The effects of ecolabels on environmentally- and health-friendly cars: an online survey and two experimental studies

The Paris Agreement, ratified globally by practically every country, except the USA and North Korea (UNFCCC 2018), states that economy-wide decisions should be introduced to achieve absolute emission reduction targets. Road transportation in Europe is a significant contributor to CO₂ emissions, as well as air pollutants such as particulate matter and nitrogen oxides (EEA 2015; Davis et al. 2010). The greenhouse gas emissions from road transportation have significant impact on the environment and on human health (Woodcock et al. 2009). Furthermore, exposure to air pollutants has been identified as a significant risk factor in a number of health conditions including respiratory infections, heart disease, stroke and lung cancer (EEA 2015; WHO 2018). Greenhouse gas emissions still remain above agreed levels and would need to fall dramatically in order to meet international agreements (UNFCCC 2018). In order to comply with the directives of the Paris Agreement, European member states will be obliged to take green initiatives (e.g. fiscal incentives, local access restrictions for particular vehicles, local environmental zones differentiating vehicle access according to emission classes or ecolabels) to reduce air pollution in cities and highly trafficked areas (Kushwaha and Sharma 2016). Ecolabels are viewed as voluntary environmental and consumer policy instruments and taken to be a simple and pragmatic option.

Increasing concern about climate change has led to calls for labelling to allow consumers to differentiate between more or less sustainable purchasing options (Teisl et al. 2008). It is assumed that with appropriate labelling information, some consumers will be motivated to purchase cars that are more sustainable (Thøgersen et al. 2010). Ecolabelling is seen an important way of enhancing transparency and consumer trust in environmental claims. It is also viewed as a method for improving the sustainability of consumption patterns without compromising freedom of choice and at the same time reducing consumers’ information search costs (Grunert and Wills 2007). The elaboration likelihood model (ELM; Petty and Cacioppo 1986) describes different ways of processing stimuli and how these affect outcomes, such as attitude change and eventually consumption behaviour. The ELM proposes two major routes that are used to process messages: the central and the peripheral routes. The central route involves an active consideration of the arguments presented in the message. By contrast, the peripheral route makes use of a simple heuristics about the merits of the advocated position, such as the credibility or attractiveness of the message. In the context of an ecolabel, the likelihood of elaboration with the central route is determined by an individual’s motivation and ability to evaluate the information presented. With the purchase of a costly product such as a car, it is assumed that consumers will adopt the ‘high-effort’ central path (Thøgersen et al. 2010) in their assessment of ecolabels.

Importantly, Thøgersen et al. highlight the importance of constructs such as ‘environmental involvement’, the credibility of ecolabels and the extent to which consumers understand the information in choice behaviour (Bamberg 2003; Gadenne et al. 2011; Thøgersen 2000, 2002, 2005; Thøgersen et al. 2012; Thøgersen and Noblet 2012). For example, the credibility of labels can be influenced also by consumers’ prior beliefs with regard to sustainable products (Teisl et al. 2003). The perceived effectiveness regarding their own behaviour and faith in the behaviour of others appear to be positively associated with increased effectiveness of labels as sources of information (Berger and Corbin 1992; Bougherara et al. 2005). As regards the labels as such, their effectiveness apparently increases when consumers can adequately rank competing products by key attributes (Teisl and Roe 2005), consumers’ prior beliefs (Teisl 2003) or when consumers are educated about the presence and meaning of ecolabels (Song et al. 2019). Comparative labels are also considered a potentially effective way of rendering complex numerical information into simple categorical scales (Peters et al. 2009). However, the use of ecolabels has been criticised because they are based on the assumption that consumers and firms behave irrationally, such as the absence of evidence of an energy-efficiency gap (Gayer and Viscusi 2013) and on the assumption that labels influence consumer to over-value energy consumption in the purchase of goods (Sahoo and Sawe 2015). Making use of ecolabels must therefore be based on robust evidence (Codagnone et al. 2016).

To encourage consumers to purchase less polluting vehicles, the European Commission is considering a voluntary emission standard, called European Union Low Emitting carS (EULES) with three complementary strategies (Ntziachristos et al. 2016): first, to promote clean transportation by providing a benchmark for local or national authorities in their development of financial or in green procurement projects; second, to provide an incentive for manufacturers to produce vehicles that deliver significant emission reductions; and third, by guiding environmentally conscious customers towards the purchase of vehicles with ‘clean’ emission profiles (low real-world driving emissions) by ecolabelling cars. The main focus of the present study is on this third strategy, the effect of ecolabelling cars.

