it is ethical for humans to genetically modify other organisms, or whether it is ethical to consume GM products (Galun, 2001; Hails, 2004; Townsend et al., 2004). Besides their concern for food safety and sensory quality, consumers pay increasingly more attention to aspects of animal welfare and the environment (Meuwissen & Van der Lans, 2005). Genetic modification seems to strengthen the importance of these issues (Laros & Steenkamp, 2004; Mclnerney, 1999). Consumers project their concerns with respect to the environment onto their own health situation (Bennett, D'Souza, Rosenberger & Smith, 2003; Bredahl, 1999; Frewer, Howard, Hedderley & Shepherd, 1997). Cook, Kerr and Moore (2002) and Mendenhall and Evenson (2002) found that those consumers concerned about their health were also worried about the environmental impact of GM foods.

22 Chapter 2

Table 4. Consumer concerns and benefits Type of concern/benefit Examples References A. Concerns Health Unknown long- and short-term Malarkey, 2003; Uzogara, 2000 effects; allergic reaction Food safety “Unsafe” GM products; unknown Grunert, 2002; Uzogara, 2000 effects Unnaturalness Going against nature; Cantley et al., 1999; Gaskell et al., modification of animals 2000; Moses, 2002; Townsend et al., 2004; Verhoog, 2003 Ethics Manipulation of other organisms Galun, 2001; Hails, 2004; Townsend et al., 2004 Animal welfare Responsibility for animals Mclnerney, 1999

Environment Danger to the environment; Bennett et al., 2003; Bredahl, 1999; release of GM organisms; Cook et al., 2002; Frewer et al., environmental destruction 1997; Laros et al., 2004; Mendenhall et al., 2002

B. Benefits Quality Healthier/leaner meat Pursel et al., 1996; Saek et al., 2004

Food safety Reduced residues/antibiotics Bonneau et al., 1999

Environment Fewer pesticides; less pollution Fernandez-Cornejo et al., 2006; Gilissen et al., 2000; Golovan et al., 2001 Animal welfare Fewer antibiotics; healthier Forano et al., 2000; Noble et al., growth; increased disease 2002; Paterson et al., 2003 resistant; less stress

Consumer benefits A large amount of research suggests that a very important consumer technology-adoption factor in food production is the “benefit realisation” (Baker & Burnham, 2001; Gaskell, Allum, Wagner, Kronberger, Torgersen, Hampel et al., 2004; Hossain, Onyango, Schilling, Hallman & Adelaja, 2003; Moses, 2002; Spetsidis & Schamel, 2002). Frewer et al. (1997) found that although the majority of respondents perceived genetic modification negatively, “they were willing to accept the process, if it were performed in order to produce particular benefits”. Moreover, only “relevant” consumer benefits were found to be important in consumer acceptance process of GM food (Bredahl, 1999). The first generation of GM crops used as feed in livestock production lacked “benefits to the consumers”. Although these crops had a significant influence on a new way of safer production by reducing amounts of agrochemicals applied, consumer acceptance of these GM crops was poor. The consumers

23 Adoption of GM technology in livestock production chains: an integrating framework

considered such crops primarily benefiting farmers, while giving them few or sometimes no benefits at all (Baker et al., 2001). Besides concerns, genetic modification in livestock production can also bring benefits to consumers (Table 4, part B: Benefits). Important consumer issues, such as food quality and safety, animal welfare and environment, can be influenced by different applications of genetic modification. Some applications can affect one of the issues, others tackle a group of them. For example, by using genetic modification, it is possible to produce less fatty and healthier meat, or meat that tastes better (Pursel et al., 1996; Saek et al., 2004). Also, compared to conventional production, genetic modification technology may reduce harmful residues in meat (Bonneau et al., 1999). It may furthermore help solve the pollution problem by producing less phosphorus in manure and requiring fewer chemicals in crops (Fernandez- Cornejo & Caswell, 2006; Gilissen et al., 2000; Golovan et al., 2001). In addition, genetic modification may improve animal welfare by helping animals grow healthier and with less stress (Forano et al., 2000; Noble et al., 2002; Paterson et al., 2003).

4.2 Producer acceptance The livestock production chain includes many producers, but information about adoption of genetic modification in chains is currently only available on farmers.

Producer concerns Similar to consumers, producers have concerns with respect to the adoption of new technology (Table 5, part A: Concerns). Some farmers in countries where planting GM crops is permitted have already experienced the technology adoption process. Although these are generally crop farmers, concerns among livestock farmers are expected to be rather similar.

24 Chapter 2

Table 5. Producer concerns and benefits Type of concern/benefit Examples References A. Concerns How well will it work Will investment be paid off Hillyer, 1999

Increased dependence Possible monopolisation Niemi et al., 2002; Traxler et al., on innovators 1999

High prices High prices for GM seeds, Anonymous, 2000; Fernandez- including technology fee Cornejo et al., 2002; Hillyer, 1999 Guaranteed level of Expected yield guarantee Anonymous, 2000; Carpenter et al., production 1999

Non-lasting technology Resistance of insects or weeds to Anonymous, 2000; Hillyer, 1999 GM crops Consumer acceptance Will consumers buy GM products ACGA, 2001; Chen et al., 2000; Fernandez-Cornejo et al., 2006 Uzogara, 2000

B. Benefits Cost savings Crops: lower weed control costs; Fernandez-Cornejo et al., 2002; reduced application of herbicides; Hillyer, 1999 reduced energy costs; Animals: use of lower-quality feed Noble et al., 2002; Owusu-Asiedu (reduced feed costs), animals et al., 2003; Paterson et al., 2003; grow faster and are healthier; Streatfield, et al., 2001 reduced costs of vaccination Feed additives and microorganisms: reduced feed Zhang et al., 2000 costs

Simplicity, flexibility Crops: one product for wide Anonymous, 2000; Carpenter et al., spectrum of weeds; more rotation 2002; Fernandez-Cornejo et al., options; timing of treatments; 2002; Hillyer, 1999 easier management

Increased growth Higher crop yields and growth Noble et al., 2002; Sang, 2003 rates of animals

Safer feed Reduced mycotoxin Falk et al., 2002; Munkvold et al., contamination 1999; Phipps et al., 2000

A main producer concern in applying genetic modification is to know “how well this technology will function in the future”. The application of genetic modification to livestock chains generally requires many specific investments: first in getting the technology into the chain and then maintaining the process of production. However, farmers do not feel sure about whether, or when, these investments will pay off (Hillyer, 1999). Furthermore, producer experience of using GM crops has revealed a new concern of increased dependency on the innovation firms. There is the opinion that the innovators of new

25 Adoption of GM technology in livestock production chains: an integrating framework

GM material will take control of the economic benefits generated by GM products (Niemi & Virolainen, 2002; Traxler & Falck-Zepeda, 1999). Farmers have to pay additionally a “technology fee” for the GM seeds (Hillyer, 1999). Therefore, prices of GM soybean seeds are for instance 35% higher than those of conventional seeds (Anonymous, 2000). For farmers sowing GM seeds, profitability will depend on the gain from the new weed-control programmes when set off against these higher costs (Fernandez-Cornejo, McBride, El-Osta, Heimlich, Soule, Klotz-Ingram et al., 2002). Another farmer concern is that the expected result from genetic modification cannot always be guaranteed and can be influenced by additional factors. GM seed varieties do not necessarily guarantee higher yields, as these still also depend on various growth factors, such as temperature and the weed control applied (Anonymous, 2000; Carpenter & Gianessi, 1999). Also, in the case of GM crops, farmers are concerned that insects or weeds may develop resistance to the new GM crop varieties (Hillyer, 1999). Specific protective actions could help to prevent or reduce such negative consequences, but that would raise the related costs of GM crops (Anonymous, 2000). The greatest concern, however, is the acceptance of genetic modification by consumers. The difference in the acceptance level of European and US consumers frightens farmers and stops them adopting the technology. As long as farmers cannot see the real market for such products, they will not plant or grow GM foods (ACGA, 2001; Chen & Buttel, 2000; Fernandez-Cornejo et al., 2006; Uzogara, 2000).

Producer benefits Besides producer concerns, there are also potential benefits. These are listed in Table 5, part B: Benefits. First, genetic modification can be a cost-reducing technology. Herbicide-tolerant crops are one example: crop farmers expect to achieve at least the same output while lowering weed-control costs (Hillyer, 1999; Fernandez-Cornejo et al., 2002). For livestock farmers edible vaccines, enzymes and antibodies are likely to reduce feed and vaccination costs (Owusu-Asiedu, Nyachoti, & Marquardt, 2003; Paterson et al., 2003; Streatfield et al., 2001). A reduction of feeding costs can also be achieved by using GM additives and microorganisms to enhance feed efficiency (Zhang, et al., 2000). Genetic modification applied to animals may also stimulate growth of animals (Sang, 2003). Genetic modification can, for example,

26 Chapter 2

increase milk production in lactating sows and, thereby, the growth rate of piglets (Noble et al., 2002). Another producer benefit of genetic modification is increased simplicity and flexibility of weed-control programmes. This technology allows growers to control a wide spectrum of weeds using one product instead of several herbicides (Carpenter & Gianessi, 2002; Fernandez-Cornejo et al., 2002). Furthermore there are more rotation options (Anonymous, 2000; Carpenter et al., 2002; Hillyer, 1999), and farmers have more flexibility in their timing of herbicide treatments (Anonymous, 2000; Carpenter et al., 2002). Insect-protected (Bt) crops are an application of genetic modification that benefits multiple producers in the chain: they not only increase yields by reducing insect damage, but they also reduce mycotoxin contamination of damaged grain (Falk, Chassy, Harlander, Hoban, McGloughlin & Akhlaghi, 2002; Munkvold, Hellmich & Rice, 1999). This reduces loss of grain and makes it safer for livestock (Phipps et al., 2000).

4.3 Regulations and labelling Regulations can stimulate or slow a technology adoption process. Different directives and regulations cover development, use and marketing of different GM and organisms. Legislation on genetic modification began in the early 1990s and continues. The European Union has introduced specific “legislation on genetic modification” having two main objectives, firstly, to protect citizens’ health and the environment and, secondly, to create and support a market for GM products in the EU. The main directives are:

• On the contained use of GMOs (Directive 90/219/EEC, as amended by Directive 98/81/EC);

• On the deliberate release into the environment of GMOs (Directive 2001/18/EC). This applies to the experimental release into the environment and marketing of GMOs;

• On the marketing of GM food and feed or food and feed products containing or consisting of GMOs (Regulation (EC)
1829/2003);

• On the intentional and unintentional movements of GMOs between Member States of the European Union and third countries (Regulation (EC)
1946/2003). GM products or GM organisms which are put on the market must also satisfy traceability conditions. Traceability requirements depend on whether the product consists of, or contains, GMOs, or has been produced from GMOs (Regulation (EC)
1830/2003). Besides traceability requirements, products consisting of or containing GMOs have to be labelled. All pre-packaged products consisting of or containing GMOs must have a special

27 Adoption of GM technology in livestock production chains: an integrating framework

label reading: “This product contains GMOs” or “This product contains genetically modified [(name of organism(s)]”. In the case of non-prepackaged products distributed to the final consumer or mass caterers, these words must appear on, or in connection with, the display of the product. The same rules apply to animal feed, including any compound feed that contains GM soy or corn. However, Regulation (EC)
1830/2003 does not require labelling of products such as meat, milk or eggs obtained from animals fed with GM feed or treated with GM medicines, nor are these products subject to the current traceability requirements. According to current regulations, the labelling of GM food and feed products is required if ingredients (considered individually) contain 0.9% or more of GM material (Regulation (EC)
1830/2003). As regards the regulations on GM outside the EU, the approaches taken in other non-EU countries differ greatly from the EU approach. Australia and New Zealand, in November 2000, adopted mandatory labelling with a 1 % threshold for the unintended presence of GM material. Exemptions from these labelling requirements are made for processing aids and food additives, highly refined food, flavours (in a concentration less than or equal to 1 % in the final food) and food prepared at the point of sale (Brent, Bittisnich, Brooke-Taylor, Galway, Graf, Healy et al., 2003). In Japan mandatory labelling is required when GM material is present in the top three row ingredients and accounts for 5 % or more of the total weight of the ingredients. Exemptions are made for feedstuffs, alcoholic beverages and processed foods (soy sauce, corn flakes and vegetable oils) (Carter & Gruere, 2003). In South Korea, regulations require mandatory labelling with a 3 % threshold for the top five ingredients used (Carter et al., 2003). In contrast to the countries mentioned with mandatory labelling, the Unites States, Canada and Argentina are the proponents of voluntary labelling (Carter et al., 2003).

4.4 Genetic modification and livestock chain issues The adoption of new technologies in food chains involves some issues with special importance for the adoption of genetic modification in livestock production chains. First there is the issue of the distribution of costs and benefits along the chain. The chain is a complex structure with many participants working towards the same goal - making profit (Chauhan & Proth, 2005). In practice, very often there is a problem of how to distribute fairly these profits among all participants in the chain. This also holds for the application of genetic

28 Chapter 2

modification, as often the costs of technology adoption occur at one stage of the chain, while the benefits are received at another stage. Another important chain issue in genetic modification is the quality of information shared among chain participants and presented to consumers. As in any other chain, there is always a risk of asymmetric information. This occurs when one party to the transaction has more or better information than the other party and the former uses this information for making decisions at the expense of the latter (Hobbs & Plunkett, 2000). Information asymmetry in the chain may, for instance, occur at the level of the farmer. Farmers may feed their animals GM crops or give them GM additives and medicines, but when selling their animals to the slaughterhouse, they may not (unintentionally) share this information with the buyer. As for the quality of information shared with the consumer, “GM” fits in the same category of credence attributes, i.e., those of on-farm production methods, country of origin and food safety. Consumers cannot judge the safety of food products or detect the presence of GMOs at the time of buying and consuming, and therefore have to believe information from sources like producers (Steenkamp, 1990). For these credence attributes, “GM” labelling could be a solution (Hobbs & Plunkett, 1999). Labels could help consumers make a deliberate choice. It gives them at least the freedom to choose. This labelling will lead to extra costs, among others the costs of traceability. However these costs are incurred under the current General Food Law anyway, which requires full traceability and the sharing of information among all chain participants (Meuwissen, Velthuis, Hogeveen & Huirne, 2003).

5. Discussion and conclusions This paper presents a framework for analysing the adoption of new technologies in food chains. The framework integrates consumers and chain participants and their concerns and benefits. We use the framework to review the literature on the adoption of genetic modification in livestock production chains. From our review we first conclude that not much research has been done yet on the adoption of genetic modification technology in livestock chains. Existing studies have investigated the adoption of GM crops, possibly because adoption of genetic modification in livestock production chains is comparatively much more complex. Second, we found that previous studies on the adoption of genetic modification are generally partial analysis, looking mainly at the role of the consumer or farmer. The use of genetic modification, however, should particularly be studied from an integrated point of view, since not only farmers need to adopt the technology, but also the other members of the

29 Adoption of GM technology in livestock production chains: an integrating framework

chain, including consumers. Integrated analysis should also enable us to study the distribution of costs and benefits of new technologies, such as GM along the chain. Third, only few adoption studies account for the uncertainty surrounding genetic modification. We regard uncertainty, for instance with respect to consumer acceptance and technology performance, as one of the major factors influencing the adoption of genetic modification. We therefore recommend future studies on food technology adoption, and, in particular, on the adoption of genetic modification, to develop integrated chain models including all chain participants as well as consumers; and for these models to explicitly account for uncertainty and the trade-offs between concerns and benefits, for a specific stakeholder as well as between stakeholders.

References Abadi Ghadim, A. K. & Pannell, D. J. (1999). "A conceptual framework of adoption of an agricultural innovation." Agricultural Economics, 21(2), 145-154. ABE (2003). European Views on Agricultural Biotechnology: An Overview of Public Opinion. Agricultural Biotechnology in Europe: Promoting an open and informed discussion: Issue Paper 7, April. ACGA (2001). "Corn Producers Survey GMO's and Markets." American Corn Growers Association, Washington, DC. Alfnes, F. (2004). "Stated preferences for imported and hormone-treated beef: application of a mixed logit model." European Review of Agricultural Economics, 31(1), 19-37. Anonymous (2000). "Economic Impacts of Genetically Modified Crops on the Agri-Food Sector; European Union, Commission of the European Communities." date accessed: 28 April 2006 at http://europa.eu.int/comm/agriculture/publi/index_en.htm Baker, G. A. & Burnham, T. A. (2001). "The market for genetically modified foods: consumer characteristics and policy implications." The International Food and Agribusiness Management Review, 4(4), 351-360. Barham, B. L., Foltz, J. D., Moon, S. & Jackson-Smith, D. (2004). "A Comparative Analysis of Recombinant Bovine Somatotropin Adoption across Major U.S. Dairy Regions." Review of Agricultural Economics, 26(1), 32-44.

30 Chapter 2

Bennett, B., D'Souza, G., Rosenberger, R. & Smith, D. (2003). Genetically Modified Fish and Seafood: Consumer Attitudes, Marketing Strategies, and Policy Implications. In 7th International Conference on Public Goods and Public Policy for Agricultural Biotechnology, Ravello, Italy, June 29 - July 3. Bonneau, M. & Laarveld, B. (1999). "Biotechnology in animal nutrition, physiology and health." Livestock Production Science, 59(2-3), 223-241. Braun, R. (2002). "People's concerns about biotechnology: some problems and some solutions." Journal of Biotechnology, 98(1), 3-8. Bredahl, L. (1999). "Consumers' Cognitions With Regard to Genetically Modified Foods. Results of a Qualitative Study in Four Countries." Appetite, 33(3), 343-360. Bredahl, L. (2001). "Determinants of Consumer Attitudes and Purchase Intentions With Regard to Genetically Modified Food - Results of a Cross-National Survey." Journal of Consumer Policy, 24(1), 23-61. Brent, P., Bittisnich, D., Brooke-Taylor, S., Galway, N., Graf, L., Healy, M. & Kelly, L. (2003). "Regulation of genetically modified foods in Australia and New Zealand." Food Control, 14(6), 409-416. Bruhn, C. M. (2003). "Consumer attitudes toward biotechnology: Lessons for animal-related applications." Journal of Animal Science, 81(E. Suppl.2), E196-E200. Burton, M., Rigby, D., Young, T. & James, S. (2001). "Consumer attitudes to genetically modified organisms in food in the UK." European Review of Agricultural Economics, 28(4), 479-498. Cantley, M., Hoban, T. & Sasson, A. (1999). "Regulations and consumer attitudes toward biotechnology." Nature biotechnology, 19(march), BV37 - BV40. Cardello, A. V. (2003). "Consumer concerns and expectations about novel food processing technologies: effects on product liking." Appetite, 40(3), 217-233. Carpenter, J. & Gianessi, L. (1999). "Herbicide tolerant soybeans: Why growers are adopting Roundup Ready varieties." AgBioForum, 2(2), 65-72. Carpenter, J. E. & Gianessi, L. P. (2002). Case Study in Benefits and Risks of Agricultural Biotechnology: Roundup Ready Soybeans. In V. Santaniello, R. E. Evenson&D. Zilberman, Market Development for Genetically Modified Foods (227-243). Wallingford, UK: CABI Publishing. Carter, C. A. & Gruere, G. P. (2003). "International Approaches to the Labeling of Genetically Modified Foods." Choices, 18(2), 1-4.

31 Adoption of GM technology in livestock production chains: an integrating framework

Chauhan, S. S., & Proth, J.-M. (2005). "Analysis of a supply chain partnership with revenue sharing." International Journal of Production Economics, 97(1), 45-51. Chen, L. & Buttel, F. H. (2000). "Dynamics of GMO Adoption Among Wisconsin Farmers." PATS Staff Paper Series. University of Wisconsin-Extension. Chung, C. & Pettigrew, J. E. (1998). "Economics of Soybean Biotechnology in the Livestock Industry." The International Food and Agribusiness Management Review, 1(3), 373- 385. Churchill, G. A. & Iacobucci, D. (2002). "Marketing research: methodological foundations".8th ed., Ohio: South-Western. Cook, A. J., Kerr, G. N. & Moore, K. (2002). "Attitudes and intentions towards purchasing GM food." Journal of Economic Psychology, 23(5), 557-572. Den Ouden, M. (1996). Economic modelling of pork production-marketing chains. Wageningen, Wageningen Agricultural University. Desquilbet, M., Lemarie, S., & Levert, F. (2002). Potential adoption of genetically modified rapeseed in France, effects on revenues of farmers and upstream companies: an ex ante evaluation. In 10th EAAE-congress: Exploring diversity in the European Agri- food System, Zaragoza, Spain, August 28-31. D’Souza, C. & Quazi, A. (2005). "The dynamics of exploring future market potential of genetically modified foods." Nutrition & Food Science, 35(2), 95-108. Darr, D. A. & Chern, W. S. (2002). Estimating Adoption of GMO Soybeans and Maize: a Case Study of Ohio, USA. In V. Santaniello, R. E. Evenson&D. Zilberman, Market Development for Genetically Modified Foods (141-159). Wallingford, UK: CABI Publishing. Enriquez, J. (2001). "Green biotechnology and European competitiveness." Trends in Biotechnology, 19(4), 135-139. Eurobarometer. (2005). "Social values, Science and technology." Special Eurobarometer 225 / Wave 63.1 -TNS opinion & Social, Brussels: European Comission, accessed 12 July 2005 at:http://europa.eu.int/comm/public_opinion/archives/ebs/ebs_225_report_en.pdf. Falk, M. C., Chassy, B. M., Harlander, S. K., Hoban, T. J., McGloughlin, M. N., & Akhlaghi, A. R. (2002). "Food Biotechnology: Benefits and Concerns." American Society for Nutritional Sciences, Issues and Opinions, 1384-1390. FAO (2004). Biotechnology applications in food processing: Can developing countries benefit? In Electronic Forum on Biotechnology in Food and Agriculture: Conference

32 Chapter 2

11, 14 June-15 July 2004, date accessed: 27 April 2006 at http://www.fao.org/biotech/c11doc.htm. Fernandez-Cornejo, J., & Caswell, M. (2006). "The First Decade of Genetically Engineered Crops in the United States". Economic Information Bulletin, Number 11 (April 2006), USDA. Fernandez-Cornejo, J., McBride, W. D., El-Osta, H., Heimlich, R., Soule, M., Klotz-Ingram, C., et al Eds. (2002). "Adoption of Bioengineered Crops." Agricultural Economic Report No. 810,Washington, DC, USDA. Fisher, D. K., Norvell, J., Sonka, S. & Nelson, M. J. (2000). "Understanding technology adoption through system dynamics modeling: implications for agribusiness management." The International Food and Agribusiness Management Review, 3(3), 281-296. Flett, R., Alpass, F., Humphries, S., Massey, C., Morriss, S. & Long, N. (2004). "The technology acceptance model and use of technology in New Zealand dairy farming." Agricultural Systems, 80(2), 199. Forano, E. & Flint, H. J. (2000). "Genetically modified organisms: consequences for ruminant health and nutrition." Annales de Zootechnie, 49, 255-271. Fortin, D. R. & Renton, M. S. (2003). "Consumer acceptance of genetically modified foods in New Zealand." British Food Journal, 105(1/2), 42-58. Frewer, L. (2003). "10. Societal issues and public attitudes towards genetically modified foods." Trends in Food Science & Technology, 14(5-8), 319-332. Frewer, L. J., Howard, C. & Aaron, J. I. (1998). "Consumer acceptance of transgenic crops." Pesticide Science, 52(4), 388-393. Frewer, L. J., Howard, C., Hedderley, D. & Shepherd, R. (1997). "Consumer attitudes towards different food-processing technologies used in cheese production--the influence of consumer benefit." Food Quality and Preference, 8(4), 271-280. Galun, E. (2001). "Genetically Manipulated Plants: Ethical and Religious Issues." In AgBioView Newsletter: Biotech Fearmongers; Jewish Perspective on GM; Syngenta, date accessed: 28 April 2006 at http://www.agbioworld.org/newsletter_wm/index.php?caseid=archive&newsid=1020. Gaskell, G., Allum, N., Bauer, M., Durant, J., Allansdottir, A., Bonfadelli, H., Boy, D., Cheveigné, S. d., Fjaestad, B., Gutteling, J. M., Hampel, J., Jelsøe, E., Jesuino, J. C., Kohring, M., Kronberger, N., Midden, C., Nielsen, T. H., Przestalski, A., Rusanen, T.,

33 Adoption of GM technology in livestock production chains: an integrating framework

Sakellaris, G., Torgersen, H., Twardowski, T. & Wagner, W. (2000). "Biotechnology and the European public." Nature biotechnology, 18(9), 935 - 938. Gaskell, G., Allum, N., Wagner, W., Kronberger, N., Torgersen, H., Hampel, J. & Bardes, J. (2004). "GM Foods and the Misperception of Risk Perception." Risk Analysis, 24(1), 185-194. Giddings, G., Allison, G., Brooks, D. & Carter, A. (2000). "Transgenic plants as factories for biopharmaceuticals." Nature biotechnology, 18, 1151-1155. Gilissen, L. J. W. J. & Nap, J.-P. (2000). A contribution to the discussion how a broad choice of GMO-free food products can be guaranteed and organized for consumers. Wageningen, Plant Research International: 34. Golovan, S. P., Meidinger, R. G., Ajakaiye, A., Cottrill, M., Wiederkehr, M. Z., Barney, D. J., Plante, C., Pollard, J. W., Fan, M. Z., Hayes, M. A., Laursen, J., Hjorth, J. P., Hacker, R. R., Phillips, J. P. & Forsberg, C. W. (2001). "Pigs expressing salivary phytase produce low-phosphorus manure." Nature biotechnology, 19(August), 741-745. Greene, W. H. (2003). "Econometric analysis". 5th ed., Upper Saddle River, NJ: Prentice Hall. Griffith, G. R., Vere, D. T. & Bootle, B. W. (1995). "An integrated approach to assessing the farm and market level impacts of new technology adoption in Australian lamb production and marketing systems: The case of large, lean lamb." Agricultural Systems, 47(2), 175-198. Grunert, K. G. (2002). "Current issues in the understanding of consumer food choice." Trends in Food Science & Technology, 13(8), 275-285. Grunert, K. G., Bech-Larsen, T., Lähteenmäki, L., Ueland, Ø. & Åström, A. (2004). "Attitudes towards the use of GMOs in food production and their impact on buying intention: The role of positive sensory experience." Agribusiness, 20(1), 95-107. Grunert, K. G., Lahteenmaki, L., Asger Nielsen, N., Poulsen, J. B., Ueland, O. & Astrom, A. (2001). "Consumer perceptions of food products involving genetic modification-- results from a qualitative study in four Nordic countries." Food Quality and Preference, 12(8), 527-542. Hails, R. (2004). "Bioethics for technology?" Current Opinion in Biotechnology, 15, 250-253. Hair, J. F. (2006). "Multivariate data analysis". 6th ed., Upper Saddle River, NJ: Pearson Prentice Hall. Hamstra, A. M. & Smink, C. (1996). "Consumers and biotechnology in the Netherlands." British Food Journal, 98(4/5), 34-38.

34 Chapter 2

Hayenga, M., Buhr, B., Skold, K., Johnson, S. & Grundmeier, E. (1989). Economic repercussions of porcine somatotropin on the pork industry. In Biotechnology for control of growth and product quality in swine: Implications and acceptability, Wageningen, the Netherlands, 12-14 December. Henriques, I. & Butler, L. J. (2002). The Importance of Feed Management Technologies in the Decision to Adopt Bovine Somatotropin: an Application to California Dairy Producers. In V. Santaniello, R. E. Evenson&D. Zilberman, Market Development for Genetically Modified Foods (91-99). Wallingford, UK: CABI Publishing. Hillyer, G. (1999). "Biotechnology Offers U.S. Farmers Promises and Problems." AgBioForum, 2(2), 99-102. Hoban, T. (1999). "Consumer Acceptance of Biotechnology in the United States and Japan." Food Technology, 53(5), 50-53. Hobbs, J. E. & Plunkett, M. D. (1999). "Genetically modified foods: Consumer issues and the role of information asymmetry." Canadian journal of Agricultural Economics, 47(4), 445-455. Hobbs, J. E. & Plunkett, M. D. (2000). "GMOs: The Economics of Consumer Food Safety Issues." Journal of Canadian Agricultural Economics Society, 1, 11-22. Hossain, F., Onyango, B., Schilling, B., Hallman, W. & Adelaja, A. (2003). "Product attributes, consumer benefits and public approval of genetically modified foods." International Journal of Consumer Studies, 27(5), 353-365. Kaye-Blake, W., Bicknell, K. & Saunders, C. (2005). "Process versus product: which determines consumer demand for genetically modified apples?" Australian Journal of Agricultural & Resource Economics, 49(4), 413-427. Koivisto Hursti, U.-K., Magnusson, M. K. & Algers, A. (2002). "Swedish consumers’ opinions about gene technology." British Food Journal, 104(11), 860-872. Lahteenmaki, L., Grunert, K., Ueland, O., Astrom, A., Arvola, A. & Bech-Larsen, T. (2002). "Acceptability of genetically modified cheese presented as real product alternative." Food Quality and Preference, 13(7-8), 523-533. Laros, F. J. M. & Steenkamp, J.-B. E. M. (2004). "Importance of Fear in the Case of Genetically Modified Food." Psychology and marketing, 21(11), 889–908. Lockie, S., Lawrence, G., Lyons, K. & Grice, J. (2005). "Factors underlying support or opposition to biotechnology among Australian food consumers and implications for retailer-led food regulation." Food Policy, 30(4), 399-418.

