Measuring consumer preferences regarding

2 downloads 0 Views 844KB Size Report
consumer preferences regarding organic labelling and the JAS label in particular, New Zealand. Journal of Agricultural ... every food product required quality labelling (MAFF. 2004). ..... tem is rated positively on the demand side, although.
New Zealand Journal of Agricultural Research

ISSN: 0028-8233 (Print) 1175-8775 (Online) Journal homepage: http://www.tandfonline.com/loi/tnza20

Measuring consumer preferences regarding organic labelling and the JAS label in particular Masaji Sakagami , Masayuki Sato & Kazuhiro Ueta To cite this article: Masaji Sakagami , Masayuki Sato & Kazuhiro Ueta (2006) Measuring consumer preferences regarding organic labelling and the JAS label in particular, New Zealand Journal of Agricultural Research, 49:3, 247-254, DOI: 10.1080/00288233.2006.9513715 To link to this article: https://doi.org/10.1080/00288233.2006.9513715

Published online: 17 Mar 2010.

Submit your article to this journal

Article views: 257

View related articles

Citing articles: 9 View citing articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tnza20 Download by: [154.16.42.67]

Date: 19 December 2017, At: 04:30

New Zealand Journal of Agricultural Research, 2006, Vol. 49: 247-254 0028-8233/06/4903-0247 © The Royal Society of New Zealand 2006

247

Downloaded by [154.16.42.67] at 04:30 19 December 2017

Measuring consumer preferences regarding organic labelling and the JAS label in particular

MASAJI SAKAGAMI Department of Human Environment and Information Technology Nihon Fukushi University Aichi, Japan [email protected] MASAYUKI SATO KAZUHIRO UETA Department of Economics Kyoto University Kyoto, Japan Abstract This study aims to examine whether Japanese consumers show a preference for particular types of food certification, particularly the organic Japanese Agricultural Standards (JAS) label, and how important certification of food products is compared to their freshness and source. Organic JAS certification is a certification system with government-backed legal foundations. To consider these issues, we conducted a willingness to pay (WTP) survey for vegetables (spinach) as food products with a choice experiment (CE). The vegetables presented in the CE had four attributes: freshness, source, price, and certification. As a result, we found that Japanese consumers valued food certifications and had a WTP for them, including the JAS label. On the other hand, certification was a smaller factor than freshness and source in vegetable purchasing criteria. Freshness was the most important factor for Japanese consumers. Keywords JAS; choice experiment; consumer preferences; organic label

A05038; Online publication date 4 July 2006 Received 15 September 2005; accepted 2 June 2006

INTRODUCTION There has been growing interest among consumers in the safety of farm products and their impact on human health and wellbeing (Mega & Kazuno 2003). Many consumers are concerned whether pesticides are used in the production of farm products (National Federation of Agricultural Co-operative Associations 2003). Organic agricultural products are therefore becoming attractive to many consumers (Mega & Kazuno 2003; Ministry of Internal Affairs and Communications 2002). Consumers want information about production methods and conditions to be displayed on, for instance, certification labels of products. According to the study of Ishida et al. (2003) and Mineki et al. (2001), certification and labelling of products is an important issue for Japanese consumers. In Japan, the organic Japanese Agricultural Standards (JAS) is one such certification and labelling system. Under the JAS regulations, only those products that meet JAS certification standards can be officially or legally labelled as organic agricultural products. It assures consumers that a product that is not observably different from non-organic food was grown, processed and packaged according to rules that limit or ban synthetic inputs and that protect the environment (Lohr 1998). Thus, this study has two research questions: (1) Do Japanese consumers show a preference for particular systems of food certification? (2) How important is certification of food products compared to their freshness and source? To consider these issues, we mainly use the example of the organic JAS certification label. JAS label The law concerning standardisation and proper labelling of agricultural and forestry products that first created JAS was established in 1950 (MAFF 2004). Later, the Consumer Protection Fundamental Law and other laws brought about amendments in 1970 to include regulations about food product quality labelling (Yoshida 2005). However, the promotion of labelling was not very aggressive, and only 64 items of food products required quality labelling by 1999