Broadly defined ecolabelling is a signalling method to encourage future consumers towards forms of sustainable consumption (Thøgersen et al. 2010). In comparison to the more general literature on ecolabels, and especially to that focusing on white goods and food, studies examining the effects of car labelling are scarce, with a small number of scientific articles and a few policy reports (Choo and Mokhtarian 2004; Kurani and Turrentine 2002; Lane and Potter 2007; Loureiro et al. 2012; Noblet et al. 2006; Teisl et al. 2008; Teisl and Roe 2005). Therefore, it is necessary to establish which factors are associated with the effectiveness of labelling cars and to examine causal effects on consumption choices.

To date, there is no single experimental study on the effect of car ecolabels on consumers’ decision-making. Teisl et al. (2008) argue that rather than focusing on either the correlation between individual characteristics and environmental friendly behaviour (eco-behaviour) or between labels’ characteristics and eco-behaviour, it is important to test a model linking individual characteristics and labels characteristics. They found that both label design and underlying psychological factors need to be taken into account. More specifically, if the label is well-designed, it can affect individuals’ perceptions of the eco-friendliness of a product.

Findings concerning consumers’ preferences and the car purchase process are reported by Achtnicht (2012), COWI (2002), Lane and Potter (2007), Noblet et al. (2006), Teisl et al. (2008), and Codagnone et al. (2013). These studies show that purchasing decisions are dominated by attributes such as price, performance and safety with eco-friendly attributes playing a secondary role. Secondly, the purchase of a car is found to be a two-stage process. Purchasers first determine the class of car they want to buy. It is only when they move to select the model that eco-friendly and fuel economy features are more seriously considered. Achtnicht (2012) finds that people in Germany say they are willing to pay more for a car that has lower CO² emissions and that willingness to pay decreases for those who reported lower price ranges for a future car purchase and that differences were found by age, gender and educational level. Younger individuals, women and individuals with a higher education entrance qualification have a significantly higher willingness to pay.

Related with the debate in environmental behavioural economics is the legitimate question whether ‘nudges’ based on information supply differ from various forms of informational instruments (Kosters and Van der Heijden 2015). According to Ölander and Thøgersen (2014), ecolabels play the role of information provision at the point of sale but change the choice architecture of consumers when (if) consumers become familiar with and consider them as credible. Following the work of Peters et al. on label that performs well in summarizing complex numerical information (Peters et al. 2009), Johnson et al. (2012) consider ‘good labels’ (as opposed to bad ones) as an instrument of attribute parsimony to reduce attribute overload and as such they qualify them as a nudge for they alter the choice architecture. A further element for considering ‘ecolabels’ as a nudge is the evidence that their design affects consumers’ perceptions (Heinzle and Wüstenhagen 2012).

In this study¹, the main purpose is to identify and test the elements that would nudge consumers towards a preference for an environmentally and health-friendly EULES car. The study focusses on two separate objectives: first, mapping the contours of public responses and the role of public authorities to a voluntary emission standard for passenger cars that would deliver real-world emission levels below the current limits; second, experimentally testing the impact of different incentives designed to support the EULES policy. The outcomes of the research will:

To assess public responses to a voluntary emission standard for passenger cars, an online survey was conducted. In addition, two online experiments tested incentives for optimal message impact (see also Fig. 1). A random sample of 6400 individuals was drawn from 8 countries Germany, Ireland, Italy, the Netherlands, Spain, UK, Czech Republic and Lithuania. The sample was divided and allocated to (i) an online survey with 3200 respondents, 400 per each of the 8 countries, and (ii) two online experiments with 3200 respondents, 400 per each of the 8 countries, giving 200 respondents for each experiment and country. The randomization was ensured at the country level, such that each country was equally represented in the survey (n = 400) and in the experiments (n = 200). Gathering the data across countries made it possible to ensure the validity and possibility to generalize findings on awareness, understanding and attitudes regarding the EULES standard as well as the comparison between EULES and CO₂ emissions.

In the online experiments, respondents were asked to choose between one of two car options. Each option consisted in the picture of the car and additional information. The first experiment featured two versions of the same small car, and the second featured two versions of a large car. Additional information featured a designed combination of one of the two conditions for the four car attributes.

The experimental data was analysed using a conditional logistic regression to estimate the increase in the probability of selecting a car when each of the attributes described above was added to the car label.