35 Adoption of GM technology in livestock production chains: an integrating framework

Loureiro, M. L. & Bugbee, M. (2005). "Enhanced GM Foods: Are Consumers Ready to Pay for the Potential Benefits of Biotechnology?" The Journal of Consumer Affairs, 39(1), 52-70. Macer, D. & Chen Ng, M. A. (2000). "Changing attitudes to biotechnology in Japan." Nature biotechnology, 18(9), 945 - 947. Malarkey, T. (2003). "Human health concerns with GM crops." Mutation Research/Reviews in Mutation Research, 544(2-3), 217-221. Marra, M., Pannell, D. J. & Abadi Ghadim, A. (2003). "The economics of risk, uncertainty and learning in the adoption of new agricultural technologies: where are we on the learning curve?" Agricultural Systems, 75(2-3), 215-234. Mclnerney, J. (1999). "Economic Pressure, Livestock Productivity and Animal Welfare- Seeking the Balance. The future developments in farm animal breeding and reproduction and their ethical, legal and consumer implications". Farm Animal Breeding and Society. Utrecht, The Netherlands, November 1999, EC-ELSA project 4th Framework Programme for RTD: 57-66. Mendenhall, C. A., &. Evenson, R. E (2002). Estimates of Willingness to Pay a Premium for Non-GM Foods: a Survey. In V. Santaniello, R. E. Evenson and D. Zilberman, Market Development for Genetically Modified Foods (55-61). Wallingford, UK: CABI Publishing. Meuwissen, M. P. M. & Van der Lans, I. A. (2005). "Trade-offs between consumer concerns: An application for pork supply chains." Acta Agriculturae Scand Section C - Food Economics,(2), 27-34. Meuwissen, M. P. M., Velthuis, A. G. J., Hogeveen, H. & Huirne, R. B. M. (2003). "Traceability and Certification in Meat Supply Chains." Journal of Agribusiness, 21(2), 167-181. Miles, S., Ueland, Ø. & Frewer, L. J. (2005). "Public attitudes towards genetically-modified food." British Food Journal, 107(4), 246-262. Moses, V. (2002). "Agricultural biotechnology and the UK public." Trends in Biotechnology, 20(9), 402-404. Munkvold, G. P., Hellmich, R. L. & L.G.Rice (1999). "Comparison of Fumonisin Concentrations in Kernels of Transgenic Bt Maize Hybrids and Nontransgenic Hybrids." Plant Disease, 83(2), 130-138.

36 Chapter 2

Nadolnyak, D. A. & Sheldon, I. M. (2002). A Model of Diffusion of Genetically Modified Crop Technology in Concentrated Agricultural Processing Markets - The Case of Soybeans.In 10th EAAE-congress: Exploring diversity in the European Agri-food System, Zaragoza, Spain, 28-31 August. Nayga Jr., R. M., Poghosyan, A. & Nichols, J. P. (2004). "Will consumers accept irradiated food products?" International Journal of Consumer Studies, 28(2), 178-185. Niemi, J. & Virolainen, M. (2002). Ex ante Economic Assessment of Adopting Genetically Engineered Crops in Finland. In V. Santaniello, R. E. Evenson&D. Zilberman, Market Development for Genetically Modified Foods (159-166). Wallingford, UK: CABI Publishing. Noble, M. S., Rodriguez-Zas, S., Cook, J. B., Bleck, G. T., Hurley, W. L. & Wheeler, M. B. (2002). "Lactational performance of first-parity transgenic gilts expressing bovine á- lactalbumin in their milk." Journal of Animal Science, 80(4), 1090-1096. Noussair, C., Robin, S. & Ruffieux, B. (2004). "Do consumers really refuse to buy genetically modified food?" The Economic Journal, 114, 102-120. O'Connor, E. & Cowan, C., Williams, G., O'Connell, J. & Boland, M. P. (2005). "Acceptance by Irish consumers of hypothetical GM dairy spread that reduces cholesterol." British Food Journal, 107(6), 361-380. O'Connor, E., Cowan, C., Williams, G., O'Connell, J. & Boland, M. P. (2006). "Irish consumer acceptance of a hypothetical second-generation GM yogurt product." Food Quality and Preference, 17(5), 400-441. Owusu-Asiedu, A., Nyachoti, C. M. & Marquardt, R. R. (2003). "Response of early-weaned pigs to an enterotoxigenic Escherichia coli (K88) challenge when fed diets containing spray-dried porcine plasma or pea protein isolate plus egg yolk antibody, zinc oxide, fumaric acid, or antibiotic." Journal of Animal Science, 81, 1790-1798. Paterson, L., DeSousa, P., Ritchie, W., King, T. & Wilmut, I. (2003). "Application of reproductive biotechnology in animals: implications and potentials. Application of reproductive cloning." Animal Reproduction Science, 79, 137-143. Phipps, R. H. & Beever, D. E. (2000). "New technology: Issues relating to the use of genetically modified crops." Journal of Animal and Feed Sciences,(9), 543-561. Pursel, V. G., Coleman, M. E., Wall, R. J., Elsasser, T. H., Haden, M., DeMayo, F. & Schwartz, R. J. (1996). "Regulatory avian skeletal [alpha]-actin directs expression of insulin-like growth factor-I to skeletal muscle of transgenic pigs." Theriogenology, 45(1), 348.

37 Adoption of GM technology in livestock production chains: an integrating framework

Qaim, M. & De Janvry, A. (2003). "Genetically Modified Crops, Corporate Pricing Strategies, and Farmers' Adoption: The Case of Bt Cotton in Argentina." American Journal of Agricultural Economics, 85(4 (November)), 814-828. Regulations (EC) on biotechnology, date accessed 28 April 2006 at http://europa.eu.int/comm/food/food/biotechnology/index_en.htm. Renko, N., Brcic´-Stipcevic´, V. & Renko, S. (2003). "Attitudes of the Croatian population toward genetically modified food." British Food Journal, 105(3), 148-161. Rogers, E. M. (2003). Diffusion of innovations. New York: Free Press. Saek, K., Matsumoto, K., Kinoshita, M., Suzuki, I., Tasaka, Y., Kano, K., Taguchi, Y., Mikami, K., Hirabayashi, M., Kashiwazaki, N., Hosoi, Y., Murata, N. and Iritani, A. (2004). Functional expression of a ∆12 fatty acid desaturase gene from spinach in transgenic pigs. PNAS 101: 3661-6366. Samaranayake, P. (2005). "A conceptual framework for supply chain management: a structural integration." Supply Chain Management: An International Journal, 10(1), 47-59. Sang, H. (2003). "6. Genetically modified livestock and poultry and their potential effects on human health and nutrition." Trends in Food Science & Technology, 14(5-8), 253-263. Schmidt, D. (2000). "Outlook for consumer acceptance of agricultural biotechnology." Nutrition, 16(7-8), 704-706. Scholderer, J. & Frewer, L. J. (2003). "The Biotechnology Communication Paradox: Experimental Evidence and the Need for a New Strategy." Journal of Consumer Policy, 26(2), 125-157. Scully, J. (2003). "Genetic engineering and perceived levels of risk." British Food Journal, 105(1/2), 59-77. Sondergaard, H. A., Grunert, K. G. & Scholderer, J. (2005). "Consumer attitudes to enzymes in food production." Trends in Food Science & Technology, 16(10), 466-474. Spetsidis, N. M. & Schamel, G. (2002). A Consumer-based Approach towards New Product Development through Biotechnology in the Agro-food Sector. In V. Santaniello, R. E. Evenson&D. Zilberman, Market Development for Genetically Modified Foods (63- 79). Wallingford, UK: CABI Publishing. Steenkamp, J.-B. E. M. (1990). "Conceptual Model of the Quality Perception Process." Journal of Business Research, 21, 309-333.

38 Chapter 2

Streatfield, S. J. (2005). "Plant-based vaccines for animal health." Scientific and Technical Review: OIE, 24(1), 189-199. Streatfield, S. J., Jilka, J. M., Hood, E. E., Turner, D. D., Bailey, M. R., Mayor, J. M., Woodard, S. L., Beifuss, K. K., Horn, M. E. & Delaney, D. E. (2001). "Plant-based vaccines: unique advantages." Vaccine, 19(17-19), 2742. Tauer, L. W. (2002). The Impact of Bovine Somatotropin on Farm Profits. In V. Santaniello, R. E. Evenson&D. Zilberman, Market Development for Genetically Modified Foods (pp. 81-90). Wallingford, UK: CABI Publishing. Traxler, G., & Falck-Zepeda, J. (1999). "The distribution of benefits from the introduction of transgenic cotton varieties." AgBioForum, 2(2), 94-98. Townsend, E. & Campbell, S. (2004). "Psychological Determinants of Willingness to Taste and Purchase Genetically Modified Food." Risk Analysis, 24(5), 1385-1393. Townsend, E., Clarke, D. D. & Travis, B. (2004). "Effects of Context and Feelings on Perceptions of Genetically Modified Food." Risk Analysis, 24(5), 1369-1384. Uzogara, S. G. (2000). "The impact of genetic modification of human foods in the 21st century: A review." Biotechnology Advances, 18(3), 179-206. Valeeva, N. I., Meuwissen, M. P. M. & Huirne, R. B. M. (2003). "Economic assessment of food safety in the dairy chain. Abstracts of the 1st Joint Annual Meeting of the American Society of Animal Science and the American Dairy Science Association with Mexican Association of Animal Production, Phoenix, Arizona, 22-26 June 2003." Journal of Animal Science, 81(Suppl.1), 132. Van der Gaag, M. A., Saatkamp, H. W., Backus, G. B. C., van Beek, P. & Huirne, R. B. M. (2004). "Cost-effectiveness of controlling Salmonella in the pork chain." Food Control, 15(3), 173. Verhoog, H. (2003). "Naturalness and the genetic modification of animals." Trends in Biotechnology, 21(7), 294-297. Vilella-Vila, M., Costa-Font, J. & Mossialos, E. (2005). "Consumer involvement and acceptance of biotechnology in the European Union: a specific focus on Spain and the UK." International Journal of Consumer Studies, 29(2), 108-118. Visscher, P., Pong-Wong, R., Whittemore, C. & Haley, C. (2000). "Impact of biotechnology on (cross)breeding programmes in pigs." Livestock Production Science, 65(1-2), 57- 70.

39 Adoption of GM technology in livestock production chains: an integrating framework

Wilcock, A., Pun, M., Khanona, J. & Aung, M. (2004). "Consumer attitudes, knowledge and behaviour: a review of food safety issues." Trends in Food Science & Technology, 15(2), 56. Winston, W. L. (2003). "Operations research: applications and algorithms". 4th ed., Belmont: Duxbury Press. Zhang, Z. B., Kornegay, E. T., Radcliffe, J. S., Wilson, J. H. & Veit, H. P. (2000). "Comparison of phytase from genetically engineered Aspergillus and canola in weanling pig diets." Journal of Animal Science, 78, 2868–2878.

40

Chapter 3

Analysis of consumer acceptance of genetically modified applications in the pork production chain

Tatiana A. Novoselova12, Ivo A.van der Lans3, Miranda P.M. Meuwissen12, Ruud B.M. Huirne12

1) Business Economics, Wageningen University, 2) Institute for Risk Management in Agriculture, Wageningen University 3) Marketing and Consumer Behaviour Group, Wageningen University

Submitted to British Food Journal

Abstract: The objective of this study is to explicitly investigate how consumers value specific GM applications that can be used in livestock production chains and whether and how consumers trade them off against the benefits that these applications can offer them. A questionnaire was designed to investigate consumer attitudes towards GM technology in pork production and trade-offs made by consumers between specific GM applications and their possible benefits. For the latter, an attribute-based stated choice modelling approach was used. The data was analysed using the nested logit model. Results indicate that conventional pork is preferred over GM pork. Among the four GM applications investigated, GM feed is most preferred. Furthermore, findings reveal that the attractiveness of different benefits offered does not influence consumer choices towards choosing GM pork over regular pork. It is only when consumers have decided to choose GM pork that their further decisions regarding which GM pork to choose depends on the benefits offered. Socio-economic and attitudinal characteristics of the consumers were found to be important determinants of consumers’ valuation of specific GM applications and their possible benefits.

Chapter 3

Introduction The number of ongoing debates in Europe concerning genetic modification (GM) has not decreased over the past few years. GM is becoming an important issue in livestock production. Different parties like consumers, producers, NGOs and policymakers engage in intense discussions about whether it is ethical, natural and safe to use new technologies such as genetic modification in food production. Several studies into the impact of GM are now available (Bonneau and Laarveld, 1999; Cunningham, 1999; Visscher et al., 2000). Many producers see – and studies indicate – great potential in such applications (Bonneau and Laarveld, 1999; Van Kempen, 2000; Ag-West Biotech Inc., 2005). Consumers, however, have concerns about genetically modified products and the technology in general (Bredahl, 1999; Cook et al., 2002; Cardello, 2003; Frewer, 2003; Fortin and Renton, 2003; Miles et al., 2005) and moral considerations also play an important role in consumer evaluation of new technologies (Eurobarometer, 2005). In Europe, the attitude towards genetic modification is still somewhat negative (Koivisto Hursti et al., 2002; Miles et al., 2005). Many studies have shown that consumers are not in favour of genetically modified organisms in their food or even of their use in food production (Bredahl, 2001). Consumer attitudes towards, and acceptance of, genetic modification has been the main topic of many studies for some years. A lot of research has been done on the driving factors influencing consumer acceptance of genetic modification (Hossain et al., 2003), on the influence of information on consumer choices – especially information related to the risks and benefits of genetic modification (Lusk et al., 2004) – and consumer willingness to pay for different genetically modified foods (Rigby and Burton, 2005). Not surprisingly, the majority of these studies explore consumers’ attitudes and perception regarding plant-based GM products since these are better accepted by consumers and supported legally by EU legislation. However, besides applications involving plants, genetic modification in livestock production has much to offer not only to the producers and industry but to the consumers and society as well (Bonneau and Laarveld, 1999; Sang, 2003). Except for the studies by Lusk et al. (2003), who investigated the demand for beef from cattle fed GM corn, and Onyango (2004) who compared acceptance of GM animal and plant food products, specific studies on consumer acceptance of GM animal-based products have not been reported. Given the great potential of genetic modification in animal production and, more importantly, the lack of consumer research on consumer acceptance of specific GM applications in this area, the objectives of this research are twofold. The first objective is to explicitly investigate how consumers value specific GM applications used in livestock

43 Analysis of consumer acceptance of GM applications in the pork production chain production chains, with the pork production chain being used as an example. Although genetic modification in livestock production chains is not restricted to the applications in pork production, the global and, particularly, European pork consumption trend suggests that pork will remain the most in-demand and consumed meat product compared to poultry products and beef (AlbertaPork, 2006; USDA, 2006). In this study, consumers were confronted with four different GM applications representing four production methods: genetic modification of the animal itself, of the feed, of the additives & medicines, and of GM bacteria used as preservatives. These applications were chosen so as to cover a whole range of possible GM applications that are currently feasible in pork production chains (Novoselova et al., 2004). The second objective is to analyse whether and how consumers make trade-offs between specific GM applications on the one hand, and the benefits that genetic modification can offer them on the other. In particular, we examine the trade-off against low price, improvements in quality, environmental impact, animal welfare and reduced residues in meat. As was observed in previous studies, consumers’ attitudes are very much driven by the particular benefits that a specific application (production method and, as a result, product) can provide them (Frewer et al., 1997b). In this study, consumer evaluations of the GM applications and their trade-off for the particular benefits are analysed using the choice-modelling framework (Louviere, 1991). The structure of the paper is as follows. The second section presents previous research on consumer acceptance of, and attitudes towards, genetic modification, and stated choice modelling studies. The third section discusses materials and methods used in this study: our sample of respondents and the design of our questionnaire, including the choice task that respondents were confronted with. In addition, this section provides a theoretical model for consumer choices. The fourth section presents empirical results of the model. In the next section the findings are discussed. The final section presents conclusions and implications for future research.

Previous consumer research on GM General consumer attitudes Previous research suggests that consumer acceptance of GM technology varies with the differences in application areas (Grunert et al., 2001). Consumers have shown more positive attitudes towards genetic modification for medical applications than for food production purposes (Gaskell et al., 2000). In general, medical applications are perceived to be more beneficial, less risky and more ethically correct than applications to food production

44 Chapter 3

(Enriquez, 2001). Frewer et al. (1997a) found that GM micro-organisms and plants were less associated with risk to health than were GM animals. According to the results of various studies (Lusk et al., 2001; Baker and Burnham, 2002; Onyango et al., 2004), consumer acceptance of genetic modification is also highly dependent on the benefits that the technology can provide. More precisely, their acceptance is strongly related to the amount of offset that they see in the offered benefits for perceived risks related to the technology (Lusk et al., 2001). Current research suggests that consumers perceive more easily the benefits that directly accrue to them than the benefits that accrue primarily to producers, such as lower production costs ( Baker and Burnham, 2002). Lusk et al. (2004) found that information about possible benefits of genetic modification can change the value that consumers place on GM foods. However, it is not only the (positive) information that is important, but consumers’ trust in organisations that disseminate information can also affect their decisions (Huffman et al., 2004). Besides studies on consumer attitudes towards various GM applications, there have been a number of studies analysing consumer willingness to pay for GM and non-GM products (Chern et al., 2002; Hossain et al., 2003; Lusk et al., 2003; Rigby and Burton, 2005). Results have suggested that consumers are willing to pay considerably more for non-GM products, although sometimes the premiums are too high (see Chern et al. (2002)).

Choice modelling studies Consumer attitude towards genetic modification has been studied by many researchers using various methods, including the grid method (Magnusson and Koivisto Hursti, 2002), experimental auctions (Lusk et al., 2004), Likert scaling (Grunert et al., 2003), contingent valuation (Chern and Kaneko, 2003) and conjoint analysis (Baker and Burnham, 2002; O'Connor et al., 2005). Among the methodologies applied to explain and predict consumer behaviour towards GM food, stated choice modelling is receiving growing recognition. The choice modelling approach represents a new type of conjoint analysis and is being widely applied in studies for new product development. For example, choice modelling has been used for analysing preferences regarding environmental issues (Horne et al., 2005), food safety improvements in the meat sector (Enneking, 2004), and hormone-treated and imported meat (Alfnes, 2004). Choice modelling is also applied to study preferences for GM foods (see, for example, studies by Hu et al. (2004) and Rigby and Burton (2005)). Burton and Pearse (2002) investigated acceptance of GM beer having reduced cost and health attributes. Using a choice modelling approach, hypothetical products were devised and

45 Analysis of consumer acceptance of GM applications in the pork production chain described for respondents as being produced using alternative GM methods. It was found that respondents were not in favour of first-generation modification in plants and micro- organisms, but some were prepared to pay more for a GM beer with medicinal benefits. In their study of consumer attitudes towards GM organisms in food in the UK, Burton et al. (2001) used choice modelling to study consumer willingness to pay to avoid GM products. In addition, they investigated the effect of different attributes such as the forms of production technology used, level of on-farm chemical use, structure of the food system, and possible food health risks on choices. Results showed that consumers preferred GM foods produced from plants modified by the introduction of genes from other plants over GM foods from plants modified by introduction of genes from both animals and plants. Hu et al. (2004) applied choice modelling to examine the trade-offs made by consumers between (perceived) risks from GM foods and potential benefits to health and environment. They identified four distinct segments of consumers: value-seeking consumers, fringe consumers, traditional consumers and anti-GM consumers. Consumers in the different segments had different perceptions of the risks associated with GM foods, and different views on the benefits.

Materials and methods The stated choice modelling approach The stated choice modelling approach is a variety of conjoint analysis, which is used in consumer research to understand and predict consumer preferences and purchasing behaviour with respect to novel products (Hair et al., 1998). It assumes that consumers’ purchase decisions are based on product characteristics, so-called attributes (e.g. price of the product), and the particular levels that that product has on these attributes, i.e. attribute levels (e.g. 0% price discount, 10% price discount and 33% price discount). The aim of the stated choice modelling approach is to estimate the contribution of attribute levels to consumer choices. To achieve this, choice sets of alternatives that are defined in terms of a combination of attribute levels are presented to the consumers. In stated choice modelling, the sets are generated according to some experimental design (Louviere, 1991; Adamowicz et al., 1998a; Adamowicz et al., 1998b).

Survey design Data was collected by means of a written questionnaire sent to 2600 random addresses. The questionnaire consisted of four parts. In the first part, respondents were asked to answer

46 Chapter 3 general questions about their buying behaviour and their consumption of pork: consumption frequency of pork as the main part of a meal, variety of pork products consumed, the price at which they typically bought pork, and the importance for them of pork attributes, for example, price, quality, nutritional value, and animal friendliness. The second part of the questionnaire asked for the respondents’ attitude towards genetic modification in pork production chains. Before answering these questions, respondents were asked to read information about genetic modification in pork production in order to become familiar with this topic and with some basic terms used later on in the questionnaire. Genetic modification was defined as “technology that involves the transfer of genes from one plant, animal or micro-organism to another plant, animal or micro-organism. The products produced with the help of genetic modification were called genetically modified products (“GM products”)”. Consumers were presented four types of genetic modification, or GM applications, in pork production, called GM animal, GM feed, GM additives & medicines and GM bacteria. GM animal was defined as an application in which genes of pigs themselves are genetically modified, so that future generations of those pigs are different. GM feed was defined as an application in which pig feed includes crops produced with the help of genetic modification. GM additives and medicines was defined as an application in which additives (e.g. vitamins, or bacteria for digestion) and medicines (e.g. vaccines, or antibiotics) are produced with the help of genetic modification. GM bacteria was defined as an application in which bacteria produced with the help of genetic modification are used for the preservation of pork after slaughter. Besides the description of four GM applications, consumers were informed about possible risks and benefits of genetic modification. Possible benefits mentioned included leaner and healthier meat, meat that tastes better or is cheaper, meat with reduced harmful residues, reduced phosphorus in hog manure, and healthier, less stressed animals. Possible risks were allergic reactions in some people, or health problems in the animals. It was also mentioned that long-term effects are not yet completely known and that genetic modification may also involve ethical concerns regarding changing the genes of animals, crops or bacteria, or concerns that GM organisms might affect wildlife biodiversity. These explanations were followed by a few questions concerning the respondent’s knowledge of genetic modification and several attitude questions on the GM applications mentioned. In this paper, we do not elaborate on the results of the first and second parts of the questionnaire. Because the objective of this study was to estimate consumer acceptance of, and willingness to, make trade-offs with respect to different applications of genetic modification

47 Analysis of consumer acceptance of GM applications in the pork production chain in livestock production chains, using methods that relied on actual market data was problematical. To eliminate this problem we used a choice modelling approach, in particular, attribute-based stated choice. An attribute-based stated choice task was presented to the consumers in the third part of the questionnaire. The design of the choice experiment, including the construction of the choice sets, is described in the next section. The last part of the questionnaire contained some socio-demographic questions used as background information to the sample description. Before sending the questionnaire out to consumers, a questionnaire pre-test was conducted to ensure that questions, additional information provided, and choice task were clear to the respondents. Moreover, the questionnaire was checked by the technical experts and industry representatives. As a result, changes to these pre-tests results were made.

Experimental design Each respondent was confronted with a number of choice sets. Each choice set consisted of three pork chops (pork chops were chosen so that respondents would know exactly what part of the pork was meant): two chops with a particular GM application (GM pork chops) and one chop to which no GM had been applied (conventional pork). In addition, the GM pork chops were varied with respect to five attributes: price, quality, animal welfare, impact on environment and amount of residues in the meat (Table 1).

Table 1. Pork attributes and attributes levels in the choice experiment Pork attributes Attribute levels Example in the questionnaire Price no reduction 10% reduction 33% reduction Quality Current Substantially improved leaner meat or longer shelf life Animal welfare Current Substantially improved less stressed animals or animals grow healthier Impact on environment Current Substantially improved less phosphorus in manure (lessening pollution problem) Presence of residues Current Substantially reduced less residues (e.g. antibiotics) in meat

The choice of these attributes was determined by two factors. First, these attributes reflect pork characteristics that can be improved with the help of genetic modification (Bonneau and Laarveld, 1999). Second, there is enough evidence to indicate that these

48 Chapter 3 attributes are important to consumers (e.g., see Verbeke and Viaene (2000), Ngapo et al. (2004) and Meuwissen and Van der Lans (2005)). Among the GM pork chops in the choice sets, the attribute “price” varied between “no price reduction”, “10% reduction” and “33% reduction”. Conventional pork always had “no price reduction.” Such attributes as “quality”, “animal welfare”, “impact on the environment” and “residues in the meat” varied between “current” and “substantially improved”. It was explained that “current” meant that the pork chop had, for example, the same quality as the pork chops one buys in the supermarket. “Substantially improved” meant, for example, the quality of the pork had been substantially improved as a result of the particular GM application. For each attribute, examples were provided to the consumers (see Table I). All five attributes varied across all the GM pork chops, except for pork chops produced with GM bacteria. By using GM bacteria it is not possible to improve animal welfare or the environment. Therefore, these attributes were excluded from the choice design for GM bacteria. For conventional pork, the attributes were always set at the current level. The choice of such general attributes and broad levels was motivated by two reasons. First, the novelty of technology – and the consequent lack of detailed data – complicate the specification of the attributes and levels. Given the limited information available on consumer acceptance of different GM applications in livestock production chains, it was difficult to select only one example and ask consumers to evaluate this specific example that still might not reflect consumer perceptions. Second, it is hard to believe that consumers have sufficient background knowledge to be able to evaluate differences between, for instance, a 75% reduction of phosphorus in manure and a 50% reduction, or a 15% reduction of pesticides and a 10% reduction, in any more objective way than they are able to evaluate “lower environmental impact”. To reduce the number of choice sets in the design, a fractional factorial main-effects-only design (Hair et al., 1998) was considered. Although by excluding interaction effects we might introduce bias into the main effect estimates, it was suggested by earlier research that main effects account for most of the stated choices (Louviere et al., 2000) and that adding interaction terms typically decreases a model’s predictive power (Hair et al., 1998). To generate the 16 choices for a particular GM application, we, first, created 16 hypothetical GM pork chops by means of a 4 x 3 x 2 x 2 x 2 x 2 orthogonal fractional factorial main-effects design (Hair et al., 1998). The first factor in this design was used as a blocking factor. The rest of the factors corresponded to the above-mentioned attributes. Second, a cyclic design was used to create 16 choice sets with two GM pork chops each, from the 16 hypothetical GM

49 Analysis of consumer acceptance of GM applications in the pork production chain pork chops. To each of these choice sets we added conventional pork as a third alternative. Conventional pork in this case was presented as a status quo and performed the role of “no- choice” option. So, for each GM application, we ended up with 16 choice sets, each with three options A, B and C (see Figure 1).

A: GM pork chop B: GM pork chop C: Conventional pork chop Price reduction of 33% No price reduction No price reduction Current quality Improved quality Current quality Current animal welfare Improved animal welfare Current animal welfare Current impact on env. Improved impact on env. Current impact on env. Current residues Reduced residues Current residues

Which pork chop do you prefer? (Tick one box) § § § Figure 1. Example of choice set

These 16 choice sets were distributed, according to the blocking factor, in four groups of four across four different versions of the questionnaire, which were administered to different groups (blocks) of respondents. In that way, each respondent saw only four choice sets per GM application, and 16 choice sets in total (4 including GM-animal alternatives, 4 including GM-feed alternatives, 4 including GM-additives & medicines alternatives, and 4 including GM-bacteria alternatives). If we had not distributed the choice sets across different versions of the questionnaire, each respondent would have had to evaluate 64 choice sets, which would have been impractical. To avoid the problem that for any version of the questionnaires a respondent could get the same block with the identical order of GM applications, we used a Greco-Latin square (Winer et al., 1991) in such a way that respondents got different choice sets for different GM applications. In addition, we varied the order of the GM applications across versions of the questionnaire.

Underlying theoretical model The analysis of the choice data was based on the random utility model (Ben-Akiva and Lerman, 1985; Louviere, 1991), which assumes that from each choice set each consumer picks the alternative with the highest utility. In this paper, a decision-making process is represented by using a two-level nested logit model. Figure 2 provides a visual depiction of the nested structure. At the upper level, consumers choose whether they would like to buy GM or non-GM pork chops, while at the lower level, they choose between two GM pork

50 Chapter 3 chops (in the case that they have chosen to buy GM pork). This process is assumed to be valid for all choice sets, regardless of GM application.