Downloaded by [154.16.42.67] at 04:30 19 December 2017

248

New Zealand Journal of Agricultural Research, 2006, Vol. 49

(Yoshida 2005). Labelling for organic agricultural products had also been chaotic. Prior to 2001, there was no accurate definition of organic products in Japan. However, concern about food product safety and reliability had been growing, and additional revisions were made in 1999, so that from April 2001 every food product required quality labelling (MAFF 2004). JAS standards for organic agricultural products based on CODEX standards were set, and now only products that meet those standards are allowed to be labelled as "organic" (MAFF 2004). These standards stipulate production principles for organic agricultural products as mentioned above, as well as set labelling rules for production method standards and a name labelling system (see MAFF 2004 for details). The certification of these standards is handled by private registered certification organisations designated by the MAFF. In order to receive certification, authorisation fees, inspection costs, meetings, training and other participation expenses are incurred, while the time and amount of effort to complete certification documents, keep records and handle other paperwork creates significant burdens (Ohyama 2003). At the time of authorisation registration, the registered certification organisation conducts onsite investigations of production methods and management and further investigation at least once within a year of authorisation (MAFF 2004). In Japan, a system has been established for normal non-organic vegetables in which farmers deliver produce to local agricultural cooperatives, which then handle further logistics automatically. No such system exists for organic vegetables, however, so the producers themselves must take care of distribution and find their own sales points (Nihon Keizai Shimbun 2005). Given the costs and paperwork complexity of organic JAS certification, producers have been passive about applying for it. As of the end of 2005, there were 95 registered certification organisations and only 5504 certifications, showing that the amount of vegetables with organic JAS certification reaching the market is still low (MAFF 2004). In recent years, the market value of organic products has been about a billion yen (Hanawa 2001) and annual production has been about 30 000 tons, a mere 0.17% of total vegetable production (Nihon Keizai Shimbun 2005). Thus, some problems exist with the organic JAS certification system. The aim of our research was to answer two questions: (1) Do Japanese consumers show a preference for particular types of food certification, particularly organic JAS vegetables, over ordinary produce?

Perhaps, they may feel ordinary vegetables are good enough because of the price premium of certified vegetables. On the other hand, they may be satisfied by vegetable certifications by other organisations, such as non-profit organisations (NPO) and supermarkets. Unlike JAS, these are not based on legal regulations. Premium prices will be paid for organic products only when there is consumer confidence in the certification label (Lohr 1998). Third-party NPO certification is assumed to certify vegetables that either use no agricultural chemicals or chemical fertilisers, or that have reduced the amounts of agricultural chemicals and chemical fertilisers used by at least 50% compared to conventional production practices. These cannot be labelled "organic", but instead are identified as "specially grown agricultural products". Japan has such a system administered by local government, but the number of products certified is small. Supermarket certifications are second-party certifications. We used the example of certification by a large Japanese supermarket house brand. These certified vegetables are produced in cooperation with the producers to reduce the amounts of agricultural chemicals and chemical fertilisers greatly. Portrait photographs of the producers and other forms of identification are shown at the sales points, and arrangements are made for consumers to hear about types and amounts of agricultural chemicals and fertilisers used directly from the producers. Furthermore, since supermarket certificated vegetables are prepared in cooperation with producers, we saw this approach as also being a kind of first-party certification and did not set a separate first-party certification standard as a criterion in our choice experiment (CE) investigation. However, first-party audits prepared by producers themselves are extremely important elements, so it would have been better to set this as an independent certification factor. This is the most important task for future research. (2) How important is certification of food products compared to their freshness and source? Perhaps consumers are satisfied with vegetables from preferred regions if they are fresh, whether or not they have organic certification. In this research, we sought to answer the above two questions using choice experiment (CE) methods to evaluate economic value for our willingness to pay (WTP) survey and adopted spinach as an example of food products. We selected spinach as it is consumed throughout the year. CE and contingent valuation methods (CVM) are stated preference methods which are effective

