Consumer stated preferences for dairy products with carbon footprint labels in Italy

Abstract

Carbon footprint (CF) labels on agri-food products represent one of the most important tools to convey information to consumers about the greenhouse gases emissions associated with their purchase behaviour.

Together with the growing interest of consumers in CF labels, the subject has gained attention also in the scientific literature, and formal evaluations of consumer response to carbon labelling have been published. Studies in this area aim at analysing consumers’ preferences for buying products with a lower CF label or their willingness to pay (WTP) for these products.

The objective of this paper is twofold. First, the study proposes a review of the literature that so far has analysed consumer WTP for CF label, focusing on Italian consumers. Second, it uses the results of two surveys of consumers’ attitudes towards dairy products with a lower CF label to analyse the factors determining a positive stated WTP. Results point out that a positive WTP for lower CF products is more likely to be declared by respondents who believe that buying products with less environmental impact can combat climate change. Conversely, highly price-sensitive consumers are less likely to be willing to pay more for CF-labelled products.

Introduction

Climate change mitigation is one of the key environmental goals of agricultural production worldwide (Gerber et al. 2013). Moreover, in Europe, climate change mitigation objectives and the contribution that agriculture is expected to provide have reached the top of the political agenda (European Commission 2016). Climate action is one of the main priorities of the Common Agricultural Policy (CAP) and agricultural greenhouse gases (GHG) emissions’ mitigation has become both an objective of the new architecture of the first pillar payments and a focus area of the actual Rural Development Policy programming period (Council of the European Union 2013a, 2013b).

According to many studies in this field, however, supply-side options alone, i.e. options that tackle production aspects of GHG mitigation, are not sufficient to reach the ambitious mitigation targets set by European and international climate policy agenda (European Commission 2011, 2016). In addition, though the most cost-effective ways to reduce GHG are carbon taxes and cap and trade systems (Nordhaus 2013; Stern 2007), these economic instruments are unlikely to be implemented in the near future in the agricultural sector, both in the EU (Coderoni and Esposti 2018) and in the United States (Shewmake et al. 2015). Thus, demand-side solutions to climate change, which consist of more sustainable consumption patterns, are becoming important tools to curb agricultural GHG emissions (Garnett 2011; Bajželj et al. 2014; Armel et al. 2011; Brunelle et al. 2017; Creutzig et al. 2016; de Boer et al. 2016).

In this respect, the so-called “sustainable labels”, i.e. types of labels that are designed to convey to the consumer concepts related to all the facets of sustainability, are the most common tools supporting changes in consumption patterns (Vermeir and Verbeke 2006; Zander and Hamm 2010). When sustainable labels try to show to consumers the overall impact of the product converting it into a standardised measure of carbon dioxide emissions, they are referred to as “carbon footprint” (CF) labels. CF labels in practice indicate the quantity (in grams) of carbon dioxide equivalent (CO_2e)¹ emitted into the atmosphere throughout all the life cycle of a product or service, which comprises production, transport, transformation, distribution and purchase (Sander et al. 2016). The rationale for these labels, when applied to food products, is that they may help to orient the consumer towards buying more GHG saving agricultural products and thus mitigating agriculture’s contribution to global warming.

Despite the potentially relevant role of demand-side options in tackling climate change, there have been a few consumer studies on WTP for carbon footprint labels (Hoek et al. 2017), especially for Italian agriculture, where the bulk of the empirical literature has focussed on the potential and effectual role of the production processes to mitigate agricultural GHG emissions.² Thus, it would be essential to analyse consumers’ preferences for purchasing products with a label showing a lower CF, to understand what drives their choices, and to recognise to what extent there is a mitigation potential deriving from Italian consumers’ choices for the Italian agricultural sector.

In this context, the objective of this paper is twofold. First, we review the literature that until 2018 has analysed consumer preferences and WTP for CF label, with a focus on Italian consumers. Second, we illustrate some of the results of two separate pilot surveys aimed at detecting whether consumers state a positive WTP for dairy products with a CF label.

The remainder of the paper is structured as follows: Section 2 introduces some definition of sustainable labels, specifically referring to CF; Section 3 presents the literature review; Section 4 shows the case studies analysed, while Section 5 presents and discusses the results of the analysis; Section 6 finally proposes some concluding policy remarks and future research guidelines.

Carbon footprint labels in the agri-food sector

According to Miranda-Ackerman and Azzaro-Pantel (2017), “New consumer awareness is shifting industry towards more sustainable practices, creating a virtuous cycle between producers and consumers enabled by eco-labelling”.

