The developed model framework is used for analysing various types of economic instruments, which are assumed to give the consumers economic incentives to change their food consumption patterns into more healthy directions. The types of instruments analysed include
• various taxes on nutrients, which are crucial with regard to obesity and nutrition-related diseases: fats, saturated fats [49
] and sugar [50
• various forms of subsidies to nutrients, which are desired to be promoted: fruits and vegetables [51
], fibres [52
• revenue-neutral combinations of taxes and subsidies
7 different regulation scenarios have been specified, cf. table . In order to make the scenarios mutually comparable, all scenarios are scaled in a way that their implications for the net economic welfare are equal across the scenarios. The point of departure for this scaling is the welfare loss due to a halved VAT rate for fruits, vegetables and potatoes (from 25% to 12.5%). In the revenue-neutral combination scenarios, the net effect on tax revenue is equal to zero.
The taxes and subsidy rates on nutrients are assumed to affect the consumer food prices according to their content of these specific nutrients [see table ]. We further assume that food supplies are price elastic, implying that the tax and subsidy changes are fully transmitted to consumer prices. As suggested above, this assumption may be justified by the fact that Denmark is part of the large EU market with several competing food suppliers. On the other hand, a high degree of market concentration in some stages of the food supply chains (e.g. the retail stage), even at the EU level, may imply a less than full impact of tax changes on consumer prices. Hence, whereas the assumption of full transmission of tax changes to consumer prices may lead to some uncertainty regarding the absolute impacts of the different scenarios it is considered less crucial for the comparison between scenarios.
Subsidies based on the fibre content (scenario 2) will benefit most foods of plant origin, as opposed to a tax reduction on fruits and vegetables (scenario 1), which will only benefit consumption of these commodities. A tax on all fats (scenario 3) or saturated fats (scenario 4) will lead to price increases for most foods – and primarily for foods of animal origin. A tax on sugar (scenario 5) will almost exclusively have effects on the prices of sugar, sweets, cakes and fruit yoghurt. To the extent that consumers adjust their consumption patterns to changed price conditions, all the considered scenarios, except scenario 5, are expected to lead to shifts away from animal-origin foods towards plant-origin foods.
The consumer price changes induced by the scenarios are used as input to the economic model in order to determine the effects on the consumption of different food categories and the intake of different nutrients. The calculated effects of the 7 scenarios on food and nutrient intake are shown in table .
Effects on the intake of selected foods and nutritional components, per cent.
Subsidies to the consumption of fruits and vegetables, e.g. in terms of reduced VAT (scenario 1) will induce an increase in the consumption of these foods, at the cost of a range of other foods, including dairy products, eggs and fish. A subsidy to the content of fibres in the foods (scenario 2) leads to an increase in the consumption of fibre-rich foods: flour/bread, potatoes, fruit and vegetables, mainly at the cost of dairy products, eggs and fats.
At the bottom of table , the subsidies' calculated effects on the intake of fats, saturated fats and fibres are shown. Both subsidy scenarios have a reducing effect on the intake of fats and a stimulating effect on the intake of fibres. The results show that the choice of taxation object has implications for these effects. For example, the effect on intake of fibres is significantly higher if a tax reduction targets fibres per se, rather than fruits and vegetables. The effects of a fibre subsidy on the fibre intake are possibly underestimated, because the subsidy induces consumers to substitute low-fibre vegetables towards high-fibre vegetables – an effect that has not been accounted for in the calculations due to the level of aggregation in the econometric model.
A tax on all fats in the foods (scenario 3) leads to a reduction in the consumption of all food categories of animal origin, except eggs. The tax induces a relatively strong reduction in the consumption of fats (butter, margarine, oils etc.) and cheese, and to some extent also the consumption of meats, and these foods are replaced by fish, fruits and vegetables, bread and especially sugar. Thus, although the tax on fats has some desired effects on the consumption of fats, it also has some undesired effects in terms of the consumption of sugar. If a fat tax is only directed towards the foods' contents of saturated fats (scenario 4), the picture changes slightly, compared with scenario 3. The reducing effect on the consumption of fats and cheese (which have a high content of saturated fats) is 10–15 per cent stronger. In contrast to taxes on fats, a tax on sugar (scenario 5) only reduces the consumption of sugarvii, but induces increases in the consumption of other food categories, including the intake of fats.
