2. Reducing climate impact; the role of informed choices

The provision of information as a measure to change consumers’ purchase patterns is a demand-side instrument, which depends on consumer preferences regarding the quality in question, such that at least a share of consumers prefer lower emitting products to higher emitting variants, all else equal. Moreover, information as a policy instrument relies on that consumers’ access and process the information and that there is product differentiation, such that at least a share of the producers supply the higher quality (lower emitting) variants (Bonroy and Constantatos, Reference Bonroy and Constantatos2015). Consumers will search and evaluate information if the effort required does not exceed the benefits, such that the expected marginal return of information exceeds the expected marginal cost of searching for the information (Stigler, Reference Stigler1961). It is also of importance that the information is seen as reliable (Darby and Karni, Reference Darby and Karni1973).

In the analysis of consumer preferences for food characteristics, a multi-attribute approach forms a theoretical point of departure (Lancaster, Reference Lancaster1966). This implies that products are viewed as bundles of characteristics and that consumers derive utility from each of these characteristics (hereafter referred to as attributes). This approach is fundamental for much microeconomic analysis of consumer demand. Consumers are assumed to be utility-maximizing, and they will choose the product that provides the combination of price and quality attributes that is most consistent with their preferences and subject to their constraints. However, such a decision process, were all product attributes in a purchase situation are evaluated, can be costly in time and effort for the consumer. In choice tasks that are low involvement, which is often the case for food choices, consumers tend to rely on heuristics that facilitate the purchase decision (Kahneman, Reference Kahneman2003; Hauser, Reference Hauser2014). In this context, food labels can simplify the information and assist consumers in the comparison of different products.

3. Ownership and mode of governance

Economic theory posits that if there are no costs of supplying information, and consumers trust the information, all producers except those with the lowest quality will voluntarily disclose the quality of their products. If such information is not disclosed, rational consumers will assume that a product has the lowest quality (Ippolito, Reference Ippolito1990). This prediction relies on the assumption that consumers are aware of that the information exists and, importantly, that the quality in question is important to at least a share of consumers. Under these assumptions, voluntary labeling will provide sufficient information for consumers to make informed choices. However, depending on the level of detail of rating, carbon labeling can be costly. Therefore, widespread voluntary disclosure relies on a large demand from consumers for this type of information.

Broadly, labels can be implemented and regulated by three types of agents: private parties, independent third parties, and government. Private parties, such as private firms and collectives of firms, can provide labels or claims based on self-declared standards. The motivation for these agents is to support product differentiation and target consumers who are willing to pay for the characteristic in question. There are several examples of such private carbon labels (Liu et al., Reference Liu, Wang and Su2016).

Labels can also be certified by independent third parties, such as NGOs, typically with the intention of promoting sustainable food consumption. Independent third party labeling can both target and shape consumer demand. An example of this is the dolphin safe label, which put an environmental problem in the spotlight, and raised awareness among consumers (Teisl et al., Reference Teisl, Roe and Hicks2002). Independent third parties, particularly NGOs, tend to have higher credibility among consumers compared to private parties (Roe and Teisl, Reference Roe and Teisl1998). Although the motives for labeling initiatives provided by private actors and independent third party NGOs vary, they are all voluntary for producers to display.

In cases where private or third party labeling is not credible, government labeling may be mandated (Sunstein, Reference Sunstein2017), to protect consumers from fraud by overseeing standards and certifications. Many countries have government-certified organic labels that are voluntary for producers to display and motivated by a desire to prevent fraud (Golan et al., Reference Golan, Kuchler and Mitchell2001). Public labeling programs can also be motivated as a measure to combat confusion, as many different sustainability-related labels have emerged on the market in the past decade.

In general, two economic situations motivate public mandatory labeling: First, when the unregulated market does not provide sufficient information for consumers to make decisions in line with their preferences. This motive for alleviating asymmetric information is not to change consumption, but to make consumers’ decisions more informed (Golan et al., Reference Golan, Kuchler and Mitchell2001). An example is that it is mandatory to display the nutritional content on products in many countries (FDA, 1994; EU, 2011). Second, and importantly for the case of climate impact, mandatory public labeling can be mandated to target externalities. When a quality is a public good, demand is societal rather than private. This implies that consumers in an unregulated market do not personally experience the full costs from consumption. In such cases, the unregulated market typically undersupplies the quality if it is a positive externality and oversupplies the quality if it is a negative externality such as CO₂ emissions (Caswell and Anders, Reference Caswell, Anders, Lusk, Roosen and Shogren2011). Mandatory labeling can then serve as an instrument to affect consumption, but we note that it relies on consumers’ preferences and willingness to pay for the public good.

