Attitudes of European farmers towards GM crop adoption

Transcription

1 Plant Biotechnology Journal (2011) 9, pp doi: /j x Attitudes of European farmers towards GM crop adoption Francisco J. Areal 1,2, *, Laura Riesgo 1,3 and Emilio Rodríguez-Cerezo 1 1 European Commission, Joint Research Centre (JRC), Institute for Prospective Technological Studies (IPTS), Edificio Expo, Seville, Spain 2 Currently at School of Agriculture, Policy and Development, University of Reading, Reading, UK 3 Currently at Department of Economics, Pablo de Olavide University, Seville, Spain Received 22 March 2011; revised 21 July 2011; accepted 21 July *Correspondence (Tel ; fax ; f.j.areal@reading.ac.uk) Keywords: genetically modified crops, herbicide tolerant, farmers attitudes, cluster analysis, coexistence. Summary This article analyses European Union (EU) farmers attitudes towards adoption of genetically modified crops by identifying and classifying groups of farmers. Cluster analysis provided two groups of farmers allowing us to classify farmers into potential or of genetically modified herbicide-tolerant (GMHT) crops. Results showed that economic issues such as the guarantee of a higher income and the reduction of weed control costs are the most encouraging reasons for potential and of GMHT crops. This article also examines how putting in place measures to ensure coexistence between GM and non-gm crops may influence farmers attitudes towards GMHT crop adoption. Results show that the implementation of a coexistence policy would have a negative impact on farmers attitudes on adoption and consequently may hamper GMHT adoption in the EU. Introduction Since their introduction in 1996, genetically modified (GM) crops have motivated a controversial debate (Pollack and Shaffer, 2001; Bernauer, 2003; Vogel, 2003; Young, 2003). A number of aspects, comprising complexity of the technology, uncertainty about long-term effects and socioeconomic issues, have stimulated the opposing views in the debate (Rey-Garcia, 2006). Scientific research on diverse points related to genetically modified organisms (GMOs), such as their agronomic and economic performance, human health and environmental risks associated with production and stakeholders attitudes on genetically modified (GM) food crops, has also contributed to the debate while supporting policy decision making. Research on attitudes towards GMOs has focused mainly on consumer s preferences overlooking other stakeholders attitudes (Bett et al., 2010). Literature on consumer s preferences on GM food shows that consumers usually value non-gm food higher than GM food (Baker and Burnham, 2001; Burton et al., 2001; Noussair et al., 2002; Lusk et al., 2004, 2005; Hall et al., 2006; Dannenberg, 2009; Zhang et al., 2010). Different reasons have been highlighted to justify this scepticism towards GM food, such as consumers uncertainty about future consequences of consuming this type of food (Lusk et al., 2004; Costa-Font et al., 2008; Dannenberg, 2009; Martinez-Poveda et al., 2009). However, consumers valuation is not homogeneous across countries and regions. Thus, studies have showed that Europeans are the most sceptical with regard to GM food, whereas consumers in Asia seem to be somewhat polarized in their attitudes and Americans are relatively supportive of GM food (Gaskell et al., 1999; Alexander and Schelman, 2003; Lusk, 2003; Ganiere et al., 2006; Dannenberg, 2009). Despite the relevance of knowing consumers and farmers attitudes towards GM crops for policy-making processes (e.g. information on the diffusion of GM technology may help in designing a European policy on coexistence between GM and non-gm crops), relatively few studies have investigated farmers attitudes towards adoption of GM crops (Baker and Burnham, 2001; Ganiere et al., 2006; Hall, 2008; Gaskell et al., 2010). On the production side, attention has been paid to identifying determinants of GM crop adoption (Marra et al., 2001; Hategekimana and Trant, 2002; Qaim and Traxler, 2005; Breustedt et al., 2008; Dill et al., 2008; Finger et al., 2009; Consmüller et al., 2010). In this regard, economic, social and or environmental issues have been identified as drivers of farmers willingness to grow GM crops. Indeed, net income gains for GM derived from the lower costs associated with lower expenditures on pesticides, labour, machinery and fuel have been reported (Bullock and Nitsi, 2001; Fernández-Cornejo et al., 2002; Phillips, 2003; Bernard et al., 2004; Qaim, 2009). In addition to profitability, nonpecuniary benefits were also reported for GMHT crops (Carpenter and Gianessi, 1999; Bullock and Nitsi, 2001; Marra and Piggott, 2006; Ervin et al., 2010). GMHT technology allows for greater flexibility and simplicity in crop management as farmers reduce the time spent in scouting and spraying fields to address weed problems. This time savings may increase farmers income by on-farm and or off-farm activities (Fernández-Cornejo et al., 2005; Keelan et al., 2009). Some authors have also reported environmental benefits associated with the adoption of GMHT crops, such as the substitution of selective herbicides (usually harmful for the environment) for less toxic broad-spectrum herbicides (e.g. glyphosate), and savings associated with low herbicide use and conservation tillage practices (Ervin et al., 2000; Wolfenbarger and Phifer, 2000; Deward et al., 2003; Nelson and Bullock, 2003; Sydorovych and Marra, 2007; Devos et al., 2008; Smyth et al., 2011). It is worth mentioning that the decrease in the total quantity of herbicides applied per unit surface area occurs only at early stages of GMHT crop adoption (Shaner, 2000; Owen Plant Biotechnology Journal ª 2011 Society for Experimental Biology, Association of Applied Biologists and Blackwell Publishing Ltd 945

