Willingness to pay for pesticide reduction in the EU: nothing but organic?

1. Introduction

Measuring the value consumers put on pesticide reduction is of interest in order to assess the variation of surplus in welfare analysis and to evaluate the potential market for farmers who could take advantage of the growing demand for pesticide-free products. In this context, it is important to know whether the different certifications and corresponding signals are clearly identified by consumers and to assess the corresponding potential willingness-to-pay (WTP). The main objective of this paper is to investigate the variations of these premiums in different European countries and to compare different kinds of signals according to the authority that guarantees the pesticide reduction (public bodies, producers or retailers).

There are many different certification programmes, which are signalled to consumers through explicit labelling or by means of a logo appearing on a sticker. Organic certification is probably the most well known. Since 2009, it has been standardised at the European level (regulation, labelling and logo). The framework of Integrated Pest Management (IPM) is also usually considered to be the agricultural production system that integrates pesticides/herbicides management. This scheme could be an alternative for consumers who care about pesticide reduction but are not, for whatever reason, ready to buy organic products. However, IPM incorporates a large range of practices and does not explicitly state the degree of pesticide reduction at the farm level. Indeed, the ‘good agricultural practices’ necessary for IPM are tailored according to regional weather conditions and sector of agricultural activity. This is why producers have not yet been able to send a clear signal about IPM to consumers.¹

To know whether there is a premium for pesticide reduction and whether it varies according to the kind of guarantee, we have elicited consumers' WTP for apples to which are attached different kinds of certification concerning pesticide use. This investigation has been conducted using the same economic experiment with 408 subjects sampled in France, Greece, the Netherlands and Portugal. Different systems of certification including pesticide use reduction have been compared with no certification (a ‘regular product’): organic, ‘IPM generic’ certification without brand name, retailers' label and PDO, which also implement IPM in agricultural practices. The focus has been placed on the impact of information provided to consumers concerning pesticide use, and sensory characteristics have been controlled as previous work showed that they have often more impact than environmental considerations and food safety in consumers' final decision.

This paper contributes in several ways to the literature on the impact of information and labels on consumers' WTP for a reduction in pesticide use. First, we show that European consumers have fairly homogeneous behaviours when valuing different certifications: they rank organic apples first, regular last and IPM certifications in between (e.g. compared with the regular product, the average premium for organic certification is 50.5 and 24.5 per cent for generic IPM certification). Moreover, except in the case of women's premium for organic products, we do not observe any systematic effect of socio-demographic characteristics of consumers (age, income, education, etc.) to explain these valuations. Second, our results suggest that consumers may not be well aware of what the different logos stand for. Surprisingly, despite its popularity, this lack of knowledge also includes the organic certification. When detailed information is provided, the value of organic products increases as people learn that they are produced without any use of synthetic pesticides. Conversely, the value of regular products decreases, as they cannot support any claim on the improvement of the production process.² Between these two limits, IPM products can be considered either a substantial or insignificant improvement, according to whether consumers are inclined to see the half-full or the half-empty glass. This may explain the limited premium for IPM apples. Third, we show how various combinations of labels in the context of IPM can achieve superior results in terms of WTP (e.g. PDO, which often achieves results comparable with organic but not deriving from pesticide reduction) or otherwise not be particularly valued by consumers (e.g. retailers' label).

Lastly, our findings confirm that environmental certification is not a guarantee of price premium when consumers experience taste disappointment. Indeed, our data suggest that consumers have strong sensory expectations specifically towards organic produce, which reinforces the need to control for the influence of organoleptic characteristic in experimental auctions of food products.

In Section 2 of the paper, some background elements from the literature concerning consumers' WTP for pesticide reduction in food products are presented. Section 3 introduces our experimental design (experimental protocol and incentive procedure to elicit WTP) and gives specific details of the experiment in each country. Section 4 reports the results. Section 5 concludes, giving the main implications of our results for production decisions and public regulation, and discussing the challenges concerning alternatives to signal pesticide use reduction.

2. Background

From the abundant literature concerning pesticide reduction, we propose to consider two groups of papers that focus on consumers' WTP for two categories of certified food products: (i) those ensuring the total elimination of certain pesticides (starting from organic and ‘pesticide residue-free’ PRF certifications, which is evaluated primarily in the United States), and (ii) those signalling the decrease in pesticide use (starting from IPM and eco-labels). Note that most of these studies on organic and PRF products find significant influence of socio-demographic characteristics on the level of the premium. Indeed, the authors point out that WTP and premiums for organic products are significantly related to income, education, gender or age (see Byrne et al., 1991; Posri, Shankar and Chadbunchachai, 2007; and also Thomson, 1998, who tempers these results for organic certification).