Downloaded by [154.16.42.67] at 04:30 19 December 2017

Sakagami et al.—Consumer preferences regarding organic labels for directly asking about resources that have little or no market values. CE allows an object to be evaluated using several of its attributes at once. And multiple alternatives can be presented (Louviere et al. 2000; Champ et al. 2003). We adopted CE because of these merits. Moreover, in our study, CE with hypothetical condition is more effective than preference analysis based on poor market data. In CE, multiple profiles comprised of sets of attributes and levels are created. However, when the number of combinations of attributes and levels is large, the number of profiles should be reduced using an orthogonal design (Louviere et al. 2000; Holmes & Adamowicz 2003). From these, questions with two, three or more alternatives are created. The survey respondents then choose which profile (alternative) is the most desirable. In actual CE surveys, these types of questions are repeated in succession. Literature review There have been some studies about WTP surveys for organic products and food certification labels. Thompson (1998) provided a comprehensive survey of studies about organic products. Many studies discovered a consumer price premium for a certain kind of organic product. Loureiro & Hine (2002) demonstrated that consumers were willing to pay 6.64 cents/lb for organic potatoes, but that locally produced potatoes had a higher WTP at 9.37 cents/lb using CVM. However, wealthier and more educated consumers preferred organic potatoes to local ones. Also, the age of respondents was negatively correlated with the estimated WTP for organic foods. Buzby et al. (1995) found a price premium of 38% for avoiding sodium ortho-phenylphenate (SOPP) using CVM which is a stated preference method. Baker (1999), using a CE, indicated that consumers have a high preference for apples with reduced pesticides. Furthermore, Roosen et al. (1998) studied WTP for apples with insecticide and those with none using an experimental auction design. They found that consumers changed their perception of product attributes if the use of pesticides was removed, causing a change in WTP. There also have been some studies about WTP surveys for food certification or label. Loureiro et al. (2003) revealed that consumers with low WTP for labelled apples were less likely to actually buy them, while those with high WTP were more likely to buy them. Blend & van Ravenswaay (1999) conducted a CVM and found the price premium for eco-labels.

249

Huffman (2003) also provided effective analysis of genetically modified food labels and found that socio-demographic characteristics of consumers didn't affect WTP for them.

MATERIALS AND METHODS CE survey In our CE survey, we created hypothetical vegetable choice questions, and analysed selection results statistically to clarify consumer vegetable preferences. The vegetables presented in the CE had four main attributes: freshness, region of production, price, and certification. Freshness of the spinach had three discrete levels. To provide respondents with some image of freshness, we provided pictorial examples. The second attribute, origin, consisted of foreign, locally produced, or produced in another region of Japan. The third attribute, price, consisted of 130,160,190 and 220 yen. The fourth attribute, certification, consisted of JAS certification, NPO certification, supermarket certification and no label. We adopted an orthogonal design to reduce profiles (the number of attributes and levels are mentioned below) and we made a question (see Appendix). There were three alternatives plus a "nochoice" option from which to choose. The results of the choices by all respondents are dependent variables, and the attribute level values are independent variables. These are analysed statistically using a discrete choice model based on random utility theory. In this way, the degree of influence of each attribute on choice behaviour can be quantified. Having prepared the description of the CE questionnaire in this way, we implemented our survey with 2000 participants. At a Kyoto branch of a leading supermarket, questionnaires were distributed to customers who agreed to participate in our research immediately after they had paid the cashier. The survey was conducted from 10 a.m. to 8 p.m. on 26 and 28 January 2003. We collected the questionnaires, giving a 200 yen merchandise coupon as a token of thanks to those who filled out the form and submitted it to the questionnaire counter. We distributed 2000 survey forms, and collected 698 (effective answers): a collection rate of 35%. Eighty-two percent of the respondents were female, 16% were male, and 2% did not give their gender. Respondents' ages and occupations are shown in Table 1.