2 946 Francisco J. Areal et al. and Zelaya, 2005; Bonny, 2008), but a rise in the quantity of herbicides is expected in late stages of adoption because of the presence of resistant weeds by intensive herbicide applications. Impacts on biodiversity or the appearance of HT volunteers are other environmental risks associated with weed management changes related to GMHT crops, already a major issue in the US (Shaner, 2000; Hayes et al., 2004; Owen and Zelaya, 2005; Graef et al., 2007; Bonny, 2008; Benbrook, 2009). The aim of this study was to illustrate the attitudes of European farmers towards GM crop adoption by examining farmers views, expressed through a survey to EU maize and oilseed rape (OSR) growers, on both adopting GM crops and the effect of coexistence measures on their views on adopting GM crops. Insight into the effects that putting in place coexistence measures to ensure coexistence between GM and non-gm crops may have on potential attitudes towards adopting GMHT technology is provided through results obtained from a set of related questions on coexistence measures. The study also provides estimates of the average adoption rates of GMHT OSR and GMHT maize per country studied and shapes farmers profiles with respect to GMHT crop adoption. To our knowledge, no existing research has addressed how regulation to ensure coexistence between GM and non-gm crops may influence farmers attitudes towards GM crop adoption. In this regard, EU legislation aiming at ensuring coexistence between GM and non-gm crops offers the basis to evaluate how farmers views towards adoption of a new technology may be affected by a set of measures designed to ensure such coexistence. Coexistence in the EU refers to the ability of European farmers to choose amongst conventional, organic and GM crop production in compliance with the legal obligations for labelling and or purity standards (European Council, 2006). The EU legislation establishes a threshold of 0.9%, below which the marketed products containing GMOs authorized to be used do not required to be labelled as GM product. The European Commission recommendation of 13 July 2010 provides guidelines for the development of national coexistence measures to avoid the unintended presence of GMOs in conventional and organic crops. This recommendation provides freedom for Member States to decide what measures to avoid the unintended presence of GMOs in conventional and organic crops should be established. A number of potential measures to avoid adventitious GM presence can be distinguished: separation distances, notification to public authorities and third parties (i.e. neighbours) of farmers intention to grow a GM crop, and segregation measures required during transportation, handling (e.g. cleaning machinery) or storage processes can be adopted to guarantee coexistence (Czarnak-Klos and Rodríguez-Cerezo, 2010). No insurance for the economic risk related to GM admixture is currently available in the European market (European Commission, 2009). However, in four countries, insurance cover or alternative types of financial guarantee are legally required (Austria, France, Italy and Luxembourg). Currently, Bt 1 maize, since 1998, and Amflora potato, since 2010, are the 1 Bt refers to insect-resistant maize. This trait is based on genes from the Bacillus thuringiensis (Bt) soil bacterium. 2 The best practice document intends to assist Member States in the development or refinement of their coexistence legislation or voluntary standards for good agricultural practice. Therefore, none of the practices included in the document are mandatory for Member States. only GM crops authorized for cultivation in the EU. Regarding maize, a best practice document 2 for coexistence has been published in 2010 with measures to limit adventitious GM content in non-gm maize to below the legal labelling threshold (Czarnak-Klos and Rodríguez-Cerezo, 2010), such as separation distances, segregation measures during transportation, handling or storage processes. No document on best practices for the coexistence of Amflora potato with non-gm potato has been published yet. The new recommendation of the European Commission gives flexibility to Member States to avoid the unintended presence of GMOs; thus, different coexistence measures are expected to be established by each Member State. In this regard, we included a broad set of coexistence measures (i.e. identification in a public register, notification of growing GM crops to neighbours and or the land owner, insurance policy to cover damages caused by adventitious presence, separation distances and segregation measures during transportation, handling or storage) to analyse how each potential measure may influence farmers attitudes towards GM crop adoption regardless of the crop or country analysed. The article focuses on GM herbicide-tolerant 3 (GMHT) OSR and GMHT maize, because they are in the EU s regulatory pipeline. Whereas only Bt maize and Amflora potato are authorized for cultivation in the EU, several GM crops, which have not yet been approved for cultivation, have been approved for import into the EU (GMHT soybean, GMHT OSR and GMHT maize).the study identifies and classifies groups of farmers per country and crop on the basis of their attitudes towards GMHT crop adoption. Groups are profiled by studying each group s views on both encouraging and dissuading statements associated with GMHT crop adoption. In addition, the article analyses both the evaluation of coexistence measures by potential and the impact of such measures on potential views on GMHT crop adoption. The article is organized as follows. Second section introduces the material (data set). Third section includes the results and aims to determine the heterogeneity of farmers regarding GMHT crop adoption as well as farmers attitudes concerning the coexistence policy. The methodology used for this research is presented in fourth section. Finally, fifth section contains a discussion of the results and the conclusions drawn. Data set A survey was conducted to 647 farmers in 6 European countries: Czech Republic, Germany, UK, France, Hungary and Spain. These countries are representative of OSR and maize production in the EU-27. Thus, in 2007, together the Czech Republic, Germany and the UK accounted for 45% and 40% of the total OSR production and area in the EU-27, respectively, while France, Hungary and Spain accounted for 46% and 37% of the total maize production and area allocated to maize in the EU-27. A total of 439 farmers were interviewed face-to-face in the Czech Republic, Germany and the UK on their attitudes towards adopting GMHT OSR, whereas 208 farmers were interviewed, also face-to-face, in France, Hungary and Spain on their attitudes towards adopting GMHT maize during March and July Within all countries, prior stratification was undertaken 3 Genetically modified herbicide-tolerant (GMHT) crops are tolerant to certain broad-spectrum herbicides such as glyphosate or glufosinate.