This consumer awareness is the foundation of sustainable consumption, which is grounded in a decision-making process that takes into account not only individual needs and wants, but also their social responsibility. In fact, as De Pelsmacker et al. (2005) have found, when dealing with sustainability concerns, an important driver for change is the inclination of the “ethical consumer [that] feels responsible towards society and expresses these feelings by means of his purchase behaviour”.³

The concept of sustainability has deeply evolved from the primer environmentalist approach (Kumar et al. 2012) and now it comprises, in its most widespread use, three different aspects: the economic, the environmental and the social one (Vermeir and Verbeke 2006). Sustainable products are those products whose characteristics respect one or more of these aspects (Vackier et al. 2002).

Eco-labelling, or environmentally sustainable labels, are a means to inform consumers of the environmental performance of either the products or the production systems they come from, and they can also inform the consumer on measures taken by the producers to minimise the product’s environmental impact.

One particular type of sustainable label is the so-called CF label, which is an indicator of the total amount of CO₂, or the equivalent of CO₂ in the case of the emission of other GHG (usually expressed in grams), emitted into the atmosphere along the whole “life cycle” of a particular product or service. Thus, the calculation comprises not only production but also transport, transformation, distribution, use and disposal.

In the agri-food sector, the European Commission has identified 129 (both public and private) information plans concerning the concept of sustainability (Grunert et al. 2014). Among these labels, the organic brand (referred to also as “bio”) is the most widely used in the Italian market. Local production, however, is gaining popularity among Italian consumers, even though a universal label for the definition of such products has not yet been established (Bazzani and Canavari 2013, 2017).

CF labels are rarely present in the agri-food market⁴, and only recently, consumers have occasionally had access to information about the CF of products, both in Italy and in most European countries. Tesco experience is exemplary in this field: the retailer, together with the Carbon Trust, has started introducing the first CF label in food retailing in 2009, claiming that they would have labelled all the 50,000 own-brand products (The Economist 2011). However, in 2012, when they only have been able to label 500 products, they had to give up the project. The reasons for this failure were that: consumers found the labels complicated and difficult to understand (so the company was looking for alternatives to replace the CF); the process of labelling the products revealed much more time consuming than planned and other retailers were slow in adopting CF labelling. Thus, the uptake of the label could not reach the desired critical mass (Financial Times 2012).

Nowadays, there are only a few CF labels that have continued in the marketplace. However, as mentioned by Peschel et al. (2016) and Grebitus et al. (2015), Eurobarometer survey’s results have found 72% of EU citizens agreeing that CF information on products should be mandatory (European Commission 2009). More recently, about 90% of EU citizens have declared that buying environmentally friendly products can bring real benefits to the environment (European Commission 2012).

In this context, it should be of much interest to investigate the drivers and the socio-economic characteristics of respondents that can influence a positive WTP of consumers towards CF-labelled products.

A literature review of WTP studies on CF for food products

Consumers’ preferences for lower CF label products have not yet been widely explored in the literature (Vanclay et al. 2011), also because of scarce market presence and uptake, and only recently, there has been a growing body of literature proposing formal evaluations of consumer response to carbon labelling.⁵

We performed a literature review to examine the works available in the Italian and international scientific literature that so far have analysed consumer preferences and WTP for CF label.⁶

Articles were selected by checking against pre-determined criteria for eligibility and relevance. Firstly, the following keywords have been identified: “footprint” (and its possible variation “foot-print”), “consumer”, “food”, together with their Italian translations. Secondly, a search for the abstracts of the articles has been done based on these keywords in the primary databases for scientific relevant literature (Scopus, Web of Knowledge, AgEconSearch, EconPapers), and thus pertinent articles have been selected.⁷

Approximately 300 articles have been consulted (including 150 references from Scopus and 130 from the Web of Knowledge, largely overlapping). Those papers went through a screening process that made emerge only 27 of them for an in-depth analysis, as they were in line with the specific goals of the review. These low figures reinforce the argument that the topic has not been widely explored in the empirical literature so far, in particular for the Italian consumers. Table 1 summarises the main aspects of the selected studies: country, products of reference, the methodology used and the main findings.

Table 1 Published articles regarding WTP evaluations for CF labels: country, products of reference, methodology used and main findings

Most of these studies show that in general consumers are responsive to CF on different products indicating lower emissions, than conventional ones. However, as Vanclay et al. (2011) found, CF labels are most effective when combined with lower prices.