The first group assesses consumers' WTP for certified products using the price of the regular counterpart as a reference. Mostly based on contingent valuation or interviews, these studies obtain different price premiums depending on the country in which the investigation is carried out, the methodology used and the date of the survey.³ Works using experimental auctions (non-hypothetical experiments) place emphasis on the control of information given to consumers to assess the premium for organic products. Gil and Soler (2006) observed that information about regular products (giving the ‘reference price’) increased the perceived value of the organic product (for olive oil in Spain). Their results also show that only the consumers that have already bought organic products were willing to pay a price premium and only 5 per cent of them would be willing to pay the corresponding market price. However, an important issue, pointed out previously by Roosen et al. (1998), concerns the strong influence of cosmetic damages. These authors investigate the impact of specific insecticides' elimination on consumers' WTP for apples and show that the elimination of insecticide increases WTP by about 50 per cent, while cosmetic damage decreases average WTP by 63 per cent. Along the same lines, Yue, Alfnes and Jensen (2009) show how the appearance and cosmetic damages affect consumers' WTP of organic fresh products more than for the regular product.

Concerning the second group of papers, it is worth mentioning that IPM products have received relatively little research attention. Anderson et al. (1996) explain this by the low notoriety of IPM, deriving from the fact that a small proportion of consumers are actually informed about the meaning of this certification. Govindasamy and Italia (1998) and Govindasamy, Italia and Adelaja (2001) evaluate the demographic characteristics that could influence consumers' WTP premium for IPM: females, younger individuals and those who frequently purchase organic produce appear to be among the most likely to pay a premium. Consequently, these authors suggest a competition between IPM and organic produce, without, however, proposing a WTP comparison of these two certifications in the same survey. They also underline the importance of the pricing of IPM products to foster the development of their market share. More recently, Tonsor and Shupp (2009) evaluate consumers' WTP for products marketed with ‘sustainably produced’ labelling claims. They conclude that US consumers are not willing to pay a positive premium for tomatoes or apples labelled as ‘sustainable production’ because this information is vague and not clearly associated with production practices. Therefore, the authors propose to conduct additional experiments designed to analyse label valuations when alternative forms and levels of information are provided to consumers. Using experimental auctions, Bougherara and Combris (2009) investigate the premium of French consumers for an eco-labelled bottled orange juice. The premium obtained was driven by selfish or altruistic motives, other than food taste or safety. Combris et al. (2010) measured the effect of information on the WTP for pears in Portugal, and the trade-off between taste and the pesticide use reduction according to IPM. The main results show that information on pesticide reduction instantly influences consumers' WTP. However, it appears that sensory intrinsic attributes related to taste beat the guarantee of food safety in driving buying behaviour.

Based on this review of the literature, our own paper proposes to assess consumers' WTP for both organic and IPM certifications using the same experimental protocol in different countries.⁴ According to the arguments summarised inter alia by Lusk and Shogren (2007), we choose to use the methodology of experimental auctions in order to control more precisely the impact of information on pesticide use reduction and to obtain a full distribution of preferences by heterogeneous consumers. Moreover, the experimental protocol was designed to carefully control the role of product taste that is clearly a major concern for food consumers.

3. Experimental design

We conducted experimental sessions in four European countries: the first set of sessions was carried out in Portugal (Lisbon) during April 2009; the second set was carried out in France (Dijon) in May 2009; the third set was conducted in Greece (Thessaloniki) in February 2010; and the last one was carried out in the Netherlands (The Hague) in October 2010. The main features of the protocol are common to all countries. However, some minor changes have been introduced in order to either improve the protocol or to adapt it to each country's particularities.

3.1. Subjects

For each country, a common set of criteria was imposed. To be selected for the experiment, subjects had to be usual consumers and buyers of apples. The recruiters asked them the price they usually paid for 1 kg of apples. If the given price was unrealistic (greater than EUR 5), then the subject was not selected. People who participated in more than three studies (marketing or consumption studies) in the last six months were rejected. The sampling process took into account the age range of participants and was designed to limit the proportion of students and retirees. If all the criteria were satisfied, recruiters presented the study as a research on food preference ending by a sale. At the end of the phone interview, the consumers were told that they would participate in an experimental session for which they would earn a lump sum between EUR 10 and 30 for participating (according to local practice). All the participants received a convocation letter with explanations of the incentive mechanism, that is the random selling-price procedure used at the end of the session for the actual sale of 1 kg of apples.