Pork chops

GM Non-GM

Option 1 Option 2 Option 3

Figure 2. Nested logit structure

It should be noted that the motivation behind the application of a nested logit model lies not only in the assumed clear structure of the process (GM vs. non-GM), but also in the violation of assumption of IIA (independence of irrelevant alternatives) 1. Therefore, to relax IIA assumption in this case, the nested logit model was considered. For a detailed description of nested logit model, see Greene, 1999, pp 865-871. The specification of a nested logit model in this case is as follows: the decision regarding which GM option to choose (the lower level choice in a nested logit model), given the condition that GM is chosen, is defined as a multinomial logit model, and can be expressed as

vGMk P(k GM ) = e ƒeVGMl l∈ DGM where P(k|GM) is the probability of choosing GM alternative k from two GM alternatives, given the condition that GM pork has been chosen at the upper level

DGM is the subset of GM-pork chops in the choice set

VGMk is the deterministic portion of utility associated with GM-pork chop k and defined as a function of the attributes.

1 IIA assumes that if a choice option of the choice set is truly irrelevant (i.e. same option twice), omitting it from the model will not change the parameters estimate and will not lead to inconsistency.

51 Analysis of consumer acceptance of GM applications in the pork production chain

In our case:

VGMk = β1 priceGMk + β2qualityGMk + β3animalwGMk +β4environGMk + β5residuesGMk

To simplify the formula, we denote this as

VGMk = βG′ M AGMk where
is the vector of model parameters A is the vector of attributes.

Therefore

βG′ M AGMk P(k GM ) = e ƒeβG′ Ml l∈ DGM The predicted probability of choosing GM pork (upper level choice probability in the nested model) can be expressed as eγ i yi +δG′ M x +τ GM IGM P(GM ) = γ y +δ ′ x +τ I e i i GM GM GM +1 where

γ i is difference in utility from a particular GM application and regular pork, representing an alternative specific constant for GM application i yi is the dummy for the alternative specific constant δ is the vector of the parameters to be estimated, which gives the contribution of both socio- economic and attitudinal characteristics x is a vector of the socio-economic and attitudinal characteristics of the individual

52 Chapter 3

For this study, δx is defined as:

δx = δ1 Male + δ 2 Age 23-39 + δ 3 Age 40-49 + δ 4 Age 50-59 + δ 5 Household 1 +

δ 6 Household 2 + δ 7 Children + δ 8 Income <12,000 + δ 9 Income 12,000-24,000 +

δ10 Income 24,000-36,000 + δ11 Income 36,000-48,000 + δ12 Education 1

+ δ13 Education 2 + δ14 Information + δ15 Familiarity + δ16 Attitude

τ GM is the inclusive value parameter

IGM is the inclusive value (or logsum) associated with the upper level alternative GM and defined as

′ I βGM AGMl GM = ln ƒe l=DGM For the consistency of the model with random utility theory, the inclusive value parameter, , should lie between zero and one (Ben-Akiva and Bierlaire, 1999). In this case, the decision of whether or not to choose GM pork is influenced by the expected utility of GM pork. If =1, the decision whether or not to choose GM pork is determined by both socio-economic and attitudinal characteristics of the individual and the expected utility of GM pork. In this case, the structure collapses to a multinomial logit model without a nested structure. If =0, the expected utility of GM pork does not influence the decision regarding whether or not to choose GM. The socio-economic and attitudinal variables were included in the model in order to examine whether the differences across the sample with respect to these characteristics might have an effect on the utility of GM applications. The descriptive statistics and definitions of these variables are presented in Table 2.

53 Analysis of consumer acceptance of GM applications in the pork production chain

Table 2. Descriptive statistics of explanatory variables Variable Description of the variable Mean Std.Dv. Socio-economic characteristics Male 1 = respondent is male; 0 = female 0.462 0.499 Age 1 23-39 1 = age between 25-39 years; 0 = otherwise 0.260 0.439 Age 2 40-49 1 = age between 40-49 years; 0 = otherwise 0.204 0.403 Age 3 50-59 1 = age between 50-59 years; 0 = otherwise 0.252 0.434 Age 4 >60* 1 = age more than 60 years; 0 = otherwise 0.284 0.451 Children 1 = children in household; 0 = no children 0.645 0.479 Household 1 1 = 1 person in household; 0 = otherwise 0.180 0.384 Household 2 1 = 2 persons in household; 0 = otherwise 0.476 0.499 Household >2* 1 = more than 2 persons in household; 0 = 0.344 0.475 otherwise Income <12 000 1 = respondent with annual income less than €12 0.067 0.251 000; 0 = otherwise Income 12 000-24 000 1 = respondent with annual income between €12 0.312 0.463 000 - €24 000; 0 = otherwise Income 24 000-36 000 1 = respondent with annual income between €24 0.363 0.481 000 - €36 000; 0 = otherwise Income 36 000-48 000 1 = respondent with annual income between €36 0.156 0.363 000 - €48 000; 0 = otherwise Income > 48 000* 1 = respondent with annual income more than 0.101 0.302 €48 000; 0 = otherwise Education 1 (Low level) 1 = respondents with low level of education; 0 = 0.161 0.368 otherwise Education 2 (Middle 1 = respondents with middle level of education; 0 0.319 0.466 level) = otherwise Education 3 (High level)* 1 = respondents with high level of education; 0 = 0.520 0.500 otherwise

Attitudinal characteristics Information 1 = respondent heard information about GM in 0.733 0.443 pork production; 0 = otherwise Familiarity 1= that respondent is familiar with GM products; 0.332 0.472 0 =otherwise Attitude Likert-scale 1 = that respondent has a negative 1.940 0.996 attitude to 5 = that respondent has a positive attitude 1) * implies that the variable is the reference level and was dropped to avoid dummy trap 2) Some categories of “Age” and “Household” characteristics were merged due to the low number of respondents in one of the categories. “Education” for this analysis was grouped according to the classification of the Dutch Association for Market and Marketing Researchers. Therefore, the categories and corresponding values presented here may differ from those in Table 3.

54 Chapter 3

Respondents In the autumn of 2004, 2600 surveys were mailed to a sample of addresses in the Netherlands obtained randomly from an electronic telephone book. Respondents were selected from 26 regions across the country. The sample distribution across regions was proportional to the population distribution among regions. After 10 days, a reminder was sent. After adjusting for undelivered surveys and excluding individuals who did not complete the questionnaire, the response rate was 11%. Such a low response rate is not surprising. The difficulty of the topic was mentioned during the pre-test of the questionnaire and it had already been observed that surveys in the Netherlands addressed to random respondents do not obtain high response rates (Stoop, 2005). In total, 253 usable questionnaires were obtained. The sample (53.8% females and 46.2% males) was representative of the Dutch population with respect to gender. The sample was not fully representative with respect to age, household size, number of children in household or education level. In our sample, respondents had on average a smaller household, fewer children in the household, and a much higher education level (Table 3). Thus, we can not rule out the possibility of some sample-selection bias in our results. Unfortunately, we can not test for any differences between respondents and non-respondents because of a lack of information on non-respondents. It is likely that the sample-selection bias is, to some extent, due to the complexity of the topic, i.e. GM technology in pork production.

55 Analysis of consumer acceptance of GM applications in the pork production chain

Table 3. Socio-economic characteristics of the sample Sample (n=253) Population P-value Number % % Gender a Female 135 53.8 50.5 0.298 Male 116 46.2 49.5

Age (years) a <24 5 2.0 8.0 0.000 25-39 60 24.0 28.9 40-49 51 20.4 20.8 50-59 63 25.2 18.5 >60 71 28.4 23.8

Children in household b Yes 87 35.5 50.5 0.000 No 158 64.5 49.5

Household size (persons) b 1 45 18 34.4 0.000 2 119 47.6 32.7 3-5 84 33.6 32.9 >5 2 0.8 --

Income (Euros) <12 000 16 6.8 -- -- 12 000-24 000 74 31.2 -- 24 000-36 000 86 36.3 -- 36 000-48 000 37 15.6 -- >48 000 24 10.1 --

Education c Primary education 15 6.0 12.5 0.000 Junior general secondary education 15 6.0 10.0 Senior general secondary education 24 9.7 6.5 Pre-vocational education 10 4.0 14.8 Senior vocational education 55 22.2 32.7 Vocational colleges 96 38.7 16.2 University education 33 13.3 7.3

a Statistics Netherlands, for 2005 year b Statistics Netherlands, for 2004 year c Statistics Netherlands, for 2002 year

56 Chapter 3

Results The analysis was based on 3583 choice sets (i.e., GM animal: 896; GM feed: 895; GM additives & medicines: 896; GM bacteria: 896). Table 4 shows that the majority of respondents (about 55%) chose the “non-GM pork chop” option when they were confronted with the applications of GM animal, GM additive & medicines and GM bacteria. However, with regard to the GM feed application, the majority of the respondents (about 54%, sum of the percentages for options 1 and 2) opted for the GM pork chop.

Table 4. Choice frequencies per GM application GM additives & Choice options GM animal GM feed GM bacteria medicines

184 210 173 170 GM pork (option 1) (20.54%) (23.46%) (19.31%) (18.97%)

232 275 230 232 GM pork (option 2) (25.89%) (30.73%) (25.67%) (25.89%)

480 410 493 494 Non-GM pork (option 3) (53.57%) (45.81%) (55.02%) (55.13%)

The data was analysed using the nested logit model according to the nested structure in Figure 2, using the MDC procedure in SAS. Two different nested logit models were estimated: i) a model in which the utility of GM pork chops depends on main effects of the specific GM application and of socio-economic and attitudinal variables at the upper level of the nested structure, and on main effects of the benefits at the lower level; ii) a model including all effects included in the previous model, and in addition as well as an interaction effect between the GM applications and each of the benefits at the lower level. The latter model was estimated to see whether the effect of each of the benefits on the utility of GM pork chops depends upon the GM application that produces it, which is something we might observe when respondents imagine different kinds of improvements, for instance in quality, animal welfare, environments and residues when confronted with different GM applications. The fit of the second model was not significantly better than that of the first model (see Table 5; difference in -2LL = 2, df = 18, P>0.05). Furthermore, both the AIC as well as Schwarz criterion favour the first model over the second. Therefore, only the estimates from the first model are presented here.

57 Analysis of consumer acceptance of GM applications in the pork production chain

Table 5. Fit of nested logit model Model 1 Model 2 (main effects only) (main effects and interactions) LL -3112 -3111 Df 3557 3539 AIC 6287 6312 Schwarz Criterion 6445 6591

The estimates of the main effects of the kind of GM application, the benefits, the socio- economic characteristics and the attitudinal variables are reported in Table 6. Note that these estimates express the contribution to overall utility relative to reference levels, of which the contribution to overall utility is set at zero by definition. For each of the benefits, ‘no improvement’ is taken as the reference level. The contribution of the different GM applications is taken relative to the contribution of conventional pork. As regards socio- economic characteristics and attitudinal variables, the last level of each variable was taken as reference level (see Table 6). The model shows plausible estimates and acceptable goodness-of-fit measures. The inclusive value parameter does not significantly differ from zero, indicating that the expected utilities of different benefits of the GM-pork alternatives do not influence consumer choices between GM and regular pork at the upper level. However, when consumers decided to choose GM pork at the upper level, their further decisions concerning which GM option to choose are based on the benefits offered. In general, results indicate that consumers attach positive utility to improvements in quality, animal welfare, environment, fewer residues and price reduction. Improvements in animal welfare have the strongest effect on consumer choice, while improvements in the environment have the weakest effect. According to our expectations, consumers attach positive utility to price discounts2. In addition, results show that consumers derive more utility from conventional pork than from GM pork, everything else being equal. All GM applications have a negative utility. These are the utilities that consumers attach to the different kinds of GM pork without benefits, relative to conventional pork. However, these utilities should be interpreted regarding consumers at the reference level of all socio-economic characteristics and attitudinal variables. For example, the utility of GM animal equals -3.0222 and is relevant for the female consumer, who is more than 60 years old, with a high income and well educated. Among the four GM applications, GM feed is most

2 In our nested logit model, we estimated a linear effect of price discounts. In Table VI the estimated utility of a price discount of 1% is presented (0.0241). From this, the estimated utilities of different price discounts can be calculated. For example, the estimated utility of a price discount of 10% is 0.0241*10=0.241

58 Chapter 3 preferred. Pairwise comparisons revealed that the utility of GM feed is significantly different from the utilities of GM animal, GM additives & medicines, and GM bacteria.

Table 6. Nested logit model parameter estimates Variable Utility estimate Standard error t-value Main effects of GM applications GM animala 2) -3.0222 0.4380 -6.90***1) GM feedabc -2.6467 0.4356 -6.08*** GM additives & medicinesb -3.0908 0.4369 -7.07*** GM bacteriac -2.9972 0.3628 -8.26***

Main effects of benefits 1% Discount 0.0241 0.0026 8.92*** Improved quality 0.3871 0.0558 6.94*** Improved animal welfare 0.9365 0.0663 14.13*** Improved environment 0.1456 0.0651 2.24* Reduced residues 0.5222 0.0564 9.62***

Socio-economic characteristics Male -0.2011 0.0812 -2.48 Age 1 23-39ab 3) 0.7335 0.1185 6.19 Age 2 40-49cd 0.8137 0.1324 6.15 Age 3 50-59ace 0.2180 0.1100 1.98 Age 4 >60bde 0 -- -- Children -0.0110 0.1495 -0.07 Household 1f 0.3217 01815 1.77 Household 2f -0.0317 0.1510 -0.21 Household 3 0 -- -- Income <12 000g 0.5409 0.2109 2.56 Income 12 000-24 000hi 0.5197 0.1495 3.48 Income 24 000-36 000hj 0.3212 0.1384 2.32 Income 36 000-48 000k 0.4263 0.1516 2.81 Income > 48 000gijk 0 -- -- Education 1 (Low level) -0.1348 0.1234 -1.09 Education 2 (Middle level)l -0.2149 0.0915 -2.35 Education 3 (High level)l 0 -- --

Attitudinal characteristics Information 0.0934 0.0670 1.39 Familiarity -0.2163 0.0821 -2.63** Attitude 0.9012 0.0426 21.17***

IV parameter 0.1563 0.1596 0.98

Goodness-of-Fit Measures Likelihood Ratio 1648.5 Df=26 McFadden’s LRI 0.2094 1) * P<0.05; ** P< 0.01; *** P<0.0001 2) All pairwise differences between GM applications have been tested. GM applications that are significantly different from each other share the same superscript character, P<0.05. 3) For each categorical variable, all pairwise differences have been tested. Variables for which there are statistically significant different effects share the same superscript character, P<0.05.

59 Analysis of consumer acceptance of GM applications in the pork production chain

With respect to socio-economic variables, the model offers plausible coefficient estimates. In general, male consumers seem less likely to choose GM pork alternatives. Younger consumers (23-39 and 40-49 years of age) show a greater propensity to choose GM pork than those aged 50-59 and above 60 years of age. Consumers aged 50-59 would opt more for GM pork than those of above 60 years old. The presence of children in the household has no significant effect on consumer choices. Regarding household size, consumers who live alone are more likely to choose GM pork than consumers in a household of two. Having a bigger household size (more than two persons) has no significant effect on consumer probability to choose GM pork. Income is found to be a significant variable in explaining consumer choices. The obtained results revealed that consumers with the highest household income (more than €48 000 a year) are least likely to buy GM pork compared to the consumers in other income categories. Consumers with a household income of less than €12 000 a year, €12 000-24 000 a year, €24 000-36 000 a year and €36 000-48 000 a year respectively, are more likely to buy GM pork. Having a household income of €12 000-24 000 a year and €24 000-36 000 a year have the same effect on consumer likelihood of buying GM pork. With respect to the level of education, consumers who completed university and vocational colleges (high level) are more likely to consume GM pork than those that attained a middle level of education. Consumers with a low level of education do not significantly differ from either of these groups in their likelihood of consuming GM pork. Information that respondents heard about GM in pork production has no significant effect on consumer choices towards GM pork. However, familiarity with GM products has a significant negative influence on the probability of buying GM pork. The attitude variable was found to have a positive significant impact on consumer choices regarding GM pork.

Discussion On average, the results indicate that conventional pork is preferred over GM pork. Among the four GM applications, GM feed is most preferred, which is not surprising given that GM plants and micro-organisms were found to be associated with less risk than GM animals (Hamstra and Smink, 1996). GM animal, GM additives & medicines and GM bacteria were found to be less preferred and did not differ from each other with regard to their contribution to overall utility. This might for instance be because consumers perhaps can hardly distinguish between these three applications. Another reason can be that consumers perhaps can distinguish between them but they are indifferent with respect to the distinctions when it

60 Chapter 3 comes to choices from among different pork chops. Regarding GM animal, our results are somewhat more positive compared to the findings of Hamstra and Smink (1996). They found that consumers in the Netherlands generally rejected products based on GM animals (pig and carp). However, in our case, GM animal application was not fully rejected by the consumers, about 46% chose this GM options over conventional pork (see Table 4). Unexpectedly, consumers preferred GM additives & medicines and GM bacteria less than GM animal. It might be the case that poor consumer knowledge about animal production systems (Kanis et al., 2003) and possible scepticism about generally used terms, e.g. additives, as was demonstrated regarding UK consumers (see FSA report, 2005) influenced consumer choices with respect to these applications. Consumers’ lack of general knowledge of production systems and technology may lead to confusion when they are presented with some technological innovations (see an example of ‘the adding of bacteriocins’ processing technology in the study by Cardello (2003)). Furthermore, the findings reveal that different benefits offered by GM pork did not influence consumer choices towards choosing GM pork over regular pork. It was only when consumers decided to choose GM pork that their further decisions regarding which GM option to choose were based on the benefits offered. The general attitude towards GM had a greater influence on the purchasing decisions, and the presence of various benefits was of secondary importance in consumer’s decision making process. These results are in line with the findings of Bredahl (2001), which demonstrate that general attitude towards genetic modification in food production influences perceived risks and benefits of the technology and, as a result, the consumer purchase decisions. Apparently, it is only when a positive general attitude towards technology has been formed, that consumer will choose some GM pork alternative and the benefits will determine which GM pork alternative they buy. In the situation in which consumers decided to choose GM pork, of all the possible benefits, consumers valued improvements in animal welfare the most, and reductions in environmental impact the least. The increasing importance of the animal welfare issue in livestock production chains was previously documented by Verbeke and Viaene (2000) and confirmed by Meuwissen and Van der Lans (2005). In line with previous studies, the socio-economic and attitudinal characteristics of the consumers were found to be important determinants in explaining consumer choices, particularly those that denote gender, age, household income, household size, education, familiarity with GM products and attitude. In contrast to previous findings (see, for example, studies by Pardo et al. (2002) and Hossain et al. (2003) our results reveal that male consumers

61 Analysis of consumer acceptance of GM applications in the pork production chain seem less likely to choose GM pork alternatives. Younger consumers (23-39 and 40-49 years of age) are more likely to choose GM pork than those of 50-59 years and of above 60 years of age. Older consumers (more than 60 years old) were, in general, less likely to choose GM pork. In the context of the American and European consumers, Hossain et al. (2003) and Pardo et al. (2002) also found younger consumers to be more willing to choose GM products. Consumers who live alone showed a greater propensity to choose GM pork than consumers in a household of two. Income is found to be a significant variable in explaining consumer choices. The obtained results revealed that consumers with the highest household income (more than €48 000 a year) are less likely to buy GM pork compared to the consumers in other income categories. Regarding level of education, consumers that completed university and vocational colleges (high level) are more likely to consume GM pork than those that attained a middle level of education. These results are in line with the findings of Pardo et al. (2002), who found that a higher level of education was positively related to consumer perception of benefits from genetic modification, and Hossain et al. (2003) who revealed that college-educated consumers are more likely to accept GM food products. Among attitudinal characteristics, familiarity with GM products has a significant negative influence on the probability of buying GM pork. It seems that people who dislike GM products are more aware of GM. Up till now, the information that was available to the consumers appeared to be more confusing and negative (Loureiro and Bugbee, 2005). On the other hand, the significant positive sign of the attitude variable suggests that respondents with a positive attitude and feeling about GM technology are more likely to buy GM pork alternatives than those respondents that have a negative attitude, which was of course to be expected. Regarding the limitations of this study, it is acknowledged that before respondents actually stated their choices they were provided with additional information on genetic modification. Although, we provided them with both positive as well as with negative information, this may still have had an influence on subsequent stated choices.

Conclusions and implications for future research This paper presents the results of a study that employs the stated choice modelling approach to evaluate consumer preferences for GM pork chops and trade-offs made by consumers between different GM applications and benefits. A major contribution of the paper is our investigation of consumers’ acceptance of genetic modification in meat production by using different GM applications that capture the whole production chain; the design explicitly

62 Chapter 3 allows for a trade-off between different alternative applications of genetic modification and benefits. In general, the present results indicate that genetic modification in livestock production has potential, particularly in the case of GM feed. However, results also reveal that consumer decisions are mainly affected by the general attitude to the technology. If the attitude towards GM is positive, then acceptance of different GM applications is more likely. Importantly, this implies that all the efforts from the producers and other chain participants who like to introduce GM should be directed to improving existing attitudes, and creating a positive image of technology by providing additional information to the consumers. Although Frewer et al. (2003) found that providing information to the consumers had little effect on attitudes towards GM foods, the two-way communication between scientific experts, industry and general public may have a significant positive influence on the formation of consumers’ attitudes (Hamstra and Smink, 1996). Therefore, much research is needed to understand how to change general consumer attitude towards GM technology. From this study, it is clear that the presence of the important consumer benefit, in this case improvements of animal welfare, had a positive influence on consumers’ choices when choosing from different GM pork alternatives. However, improvements of animal welfare were formulated at a very general and abstract level in this study. Therefore, it would be worthwhile to study consumers’ preferences for more important benefits like animal welfare by using experiments in which, specific examples of possible improvements of animal welfare are applied.

References Adamowicz, W., Louviere, J. and Swait, J. (1998a), “Introduction to Attribute-Based Stated Choice Methods”, Final Report Submitted to Resource Valuation Branch, Damage Assessment Center, NOAA, Edmonto, Alberta. Adamowicz, W., Boxall, P., Williams, M. and Louviere, J. (1998b), "Stated Preference Approaches for Measuring Passive Use Values: Choice Experiments and Contingent Valuation", American Journal of Agricultural Economics, Vol. 80, pp. 64-75. Ag-West Biotech Inc. (2005), "Feed Biotechnology: The Science of Animal Nutrition", Saskatchewan Agricultural Biotechnology Information Centre (SABIC), date accessed: August 2006 at http://www.agwest.sk.ca/publications/ls_feed.php.

63 Analysis of consumer acceptance of GM applications in the pork production chain

AlbertaPork (2006), "Global pork production: meeting the challenge in a changing world", Industry Report, date accessed: July 2006 at http://www.albertapork.com/Uploads/Objects/IndRptFeb2006.pdf. Alfnes, F. (2004), "Stated preferences for imported and hormone-treated beef: application of a mixed logit model", European Review of Agricultural Economics, Vol. 31 No. 1, pp. 19-37. Baker, G. A. and Burnham, T. A. (2002), "The market for genetically modified foods: consumer characteristics and policy implications", The International Food and Agribusiness Management Review, Vol. 4 No. 4, pp. 351-60. Ben-Akiva, M. and Bierlaire, M. (1999), "Discrete choice methods and their applications to short-term travel decisions", In Hall, R. (ed) Handbook of Transportation Science, Kluwer, pp.5-34. Ben-Akiva, M. and Lerman, S. R. (1985), Discrete Choice Analysis: Theory and Application to Travel Demand, MIT Press, Cambridge. Bonneau, M. and Laarveld, B. (1999), "Biotechnology in animal nutrition, physiology and health", Livestock Production Science, Vol. 59 No. 2-3, pp. 223-41. Bredahl, L. (1999), "Consumers' Cognitions With Regard to Genetically Modified Foods. Results of a Qualitative Study in Four Countries", Appetite, Vol. 33 No. 3, pp. 343-60. Bredahl, L. (2001), "Determinants of Consumer Attitudes and Purchase Intentions With Regard to Genetically Modified Food - Results of a Cross-National Survey", Journal of Consumer Policy, Vol. 24 No. 1, pp. 23-61. Burton, M. and Pearse, D. (2002), "Consumer Attitudes Towards Genetic Modification, Functional Foods, and Microorganisms: A Choice Modeling Experiment for Beer", AgBioForum, Vol. 5 No. 2, pp. 51-58. Burton, M., Rigby, D., Young, T. and James, S. (2001), "Consumer attitudes to genetically modified organisms in food in the UK", European Review of Agricultural Economics, Vol. 28 No. 4, pp. 479-98. Cardello, A.V. (2003), "Consumer concerns and expectations about novel food processing technologies: effects on product liking", Appetite, Vol. 40 No. 3, pp. 217-33. Chern, W. S. and Kaneko, N. (2003), "Contingent valuation of the willingness to pay for selected genetically modified foods: a multicountry analysis", In 7th International Conference on Public Goods and Public Policy for Agricultural Biotechnology, June 29-July 3, Ravello.

64 Chapter 3

Chern, W. S., Rickertsen, K., Tsuboi, N. and Tsu-Tan-Fu (2002), "Consumer Acceptance and Willingness to Pay for Genetically Modified Vegetable Oil and Salmon: A Multiple- Country Assessment", AgBioForum, Vol. 5 No. 3, pp. 105-12. Cook, A. J., Kerr, G. N. and Moore, K. (2002), "Attitudes and intentions towards purchasing GM food", Journal of Economic Psychology, Vol. 23 No. 5, pp. 557-72. Cunningham, E. P. (1999), "The application of biotechnologies to enhance animal production in different farming systems", Livestock Production Science, Vol. 58 No. 1, pp. 1-24. Enneking, U. (2004), "Willingness-to-pay for safety improvements in the German meat sector: the case of the Q&S label", European Review of Agricultural Economics, Vol. 31 No. 2, pp. 205-23. Enriquez, J. (2001), "Green biotechnology and European competitiveness", Trends in Biotechnology, Vol. 19 No. 4, pp. 135-39. Eurobarometer (2005), "Social values, Science and technology", Special Eurobarometer 225 / Wave 63.1 -TNS opinion & Social, Press release, June 2005, European Commission, Brussels, date accessed: July 2006 at http://europa.eu.int/comm/public_opinion/archives/ebs/ebs_225_report_en.pdf. Food Standards Agency (FSA) (2005), "Marketing terms", Report, date accessed: August 2006 at http://www.food.gov.uk/multimedia/pdfs/marktermfinalrep0106.pdf. Fortin, D. R. and Renton, M. S. (2003), "Consumer acceptance of genetically modified foods in New Zealand", British Food Journal, Vol. 105 No. 1/2, pp. 42-58. Frewer, L. (2003), "10. Societal issues and public attitudes towards genetically modified foods", Trends in Food Science & Technology, Vol. 14 No. 5-8, pp. 319-32. Frewer, L. J., Howard, C. and Shepherd, R. (1997a), "Public Concerns in the United Kingdom about General and Specific Applications of Genetic Engineering: Risk, Benefit, and Ethics", Science, Technology, & Human Values, Vol. 22 No. 1, pp. 98-124. Frewer, L. J., Howard, C., Hedderley, D. and Shepherd, R. (1997b), "Consumer attitudes towards different food-processing technologies used in cheese production--the influence of consumer benefit", Food Quality and Preference, Vol. 8 No. 4, pp. 271- 80. Frewer, L. J., Scholderer, J. and Bredahl, L. (2003), "Communicating about the Risks and Benefits of Genetically Modified Foods: The Mediating Role of Trust", Risk Analysis, Vol. 23 No. 6, pp. 1117-33.

65 Analysis of consumer acceptance of GM applications in the pork production chain

Gaskell, G., Allum, N., Bauer, M., Durant, J., Allansdottir, A., Bonfadelli, H., Boy, D., Cheveigné, S. d., Fjaestad, B., Gutteling, J. M., Hampel, J., Jelsøe, E., Jesuino, J. C., Kohring, M., Kronberger, N., Midden, C., Nielsen, T. H., Przestalski, A., Rusanen, T., Sakellaris, G., Torgersen, H., Twardowski, T. and Wagner, W. (2000), "Biotechnology and the European public", Nature biotechnology, Vol. 18 No. 9, pp. 935-38. Greene, W.H. (1999), Econometric Analysis, Prentice Hall, New Jersey. Grunert, K. G., Bredahl, L. and Scholderer, J. (2003), "Four questions on European consumers' attitudes toward the use of genetic modification in food production", Innovative Food Science & Emerging Technologies, Vol. 4 No. 4, pp. 435-45. Grunert, K. G., Lahteenmaki, L., Asger Nielsen, N., Poulsen, J. B., Ueland, O. and Astrom, A. (2001), "Consumer perceptions of food products involving genetic modification-- results from a qualitative study in four Nordic countries", Food Quality and Preference, Vol. 12 No. 8, pp. 527-42. Hair, J. F., Anderson, R. E., Tatham, R. L. and Black, W. C. (1998), Multivariate data analysis, Prentice-Hall International Inc., London. Hamstra, A. M. and Smink, C. (1996), "Consumers and biotechnology in the Netherlands", British Food Journal, Vol. 98 No. 4/5, pp. 34-38. Horne, P., Boxall, P. C. and Adamowicz, W. L. (2005), "Multiple-use management of forest recreation sites: a spatially explicit choice experiment", Forest Ecology and Management, Vol. 207 No. 1-2, pp. 189-99. Hossain, F., Onyango, B., Adelaja, A., Schilling, B. and Hallman, W. (2003), "Consumer Acceptance of Food Biotechnology: Willingness to Buy Genetically Modified Food Products", Journal of International Food & Agribusiness Marketing, Vol. 15 No. 1/2, pp. 53-76. Hu, W., Hünnemeyer, A., Veeman, M., Adamowicz, W. and Srivastava, L. (2004), "Trading off health, environmental and genetic modification attributes in food", European Review of Agricultural Economics, Vol. 31 No. 3, pp. 389-408. Huffman, W. E., Rousu, M., Shogren, J. F. and Tegene, A. (2004), "Who Do Consumers Trust for Information: The Case of Genetically Modified Foods?", American Journal of Agricultural Economics, Vol. 86 No. 5, pp. 1222-29. Kanis, E., Groen, A. F. and Greef, K. H. D. (2003), "Societal Concerns about Pork and Pork Production and Their Relationships to the Production System", Journal of Agricultural and Environmental Ethics, Vol. 16 No. 2, pp. 137-42.

