A Model of Welfare and Cost

The data for this model mainly came from the following sources: Säll and Gren (2015), Säll (2018) and Springmann (2018). The author only chose peer-reviewed sources that were also providing data relevant to Sweden for this analysis.

As discussed in Chapter 3.1.3, one way for governments to prevent a consumer's loss in utility is by compensating them the amount needed to propel them back onto their original utility curve. This thesis, therefore, uses the compensating variation (CV) estimates of Säll (2018) to establish a potential loss of consumer welfare. Säll's research is based in Sweden, making her research the most relevant for this research paper.

Compensating variation calculates the monetary amount needed to compensate the consumer surplus loss and keep utility before and after taxation at the same level. The research by Säll (2018) uses the Hicksian demand, which means that only the substitution effect and not income effect is analysed. Tax levels and elasticities are generated from her previous research with Gren, conducted in 2015 (Säll & Gren, 2015).

Household consumption data was drawn from Vingren and Kruse's data for the years 2007-2009, and the income levels were split into quintiles, and then recalculated into per capita levels, to be better compared to the given meat consumption data. An important side-note is that the referenced paper did not receive grants from any agency, and hence is likely to be more objective.

The values for the environmental damage costs are taken from Säll and Gren's 2015 research also based in Sweden. They analysed and included not just GHG emissions, but a variety of pollutants, namely GHG, nitrogen, ammonia and phosphorus. They do note, however, that their estimates on pollutants are likely to be lower than the actual emission generated by meat production. The taxes, which should equal the externality cost as they are based on Pigouvian taxes, were calculated by computing the average damage cost from each pollutant multiplied by the average emission per kg of meat.

Transforming the cost into monetary terms was done by Säll and Gren by using measurements from a variety of sources, such as the IPCC report, and recalculating them accordingly. GHG emission cost was based on Sweden's political cost, due to a lack of consensus in previous research (Säll & Gren, 2015).

Finding a source of the externality cost of health was more complicated. This may be due to the increasing complexity of understanding dietary-related illnesses, as well as a general lack of awareness and interest. Interestingly, this can also be noted in the questionnaire responses, where, while people are somewhat aware of meat’s dietary impacts, the knowledge level was lower than that of environmental impact. Ultimately, the author decided to use Springmann et al. (2018) study of health motivated taxation on red and processed meat. The study was chosen due to it being based on highly reliable meta-analyses. The same analyses were also used by researches of the WHO, the World Cancer Research Fund and the American Institute for Cancer Research. Furthermore, the research also included data directly relating to Sweden (Springmann, et al., 2018).

The external sources of Springmann et al. (2018) were all based on meta-analyses or extended meta-analyses. They first estimated the mortality and disease risks burdens on society, also noting weight-related risks, then estimated the costs imposed on health care and society. This was done by using cost-of-illness approaches, estimating medical health care costs and loss of working days in the European Union. Furthermore, the cost of death, using EU statistics, was assessed, as well as productivity losses for those in working age. While, the paper only analyses the health risk and related cost of red and processed meat; ignores poultry, and does not differentiate between pork and beef, it was still chosen due to its research methodology and for lack of better quality data.

4.4.2. Data Manipulation and Adaptation

This paper uses the calculated CV as a measure of welfare loss to be compared to measures of externality cost. To effectively compare, certain adjustments had to be made. The compensating variation as analysed by Säll (2018) was given for households of different income groups and per year. To be comparable to the environmental damage costs per kilogram of meat, the household levels were transformed to per capita levels and further divided by the average amount of meat (in kilograms) consumed. The CV was calculated for an average income level.

Figure 25: Modification of data and computation outputs for the compensating variation data.

The CV per meat type was calculated from the CV shares, as indicated by Säll (2018) in her paper. The weight (in kg) consumed were further taken from Säll. The final per kilogram CV is a computation of the annual CV per meat type, divided by the average amount of people per household and the average amount consumed of this meat type per year. Some calculation outcomes can be viewed in Figure 24.

The environmental damage costs did not have to be modified, as they were already given in per kilogram measures. The health damage costs were taken from Springmann et al.

(2018). They calculated the tax value of a cost-compensating tax as for red and processed meat. This number is not well fitted to the model, but for lack of more precise data, and showing some insights, was used. It was translated from dollar amounts to SEK. Any further modification was not deemed justified, as it could only be done on the author’s guesswork and would reduce the reliability of the study.

4.4.3. Data Analysis and Results

The result of the computation is split into two parts. First, the environmental damage cost and the CV are compared. The results show that clearly the environmental damage costs to society per kilogram of meat are greater than the compensating variation needed to bring the consumer back to the same level of utility.

Figure 26: A summary of the results: CV as opposed to environmental and health damage cost.

This result is valid for the kilogram levels of meat consumed in 2009 and may very well be different with lowered or heightened consumption. It is also relevant to note that for poultry, the cost and the welfare loss are almost equal, and that beef has the most significant discrepancy, which also goes in line with the previous literature review.

Figure 25 gives more detailed insight, and Figure 26 visualises these findings.

Figure 27: CV is significantly lower than environmental damage cost for beef and pork, and slightly lower for chicken in Sweden.

As stated, assessing the health damage in a comparable matter was not possible.

However, it is interesting to note that measuring the health damage in Sweden stemming from red and processed meat is nearly as costly as the kilogram environmental damage of beef, which is further visualised in Figure 27. Whilst there would need to be further research and measurement done to present more comparative data, these results already indicate the relevancy that future research could provide.

- SEK 5,00 SEK 10,00 SEK 15,00 SEK 20,00 SEK 25,00 SEK 30,00 SEK 35,00 SEK

Beef Pork Chicken Total

CV and Environmental Damage Cost

per kilogram CV - for average income groupe 2009 per kilogram Environmental Damage Cost

Figure 28: Health damage cost for processed and red meat are nearly as high as the environmental damage costs of beef per kg.

A sensitivity analysis was performed on the different income groups’ compensating variation. The results for the income groups revealed that the smallest income group would require significantly more compensation if meat were taxed than the average and highest income group. In the case of chicken, the welfare loss even is higher than the gain from a reduction of environmental damage costs. This goes in line with Säll, Caillavet, and other researchers conclusions that food taxes, and in particular a meat tax, are regressive (Caillavet, Fadhuile, & Nichele, 2019; Säll, 2018). The full model can be found in Appendix C.

Figure 29: Sensitivity analysis – different income groups require different compensating variation.

- SEK

per kilogram CV - for average income groupe 2009 per kilogram Environmental Damage Cost

per kilogram CV - for average income groupe 2009 per kilogram Environmental Damage Cost per kilogram CV - for smallest income group per kilogram CV - for highest income groupe