3.2. Products

We chose to use apples for our experiments. Apples are the most widely produced fruit and also the most widely consumed in many European countries throughout the year. This product has already been much used in previous experiments (see, for example, Baker, 1999, and Loureiro, McCluskey and Mittelhammer, 2002). The familiarity of this product is in itself an advantage which permits eliciting behaviours of general interest while simplifying the experiments, apples being easy to handle and store, and being sold all the year along.

In each country, apples produced according to three different types of pest management, corresponding to different levels of chemical pesticide use, were proposed to the participants. The first level of pesticide use is defined by the European minimum quality standard. In this category, the apples, named ‘regular’ (REG) apples in the rest of the paper, are produced according to the European regulation for pesticide use. The second level of pesticide use is defined by a controlled reduction of pesticide use compared with the existing legislation. In our experiment, we chose to test three of these certifications: Finally, organic apples (ORG), for the production of which chemical inputs are prohibited, represent the third and lowest level of pesticide use. In the Netherlands, an additional organic alternative was tested in the experiment. This alternative, called ‘Organic+’, forbids any use of pesticides, even those of organic origin.

• Integrated Pest Management (IPM) is a generic neutral certification, without explicit statement of a public or private brand.

• Protected Denomination of Origin (PDO) is a guarantee that producers are engaged in a controlled process with production rules including IPM. Note that the environmental requirement is not communicated to the consumers.

• Retail label (RET) is the guarantee that retailers require their producers to follow IPM rules.

Table 1 presents the different apples and certifications proposed to participants in each country. Apples were selected according to national preferences and supply constraints.⁵

To signal the characteristics of the apples, we used stickers with logos. The regular apple had no sticker. However, in Greece, we added a regular apple with a sticker signalling the name of the variety. The objective was to control for a possible ‘sticker effect’. In France, the available organic apples were smaller than the other ones; so to control for the ‘size effect’, we included a small apple without any environmental characteristic. The IPM alternatives were signalled with the logos currently used in each country. To signal the retailers' brand, we used the actual logos available on national markets.⁶ Finally, to signal the ORG apples, we choose the logo commonly used in each country (national logos in France, Greece and the Netherlands, and the European logo in Portugal).

3.3. Procedure

For all experiments, sessions took place in sensory analysis rooms and gathered between 8 and 20 participants. Upon arrival, participants received a show-up fee as a compensation for the time spent in the session. Then participants were not endowed with extra cash to pay for the apples they might buy. Each session began with an oral presentation of the procedure. At the outset of each session, an experimenter presented the objective of the experiment as ‘an evaluation of apples of the same variety but coming from different production systems’. Participants were informed that they would carry out their evaluations in different information contexts. In every one of them, they would have to indicate the maximum buying price they would pay for 1 kg of each of the tested apples given the information they had on them. Participants were also told that a real sale would take place at the end of the session. Following the BDM procedure (Becker, DeGroot and Marschak, 1964), one of each participant's evaluations would be randomly chosen and compared with a random price drawn from a uniform distribution actually wider than the distribution of market prices. Then the participant would have to buy 1 kg of the corresponding apple at the random price in case it was lower or at most equal to her evaluation. A trial random-price sale took place at the beginning of each session, to check that the procedure was properly understood by all participants.

After all explanations had been given and questions answered, the experiment started. Four steps were common to all experiments. Steps were defined according to the information given to participants to evaluate the apples. Participants did not know in advance the series of steps they would go through. They only knew that they would have to carry out several evaluations and that each one would be made independent by a random selection before the actual sale. So, strategic WTP revelation is minimised because each evaluation, even those made with limited information, could be the one selected for the final sale. Participants who did not like a specific type of apple, or suspected unwanted characteristics, were always able to indicate a buying price of zero.

At the beginning of the first step of the experiment (‘Tasting step’), the different apples were presented simultaneously to each participant. Participants did not have any information about the apples except for the variety's name. To compare the apples, the participants could only look and taste them. As explained above, all the participants had to evaluate five apples in Portugal and the Netherlands, six in France and Greece (see Table 1).