66 Chapter 3

Koivisto Hursti, U.-K., Magnusson, M. K. and Algers, A. (2002), "Swedish consumers’ opinions about gene technology", British Food Journal, Vol. 104 No. 11, pp. 860-72. Loureiro, M. L. and Bugbee, M. (2005), "Enhanced GM Foods: Are Consumers Ready to Pay for the Potential Benefits of Biotechnology?", The Journal of Consumer Affairs, Vol. 39 No. 1, pp. 52-70. Louviere, J. J. (1991), "Experimental Choice Analysis: Introduction and Overview", Journal of Business Research, Vol. 23, pp. 291-97. Louviere, J. J., Hensher, D. A. and Swait, J. D. (2000), Stated Choice Methods: Analysis and Application, Cambridge University Press, Cambridge. Lusk, J. L., House, L. O., Valli, C., Jaeger, S. R., Moore, M., Morrow, J. L. and Traill, W. B. (2004), "Effect of information about benefits of biotechnology on consumer acceptance of genetically modified food: evidence from experimental auctions in the United States, England, and France", European Review of Agricultural Economics, Vol. 31 No. 2, pp. 179-204. Lusk, J. L., Moore, M., House, L. O. and Morrow, B. (2001), "Influence of brand name and type of modification on consumer acceptance of genetically engineered corn chips: a preliminary analysis", The International Food and Agribusiness Management Review, Vol. 4 No. 4, pp. 373-83. Lusk, J. L., Roosen, J. and Fox, J. A. (2003), "Demand for Beef from Cattle Administered Growth Hormones or Fed Genetically Modified Corn: A Comparison of Consumers in France, Germany, the United Kingdom, and the United States", American Journal of Agricultural Economics, Vol. 85 No. 1, pp. 16-29. Magnusson, M. K. and Koivisto Hursti, U.-K. (2002), "Consumer attitudes towards genetically modified foods", Appetite, Vol. 39 No. 1, pp. 9-24. Meuwissen, M. P. M. and Van der Lans, I. A. (2005), "Trade-offs between consumer concerns: An application for pork supply chains", Acta Agriculturae Scand Section C - Food Economics, Vol. 2 No. 1, pp. 27-34. Miles, S., Ueland, Ø. and Frewer, L. J. (2005), "Public attitudes towards genetically-modified food", British Food Journal, Vol. 107 No. 4, pp. 246-62. Ngapo, T. M., Martin, J. F. and Dransfield, E. (2004), "Consumer choices of pork chops: results from three panels in France", Food Quality and Preference, Vol. 15 No. 4, pp. 349-59. Novoselova, T.A., Meuwissen, M.P.M. and Huirne, R.B.M. (2004), "New technology adoption in food chains: a review with special reference to GMO applications in

67 Analysis of consumer acceptance of GM applications in the pork production chain

livestock production", In Bremmers, H. J., Omta, S. W. F., Trienekens, J. H., Wubben, E. F. M. (Eds), Dynamics in chains and networks, Proceedings of the 6th International Conference on chain and network management in agribusiness and the food industry, 27-28 May, pp. 33-39, Wageningen Academic Publishers, Wageningen. O'Connor, E., Cowan, C., Williams, G., O'Connell, J. and Boland, M. (2005), "Acceptance by Irish consumers of hypothetical GM dairy spread that reduces cholesterol", British Food Journal, Vol. 107 No. 6, pp. 361-80. Onyango, B. (2004), "Consumer acceptance of genetically modified foods: the role of product benefits and perceived risks", Journal of Food Distribution Research, Vol. 35 No. 1, pp. 154-61. Onyango, B., Govindasamy, R. and Nayga Jr., R. M. (2004), "An Application of Choice Modelling To Measure U.S. Consumer Preferences for Genetically Modified Foods", In American Agricultural Economics Association, August 1-4, Denver, Colorado. Pardo, R., Midden, C., Miller, J.D. (2002), "Attitudes toward biotechnology in the European Union", Journal of Biotechnology, Vol. 98 No. 1, pp. 9-24. Rigby, D. and Burton, M. (2005), "Preference heterogeneity and GM food in the UK", European Review of Agricultural Economics, Vol. 32 No. 2, pp. 269-88. Sang, H. (2003), "6. Genetically modified livestock and poultry and their potential effects on human health and nutrition", Trends in Food Science & Technology, Vol. 14 No. 5-8, pp. 253-63. Stoop, I. A. L. (2005), The hunt for the last respondent: nonresponse in sample surveys, Social and Cultural Planning Office of the Netherlands, The Hague. USDA (2006), "Livestock and Poultry: World Markets and Trade", Circular Series DL&P 1- 06, March 2006, United States Department of Agriculture, date accessed: July 2006 at http://www.fas.usda.gov/dlp/circular/2006/06- 03LP/toc.htm. Van Kempen, T. (2000), "Genetically Modified Organisms in Animal Production", Swine News, North Carolina Cooperative Extension Service, Vol. 23 No. 5, date accessed: August 2006 at http://mark.asci.ncsu.edu/Swine_News/2000/june00.pdf. Verbeke, W. A. J. and Viaene, J. (2000), "Ethical Challenges for Livestock Production: Meeting Consumer Concerns about Meat Safety and Animal Welfare", Journal of Agricultural and Environmental Ethics, Vol. 12 No. 2, pp. 141-51.

68 Chapter 3

Visscher, P., Pong-Wong, R., Whittemore, C. and Haley, C. (2000), "Impact of biotechnology on (cross)breeding programmes in pigs", Livestock Production Science, Vol. 65 No. 1- 2, pp. 57-70. Winer, B. J., Brown, D. R. and Michels, K. M. (1991), Statistical principles in experimental design, McGraw-Hill, New York.

69

Chapter 4

Heterogeneity in consumer acceptance of GM applications in pork production: attitudinal and socio-economic determinants

Tatiana A. Novoselova1,2, Ivo A. van der Lans3, Miranda P.M. Meuwissen1,2, Ruud B.M. Huirne1,2 1) Business Economics, Wageningen University 2) Institute for Risk Management in Agriculture, Wageningen University 3) Marketing and Consumer Behaviour Group, Wageningen University

Submitted to Food Quality and Preference

Abstract: This paper examines consumer preferences regarding different GM applications used in the production of pork and explains observed heterogeneity by using attitudinal and socio- demographic determinants. Choice experiment design was used to capture consumers’ preferences for GM benefits of GM pork chops produced by using four GM applications. Finite-mixture regression analysis was used to obtain five consumer segments. About 49% of the respondents were found to be rejecters of GM technology, including 37.2% of “Anti-GM” consumers and 12.3% of those that occasionally would try GM pork, but still preferred conventional or any other types of pork (“Occasional eaters”). A further 29.3%, “GM proponents”, were found to accept GM pork from all GM applications with benefits. The “Benefit seekers” (9.1%) based their buying decision on the benefits offered, while “Application sensitive” (12.3%) accepted GM pork with benefits depending on the type of GM applications used.

Chapter 4

1. Introduction In the literature consumer acceptance of products from emerging technologies is referred to as a key success factor for technology adoption (Frewer, 2003; Hossain, Onyango, Schilling, Hallman & Adelaja, 2003; Novoselova, Meuwissen & Huirne, 2007). Therefore, it is important for producers and the whole supply chain to understand the complex problem of how consumers evaluate and consequently choose foods produced with the help of new technologies. This is especially true for genetic modification. As a result, much research has been done to understand consumer behaviour issues regarding genetic modification and genetically modified products (Bredahl, 2001; Bredahl, Grunert & Frewer, 1998; Frewer, Howard, Hedderley & Shepherd, 1997a; Frewer, Howard & Shepherd, 1997b; Grunert, Bredahl & Scholderer, 2003). All of these studies show that the current level of public acceptance of genetic modification is low. GM foods were found to conflict with “both individual and social values” of the consumers (Bredahl, 1999) and associated with “negative consequences and risks, but mostly with uncertainty and unhealthiness” (Grunert, Lahteenmaki, Asger Nielsen, Poulsen, Ueland & Astrom, 2001). Consumer acceptance and choice of new food products of emerging technologies are believed to depend on many factors. It was previously assumed that “people’s perception of risk and benefit associated with particular products and applications will determine acceptance” (Frewer, Howard & Aaron, 1998). Also, risk perception has been seen as the result of a “potential disutility as well as lack of the benefits from the new technologies” (Costa-Font & Mossialos, 2007). This perceived absence of a special attractive consumer benefit in some GM foods has been referred to as the “Achilles heel” of GM foods (Gaskell et al., 2004). It is generally expected that tangible consumer benefits of GM foods will considerably increase their acceptance (Hossain et al., 2003; Loureiro & Bugbee, 2005; Mather, Knight & Holdsworth, 2005). Yet, the research has evidenced that the consumer perception of benefits and risks associated with GM foods do not primarily influence consumer acceptance. The general attitudes that consumers form towards technology are the significant influencing factor (Lahteenmaki, Grunert, Ueland, Astrom, Arvola & Bech-Larsen, 2002; Loureiro et al., 2005). General attitudes towards technology were found to influence perceived risks and benefits of the technology (Bredahl, 2001). Also, it was suggested that the perceived risk may be a reflection of more general attitudes (Eiser, Miles & Frewer, 2002). In addition, it was found that depending on the character of the pre-existing attitudes, consumers interpreted events

73 Heterogeneity in consumer acceptance of GM applications in pork production concerned with genetic modification either in a positive or a negative way. However, for consumers that have intermediate attitudes, the negative factors have a greater impact than the positive ones (Poortinga & Pidgeon, 2004). In line with this, Lusk et al. (2004) showed that initial attitudes have a significant effect on how people respond to new information about genetic modification. To summarize, previous consumer studies have yet evidenced that consumer acceptance and choice of GM products mainly depend on the interactions of the product benefits with more general attitudes to the production technology and perception of risks associated with this technology (Hu, Hünnemeyer, Veeman, Adamowicz & Srivastava, 2004). It was suggested that having a relevant consumer benefit (Fortin & Renton, 2003) and a positive experience with real GM products (Lahteenmaki et al., 2002) may help to change consumer attitudes and therefore the acceptance of genetic modification. Different areas of application and the type of modification is another issue frequently associated with acceptance of technology (Grunert et al., 2001). Consumers mitigate their negative attitudes in favour of genetic modification for medical purposes, for example “GM bacteria for producing insulin” or “GM for hereditary diseases” (Koivisto Hursti, Magnusson & Algers, 2002). Also, GM animals are less accepted compared to GM micro-organisms and GM plants (Hamstra & Smink, 1996). Studies also show that a higher level of knowledge about genetic modification does not necessarily lead to acceptance (Hamstra et al., 1996). On the contrary, well-informed consumers are likely to be more concerned about genetic modification (Bucchi & Neresini, 2002). An examination of existing literature suggests that previous research mainly concentrated on identifying factors that influence consumer acceptance, studying the effect of provision of additional information to the consumer, exploring the influence of the benefits and the risks associated with technology and proposing new communication strategies. All this is highly relevant for the understanding of consumer behaviour towards modern technologies and GM foods, in particular. Yet, as was mentioned earlier by Hamstra et al. (1996) and Hobbs (2001), consumers are not a homogeneous mass and not all people will react similarly to genetic modification. There are different consumer segments with different attitudes (Hobbs, 2001). However, only a limited number of studies addressed this issue by segmenting and profiling consumers (Burton, Rigby, Young & James, 2001; Frewer et al., 1997a; Hu et al., 2004; Kuznesof & Ritson, 1996; O'Connor, Cowan, Williams, O'Connell & Boland, 2006; Verdurme & Viaene, 2003). This paper aims at filling up the knowledge gap with respect to different consumer segments in preferences for different applications of GM technology in

74 Chapter 4 pork production. Therefore, the main objectives of this study are (1) to analyse how heterogeneous consumers’ preferences are regarding different GM applications used in the production of pork, and (2) to explain this heterogeneity with attitudinal and socio-economic determinants. The paper proceeds as follows. Section 2 presents materials and methods used in this study. It introduces the questionnaire design and the design of the choice experiment. The description of the sample is also presented in this section. Section 3 presents results of consumer segmentation and profiling. The last section outlines conclusions and suggests some issues for further research.

2. Materials and methods 2.1 Questionnaire design A consumer questionnaire was designed, which incorporated a choice experiment task to explore consumer acceptance of different GM applications used for the production of pork chops. Next to the choice experiment, the questionnaire also contained questions related to general pork purchasing behaviour of consumers, familiarity with GM technology, attitudes towards GM technology, GM processes and GM products, consumers’ opinion about the threshold values for and labelling of GM products, and socio-economic characteristics of the respondents. Before starting to answer these questions, respondents were informed that “pork” referred to fresh pork. The questionnaire consisted of four parts. In the first one, consumers were asked questions regarding their actual buying behaviour of pork products. These were questions identifying whether a respondent was the main shopper in the household, whether he/she had bought pork during the past year, the type of pork consumed, the place of purchase, important buying aspects, importance of information provided on the label, the brand of pork purchased and the consumption frequency of pork. Respondents who did not buy pork during the past year, did not have to answer any questions concerning the first part and were asked to continue with the second part of the questionnaire. The second part of the questionnaire included questions regarding consumer attitudes towards genetic modification in pork production. Before answering these questions, consumers were provided with a definition of genetic modification and they were introduced to four possible GM applications in pork production. These will be explained in Section 2.1.1 “GM applications”. Several issues were addressed in the second part of the questionnaire:

75 Heterogeneity in consumer acceptance of GM applications in pork production

Familiarity with genetic modification. Consumers were asked whether they had heard information (before reading the short introduction of Part 2) about genetic modification (“I have heard about it a lot”, “… something” and “…nothing”). Next, consumers were asked whether they knew some GM products (“Yes/No” options). If the answer was “Yes”, consumers had to name the products they knew. Attitude towards genetic modification. Attitude towards the use of GM technology has been measured in several ways. First, consumers’ general attitude towards genetic modification was assessed on a 5-point scale (from “negative” to “positive”). Next, consumers had to assign a score from 1 to 5 (1 = “totally disagree” to “5 = totally agree”, with “do not know” option) to 12 statements specified for four GM applications (so in total 48 statements) regarding consumers’ beliefs about production and consumption of GM food. Consumers were also asked about their willingness to consume GM pork, produced by using four GM applications (from 1 = “definitely not” to 5 = “definitely yes”). Labelling and maximum allowable level of GMO material in food. According to current regulations, the labelling of GM food and feed products is required if ingredients (considered individually) contain 0.9% or more of GM material (Regulation (EC)
1830/2003). Consumers were asked to indicate “how important was for them that food containing more than 0.9% of GMO should be labelled” (1 = “not important”, 5 = “very important”) and whether this level of 0.9% was acceptable to them. As to the latter question, consumers had three options to choose from: “yes”, “no” and “do not know”. If option “no” was chosen, consumers had to provide their own threshold value. With respect to the total maximum allowable percentage of GMO in food, this is still very arbitrary, since there is no general legislation. Therefore, consumers had an opportunity to provide their own acceptable maximum level of GMO in food by giving percentages. For those respondents who had difficulty in providing a percentage, there was an option “do not know”. The third part presented the choice experiment task, which is considered the most important part of the questionnaire and will be explained in section 2.2 “Choice experiment”. Part 4 contained questions regarding the socio-economic background of respondents, such as gender, household composition, the presence of children in the household, household income, education and age.

76 Chapter 4

2.1.1 GM applications To evaluate consumer acceptance of GM technology in the pork production chain, we used four GM applications that are possible in the pork production chain. For a detailed description of these applications, please refer to Novoselova et al. (2007). To summarise, we defined the following applications (Novoselova, van der Lans, Meuwissen & Huirne, 2005): 1) GM animal as a pig produced with the help of genetic modification, which change the genes of the pig itself, so the next generation of those pigs will be different. 2) GM feed as feed composed of the crops produced with the help of genetic modification. 3) GM additives (such as vitamins, bacteria for digestion) and medicines (such as vaccines and antibiotics) as additives & medicine that are produced with the help of GM technology. 4) GM bacteria as special bacteria used after slaughtering of the pigs, during processing of the meat, for preservation of meat. Bacteria are produced with the help of GM technology. Therefore, pork produced with the help of these GM applications is called “GM pork”.

2.1.2 GM benefits Benefits from applying GM technology in pork production were selected based on the technological feasibility to obtain those benefits and their relevance to consumers. The following benefits were chosen: 1) Price of the final product. This attribute was presented having three levels, i.e. no price reduction, 10% price reduction and 33% price reduction. The attributes listed below had two levels: current level or improved/reduced level. 2) Meat quality 3) Animal welfare 4) Impact on the environment 5) Residues in meat For a detailed description and justification, see Novoselova et al. (2005).

2.2 Choice experiment In the choice experiment, respondents were confronted with a number of choice sets including conventional pork and two pork chops varying with respect to particular GM application and

77 Heterogeneity in consumer acceptance of GM applications in pork production the five consumer benefits. The conventional pork chop did not include any of the five consumer benefits, i.e. no price reduction nor improvements in quality, animal welfare, reduction of impact on the environment and reduction of residues in meat. For the GM pork chops each of the benefits could either be present or absent, except for the price reduction benefit, which was varied across three levels. To reduce the number of choice sets that had to be evaluated by each respondent, we used an orthogonal fractional-factorial main-effects design on the consumer benefits in combination with a cyclic design to generate 16 hypothetical choice sets for each GM application. The orthogonal fractional-factorial main- effects design also included a blocking factor that split the 16 hypothetical choice sets into four blocks consisting of four choice sets each. Subsequently, we created four different versions of the questionnaire. In each version, four choice sets were included for each of the four GM applications according to the blocking factor. As such, each respondent was asked to make choices from 16 choice sets (4 including GM-animal alternatives, 4 including GM-feed alternatives, 4 including GM-additives & medicines alternatives, and 4 including GM-bacteria alternatives).

2.3 Data collection 2600 surveys were mailed to a sample of addresses in the Netherlands, which were obtained randomly from an electronic telephone book. Respondents were selected from 26 regions across the country. After two weeks a reminder was sent. In total 253 usable questionnaires were obtained. The average age of the respondents was 50.5 years (s.d. 14.3), of whom 53% were female. Comparing socio-economic characteristics of our sample with the Dutch population (Statistics Netherlands, 2002, 2005, 2006), we conclude that the population in our sample is relatively older, more educated, with more households of two persons and without children. Regarding gender, the sample was considered to be representative of the Dutch population. Only 3% and 10% of consumers were not responsible for shopping in the household and were not pork purchasers, respectively. Most consumers purchased pork in the supermarket (77%) or at the butcher (20%). Only 3% purchased pork somewhere else. The majority of consumers (85%) purchased pre-packed pork. The most frequently purchased types of pork are pork chop (20%), pork steak (22%) and pork fillet (17%). Most consumers ate pork 2-3 times a week (31%), or once a week (28.5%); only 10% ate pork more than 3 times a week and 24% less than once a week.

78 Chapter 4

2.4 Data analysis The heterogeneity in consumer acceptance of GM pork is captured by assuming that there are several consumer segments, in each of which the type of GM application and the consumer benefits exert a different influence on consumer decision making. The analysis is based on the data obtained from the choice experiment part of the questionnaire. To obtain the segments, we used finite-mixture regression analysis (GLIMMIX 2.0). Two main selection criteria CAIC (Consistent Akaike Information Criterion) and BIC (Bayesian Information Criterion) statistics were used to identify the optimal number of segments (Wedel & Kamakura, 1999). The solution that minimizes the values of CAIC and BIC was chosen after ensuring its managerial interpretability. The regression equations were specified in such a way that the regression coefficients obtained were the utilities of i) a price decrease of 1%, ii) the presence of each other’s benefit compared to its absence, and iii) each GM application compared to the situation in which no GM was used. Segments were compared in terms of the attitudinal and socio-economic differences using cross-tabulations. Also, differences between segments in terms of consumer buying behaviour, general attitude to the technology and knowledge about technology were tested by using analysis of variance (ANOVA). When significant differences were observed, Tukey’s HSD post hoc procedure was used to find out which segments were different. Factor analysis using principal components was performed on 12 statements for four GM applications in order to assess consumers’ attitudes towards consumption and production of GM pork by using different applications. The factor analysis was carried out on a so-called “extended” data matrix where each row of the data matrix gives scores of a respondent for a GM-application on 12 statements (Dillon, Frederick & Tangpanichdee, 1985). To aid the interpretation of factors, first VARIMAX and then OBLIMIN rotations were performed. The latter was considered more appropriate regarding the interpretability of the solution and strong relation between two factors. Principal component extraction and eigenvalue greater than 1 were used as the cut-off values for determining the number of factors to be extracted (Hair, Anderson, Tatham & Black, 1998). To investigate differences within the consumer attitudes towards four applications across five segments, a repeated measures ANOVA using the Sidak corrected post hoc procedure was conducted. To test the assumption of sphericity for our repeated measures analysis, Mauchly’s test of sphericity was performed. For cases that yielded a significant result, the Huynh-Feldt correction was used. In addition, MANOVA (multivariate analysis of variance) was conducted, because of its independency upon the assumption of sphericity (Field, 2000).

79 Heterogeneity in consumer acceptance of GM applications in pork production

3. Results 3.1 Identifying segments Before running finite-mixture regression analysis respondents who had constantly selected conventional pork for all sixteen choice sets were assigned to a special segment and excluded from the analysis, because of computation problems. Further segmentation was based on 2544 choice sets (159 respondents). Based on the minimum CAIC and BIC (see Table 1) we have selected a five-segment solution (including the special segment mentioned above).

Table 1. Statistics for determining optimal number of segments Number of DF Log R-square CAIC BIC segments likelihood at convergence 3 19 -2119.09 0.23 4406.15 4387.15 4 29 -1980.44 0.28 4217.27 4188.27 5 39 -1888.97 0.31 4122.75 4083.75

Table 2 presents the estimated utilities for each of the five segments. The estimated utilities of benefits are expressed relative to a reference level (current level) which was set at zero. The utilities for the different GM applications are taken relative to the utility of conventional pork, which was set at zero. Examination of the estimated utilities of each segment indicates that consumers in each segment value product benefits and GM applications very differently. The first segment presents 12.3% of the consumers. In this segment consumers have preferences for two GM applications: GM animal and GM feed over conventional pork. GM additives & medicines and GM bacteria have a negative influence on the consumers’ choice. Regarding benefits, price reduction and improvements in animal welfare have a strong positive influence on the consumers’ choice. Reduced impact on the environment and residues do not interest these consumers. This segment was named “Application sensitive”. The second segment also comprises 12.3% of the consumers. In this segment consumers are in general negative about GM. None of the utilities for four applications has a positive sign. Consumers in this segment do not care about price discounts. Improvements in other characteristics do not play an important role, as all utilities for benefits are not significant. However, these consumers occasionally choose GM pork and do not completely reject it. On this basis, this group was designed as “Occasional eaters”.

80 Chapter 4

Table 2. Estimated utilities for finite-mixture regression analysis Application Occasional Benefit GM Anti-GM sensitive eaters seekers proponents Benefits 1% price discount 1 0.020*** 0.002 0.022** 0.025*** (0.006) (0.007) (0.008) (0.003) Improved quality 0.346** -0.020 0.455** 0.460*** (0.126) (0.162) (0.175) (0.068) Improved animal welfare 1.129*** 0.190 1.769*** 0.940*** (0.140) (0.191) (0.243) (0.082)

Improved environment 0.017 0.073 0.300

0.171* t n

(0.135) (0.190) (0.218) (0.080) e m g

Reduced residues 0.176 0.061 1.063*** 0.567*** e s (0.125) (0.162) (0.189) (0.069) M G - N

GM applications O

GM animal 2.917* -1.980*** -7.067*** 2.755*** N (1.026) (0.290) (0.808) (0.718) GM feed 2.205* -0.729** -2.272*** 2.371*** (0.746) (0.271) (0.404) (0.590) GM additives & -2.340* -2.826*** -0.305 2.817*** medicines (0.260) (0.340) (0.568) (0.730) GM bacteria -1.993* -1.374*** -0.947** 3.211*** (0.229) (0.236) (0.308) (0.645)

Choices, % GM pork (options 1&2) 67.9 31.2 59.8 99.2 0 Non-GM pork (option 3) 32.1 68.8 40.2 0.8 100

Segment size, % 12.3 12.3 9.1 29.3 37.2 *P < 0.05, ** P < 0.01, *** P < 0.001 1 The estimated utility of a price discount of 1% is presented. From this, the estimated utilities of different price discounts can be calculated. For example, the estimated utility of a price discount of 10% for Segment 1 is 0.02*10=0.20

The third segment is the smallest segment, its size being 9.1%. Consumers of this segment are “against” GM (all GM applications have negative utilities). Consumers are especially negative about GM animals, followed by GM feed. They have a less negative utility towards GM additives & medicines and GM bacteria. The third-segment consumers have the same preference to the price reduction as consumers in the first segment. It seems that improvements in animal welfare and reduced residues have an important positive influence on the consumer’s utility. These two benefits in the third segment have the highest coefficients compared to the same benefits in the other segments. Therefore, this segment was termed “Benefit seekers”.

81 Heterogeneity in consumer acceptance of GM applications in pork production

The fourth segment represents about 29.3% of the respondents. The positive and high utilities for all GM applications indicate that consumers are positive about genetic modification being used in pork production and willing to consume GM pork. Presence of GM bacteria has the strongest positive influence on the consumer choices. The respondents in this segment also appear to appreciate the benefits. Compared to the respondents in other segments, consumers in this segment are price sensitive - the utility coefficient for price discount is much higher compared to other segments. Also, improvements in all pork characteristics have a significant positive influence on consumer choices. This group was named “GM proponents” The last, fifth segment is the largest segment in our sample (37.2%). It includes consumers who always choose the conventional pork chop. It means that these consumers do not accept GM pork at all, no matter what kind of modification was applied and what kind of benefits this pork can offer to the consumers. This group was termed “Anti-GM”.

3.2 Explaining segment heterogeneity The next step in the analysis is to relate observed heterogeneity to attitudinal and socio- economic characteristics.

3.2.1 General pork purchasing behaviour of consumer Differences exist between segments in terms of their attitude towards important buying aspects of meat, the importance of the information provided on the label of packed meat and the brand of pork that consumers regularly buy (Table 3).

82

Table 3. Mean ratings of the importance of buying aspects, information on the label, preferred brand of meat and consumer attitude towards GM technology Importance of buying aspects Importance of information on Preferred brand of meat Attitude (1= not important to 5=very the label (yes, %) (1= negative important) (1= not important to 5=very to 5=positive) important) Animal Naturalness Country of E-numbers EKO Environmental No specific welfare origin (organic) quality brand F=2.877b* F=4.112** F=3.362* F=2.350
2=9.759*
2=4.652*
2=19.777** F=14.863**

Application 3.15ab 3.00a 1.52a 1.22a 3.6a 0.0a 57.1a 2.13ac sensitive Occasional eaters 3.08ab 3.35ab 1.91ab 1.55ab 15.4ab 15.4b 46.2ac 1.87ab

Benefit seekers 3.04ab 3.17ab 1.63ab 1.84ab 21.7bc 4.3ab 4.3b 1.70ab

GM proponents 3.05a 3.31a 1.83ab 1.69ab 13.8ac 3.1a 27.7c 2.54c

Anti-GM 3.65b 3.89b 2.21b 1.91b 27.4b 13.1b 28.6c 1.48b a Tukey HSD-test: mean scores in the same column with unlike letters (a, b, c) are significantly different (P<0.05) b F-test and 2-test significance levels: *P<0.05 and **P<0.01

Heterogeneity in consumer acceptance of GM applications in pork production

Overall, all buying attributes of meat (price, taste, nutritional value, animal welfare, naturalness, no additives, expiration date) are important to the consumers but still differences exist among different attributes and segments (Figure 1). “Taste” is the most important attribute compared to all attributes. Next to this, “no additives” and “expiration date” are considered more important attributes compared with price, nutritional value, animal welfare and naturalness.