4. Carbon label characteristics

Generally, carbon labels on the market and in experimental studies fall into four categories (Thøgersen and Nielsen, Reference Thøgersen and Nielsen2016): (i) Labels that provide information about the CO₂ equivalents emitted from a unit of the product, based on lifecycle analysis. (ii) A symbol that indicates that the product is low in carbon emissions. The criteria that the certifying agent applies are typically not presented to the consumer at the point of purchase. (iii) Labels declaring that the producer has committed to reduce emissions. This type of label does not inform about the absolute level of emissions, but rather about an improvement compared to historical levels. This type of label is potentially most beneficial for high emitting producers that have more room for improvement. (iv) Carbon neutrality was the carbon emissions caused by the production of the product that has been compensated by the company, for example by paying for forestation (Thøgersen and Nielsen, Reference Thøgersen and Nielsen2016). In the following, we focus on labels that inform consumers about carbon emissions from the product in some form. The third and fourth types of labels are not related to the absolute emission levels and are hence not educative for the consumer. Since a shift in consumption patterns is needed to achieve significant reductions in emissions from food, we will focus on the first two label types only.

4.2 Detail of rating

The detail of rating describes the degree of differentiation between products that a label enables. The detail of rating ranges from continuous scales to binary ratings. Similar to the Assessment dimension, the Detail of rating dimension is related to the ease of use and understanding for the consumer. Both dimensions are illustrated in Figure 1. Among existing food labels related to health and sustainability, there is a correlation between the detail of rating and the type of assessment, although such correlation is not necessary. While purely descriptive labels (D) commonly are in continuous forms, these values can be presented in a categorical form, to reduce the cost on producers and certifiers of measuring precise values. The categorical form will also reduce the cognitive burden for the consumers. Evaluative labels are commonly in the form of traffic light systems or logos, although the EU energy label is an example of an evaluative label in a detailed format (European Commission, n.d.).

Figure 1.

Assessment and detail of rating.

Note: Assessment ranges from Descriptive to Evaluative (horizontal axis). Detail of rating ranges from Binary to Continuous (vertical axis).

The type and volume of information on a label will affect the amount of time and effort required, in order to evaluate product attributes. In general, consumers find categorical labels easier to understand and use compared to continuous scales (Loewenstein et al., Reference Loewenstein, Sunstein and Golman2014). There is also evidence that consumers prefer the less detailed nutrition and health labels and that consumers are more successful in identifying healthy products when using labels with fewer categories (Campos et al., Reference Campos, Doxey and Hammond2011). Hence, more information is not always better, as this increases the likelihood that consumers do not find it worthwhile to spend time and effort to use the information and rather choose to ignore it. Taken together, theory does not guide on the optimal amount of information and level of detail, as this likely varies between individuals and products, depending on their prior knowledge, cognitive ability, interest, and preferences for the attribute that the label concerns (Teisl and Roe, Reference Teisl and Roe1998).

4.3 Level of reference

For all labels that are evaluative to any degree, there must be a level of reference. This implies that a subjective judgment must be made on how the reference is defined. For example, how should the reference intake be calculated for the reference intake (RI) type of label? Which thresholds should be used for the traffic light system (TLS)? And which requirements should there be for the display of a positive low carbon logo, L(p), or a negative warning logo, L(w)? The choice of level of reference has important implications for which type of information the consumers retrieve from a label. This is particularly decisive for the most evaluative labels such as TLS and L labels, which identify if a product fulfills certain criteria. In particular, it must be decided if these criteria identify carbon emissions thresholds that apply across all products or if there should be different thresholds within each food category. For the purpose of illustration, we assume two product categories: one high emitting and one low emitting product category. The high emitting category can be red meat, while the low emitting category can be crop based. Further, we assume two products in each category: one high emitting variant and one low emitting variant. This depends on aspects such as technology in production and management practices and can therefore not be evaluated by the consumer upon inspection or consumption. The implications from the choice of reference are illustrated in Figure 2.

Figure 2.

Level of Reference for evaluative labels.

Note: TLS=Traffic Light System, L(p)=Logo (positive). Red footprint indicates high emissions, and green footprint indicates low emissions. Amber footprint indicates medium high emissions, but is not included in this simplified example.