Moreover, Akaichi et al. (2013) and Onozaka and Mcfadden (2011) highlighted that consumers have been particularly likely to buying low-CO₂ products in case they were also labelled with local origin, and according to Hoek et al. (2017), the combination of a health and carbon logo has a more positive effect than the logos separately or no logo. An interesting result is that from Shewmake et al. (2015) that have shown how even if CF labels can lower GHG emissions, they can also have the potential to incur in the opposite effect if their implementation does not account for consumer beliefs as well as complementary and substitute relationships among different products.

Among the sorted articles, only the works by Caputo and co-authors (Caputo et al. 2012, 2013b), Vecchio (2013), Vecchio and Annunziata (2015), Lombardi et al. (2017) and Colantuoni et al. (2016) focused specifically on the Italian market.

Caputo et al. (2012) provided information on the presence of food miles’ labels and the level of GHG emissions related to transport, finding a positive influence of both information on consumers’ utility. Caputo et al. (2013b) found that Neapolitan consumers interviewed have shown a greater WTP for transport distance information label (food miles) rather than for the more comprehensive CF label. This finding suggests that the local origin of the product might have an impact on the Italian consumers’ purchasing choice. Italian consumers could thus be more concerned with labels related to a concept of sustainability together with the local origin.

Vecchio (2013) and Vecchio and Annunziata (2015) evaluate young consumers’ attitude towards sustainability labels. Vecchio (2013) found a positive young adult wine drinkers’ WTP for CF-labelled wine. Vecchio and Annunziata (2015) found a positive (1.41€) WTP for CF-labelled chocolate bars and identified some factors positively affecting WTP for CF label: age (younger individuals express a higher WTP); gender (female respondents); intensity of trust in the specific labels and the preference for food obtained in an environmentally friendly way.

Lombardi et al. (2017) analyse consumers’ preferences when buying fresh milk and find an average premium price of 0.55€ per litre.

On the contrary, Colantuoni et al. (2016) explore the market potential of domestic early potato and find that Italian (and German) respondents were unwilling to pay more for CF certification. Marginal WTP estimated was, in fact, negative and higher for Italians than for Germans.

As regards the type of product, the preference for low CF product has been found for both livestock and vegetable foodstuff. Echeverría et al. (2014) have analysed the WTP of Chilean consumers for both a product of vegetable origin (bread) and an animal product (milk) and found that respondents showed greater sensitivity when evaluating animal products as they were more likely to pay for lower CF for milk than for bread. To this respect, product origin (animal or vegetal) can be acknowledged as an additional aspect that could potentially affect the preferences of Italian consumers for CF labels: e.g. Canavari and Nayga (2009) have shown that Italian consumers exhibit differentiated behaviours when consumer choice is related to GMO products of vegetable origin rather than of animal origin.

As regards the methodological aspect, the WTP for low CF products was primarily estimated by hypothetical choice experiments. Only five studies out of 27 have used non-hypothetical methods (i.e. experimental auctions in four cases and a real choice experiment in one case).

The two case studies in the dairy sector: data and method

The two case studies carried out dealt with consumers’ habits related to dairy foodstuffs purchasing and were performed through two different surveys aimed at evaluating consumer understanding, knowledge, and preference for low CF products. Besides, they aimed at identifying the products’ characteristics that influence purchasing behaviour and the consumers’ WTP for the purchase of 1 litre of fresh milk with a lower CF label in comparison to a conventional one. The focus on dairy foodstuff was driven by the importance of livestock products in the debate at international level for their higher contribution to climate change with respect to vegetable foodstuff production (Gerber et al. 2013; GRAIN and IATP 2018).

The two studies were conducted among Italian consumers from December 2016 to February 2017, in both cases using an online questionnaire gathering information on consumption choices and socio-economic characteristics for 393 consumers interviewed (215 in case study A and 178 in case study B, respectively). The questionnaires were similar but not identical, and they were composed of four sections: the first on consumers’ habits, the second on their environmental awareness, the third about their knowledge of environmental and CF labels, and a section dealing with personal socio-demographic information (Author1 et al. 2018).

Though the use of web instruments to administer the questionnaire has facilitated reaching a high number of respondents, it raises the issue of representativeness of the sample, because this sampling method tends to gather self-selected respondents. Consequently, it usually generates a biased sample, in which younger people with a higher level of education or web literacy are overrepresented (Canavari et al. 2005). Therefore, the samples cannot be considered representative of Italian consumers; nevertheless, they allow obtaining quite interesting information about the relationship among the variables analysed. Though WTP estimations based on a non-representative sample cannot be used to extend WTP results to the population analysed, as figures would be biased, the relationships among the socio-economic characteristics of respondents the positive WTP eventually expressed, remain valid.