Scale (1=least important, 5=most important)

5

4.5 price taste 4 nutritional value animal welfare 3.5 naturalness no additives expiration date 3

2.5 Application Occasional Benefit GM Anti-GM sensitive eaters seekers proponents Segments

Figure 1. Means of different buying aspects

Differences between consumer segments were observed for “animal welfare” and “naturalness” (Table 3). Consumers in segment “GM proponents” consider these aspects less important than those in the “Anti-GM” segment. Also, consumers in segment “Application sensitive” think naturalness of meat less important than consumers in the “Anti-GM” segment. As regards “no additives”, no significant differences between segments were found; however, the differences were large enough to give overall significant value. With respect to the important information provided on the label of meat, no differences were observed between segments regarding price, expiration date, weight, nutritional value,

84 Chapter 4 and other (provided by the consumer). However, segments are different regarding two aspects: country of origin and E-numbers (Table 3). The differences are observed between consumers in segment “Application sensitive” and “Anti-GM”, where in the “Application sensitive” segment consumers hardly ever pay attention to these aspects, whereas in “Anti- GM” segment consumers often check this information on the meat label. Among different brands of pork meat (IKB, EKO (organic), Bio-dynamic (also organic), Environmental quality, Free range, Supermarket-label, No special brand and other), differences exist with respect to the EKO (organic), Environmental quality and No specific brands (Table 3). Consumers in “Application sensitive” segment buy less EKO meat compared with consumers in the “Anti-GM” and “Benefits seekers” segments. Also, consumers in the “GM proponents” segment buy twice less EKO meat than consumers in the “Anti-GM segment”. Regarding “Environmental quality” brand, consumers in the “Application sensitive” and “GM proponents” segments do not or hardly buy meat of this brand compared to the consumers in the “Occasional eaters” and “Anti-GM” segments. Consumers in the segment “Benefit seekers” do not buy non-brand meat, whereas consumers in the “Occasional eaters”, “GM proponents” and “Anti-GM” segments do buy. On the contrary, the majority of the consumers in the segment “Application sensitive” (57.1%) buy relatively more meat with no special brand compared to the consumers of the “Benefits seekers” (4.3%), “GM proponents” (27.7%) and “Anti-GM” (28.6%) segments.

3.2.2 Familiarity with GM technology No differences were observed between the segments in terms of their familiarity with the information about GM products and the current (available on the market or at lab level) GM products or innovations. Only 26% of the respondents have indicated to be familiar with GM products. The most familiar GM products to the respondents were GM soy and GM maize.

3.2.3 General attitude to GM technology Consumers in segments are different in their general attitudes towards GM technology (see Table 3). The segments “Application sensitive” and “GM proponents” have a more neutral attitude, whereas the “Anti-GM” segment is negative about technology. Also, consumers in segment “Benefit seekers” are different from those in the “GM proponents” segment, with the latter having a more positive attitude.

85 Heterogeneity in consumer acceptance of GM applications in pork production

3.2.4 Attitude to GM technology: factor analysis As was mentioned earlier, factor analysis (PCA with an OBLIMIN rotation) was performed on the statements assessing consumers’ attitudes towards consumption and production of GM pork by different GM applications. Bartlett’s test of sphericity (p<0.0001) and the Kaiser- Meyer-Olkin measure of sampling adequacy (0.9) indicate that conducting factor analysis on the data is appropriate. The eigenvalue greater than 1 criterion indicated a two-factor solution accounting for 65% of the variance. A possible interpretation of factor 1, accounting for 55% of the total variance (after OBLIMIN rotation), is “Concerns”. This factor groups statements that refer to possible consumer concerns regarding both consumption and production of GM products for all four GM applications. “Long-term health problems”, “Moral/religious obligations”, “Artificial nature of GM pork” “Threats to the environment” and “Ethical considerations” load high on this “Concerns” factor. Factor 2 accounts for 10% of total variance (after OBLIMIN rotation) and can be labelled as “Benefits” factor. It is composed of the following statements regarding both consumption and production of GM products for all four GM applications: “More chances than threats (from consumption and production)”, “Useful technology”, “Technology that helps to solve environmental problems”, “Consumption is not risky for humans”, “Production is not dangerous for people” and “Production of GM products increases food security”. Data on the scores of factor 1 “Concerns” violated the sphericity assumption. Therefore Huynh-Feldt corrected F-statistics was used (main effects of applications: F(3, 221)=7.531 p<0.0001, interaction effect of applications and segments: F(10, 221)=3.895 p<0.0001). This suggests that consumers differ in their perceptions of concerns related to the four different GM applications in general and across five segments. Also, the results of MANOVA (main effects of applications: Wilks’ Lambda=0.792, F(3, 86)=7.549, p<0.0001; interaction effect of applications and segments: Wilks’ Lambda=0.714, F(12, 228)=2.6, p<0.003) show the same effect of the type of application across the five segments. Among the four GM applications, GM feed arouses fewer concerns associated with GM technology (all pairwise comparisons of GM animal, GM additives & medicines, and GM bacteria with GM feed are significant). As to the differences in segments, consumers in the “Occasional eaters” segment perceive more concerns associated with technology than consumers in the “GM proponents” segment (see Fig. 2 A). On the contrary, consumers in the segments “Benefit seekers” and “GM proponents” perceive fewer concerns associated with the four types of GM applications compared to the consumers in the “Anti-GM” segment.

86 Chapter 4

Factor 1 "Concerns"

GM animal GM feed GM add.&med. GM bacteria

0,9

s 0,7 n a e 0,5 M

l a

n 0,3 i g r

a 0,1 M

d -0,1 e t a

m -0,3 i t s

E -0,5 -0,7 Application Occasional Benefit GM Anti-GM sensitive eaters seekers proponents Segments

Figure 2 A: Consumer attitudes towards GM pork produced by using different GM applications: Concerns

As regards factor 2 “Benefits”, data also violated the sphericity assumption and therefore Huynh-Feldt corrected F-statistics was used (main effect of applications: F(3, 225)=1.153 p<0.325, interaction effect of applications and segments: F(10, 225)=2.596 p<0.005). This suggests that in general the type of application does not have an effect on consumers perceptions of benefits related to the four different GM applications. However, there is a significant interaction effect of type of application and segment membership on the perception of benefits related to the four different GM applications. Also, the results of MANOVA (main effect of applications: Wilks’ Lambda=0.944, F(3, 86)=1.697, p<0.174; interaction effect of applications and segments: Wilks’ Lambda=0.734, F(12, 228)=2.353, p<0.007) show a significant effect of type of the application across five segments on the consumers’ attitude related to the benefits associated with GM technology. As to the observed differences in segments, consumers in the “GM proponents” segment perceive more benefits associated with technology than consumers in the “Application sensitive”, “Occasion eaters” and “Anti-GM” segments (see Fig. 2 B).

87 Heterogeneity in consumer acceptance of GM applications in pork production

Factor 2 "Benefits"

GM animal GM feed GM add.&med. GM bacteria

0,9 s

n 0,7 a e

M 0,5

l a

n 0,3 i g r

a 0,1 M

d -0,1 e t a

m -0,3 i t s

E -0,5 -0,7 Application Occasional Benefit GM Anti-GM sensitive eaters seekers proponents Segments

Figure 2 B: Consumer attitudes towards GM pork produced by using different GM applications: Benefits

3.2.5 Labelling and threshold values Consumers in different segments have different views on the importance of labelling (F(4, 246)=7.832, p<0.0001). There are two sub-sets of consumer segments, based on the means, those that find labelling very important (segments “Occasional eaters”, “Benefit seekers”, and “Anti-GM”) and those that find labelling somewhat important (segments “Application sensitive” and “GM proponents”). Opinions of the consumers in the “GM proponents” segment are significantly different from those in the “Occasional eaters”, “Benefit seekers”, and “Anti-GM” segments. Also, the “Application sensitive” segment is different from the “Anti-GM” segment in its views on importance of labelling. However, no differences were observed between consumers in the “Application sensitive” segment compared to the segments “Occasional eaters” and “Benefit seekers”. Overall, about 42% of consumers find the currently required level of 0.9% for labelling acceptable, 30% of the consumers have difficulty answering this question and choose option “do not know”, about 28% disagree with the current regulations and propose their own values. Among these consumers 10.3% prefer no GMO in food and 17.7% propose value in a range from >0.0 to > 10% (Figure 3). Consumers in the “Anti-GM” segment are different from the

88 Chapter 4 consumers in the segments “Application sensitive”, “Occasional eaters” and “GM proponents”.

80 70 60 Application sensitive 50 Occasional eaters 40 Benefit seekers 30 GM proponents Anti-GM 20 10 0 0.0% >0.0%- 0.1%- 0.9% > 10% Don't 0.05% 0.8% know

Figure 3. Consumers opinion about the acceptable boundary value of GMO in food above which it has to be labelled

As regards the consumer opinion on the maximum allowable percentage of GMO in food, the majority of the consumers (60.2%) do not have a firm opinion about this, 19.5% choose for no GMO in food, and about 20% indicate a number from a range from 0.1% to >50% (Figure 4). Differences between segments are similar to those observed previously: consumers in the “Anti-GM” segment are significantly different in their opinion (more consumers chose the “No” option and less opt for “Do not know”) compared to all other segments. There are no differences among the other segments.

89 Heterogeneity in consumer acceptance of GM applications in pork production

80 70 60 Application sensitive 50 Occasional eaters 40 Benefit seekers 30 GM proponents Anti-GM 20 10 0 0.0% 0.1%- 1%-5% 10%- > 50% Don't 0.9% 25% know

Figure 4. Consumers opinion about the allowable maximum level of GMO in food

3.2.6 Socio-economic profile With respect to socio-demographic characteristics, differences were observed between segments regarding age of the consumers (Table 4). A significant number of younger consumers belong to the “Benefit seekers” segment compared to all other segments. In this segment the majority of the consumers fall in the age categories of 25-39 and 39-49 years (in total 78.2% of consumers), whereas only 17.4% belong to the 50-59 and >60 age categories. Although there were no overall significant differences found between segments with respect to the education level, household size and presence of children in the household, pairwise comparisons revealed differences for the “Occasional eaters”, “Benefit seekers” and “Anti GM” segments. Consumers in the “Benefits seekers” segment are more highly educated; there are no respondents in the lower level of education and about 68% have a high level of education (Mann-Whitney U: 234.5, p<0.047). Contrary, in the “Occasional eaters” segment, the number of consumers with a high level of education is lower (46.6%) and 23.3% have a lower level of education. As regards the household size, “Benefit seekers” have a larger household size than “Occasional eaters” (Mann-Whitney U: 241, p<0.047). The majority of “Benefit seekers” (54.5%) have a household size of 3-5 persons, whereas the majority of the “Occasional eaters” (58.1%) have a household of 2 persons. Also, “Benefit

90 Chapter 4 seekers” more often live in households with children (57.1%) compared to “Occasional eaters” (17.2%) (Pearson
2 =8.642, p<0.003) and compared to consumers who are “Anti- GM” (34.1%) (Pearson
2 =3.842, p<0.05).

Table 4. Socio-economic characteristics (%) Application Occasional Benefit GM Anti-GM sensitive eaters seekers proponents Gender Male 35.5 45.2 47.8 49.3 47.3 Female 64.5 54.8 52.2 50.7 52.7 Pearson
2 = 1.801, df=4, P.=0.772

Household composition 1 person 12.9 16.1 9.1 23.0 18.5 2 persons 54.8 58.1 36.4 39.2 50.0 3-5 persons 32.3 25.8 54.5 36.5 30.4 >5 persons ------1.4 1.1 Kruskal-Wallis Test
2=4.325, df=4, P.=0.364

Children (yes) 38.7 17.2 57.1 37.0 34.1 Pearson
2 = 8.809, df=4, P.=0.066

Household income <1000 14.8 -- 9.5 2.8 9.0 1000-2000 33.3 37.9 33.3 31.9 27.0 2000-3000 29.6 31.0 23.8 38.9 40.4 3000-4000 14.8 17.2 19.0 18.1 12.4 >4000 7.4 13.8 14.3 8.3 11.2 Kruskal-Wallis Test
2=2.018, df=4, P.=0.732

Education Low level 16.2 23.3 -- 16.4 17.2 Middle level 32.3 30.0 31.8 30.1 33.3 High level 51.6 46.6 68.2 53.4 49.5 Kruskal-Wallis Test
2=4.574, df=4, P.=0.334

Age <24 -- 3.2 4.3 4.1 -- 25-39 32.3 16.1 47.8 21.6 19.8 40-49 29.0 19.4 30.4 20.3 16.5 50-59 6.5 35.5 8.7 31.3 27.5 >60 32.3 25.8 8.7 23.0 36.3 Kruskal-Wallis Test
2=16.610, df=4, P.=0.002

91 Heterogeneity in consumer acceptance of GM applications in pork production

4. Discussion and conclusions The results of this study indicate that consumer preferences towards GM foods are heterogeneous. We have identified five possible segments of consumers which are distinct with respect to consumer preferences towards GM applications and benefits that these applications can offer to the consumers. While the majority of the consumers expressed their overall negative attitude towards GM technology and GM foods, consumers in some segments would be willing to purchase GM pork depending on the type of application being used in the production and benefits offered. About one-third of the consumers (29.3%, “GM proponents”) are found to accept GM pork from all GM applications with benefits. These results are similar to those found by O’Connor et al. (2006), where 21.2% of Irish consumers were found to have an overall preference for GM products with health benefit. Further, 49% of the respondents were found to be rejecters of GM technology including 37.2% of “Anti- GM” consumers and 12.3% of those that occasionally try GM pork, but still prefer conventional or any other types of pork (“Occasional eaters”). Hu et al. (2004) found that 46% of consumers, i.e. “Traditional” and “Anti-GM”, will not purchase GM food, no matter what benefits it might have. The “Application sensitive” and “Benefit seekers” are special consumer segments, as consumers in the first segment purchase GM pork with benefits depending on the type of GM applications used, whereas consumers in the latter make their buying decisions, based on the benefits offered. As to the benefits chosen by consumers in this segment, “animal welfare” and “reduced residues in meat” have an important positive influence on the consumers’ choices and have the highest coefficients compared to other segments. Greater insight into consumer acceptance of GM pork was provided by profiling consumers regarding attitudes and socio-economic characteristics. For example, consumers in segments “Applications sensitive” and “GM proponents” have a more neutral attitude towards GM technology than consumers in other segments. Also, they buy meat with no specific brand. “GM proponents” perceive fewer concerns and more benefits associated with GM technology in general, whereas “Application sensitive” consumers associate fewer concerns and more benefits only with respect to GM animal and GM feed applications. It appears that labelling is less important to the consumers in these segments. The segments “Occasional eaters” and “Anti-GM” are somewhat similar in their attitudes towards GM technology, with a difference that “Anti-GM” consumers are more critical, perceiving fewer benefits and more concerns. As to the consumers in the “Benefit seekers” segment, it seems likely that consumers in this segment prefer meat with guaranteed quality, as they hardly buy non-brand

92 Chapter 4 meat . Also, this segment is significantly different compared to other segments in terms of socio-economic characteristics and it is represented by highly educated, young consumers with children. Similar results were found in the study by Hu et al. (2004), who identified an analogous segment of consumers, i.e. Value-Seeking consumers, who value additional benefits of GM bread, regardless of the presence of GM ingredients and who have young families with children. Differences between the utilities of the GM applications across segments are consistent with mean factor scores for concerns and for perceived benefits. Interestingly, general consumer concerns are influenced by the type of GM applications used, while general perception of GM benefits is not. Also, although consumers perceive labelling as important and could state their opinion about the acceptable level of GM required for labelling, the majority do not have a firm opinion about the allowable level of GM in food products. A potential limitation of this study is that we cannot exclude the possibility of some sample-selection bias in our results. Unfortunately, due to unavailability of data on the non- respondents we cannot test for any differences between respondents and non-respondents. Non-response bias could occur due to the difficulty and sensitivity of the topic, as genetic modification in animal production, in general, causes more discussion than GM in plants. Also, we could argue that potential bias can be reduced to some extent due to the segmentation of consumers. Results show that the consumers segments obtained replicate the findings from previous studies, which justifies certain patterns in consumers’ preferences across different countries and with respect to different GM products. By examining choices that consumers make with respect to GM pork and possible segments of the consumers, this study adds new knowledge about consumer preferences towards GM food and new opportunities for pork production chains, which is important for producers. Further research could examine consumers’ acceptance by including other benefits in GM pork. For example, “Taste” would be an interesting option, as consumers in all segments considered this attribute above all others most important.

References Bredahl, L. (1999). Consumers' Cognitions With Regard to Genetically Modified Foods. Results of a Qualitative Study in Four Countries. Appetite, 33(3), 343-360.

93 Heterogeneity in consumer acceptance of GM applications in pork production

Bredahl, L. (2001). Determinants of Consumer Attitudes and Purchase Intentions With Regard to Genetically Modified Food - Results of a Cross-National Survey. Journal of Consumer Policy, 24(1), 23-61. Bredahl, L., Grunert, K. G. & Frewer, L. J. (1998). Consumer Attitudes and Decision-Making with Regard to Genetically Engineered Food Products - A Review of the Literature and a Presentation of Models for Future Research. Journal of Consumer Policy, 21, 251-277. Bucchi, M. & Neresini, F. (2002). Biotech remains unloved by the more informed. Nature, 416, 261. Burton, M., Rigby, D., Young, T. & James, S. (2001). Consumer attitudes to genetically modified organisms in food in the UK. European Review of Agricultural Economics, 28(4), 479-498. Costa-Font, J. & Mossialos, E. (2007). Are perceptions of `risks' and `benefits' of genetically modified food (in)dependent? Food Quality and Preference, 18(2), 173. Dillon, W. R., Frederick, D. G. & Tangpanichdee, V. (1985). Decision Issues in Building Perceptual Product Spaces with Multi-Attribute Rating Data. Journal of Consumer Research, 12, 47-63. Eiser, J. R., Miles, S. & Frewer, L. J. (2002). Trust, Perceived Risk, and Attitudes Toward Food Technologies. Journal of Applied Social Psychology, 32(11), 2423-2433. Field, A. (2000). Discovering statistics using SPSS for Windows. London: Sage. Fortin, D. R. & Renton, M. S. (2003). Consumer acceptance of genetically modified foods in New Zealand. British Food Journal, 105(1/2), 42-58. Frewer, L. (2003). 10. Societal issues and public attitudes towards genetically modified foods. Trends in Food Science & Technology, 14(5-8), 319-332. Frewer, L. J., Howard, C. & Aaron, J. I. (1998). Consumer acceptance of transgenic crops. Pesticide Science, 52(4), 388-393. Frewer, L. J., Howard, C., Hedderley, D. & Shepherd, R. (1997a). Consumer attitudes towards different food-processing technologies used in cheese production--the influence of consumer benefit. Food Quality and Preference, 8(4), 271-280. Frewer, L. J., Howard, C. & Shepherd, R. (1997b). Public Concerns in the United Kingdom about General and Specific Applications of Genetic Engineering: Risk, Benefit, and Ethics. Science, Technology, & Human Values, 22(1 (Winter), 98-124.

94 Chapter 4

Gaskell, G., Allum, N., Wagner, W., Kronberger, N., Torgersen, H., Hampel, J. & Bardes, J. (2004). GM Foods and the Misperception of Risk Perception. Risk Analysis, 24(1), 185-194. Grunert, K. G., Bredahl, L. & Scholderer, J. (2003). Four questions on European consumers' attitudes toward the use of genetic modification in food production. Innovative Food Science & Emerging Technologies, 4(4), 435-445. Grunert, K. G., Lahteenmaki, L., Asger Nielsen, N., Poulsen, J. B., Ueland, O. & Astrom, A. (2001). Consumer perceptions of food products involving genetic modification-- results from a qualitative study in four Nordic countries. Food Quality and Preference, 12(8), 527-542. Hair, J. F., Anderson, R. E., Tatham, R. L. & Black, W. C. (1998). Multivariate data analysis. London: Prentice-Hall International Inc. Hamstra, A. M. & Smink, C. (1996). Consumers and biotechnology in the Netherlands. British Food Journal, 98(4/5), 34-38. Hobbs, J. (2001). Consumer Responses to Food Quality, Food Safety, and Health Concerns. In M. Fulton, H. Furtan, D. Gosnellet al, Transforming Agriculture: The Benefits and Costs of Genetically Modified Crops. Prepared for The Canadian Biotechnology Advisory Committee, Project Steering Committee on the Regulation of Genetically Modified Foods. Hossain, F., Onyango, B., Schilling, B., Hallman, W. & Adelaja, A. (2003). Product attributes, consumer benefits and public approval of genetically modified foods. International Journal of Consumer Studies, 27(5), 353-365. Hu, W., Hünnemeyer, A., Veeman, M., Adamowicz, W. & Srivastava, L. (2004). Trading off health, environmental and genetic modification attributes in food. European Review of Agricultural Economics, 31(3), 389-408. Koivisto Hursti, U.-K., Magnusson, M. K. & Algers, A. (2002). Swedish consumers’ opinions about gene technology. British Food Journal, 104(11), 860-872. Kuznesof, S. & Ritson, C. (1996). Consumer acceptability of genetically modified foods with special reference to farmed salmon. British Food Journal, 98(4/5), 39-47. Lahteenmaki, L., Grunert, K., Ueland, O., Astrom, A., Arvola, A. & Bech-Larsen, T. (2002). Acceptability of genetically modified cheese presented as real product alternative. Food Quality and Preference, 13(7-8), 523-533.

95 Heterogeneity in consumer acceptance of GM applications in pork production

Loureiro, M. L. & Bugbee, M. (2005). Enhanced GM Foods: Are Consumers Ready to Pay for the Potential Benefits of Biotechnology? The Journal of Consumer Affairs, 39(1), 52-70. Lusk, J. L., House, L. O., Valli, C., Jaeger, S. R., Moore, M., Morrow, J. L. & Traill, W. B. (2004). Effect of information about benefits of biotechnology on consumer acceptance of genetically modified food: evidence from experimental auctions in the United States, England, and France. European Review of Agricultural Economics, 31(2), 179- 204. Mather, D., Knight, J. & Holdsworth, D. (2005). Pricing differentials for organic, ordinary and genetically modified food. Journal of Product & Brand Management, 14(6), 387- 392. Novoselova, T. A., Meuwissen, M. P. M. & Huirne, R. B. M. (2007). Adoption of GM technology in livestock production chains: an integrating framework. Trends in Food Science & Technology, doi: 10.1016/j.tifs.2006.12.005 (in press). Novoselova, T., van der Lans, I. A., Meuwissen, M. P. M. & Huirne, R. B. M. (2005). Consumer acceptance of GM applications in the pork production chain: a choice modelling approach. In The Future of Rural Europe in the Global Agri-Food System. Copenhagen, Denmark, in: August 24-27, 2005, EAAE. O'Connor, E., Cowan, C., Williams, G., O'Connell, J. & Boland, M. P. (2006). Irish consumer acceptance of a hypothetical second-generation GM yogurt product. Food Quality and Preference, 17(5), 400-441. Poortinga, W. & Pidgeon, N. F. (2004). Trust, the Asymmetry Principle, and the Role of Prior Beliefs. Risk Analysis, 24(6), 1475-1486. Regulation (EC)
1830/2003, date accessed 03 February 2007: http://europa.eu.int/comm/food/food/biotechnology/index_en.htm Verdurme, A. & Viaene, J. (2003). Consumer Beliefs and Attitude Towards Genetically Modified Food: Basis for Segmentation and Implications for Communication. Agribusiness, 19(1), 91-113. Wedel, M. & Kamakura, W. (1999). Market Segmentation: Conceptual and Methodological Foundations. Boston, Dordrecht, London: Kluwer Academic Publishers.

96

Chapter 5

A feasibility study of GM pork production in the Netherlands

Tatiana A. Novoselova1,2, Miranda P.M. Meuwissen1,2, Age W. Jongbloed3, Ruud B.M. Huirne1,2 1) Business Economics, Wageningen University 2) Institute for Risk Management in Agriculture, Wageningen University 3) Animal Sciences Group, Wageningen University, Lelystad

Submitted to Livestock Science

Abstract: The purpose of this paper is to quantify technical and economic performance of genetic modification (GM) applications in the pork production chain. In order to capture these relationships, we used an economic farm model with farrowing and fattening stages. Input values concern both technical and economic parameters. Model outputs present the cost price of a 25 kg piglet and 1 kg of pork produced, using several GM applications. A sensitivity analysis was carried out to show the effects of various input values across the GM applications on the output result. The results showed that although GM applications improved output results in both stages, the GM applications had a limited economic impact. The sensitivity analysis demonstrated that the changes in the main input parameters had little effect on the economic output results in both stages. The results obtained suggested that the current possible range of improvements in output results due to the GM applications was rather limited and most likely would not be able to motivate farmers to introduce GM into their production system. Chapter 5

1. Introduction Genetic modification represents a type of emerging technologies that are used in the areas of food and feed production (Cardello, 2003; Novoselova, Meuwissen & Huirne, 2007a; Sang, 2003). It is believed that genetic modification has a great potential (Ag-West Biotech Inc., 2005; Bonneau & Laarveld, 1999; Van Kempen, 2000). It offers a wide range of benefits to the producers, consumers and society. For example, producers may profit from an enhanced growth performance of animals and from reduced production costs, e.g. animal health, feeding, and manure costs (Bonneau et al., 1999). As to the consumers, benefits may be found in an increased quality of meat (less fat, fewer residues) and lower prices (Solomon, Pursel, Campbell & Steeleae, 1997; De Vries, Faucitano, Sosnicki & Plastow, 2000). Additionally, society as a whole may benefit from a reduced environmental impact (reduction in quantity of waste and its composition) and improved animal welfare (Bonneau et al., 1999). In general, consumers in Europe are rather sceptical about genetic modification and products produced with its help (Frewer, 2003; Koivisto Hursti, Magnusson & Algers, 2002; Miles, Ueland & Frewer, 2005). Yet, previous research showed that not all consumers react similarly to the GM products (Frewer, Howard, Hedderley & Shepherd, 1997; Kuznesof & Ritson, 1996; Noussair, Robin & Ruffieux, 2004; Novoselova, Van der Lans, Meuwissen & Huirne, 2007b; O'Connor, Cowan, Williams, O'Connell & Boland, 2005; O'Connor, Cowan, Williams, O'Connell & Boland, 2006). Consumers are willing to consider a purchase of GM products if these products have special benefits for consumers. For example, in France 42% of the consumers were willing to buy products made with GMO, if the products were “sufficiently inexpensive” (Noussair et al., 2004). About 21% of a sample of Irish consumers was found to accept GM yogurt that offered a specific consumer benefit (O'Connor et al., 2006). Similar to these results, Novoselova et al., 2007b demonstrated in their study in the Netherlands that 29.3% of the participants would buy GM pork produced with different GM applications with clear benefits. Moreover, some of the consumers in the sample (9.1%) based their decisions on the benefits offered and another 12.3% would accept GM pork depending on the type of GM applications used. It seems that there might be a market for GM products. But why would producers step in this uncertain market? Although the possible benefits of applying GM are recognised, it is remarkable that only a few studies have quantified them (Phillips, 2002). Our study takes the chain approach and various possible GM applications in the chain into account. Therefore, the GM applications are summarised according to the type of the application and affiliation to the benefits. The current study examines the feasibility of adopting several GM applications in

99 A feasibility study of GM pork production in the Netherlands the pork production chain. In more detail, the purpose of this paper is to quantify technical and economic performance of GM applications in the pork production chain. In order to capture these relationships, we used an economic farm model where a pork production chain includes farrowing and fattening stages. The paper is structured as follows. Section 2 presents materials and methods. More specifically, it introduces GM applications used in the analysis, the set-up of the economic farm model, and the technical input values per stage. Data collection is also presented in this section. Section 3 presents results of the model regarding default situation and situations in which GM application is used. Moreover, the results of the sensitivity analysis are presented here. The last section outlines the discussion and future outlook.

2. Materials and methods GM applications and sub-applications in livestock production chains Genetic modification can be applied for many purposes e.g. medication, test-models, plant and animal production. The last-mentioned is described in this paper. Three types of genetic modification were selected for the analysis to cover the farrowing and fattening stages (see Novoselova et al., 2007b). These applications were GM animal, GM feed, GM additives & medicines1. GM animal refers to a GM pig with modified genes. GM feed presents feed/crops that were produced with the help of gene technology. GM additives & medicines refer to additives (phytase, vitamins) and medicines (antibiotics, vaccines) that were produced with the help of genetic modification. Per GM application three concrete sub-applications were chosen for the analysis (Table 1). Table 1. GM applications and sub-applications Producer Society Consumer GM animal Bovine gene Enviropig Spinach gene* GM feed GM crops Low phytate plants Edible vaccines* GM feed additives Antibodies Microbial phytase Immunocastration & medicines * are not used in the analysis

Sub-applications are grouped according to the party they are expected to be beneficial for (producer, society, and consumer). These sub-applications aim at different benefits important to producer, society and consumers. Regarding sub-applications relevant to producers, these are “Bovine gene”, “GM crops” (GM soy and maize) and “Antibodies”. These sub-

1 The “GM bacteria” application was not included in this analysis. This application is assumed to be used in the processing stage

100 Chapter 5 applications decrease production costs by improving animal growth performance or reducing feed costs. Sub-applications relevant to “society” reduce impact of the pork production on the environment by lowering the excretion of phosphorus in the manure. In this study these are “Enviropig”, “Low phytate (LP) plants” (GM soy and maize) and “Microbial phytase”. The last group of sub-applications, i.e. those that are important to consumers, aims at production of (healthier) leaner meat (“Spinach gene”), reduction in the use of antibiotics (“Edible vaccines”) and improvement of animal welfare (“Immunocastration”). Due to the limited available information on the sub-applications relevant to the consumers only “Immunocastration” is analysed in this study.