As visible in the middle column of Figure 2, labels with between-category evaluation inform about the overall emission levels. This type of label informs consumers about large differences. A comparison to the health label area could be that unhealthy categories such as soda, chips, and candy are labeled with red TLS labels (or cannot display the positive logo), while all products in healthy categories, such as vegetables, are labeled with green TLS labels or displays the positive logo. An important benefit with the between-category evaluation is that it is educative, as consumers will learn about the larger differences in emissions. Largely, the same information could be disseminated without specific product labels, as the between-category evaluation could be communicated in general information campaigns, in posters in the purchase setting, and through educational efforts in other environments. Yet, while this is true in general terms, and regarding the large differences between product categories, the large number of product categories available in many grocery stores implies that it will be challenging to communicate the necessary information through general campaigns only.

For the within-category evaluation type of label (right column in Figure 2), consumers are informed about the relative climate impact within a specific category. The information conveyed with this level of reference type is more likely to be a credence quality; consumers cannot judge upon inspection or consumption if the product is low or high in emissions compared to other similar products. For this reason, contrary to a between-category evaluation, the information conveyed from within-category evaluation is not possible to provide in general information campaigns. Importantly, the information is less educative compared to the between-category evaluation, in the sense that consumers will not learn about which type of substitutions that have the largest impact on their climate impact. Since today’s consumers usually buy a mix of food products, from both high and low emitting product groups, within-category labeling can still have an important impact on consumer choices and carbon emissions. Consumers usually find it easier to substitute products within a product group than between products in different product groups. The within-category evaluation type of label is available in the health area, including the health-tick, the Nordic keyhole, and the Choices logo, where consumers are informed about the healthier choice within a certain category. Notably, the literature that evaluates the effects of these label focus on within-category substitutions, and not on larger improvements toward healthy diets, a shortcoming that has been noted in the literature (Nikolova and Inman, Reference Nikolova and Inman2015). As can be seen from Figure 2, low emitting products within both high and low emitting categories will have the same logo.

In addition to the type of assessment, the level of detail, and the choice of reference level, other label characteristics related to the appearance of the label can affect the attention and use of the label. For example, front of pack food labels with more color receive more attention, and the placement on the package, and how many other labels there are on the product also affect attention and use of labels. Hence, these aspects are of relevance in the design of a label and for producers in the decision on how and where to display a label (Graham et al., Reference Graham, Orquin and Visschers2012).

5. How does labeling strategy affect consumer perception?

Both the mode of governance and the labeling characteristics will affect the type and amount of information a carbon label provides. For illustration, we extend on the example with two food categories presented in Figure 2. The columns in Figure 3 represent different types of assessment and varying degree of details of rating, ranging from the purely descriptive and continuous Digit label to the fully evaluative and binary labels (logos). Further, the evaluative labels vary by the level of reference, including both between- and within-category evaluation. The two vertical panels in Figure 3 represent the mode of governance, with voluntary and mandatory labeling.

Figure 3.

Dimensions of carbon labeling systems.

Note: D = Digit, RI = Reference Intake, TLS = Traffic Light System, L(p) = Logo(positive), L(w) = Logo(warning).

Figure 3 illustrates that with voluntary labeling, only the low emitting products will display information. This holds across different types of assessment, details of rating, and levels of reference. Rational and sophisticated consumers will infer that the absence of low carbon emission labels implies that a product is high in carbon emissions (Ippolito, Reference Ippolito1990). If this assumption holds, private and voluntary labeling should be sufficient. Yet, if consumers lack this awareness, or fail to make the conclusions, voluntary labels are not sufficient, and mandatory labeling is justified (Loewenstein et al., Reference Loewenstein, Sunstein and Golman2014). In sum, given that climate impact is a public good and that consumers may not make the inference from the absence of labels, mandatory carbon labeling is likely needed to ensure that sufficient information is provided to enable consumers to make informed choices.

Mandatory labeling, which includes all products, is expected to have a larger impact on consumer behavior since such labeling systems also identify products with high emissions. As mentioned, consumers tend to react more strongly to negative information than to positive information. Inspection of the mandatory labeling (lower panel of Figure 3) reveals that the most descriptive and detailed label formats (Digit and Reference Intake) provide greater amount of information and enables comparison both between and within categories. For evaluative labels, the choice of reference level greatly affects the type of evaluation consumers can make from the label. There is a big difference between mandatory between-category and within-category evaluative labels, if one considers total CO₂ emissions. The last three columns in Figure 3 show that the within-category evaluative logos provide the consumer with exactly the same information for products with high and low climate impact, although their climate impact is very different. Hence, consumers must be knowledgeable about that such within-category labels only guide decision within categories and that it is not to be used for substitution decisions between categories.