3 Attitudes of EU farmers towards GM crop adoption 947 with respect to growing areas that were centres for significant OSR and maize cultivation. Thus, the regions selected for the Czech Republic were Strední Cechy, Jihozápad, Severovýchod and Jihovýchod; the German regions were Mecklenburg- Vorpommern, Brandenburg, Sachsen-Anhalt, Thüringen, Sachsen, Schleswig-Holstein, Nordrhein-Westfalen, Rheinland-Pfalz and Bayern; the UK counties were Norfolk, Suffolk, Cambridgeshire, Essex, Hertfordshire, Bedfordshire, Lincolnshire, Leicestershire, Rutland, Northamptonshire, Nottinghamshire and Derbyshire; Spanish provinces surveyed were Sevilla, Badajoz, Cáceres, Zaragoza and Lleida; Hungarian farmers from Baranya, Somogy, Tolna, Hajdú-Bihar, Jász-Nagykun-Szolnok and Szabolcs-Szatmár-Bereg regions were interviewed, while the French regions of Aquitaine, Midi Pyrénées, Poitou-Charentes, Alsace and Lorraine were targeted to conduct the survey in France. A set of 24 statements reflecting attitudes towards GM adoption for the two crops studied (GMHT OSR and GMHT maize) were posed to farmers to evaluate the relevance of several factors in farmer s decision process (Table 1). Amongst the 24 statements, the first 13 statements illustrate reasons that may encourage the use of GMHT crops, whereas the rest represent reasons that may dissuade farmers from cultivating GMHT crops. 4 Statements comprise economic, environmental, social and personal views of adopting the new technology. The statements were presented to the respondents in the order given in Table 1. Correlation results suggest no influence of the statements order in their evaluation (see Tables S1-S6 in Supporting Information I). Farmers were asked to evaluate these statements using a Likert scale from 1 to 5, where 1 meant completely unimportant in deciding to adopt the new technology (definitely not reject the idea of cultivating GM crops in the case of dissuading statements) and 5 meant very important (definitely reject the idea of cultivating GM crops in the case of dissuading statements); hence, all evaluations contribute equally to the distance or similarity between statements. In addition to encouraging and dissuading statements, a question regarding farmer s willingness to adopt GMHT OSR GMHT maize was posed to farmers. Respondents could choose from five alternatives: (i) It s very unlikely I would change to GMHT OSR (GMHT maize); (ii) It s somewhat unlikely I would change to GMHT OSR (GMHT maize); (iii) It s uncertain I would change to GMHT OSR (GMHT maize); (iv) It s likely I would change to GMHT OSR (GMHT maize); and (v) It s very likely I would change to GMHT OSR (GMHT maize). Results Descriptive statistics Table 2 shows the potential adoption rates per crop and country obtained from the question related to farmer s willingness to adopt GMHT OSR maize. Approximately half of farmers questioned stated that they were likely or very likely to adopt GMHT OSR, whereas over a third of farmers questioned in Spain, France and Hungary were likely or very likely to adopt GMHT maize. This shows that a significant part of European farmers would be keen on adopting GMHT crops if they were allowed. 4 Please note that the language used in the encouraging and dissuading statements differs. The term guarantee is used 4 times in encouraging statements, being this a strong term, whereas dissuading statements uses terms much softer, such as I do not think. Table 1 Statements reflecting attitudes towards GMHT adoption Number Attitude statement Encouraging adoption of GMHT 1 It guarantees the reduction in losses caused by weed growth 2 I practise direct sowing and it facilitates this process 3 It guarantees greater yields during harvest 4 It guarantees a higher income 5 It reduces weed control costs 6 My regular distributor recommended it 7 It guarantees a higher quality of harvest 8 The environmental impact on my farm is reduced because it involves a cut down in herbicides 9 It seems to involve less risk 10 It makes me feel as though I were at the forefront of biotechnological progress 11 All of the farmers in my area are using it 12 The technician s that I have consulted have recommended it 13 It facilitates my work being a technology that makes cultivation easier Dissuading adoption of GMHT 14 I do not have a huge weed problem in my land 15 The seeds are much more expensive than the conventional seeds 16 I do not think there would be an improvement in yield 17 I do not believe in these new products 18 I do not think there would be an improvement in financial returns 19 I prefer not to change the type of crop. I do not really like change 20 I have been advised not to use this type of maize rape 21 I think it would be difficult to market the grain 22 I have more faith in the use of herbicides to avoid weed infestation than in this type of crop 23 It is not well received by society in general 24 It is a complicated technology to use (one has to conform to coexistence or buffer zone regulations) Figures 1 and 2 show the average scores given by farmers to a number of encouraging and dissuading statements related to the possibility of adopting GMHT maize and GMHT OSR, respectively. Figures effectively show which aspects farmers perceive as relevant and unimportant when deciding whether or not to adopt a GMHT crop. To analyse the scores given by farmers, we consider that scores from 1 to 2.50 mean that the statement is an unimportant issue for farmers in their decision to adopt (positive statements, 1 13) or reject (negative statements, 14 24) GMHT crops; scores from 2.51 to 3.50 mean that farmers are undecided or do not know whether the statement is important or unimportant; and scores from 3.51 to 5 mean that farmers consider the statement as important in their decision. Following this interpretation, the higher the bar, the higher the importance of the statement to adopt (for statements from 1 to 13) or reject (for statements from 14 to 21) the adoption of GMHT crops. The figures show that in most cases, positive statements are more persuading to encourage the adoption of GMHT crops than negative statements to reject that adoption. Positive attitudes towards GMHT crop adoption Illustrations show that guaranteeing a higher income (statement 4) is considered on average to be the most important factor determining adoption in both crops and all countries

4 948 Francisco J. Areal et al. Table 2 Potential adoption rates (%) of GMHT OSR and GMHT maize crops per country Willingness to adopt GMHT maize Spain France Hungary Very unlikely Somewhat unlikely Uncertain Likely Very likely Willingness to adopt GMHT OSR Czech Republic Germany UK Very unlikely Somewhat unlikely Uncertain Likely Very likely Potential adoption rates show the percentage of respondents who answered one of the five alternatives (i.e. very unlikely, somewhat unlikely, uncertain, likely, very likely) to the question regarding farmer s willingness to adopt GMHT OSR maize. except for France where reducing weed control costs (statement 5) is considered the most important factor in deciding to adopt GMHT maize (average scores above 4). Other relevant factors determining GMHT maize adoption have also an economic character such as guaranteeing the reduction in losses caused by weed growth (statement 1) and guaranteeing greater yields during harvest (statement 3). On the other hand, least persuading statements for farmers regarding GMHT maize adoption were those related to farmer s perception of both social pressures on him to adopt such as my regular distributor recommended it (statement 6) or all farmers in my area are using it (statement 11) as well as statements related to personal satisfaction derived from using the new technology since It makes me feel as though I were at the forefront of biotechnological progress (statement 10; average scores below 3). Spanish and Hungarian farmers tend to give similar scores to encouraging statements, whereas French farmers tend to be less persuaded by the encouraging statements. In contrast, while Hungarian and French farmers give similar scores to dissuading statements, Spanish farmers are relatively highly persuaded by negative statements on the rejection of GMHT maize. Spanish farmers hold the strongest views about both encouraging and dissuading statements, showing that almost all statements (positive and negative) were relevant for them regarding adoption or rejection of GMHT maize. 5 Positive statements Negative statements Average score Number of statement Spain Hungary France More Less persuading Indecisive persuading Figure 1 Attitudes towards GMHT maize: average score per statement. Average score Positive statements Negative statements More Less persuading Indecisive persuading Attitudinal statement number Czech Rep UK Germany Figure 2 Attitudes towards GMHT OSR: average score per statement.