Data collection: technical and economic data A panel of experts was used to collect and verify necessary technical and economic data on the GM applications selected. Experts were selected depending on the area of their expertise; these were nutritionists, animal physiologists, animal and plant breeders, veterinarians, and representatives of the feed industry. The interviews were performed on a face-to-face basis. First, before discussing GM applications with the experts, the scientific literature on the topic was reviewed. We selected several technical parameters (such as average growth rate, feed conversion ratio, litter size) that were possible to change due to application of GM, and linked them to the chain input parameters. Moreover, from the scientific results, the possible ranges of the expected effects were identified. Secondly, these effects were discussed with the experts selected. Experts had to analyse results obtained from scientific literature and then provided their personal estimates (if it was possible) on the effects of sub-applications. In general, experts had some difficulties in responding, due to a wide range of uncertainty of some GM applications that do not exist yet in the production. Also, in the experts’ view, information/numbers extracted from scientific literature were somewhat different for Dutch situations. Most research articles describe experiments performed in the USA or Canada, using, for example, special feed formulations that are quite different from what is used in the Netherlands.

Model: general outline The economic farm model consists of two stages: the farrowing and fattening stages. In the farrowing stage piglets are kept until their live weight has reached 25 kilograms, after which they are transferred to the fattening stage. In the fattening stage, piglets grow until an approximate live weight of 115 kilograms. Then the hogs are transported to the

101 A feasibility study of GM pork production in the Netherlands slaughterhouse. Optionally, the farrowing and fattening stages can either be specialised as separate stages (different farms) or be integrated into one stage. The first option is assumed in this study. In the model three types of parameters are distinguished: input parameters, output parameters and parameters that represent inter-stage relations. On the one side, the model input consists of technical parameters that originate from literature and expert opinions, on the other it is composed of relevant economic parameters of production that are taken from the Agricultural Information and Knowledge Centre and the Research station for Animal Husbandry (2006a,b). One example of technical parameters is the number of weaning days or kg of feed, as to the economic parameters, this is a price per kg of feed or value of sows. The output parameters are summarised in costs categories and differ depending on the stage in the chain. For the farrowing stage, these are labour costs, housing, interest, feed costs, health care costs (including medicines), insemination costs, additional returns and other costs (including water, heating, electricity, manure and other). For the fattening stage, these are labour costs, housing, interest, purchase of piglets, transportation, delivery costs, feed costs, healthcare costs (including medicines), mortality and other costs. The total costs are presented on an animal basis, i.e. costs per sow per year (farrowing stage) and costs per hog delivered. The inter-stage parameter used in this model is the cost price of piglet sold for fattening. The main output parameter is the cost price per kg of pork. In the model we distinguish a default situation in which the regular pork parameters are analysed and a GM situation when one of the applications is applied. The economic and technical data of alternative situations (a specific GM application) are compared with a default situation (no GM). The default situation in this study is a Dutch pork production chain. To determine the impact of the main input parameters in the situations with GM applications, a sensitivity analysis was performed. In some cases, depending on the GM application, it was reasonable to vary one parameter at a time, in other cases two parameters at a time. The results give insight into the impact of GM applications on (the distribution of) costs and revenues throughout the pork production chain. All calculations are made for a farm of a size of 540 sowplaces and 3300 hogplaces with an occupation rate of 93%. For the farrowing stage, a farmer (manager) and one co-worker is assumed, which in total represents 4307 working hours. For the farrowing stage, one co- worker with 2349 working hours is assumed (Agricultural Information and Knowledge Centre and the Research station for Animal Husbandry, 2006b).

102 Chapter 5

Farrowing stage: default and GM applications In the farrowing stage, regarding both default and GM situations, the main important parameters concern the sow’s feed production cycle and productivity (Table 2). In particular, these are the weaning and rearing days, litters and piglets per sow per year and piglet mortality (both pre- and post-weaning).

Table 2. Major technical input values (farrowing) Variable Default Bovine GM Anti- Enviro- LP Micro- Immuno- gene crops bodies pig plants bial castration phytase Duration of sow 157 155.5 150 cycle (days) Litter size (piglets) 11.9 Piglets’ mortality (%) pre-weaning post-weaning 12.1 1.9 Weaning at (days) 28 26.5 21 Delivering at (days) 70 68.3 63 Piglets weight (kg) 25 Litters per sow 2.33 2.39 2.43 per year Price feed (€/100kg) piglets 28.5 28.24* sow, boar 18 17.8 P inorganic added in 3.18 0.48 2.83 1.39 diets (kg) P excreted (kg) 7.68 4.98 7.32 5.89 Time for castration 14.73 (minutes per all piglets per sow per year) * prices include GM maize and GM soy In the default situation the sow’s production cycle is assumed to be 157 days. It is based on a gestation period of 115 days, a lactation period of 28 days and a 14 day-interval from weaning until breeding. Piglets are weaned at 28 days and then housed in the separate rearing pens. In the pens they are reared, approximately up to an age of 70 days until a live weight of 25 kilograms. Of the 11.9 piglets born alive per litter, 12.1% die before weaning and 1.9% die after weaning. Therefore, the number of piglets per sow per year is corrected according to the mortality rate. In a default situation, it is assumed that a total number of 4307 working hours is available (Agricultural Information and Knowledge Centre and the Research station for Animal Husbandry, 2006b). The amount of feed consumed per animal group (piglets, lactating sows, breeding sows, boars) was taken according to the standard ration with average net energy

103 A feasibility study of GM pork production in the Netherlands content (Central Feeding Bureau, 2006). Thus, for example, for a piglet 28.1 kg of feed is required for the whole farrowing period (Agricultural Information and Knowledge Centre and the Research station for Animal Husbandry, 2006a). As for the GM situation, based on the literature and experts evaluations, depending on the nature of GM application one or more input parameters can be different from the default2. For example, in the case of “Bovine gene”, this application has a direct effect on the number of weaning days, due to the increased growth rate of piglets, which is stimulated by the increased amount of sow milk produced. The expected weaning days for the “bovine” piglets is 26.5 days, which is 5.4% lower than in the default situation. Decrease in the weaning days has a consequent effect on the production cycle of sows and the number of litters and piglets per sow per year. To produce more milk, sows need additional feed that is calculated based on the assumption that 1 kg of extra milk requires 0.5 kg of feed. Given the better growth rate of “bovine” piglets, the number of rearing days has decreased. That leads to the reduction in the amount of piglet feed and medicines required. For “GM crops”, only prices for feed that contains GM soy or GM maize are affected, implying the reduction in feed prices. The “Antibodies” application affects the same technical input parameters as “Bovine gene”: weaning and rearing days, sow cycle and number of litters and piglets per sow per year. Only this effect, however, is due to the antibodies that protect early-weaned piglets from diseases, in particular from E.coli. The GM applications “Enviropig”, “LP plants” and “Microbial phytase” influence the amount of inorganic phosphorus (P) supplemented to the diet, which is linked to the price of feed and P excreted in the manure. Regarding “Immunocastration”, because no castration of piglets is required, the amount of labour is decreased in the farrowing stage.

Fattening stage: default and GM applications At the starting live weight of 25 kg, piglets are transferred to the fattening stage, where they stay for approximately 120 days, until a live weight of 115 kilograms. As in the farrowing stage, we defined also key input parameters here. These are the average daily growth rate (in default: 774 grams/day), which is related to the feed conversion ratio (in default: 2.65 kg feed/kg gain). The feed conversion ratio is the amount of feed per kilogram of live weight gain. The number of deliveries per year is related to the average daily growth rate implying that a higher number of deliveries per year can be achieved, if the average daily growth rate

2 In the GM situation, regarding “Bovine gene” and “Enviropig” applications, prices for breeding sows are assumed to be the same as in a default situation.

104 Chapter 5 increases. The amount of feed for hogs in the default and GM situations was calculated according to the standard ration with average energy content (Central Feeding Bureau, 2006) (Table 3). Table 3. Major technical input values (fattening) Variable Default Bovine GM Anti- Enviro- LP Micro- Immuno- gene** crops bodies** pig plants bial castration phytase Av. daily gain 774 794 (grams/day) Feed conversion 2.65 2.58 ratio Mortality rate 3 (%) Number of 3.13 3.21 deliveries per cycle Prices (€/100kg) feed 18 17.8* P inorganic added 0.26 0.00 0.12 0.11 in diets (kg) P excreted (kg) 0.84 0.58 0.80 0.62 Price of vaccine 1.42 1.42 (€/2 pieces) Twice vaccination 0.54 (minutes per pig) * prices include GM maize and GM soy ** no technical effects are expected in this stage

In the fattening stage, the “Bovine gene” and “Antibodies” applications do not have an effect on technical input parameters. In case of “GM crops”, the same effects are expected as in the fattening stage, i.e. prices for GM soy and maize ingredients in diets are lower. As to “Enviropig”, “LP plants” and “Microbial phytase”, the same input parameters are influenced in this stage: P inorganic supplemented and P excreted. With respect to the “Immunocastration”, hogs are vaccinated twice in the fattening stage. This involves additional labour and GM vaccines. Due to the effects of “Immunocastration”, the hogs’ average daily gain and feed conversion ratio increase by 2.6% and 2.5%, respectively.

3. Results Default The analysis of production costs in the farrowing and fattening stages resulted in a cost price of €46.20 per piglet sold and €1.39 per 1 kg of pork produced (Tables 4 and 5). These results are based on the input values described in the previous section, where in the farrowing stage,

105 A feasibility study of GM pork production in the Netherlands a sow produces 2.33 litters per year on average, resulting in 23.8 piglets sold for fattening per sow per year. In the fattening stage, 3.08 (adjusted for mortality) production cycles were realised. Comparing the categories of production costs, feed is the main cost component in the farrowing stage, representing 36.9% of the total production costs. In the fattening stage feed costs are the second important cost component, representing 34.2% of total production costs. The major costs in this stage include the purchase of 25 kg-piglets, representing 36.7% of the total production costs in the farrowing stage. The farrowing stage is relatively labour- and capital intensive, compared with the fattening stage. Also, other costs including manure costs are higher in the farrowing stage (15.4%) than in the fattening stage (7.7%).

Effect of GM application on economic performance of the pork production chain As to the GM applications, the representation of the cost categories in both stages is about the same as in the default situation. Because some costs are the same in default and all GM applications, only the costs components that change due to GM applications are reported here. In general, small differences in the outputs per stage can be observed (Table 4 and 5).

Table 4. Cost price per 25 kg piglet in default situation and for GM applications (euros) Costs Default Bovine GM Anti- Enviro- LP Micro- Immuno- gene crops bodies pig plants bial castration phytase Costs per sow per year Labour 157.4 157.4 157.4 157.4 157.4 157.4 157.4 152.4 -castration 0 0 0 0 0 0 0 -4.5** Feed 406.1 402.1 402.3 414.6 405.3 406.0 405.5 406.1 Health 48.0 45.3 48.0 45.5 48.0 48.0 48.0 48.0 Other* 169.5 169.5 169.5 169.5 143.2 166.0 152.1 169.5 -manure 0 0 0 0 -26.4 -3.5 -17.5 0 Total 1100.9 1094.2 1097.0 1106.8 1073.6 1097.2 1080.2 1096.4 Cost price per 25 kg piglet Euro/piglet 46.2*** 45.6 46.0 44.5 45.0 46.0 45.3 46.0 % change 100 -1.3 -0.4 -3.7 -2.5 -0.3 -1.9 -0.4 * cost category “other costs” includes water, heating, electricity, telephone, insurance, manure costs ** negative sign means the cost saving; for example, due to the GM €4.50 are the cost savings on labour; this number is already included in total labour *** based on the number of piglets per sow per year; in the default 23.8 piglets per sow per year

Table 5. Cost price per kg pork in default situation and for GM applications (euros)

106 Chapter 5

Costs Default Bovine GM Anti- Enviro- LP Micro- Immuno- gene crops bodies pig plants bial castration phytase Costs per hog delivered Labour 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.1 -vaccination +0.2* Piglet price 46.2 45.6 46.0 44.5 45.0 46.0 45.3 46.0 Feed costs 43.1 43.1 42.8 43.1 43.1 43.1 43.1 42.0 Mortality 2.5 2.5 2.5 2.4 2.4 2.5 2.5 2.5 Other costs** 9.7 9.7 7.1 9.70 7.0 7.9 7.1 11.0 -vaccine price 0 0 0 0 0 0 0 +1.4*** -manure 0 0 0 0 -1.7*** -0.3 -1.4 0 Total 125.9 125.4 125.5 124.2 122.0 124.1 122.5 127.9 Costs per kg pork Euro/kg 1.393 1.387 1.387 1.374 1.360 1.389 1.368 1.390 % change 100 -0.4 -0.4 -1.4 -2.4 -0.3 -1.8 -0.2 * positive sign means that labour costs increased due to vaccination; this number is already included in total labour ** cost category “other costs” includes water, heating, electricity, telephone, insurance, manure costs *** “other costs”: positive or negative sign means the increase/decrease of costs due to GM applications; these numbers are already included in “other costs”

The production with GM applications is cheaper than the default. The differences in cost prices (decrease) range from 0.3% to 3.7% in the farrowing stage and from 0.2% to 2.4% in the fattening stage. The differences in the output results can be explained by the changes in the technical values in the input. In “Bovine gene” application, the feed costs and healthcare costs per sow per year are decreased by €4 and €2.70, respectively, due to the increased growth of piglets and, as a consequence the decreased number of weaning and rearing days. This resulted in the decrease of 1.3% of cost price of piglet produced. In the fattening stage, however, no technical improvements are expected and, therefore, the cost price of 1 kg of pork is only 0.4% lower compared to the default, due to the prices for purchasing piglets. Among all GM applications, the largest effect on the cost price of piglets produced in the farrowing stage can be expected with respect to the “Antibodies” application (3.7%). However, in the fattening stage this application has only resulted in the decrease of the cost price of 1 kg of pork by 1.4%. In the farrowing stage, the decrease in the cost price of piglets produced is explained by the decrease in weaning days (by 7 days) and, thus, the increase of piglets per sow per year (24.9 piglets). Although in the fattening stage, the improved results of

107 A feasibility study of GM pork production in the Netherlands the farrowing stage are presented by purchasing prices of piglets, this does not reduce much (only by 1.4%) the cost price of 1 kg of pork. In the case of “Enviropig”, “LP plants” and “Microbial phytase”, the improvements in the output results are due to decreased feed costs (less inorganic P is supplemented in the feed) and savings as to the manure costs (less P excreted). Among these three applications, “Enviropig” has the highest effect on the output in both stages. With this application, the cost price of piglets produced is €45.03, which is 2.5% lower than in the default situation. The cost price of 1 kg of pork produced is €1.36, which is 2.4% lower. The decrease in cost prices in both stages is due to the decrease in feed costs and other costs, more specifically manure costs. This application has the highest saving in manure costs in both the farrowing (€26.40) and fattening (€1.70) stages compared to the “LP plants” and “Microbial phytase”. The second highest effect among these three applications on the output results has the “Microbial phytase”. Due to the decrease in feed costs by €0.60 and manure costs by €17.50 in the farrowing stage and manure costs by €1.40 in the fattening stage, the cost price per piglet and per 1 kg of pork is decreased by 1.9% and 1.8% in both stages respectively compared to the default situation. “LP plants” application has the smallest effect on the output, also among all GM applications, resulting in a 0.4% difference (decrease) in cost prices. This can be explained by the fact that in the so-called “European diet” only a limited amount of GM soy and maize is used, due to the diet composition (Gâtel & Porcheron, 2003; Hazzledine, 2003; Van Cauwenberghe, Peisker & Pack, 2003). Therefore, the LP plants consisting of GM soy and maize have only a limited effect on the production cost prices. Similar to “LP plants”, the “GM crops”, used in a limited amount in the diets, have a small effect on the cost prices, which are decreased by 0.4% in both stages. The difference observed is due to the lower prices for GM soy and GM maize. The last application, “Immunocastration”, resulted in a decrease in cost price per boar delivered by 0.4%, which is mainly due to the savings on labour costs. In the farrowing stage the labour costs are reduced by €4.50 per sow per year, because the farmer does not need to castrate piglets. However, in the fattening stage “Immunocastration” requires vaccination of the animals, which results in slightly increased labour costs. Moreover, the costs of vaccines increase the total production costs. The “Immunocastration” also increases the growth rate of piglets by 2.6%, causing the number of fattening cycles to rise from 3.08 (in the default) to 3.16 (+0.08) per year. Because daily gain is related to the feed conversion ratio, the latter decreased from 2.65 (in the default) to 2.58. Cost price per hog delivered (male) decreased by 0.2% compared to the default.

108 Chapter 5

Sensitivity analysis: GM applications The results of the sensitivity analysis for both stages are presented in Table 6. The analysis is based on the minimum (min) and maximum (max) values, which represent the possible ranges for the impact of GM applications of certain input parameters on the output results. In Table 6 the default values represent the values for the certain input parameters regarding the situation with no-GM applied and GM situation (most likely) with the same input as in Table 3. Minimum and maximum values are the ranges of the most likely GM situation; however, the results are compared to the default with no-GM situation.

Table 6. Results of sensitivity analysis (relative to the default cost price results, in % of change, Min/Max values) Farrowing Fattening Default Change Change Default Change Change parameters parameters wrt cost parameters parameters wrt cost (value) (%) price (value) (%) price (%) (%) Bovine gene weaning days 28/26.5* -10/+10 -4.0/+3.0 n.a. n.a. -1.4/+1.1 GM crops Prices: feed Piglet 28.5/28.24 -1,-3/-3,-5** -0.4/-0.5 n.a. n.a. n.a. Sow 18.0/17.8 -1,-3/-3,-5 18.0/17.8 -1,-3/-3,-5 -0.4/-0.6 g of soy/maize 200/200 800*** -1.0 200/200 800 -1.2 in kg of feed Antibodies weaning days 28/21 -14.0/+14.0 -5.4/-2.1 n.a. n.a. -1.4/-0.7 Enviropig P supplemented 3.18/0.48 -10/+10 -2.9/-2.0 0.26/0.00 -10/+10 -2.7/-1.8 LP plants P supplemented 3.18/2.83 -10/+10 -1.0/+0.3 0.26/0.12 -10/+10 -0.9/+0.4 Microbial phytase P supplemented 3.18/1.39 -10/+10 -4.0/-1.4 0.26/0.11 -10/+10 -2.9/-1.2 Immunocastration Av. daily gain n.a. n.a. n.a. 774/794 +1.6/+3.6 +0.4/-0.7 (hogs) * 28/26.5 means that 28 weaning days are in the default with no-GM situation and 26.5 weaning days with GM situation ** the reduction in prices for GM crops; comma separates the reduction in prices for GM maize and GM soy *** only the increase of GM crops in diets is assumed here; change of the parameter (in this case) is presented in total grams of GM soy/maize in 1 kg of feed

Some peculiarities of the sensitivity analysis should be mentioned here. In some cases, the changes in the input parameters are interrelated like in the “Bovine gene”. For example, in the case of “Bovine gene”, the important input parameter is the growth rate of piglets, which in the model is related to the weaning and as a consequence to the rearing days. Thus, in the

109 A feasibility study of GM pork production in the Netherlands sensitivity analysis, the number of weaning days is taken as a variable input parameter that is connected to the rearing days and amount of feed. In the case of GM crops not only the prices for the crops can be varied, but also the amount of GM crops in the diet. Although in European diets, wheat is prevailing over other cereals and particularly over the use of soy and maize (Gâtel et al., 2003; Hazzledine, 2003; Van Cauwenberghe et al., 2003), in the sensitivity analysis the amounts of maize and soy have been increased approximately from 20% (in the default) to 80% for all diets used in the farrowing and fattening stages. The highest effects of 2.1% (min) and 5.4% (max) of the cost price reduction in the farrowing stage can be expected due to the “Antibodies”. However, in the fattening stage the effect of “Antibodies” is diminished compared to other applications. In the fattening stage, the highest effects of 2.7% (min) and 1.8% (max) and 2.9% (min) and 1.2% (max) can be expected, due the “Enviropig” and “Microbial phytase”, respectively. All in all, the sensitivity analysis of the effects of GM applications on the cost price of piglet and 1 kg of pork produced showed that the trend of the outcomes does not change and the economic effects of applying GM applications is still rather limited.

4. Discussion and future outlook The goal of this paper was to assess the potential economic impact of adoption of GM technology on Dutch farms. To achieve this goal, an economic farm model for two production stages, i.e. farrowing and fattening, was developed in which the impact of different GM applications could be analysed and compared with the default situation. From the results obtained it can be concluded that although the GM applications improved output results in both stages, the GM applications had a limited economic impact. Overall, pork producers can expect reductions in their cost prices by 0.3% to 3.7% in the farrowing stage and by 0.2% to 2.4% in the fattening stage by using GM applications. Of all the GM alternatives, “Antibodies” showed the largest potential effect (3.7% cost price reduction) on the output results in the farrowing stage, whereas “LP plants” had the lowest (0.3% cost price reduction). In the fattening stage, “Enviropig” had the largest effect (of 2.4% reduction) on the cost price of 1 kg of pork and “Immunocastration” had the lowest of 0.2% cost price reduction. Generally, cost savings and improvements in the output results were larger in the farrowing stage than in the fattening stage. Moreover, in many GM cases, changes within one stage did not affect much the economic performance in another stage.

110 Chapter 5

The sensitivity analysis showed that the changes in the main input parameters had little effect on the output results in both stages. The results obtained suggest that the current possible range of improvements in output results due to the GM applications is rather limited and most likely will not be able to motivate farmers to introduce GM into their production system. Results presented in this article should be interpreted carefully, especially since a complete insight into the effects of GM applications is hard to obtain, due to the limited quantitative information about the expected effects and benefits in production stages. In Europe, based on the consumer resistance to GM technology and GM foods and EU legislation, many food processors and retailers are trying to restrict or avoid the production or supply of GM foods and GM ingredients (Anonymous, 2000). These have an effect on the speed of the developments of new GM applications. Especially for some GM applications, developments are still at laboratory level and being discussed whether it is ethical and safe to use them in food and feed production. “GM animals” is one of the examples (Sang, 2003). In this article the GM data originated from scientific articles and evaluated by experts. In the majority of the cases, the scientific experiments are done in the US or Canada, implying that some conditions such as feed diets and general breeding requirements can be different from those used in Europe. This evolved some difficulties for experts to evaluate the real effect of GM application in European conditions. Benefit realisation is considered an important issue in technology adoption for the society as whole, producers, and consumers (Novoselova et al., 2007a; Frewer, 2003). For the last- mentioned, however, two important issues should be mentioned in this study. The first topic concerns the research goal of this study, implying the evaluation of economic effects of GM applications. However, given the nature of GM applications used in this study, the technology does not ultimately aim at improving economic results, but some applications, such as “Immunocastration” or “Enviropig”, were developed to improve animal welfare and environment. Therefore, only reporting economic benefits should not diminish the value of GM applications. For example, in the case of “Enviropig”, “LP plants” and “Microbial phytase”, reduction in the excretion of phosphorus into the environment compared with the current practice are expected to vary from 5% to 35% (this study) or in some cases up to 75% (Golovan, Meidinger, Ajakaiye, Cottrill, Wiederkehr, Barney et al., 2001). These improvements can be important to farmers, because they can save on the payments for manure, but most importantly, to the society. Another example, “Immunocastration”, is considered a potential alternative for the castration of piglets, which causes pain and stress to

111 A feasibility study of GM pork production in the Netherlands the animals. In the EU, there are serious discussions about the “animal-unfriendly way of surgical castration” (Turkstra, 2005). In Norway, for example, stricter regulation will come into force, which from January 2009 will forbid surgical castration of piglets (Ten Hooven, 2005). Therefore “Immunocastration” can be a new animal-friendly alternative to surgical castration. The second issue is related to the availability of GM applications with special consumer benefits. The first generation GM applications (mainly GM crops) has been widely criticised for not having special attractive benefits for the consumers (Gaskell, Allum, Wagner, Kronberger, Torgersen, Hampel et al., 2004). Although the presence of attractive benefits can be one of the possible positive factors in acceptance of GM foods, the number of GM applications developed for the needs and concerns, i.e. animal welfare concerns, of consumers is still very limited and still in its infancy. Therefore, future research should be directed towards the developments of new products with desirable consumer benefits.

References Ag-West Biotech Inc. (2005). "Feed Biotechnology: The Science of Animal Nutrition." Retrieved 10.04 http://www.agwest.sk.ca/publications/ls_feed.php, 2006. Agricultural Information and Knowledge Centre and the Research station for Animal Husbandry (2006a). Quantitative information on livestock industry 2005-2006. Lelystad, (in Dutch). Agricultural Information and Knowledge Centre and the Research station for Animal Husbandry (2006b). Explanation of the calculation of the national recommended price for feeder pigs. The Hague, (in Dutch). Anonymous. (2000). "Economic Impacts of Genetically Modified Crops on the Agri-Food Sector; European Union, Commission of the European Communities." Retrieved 28.07, 2004, from http://europa.eu.int/comm/agriculture/publi/index_en.htm. Bonneau, M. & Laarveld, B. (1999). Biotechnology in animal nutrition, physiology and health. Livestock Production Science, 59(2-3), 223-241. Cardello, A. V. (2003). Consumer concerns and expectations about novel food processing technologies: effects on product liking. Appetite, 40(3), 217-233. Central Feeding Bureau (2006). Table booklet Animal Nutrition: feed standards livestock and nutritional values of feed. Lelystad, (in Dutch).

112 Chapter 5

De Vries, A. G., Faucitano, L., Sosnicki, A. & Plastow, G. S. (2000). The use of gene technology for optimal development of pork meat quality. Food Chemistry, 69(4), 397-405. Frewer, L. (2003). 10. Societal issues and public attitudes towards genetically modified foods. Trends in Food Science & Technology, 14(5-8), 319-332. Frewer, L. J., Howard, C., Hedderley, D. & Shepherd, R. (1997). Consumer attitudes towards different food-processing technologies used in cheese production--the influence of consumer benefit. Food Quality and Preference, 8(4), 271-280. Gaskell, G., Allum, N., Wagner, W., Kronberger, N., Torgersen, H., Hampel, J. & Bardes, J. (2004). GM Foods and the Misperception of Risk Perception. Risk Analysis, 24(1), 185-194. Gâtel, F. & Porcheron, E. (2003). 7.- The role of cereals in the European protein supply. In FEFANA: Protein requirements and supply for a competitive European pig production in 2010 (23-27). Brussels, March 18, 2003. Golovan, S. P., Meidinger, R. G., Ajakaiye, A., Cottrill, M., Wiederkehr, M. Z., Barney, D. J., Plante, C., Pollard, J. W., Fan, M. Z., Hayes, M. A., Laursen, J., Hjorth, J. P., Hacker, R. R., Phillips, J. P. & Forsberg, C. W. (2001). Pigs expressing salivary phytase produce low-phosphorus manure. Nature biotechnology, 19(August), 741-745. Hazzledine, M. (2003). 9.- The opportunities offered by amino acids in practical feed formulation.In FEFANA: Protein requirements and supply for a competitive European pig production in 2010 (33-38). Brussels, March 18, 2003. Koivisto Hursti, U.-K., Magnusson, M. K. & Algers, A. (2002). Swedish consumers’ opinions about gene technology. British Food Journal, 104(11), 860-872. Kuznesof, S. & Ritson, C. (1996). Consumer acceptability of genetically modified foods with special reference to farmed salmon. British Food Journal, 98(4/5), 39-47. Miles, S., Ueland, Ø. & Frewer, L. J. (2005). Public attitudes towards genetically-modified food. British Food Journal, 107(4), 246-262. Noussair, C., Robin, S. & Ruffieux, B. (2004). Do consumers really refuse to buy genetically modified food? The Economic Journal, 114, 102-120. Novoselova, T. A., Meuwissen, M. P. M. & Huirne, R. B. M. (2007a). Adoption of GM technology in livestock production chains: an integrating framework. Trends in Food Science & Technology, doi: 10.1016/j.tifs.2006.12.005 (in press). Novoselova, T., van der Lans, I. A., Meuwissen, M. P. M. & Huirne, R. B. M. (2007b). Heterogeneity in consumer acceptance of GM applications in pork production:

113 A feasibility study of GM pork production in the Netherlands

attitudinal and socio-economic determinants. Submitted to Food Quality and Preference O'Connor, E., Cowan, C., Williams, G., O'Connell, J. & Boland, M. (2005). Acceptance by Irish consumers of hypothetical GM dairy spread that reduces cholesterol. British Food Journal, 107(6), 361-380. O'Connor, E., Cowan, C., Williams, G., O'Connell, J. & Boland, M. P. (2006). Irish consumer acceptance of a hypothetical second-generation GM yogurt product. Food Quality and Preference, 17(5), 400-441. Phillips, P. W. B. (2002). Biotechnology in the global agri-food system. Trends in Biotechnology, 20(9), 376-381. Sang, H. (2003). 6. Genetically modified livestock and poultry and their potential effects on human health and nutrition. Trends in Food Science & Technology, 14(5-8), 253-263. Solomon, M. B., Pursel, V. G., Campbell, R. G. & Steeleae, N. C. (1997). Biotechnology for porcine products and its effect on meat products. Food Chemistry, 59(4), 499-504. Ten Hooven, M. (2005). Stress remains inspite of expensive shot. Pig progress; www.agriworld.nl, 21(1). Turkstra, J. (2005). Active immunization against gonadotropin-releasing hormone: an effective tool to block the fertility axis in mammals, Utrecht University; Ridderprint: Ridderkerk. Van Cauwenberghe, S., Peisker, M. & Pack, M. (2003). 10.- The role of amino acids in an environmentally friendly competitive European pig production. In FEFANA: Protein requirements and supply for a competitive European pig production in 2010 (44-48). Brussels, March 18, 2003. Van Kempen, T. (2000). Genetically Modified Organisms in Animal Production. Swine News, North Carolina Cooperative Extension Service, 23(5).