6. Benefits and costs of labeling

When deciding on labeling strategy, policy makers should weigh the benefits against the costs of labeling, and how these are distributed (Golan et al., Reference Golan, Kuchler and Mitchell2001; Sunstein, Reference Sunstein2017). For consumers, the main benefit of a labeling is that they are enabled to make informed choices, which are in line with their preferences. The social benefits from carbon labeling depend on the reductions in carbon emissions that the labeling program achieves. Such effects on carbon emissions, hence, depend on the changes in consumption that occur after the label is implemented, which in turn depends on consumers’ preferences.

If few consumers have interest in making choices that are low in carbon emissions, and if they do not value this much, the labeling benefits will be small. Policy makers can be informed about this by quantifying consumers’ willingness to pay for reduced carbon emissions. Consumer valuation studies, where the willingness to pay is measured in experimental studies, can be used for estimating this (Sunstein, Reference Sunstein2017). Experimental studies have found support for a positive willingness to pay for products that are lower in carbon emissions, but the climate impact is relatively unimportant for most consumers compared to other food qualities (Rondoni and Grasso, Reference Rondoni and Grasso2021). A challenge with this approach is that since carbon labeling has not been implemented on a large scale, it may be difficult for consumers to evaluate the use and benefits with carbon labeling (Sunstein, Reference Sunstein2021; Reisch et al., Reference Reisch, Sunstein and Kaiser2021). Ex post estimates of the willingness to pay for carbon labeling may therefore give better assessments, and this could be done via lab or field experiments (Sunstein Reference Sunstein2021). Further, it is of relevance for policy makers to reflect on the distribution of willingness to pay among consumers.

One potential benefit from labeling is education. This is particularly true for government-mandated labeling with a labeling format that enables comparison between product categories. While general knowledge about carbon emissions from different product categories is possible to communicate in other channels, such as information campaigns and dietary guidelines, consumers may find it difficult or misinterpret the connection in the purchase situation. Hence, for information that is in the between-category evaluation format, mandatory labeling could be motivated by a need to make such information more salient and easier to access (Teisl and Roe, Reference Teisl and Roe1998). As a sustainability education intervention, mandatory carbon labels could have a broad reach if they are displayed on products, as they will be presented to consumers upon purchase, but also when food is prepared and consumed. The information may also affect the social norm such that this information becomes more demanded and important for consumers in their purchase decisions.

A further benefit that may arise from government-mandated labeling is that it can motivate producers to reduce their emissions to obtain labels that are more favorable. Evidence from other areas shows that the implementation of mandatory calorie information disclosure caused restaurants to make their menus healthier, and when an energy efficiency label was mandated, the energy efficiency innovations increased substantially (Golman et al., Reference Golman, Hagmann and Loewenstein2017). Labeling benefits in the form of reformulation effects are more likely to occur if the more descriptive (D and RI) or the evaluative within-category evaluation types of labels are implemented. For example, a producer of high emitting products in the high emitting category can change technology and management practices to reduce emissions and improve their digit, RI or obtain a positive within-category logo. In contrast, a between-category evaluation will not enable producers in the high emitting categories to obtain a positive logo by technological advances, since emission levels are connected to the product category.

Finally, if firms anticipate stringent requirements in the future, such as implementation of a carbon tax or regulations of maximum amount of carbon emissions per product, they may voluntarily display carbon labels. By voluntarily displaying, and improving their formulas, they show to the consumers (and policy makers) that they are making an effort in the area. This potential benefit from a carbon label is, however, difficult to measure (Yokessa and Marette, Reference Yokessa and Marette2019).