5 Attitudes of EU farmers towards GM crop adoption 949 With regard to GMHT OSR, results are similar to the ones obtained for GMHT maize. In this case, the most important factors for all countries are the guarantee of a higher income (statement 4), reduction of weed control costs (statement 5) and obtaining greater yields during harvesting (statement 3) being farmer s perception of social pressures on him as well as farmer s own satisfaction derived from the use of the new technology the least persuading motivations (statements 6, 11 and 10, respectively). Negative attitudes towards GMHT crop adoption Amongst the statements that may cause farmer s rejection of adopting GMHT crops, I have more faith in the use of herbicides to avoid weed infestation than in this type of crop (statement 22) is the most important on average for French and Spanish (GMHT maize) and British (GMHT OSR) farmers showing a preference for current practices against the new technology. While for Spanish and British farmers, this is a key factor that would persuade them to reject adoption, the average score (3.29) given by French farmers indicates that they are unsure about whether they would reject or adopt GMHT maize for this reason. On the other hand, the most important statement that may cause farmer s rejection in Hungary (GMHT maize), the Czech Republic and Germany (GMHT OSR) is the higher price of GM seeds when compared with their conventional counterparts (statement 15). It is also worth noting that the average score given to seed price by Hungarian (3.32) and Czech farmers (3.20) shows that farmers were unsure about whether higher seed price was enough reason to be dissuaded from adopting GMHT maize or GMHT OSR. With regard to the least relevant statement to EU farmers, farmer s disbelief in an improvement in yield and being advised not to use GMHT crops are the least relevant statements regarding rejecting GMHT OSR adoption on average (statements 16 and 20). From Czech farmers scores, it can be concluded that they would be the keenest on adopting GMHT crops because they tend to be the most persuaded by the encouraging statements and the least persuaded by the rejecting issues concerning the adoption and rejection of GMHT OSR. Cluster analysis We conducted a cluster analysis using all 24 statements related to GMHT crop adoption. It is worth noting that in case we had grouped farmers by the simple examination of responses to the question about farmer s willingness to adopt GMHT OSR GMHT maize, three groups would had been obtained including potential, potential and indecisive farmers (see Table 2). Cluster analysis allowed us to find groups of farmers with similar views on GMHT crop adoption. According to this technique, farmers were progressively grouped into smaller groups. At the beginning, each farmer is a cluster itself. Its closest neighbour in terms of Euclidean distance (L2 squared) is then aggregated to a new cluster. This aggregation procedure produces a dendrogram for each crop and country (Figure 3) from which the researcher determines the number of clusters, seeking a trade-off between the number of clusters and their internal homogeneity. In this particular case, we found appropriate to cut the dendrogram to group the farmers into two groups of farmers by crop and country, i.e. potential and potential. Taking into account the number of clusters obtained by crop and by country, adoption rates can be calculated (see Table 3). Comparing with the adoption rates shown in Table 2, we can appreciate that the latter results showed similar rates of adoption (i.e. adding the responses likely and very likely ) and rejection (i.e. adding the responses somewhat unlikely and very unlikely ) when facing the adoption of GMHT maize. This similarity between both rates made interpretation vague by country, and no conclusion can be made regarding the position of the country when facing the cultivation of new GMHT crops Spain Dendrograms for maize farmers Hungary France Czech Republic Dendrograms for OSR farmers UK Germany Figure 3 Dendrograms for maize and OSR farmers by country.

