According to this research paper German consumers are more likely to support a meat tax on animal products for animal welfare reasons.
A tax on meat could help address the climate impact and animal welfare issues associated with the production of meat.
The animal farming industry is in the public eye. Consumption and production of meat and dairy products and their consequences are discussed in society and politics alike. The livestock sector accounts for 14.5% of all human-induced greenhouse gas (GHG) emissions. Breeding and husbandry conditions, especially in intensive livestock farming, lead to animal diseases or painful disease-prevention measures such as tail-docking pigs
Working conditions in meat processing firms have also drawn increasing attention of policymakers, which has partly resulted in legislative amendments.
From a health perspective, meat consumption levels are too high in industrialized nations, leading to increased risks for colorectal cancer and cardiovascular diseases, and eventually straining public health systems.
Given the diverse deficiencies of the animal farming and meat production systems, policymakers are increasingly accounting for them, such as in the European Commission’s Farm to Fork strategy. Alongside setting stricter rules and standards for producers, one potential intervention could be the introduction of a tax on meat and animal products. Modelling studies show that taxing meat and animal products could have strong steering effects, thus improving public health and reducing the environmental impact.
In Germany, policymakers are discussing a tax on meat to address two of the issues named above, namely the climate and animal welfare aspects. In the context of introducing a carbon price for fossil fuels in the heating and transportation sector, the German Green Party suggested a climate charge on animal products. In addition, an expert commission set up by the then German Minister of Food and Agriculture suggested implementing a fixed animal welfare consumption tax, the so-called Tierwohlabgabe, on every kilogram of meat sold, with revenues intended to support farms in improving husbandry conditions.
In April 2022, the expert commission reminded the new government of its recommendation. The climate change and animal welfare debates are conducted rather independently of one another, although they concern the same industry and the same products. We therefore focus on these two aspects while acknowledging that there are other reasons to motivate meat taxation such as biodiversity loss, water pollution and health concerns.
The introduction of taxes on food is undoubtedly a political challenge, particularly in times of high inflation and globally rising food prices. Numerous surveys and choice experiments have examined individuals’ preferences regarding (carbon) tax schemes in general, and animal products in particular. Several policy characteristics have been found to increase public support, for example, refraining from calling the charge a tax, earmarking revenues, establishing progressive taxation and clearly explaining the tax’s impact.
In this article, we varied additional tax attributes to determine their impact on support for meat taxation. Motivated by the two justifications discussed by policymakers in Germany, we tested if support rates for a tax on meat differ depending on whether the tax is levied to mitigate climate change or to improve animal welfare. On the basis of previous findings on the effectiveness or stated importance of different reasons to reduce meat consumption, we hypothesized that support is higher for a tax aiming to promote animal welfare.
In addition, we compared two versions of a per-unit excise tax varying in their degree of differentiation. The uniform variant charges a fixed amount on every kilogram of meat sold, independent from the meat’s carbon footprint or the husbandry conditions. Examples for such a tax type are the proposed Tierwohlabgabe of the German expert commission and the German electricity tax. The second variant is, in the spirit of a Pigouvian tax, differentiated to represent differences in external damages associated with the product, such as alcohol or tobacco taxes and the German CO2 price on fuels.
Meat types with a higher carbon footprint in case of a climate tax, or produced by farms with poorer husbandry conditions in case of an animal welfare tax, are charged a higher tax rate per kilogram than those with lower emissions or better husbandry conditions, respectively. The two tax types are expected to affect consumption differently. A uniform tax primarily reduces meat consumption overall as it does not change relative prices within meat categories. A differentiated tax is expected to affect both the level as well as the composition of meat products consumed. The latter is due to increased prices of products associated with higher damages to other human and non-human beings. The additional steering effect of a differentiated tax helps to reduce these damages and is hence typically considered to better improve human and animal welfare compared with a uniform tax. We tested whether voters appreciate the Pigouvian idea once all other tax attributes, including earmarking of revenues, are held constant.
We presumed voters’ perceptions of the tax’s impact on consumption patterns to affect support rates. While there are, a priori, no reasons to expect that the justification of a tax influences its impact on consumption patterns, we would anticipate such effects for the degree of differentiation. However, whether consumers anticipate this difference and how it might affect their stated support remains to be seen. Research on the acceptance of congestion charges, waste taxes and a carbon tax finds that trial periods increase support and people update their beliefs regarding the tax. Thus, we tested whether varying the salience of expected behavioural effects on consumption affects support rates. If participants anticipate the stronger steering effect of a differentiated tax and appreciate it, then higher support rates would be expected if this is made more salient. We increased salience for a subgroup by asking participants to reflect upon the tax’s potential impact on consumption behaviour before eliciting their support.