Labeling is also associated with costs for consumers, producers, and the certifying body. For consumers, the time and effort required for searching and interpreting the label is associated with alternative costs (it could be used for something else). The more evaluative labels are less costly for the consumer to interpret compared to the most descriptive and detailed label formats. Moreover, the introduction of a carbon label can contribute to choice overload, where the number of sustainability-related labels is already high (Cohen and Vandenbergh, Reference Cohen and Vandenbergh2012; Brown et al., Reference Brown, Harris, Potter and Knai2020). An overload of labels may thus contribute to reduced benefits from other labels. Information may also create a psychological cost or discomfort for consumers (Thunström et al., Reference Thunström, Nordström, Shogren, Ehmke and vant Veld2016, Reference Thunström, van’t Veld, Shogren and Nordström2014; Nordström et al., Reference Nordström, Thunström, van’t Veld, Shogren and Ehmke2020). If climate information causes consumers great disutility, this could constitute a non-marginal cost, and alternative policy instruments may be regarded more socially optimal (Golman et al., Reference Golman, Hagmann and Loewenstein2017).

Further, while producers face labeling costs associated with providing information related to administration, testing, reformulation, printing, and certifying bodies face costs of enforcing and controlling label use, such costs may be forwarded to the consumers (Golan et al., Reference Golan, Kuchler and Mitchell2001; Crespi and Marette, Reference Crespi and Marette2003). With voluntary labeling schemes, the cost falls on the labeled products, while mandatory labeling will distribute the costs among all products. It is thus important to take distributional aspects of the labeling costs into account when evaluating different labeling schemes. Importantly, the dimensions of a carbon label will influence the costs, where the cost for a digit type of label is relatively high, since it is based on life cycle analysis. The early experiences from such labels by Tesco highlight that such costs are considerable (Vaughan, Reference Vaughan2012). The evaluative labels can be assumed to be cheaper to implement since these labels do not require the same degree of detailed information.

In addition, the benefits of carbon labeling depend on the consumers’ behavioral changes and associated reduction in the carbon footprint. As this can be difficult to estimate a priori, well-designed field studies can provide valuable information before decisions about mandatory carbon labeling are taken.

7. Effects from carbon labeling on market structure

The implementation of a credible carbon label enables firms producing higher quality (lower emitting) products to differentiate their products from that of lower quality competitors. Importantly, the introduction of a mandatory or voluntary label may affect the market structure, by affecting different types of firms differently.

The costs associated with a carbon label are one aspect that may result in changes in the market structure. Costs for firms associated with a labeling scheme consist of certification costs and costs for producing products that fulfill requirements for displaying a high-quality label. Cost for certification can be fixed or variable, where fixed costs may burden smaller firms disproportionally, leaving them disadvantaged relative to larger firms (Golan et al., Reference Golan, Kuchler and Mitchell2001). Fixed costs may also work as an entry barrier for new firms. Further, the costs for firms to achieve higher quality products may affect firms differently. Reducing carbon emissions, and thus fulfilling the requirements for a binary logo or higher levels on a multi-level label, may be fixed (e.g. investing in R&D or new technology) or variable (increased reliance on low emitting input) (Roe and Deans, Reference Roe and Deans2014).

In addition to costs associated with labeling, firm welfare effects and implications on the market structure depend on a number of aspects, including the market structure prior to the introduction of a label, the design of the labeling scheme, and the preference structure among consumers (Bonroy and Constantatos, Reference Bonroy and Constantatos2015; Sheldon, Reference Sheldon2017). In the industrial organization literature on food labeling, it is typically assumed that there are high-quality and low-quality products, and this is a scenario that describes how many credence qualities are represented on the market (e.g. organic, non-GMO, MSC, locally produced). Consumers are assumed to prefer the high quality, but have different willingness to pay for it. In the case of carbon impact, this holds if a within-category evaluative type of labeling scheme is applied. All else equal, consumers can be assumed to prefer lower emitting product variants but their willingness to pay for lower emissions and the cost of labeling is an empirical question that needs to be studied before firm conclusions can be given on the market outcome (see e.g. the results in Mattoo and Singh (Reference Mattoo and Singh1994)). Looking at the marker structure as such, results in Zago and Pick (Reference Zago and Pick2004) indicate that if the labeled high-quality sub-market remains sufficiently competitive, the introduction of a labeling scheme is welfare improving. However, if the introduction of labeling increases concentration in the sub-market, it reduces welfare. Bonroy and Constantatos (Reference Bonroy and Constantatos2015) study the welfare effects of labeling under price competition and find that labeling is welfare improving (under a uniform consumer distribution), driven by higher profits for the firms. It is thus case-specific whether a labeling reform will improve or reduce welfare, and this requires ex post analysis. We note that the external benefit of lower CO₂ emissions is not accounted for in these studies, and since this is an important motivation for introducing mandatory carbon labeling, it should be included in the overall welfare analysis.