For each of the four attributes, the tables show coefficients, odds ratios, standard errors, 95% confidence intervals, and the outcome of the effect likelihood ratio tests expressed by the L-R chi-square (the value of the likelihood ratio chi-square statistic for a test of the corresponding effect; DF, the degrees of freedom for the chi-square test Prob > ChiSq, the p value for the chi-square test). We also report the models likelihood, AICc and BIC². In discrete choice models, each coefficient is a ‘part-worth’ estimate or the utility associated with that attribute. In the analysis, the ‘Taxes’ message was used as a reference point for the categorical variable ‘EULES Message’ and therefore does not appear in the output tables. ‘Taxes’ is used as a reference. ‘Access’ yields a negative part-worth compared to ‘Taxes’, while ‘Health’ has a positive part-worth. By default, estimates are based on the Firth bias-corrected maximum likelihood estimators (MLEs) and therefore are considered to be more accurate than MLEs without bias correction.

The result of the statistical analysis for the two experiments shows that all four attributes considered have a significant impact (P < 0.05) on the participants’ choices (see Tables 2 and 3). The presence of logos, either EULES or on CO₂ emissions, has a small negative effect, 12% less likely for both logos for small cars and 7% (EULES) and 6% (CO₂) for large cars. For small and large cars, the additional price had a marginal positive effect in terms of increasing probabilities of selection by respondents (17% for small cars and 38% for large cars). The effect of the added messages is slightly positive in the case of ‘health’, increasing odds of selection of 13% for small cars and 11% for large cars while negative for ‘access’, decreasing odds of selection by 50% for small cars and 40% for large cars.

Of significance is the finding that the single most important predictor affecting choices is the message conveying the benefits of a low-emission car that accompany the logo. Tables 4 and 5 show how the different benefits captured in the messages (taxes, urban access and health) impact on choices.

Table 5, focusing on larger cars, is based on the same analysis as Table 1. The odds ratio was 2.48 versus urban access and 1.34 versus health benefits. This means that information about the tax advantages of low-emission cars makes the choice of selection 1.48 times more likely than the urban access benefit and 0.30 more likely than a health benefit.

Overall, this shows that information attached to a low-emission logo is the most important attribute in in increasing the likelihood of choice. But, crucially a message about tax advantages has considerably more impact than messages about the benefits of urban access and health. This effect is particularly strong for those choosing small cars and to a lesser extent for those choosing large cars.

In order to comply to the Paris Agreement (UNFCCC 2018), greenhouse gas emissions have to reduce significantly. Current levels of air pollution are above international agreements and an important factor that contributes to air pollution is road transportation (Davis et al. 2010; EEA 2015). It is therefore of great importance to examine effectiveness of behavioural techniques that might improve consumers’ choices when buying cars. The main objectives of the current study were (1) to assess and explore public responses and the role of public authorities to a EULES in order to deliver real-world emission levels below the most stringent current emission limits and to (2) experimentally test the incentives for this optimization.

First, as the results from the survey show, Europeans are aware of the health and environmental impact of cars. The survey points to a general understanding of the adverse health effects of pollution. Results showed that the respondents recognize that there is a link between the environmental protection and human health. Second, people think that the pollution in their cities or neighbourhoods is rather high. Third, the result also shows respondents are aware of the environmental impact of polluting vehicles. More than half of respondents (54%) believe that ‘cars contribute significantly to the air pollution’ in their city or neighbourhood. Despite the high environmental awareness, the main factor that subjects take into account when buying a car is price (50% of respondents), followed by road safety (48%) and fuel consumption (46%). Most importantly, only 1 in 10 (11%) is ready to pay more for a more environment-friendly car. Finally, results also confirm the gap between self-reported attitudes/intentions and actual behaviours: health and environmental concerns come only after many other attributes (price, safety, performance, etc.) in terms of importance in influencing car purchase decisions.

In fact, many consumers perceive buying an eco-friendly car as entailing a loss/sacrifice in terms of performance. There is a quite widespread perception that less polluting or lower consumption vehicles are associated with higher prices and that, to some extent, also compromise performance, which is in line with previous research (Gould and Golob 1998; Zhang et al. 2015). In this regard, citizens do not clearly understand the difference between air pollution and greenhouse gas emissions. Nearly half of the respondents (47%) agree with the statement ‘Vehicles that produce less pollution consume less fuel’. Financial incentives are the most likely to change consumers’ behaviour towards low-emission cars. According to 30% of survey respondents, providing higher financial incentives for low-emission products would be an effective strategy to tackle air pollution. Financial measures (tax breaks, subsidies) are more likely to nudge consumers towards low-emission cars than are non-financial incentives. Among the latter, low-emission zones seem more appealing, especially for those consumers who drive to work on a daily basis. Citizens are likely to prefer tax exemptions and subsidies rather than environmental taxes or road pricing. We draw this conclusion from the survey, in which respondents believe they are paying already too much of environmental taxes for their car and are not in favour of applying ‘road pricing’ (e.g. road tax, registration tax) for major metropolitan areas (e.g. cities with over 100,000 inhabitants and their suburbs). At the same time, they welcome proposals in which the government would provide either tax exemptions or subsidies for those purchasing a low-emitting car. European consumers are not too acquainted with car labels. More than 43% of respondents admit to be ‘unfamiliar with’ existing car labels. Around 37% disagree with the statement that labels are ‘easily recognizable’. As many as 39% do not trust the information displayed, and reportedly, the most common use of labels is ‘to check if what said in the advertisement is true’. Many even misunderstand their purpose: nearly 38% believes that labels symbolize a product’s reliability. In any case, the element of car labels that most attracts the attention of candidate buyers is the message.