114

Chapter 6

General Discussion

General Discussion

1. Introduction Pork production chains are being restructured as a result of consumer concerns about aspects such as food safety, animal welfare, and environmental sustainability. A further development in the pork chain is the use of genetic modification applications. In principle GM applications provide considerable technical and economic opportunities, but they also comprise concerns for (specific segments of) consumers. The overall objective of this study was to gain insight into the adoption aspects of genetic modification in the pork production chains. Chapter 2 reviews the available literature on adoption of genetic modification in livestock production chains. In this chapter, a conceptual framework for analysing the adoption of new technologies in food chains is presented. This provides a basis for a new integrated chain approach for analysing genetic modification in pork production chains. The main part of the framework is the consumers whose opinions play an important role in the process of successful technology adoption. The analysis of consumer attitudes towards GM technology in general and consumer preferences for specific GM products (pork) are described in Chapter 3. In Chapter 4 the analysis of consumer preferences for GM pork is extended by identifying distinct segments of consumers with different preferences for GM pork. Furthermore, the heterogeneity in the segments obtained is explained by attitudinal and socio-economic characteristics of consumers in each segment. For a selected number of GM applications, (the distribution of) costs and revenues in the pork production chain are studied in an economic farm model in Chapter 5. This final chapter, i.e. the General Discussion, discusses different emerging issues related to the development of genetic modification in food chains and further implications of the project. Section 2 presents the chain approach used in this study and the assumptions regarding the use of GM. In Section 3, issues related to the data collection regarding consumer attitudes and preferences for GM foods and expert evaluations of GM applications are discussed. In Section 4, the practical implications of the results obtained from the consumer study and the economic farm model are presented. Section 5 addresses an outlook for the future research on the adoption of genetic modification. The last section presents the main conclusions of the thesis.

2. Pork production chain The present research focuses on the adoption of genetic modification in the pork production chains. The pork production chain presents stages that are connected and aimed at the production and supply of pork to the end-user: the consumers. The concept of a consumer-

116 Chapter 6 driven chain is used in this research. The pork chain consists of many stages. Due to the limited information available, it was not feasible in the current project to include all stages. Focus was on three stages in this research: two belong to the direct production of pork (farrowing and fattening) and the third to consumption (consumers). This research has covered an important part of the chain with a few selected GM applications, but there are many other stages, such as feed production and breeding, that could be considered. In this research these were indirectly included through feed prices and animal performance. In this study two production situations are analysed: the current situation (default), when no GM is applied and a GM situation when one of the GM applications is used. However, in reality the production situation is not “black” and “white”; it is more complex. It is possible that some GM applications, in terms of feed or feed ingredients, are being used in some production chains or their use is being partially restricted to no-use at all. In the EU, given the strong consumer opposition to GM foods, the current strategy of many food producers and retailers is to avoid or restrict GM foods or GM ingredients used in food production (Anonymous, 2000; Levidow & Bijman, 2002). But a 100% of no GM in the chain is probably not possible.

3. Data availability To consider the evaluation of technology adoption in food chains, we need data on consumer preferences for GM pork and technical and economic information on possible impact of GM applications in the pork production chain. Data are the vital basis for the research. Although good quality data are very important, the collection process is not always an easy straightforward task. Several issues related to data collection for both consumer and farmer (producer) are discussed below.

Consumer preferences for GM pork In this study data on consumer attitudes towards GM technology and preferences for GM food in general, and GM pork in particular, were collected via a mail questionnaire. This resulted in 253 (11%) fully filled-in questionnaires. The questionnaire did not require any special knowledge on the topic of genetic modification. However, for some consumers, the topic was considered “very difficult and somewhat unknown”, some respondents were “afraid not to have special knowledge that would help them to fill in this questionnaire”. Apparently, it is possible that some consumers were not able to complete all questions, or were very conservative or attached themselves incorrectly to the group of “No-GM” consumers.

117 General Discussion

Design consideration: In the questionnaire consumers were faced with four types of GM applications considered in the pork production chain. These four GM applications, i.e. GM animal, GM feed, GM additives & medicines and GM bacteria were presented in a general way. The GM product researched was GM pork produced with the help of one of these four GM applications. The difference between conventional pork and GM pork was in the assumed attractive consumer benefits attached to the GM pork. These were price reduction, improvements in animal welfare, impact on the environment, quality, and level of residues in meat. However, throughout the design of the consumer survey, there was always the issue of how to present benefits achieved by the GM applications to the consumers in a clear and understandable way. The study ultimately opted for a general way of presenting the benefits, for instance “improved animal welfare”. The general nature of the attributes considered (Chapter 3) was chosen for two reasons. Firstly, the delicate nature of the topic and as a consequence a lack of appropriate information complicated the selection of only one particular example that could be presented to the consumers in a more detailed description (attributes levels). Secondly, and far more importantly, is that consumers do not possess special technical knowledge and will not be able to evaluate properly the improvements in certain attributes. It is always a challenge in new research to find a group of representative consumers that will be able to understand the technical side of the improvements. To overcome this problem, several options can be considered, such as substituting consumers’ opinions with those of relevant consumer organisations or providing consumers with the improvements quantified in terms of well-understood examples. As to the first option, for example, Den Ouden, Nijsing, Dijkhuizen. & Huirne (1997) used consumer- related groups (retailers) to evaluate the perception of different categories of animal welfare in pork production. Van Calker, Berentsen, Romero, Giesen, & Huirne (2006), used the animal protection organisation, nature organisations and consumer organisations, due to the difficulty to find a representative group of consumers that was able to judge the sustainability of dairy farming systems. The motivation to use these groups was that these groups in the debates about sustainable dairy farming represent themselves as consumers. However, these groups may have a “stronger” view than the “average” consumer (Van Calker et al., 2006). Within this research, however, it would be difficult to find a representative neutral (unbiased) group of “consumer-related” organisations. The opinions of those groups are very diverse with a prevalence of a strong negative attitude towards genetic modification. Another option in overcoming the above-mentioned research difficulty is to present consumers with well- understood examples. For example, Travisi & Nijkamp (2004) presented a biodiversity

118 Chapter 6 indicator related to the environmental impacts in terms of number of endangered farmland bird species. In the current research, this approach would be possible if additional information on how consumers evaluate different specific improvements caused by GM, for example, in animal welfare, was available. The identification of such specific levels of attributes interesting to the consumer is an issue for future research. Representativeness of consumer sample: The consumer sample obtained consisted of consumers who had on average a smaller household, fewer children in the household and a much higher education level (Chapter 3). These make the results less representative of the Dutch population. However, these results could be a reflection of those consumers that have an opinion about genetic modifications and who are able to express their attitudes towards the technology. The results obtained showed that general attitudes towards technology were important determinants in the valuation of GM applications and benefits. One could argue that in reality the population consists of several groups that are different in their preferences. Therefore, the subsequent segmentation in this study identified five groups (segments) of consumers who share similar needs and demonstrate similar preferences (Chapter 4), and who can be considered a representation of consumer groups in the Dutch population.

Experts For the economic farm model, technical and economic data on GM applications were collected to study the effects of GM applications. The topic of genetic modification is very delicate and not so much information at farm-level impact could be made available. In research, when data availability is limited, expert’s opinions and experiences are considered a proper tool for collecting additional data (Horst, 1998; Van der Gaag, 2004). But who can be considered an expert in the field? According to Niels Bohr, the Nobel Prize Winner in physics (1922), “an expert is a person who has made all the mistakes that can be made in a very narrow field”. Another definition that can be found in the Free Encyclopedia, Wikipedia, “an expert is someone widely recognized as a reliable source of knowledge, technique, or skill whose judgment is accorded authority and status by the public or their peers”, is more useful in this research. Following Bohr’s definition, it would be difficult to find a person in the field of genetic modification in livestock production who experienced all the mistakes, given that this field is still in its development. The great opposition to genetic modification has resulted in the limited and decreasing funding for research in animal and meat science that “has left the meat production sector lagging behind other sectors of life science” in this field of research (Garnier, Klont & Plastow, 2003).

119 General Discussion

For this study, experts were selected to represent several research fields from which GM applications originated (Chapter 5). Due to the variability of GM applications in the pork production chain, it was difficult to find experts that could provide information for all applications in the chain. Experts, depending on the area of expertise, were nutritionists, animal physiologists, animal and plant breeders, veterinarians, and representatives of the feed industry. Private companies, such as feed and additive producers, were reluctant to provide information, since they were not keen on being involved in research concerning genetic modification. Another difficulty relating to the experts interviews was that the majority of the GM applications originated from research conducted in the USA and Canada, which complicated the translation of the findings to the situation in Europe. Therefore, European experts might have been conservative in providing their opinions and estimations. Even when performing sensitivity analysis on the wider range of values given by experts, results still show a relatively limited economic impact for all GM applications (Chapter 5).

4. From producer to consumer: integrated approach Chapter 2 of this study has advocated for a missing integrated approach when adoption of genetic modification in food chains is analysed. In such an approach, consumers’ preferences for GM products and producers’ perspectives should be taken into account. Therefore, the whole thesis was organised starting from the analysis of consumer preferences and identification of consumer segments followed by an economic analysis at farm level. The analysis of adoption of genetic modification in pork production chains as described in this thesis can be considered a first step. A few considerations regarding this approach and results obtained are discussed in this section. Meeting consumer preferences: As a result of the consumer survey, consumers’ preferences for GM pork produced with the help of various GM applications and benefits offered were analysed (Chapter 3). Furthermore, consumer data were used to identify five distinct consumer segments (Chapter 4). Among the consumer benefits offered, the improvements in animal welfare was valued most and reduction in environmental impact least. The latter result was surprising, given that consumers often mention the importance of reducing pollution from modern agriculture (Bonny, 2000). Also, looking at the GM applications directed to reduction of the environmental impact, it seems that consumers who base their decisions on the benefits offered by GM would not opt for these applications. This especially concerns applications where animals are modified, for example “Enviropig”, given the strong anti-preferences for GM animals. On the other hand, GM applications aiming at

120 Chapter 6 improvements of animal welfare would have potential, indicated by the strong consumer interest in animal welfare improvements. Another possible benefit that could be a motive to buy GM foods is taste. It was not used in the evaluated product profiles offered to the consumers, due to the absence of data on the developments of genetic modification in livestock regarding taste. From the results on the general question on important buying aspects of meat, however, this attribute was considered the most important by all consumers in the sample. Also, from another study in the Netherlands, taste was found to be the main purchase motive, strongly correlated with a high willingness to pay (Baltussen et al., 2006). Another discussion point regarding benefits analysed is the price levels used, i.e. 0%, 10% and 33% reduction. The choice for these high levels was motivated by three reasons. Firstly, the design and suggestions of previous consumer research were taken into account (see, for example, Gath & v. Alvensleben, 1998; Hu, Hünnemeyer, Veeman, Adamowicz, & Srivastava, 2004). For example, in the study by Gath et al. (1998), the prices of GM products (cheese, tomato soup, brand cheese, bread) were suggested to be “30-40% lower than the prices of competitive products”. Secondly, given previously observed consumer anticipation of GM products originating from animal production, these price levels were assumed to be necessary to motivate the buying behaviour of consumers. The third reason was the prior expectations as to considerable effects of GM technology (for example, Bonneau & Laarveld (1999)). Later in this study, the assumptions for price levels from the current GM applications did not seem to be justified by the economic results of the farm-level model. Consumers equal citizens? A so-called “consumer-citizen dichotomy” is sometimes observed in the behaviour of people (Ovaskainen & Kniivil, 2005; Salmela & Varho, 2006). Depending on the “nature” of a decision a person can act as citizen or consumer. It was acknowledged by many studies that individuals may act as citizens when making “hard” decisions, for example, concerning environmental issues, and as consumers when these decisions concern “personal or self-regarding wants and interests” (Sagoff, 1988; Salmela et al., 2006, p.8). With respect to the consumer results in this study, it was not clear whether respondents who showed little interest in GM pork or none at all (about 49% of respondents) were acting as consumers or citizens showing their social preferences? In reality, when the product is available at the supermarket, “a consumer can be mainly concerned with getting the cheapest product” (Salmela et al., 2006), but the same person in the role of citizen can have negative attitudes to GM technology and products produced with its help. Therefore, the possibility that some respondents in this study acted as citizens could not be ruled out. This may have had a negative impact on the attractiveness of GM applications.

121 General Discussion

Economics (cost considerations): According to Belli, Anderson, Barnum, Dixon and Tan (2001), identifying costs and benefits is the first and most important step in economic analysis of investment operations of any new project or investment. This also applies to the adoption of GM. Some economic costs resulting from GM applications are direct and not difficult to identify. In this study these were, for example, feed costs or housing costs. However, there is another group of costs which is difficult to identify and measure “especially if the project generates side effects that are not reflected in the financial analysis” (Belli et al., 2001, p.25). These costs can also be found when analysing adoption of GM technology; they are very important and should be kept in mind, but due to the difficulty of measuring them they were not included in the current economic analysis. For example, this could be contributions of the technology to the sustainability of society, costs of (perceived) food safety, labelling, transparency and environmental impact (FAO, 2001b; Barling et al., 1999; Enriquez, 2001). Labelling issue: Also in this research, labelling was an important issue. With the development of genetic modification, the argument that consumers have rights to know about GM used in their foods has evoked many discussions (Enriquez, 2001). As the technology component of GM foods is not distinguishable to the consumers before and after consumption, labelling has been recommended as a possible tool to inform consumers. Consumers, in turn, are also very supportive of the idea of having labels in which distinction between GM and non-GM can be seen (see, for example, the results of this study (Chapter 4); Carneiro, Minim, Deliza, Silva, Carneiro & Leao, 2005; Mucci & Hough, 2004; Poortinga & Pidgeon, 2004). As a result the EU has adopted new regulations on labelling GM foods. According to these regulations, all products containing or consisting of genetically modified organisms or containing ingredients produced from genetically modified organisms must be labelled, if the ingredients, considered individually, contain 0.9% or more of GMOs (Regulation (EC)
1830/2003). In contrast, no such requirements for labelling currently apply to products such as milk, meat, cheese and eggs from animals fed a diet with GM ingredients. The practical implications of GM labelling concern issues of quality of information needed and the expected implementation costs of GM labelling. The quality of information provided on the label will mainly depend on consumers (segments) to whom this information is addressed. Those consumers who reject or select products based on the use of genetic modification will need labelling that allows them to make this distinction (Teisl & Caswell, 2003). For these consumers a simple format “yes/no GM” of the information provided on a label is sufficient. Those consumers who base their accepting or rejecting decisions on the

122 Chapter 6 attributes provided or those who have no concerns about genetic modification and will compare the benefits of GM food with benefits of conventional food, will need more extended information on the label. Only providing a “yes/no GM” label to these consumers will be irrelevant or inadequate information (Teisl et al., 2003). Labelling would require a major and costly change in agricultural production. A variety of costs required, among other things for segregation and identity preservation, were described in the literature (Anonymous, 2000; Teisl et al., 2003). A few attempts were made to estimate some of these costs, but only for GM crops. Therefore, more information is needed to examine the costs of labelling for different products and scenarios of implementation of labelling systems along the production chains.

5. Outlook The integrated chain approach developed in this study can be used to further investigate the adoption of genetic modification in food chains. Furthermore, besides analysing consumer preferences and economic possibilities of genetic modification, this research urges for the necessity of studying this important and doubtful topic of genetic modification very carefully and also to consider other issues, such as consumer and producer concerns. This research can be developed further in several ways. The chain approach can be extended by considering more stages that are interlinked. For example, slaughtering would be a feasible option. Needless to say, this kind of chain approach can also be used to analyse the adoption of other GM applications in different chains or it can be applied to the analysis of adoption of another emerging technology. With respect to the consumers, the results of this study suggest that consumers’ acceptance of GM applications in livestock production chains is not one-sided. Negative as well as positive/neutral attitudes towards technology exist. Although consumers’ buying decisions are primarily influenced by attitudes towards GM technology, the GM benefits provided are also important. With respect to this, the current research provides useful insight into consumer acceptance of different GM applications used in pork production and benefits. However, future research can continue by investigating consumer acceptance of more specific GM applications and detailed levels of GM benefits. With regard to the “GM consequences” for producers, the uncertainty around technical input parameters of GM applications is still very high and the number of stages analysed is limited for drawing definite conclusions on the economic impact of GM applications to the

123 General Discussion whole pork production chain. Nevertheless, the results obtained provide valuable insights into the expected costs and benefits of the GM applications analysed.

6. Main Conclusions The following main conclusions can be drawn from this study: • Analysis of new technology adoption is not an easy task, especially in food chains, where consumers are generally very sensitive to change and evaluate new developments with special attention. Genetic modification is no exception. Not much research has been done yet on the adoption of genetic modification technology in livestock chains. The use of genetic modification, however, should particularly be studied from an integrated point of view, since not only farmers need to adopt the technology, but also the other members of the chain, including consumers. Uncertainty, for instance with respect to consumer acceptance and technology performance, is one of the major factors influencing the adoption of genetic modification (Chapter 2).

• Consumer decisions with respect to GM pork are mainly affected by the general attitude to the technology. If the attitude towards GM is positive, then acceptance of GM applications is more likely. Importantly, this implies that all the efforts of the producers and other chain participants who want to introduce GM should be directed to improving existing attitudes, and to creating a positive image of the technology by providing additional information to the consumers. Genetic modification in livestock production has potential, particularly in the case of GM feed. The presence of important consumer benefits, such as improvements in animal welfare, had a positive influence on consumers’ choices when choosing from different GM pork alternatives. Socio-economic and attitudinal characteristics of the consumers were found to be important determinants of consumers’ valuation of specific GM applications and their possible benefits (Chapter 3).

• Consumer preferences towards GM foods are heterogeneous. Five possible segments of consumers were identified. Segments are distinct with respect to consumer preferences as to GM applications and benefits these applications can offer to the consumers. While the majority of the consumers expressed their

124 Chapter 6

overall negative attitude towards GM technology and GM foods, consumers in some segments would be willing to purchase GM pork depending on the type of application used in the production and benefits offered. Greater insight into consumer acceptance of GM pork was provided by profiling consumers regarding attitudes and socio-economic characteristics (Chapter 4).

• Although the GM applications studied improved output results in the farrowing and fattening stages, economically the GM applications do not have much impact. Overall, pork producers can expect reductions in their cost prices of 0.3% to 3.7% in the farrowing stage and of 0.2% to 2.4% in the fattening stage by using GM applications. Of all the GM alternatives “Antibodies” showed the largest potential effect (3.7% cost price reduction) on the output results in the farrowing stage, whereas “LP plants” had the lowest (0.3% cost price reduction). In the fattening stage, “Enviropig” had the largest effect (2.4% reduction) on the cost price of 1 kg of pork and “Immunocastration” had the lowest effect (0.2% cost price reduction). Generally, cost savings and improvements in the output results were larger in the farrowing stage than in the fattening stage. In many GM cases, changes within one stage did not much affect the performance in another stage. The sensitivity analysis showed that changes in the main input parameters had little effect on the economic output results in both stages. The economic results obtained suggest that the current possible range of improvements in output results due to the GM applications is rather limited and most likely will not be able to motivate farmers to introduce GM in their production (Chapter 5).

References Anonymous. (2000). "Economic Impacts of Genetically Modified Crops on the Agri-Food Sector; European Union, Commission of the European Communities." Retrieved 28.07, 2004, from http://europa.eu.int/comm/agriculture/publi/index_en.htm. Baltussen, W. H. M., Wertheim-Heck, S. C. O., Bunte, F. H. J., Tacken, G. M. L., Galen, M. A. v., Bakker, J. H. & Winter, M. A. d. (2006). Een biologisch prijsexperiment; Grenzen in zicht? LEI [in Dutch]. Barling, D., de Vriend, H., Cornelese, J. A., Ekstrand, B., Hecker, E. F. F., Howlett, J., Jensen, J. H., Lang, T., Mayer, S., Staer, K. B. & Top, R. (1999). "The social aspects

125 General Discussion

of food biotechnology: a European view." Environmental Toxicology and Pharmacology, 7(2), 85. Belli, P., Anderson, J. R., Barnum, H. N., Dixon, J. A. & Tan, J.-P., Eds. (2001). Economic Analysis. Economic Analysis of Investment Operations: Analytical Tools and Practical Applications. Washington, D.C., World Bank Publications. Bonneau, M. & Laarveld, B. (1999). "Biotechnology in animal nutrition, physiology and health." Livestock Production Science, 59(2-3), 223-241. Bonny, S. (2000). "Consumer concerns about industrialized agriculture and food safety: importance, origin and possible solutions." Annales de Zootechnie, 49, 273–290. Carneiro, J. d. D. S., Minim, V. P. R., Deliza, R., Silva, C. H. O., Carneiro, J. C. S. & Leao, F. P. (2005). "Labelling effects on consumer intention to purchase for soybean oil." Food Quality and Preference, 16(3), 275. Den Ouden, M., Nijsing, J. T., Dijkhuizen, A. A. & Huirne, R. B. M. (1997). "Economic optimization of pork production-marketing chains: I. Model input on animal welfare and costs." Livestock Production Science, 48(1), 23. Enriquez, J. (2001). "Green biotechnology and European competitiveness." Trends in Biotechnology, 19(4), 135-139. FAO (2001b). "Genetically modified organisms, consumers, food safety and the environment." Series title: FAO Ethics Series - 2, ISBN: 9251045607, available on- line at http://www.fao.org/DOCREP/003/X9602E/x9602e00.htm. Garnier, J.-P., Klont, R. & Plastow, G. (2003). "The potential impact of current animal research on the meat industry and consumer attitudes towards meat." Meat Science, 63(1), 79. Gath, M. & v. Alvensleben, R. (1998). "The Potential Effects of Labeling Genetically Modified Foods on the Consumer Decisions - Preliminary Results of Conjoint Measurement Experiments in Germany." Effective Communication and GM Foods, AIR-CAT 5th Plenary Meeting, 4 (3), 18-28. Horst, H. S. (1998). Risk and economic consequences of contagious animal disease introduction. Wageningen, Wageningen University. Hu, W., Hünnemeyer, A., Veeman, M., Adamowicz, W. & Srivastava, L. (2004). "Trading off health, environmental and genetic modification attributes in food." European Review of Agricultural Economics, 31(3), 389-408. Levidow, L. & Bijman, J. (2002). "Farm inputs under pressure from the European food industry." Food Policy, 27, 31-45.

126 Chapter 6

Mucci, A. & Hough, G. (2004). "Perceptions of genetically modified foods by consumers in Argentina." Food Quality and Preference, 15(1), 43. Ovaskainen, V. & Kniivil, M. (2005). "Consumer versus citizen preferences in contingent valuation: evidence on the role of question framing." The Australian Journal of Agricultural and Resource Economics, 49, 379-394. Poortinga, W. & Pidgeon, N. F. (2004). "Trust, the Asymmetry Principle, and the Role of Prior Beliefs." Risk Analysis, 24(6), 1475-1486. Sagoff, M. (1988). The economy of the earth: Philosophy, law, and the environment, Cambridge, Cambridge University Press. Salmela, S. & Varho, V. (2006). "Consumers in the green electricity market in Finland." Energy Policy, 34(18), 3669. Teisl, M. F. & Caswell, J. A. (2003). Information Policy and Genetically Modified Food: Weighting the Benefits and Costs. Amherst, University of Massachusetts: 1-24. Travisi, C. M. & Nijkamp, P. (2004). Are Italians Willing to Pay for Agricultural Environmental Safety? A Stated Choice Approach.In 84th EAAE seminar, Zeist, the Netherlands. Van Calker, K. J., Berentsen, P. B. M., Romero, C., Giesen, G. W. J. & Huirne, R. B. M. (2006). "Development and application of a multi-attribute sustainability function for Dutch dairy farming systems." Ecological Economics, 57(4), 640. Van der Gaag, M. (2004). Epidemiological and economic simulation of Salmonella control in the pork supply chain. Wageningen, Wageningen University.

127

Summary

Summary

1. Introduction During the past decades the agro-food sector has changed and developed very drastically and rapidly. The undergone major changes involved the creation of new and expanded international markets, applying “sustainable” farming systems, improving food safety and adopting new technological developments. The nature of agro-production has changed and will change further due to the impact of new emerging technologies. A considerable number of new technological advances are related to food production. These include, for instance, irradiation, nutriogenomics and genetic modification, all of which aiming at improving food safety, animal welfare and the environment. Other technologies reduce production costs and the prices of final products. Although new food technologies hold the promise of such benefits to producers, consumers and society, not all of them are readily adopted and accepted. Technology adoption presents a complex process which has been widely studied from different perspectives. It has been recognised that many different factors play an important role in the adoption of existing and emerging food technologies. Many studies have been done on technology adoption at individual farm level or on acceptance at consumer level, usually with a focus on a single application of the technology or a single concern. However, there is a lack of insight into the integrated effects of technology applications at chain level. The latter is important, because for some technology applications, costs occur at one stage of the chain, while the benefits are enjoyed in another. Although much work has been done to date, more investigation is needed to study technology adoption from an integrated chain perspective. This thesis analyses the GM technology adoption from a chain perspective. The main objective of this study was to gain insight into adoption of genetic modification in the pork production chain. The following research questions were formulated to address the main objective in this study: 1. What aspects are relevant in the adoption process of genetic modification in pork production chains? 2. How do consumers value and make trade-offs between specific GM applications in pork production and benefits GM can offer them? 3. How heterogeneous are consumers’ preferences regarding different GM applications and how can heterogeneity be explained by attitudinal and socio-economic determinants? 4. What is the farm-level economic impact of selected GM applications in pork production chains?

130 Summary

2. Adoption of GM technology in livestock production chains: an integrating framework Chapter 2 defines the main terms used throughout this research, i.e. genetic modification and genetically modified organisms. The term “genetic modification” covers a range of ways of manipulating genetic material and means a process of moving genes into a new species and getting them to function. Although each plant and animal variety is a “genetically modified organism”, the term GMO refers to the product or technique which recombines specific sections of the genetic code from one organism to produce a new plant, animal, or microbe. Main developments of genetic modification in livestock production chains are also discussed. An integrating framework with important aspects for analysing the adoption of new technologies is presented. The framework proposed integrates consumers and chain participants and their concerns and benefits. This framework is used to review the literature on the adoption of different new technologies in food chains in general and genetic modification in livestock production chains in particular. The chapter concludes that in previous research an integrated chain approach is lacking. The use of genetic modification, however, should particularly be studied from an integrated point of view, since not only farmers need to adopt the technology, but also the other members of the chain, including consumers. Thus, following this approach, consumers and chain participants and their concerns and benefits should be analysed in an integrated way. The subsequent analysis in the thesis is organised according to this approach.

3. Consumer acceptance of GM applications in the pork production chain Chapter 3 investigates how consumers value specific GM applications* (GM animal, GM feed, GM additives & medicines and GM bacteria) that can be used in livestock production chains and whether and how consumers trade them off against the potential benefits (price reduction, improvements in quality, environment, animal welfare and reduced residues in meat) these applications can offer them. For this, an attribute-based stated choice modelling approach (choice experiment) was used. Data were collected by means of a written questionnaire sent to 2600 randomly selected addresses. In total, 253 usable questionnaires were obtained for analysis. The sample (53.8% females and 46.2% males) was representative of the Dutch population with respect to gender. The sample was not fully representative with

* GM animal is defined as an application in which genes of pigs themselves are genetically modified, so that future generations of those pigs are different. GM feed is defined as an application in which pig feed includes crops produced with the help of genetic modification. GM additives & medicines is defined as an application in which additives (e.g. vitamins, or bacteria for digestion) and medicines (e.g. vaccines, or antibiotics) are produced with the help of genetic modification. GM bacteria is defined as an application in which bacteria produced with the help of genetic modification are used for the preservation of pork after slaughter.