We addressed all three attributes discussed above in a referendum choice experiment, in which a sample representative of the German adult online population was asked to vote on a tax on meat. The referendum setting was chosen as previous studies find that referendum surveys are externally valid and incentive compatible if perceived to be consequential. To increase consequentiality, participants knew that referendum results of this study will be sent to the committees of the German parliament responsible for agriculture and the environment, allowing policymakers to update their beliefs about public support for a tax on meat. We randomly assigned participants to one of two tax purposes (animal welfare versus climate), one of two tax types (uniform versus differentiated tax) and one of two salience levels (low versus high salience of the tax’s effect), that is, eight treatment groups. Within subjects, proposals differed only in tax level, which was gradually increasing from the first to the last proposal. Participants had to make a decision on six consecutive proposals.
Our results contribute to the delicate topic of how to reduce meat consumption as one of the big societal, environmental and ethical challenges humanity faces. As the paper focuses on public support and, in particular, on hypothetical voting in a referendum, the approach is, by design, anthropocentric, as only the preferences and values held by participants drive the results of the study. The paper is not concerned with why society should tax meat, but rather on how specific features, including justifications, affect support rates for such a tax. We add to the literature on instruments to influence meat consumption, more specifically on what affects people’s support for the rather heavy-handed fiscal intervention of a tax on meat. This complements studies looking into consumers’ preferences regarding information provision and labels on meat products.
By considering two different rationales for a tax on meat (climate protection versus animal welfare), we broaden research on the acceptance of carbon taxes by the animal welfare aspect. We thereby address different arguments for meat taxation as requested by Fesenfeld et al. and extend the findings by Fesenfeld et al. on willingness to pay for a tax for animal welfare, climate, local environment and health frames in Germany. Moreover, we complement the emerging literature on the link between tax support and use of tax revenues. We also take current policy discussions into account by comparing a Pigouvian tax – which is usually favoured by economists – to a uniform tax debated in Germany. Empirical evidence on whether support rates differ between a uniform and a differentiated tax on meat remains limited.
Tax justification and level drive support rates
We tested pre-registered hypotheses on the impact of the attributes of a tax on meat on support by voters. The attributes considered are: tax level and differentiation thereof, justification and salience. As in a real referendum, we counted only valid responses, that is, Yes and No votes. The support rate thus equals the share of Yes votes among valid votes. Rates of abstention are similar across all tax levels and schemes, ranging from 6% to 8%
The percentage of votes in favour of the proposed tax on meat monotonically decreases by 2.6 percentage points (here, and in the following, we report 95% confidence intervals (CI) (−2.49 pp, −2.78 pp)) for each €0.10 kg increase in the tax rate. The average support rate is 62% at the lowest tax level of €0.19 kg, corresponding to a carbon price of €25 t CO2. At this level only, every proposed tax scheme would receive a simple majority. Support monotonically decreases in the tax rate and reaches on average 23% at the highest tax rate of €1.56 kg, corresponding to €200 t CO2. This confirms our hypothesis that support is decreasing in the tax level. Fifty per cent of participants would still support a tax level of €0.39 kg if linearly interpolated.
Support for climate-justified taxes is significantly lower than for otherwise identical animal welfare-justified taxes across all tax levels. On average, an animal welfare tax receives 11.1 percentage points (8.3 pp, 14.0 pp) more Yes votes than an otherwise identical carbon tax. This again is in line with the pre-registered hypothesis. All estimates are similar and highly statistically significant across models. Interestingly, the degree of differentiation of the tax has at most a minor and statistically not significant impact on support rates (β = 0.024, (−1.6 pp, 6.4 pp)), which counters our hypothesis.
High salience increases the support rate by 4.0 percentage points (0.0 pp, 8.1 pp). Participants who were induced to think about the potential effect of the proposed tax before they vote are thus more likely to support the scheme. However, we find no significant interaction between salience and the degree of differentiation (β = −0.007, (−6.5 pp, 5.0 pp)). The interaction term is close to zero and statistically insignificant. Counter to our pre-registered hypothesis, the effect of a differentiated tax is not more pronounced in the case of high salience.