To find that out, we ran a series of experiments in which individuals were asked to choose a car between two options. Each option consisted in a label depicting a car and a series of additional elements—logos, price tags and a series of messages. As the results show, all attributes are significant, and they all contribute to the choice. However, the strongest utility is given by the messages. Respondents were more likely to choose EULES-friendly cars when they the label shows information on lower costs or lower taxes. When it comes down to one car purchase decision, consumers are less interested in knowing about health-related benefits, convenient parking or access fees. A minor increase in price has a negligible effect. Finally, combinations of one message with other elements—EULES logo, CO₂ logo or both—within the same label have a small but positive effect on participants’ choices.

Ecolabelling is considered an effective strategy to encourage future consumers towards forms of sustainable consumption (Thøgersen et al. 2010). Compared to the more general literature on ecolabels and especially to that focusing on white goods and food, the study of car labelling is considerably less developed with only a limited amount of studies conducted to examine its effects (Choo and Mokhtarian 2004; Kurani and Turrentine 2002; Lane and Potter 2007; Loureiro et al. 2012; Noblet et al. 2006; Teisl et al. 2008; Teisl and Roe 2005). The first strength of the current study is that it established which factors are associated with the effectiveness of labelling cars. Ecolabels are considered as voluntary environmental and consumer policy instruments and therefore considered as a pragmatic option for policy developers. It is a form of choice architecture that does not affect the perceived autonomy of consumers, being an ideal example of a ‘nudge’. The second strength of the study is that we are the first that experimentally examined causal effects on consumption choices. Until now, there has been no experimental study conducted on the effects of car labels on consumers’ decision-making, although it is of great importance to reduce the negative health consequences of road transport (EEA 2015). Teisl et al. (2008) showed that it is important to test a model linking individual characteristics and label characteristics in order to create effective ecolabels that affect individuals’ perceptions of the eco-friendliness of products. Existing studies on ecolabels have been focusing on the relationship between individual characteristics and eco-behaviour or between label characteristics and eco-behaviour but not on the interaction between these two factors. It is therefore important to extend this knowledge by examining both individual characteristics and the manipulation of ecolabels simultaneously and test the effectiveness of the ecolabels. As Teisl et al. (2008) already have shown, it is highly relevant to examine underlying psychological factors and individuals’ priors of the product and of the environmental problem, because it plays a strong role for the long-run provision of eco-information, in particular in context where individual consumers hold incorrect perceptions.

One limitation of the current study is that participants conducted the survey and experiment online, whereby we did not control the situation in which they participated. This is a potential harm of internal validity. However, it could also be argued that such settings could reflect the real purchasing decisions to a larger extent. One could imagine that a typical consumer usually chooses its vehicle on the basis of a number of attributes by browsing the Internet, and subsequently, the person visits a car dealer to obtain further details. Additionally, future research could investigate the effects of ecolabels on consumption behaviour in real-life car shops. Secondly, only a limited amount of car types were used in the experiments so it is difficult to generalize the results to other type of cars. Nonetheless, the same mechanism that we have found in the current study seems to be applicable to other cars as well, in particular because we assessed public responses to possible incentives in general that might affect the perception of the ecolabels.

The current study is of great importance because it will provide the standard that could be presented to consumers either in a form of a label, a message or both. In this study, our research goal was to identify and test the elements which would nudge consumers towards an environmentally and health-friendly EULES car. Our findings will support governmental decision-making process with respect to the EULES format that should have the greatest impact on consumer purchasing decisions. In addition, we also assessed the likely consumers’ preference between the CO₂ label and the EULES label. The results will help to guide environmental conscious customers towards the purchase of vehicles with clean emission profiles. Furthermore, it will provide a benchmark for local or national authorities when developing financial or access and demand policies to promote clean transportation. Finally, it will provide an incentive for manufacturers to produce vehicles that deliver significant emission reductions on the road.