New Zealand Journal of Agricultural Research, 2006, Vol. 49

250

Downloaded by [154.16.42.67] at 04:30 19 December 2017

Model Discrete choice models such as the logit model are often used in CE. Logit models are consistent with random utility theory, and are applicable to welfare economic topics. The random utility function consists of a nonstochastic term V(·) and the stochastic term ε. When individual n chooses alternative i, the utility is Uin = Vin(xin,min)+εin (1) where xi is a vector of attributes of profile I, and mi is a monetary attribute (e.g., price). Including at least one monetary attribute like this makes it possible to evaluate each attribute in monetary terms. If individuals choose an alternative to obtain maximum utility, the probability of

_

(2)

exp(Vin)

The log likelihood function is given by:

(3) din is a dummy variable, equal to 1 if individual n chooses profile i, and 0 otherwise. We defined V as a linear function of attributes, where Xk is the k-th attribute, Xm is the monetary attribute and each β is a coefficient: (4)

v _i_ n v

Though the taste parameters are homogeneous in the CL, they vary in the population in the random parameter logit (RPL) model. Train (2003) shows Pin=prob(Vin + εin>Vjn + εjn) the RPL model as below: =prob{sin-sjn>Vjn-Vin In t time period, respondent chooses alternative As McFadden (1974) has shown, if we assume ε as IIDEVI, the choice probability Pin is given by the j . The utility of alternativej in t time period is conditional logit (CL) model: Ujt=βxjt + εjt (5)

Table 1 Characteristics of the respondents in the choice experiment survey (sex, age and occupation).

Age 10-19 20-29 30-39 40-49 50-59 60-69 70-79 80-89 No reply Total

Occupation Office worker Part-timer Selfemployment Full-time housewife Student Unemployed Other No reply Total

Total

Female

Male

No reply

No. of Percent respondents of total

No. of Percent respondents of total

No. of Percent respondents of total

No. of Percent respondents of total

3.4 19.2 14.0 15.3 19.8 16.2 8.8 1.1 2.2 100.0

28 160 117 128 166 135 73 9 18 834

21 100 108 116 150 120 59 6 1 681

3.1 14.7 15.9 17.0 22.0 17.6 8.7 0.9 0.1 100.0

7 60 9 12 15 14 13 2 0 132

5.3 45.5 6.8 9.1 11.4 10.6 9.8 1.5 0.0 100.0

0 0 0 0 1 1 1 1 17 21

0.0 0.0 0.0 0.0 4.8 4.8 4.8 4.8 80.8 100.0

Total

Female

Male

No reply

No. of Percent respondents of total

No. of Percent respondents of total

No. of Percent respondents of total

No. of Percent respondents of total

82 124 71

9.8 14.9 8.5

66 120 50

9.7 17.6 7.3

16 4 20

12.1 3.0 15.2

0 0 1

0.0 0.0 4.8

270

32.3

270

39.6

0

0.0

0

0.0

134 87 38 28 834

16.1 10.4 4.6 3.4 100.0

73 66 27 9 681

10.7 9.7 4.0 1.3 99.9

61 20 10 1 132

46.1 15.2 7.6 0.8 100.0

0 1 1 18 21

0.0 4.8 4.8 85.6 100.0

Sakagami et al.—Consumer preferences regarding organic labels The probability that a respondent chose an alternative sequentially is the logit form. Then the unconditional probability is where

and then the simulated log-likelihood function is introduced. (7)

Downloaded by [154.16.42.67] at 04:30 19 December 2017

Simulated LL=

L where Pˆjt = R1 jt We specifiedf(β) as Normal and the probabilities were estimated through simulation. The parameter of monetary attribute was fixed to calculate the WTP easily (Train 2003). We estimated the parameters of the CL and RPL model using the LIMDEP statistical package version 8.0 (William. H. Greene, Econometric Software, Inc.). The WTP (or compensating variation) which is calculated from CL was used here as a welfare measure when utility function was linear as in (4). WTP was given by (Hanemann 1984; McConell 1995):

(8) 0

exp(V )j

251

βm is the parameter of monetary attribute. V° is the profile before the change and V1 is the profile after the change. We can interpret WTP as the ratio of the coefficient of an attribute to one of monetary attribute when a single attribute change is considered (Rolfe et al. 2000).