During the second step (‘Sticker step’), the different apples with their stickers were presented simultaneously to each participant. No information was given to the participants about the actual meaning of the stickers, and they were not allowed to taste the apples. In Portugal, five variants of the variety which was preferred in the first step were used for the second, third and fourth steps.

At the beginning of the third step (‘Information step’), each participant received an information sheet explaining the meaning of the different logos. For REG apples (with or without sticker), the information sheet stated that these apples were produced according to national rules regarding pesticide use. For IPM apples, we specified who guarantees that pesticide use had been reduced: public authorities (generic IPM in Portugal, PDO in France and Greece), producers (generic IPM in France, Greece and the Netherlands, PDO in Portugal and the Netherlands) or retailers. Concerning ORG apples, the information sheet indicated that public authorities guaranteed that no chemical pesticides had been used (no chemical and no organic pesticides for ORG+ in the Netherlands).

During the fourth step (‘Full information step’), each participant evaluated each apple with the same information as in the third step but they were also invited to taste the apples before giving their evaluations.

During each step, apples were presented together on a tray and evaluated simultaneously.⁷ A maximum purchase price for each apple was recorded for every participant at the end of each step. When a new step started, participants could not go back or change the prices they had given.

4. Results

4.1. Samples and data

Four hundred and eight adults, aged between 18 and 73, participated to our experimental sessions in four different countries. Table 2 reports summary statistics of socio-demographic variables common to all locations. As expected, some differences appear in our recruited samples. Some of these differences simply reflect specificities of each country's population. For example, income is lower and family size is larger in the Greek and Portuguese samples (see Table 2). This is in line with national statistics, showing that average monthly income per capita was EUR 891 in Greece and EUR 1,034 in Portugal compared with EUR 1,567 in France and EUR 1,667 in the Netherlands.⁸ According to the same national statistics, average family size was 2.8 in Greece and 2.6 in Portugal, but 2.3 in France and 2.2 in the Netherlands. Some other differences (over-representation of young people in the Greek sample and of women in the Dutch sample) result from local recruitment difficulties in spite of the recommendation we made to all agencies to avoid recruiting too many economically ‘inactive’ or part-time active people.

In all experiments and locations, participants followed the instructions and indicated buying prices consistent with local prices. Each Portuguese participant revealed 21 prices (six during the tasting step and then five at each of the remaining three steps), the French and the Greek participants revealed 24 prices (six apples and four steps) and the Dutch participants revealed 20 prices (five apples and four steps). So, we collected a total of 9,090 prices, ranging from EUR 0 to 5 for 1 kg of apples, with a median of EUR 1 and a mean of EUR 1.12. On average, 6.3 per cent of the prices were null, but no participant systematically refused to buy. They all gave some positive prices: the total per participant ranging from 7 to 24 positive prices, with a mean and a median of 20. The mean of the positive prices across all steps and apples was EUR 0.91 in Portugal, EUR 1.38 in France, EUR 1.15 in Greece and EUR 1.28 in the Netherlands.

4.2. WTP for pesticide reduction: premiums for IPM and organic apples

The main point of this paper is to analyse whether, and how much, European consumers are ready to pay for pesticide use reduction when they are fully informed of production methods. So, we started with the analysis of the WTP at the third step of each experiment, when participants had been informed of the meaning of the different stickers attached to the apples in terms of pesticide use. First, we tested the difference in WTP between REG, IPM and ORG. To test the hypothesis that WTP increases when pesticide use decreases, we used both parametric paired t-tests and non-parametric Wilcoxon tests. Table 3 reports the results of these tests, comparing IPM and REG apples (3.1), ORG and REG (3.2) and finally ORG and IPM (3.3).

For all three comparisons and all experiments, both tests lead to the same conclusions. The first section of Table 3 shows that WTP for IPM apples is always significantly higher than WTP for REG apples (24.5 per cent higher in average), except in the Dutch experiment. The second section shows that WTP for ORG apples is always higher than for REG apples (50.5 per cent in average). The third section also displays a positive difference between ORG and IPM apples in all four countries.