Expected tax impact varies by justification and differentiation
We conducted an analysis of participants’ beliefs about the behavioural impacts of the tax schemes. This analysis is exploratory given the hypotheses tested were not pre-registered. It aims at providing insights on what might drive the main results presented in the previous section. Participants stated their expectation about the market-wide development of meat consumption if the proposed tax scheme was to be implemented. Figure 3 shows average marginal effects on the probability of choosing the three possible answer categories (decrease, remain the same or increase) from generalized ordered logistic regressions for overall meat consumption and consumption in the subcategories beef/husbandry level 1, lamb/husbandry level 2, pork/husbandry level 3 and poultry/husbandry level 4, respectively.
Looking at tax types, we find that participants expect the differentiated tax to be significantly more effective in steering meat consumption towards lower-impact meat compared with the uniform one. For the two meat types/husbandry levels that are taxed the most under a differentiated tax, the probability of choosing ‘decrease’ is significantly higher for those facing a differentiated rather than a uniform tax. The opposite applies for the two meat types/husbandry levels that are taxed the least under a differentiated tax. Looking at answer option ‘increase’, the marginal effects are reversed. In addition, participants expect overall meat consumption not to be impacted by the degree of differentiation, which is consistent if effects from the four subcategories cancel each other out.
Regarding the tax’s justification, we find that participants expect the climate tax to be significantly more likely to decrease consumption in all meat type/husbandry level subcategories compared with the animal welfare tax. Even if we look at the uniform tax subsample only, we find the same differences. For a uniform tax, prices of all meat products on the market rise by the same amount, independent of whether the levy is raised for climate or animal welfare purposes. Thus, effects cannot be driven by perceived or real differences in the market shares of husbandry/meat type categories or different degrees of substitutability between them. Moreover, participants do not expect a significantly different effect of the climate tax compared with the animal welfare tax for overall meat consumption, which contradicts responses for the subcategories of consumption.
Our study provides important insights for policymakers on how to design a tax on meat to receive public support. First, supported tax levels are found to be rather low in our experiment. At the lowest tested tax rate of, on average, €0.19 kg (equivalent to €25 t CO2), a simple majority of participants votes in favour of a tax on meat in every tax scheme suggested. For the second lowest tax rate of, on average, €0.39 kg (or €50 t CO2), only taxes justified by animal welfare win a referendum. This level of an animal welfare tax matches the proposal by the expert commission reporting to the previous German government. Thus, the proposal is backed by voters at the time of the experiment. We acknowledge that support for the actual tax rates tested represents a snapshot given participants’ current disposable income, recent societal debates and other structural and individual factors. Nonetheless, given that the rate of support for a tax on meat is strongly decreasing, in particular, at the lower end of the range tested in our study as well as in the extant literature, we recommend starting with a low rate when introducing a tax on meat. Following a ratcheting-up strategy is likely to receive more support than trying to go full scale initially. However, more research is needed to determine the exact relationship between (dynamic) tax rates and public support.
We find that participants are more willing to vote for a tax if its purpose is to improve animal welfare as opposed to reducing the climate impact of meat products. This complements results from (choice) experiments and surveys on labels and information provision, in which animal welfare arguments are found to be more important or effective in inducing intrinsically motivated behavioural change than climate protection arguments. The stronger appeal of animal welfare motives is also present in the context of the more intrusive intervention of a tax on meat.
Our result is, however, in contrast to Fesenfeld et al. who find no significant differences between the two framings. This difference in findings might be driven by the naming and description of the tax schemes in the two studies. They tested how different independent frames (climate change mitigation, animal welfare and health benefits) affect support for a tax on meat. In contrast, we made the frame explicit in the tax name, calling it ‘animal welfare’ or ‘climate levy’. The explicit framing in the tax name might send a more credible signal to participants that animal welfare is actually addressed with the tax, increasing support. Moreover, the lack of information in Fesenfeld et al. on how the tax revenues would be spent might have substantially reduced support for a tax in their study, and hence made it more difficult to detect differences between frames. In contrast, we stated that tax revenues are earmarked and provided detailed information on which meat types or animal welfare levels are taxed and why. Especially for animal welfare, voters might be more supportive if they have a concrete idea of how animals might benefit from a tax. While earmarking seems to be less important when considering combined support of several food policies, it is found to be a crucial success factor for acceptance of a stand-alone carbon tax, and hence maybe also for a stand-alone tax on meat. The design of our study does not allow distinguishing between the framing and the earmarking aspect. Specifying their relative importance is left for future research.