6 950 Francisco J. Areal et al. Table 3 Cluster profile: Potential versus potential GMHT maize Spain Hungary France Cluster 1 Cluster 2 Cluster 1 Cluster 2 Cluster 1 Cluster 2 Characteristic SP1-Potential SP2-Potential HU1-Potential HU2-Potential FR1-Potential FR2-Potential No. of farmers (potential adoption rate, %) 50 (54) 43 (46) 27 (41) 39 (59) 27 (55) 22 (45) Average farm size (ha) 41* 29* Average income ( ) n.a. n.a Average age 47** 52** ** 41** Willingness to adopt GMHT maize (score) 3.64*** 1.84*** 3.93*** 2.44*** 4.07*** 2.32*** GMHT OSR Czech Republic UK Germany Cluster 1 Cluster 2 Cluster 1 Cluster 2 Cluster 1 Cluster 2 Characteristic CZ1-Potential CZ2-Potential UK1-Potential UK2-Potential GE1-Potential GE2-Potential No. of farmers (percentage) 52 (50) 51 (49) 96 (67) 47 (33) 107 (55) 86 (45) Average farm size (ha) Average income ( ) ** 7550** *** *** * * Average age Willingness to adopt GMHT OSR (score) 3.85*** 2.82*** 3.57*** 2.68*** 3.93*** 2.19*** *, ** and *** mean that Cluster 1 characteristic is statistically different from that of Cluster 2 at 0.1, 0.05 and 0.01 levels, respectively. (e.g. 37% of farmers were willing to adopt and 38% were willing to reject GMHT maize in Spain, while 35% of Hungarian farmers were willing to adopt and 38% were willing to reject GMHT maize). Cluster analysis then contributes to elucidate the interpretation of potential adoption rates by crop and country (Table 3). Thus, according to the values assigned by farmers to encouraging or dissuading statements regarding GMHT crop adoption, indecisive farmers were grouped into potential or potential, and the difference between adoption and rejection rates per country and crop was higher than the rates shown in Table 2 for most cases. The characterization of these clusters is shown in detail in Table 3. Statistical tests were conducted to ensure that groups obtained from the cluster analysis differ on average from each other. Kolmogorov Smirnov and Levene s tests were conducted to test for normality and equality of variances, respectively. For those which failed one of the assumptions, nonparametric Mann Whitney tests were conducted. To give a profile of each group of farmers in each country, we took into account socioeconomic variables such as farm size, income and farmer s age as well as the average score given to the statement on the willingness to adopt GMHT crops (maize or OSR). See Supporting Information II for a distribution of socioeconomic data per country and crop. Amongst all socioeconomic variables, only willingness to adopt shows statistically significant differences between clusters in each country and for both crops. Therefore, the clusters were classified as potential or potential of GMHT crops. We consider that scores from 1 to 2.50 mean that farmers are willing to reject GMHT crops; scores from 2.51 to 3.50 mean that the farmer is indifferent; and scores from 3.51 to 5 mean that farmers are willing to adopt GMHT crops. Consequently, clusters 1 include potential of GMHT crops, whereas clusters 2 include potential of this type of crops. Farmers included in cluster 2 in the Czech Republic and Hungary have a score slightly higher than 2.50, meaning that these farmers are indifferent but closer to a potential rejection of GMHT OSR. Mean comparison tests were conducted for each encouraging and dissuading statement to complete the profile of each group of farmers in each country. Tables 4 and 5 show the average scores per country and group related. The statement scores between clusters within a country are mostly statistically different from each other. Taking into account the methodological approach, farmers were grouped into potential or potential on the basis of their attitudes towards GMHT crops (see Figure 4, step 1). Therefore, each cluster shares similar scores of statements and, as it was expected, farmers included in clusters 1 (potential ) show, in average, higher values for encouraging statements than farmers of clusters 2 (potential ). Some exceptions were found in the Czech Republic when statements 6, 10, 11 and 12 were evaluated by OSR farmers. In this latter case, farmers included in clusters 1 show lower values than farmers of clusters 2. It is worth mentioning the low importance of these statements for the rest of clusters in France, Hungary, Germany and UK. In all these cases, the low valuation of the statements by farmers included in clusters

7 Attitudes of EU farmers towards GM crop adoption 951 Table 4 Encouraging statement to adopt: average score by cluster and by country, GMHT maize and GMHT OSR Table 5 Dissuading statement to adopt: average score by cluster and by country, GMHT maize and GMHT OSR GMHT maize GMHT OSR GMHT maize GMHT OSR Statement number Cluster Spain France Hungary Germany Czech Republic UK Statement number Cluster Spain France Hungary Germany Czech Republic UK a 4.15 a 4.52 a 3.86 a 4.46 a 4.21 a b 2.68 b 4.21 a 2.62 b 3.90 b 3.34 b a 3.19 a 3.81 a 3.58 a 4.08 a 3.07 a b 1.77 b 2.62 b 2.30 b 3.00 b 1.98 b a 4.48 a 4.59 a 4.19 a 4.58 a 4.48 a b 2.68 b 4.00 b 2.88 b 4.25 a 3.34 b a 4.48 a 4.59 a 4.42 a 4.71 a 4.76 a b 2.73 b 4.28 a 3.19 b 4.29 b 3.81 b a 4.52 a 4.63 a 4.39 a 4.48 a 4.58 a b 3.05 b 3.90 b 3.13 b 4.35 a 3.81 b a 3.22 a 2.52 a 2.53 a 2.21 a 3.08 a b 1.64 b 1.77 b 1.91 b 2.51 a 2.43 b a 4.26 a 4.52 a 3.91 a 4.29 a 4.32 a b 2.59 b 3.18 b 2.57 b 4.08 a 3.15 b a 4.48 a 4.41 a 4.00 a 4.02 a 4.43 a b 2.73 b 3.97 a 2.70 b 3.94 a 3.26 b a 4.15 a 4.07 a 4.07 a 3.75 a 4.21 a b 2.23 b 3.59 a 2.55 b 3.69 a 3.09 b a 3.37 a 3.33 a 3.07 a 1.75 a 2.89 a b 2.00 b 3.00 a 1.84 b 2.27 b 1.94 b a 2.63 a 3.19 a 2.17 a 1.73 a 2.55 a b 1.73 b 2.03 b 1.38 b 2.51 b 1.91 b a 3.41 a 3.59 a 2.65 a 1.94 a 3.48 a b 1.59 b 2.36 b 1.81 b 2.92 b 2.34 b a 4.26 a 4.11 a 3.96 a 3.75 a 3.89 a b 2.64 b 3.41 b 2.36 b 3.75 a 2.79 b This table shows the average score given by each cluster (cluster 1 or potential and cluster 2 or potential ) for encouraging statements (statements 1 13) on the potential adoption of GMHT crops. Average scores below 2.50 mean that farmers were not persuaded by encouraging statements to adopt GMHT crops, whereas scores above 3.50 mean that farmers were highly persuaded by encouraging statements to adopt GMHT crops. Note: Average scores between clusters for each country were statistically tested for significant difference. Scores followed by the same letters are not statistically different at 0.05 level. 