New figures reinforce the role taxes could play in improving diets, but which approach is best?

Recipe for Change has identified two approaches that could be taken for a new levy on food, building on the success of the Soft Drinks Industry Levy (SDIL).

1. An ingredient levy, as designed by the National Food Strategy (NFS), hereafter “NFS ingredient levy”. This policy option proposes an upstream sugar and salt levy on manufactured food sold in retail, restaurants and catering. The tax would be applied per kg of salt and sugar in products.
2. A category levy, for instance like one designed by Nesta, hereafter “NPM category levy”, focused on key categories of less healthy food products sold in retail for home consumption,  with an NPM score of 4 or higher that are also in scope of the existing regulations on product placement, promotion and advertising. This is just one option for a category focused levy (see further here).

Working with London School of Hygiene and Tropical Medicine (LSHTM), Recipe for Change has modelled the potential impact that could result specifically from the potential salt and sugar reductions achieved from the two levy options.

In the case of the NFS ingredient levy, we have modelled the health impacts of the sugar and salt reductions based on previous work by IFS. In the case of the NPM category levy we have estimated the potential sugar reduction based on the calorie reductions published by Nesta and then modelled the health impacts alongside salt. This analysis does not capture potential health impacts from calorie reductions beyond sugar (e.g. saturated fat), and therefore likely underestimates the total benefits of an NPM category levy. 

How do the impacts on calories compare?

Overall, both levy approaches have comparable impacts on health and obesity, with significant reductions in food-related ill health such as cardiovascular disease. The NFS ingredient levy model is broader in scope, applying to all salt and sugar used in manufactured food in a greater range of settings, rather than a specific category of foods, and so as expected, when comparing the health impacts from salt and sugar reductions specifically, the population health impacts are more far-reaching.

Both approaches would reduce calorie intake, between 13 and 26 kcal per adult per day (from sugar reduction only) depending on the exact scenario and model used for the NPM category model, and between 15-38kcal per day for the NFS ingredient model.  Whereas Recipe for Change analysis looks specifically at salt and sugar-specific reductions, Nesta’s calorie-based modelling for the NPM category levy looked more broadly at NPM score reductions, estimating calorie reduction of between 37 and 75kcal per adult per day, in turn reducing the prevalence of adult obesity by up to 24% in the UK. These results cannot be directly added due to overlap in the Recipe for Change sugar and Nesta calorie estimates.

We see similar calorie reductions between the NPM category model, and a category version of the NFS levy that Recipe for Change and LSHTM previously modelled, whereby a levy is applied to the sugar used in a selection of discretionary food categories that contribute a high amount of sugar to our diet. 

How do they diverge on nutrient shifts?

Overall, the biggest divergence in the impacts of the two policy designs is on the shifts in specific nutrients including salt, sugar and saturated fat.

The NFS ingredient levy has greater impacts on sugar and salt specifically, which is to be expected due to the design which targets these ingredients. Reductions in these nutrients, alongside calories, have been previously prioritised by government as part of voluntary reformulation programmes.

In the NPM category model, the sugar reductions are still substantial, though lower than for an NFS levy. Salt reduction is particularly low in an NPM category model – this is to be expected due to the tax design focusing on categories of food previously identified as contributing high amounts of sugar and calories to children’s diet, which are not typically the same categories contributing the most salt.

Meanwhile, in the NPM category approach, it is to be expected that companies would also reduce saturated fat levels in food as part of reformulation efforts to lower their NPM score and therefore levy amount, which would not be expected with an NFS ingredient tax focused specifically on salt and sugar. As noted previously, saturated fat is also an important consideration when looking at diets and therefore policy design. The Recipe for Change analysis of the NPM-category levy does not capture potential health impacts from calorie reductions beyond sugar (e.g. saturated fat), and therefore likely underestimates the total benefits to health.

How feasible would these levies be to implement?

The NPM category levy has the advantage that both NPM and the categories included are covered by existing policies, including the promotion and placement, and advertising restrictions, meaning that policy alignment is clear and easier to legislate. Conversely, it does little to improve the healthiness of foods that currently sit out of scope of these categories, such as chocolate spread, cake and ice-cream decorations and savoury crackers.

The NFS ingredient tax is based on sugar and salt as ingredients, which offers a degree of simplicity for application. However, this could also create confusion about which stage of production the tax should be paid and by who, as the inclusion of salt and sugar in a product may take different forms and be added at various stages.

What about the out-of-home sector?

Existing nutrition policies, such as restrictions on multi-buy price promotions, largely focus on food sold within retail environments. However, incentivising improvements in the out of home sector, which contributes to approximately 14% of an adult’s overall calories and is often relatively high in fat, salt and sugar, is an important part of shifting diets.

Based on the designs explored here, the NFS ingredient levy would be applied to out of home in the same way as it is applied to food sold in retail and the feasibility of doing so would not change as it is based on the ingredients.

The NPM category model as proposed does not apply to out of home, although the design could be applied in principle to out of home. The challenge here would be the availability of information to calculate the NPM as nutrition information is less widely available in the out of home sector.

How strong is the incentive of each approach?

Both levies are likely to incentivise reformulation across a range of products but would work in different ways. An advantage of the NPM category approach, is that there is a ‘get out’ built into the design, in that the levy is only applied to products with an NPM score of 4 and over. Similarly to the SDIL, this creates a target for reformulation. Above this threshold, manufacturers reduce their tax burden for every 1 improvement in the NPM score and therefore even small changes, up to the threshold, would be incentivised.  However, for some categories (such as chocolate) the reformulation success to date and potential for reformulation is more limited.

In the case of an NFS ingredient levy there is no tax-free threshold. However, fact that all categories would be impacted means that there is incentive for improvement across an even greater range of food people eat, and there is also greater incentive for salt reformulation.

How much revenue could be raised?

A levy on unhealthy food is intended to encourage reformulation and therefore benefit health. However, where reformulation does not occur, a levy has the potential to raise tax. Under the NPM category levy, it is estimated that the government could raise up to £1.6 billion annually in direct tax revenue. The NFS ingredient levy could raise £2.9–£3.9 billion per year (£2.3–£2.8 billion from sugar and £570–£630 million from salt).

The revenue raised from either levy provides an opportunity to invest in nutritional safety nets that improve diets, particularly for low-income families, such as the Healthy Start Scheme or getting free fruit and vegetables into schools, which is particularly important in light of recent and projected food price increases in the coming months.

Both approaches also have the potential for food price rises, depending on the level of reformulation that happens. Nesta's model assumes manufacturers pass the full tax cost to consumers, though in practice they may not. If they did, an NPM category tax would raise the average home food basket by around just 1%, with increases limited to less healthy products. The NFS ingredient levy analysis does not assume full cost pass-through, but predicts price rises on foods not fully reformulated of 16–20p per adult per day. In either scenario, lower-income households - being more price-sensitive - are most likely to cut consumption of taxed foods.

Summary of strengths and weaknesses of the two approaches

The best way forward?

Both models show that taxes could play a role in incentivising dietary change and improving health, reducing diet-related chronic disease and improving life expectancy. The exact approach will depend on what the priorities are, weighed against the feasibility of the different pathways to reducing ill health and improving diets. The public support is there, with nearly three-quarters of the public supporting bringing in regulations on the amount of sugar, salt and saturated fat that is allowed in food and drink.

Overall, the evidence base presented by the Recipe for Change coalition indicates that fiscal levers should be considered by Government as part of their package of policies to improve nutrition and health across the population.