The specificity of the results obtained for the IPM variant in the Netherlands may derive from differences in the set of apples that were tested. Unlike the other three experiments, the REG apple in the Dutch experiment was identified with a sticker indicating the apple variety (‘Elstar’). To check for a possible effect of this sticker, we ran another experiment under the same conditions, with a smaller sample (n = 51), and found a significant difference (p = 0.026 for the paired t-test, and p = 0.001 for the Wilcoxon test) between a REG apple without a sticker and an IPM apple with the same sticker as in the first experiment. Another difference between the Dutch and the other three experiments was the presence of an ORG+ apple in the set of tested variants. These apples are produced with no pesticide at all, either synthetic or natural, while standard organic production uses natural pesticides. The ORG+ appeared clearly as the highest valued option (the mean WTP was EUR 1.46/kg compared with EUR 1.33/kg for the standard ORG). The ORG was the second best option.

To compare the WTP for pesticide use reduction across countries and control for differences in socio-demographic characteristics of participants, absolute premiums (difference in WTP between IPM and REG, and between ORG and REG) have been regressed on dummy variables for country, gender, age, income and household size. The estimation of premiums for IPM and ORG apples, relative to REG, may be affected by censoring issues if participants did not disclose the full range of their WTP. Censoring can affect both WTP used to calculate the premiums and could concern lower and upper values as well.

Censoring from below (at zero) is a frequent issue, even when selected participants are usual consumers of the tested products. Thus, bidding zero does not mean that they are not engaged (and actually no participant bid zero for all apples), it means that they strongly reject the corresponding variant. It is thus not meaningless to suppose that some of the participants actually have a negative value for some of the variants. In this case, premiums are either under- or overestimated according to which of the two prices used to calculate the difference is censored at zero. If the price of the REG apple is zero and the price of the IPM (ORG) apple is positive, the premium may be underestimated and thus right-censored. If the price of the IPM (ORG) apple is zero and the price of the REG is positive, then the premium is negative and left-censored. If both prices are censored at zero, the ‘true’ premium can take any positive or negative value and as such drops out of the likelihood function (see Lusk et al., 2004).

Censoring from above at the market price of a field substitute could also be an issue when there is a clear outside option with a well-known price, or a price made public during the experiment as in the Hoffman et al. (1993) study discussed by Harrison, Harstad and Rutström (2004). In this case, the censoring of individual values higher than the field price of the substitute entails a downward bias in the mean value of the WTP elicited in the experiment. However, in our experiment, such a reference price is unlikely to exist given that apples are supplied in the market in many different varieties and qualities and that their prices may vary significantly according to shopping place and season. So, as underlined by Coller and Williams (1999), who first raised this concern, it may be difficult for the experimenter to know which reference price has been considered by participants, as it can be unclear for participants themselves what the prices of field substitutes are, when compared with the different products they evaluated during the experiment.

We thus chose to account only for censoring at zero of participants' values for the different apples. A Tobit model with left and right censoring was estimated to accommodate the different cases of censoring of the premiums. The same has been done for relative premiums (the ratio of WTP for IPM to WTP for REG, and the ratio of WTP for ORG to WTP for REG). Results are reported in Table 4.

The third column of Table 4 shows that absolute premiums for ORG apples are not different in France, Greece and Portugal, but are significantly lower in the Netherlands. In relative terms (fourth column), premiums are significantly higher in Greece and Portugal compared with France (the reference), and significantly lower in the Netherlands. Premiums for ORG apples show some sensitivity to socio-demographics: the absolute premium is higher for women and participants above the second tercile of income. The relative premium is also higher for women. Except for women, who will pay significantly more for ORG apples, there is no systematic effect of age, income or household size.⁹ Compared with ORG, IPM certification induces far fewer differences. The first and second columns of Table 4 show that socio-demographic characteristics of participants have no impact on either absolute or relative premiums for IPM apples. For reasons explained above, Dutch participants were not willing to pay more for IPM apples either in absolute or relative terms. The only remaining difference is a significantly lower absolute premium in Greece compared with the French reference.

Turning to the different variants of IPM apples, section 4 of Table 3 shows that in all locations where the retailers' option has been proposed (France, Greece and Portugal), the retail label is less valued than the generic IPM label. This may explain why private labels do not communicate much on pesticide reduction. Concerning PDO apples, section 5 of Table 3 reveals a positive premium compared with generic IPM apples in France, Greece and Portugal, but not in the Netherlands. In this last case, PDO certification actually does not exist, but, as explained above, we decided to create a fake label referring to a popular production area to check whether Dutch consumers could be responsive to origin. Clearly they are not. In the other three countries, origin seems to have a value in itself independent of the issue of pesticide use reduction. There is a significant premium for PDO apples compared with generic IPM apples, and, moreover, as the next section will show, it is not influenced by information on pesticide use.