1 (scores below 3.5) means that farmer s own satisfaction derived from the use of the new technology and neighbours position towards GMOs are poor incentives to encourage the adoption of GMHT crops. Regarding dissuading statements, and owing to the design of the cluster analysis, farmers included in clusters 2 (potential ) show, in average, higher values than farmers of clusters 1. Only one exception was found in the UK for the statement 21 with similar values for both clusters (scores around 3.10). In this case, this statement shows that farmers were indifferent towards the difficulty to market GMHT OSR grain. Results also show that the higher valuation of the statements (i.e. value given by farmers included in clusters 2) was below 3.6 for statements 16, 19, 20 and 21 in all countries except Spain. Consequently, farmer s disbelief in an improvement in yield, the difficulty associated with new crops and being advised a 3.07 a 2.48 a 2.77 a 2.06 a 2.60 a not to use GMHT crops are the least relevant statements regarding rejecting GMHT crop adoption. GMHT maize b 3.36 a 3.87 b 4.08 b 2.90 b 3.21 b a 3.00 a 2.93 a 3.53 a 2.21 a 3.02 a b 3.55 a 3.59 b 4.48 b 3.59 b 3.89 b a 2.07 a 1.78 a 2.65 a 1.54 a 2.49 a b 3.00 b 3.10 b 3.30 b 2.57 b 2.94 a a 2.59 a 1.96 a 2.94 a 1.63 a 3.06 a b 3.73 b 3.92 b 3.99 b 3.08 b 3.60 b a 2.63 a 2.00 a 3.12 a 1.54 a 3.59 a b 3.36 a 4.05 b 4.08 b 3.10 b 3.94 a a 2.11 a 1.74 a 2.88 a 1.58 a 2.56 a b 3.00 b 3.10 b 4.34 b 2.94 b 3.64 b a 2.00 a 1.81 a 2.06 a 1.44 a 2.52 a b 3.23 b 2.82 b 2.98 b 2.75 b 3.00 b a 2.07 a 1.85 a 2.72 a 1.87 a 3.15 a b 3.09 b 3.59 b 3.53 b 2.80 b 3.09 b a 3.22 a 2.15 a 3.24 a 2.04 a 3.64 a b 3.45 a 4.00 b 4.36 b 3.08 b 4.04 a a 2.33 a 1.85 a 2.79 a 2.13 a 3.21 a b 3.55 b 3.79 b 4.21 b 2.78 b 4.02 b a 2.33 a 2.63 a 3.50 a 2.29 a 2.98 a b 3.32 b 3.62 b 4.58 b 3.06 b 3.64 b This table shows the average score given by each cluster (cluster 1 or potential and cluster 2 or potential ) for dissuading statements (statements 1 13) on the potential adoption of GMHT crops. Average scores below 2.50 mean that farmers were not persuaded by encouraging statements to adopt GMHT crops, whereas scores above 3.50 mean that farmers were highly persuaded by encouraging statements to adopt GMHT crops. Note: Average scores between clusters for each country were statistically tested for significant difference. Scores followed by the same letters are not statistically different at 0.05-level. Table 4 shows that Spanish potential (SP1) evaluated positively all encouraging statements (all above 3.5). Even potential (SP2) evaluated economic and agro-economic factors (statements 1, 3, 4 and 5) above 3.5, indicating that Spanish farmers as a whole perceived good reasons to adopt GMHT maize on the basis of economic reasoning. This result is consistent with previous research as economic and agronomic benefits were pointed out as drivers to explain the rapid adoption of GMHT crops worldwide by Fernández-Cornejo et al. (2002), Phillips (2003), Brookes and Barfoot (2005), Dill et al. (2008) or Finger et al. (2009) amongst others. Regarding dissuading statements (Table 5), the potential of GMHT maize in Spain tended to score values above 3.5, which indicate that the potential reasons to reject GMHT maize cultivation highlighted in those statements would dis-

8 952 Francisco J. Areal et al. Statements reflecting attitudes towards GM crops SP farmers GMHT FR farmers maize HU farmers GMHT OSR CZ farmers UK farmers GE farmers Cluster t analysis Cluster 1 Cluster 2 Cluster 3 Willingness to adopt GMHT crops Socio-economic variables Cluster profile Potential Potential Comparative analysis of clusters attitudes towards GM crops Evaluation of coexistence measures by 1st step 2nd step 3rd step Figure 4 Outline of the methodological approach. suade this group to cultivate GMHT maize, the principal reason of rejection being the relative high GMHT maize seed price. Farmers concerns about the perception of GM crops by society (statement 23), their dislike for change (statement 19) and being advised not to use the new technology (statement 20) are amongst the statements that the potential of GMHT maize considered as causes that they would probably not cause rejection of GMHT maize adoption. French potential of GMHT maize (FR1) also deemed economic (statements 1, 4 and 5), agronomic (statements 3 and 7), technical (i.e. facilitation of work), environmental (statement 8) and risk issues (statement 9) as decisive factors to adopt the new technology. Amongst the least valued reasons to encourage adoption were those associated with social and peer pressure to adopt (statements 6, 10, 11 and 12). The relative high price of GMHT maize seeds with respect to the conventional seed price (statement 15), disbelief for the new technology (statement 17) and social views on GM crops (statement 23) were deemed as the main reasons to reject GM adoption by the French potential (FR2). Amongst, results show that social factors appear to play a role in the rejection of GM crops, which is consistent with previous research by Morin (2008) and Smyth et al. (2011). However, their scores do not show a view as negative on these issues as the one shown by the Spanish potential of GMHT maize adoption. On the other hand, potential in France considered negative agronomic reasons (statement 16), dislike to changes; difficulties to market the grain (statement 21); social views on GM and the complexity of the new technology in its use (statement 24) as reasons that would not cause them rejection towards GMHT maize adoption. The evaluation of encouraging statements by the Hungarian potential group (HU1) is similar to their French colleagues. Amongst their least convincing statements to encourage GMHT maize adoption for the Hungarian potential were the recommendation of the new technology by their regular distributor (statement 6); all farmers in my area are using it (statement 11); and it facilitates my work being a technology that makes cultivation easier (statement 13). Focussing on the