RESULTS AND DISCUSSION The statistical results are shown in Tables 2 and 3 and the results of calculated WTP are shown in Table 4. In the results of the CL (Table 2), all of the coefficients were statistically significant. The coefficient on the price was negative. This corresponded with the theory of economic behaviour. We found that the greatest importance was attached to "freshness" and that the next was "source (the region of production)". Respondents attached importance to both traditional attributes in purchasing vegetables in Japan. Respondents preferred certified or labelled organic vegetables to non-certified ones from the fact that they perceived all the parameters of the "certified label" attributes as being positive. They required certification of foods by some organisation and to obtain the benefit of labels reducing asymmetric information. They set an especially high value on the NPO label, followed by government labels (the JAS label), with supermarket labels ranking second and

Table 2 Estimated results of choice experiment for purchasing vegetables using conditional logit model. LL, log likelihood; ρ2, likelihood ratio index; AIC, Akaike Information Criterion; SE, standard error; JAS, Japanese Agricultural Standards; NPO, non-profit organisation. Attributes and levels

Coefficient

SE

-value

Fresh Slightly shrivelled

1.2249 -2.7061

0.0653 0.1081

18.7440 -25.0269

Local products Foreign products

0.3037 -2.7789

0.0666 0.1406

4.5580 -19.7646

0.8110 1.1105 0.6496

0.0680 0.0980 0.0926

11.9331 11.3368 7.0180

-0.0421

0.0012

-34.8081

JAS certification NPO certification Supermarket certification Price LL ρ2 N AIC

-3032.643 0.31613 4044 6081.286

252

New Zealand Journal of Agricultural Research, 2006, Vol. 49

Downloaded by [154.16.42.67] at 04:30 19 December 2017

Table 3 Estimated results of choice experiment for purchasing vegetables using random param2 eter logit model. LL, log likelihood; ρ , likelihood ratio index; AIC, Akaike Information Criterion; SE, standard error; SD, standard deviation; JAS, Japanese Agricultural Standards; NPO, non-profit organisation. SE

Attributes and levels

Coefficient

Fresh Fresh (SD) Slightly shrivelled Slightly (SD)

1.66554 1.51249 -3.54391 1.2864

0.104846 0.082557 0.169663 0.213031

15.8855 18.3206 -20.888 6.03855

Local products Local products (SD) Foreign products Foreign products (SD)

0.412674 0.620971 -3.54738 1.58132

0.085082 0.120103 0.268157 0.335977

4.85031 5.1703 -13.2288 4.70665

1.1625 1.0006 1.46632 0.019873 0.982015 0.270556

0.094404 0.116487 0.120838 0.235963 0.116093 0.272699

JAS certification JAS (SD) NPO certification NPO (SD) Supermarket certification Supermarket (SD) Price

-0.05219

LL ρ2 N AIC

CL

RPL

Fresh Slightly shrivelled

29.1 -64.3

32 -68.5

Local products Foreign products

7.2 -66.1

9 -67.7

19.3 26.4 15.4

22 28.7 19.1

JAS certification NPO certification Supermarket certification

12.3141 8.58977 12.1345 0.08422 8.45885 0.992142 -30.5212

-2835.74 0.3605 4044 5641.48

Table 4 Willingness to pay (WTP) for vegetables calculated by estimated coefficients of the conditional logit model (CL) and random parameter logit model (RPL) in the choice experiment (unit: yen). JAS, Japanese Agricultural Standards; NPO, non-profit organisation. Attributes and levels

0.00171

-value

third. Japanese consumers had more respect for the NPO than for the JAS in spite of it not being a legal standard. We introduced the RPL model to explore these causes more thoroughly. Using the RPL model which allows parameters to vary in the population, the Akaike Information Criterion (AIC) was improved (Table 3); namely, there is

heterogeneity in the data. The RPL had more explanation power than the CL. Although the coefficients on the standard deviation of the parameter of NPO and supermarket were not statistically significant, the coefficient on the standard deviation of the parameter of "JAS certification" was significant. This implied that the parameters of "NPO certification" and "supermarket certification" did not vary and that respondents'preference for "JAS certification" had higher variation. In other words, while some people may rely heavily on JAS or the government, others may not place much trust in JAS or the government at all. Thus, there can be said to be considerable heterogeneity among consumers in valuing the JAS label. Consumers have various impressions of the Japanese government and exercise various levels of trust towards it based on their past experiences. For this reason there is considerable heterogeneity in the recognition of JAS certification. It is in such a situation that the NPO label can come to be favoured. From the results of estimated WTP of RPL as a welfare measure, the attribute of freshness was valued more highly than other attributes (Table 4). Consumers were willing to pay 32 yen (i.e., 15-25% of