The estimated effects on the consumption of individual types of foods in the tax scenarios (3–5) are presumably over-estimated due to consumers' option of changing towards e.g. more low-fat varieties of the individual foods. By shifting from e.g. high-fat milk products towards more low-fat milk products, the consumer may avoid part of the price increase due to the tax, and may thus be less likely to reduce the overall consumption of milk than the above results suggest. For example, by means of estimated detailed price elasticities, it has been calculated that the considered fat tax will lead to an 8–10 per cent reduction in the average fat content in consumed fluid milk, because consumers replace high-fat milk with more low-fat varieties [53
]. On the other hand, the effects on fat intake may be underestimated due to these within-aggregate substitution effects.
Combinations of tax reductions on fibres or fruits and vegetables on the one hand, and increased taxes on the most unhealthy fats on the other hand (scenarios 6 and 7) are seen to have desirable effects on the intake of fruit and vegetables, and thus the amount of fibres, while at the same time reducing the intake of fats and sugar. With regard to objectives of reducing the intake of fats and sugar and increasing the intake of fibres, scenario 6 is up to 40 per cent (and for sugar even more than 100 per cent) more effective than scenario 7.
The scenarios also have economic implications for consumers and the government budget, and hence for society as a whole. A measure of the welfare loss is the sum of lost consumers' surplus and net revenue losses for the government. The loss in consumers' surplus is measured in terms of equivalent variation, which measures the food budget change necessary to obtain the initial utility level at the changed prices, taking into account the changed composition of consumption. For instance, if the price of one commodity increases, there will be a need for a budget increase in order to obtain the same utility level as before the price increase. The estimated consequences of the considered scenarios on consumers' surplus, government revenues and economic net welfare are displayed in table .
Economic consequences of economic food policy instruments
As mentioned above, the 7 scenarios are scaled to yield the same welfare loss (41 million DKK – corresponding to 5.5 million euros – per year) in order to make the scenarios comparable. However, the distribution of this loss between consumers and the government sector varies considerably across scenarios. Thus, a general fat tax (scenario 3) implies a relatively large redistribution from consumers towards the public sector, whereas the redistribution effect of a sugar tax (scenario 5) is less, because the sugar tax affects a smaller share of the food budget.
As expected, the redistributive effect goes in the opposite direction in the two subsidy scenarios, where consumers gain while the government sector suffers a revenue loss. The extent of redistribution is larger for a fibre subsidy (scenario 2) than for the VAT reduction on fruit and vegetables (scenario 1), as the fibre subsidy concerns a larger share of the food budget than fruits and vegetables. Due to construction of the scenarios, the two revenue-neutral combination scenarios (scenario 6 and 7) only affect the consumers, and the welfare loss equals the loss of consumers' surplus.
As was the case with the consumption responses above, the indicated economic effects of taxes are probably over-estimated, because consumers to some extent are able to reduce tax payments beyond those represented in the price elasticities by shifting toward "light" varieties of the products, e.g. from whole-milk to skimmed milk or from high-fat towards low-fat cheeses [53
]. On the other hand, the revenue effect of a fibre subsidy is probably under-estimated, because consumers will tend to substitute towards more fibre-rich (and thus more eligible for subsidies) food varieties, when the prices of these are reduced as a consequence of the subsidy.
Consumers' possibilities for substitution between foods and other consumption goods are ignored in the calculations. To the extent such substitution takes place, the costs are over-estimated. However, this is not considered to have serious implications for the comparisons across regulation scenarios in the present context.
As mentioned, an assessment of the cost-effectiveness of the considered regulations is based on a comparison of the economic consequences in table with the nutrient intake effects in table . A difficulty in this respect is however the multidimensional character of the nutritional effects (fats, sugar, fibres, etc.). Which of two tax instruments is the most cost-effective from an overall perspective depends on the weighting of the respective nutritional effects. From table it is however seen that a combined regulation, where the instruments are specifically targeted towards the critical nutritional components (scenario 6), has a relatively strong impact on the intake of all the considered components. So even if a precise evaluation of the relative cost-effectiveness of the considered instruments is difficult, there seems to be no doubt that the cost-effectiveness is relatively high for this combination of economic regulation instruments.