Surprisingly, participants seem to attribute a stronger steering effect to a climate tax compared with an animal welfare tax, even if they are identical in all other respects. We can only hypothesize why this is the case. Preferences for animal welfare taxes might not be driven by beliefs in their ability to reduce meat consumption, but potentially by beliefs in their effectiveness of promoting animal welfare independent of the amount of meat consumed. This is plausible if consumers consider the lives of farm animals to be worth living and are not primarily concerned about the fact that animals have to be killed to produce meat. Moreover, participants might expect additional individual benefits from paying an animal welfare tax because they associate healthier or tastier products with higher animal welfare standards. In the latter case, participants would consider animal welfare not only as a public good, but would also derive private benefits from improving rearing conditions (for similar thoughts regarding labelling antibiotic use on meat products, see refs.). Future research could look into drivers behind preferences for an animal welfare tax. For policymakers, this shows that justifications matter, potentially more so than expected impacts on behaviour. Our study does not shed light on the question of whether combining justifications (and splitting revenues) would improve or weaken support for the measure.
Our findings show that the degree of differentiation does not play an important role in shaping support for a tax on meat. Simulation studies in other contexts, namely sugar-sweetened beverages, suggest that a differentiated tax is more effective in reducing externalities. As answers to the belief questions show, participants on average understand the mechanism behind a differentiated tax and also expect a stronger steering effect from this tax type. However, we only find a minor and mostly statistically insignificant positive effect on support compared with a uniform tax. Raising the salience of the stronger steering effect has no impact on support rates. We conclude that voters might well understand that Pigouvian taxes are more effective in changing consumption patterns than uniform ones, but that they do not appreciate this. This finding is in line with empirical results by Kallbekken et al. who find that support rates for a Pigouvian tax in a laboratory experiment do not increase if participants are informed about its benefits. Our results confirm their findings and extend them in two directions. The lack of a significant interaction effect between raising the salience of a proposed tax scheme and the degree of differentiation is analogous to their observation that educating participants about the additional steering effect does not systematically change support rates. This builds our first extension, that is, that, on average, participants are able to qualitatively anticipate the steering effect of differentiated taxes in a more complex real-world setting without being educated about them by the experimenter. Second, we directly compare support for a Pigouvian with support for a uniform tax. Our results show that adding a steering effect does not increase support rates compared with a tax that is identical in all other features. Overall, the results substantiate the point that the indifference found between uniform and differentiated taxes is not primarily driven by participants who do not understand how the tax schemes differ, but it is rather caused by a lack of caring about this difference. This provides relevant insights for policymakers. The indifference between Pigouvian and uniform tax is at least partially good news for them as there is low risk in implementing the more effective differentiated tax. The recommendation is weakly backed by comments in the Remarks fields of our survey. Thirty-eight participants who had been assigned to a uniform tax treatment criticize the lack of differentiation or state that they would prefer a differentiated tax. On the other hand, only one participant in the differentiated treatments asks for a uniform tax.
Given that we find a positive effect of high salience on support rates, we additionally recommend communicating the tax’s desired behavioural impact very clearly to win the public over. Our result supports previous findings that experiencing the effect of a tax in trial periods makes people more likely to support it if thinking the effect through is indeed a proxy for such a trial experience.
To conclude, there is support for a tax on meat in Germany, but only under certain conditions that policymakers would benefit from taking into account. The version recently suggested by a government-installed expert commission meets these criteria, but more effective taxes would also be supported by voters. While we focus on Germany, other countries have been, or are currently, discussing different forms of a tax on meat as well. In October 2022, New Zealand’s government proposed to price livestock emissions at the farm level with revenues used to support farmers in their efforts to reduce emissions. This corresponds to the differentiated tax treatment in our study as the price impact will differ in line with the emission intensity of meat types. In the Netherlands, policymakers presented concrete proposals to implement a tax but so far have not been able to convince a majority in parliament. In the UK, meat taxation was discussed, but despite being found to have a substantial potential impact on GHG emissions and public health, it was explicitly disregarded from the National Food Strategy published in 2021 due to potential lack of acceptance among citizens. The Danish Council of Ethics, a Danish think tank, recommended a tax on red meat for Denmark in 2016, which was refused by politicians. Our findings could be particularly relevant for the failed proposals by checking if the taxes could have been defined or framed differently. Future research could leverage our design and compare support rates internationally.
Original source: https://www.nature.com