Sakagami et al.—Consumer preferences regarding organic labels

Downloaded by [154.16.42.67] at 04:30 19 December 2017

the total price) more for freshness. And they would pay a premium of 8-22% for an organic certification label. Japanese consumers preferred organic foods whose quality has been approved by some organisation. They were also willing to pay 22 yen (i.e., 10-17% of the total price) more for the JAS label. It was clear that most, although not all, of them were willing to pay a premium for JAS organic foods, although not all. They attached a positive value to JAS-certified foods and the JAS certification system. The WTP for a quality label approved by an NPO was higher than for other labels. They were willing to pay 28.7 yen (i.e., 13-22% of the total price) for an NPO label.

CONCLUSION As a result of our CE study, we learned that the top criterion of Japanese consumers in purchasing vegetables was freshness. Our research reconfirmed many similar results found by other attitude investigations in Japan (Ishida et al. 2003; Mineki et al. 2001). Japanese consumers tend to dislike vegetables that are not fresh more than anything else. Such traditional preferences are to be taken into account in organic foods—a result similar to that of Torjusen et al. (2001). Consumers concerned about freshness are willing to pay more for organic foods (Loureiro & Hine 2002). Source (the region of production) is about as important as freshness. In spite of recent increasing imports and declining Japanese food self-sufficiency, Japanese consumers have come to strongly dislike foreign-produced vegetables, preferring the domestic product, especially local ones. From the results that 90% of housewives as respondents were anxious about the quality of imported foods (National Federation of Agricultural Co-operative Associations 2002), this was also confirmed. This may result from concern about the freshness of vegetables, that is, their perceived effect on health. Concern for health is a key issue in the ordinary purchase of organic products (Torjusen et al. 2001). This is probably influenced by recent popular concern about persistent agricultural chemicals in imported vegetables, and BSE. Certification was a smaller factor than freshness and source (the region of production) in vegetable purchasing criteria. However, we found that consumers preferred certified vegetables to ordinary vegetables and they would pay a price premium for them. Despite the fact that JAS quality is legal and

253

should inspire the most confidence, consumers most preferred NPO certification even though it cannot use the "organic" label. Having found that some system of certification is considered desirable, and that the JAS certifying system is rated positively on the demand side, although not all Japanese consumers trust it, the market for JAS organic foods should expand with the settlement of some problems described before. For example, establishing a business model that includes subsidies to JAS organic food producers, and adopting a sales distribution system like that used for ordinary foods, will stabilise the distribution of JAS organic foods and expand the market.

ACKNOWLEDGMENTS The project was funded by the Japan Science and Technology Agency (the study on sustainable society of 2001). We thank Hiroshi Takatsuki, Yasufumi Suzuki, Seiichi Ueyama, Harumi Koizumi, Masako Fukuoka and other project members, and anonymous referees for many useful comments.

REFERENCES Baker GA 1999. Consumer preference for food safety attributes in fresh apples: market segments, consumer characteristics, and marketing opportunities. Journal of Agricultural and Applied Economics 24(1): 80-97. Blend J, van Ravenswaay E 1999. Measuring consumer demand for ecolabelled apples. American Journal of Agricultural Economics 81: 1072-1077. Buzby JC, Ready RC, Skees JR 1995. Contingent valuation in food policy analysis: a case study of a pesticide-residue risk reduction. Journal of Agricultural and Applied Economics 27: 613-625. Champ PA, Boyle KJ, Brown TC ed. 2003. A primer on nonmarket valuation. Dordrecht, Kluwer Academic Publishers. Hanawa A 2001. JA Sninteiseido no kanzensikou to yuukishokuhinsijo. JETRO Sensor 2001.9: 32-36. Hanemann WM 1984. Welfare evaluation in contingent valuation experiments with discrete responses. American Journal of Agricultural Economics 66: 332-341. Holmes TP, Adamowicz WL 2003. Attribute-based methods. In: Champ PA, Boyle KJ, Brown TC ed. A primer on nonmarket valuation. Dordrecht, Kluwer Academic Publishers. Pp. 171-219.