131 Summary respect to age, household size, number of children in household or education level. In our sample, respondents had on average a smaller household, fewer children in the household, and a much higher education level. The data were analysed by using a nested logit model. Results indicate that conventional pork is preferred over GM pork. Among the four GM applications investigated, GM feed is preferred most. Furthermore, findings reveal that the attractiveness of different benefits offered does not influence consumer choices towards choosing GM pork over regular pork. It is only when consumers have decided to choose GM pork that their further decisions regarding which GM pork to choose depend on the benefits offered. Socio- economic and attitudinal characteristics of the consumers were found to be important determinants of consumers’ valuation of specific GM applications and their possible benefits.

4. Heterogeneity in consumer acceptance of GM applications in pork production An examination of existing literature suggests that there are different consumer segments with different attitudes. Therefore, Chapter 4 deals with heterogeneity of consumers’ preferences regarding different GM applications used in the production of pork. Specifically, this chapter identifies distinct consumer segments and further profiles consumers in the segment using attitudinal and socio-economic characteristics. The heterogeneity in consumer acceptance of GM pork is captured by assuming that there are several consumer segments, in each of which the type of GM application and the consumer benefits exert a different influence on consumer decision making. The analysis is based on the data obtained in Chapter 3 from the choice experiment. The consumer segments were obtained using finite-mixture regression analysis. Segments were compared in terms of the attitudinal and socio-economic differences using cross-tabulations. Also, differences between segments were tested by using analysis of variance (ANOVA). Factor analysis using principal components was performed to assess consumers’ attitudes towards consumption and production of GM pork by using different applications. Five distinct consumer segments were obtained. About 49% of the respondents were found to be rejecters of GM technology, including 37.2% of “Anti-GM” consumers and 12.3% of those who occasionally would try GM pork, but still preferred conventional or any other types of pork (“Occasional eaters”). A further 29.3%, “GM proponents”, were found to accept GM pork from all GM applications with benefits. The “Benefit seekers” (9.1%) based their buying decision on the benefits offered, while “Application sensitive” (12.3%) accepted GM pork with benefits depending on the type of GM applications used. Greater insight into consumer acceptance of GM pork was provided by profiling consumers regarding attitudes and socio-economic characteristics. For example, consumers in segments “Applications

132 Summary sensitive” and “GM proponents” have a more neutral attitude towards GM technology than consumers in other segments. Also, they buy meat of no specific brand. “GM proponents” perceive fewer concerns and more benefits associated with GM technology in general, whereas “Application sensitive” consumers associate fewer concerns and more benefits only with respect to GM animal and GM feed applications. It appears that labelling is less important to the consumers in these segments. The segments “Occasional eaters” and “Anti- GM” are somewhat similar in their attitudes towards GM technology, with a difference that “Anti-GM” consumers are more critical, perceiving fewer benefits and more concerns. As to the consumers in the “Benefit seekers” segment, it seems likely that consumers in this segment prefer meat with guaranteed quality, as they hardly buy non-brand meat. Also, this segment is significantly different compared to other segments in terms of socio-economic characteristics and it is represented by highly educated, young consumers with children.

5. Economic performance of GM applications in the pork production chain Chapter 5 quantifies technical and economic performance of GM applications in the pork production chain. Per GM application, three concrete sub-applications were chosen for the analysis. Sub-applications are grouped according to the party they are expected to be beneficial for (producer, society, and consumer). These sub-applications aim at different benefits important to producer, society and consumers respectively. Regarding sub- applications relevant to producers, these were “Bovine gene”, “GM crops” (GM soy and maize) and “Antibodies”. Sub-applications relevant to “society” were “Enviropig”, “Low phytate (LP) plants” (GM soy and maize) and “Microbial phytase”. Due to the limited available information on the sub-applications relevant to the consumers, only “Immunocastration” was analysed in this study. An economic farm model with farrowing and fattening stages was used to analyse economic performance of GM applications in the pork production chain. Input values concerned both technical and economic parameters. Model outputs present the cost price of a 25 kg piglet in the farrowing stage and the cost price of 1 kg of pork produced in the fattening stage using several GM applications. A panel of experts was used to collect and verify necessary technical and economic data and assumptions on the GM applications selected. Experts were selected depending on the area of their expertise. A sensitivity analysis was carried out to show the effects of various input values across the GM applications on the output result. The results showed that although GM applications improved output results in both stages, the GM applications have a limited economic impact. Overall, pork producers can expect reductions in their cost prices of 0.3% to 3.7% in the farrowing

133 Summary stage and of 0.2% to 2.4% in the fattening stage by using GM applications. Of all the GM alternatives “Antibodies” showed the largest potential effect (3.7% cost price reduction) on the output results in the farrowing stage, whereas “LP plants” had the lowest effect (0.3% cost price reduction). In the fattening stage, “Enviropig” had the largest impact on the cost price of 1 kg of pork (2.4% reduction) and “Immunocastration” the lowest effect, i.e. 0.2% cost price reduction. Generally, cost savings and improvements in the output results were larger in the farrowing stage than in the fattening stage. In many GM cases, changes within one stage did not affect much the performance in another stage. The sensitivity analysis demonstrated that the changes in the main input parameters had little effect on the economic output results in both stages. The results obtained suggest that the current possible range of improvements in economic results due to the GM applications was rather limited and most likely would not be able to motivate farmers to introduce GM into their production system.

6. Main conclusions The following main conclusions can be drawn from this study: • Analysis of new technology adoption is not an easy task, especially in food chains, where consumers are generally very sensitive to change and evaluate new developments with special attention. Genetic modification is no exception. Not much research has been done yet on the adoption of genetic modification technology in livestock chains. The use of genetic modification, however, should particularly be studied from an integrated point of view, since not only farmers need to adopt the technology, but also the other members of the chain, including consumers. Uncertainty, for instance with respect to consumer acceptance and technology performance, is one of the major factors influencing the adoption of genetic modification (Chapter 2).

• Consumer decisions with respect to the GM pork are mainly affected by the general attitude to the technology. If the attitude towards GM is positive, then acceptance of GM applications is more likely. Importantly, this implies that all the efforts of the producers and other chain participants who want to introduce GM should be directed to improving existing attitudes, and creating a positive image of the technology by providing additional information to the consumers. Genetic modification in livestock production has potential, particularly in the case of GM

134 Summary

feed. The presence of important consumer benefits, such as the improvements in animal welfare, had a positive influence on consumers’ choices when choosing from different GM pork alternatives. Socio-economic and attitudinal characteristics of the consumers were found to be important determinants of consumers’ valuation of specific GM applications and their possible benefits (Chapter 3).

• Consumer preferences towards GM foods are heterogeneous. Five possible segments of consumers were identified. Segments are distinct with respect to consumer preferences towards GM applications and benefits these applications can offer to the consumers. While the majority of the consumers expressed their overall negative attitude towards GM technology and GM foods, consumers in some segments would be willing to purchase GM pork depending on the type of application used in the production and benefits offered. Greater insight into consumer acceptance of GM pork was provided by profiling consumers regarding attitudes and socio-economic characteristics (Chapter 4).

• Although the GM applications studied improved output results in the farrowing and fattening stages, economically the GM applications do not have much impact. Overall, pork producers can expect reductions in their cost prices of 0.3% to 3.7% in the farrowing stage and of 0.2% to 2.4% in the fattening stage by using GM applications. Of all the GM alternatives “Antibodies” showed the largest potential effect (3.7% cost price reduction) on the output results in the farrowing stage, whereas “LP plants” had the lowest (0.3% cost price reduction). In the fattening stage, “Enviropig” had the largest effect (of 2.4% reduction) on the cost price of 1 kg of pork and “Immunocastration” had the lowest (0.2% cost price reduction). Generally, cost savings and improvements in the output results were larger in the farrowing stage than in the fattening stage. In many GM cases, changes within one stage did not affect much the performance in another stage. The sensitivity analysis showed that changes in the main input parameters had little effect on the economic output results in both stages. The economic results obtained suggest that the current possible range of improvements in output results due to the GM applications is rather limited and most likely will not be able to motivate farmers to introduce GM in their production (Chapter 5).

135

Samenvatting

Samenvatting

1. Introductie De afgelopen decennia is de agri-food sector drastisch veranderd, bijvoorbeeld op het gebied van nieuwe exportmarkten, de ontwikkeling van duurzame landbouwsystemen, de verbetering van de voedselveiligheid en de ontwikkeling en het gebruik van nieuwe technologieën. Nieuwe technologieën zullen er ook in de toekomst toe blijven leiden dat de agrarische sector nog verder zal veranderen. Technologieën met betrekking tot de voedselproductie en daarmee gerelateerde zaken als voedselveiligheid, milieu, en dierenwelzijn zijn onder andere doorstraling, nutrigenomics en genetische modificatie. Andere technologieën richten zich meer op het verlagen van productiekosten en de uiteindelijke prijs in de supermarkt. De adoptie van een nieuwe technologie is complex en kan vanuit verschillende perspectieven bestudeerd worden. Over het algemeen wordt aangenomen dat een groot aantal verschillende factoren een rol speelt bij de adoptie van nieuwe en opkomende technologieën. Vaak wordt het adoptieproces alleen bestudeerd vanuit de individuele agrarische ondernemer, of vanuit de consument, en alleen gericht op een specifieke toepassing of wens van de consument; onderzoek met betrekking tot de keten als geheel ontbreekt. Juist dergelijke ketenstudies zijn van belang omdat voor een aantal technologieën bijvoorbeeld de kosten worden gemaakt door de ene ketenpartij terwijl de revenuen terecht komen bij een andere partij. Deze promotiestudie analyseert de adoptie van nieuwe technologieën vanuit een geïntegreerd ketenperspectief. Het doel van het onderzoek is het verkrijgen van inzicht in de adoptie van genetische modificatie in varkensketens. Dit doel is als volgt opgesplitst: 1. Welke aspecten zijn van belang bij het adoptieproces van genetische modificatie in de varkensketen? 2. Welke waarde geven consumenten aan specifieke toepassingen van genetische modificatie en welke afwegingen maakt men tussen toepassing en gerelateerde voordelen? 3. Hoe heterogeen zijn consumenten ten aanzien van verschillende toepassingen van genetische modificatie en in welke mate kan dit verklaard worden door sociaal- economische variabelen? 4. Wat is de te verwachten impact op boerderijniveau van de toepassing van genetische modificatie in de varkensketen?

138 Samenvatting

2. Adoptie van genetische modificatie in veehouderijketens: een geïntegreerde aanpak In hoofdstuk 2 zijn de belangrijkste termen van dit onderzoek nader gespecificeerd, i.e. genetische modificatie en genetisch gemodificeerde organismen. Genetische modificatie (GM) heeft betrekking op verschillende manieren om genetisch materiaal te manipuleren. Hoewel ieder organisme een “genetisch gemodificeerd organisme” is, betreft het bij GMO een zodanige nieuwe combinatie van genetische informatie dat er een nieuw dier, plant of micro- organisme ontstaat. Hoofdstuk 2 gaat in op de belangrijkste ontwikkelingen rond genetische modificatie in de veehouderijketen. Daarnaast wordt een raamwerk gepresenteerd voor de analyse van de adoptie van nieuwe technologieën in de keten. Hierin worden consumenten en (andere) ketenpartners geïntegreerd, alsook hun wensen en te behalen voordelen. Aan de hand van dit raamwerk is de literatuur rond technologie-adoptie en genetische modificatie in ketens besproken. Hieruit wordt geconcludeerd dat een geïntegreerde ketenbenadering momenteel ontbreekt. Een dergelijke ketenbenadering is echter met name voor zaken als genetische modificatie van belang omdat het niet alleen gaat om de adoptie van de technologie door de veehouder maar ook door die van de consument. Hierop voortbouwend wordt aanbevolen om de consumenten- en producenten wensen en voordelen van nieuwe technologieën te bestuderen volgens een geïntegreerde methode. De rest van het onderzoek is dan ook volgens deze wijze opgezet.

3. Consumentenacceptatie van toepassingen van GM in de keten van varkensvlees Hoofdstuk 3 analyseert hoe consumenten specifieke toepassingen van GM (GM-dier, GM- voer, GM-additieven en -medicijnen, GM-bacteriën) waarderen en of ze bepaalde afwegingen maken tussen toepassing en voordeel, i.e. een lagere prijs, betere kwaliteit, voordelig voor milieu, goed voor dierenwelzijns en minder residuen. Een en ander is geanalyseerd met behulp van een “choice experiment”, een zogenaamd attribute-based stated choice modelling approach. Data zijn verzameld met behulp van een schriftelijke vragenlijst die is toegestuurd aan 2600 willekeurig geselecteerde adressen. Van de teruggestuurde vragenlijsten waren er 253 bruikbaar voor verdere analyses. De steekproef (53.8 vrouw, 46.2 man) was alleen representatief ten aanzien van geslacht, niet ten aanzien van leeftijd, gezinsgrootte, aantal kinderen en opleidingsniveau. In de steekproef was het gemiddelde gezin kleiner, waren er minder kinderen en was het opleidingsniveau hoger. Analyses zijn uitgevoerd met behulp van een “nested logit model”. Resultaten laten zien dat consumenten over het algemeen de voorkeur geven aan traditioneel varkensvlees boven genetisch gemodificeerd varkensvlees. Van de verschillende GM-toepassingen gaat de grootste voorkeur uit naar GM-voer. De

139 Samenvatting resultaten geven ook aan dat de gepresenteerde voordelen (lagere prijs etc.) de keuze van consumenten niet beïnvloeden. Alleen als consumenten al kiezen voor GM-varkensvlees, dan laten ze verdere keuzes wel van de voordelen afhangen.

4. Heterogeniteit in de consumentenacceptatie van GM in de varkensketen In de literatuur wordt er over het algemeen vanuit gegaan dat consumenten kunnen worden ingedeeld in verschillende segmenten. Hoofdstuk 4 identificeert consumentensegmenten rond de productie van varkensvlees waarbij verschillende methoden van genetische modificatie worden toegepast. Segmentatie vindt plaats op basis van aspecten van houding ten aanzien van genetische modificatie en een aantal sociaal-economische kenmerken van consumenten. De analyse bouwt verder op de data van hoofdstuk 3 en maakt gebruik van de technieken “finite mixture regression analysis”, vergelijkingen in kruistabellen en variantie-analyse (ANOVA). Voor wat betreft een analyse van de houding van consumenten ten aanzien van consumptie en productie van GM-varkensvlees is gebruik gemaakt van factor-analyse. In totaal zijn 5 segmenten onderscheiden. 49% van de consumenten is tegen GM, waarvan 37.2% echt “Anti-GM” is en 12.3% “het wel eens zou willen proberen”, maar met een voorkeur voor traditioneel varkensvlees. 29.3% daarentegen is “voor GM” zolang er voordelen mee verbonden zijn. Een verdere 9.1%, de “voordeelzoekers”, baseert hun keuze voor GM op basis van het geboden voordeel. 12.3% stelt zich “afhankelijk van de specifieke toepassing”. Segmenten zijn beschreven op basis van hun kenmerken rond houding ten aanzien van GM en sociaal-economische aspecten. Zo hebben consumenten in de segmenten “afhankelijk van toepassing” en “voor GM” een meer neutrale houding ten aanzien van genetische modificatie dan consumenten in andere segmenten. Bovendien kopen zij meestal varkensvlees zonder specifiek merk. De “voor GM” groep heeft ook minder zorgen en ziet meer voordelen rond GM. Voor de groep “afhankelijk van toepassing” geldt dit alleen voor de toepassingen van GM-dier en GM-voer. De twee segmenten die tegen GM zijn zitten hier voor wat betreft hun houding dichtbij, met dien verstande dat “anti-GM” het meest kritisch is en minder voordelen en meer nadelen ervaart. Het segment “voordeelzoekers” prefereert vlees van gegarandeerde kwaliteit, aangezien zij met name merkvlees kopen. Dit segment onderscheidt zich ook van de anderen voor wat betreft sociaal-economische aspecten: het kenmerkt zich door hoog-opgeleide en jonge consumenten met kinderen.

140 Samenvatting

5. Economische consequenties van GM in de varkensketen Hoofdstuk 5 kwantificeert de technische en economische gevolgen van GM in de varkensketen. Hiertoe zijn per GM-toepassing (GM-dier, GM-voer, GM-additieven en - medicijnen, GM-bacteriën) drie concrete applicaties gedefinieerd. Deze zijn vervolgens weer gegroepeerd naar de partij die er waarschijnlijk een voordeel mee kan behalen. Dit zijn de consument, de producent en de maatschappij. Voor de producent zijn de gekozen GM- applicaties: “bovine gene”, “GM crops” (GM soja en mais), en “antibodies”. Voor de maatschappij zijn dit “enviropig”, “low phytate (LP) plants” (GM soja en mais) en “microbial phytase”. Voor wat betreft de consument is vanwege de beperkte hoeveelheid informatie alleen gekeken naar “immunocastratie”. Analyses zijn uitgevoerd met behulp van een economisch bedrijfsmodel voor de zeugen- en vleesvarkenshouderij. Inputparameters zijn zowel technisch als economisch van aard. De output van het model bestaat uit de kostprijs van een 25-kg big en de kostprijs per kg varkensvlees in de vleesvarkenshouderij. Input rond de GM-applicaties is geschat met medewerking van een expertpanel. Resultaten geven aan dat de GM-applicaties in zowel de zeugenhouderij als de vleesvarkenshouderij een positieve impact hebben, maar dat het totale economische effect gering is. In de zeugenhouderij gaat door het toepassen van GM de kostprijs met zo’n 0.3% tot 3.7% naar beneden. In de vleesvarkenshouderij is dit 0.2% tot 2.4%. Van de geanalyseerde applicaties hebben in de zeugenhouderij “antibodies” het grootste effect (kostprijsverlaging van 3.7%) en “LP plants” het minste (0.3% lagere kostprijs). In de vleesvarkenshouderij heeft “enviropig” het meeste effect (kostprijs min 2.4%) en “immunocastratie” het minste (0.2% kostprijsverlaging). Over het algemeen zijn de technische en economische effecten in de zeugenhouderij groter dan in de vleesvarkenshouderij. Ook blijken de gevolgen van GM-applicaties in het ene deel van de keten weinig effect te hebben op de prestaties in het andere deel van de keten. Gevoeligheidsanalyses voor wat betreft de belangrijkste aannames rond de input veranderen de conclusies nauwelijks. Bovenstaande resultaten laten zien dat de verbeteringen door GM- applicaties zoals die op dit moment worden ingeschat relatief beperkt zijn en mogelijk te gering om veehouders daadwerkelijk over te halen dergelijke technieken te gebruiken.

141 Samenvatting

6. Belangrijkste conclusies De belangrijkste conclusies van dit onderzoek zijn: • De analyse van de adoptie van nieuwe technologieën is niet eenvoudig, met name niet in voedselproductieketens waarbij consumenten zeer gevoelig zijn voor veranderingen. Dit geldt ook voor genetische modificatie. Voor wat betreft de adoptie van GM in veehouderijketens is nog relatief weinig onderzoek gedaan. Juist voor dergelijke zaken zijn analyses vanuit een geïntegreerd ketenperspectief van belang omdat de techniek niet alleen geadopteerd moet worden door de veehouder maar ook door de andere ketendeelnemers, inclusief de consument. Onzekerheid, onder ander met betrekking tot consumentenacceptatie en technologieprestatie, is een van de belangrijkste factoren die de adoptie van GM bepalen (Hoofdstuk 2).

• Beslissingen van consumenten rond GM-varkensvlees worden in belangrijke mate beïnvloed door hun algemene houding ten aanzien van genetische modificatie. Als de houding ten aanzien van GM positief is, is het waarschijnlijker dat GM-aplicaties zullen worden geaccepteerd. Hiertoe is het van groot belang dat producenten die GM- applicaties willen introduceren er voor zorgdragen dat consumenten een zo positief mogelijk imago van de technologie ontwikkelen en goed geïnformeerd zijn. Van de verschillende toepassingen van genetische modificatie in de veehouderij lijkt met name GM-voer kansrijk. Verder bleek ook dat de aanwezigheid van specifieke voordelen, bijvoorbeeld een lagere prijs of een betere voedselveiligheid, de acceptatie van GM-toepassingen gunstig kan beïnvloeden. Ook sociaal-economische kenmerken en de houding van consumenten ten aanzien van de GM-technologie bleken belangrijke factoren bij de keuze rond GM (Hoofstuk 3).

• Consumentenvoorkeuren rond GM-voedsel zijn niet homogeen. In dit onderzoek rond genetische modificatie in de varkensketen zijn vijf segmenten onderscheiden. De segmenten verschillen onderling in hun voorkeur ten aanzien van GM-applicaties en de waardering van gerelateerde voordelen. Hoewel het merendeel van de consumenten tegen GM-technologie en GM-voedsel is, is er ook een groep die GM-varkensvlees wil kopen, afhankelijk van specifieke applicaties in de keten en gerelateerde voordelen. Segmenten zijn beschreven aan de hand van een aantal sociaal- economische kenmerken en de houding van consumenten ten aanzien van GM- technologie.

142 Samenvatting

• Hoewel de technische en economische resultaten in de zeugen- en de vleesvarkenshouderij door de toepassing van GM verbeteren, is het totale economische effect gering. In de zeugenhouderij daalt de kostprijs door het toepassen van GM met zo’n 0.3% tot 3.7%. In de vleesvarkenshouderij is dit 0.2% tot 2.4%. Van alle applicaties hebben in de zeugenhouderij “antibodies” het grootste effect (kostprijsverlaging van 3.7%) en “LP plants” het minste (0.3% lagere kostprijs). In de vleesvarkenshouderij heeft “enviropig” het meeste effect (kostprijs min 2.4%) en “immunocastratie” het minste (0.2% kostprijsverlaging). Over het algemeen zijn de technische en economische effecten in de zeugenhouderij groter dan in de vleesvarkenshouderij. Gevoeligheidsanalyses veranderen de conclusies nauwelijks. Bovenstaande resultaten laten zien dat de te behalen verbeteringen door GM- applicaties zoals die op dit moment worden geschat relatief beperkt zijn, en mogelijk te gering om veehouders daadwerkelijk over te halen dergelijke technieken te gebruiken (Hoofdstuk 5).

143

List of Publications

Refereed Scientific Papers Novoselova T.A., Meuwissen M.P.M., Huirne R.B.M (2007). Adoption of GM technology in livestock production chains: an integrating framework. Trends in Food Science and Technology 18: 175-188.

Novoselova, T., van der Lans, I.A.C.M., Meuwissen, M.P.M., Huirne, R.B.M. Analysis of consumer acceptance of genetically modified applications in the pork production chain. Submitted to British Food Journal.

Novoselova, T., van der Lans, I.A.C.M., Meuwissen, M.P.M., Huirne, R.B.M. Heterogeneity in consumer acceptance of GM applications: attitudinal and socio-economic determinants Submitted to Food Quality and Preference.

Novoselova, T., Meuwissen, M.P.M., Jongbloed, A.W., Huirne, R.B.M. A feasibility study of GM pork production in the Netherlands. Submitted to Livestock Science.

Conference Papers Novoselova, T., Meuwissen, M.P.M., Huirne, R.B.M. (2006). An integrated consumer and supply chain risk analysis of genetic modification in livestock production chain. Presented at American Agricultural Economic Association (AAEA) Annual Meeting in Long Beach, July 23-26; on-line at http://agecon.lib.umn.edu/cgi-bin/view.pl

Novoselova, T., van der Lans, I.A.C.M., Meuwissen, M.P.M., Huirne, R.B.M. (2005). Consumer acceptance of GM applications in pork production chain: a choice modeling approach. In: Proceedings of the 11th congress of the European Association of Agricultural Economists (EAAE), Copenhagen, Denmark, 24-27 August, on-line at http://agecon.lib.umn.edu/cgi-bin/view.pl

Novoselova, T., van der Lans, I.A., Meuwissen, M.P.M.; Huirne, R.B.M. (2005). An application of choice modeling to measure consumer preferences for GM applications in pork production chain. In: European/EAAE PhD Workshop, Mansholt Graduate School of Social Sciences Wageningen, 22-23 September.

Meuwissen, M.P.M.; Novoselova, T.; van der Lans, I.A. (2004). Ethical concerns about pork production: a comparison of elicitation techniques In: Science, Ethics and Society, 5th Congress of the European Society for Agricultural and Food ethics (EURSAFE), Leuven, 2-4 September, pp.284-289.

Novoselova, T.; Meuwissen, M.P.M.; Huirne, R.B.M. (2004). Consumer ethical concerns relating to biotechnology in livestock production In: Science, Ethics and Society, 5th Congress of the European Society for Agricultural and Food ethics (EURSAFE), Leuven, 2-4 September, p.346.

145 Novoselova, T.; Meuwissen, M.P.M.; Huirne, R.B.M. (2004). New technology adoption in food chains: a review with special reference to GMO applications in livestock production In: Proceedings of the 6th International Conference on chain and network management in agribusiness and the food industry, Ede, 27-28 May: Wageningen Academic Publishers, 2004 - ISBN 907699840X - pp.33-39.

Meuwissen, M.P.M., Novoselova, T., van der Lans, I.A.C.M. (2003) Opportunism in the pork market In: Proceedings of the 83rd EAAE Seminar, Chania, Greece, September 4-7.

Novoselova, T.A., Meuwissen, M.P.M., Huirne, R.B.M. (2003). GM opportunities in pork production chains, In: 7th International Conference on: Public Goods and Public Policy for agricultural Biotechnology (ICABR), June 29-July 3 2003, Ravello, Italy.

Novoselova, T., Meuwissen, M.P.M., van der Lans, I.A., Valeeva, N. (2002). Consumers' perception of milk safety In: 13th International IFMA Congress of Farm Management: Book of Abstracts Papers and Posters, Arnhem, 7-12 July, 2002 / R.B.M. Huirne, J.A. Renkema and J. Niejenhuis

146 Curriculum vitae

Tatiana Alexandrovna Novoselova was born on July 26, 1979 in Losino-Petrovskiy town, Russia. In 1996 she finished high school No1 of Losino-Petrovskiy, with distinction. In the same year she entered Moscow Timiryazev Agricultural Academy, where she enrolled in the programme of Agricultural Economics at the Faculty of Economics. In June 2000 she obtained her BSc diploma with distinction. In the same year she was granted a TEMPUS scholarship covering a full tuition fee for an MSc programme at Wageningen University. In 2002 she received her MSc diploma in Agricultural Economics and Management, with specialisation Marketing and Consumer Behaviour. In April 2002 she was appointed as a PhD researcher (AIO) at the Institute for Risk Management in Agriculture (IRMA) and Business Economics Group of Wageningen University. Also during her PhD period she was a member of Mansholt Graduate School Council and a member of PromoV, representing temporary staff members of the university at Student - Staff Council and Wageningen PhD Council.

147

Completed Training and Supervision Plan

Description Institute / Department Year Credits* Courses Writing and Presenting a Scientific Mansholt Graduate School of 2002 1 Paper Social Sciences (MG3S) Written English Language Centre, 2003 1.5 Wageningen University Scientific Writing Language Centre, 2003 1.5 Wageningen University Career Perspectives Wageningen Graduate 2005 1.3 Schools Mansholt Introduction Course MG3S 2002 1 Critical Reflection on MG3S 2002 1 Science/Technology, Values and Sustainability Behavioural Economics MG3S 2003 3 Advanced Consumer Behaviour Wageningen University 2003 4 Supply Chain Economics MG3S 2003 2 Food Safety Risk Analysis MG3S 2003 2 Food Perception and Food VLAG Graduate School, 2003 0.7 Preference Wageningen University PhD Discussion Groups Business Economics, 2002/ 4 Wageningen University 2006 Mansholt Multidisciplinary Seminar MG3S 2005 1 (PhD Workshop)

Presentations at international conferences 2 IFMA Congress, Arnhem, The Netherlands 2002 EAAE Seminar Chania, Greece 2003 International Conference on Chain and Network Management in 2004 Agribusiness and the Food Industry, Ede, The Netherlands EAAE Congress, Copenhagen, Denmark 2005 AAEA Annual Meeting, Long Beach, USA 2006

Total (minimum 20 Credits) 26.0 *One credit represents (on average) 40 hours of work IFMA stands for International Farm Management Association EAAE stands for European Association of Agricultural Economists AAEA stands for American Agricultural Economics Association

149

Printing establishment Ponsen & Looijen Printers BV, Wageningen

Cover Igor Bezlepkin and Evgeny Novoselov

Financing The research described in this thesis was funded by Mansholt Graduate School and Institute for Risk Management in Agriculture (IRMA).

The attendance of conferences was supported by the LEB fund, Wageningen University.

150