Table 2 and Table 3 show the descriptive statistics of the surveys analysed. As mentioned, the questionnaires were similar, but not identical. Thus, Table 2 shows the replies to the question that differed among surveys, while Table 3 shows only the shared items analysed.

Table 2 Descriptive statistics of the questions differing between the two surveys

Table 3 Descriptive statistics of the questions asked in both surveys

As regards methodological aspects, WTP analysis was conducted with different approaches, both relying on the contingent evaluation (CV) method. CV is a method of estimating the value that a respondent gives to a specific good or a good attribute, by asking to directly report his WTP, rather than inferring it from observed behaviours in market transactions. For this reason, CV is often referred to as a “stated preference” method, on opposite to “revealed preference” method based on existing prices. There are different CV approaches that depend on the way consumers are asked about their preferences. In the present analysis, in the case of study A, the survey used the open-ended contingent valuation method that relies upon asking directly to consumers to state their WTP for the product considered. An open-ended question is a question that is designed to encourage a full, meaningful answer, using the subject’s personal knowledge or feelings. In the case of study B, we relied upon a dichotomous choice contingent valuation: a hypothetical purchase situation has been proposed to estimate the WTP, by comparing product 1 (milk bottle with 200 g of CO_2e emissions) at the fixed price of 1.30€ with product 2 (bottle of milk with 150 g of CO_2e emissions). Respondents were asked to state their preference between the two products according to a price variation of 0.10€ of product 2, up to a maximum value of 2.00€. Given these different approaches for the elicitation of the WTP in the two studies, the consumer’s preference has been evaluated considering whether the choice of the respondent (individual outcome variable) was to state a positive WTP for lower CF products or not. The dichotomous choice contingent valuation, proposing such a take-it-or-leave-it survey valuation question, is more likely to reflect real individual purchase decisions. Moreover, such elicitation format has shown to be less susceptible to strategic bidding behaviour than the open-ended one.

A conditional logit model has been estimated to investigate the determinants of the probability for consumers to declare a positive WTP for products with lower CF, based on explanatory variables, as responses shared to both surveys, expressing some socio-demographic characteristics and attitudes of the consumers interviewed. The general equation for the conditional logit model estimated is:

$$ {P}_i\left({y}_i\ne 0\ |{X}_i\right)=\frac{\exp \left({X}_{i\kern0.5em }\beta \right)}{1+\exp \left({X}_{i\kern0.5em }\beta \right)\ } $$

(1)

where i indicates the generic individual; P_i is the predicted probability of individual i to make a specific choice; β is a vector of unknown parameters and X a vector of explanatory variables expressing the characteristics and choices of the individual expected to influence the respective choice. By including among these variables some features of the individuals interviewed, we assume heterogeneous respondents’ preferences.

For the sake of brevity, we only report the variables that were significant (Table 4).

Table 4 Estimations results

Results discussion

Table 2 shows the answers to both the surveys to the different questions analysed. As regards survey A, interestingly, almost all the respondents declare to know the climate change phenomenon, are (on average) interested in it and think that the consumption of products with an environmental label helps contrasting climate change. This survey also reported a set of questions on which tools could be used to promote the knowledge and dissemination of CF labels and the web instruments and education were judged the most important, followed by campaigns, advertising, the label itself and newspapers.

As regards survey B, the majority of respondents consider it important to have a CF label to inform purchase decisions in an environmental sense and thinks that buying organic food helps to reduce GHG emissions. However, it seems that the majority of the respondents do not read the label but gives importance to the sensory quality or expiration date when buying food.

Table 3 summarises some descriptive statistics of the pooled sample analysed, made by the common questions. As mentioned, being the sample self-selected and based on an online survey, some demographics reflect the nature of the data source. The respondents are 64% female. Almost all the respondents have at least a high school diploma, and 52% have a university degree (or higher). Despite the low presence of CF label in the Italian market, a majority of subjects declared to know the concept of CF labels.

As regards WTP, results indicate that only 24% of the total sample states not to be willing to pay more for a litre of milk with lower CF. This figure is likely underestimated since it is based on a stated preference survey, and the sample is self-selected.

For respondents that declared a positive WTP, in case study A, an average 9% premium price for lower CF milk has emerged, with maximum values of 50%. The premium price was on average 0.19€, assuming an average price of 2€. In case study B, the average WTP was more than 30%. The consumer is likely to pay € 1.68 per bottle of low CF milk and therefore, compared to the high-impact product proposed at the price of € 1.30, the surplus difference is € 0.38 (Author1 et al. 2018).