4.3. Impact of information on pesticide use reduction

In the previous analysis, participants' WTP has been evaluated after they received information about the use of pesticides in the different production systems. In fact, such information is not readily available to consumers. So, by assessing its impact on WTP, our objective is to know whether better-informed consumers could contribute to the reduction of pesticide use by rewarding producers for their efforts. To answer this question, we compared the WTP before (step 2) and after informing participants on pesticide use in the production of the different apples (step 3). As previously, we used both parametric and non-parametric tests. Results are reported in Table 5.

The first thing to notice is that giving information on pesticide use to participants decreased the WTP for REG apples in all experiments. This decline in WTP is always significant, except in the Dutch experiment. The second important result is that information had no impact on WTP for IPM apples. Tests always conclude to no difference in WTP, before and after information has been provided to participants. The third section of Table 5 shows a general increase in WTP for ORG apples when participants are fully informed. This increase is highly significant in all countries and experiments.

The last section of Table 5 confirms that in France, Greece and Portugal, where PDO certification is well known, information on producers' commitment to reduce pesticide use has no effect on WTP. It is also interesting to observe that in the Netherlands, where PDO is unfamiliar, the same information significantly increases WTP. An interview with French participants showed that only 31 per cent of them knew about the requirement of controlling pesticide use to obtain the PDO certification Pomme du Limousin. In spite of that, releasing information on pesticide monitoring did not increase their WTP, which tends to prove that their interest in PDO products is more linked to the knowledge of origin than to a guarantee on agricultural practices.

4.4. Impact of sensory characteristics

One last and important question is to know whether sensory characteristics of apples could affect consumers' WTP for pesticide use reduction. As mentioned in Section 2 of the paper, cosmetic defects have been shown to impact WTP for pesticide reduction and for ORG apples. Taste being another major sensory concern when choosing food products, an unpleasant taste may reduce, or even cancel out, the premium consumers would be willing to pay for increased safety.

Though our experiments were not designed to test conflicts between taste and pesticide reduction, we considered it important to control for taste influence in the assessment of WTP. So, in the first step of the experiment, we measured WTP after blind tasting in order to evaluate the perceived differences between apples independent of any information other than taste. Then, during steps 2 and 3, participants were given a new set of the same apples in a different order and evaluated them without being able to establish a connection with those they tasted during step 1. Consequently, WTP at step 3 reflects individual values built from a visual aspect, objective information and expectations on taste resulting from step 1. Finally, in order to assess individual values with full actual information, participants were asked to taste each apple and give their final WTP. Tasting the different apples while knowing all their characteristics led to variations in WTP compared with the evaluations made at step 3. Relative variations for REG, IPM and ORG apples in each country are shown in Table 6.

A large majority of these variations (9 out of 12) are significant, and in particular WTP for ORG apples decreases significantly in all countries, from 11.2 per cent in France to 19.5 per cent in the Netherlands. This means that, at step 4, participants eventually found out that these apples tasted worse than they expected when they evaluated them at step 3. To assess to what extent these changes in WTP were related to difference in taste, we used the WTP elicited at step 1, which gives a measure of sensory perception not affected by previous knowledge of the apple's type.

The first three lines of Table 7 show that a majority of coefficients are significant and positive, which means that WTP decreased between steps 3 and 4 when taste was perceived as below the best. Significant coefficients range from 0.29 to 0.58, indicating that the impact on WTP due to taste alone is finally reduced when tasting occurred with full knowledge of the apples' characteristics. The impact of taste on WTP for ORG apples is always significant, which can be viewed as confirmation that consumers will put lower values on organic products if they are disappointed by their sensory characteristics.

On average, the WTP for ORG apples increased by 24.4 per cent when pesticide information was released and then decreased by 15.2 per cent when participants were allowed to taste apples. This means that informing consumers about pesticide use reduction should not be taken as a guarantee that they will definitely place more value on the corresponding product. Consequently, strategies to reduce pesticide use should not overlook sensory issues.