dissuading statements, the Hungarian potential of GMHT maize (HU2) considered most of the statements as dissuading enough to cause rejection of GMHT maize cultivation (scores above 3.5). GMHT OSR German, Czech and British farmers show similar attitudes regarding encouraging statements to adopt GMHT OSR (see Table 4). Thus, potential in these countries (GE1, CZ1 and UK1) considered economic, agronomic, technical, environmental and risk issues as important factors encouraging them to adopt GMHT OSR, whereas factors related to social and peer views on GM crops (statements 10, 11, 12 and 13) were irrelevant for them. It is worth highlighting the high scores (above 3.9) given to economic, agronomic, technical and environmental and risk issues also by the potential group in the Czech Republic (CZ2). The least encouraging reason to adopt GMHT OSR for the potential in the Czech Republic is the own satisfaction derived from being part of the biotechnological progress (statement 10 with an average score of 2.27). German and British farmers (GE2 and UK2) also pointed this as a low encouraging statement in addition to other statements such as facilitating the process throughout practising direct sowing, recommendation of use the new technology by their regular distributor or technicians and being the new technology used by farmers in the area (statements 2, 6, 12 and 11). Larger differences were found between average evaluation scores per countries, particularly dissuading statements seem not to be convincing enough for Czech farmers. The potential in the Czech Republic (CZ1) considered every single dissuading statement as not being enough to cause rejection on GMHT OSR cultivation (see Table 5). On the other hand, the Czech potential of GMHT OSR pointed out the relative high seed price of the new technology as a dissuading driver on GMHT OSR adoption (statement 15). German and British farmers tend to give similar evaluations to the dissuading statements. The potential in both countries (GE2 and UK2) deem most of the dissuading statements as dissuasive to adopt GMHT OSR. Exceptions for German and British farmers thought that adopting would not lead to a yield improvement (statement 16) and being advised not to use this type of crop (statement 20) showing scores below 3.5. In addition, not having a huge weed problem in their land (statement 14) and thinking that it would be difficult to market the grain (statement 21) are also exceptions for British farmers. The potential in the United Kingdom (UK1) disregard agro-

9 Attitudes of EU farmers towards GM crop adoption 953 nomic reasons such as not having an improvement in yield (statement 16) as a discouraging reason, whereas German farmers (GE1) disregard being advised not to use this type of OSR (statement 20). Coexistence measures In addition to the analysis conducted to identify and profile farmers based on their attitudes towards GMHT crop adoption, we assessed how different coexistence measures may influence potential views on adopting GMHT crops. Because farmers included in clusters 1 are the keenest groups on adopting GMHT crops (potential ), results were only referred to these groups of farmers. To analyse how potential value the implementation of the coexistence policy in the EU, a set of coexistence measures was included in the questionnaire (Table 6). Farmers were asked to evaluate these statements using a Likert scale from 1 to 5, where 1 means completely unimportant coexistence measure in deciding to reject GMHT crops and 5 means very important. To interpret the results, we consider that scores from 1 to 2.50 mean that the statement represents an unimportant measure for farmers when deciding whether or not to adopt a GMHT crop; scores from 2.51 to 3.50 denoted that the coexistence measure does not matter for farmers decision process; and scores from 3.51 to 5 mean that the coexistence measure is an important issue for farmers and may cause the rejection of GMHT crops. Figures 5 and 6 show the average scores given to statements related to coexistence measures given by potential of GMHT maize and GMHT OSR, respectively. Figure 5 shows that the highest valuation of coexistence measures (i.e. score above 3.5) was given to those measures related to a monetary compensation to neighbours (Co5 and Co6) and the implementation of a tax policy for sowing GM crops (Co9). As GM farmers are responsible to implement coexistence measures, the application of these measures would increase their costs and subsequently may cause the rejection of GMHT maize. Table 6 Statements reflecting attitudes towards coexistence measures Number Co1 Co2 Co3 Co4 Co5 Co6 Co7 Co8 Co9 Co10 Co11 Co12 Co13 Coexistence statement The plot must be identified in a public register Neighbouring farmers must be notified The genetically modified crops must be identifiable If the land is rented, the owner must be informed Neighbours would have been compensated in case of adventitious presence An insurance policy must be underwritten to cover potential claims from neighbours The farmer must follow a training course The combine harvester must be thoroughly cleaned before and after harvesting A tax must be paid for sowing genetically modified crops Genetically modified crops and conventional crops must be sown at different times of the year The crop must be labelled as GMO when it is marketed, making it visible to buyers Compulsory separation distance or buffer zones Compulsory of separate harvesting and transportation Spanish farmers were the most concerned with the implementation of the coexistence policy because they gave scores greater than 3.5 to most of the coexistence measures. In contrast, Hungarian seem to be the least worried with this type of measures to ensure coexistence between GM and non- GM crops. The average farm size of maize potential in Spain and Hungary (see Table 3) may be a reason to explain this different attitude towards coexistence measures. Thus, Hungarian potential have an average farm size 18 times larger than Spanish farmers. A number of studies have found that smaller farm sizes were more concerned with coexistence measures as these measures may imply a reduction in their competitiveness (Beckmann et al., 2006; Messean et al., 