The results of the logit model estimation summarised in Table 4 allow identifying the drivers behind this positive WTP.⁸

Results indicate that in the sample analysed, if a respondent gives high importance to low impact products to tackle climate change, this trait positively affects the probability to be willing to pay more for CF-labelled milk. Also, the format of the different surveys may matter: survey B respondents are more likely to show a positive WTP compared to survey A respondents.

As regards socio-demographic variables, respondents who are more sensitive to price when buying products (about 40% of the sample) are less likely to be willing to pay more for products with a lower CF label; this result is consistent with what other authors in this field have found (see among others Vanclay et al. 2011).

Instead, the only knowledge of the CF concept does not seem to be relevant in determining the stated perception of value. Also, age and education do not affect the WTP of consumers, similarly to what was found from the detailed analysis of case study A published in another article (Author1 and Author2 2019).

As regards gender, females show a slightly higher WTP than males, but this result is significant only at the 0.10 level. Hence, even if this result goes in the direction of what found in previous work (e.g. Steiner et al. 2017; Vecchio and Annunziata 2015), its statistical significance is quite poor.

Conclusions and policy implications

CF labels represent one of the most important tools to help to tackle climate change through consumers’ informed purchases behaviour.

Despite their relevance for demand-driven mitigation options, their presence is still scarce in the Italian food sector, and so it is also for studies aimed at investigating Italian consumers’ WTP for products with lower CF. From the literature review, a positive WTP for lower CF products seems to emerge, though not for all products and respondents’ socio-economic characteristics.

The two pilot case studies presented, focused on Italian consumers’ habits when purchasing milk, allowed us to make a rough evaluation of their preferences for low-carbon-labelled dairy products.

Results, though based on convenience and probably biased samples and stated preferences, suggest that the interest of consumers in CF labels may exist. Findings are generally in line with previous studies indicating that respondents that give high importance to foodstuff produced with low environmental impact to mitigate GHG emissions have shown to be more willing to attribute a positive premium price to CF-labelled products. Also, the data confirm that CF labels could be most effective when combined with prices lower than (or at least equal to) conventional products (Vanclay et al. 2011), as more price-sensitive consumers are less prone to perceive a higher value for lower CF products.

Those results, if confirmed by larger and representative samples, may have interesting policy implications. In fact, they would suggest that a policy framework aiming at promoting demand-side mitigation options in the agricultural sector should tackle both the consumers’ side, informing consumers about the environmental impact of food production and the potential of environmental label in reducing it, and the producers’ side, helping the food supply chain reducing its GHG emission in a cost-effective way.

About the consumers’ side, policies should aim at both enhancing consumers’ awareness about climate change challenge and ensure that the system of certification is reliable and easily interpretable by consumers. To this respect, the initiative of the European Commission (2013) on “building the single market for green products facilitating better information on the environmental performance of products and organisations”, is of utmost importance.

As regards the production side, results would suggest to producers that a lower CF would be appealing if offered at the same or a lower price. Indeed, the possibility to couple lower prices with lower GHG emissions in the agricultural sector is not rare, because technical studies on the mitigation potential of agri-food productions have found many of the so-called “win-win solution” to climate change, i.e. strategies that allow saving both GHG emissions and production costs (Coderoni et al. 2015). When a win-win solution is adopted, thus, lower CF products can be produced at lower costs⁹ that could, in turn, be translated into lower selling prices, as entrepreneurs participating in the CF labelling scheme have declared to be willing to do (Coderoni and Pontrandolfi 2016). If these solutions are applied, thus, CF product uptake could be easily enabled.

Also, Rural Development Programmes funds could be used to reduce farmers’ costs of adopting GHG saving techniques, as they provide incentives for both GHG calculation and certification and farms’ investments to implement mitigation strategies identified. This should be made taking into account the likely evolution of the food systems as a whole (Macombe 2018).

Given the limitations of this study, a more in-depth analysis is needed to estimate Italian consumers’ WTP for CF labels accurately. Future research should on one side, rely upon a larger and nation-wide representative sample to avoid the problems linked to self-selected and biased samples; on the other side, it should focus on non-hypothetical techniques, such as experimental auctions to obtain reliable estimations of WTP (Lusk and Shogren 2007). In fact, the studies based on hypothetical choices, generally, tend to overestimate the WTP and the experience of Tesco with CF-labelled products seems to confirm this gap between stated and real behaviour. A further research avenue could be the consideration of a more comprehensive framework for the analysis of the environmental impact of food consumption, covering not only the GHG emissions generated, but also the use of resources such as water and land, and the generation of waste (Candy et al. 2018).