5. Conclusion

The results obtained in this paper suggest a relatively homogeneous behaviour of European consumers. It appears that, in all four countries, there is a significant premium for apples produced with reduced pesticide use, and that the consumers' behaviours vis à vis the quality signals, in the different situations of information, are similar. In all our experiments, the highest WTP after information on pesticide use always goes to ORG apples. Except in the Netherlands, there is also a premium for IPM apples compared with REG apples, which is significantly less than the premium for ORG apples. However, sensory characteristics may change the magnitude of WTP differences due to information alone. In particular, WTP for ORG apples appeared to be sensitive to taste in all countries. This did not reverse the hierarchy of WTP, but confirmed that taste is always a major determinant of choices.

Additionally, we show how informing participants of pesticide use reduction has a highly significant impact on the WTP for the ORG apple and has no impact on the WTP for the IPM apple. As explained above, we obtain a significant decrease of the WTP for the REG apple. Therefore, while the labels may convey positive messages to consumers about the production conditions, they may simultaneously stigmatise the conventionally produced product by highlighting perceived problems. The net economic result for producers can be negative since consumers may decrease their WTP for the regular product that has the largest market share.

Organic products are well valued by all consumers, and, as our experiments show, this valuation may increase, should information on the various certifications existing on the market be disseminated. Consequently, and without underestimating sensory issues, organic products may appear as a safe haven for under-informed consumers. However, in Europe, market prices of organic products are generally twice as high as prices of their regular counterparts. This gap is larger than the premium consumers are willing to pay, which is the main reason for the small market shares that are observed for organic products in Europe (generally below 5 per cent). Our results could be further used along with detailed information on cost differential between organic and conventional products to assess potential market shares under different pricing hypotheses.

The partial reduction of pesticide use permitted within the framework of IPM does not lead to the same clear results as those we obtained for organic products, which guarantee no use at all of chemical pesticides.¹⁰ However, contrary to the ORG case, market prices of IPM-certified products are more in line with the WTP we have elicited. Consequently, purchasing IPM-certified products could increase consumers' surplus and lead more easily to mass consumption. This point could now be tested along with the comparison of surpluses according to different levels of market prices with the objective of comparing the efficiency of different policy alternatives combining taxes on pesticides or subsidies to support a decrease in their use.

Another issue is to clearly assess the best way to ensure the reduction of pesticides using IPM. Should certifications from public authorities, producers and retailers be used? Or, is it preferable to include another kind of referential incorporating pesticide monitoring in requirements from farmers? Our work brings some results on that point, showing, for instance, that retailers are not seen as a trusted third party to guarantee the reduction of pesticide use within the IPM frame. In the same context, we have shown that WTP for PDO certification is almost equal to that for ORG (the WTP for ORG apples is only 8 per cent higher than for PDO), while information about the inclusion of IPM rules in the PDO certification does not add more value to the PDO signal. Indeed, our results show that in countries where PDO are familiar to consumers, they are valued for themselves (i.e. for origin rather than for control of pesticide use) and their value for consumers is often close to that of ORG products.¹¹

Concerning socio-demographic influences on WTP, apart from the higher WTP of women for ORG apples, our results do not reveal any other systematic influence (there is no gradient for age, income or education, for example). This confirms previous results concerning women's WTP, a premium for products with a guarantee of less pesticide use, but also suggest that other socio-demographic influences are less significant than they appeared in previous studies. This could mean that all categories of consumers show increasing interest for ORG products, and that there is a potential market that may include a much larger target than the initial users. Within this subject, an interesting question is to understand how organic product consumption progressively spreads over the market, and to better characterise the ‘organic food consumer’ of today and tomorrow, a research issue already on the agenda of sociologists (see Hughner et al., 2007).

A last point concerns public information on pesticide use. Our multi-step protocol was designed to test the impact of information clarifying the way producers monitor pesticide use. Releasing information on certified products had two unexpected effects: a decrease in WTP for REG apples and an increase for ORG apples. This raises the issue of public control of information on those product characteristics which are meaningful to consumers. Would organic produce become a safe haven if consumers became conscious of a sanitary hazard? Including risk aversion in further studies would be of great interest, given that consumers are very sensitive to negative information.

Acknowledgements

The research leading to this article has received funding from the European Community's Seventh Framework Programme (FP7/2007-2011) under grant agreement ‘TEAMPEST 212120’. The authors would like to thank particularly Konstadinos Mattas, who initiated and coordinated the project. They would also like to thank Manuela Berjano, Raquel Maia, Caroline Hanus and Jonathan van't Riet for the valuable assistance they have given for the setting of the different experimental auctions.

References