254

New Zealand Journal of Agricultural Research, 2006, Vol. 49

Huffman W 2003. Consumers' acceptance of genetically modified foods in high-income countries: effects of labels and information in an uncertain environment. American Journal of Agricultural Economics 85: 1112-1118. Ishida H, Kawai J, Saito Y, Kishida M, Nakajima F, Okumura M 2003. Study on marketing of organic vegetables. Hyogo Kenritsu Seikatsukagaku Kenkyusyo KenkyuHokoku 17: 183-208.

Downloaded by [154.16.42.67] at 04:30 19 December 2017

Lohr L 1998. Implications of organic certification for market structure and trade. American Journal of Agricultural Economics 80: 1125-1129. Loureiro ML, Hine S 2002. Discovering niche markets: a comparison of consumer willingness to pay for local (Colorado grown), organic, and GMO-free products. Journal of Agricultural and Applied Economics 34: 477-487. Loureiro ML, McCluskey JJ, Mittelhammer RC 2003. Are stated preferences good predictors of market behavior? Land Economics 79: 44-55. Louviere JJ, Hensher DA, Swait JD 2000. Stated choice methods; analysis and application. Cambridge. 402 p. MAFF (Ministry of Agriculture, Forestry and Fisheries of Japan) 2004. The inspection certification system for organic agricultural products (Food Safety and Consumer Affairs Bureau). http://www.maff.go.jp/ (accessed 30 March 2006). McConnell KE 1995. Consumer surplus from discrete choice models. Journal of Environmental Economics and Management 29: 263-270. McFadden D 1974. Conditional logit analysis of qualitative choice behavior. In: Zarembka P ed. Frontiers in econometrics. New York, Academic Press. Pp. 105-142. Mega A, Kazuno K 2003. The research trends and investigation of consumers' consciousness about agricultural products cultivated by organic fertilizer. Bulletin of University of Yamanashi, Faculty of Education Human Sciences 5(1): 30-39.

Appendix

Mineki M, Sakamoto K, Isihi Y, Fujii A, Niizawa Y, Kawai T, Kanatani A 2001. The relationship between food market and consumer behavior in food environment. Distribution and consumption of organic agriculture products. The Japan Society of Cookery Sciences 34(2): 214-223. Ministry of Internal Affairs and Communications 2002. The questionnaire result about food labelling (outline). http://www.soumu.go.jp/s-news/ (accessed 30 March 2006). National Federation of Agricultural Co-operative Associations 2002. Shufu no 'Shoku to Noh' heno Kansindochousa. http://www.zennoh.or.jp/ZENNOH/FOODS/ (accessed 31 March 2006). National Federation of Agricultural Co-operative Associations 2003. Shufu no 'Shoku to Noh' heno Kansindochousa. http://www.zennoh.or.jp/ZENNOH/FOODS/ (accessed 31 March 2006). Nihon Keizai Shimbun 2005. 'Shoku; Anshin to Kakaku', June 28. Ohyama T 2003. Yuukishokuhin system no Kokusaitekikenshou. Tokyo, Nihon Keizai Hyoronsha. Rolfe J, Bennett J, Louviere J 2000. Choice modelling and its potential application to tropical rainforest preservation. Ecological Economics 35: 289-302. Roosen J, Fox JA, Hennessy DA, Schreiber A 1998. Consumers' valuation of insecticide use restrictions: an application to apples. Journal of Agricultural and Resource Economics 23(2): 367-384. Torjusen H, Lieblein G, Wandel M, Francis CA 2001. Food system orientation and quality perception among consumers and producers of organic food in Hedmark Country, Norway. Food Quality and Preference 12: 207-216. Thompson GD 1998. Consumer demand for organic foods: what we know and what we need to know. American Journal of Agricultural Economics 80: 1113-1118. Train KE 2003. Discrete choice methods with simulation. Cambridge, Cambridge University Press. Yoshida T 2005. Tabetemo Heiki?BSE to Shokuhinhyouji. Tokyo, Shueisha.

An example of choice experiment question.

Question: There are three types of spinach which have different freshness, source, certification and price. Please check the one alternative (A-C and D) which you want to purchase.

Freshness Origin Certification Price (yen)

A

B

C

D

Fresh Foreign JAS 160

Standard Local NPO 190

Standard Foreign Supermarket 130

No choice