2006; Demont et al., 2008). Analysing the attitudes of potential of GMHT OSR results was similar to those obtained for GMHT maize because the coexistence measures related to neighbours and tax policy were the most valued by farmers (scores above 3.50 for all countries). Therefore, the implementation of any of these measures may cause the rejection of GMHT OSR. While German farmers were the most worried by the implementation of almost all coexistence measures (with the exception of following a training course), Czech farmers showed the least concern with these measures. German farmers concerns on coexistence measures have been analysed by Consmüller et al. (2010) showing that farmers willingness to adopt will strongly depend on coexistence costs. Methodology A flow chart summarizing the main steps of the methodological approach is shown in Figure 4. In a first step, a cluster analysis, a multivariate statistical technique, was performed for each crop and in each country to group individuals into mutually exclusive groups. A hierarchical cluster analysis using a Ward s method squared Euclidean distance followed by a k-means cluster analysis was conducted on the farmers evaluations of the 24 statements. Thus, groups of farmers can be identified on the basis of their attitudes towards GMHT crops (i.e. encouraging and dissuading statements associated with GMHT adoption). Therefore, each cluster or group of farmers shares particular characteristics providing insight into producer s underlying attitudes towards GM crops. The cluster analysis produced a two-cluster solution per country studied. Once groups were obtained, we applied statistical tests to elucidate farmers attitudes towards GM crops as well as any differences that may be found between groups, countries and crops. To ensure that groups obtained from the cluster analysis differ from each other, two types of tests (parametric t- tests and nonparametric Mann Whitney tests) were conducted. Farmers evaluations of each statement were analysed by testing differences by country using these tests. In a second step, a profile for each group in each country was identified. This was conducted by considering both the average scores given to the statement on the willingness to adopt GMHT crops and farmer s socioeconomic variables. Once a profile was obtained, a last step involving a comparative analysis of clusters attitudes towards GMHT crops, as well as the evaluation on how different coexistence measures may influence on potential of GMHT crops, was carried out. The analysis of coexistence measures was performed only for potential as they are the keenest groups on adopting GMHT crops.

10 954 Francisco J. Areal et al. 5 Average score Co1 Co2 Co3 Co4 Co5 Co6 Co7 Co8 Co9 Co10 Co11 Co12 Co13 Coexistence statement number Spain France Hungary More Less persuading Indecisive persuading Figure 5 Attitudes of potential of GMHT maize towards coexistence: average score per statement. 5 Average score Co1 Co2 Co3 Co4 Co5 Co6 Co7 Co8 Co9 Co10 Co11 Co12 Co13 Coexistence statement number Czech Rep UK Germany More Less persuading Indecisive persuading Figure 6 Attitudes of potential of GMHT OSR towards coexistence: average score per statement. Discussion and Conclusions This article contributes to the literature on stakeholders views on GM crops by analysing the views of one of the stakeholders least studied, farmers. Farmers views expressed through a survey to EU maize and OSR growers allowed us to estimate the average adoption rates of GMHT OSR and GMHT maize per country studied and shape farmers profiles with respect to GMHT crop adoption. The study provides insight into the effects that putting in place coexistence measures to ensure coexistence between GM and non-gm crops may have on potential attitudes towards adopting GMHT technology. Potential and potential of GMHT crops were identified for each of the countries studied. Economic aspects either associated with production or legislation on coexistence were found to be relevant for when deciding whether or not to adopt GMHT crops. Economic aspects were the most important factors profiling the attitudes of both potential and potential towards GMHT crops. Economic issues such as the guarantee of a higher income or the reduction in weed control costs were found to be the most encouraging reasons for both potential and potential of adopting GMHT crops. Furthermore, economic issues such as the relative high price of GMHT seeds were also amongst the most important reasons found by farmers to reject adoption in addition to the disbelief or the complexity of the new technology. Overall, a significant number of European farmers would be keen on adopting GMHT crops, especially GMHT OSR in case cultivation of GMHT crops in the EU was allowed. Currently, GMHT crops are allowed to be imported in the EU but their cultivation has not been approved yet. We found that over half of German farmers and almost half of Czech and UK farmers would be keen on adopting GMHT OSR, whereas over a third of Spanish, French and Hungarian farmers would be keen on adopting GMHT maize. The fact that GMHT OSR is relatively more appealing to EU farmers than GMHT maize is probably reflecting a relatively higher complexity of treating weeds in the case of OSR crops than maize crops. The willingness to adopt GMHT crops from a number of European farmers would be significantly affected by the implementation of coexistence measures, especially those that mean an economic burden for the farmer. The European Coexistence Bureau (ECoB) 5 has recently published the best practices for coexistence of GM with conventional maize, which serves as a 5 The ECoB was established jointly by the Directorate General for Agriculture and Rural Development and the Joint Research Centre (JRC) of the EC to organize the exchange of technical-scientific information on best agricultural management practices for coexistence and, on the basis of this process, develop consensus agreed crop-specific guidelines for coexistence measures. The ECoB does not cover any monetary measure (e.g. taxes and compensations) as implementing such measures (fiscal policies) is competence of each Member State and hence cannot be harmonized at the EU level.