Our analysis uses nutrient profiling - a holistic approach to evaluating the healthiness of foods. In this analysis, we have used the FSA/Ofcom Nutrient Profiling Model. This model was developed to identify products that should not be advertised to children. It assigns products a score by adding points relating to calorie, sugar, saturated fat and salt/sodium density, and subtracting points relating to protein, fibre and fruit, vegetable and nut content. Higher NPM scores on the raw scale indicate unhealthier products and scores can be any integer between -15 to 40. A raw NPM score of 4 is the threshold used to define a product as ‘unhealthy’ according to current Government regulations. NPM scores are regularly used in academic research as proxy measures of healthiness.

The raw FSA/Ofcom NPM scale can be difficult to interpret as it ranges from negative to positive numbers with a lower score indicating a healthier product. Therefore, to ease the interpretability of the NPM we have scaled the raw NPM scores to be between 0-100, where the healthier the product, the closer to 100 the score is. To do so, we followed a formula developed by the University of Oxford which involves multiplying the raw NPM score by -2 and adding 70. Using this formula, the raw NPM score of 4 is equal to a converted NPM score of 62 (the threshold for a low converted NPM score or ‘unhealthy’ classification).

The NPM was originally developed to assist binary decisions on whether a product should be classified as high fat, salt or sugar (HFSS) and therefore not suitable to be advertised to children. Since then, legislation based on the HFSS classification (and the underlying NPM) has undergone proposed expansions to restrict the placement of HFSS products in prominent in-store and online locations and the promotion of HFSS products by price or volume. As a result, all major UK retailers will be familiar with the NPM score metric to ensure they are complying with existing and potential future HFSS legislation. Aggregating NPM scores across all products sold by a retailer rather than applying the score to individual products is a new use of the metric. Nevertheless, NPM scores are regularly used in public health, industry and academic research as proxy measures of healthiness (see source 1, 2, 3). It should be noted that whilst NPM scores can be used as a proxy for healthiness, the relationship between NPM scores and healthiness is not necessarily linear. For example, a product with an NPM score of 2 is not necessarily twice as healthy as a product with a score of 4. However, we can say that overall the product with score 4 has a nutritional profile that is less healthy than that of score 2.

We stress tested our headline finding to measure how robust it is to variations in the main sources of uncertainty in our methodology. These sources include assumptions used in the model as well as uncertainty around the calculation of NPM scores. We found that when key factors in the model are adjusted by different amounts, the resulting outcome varies between 20.1% and 25.3%, with most outcomes concentrated around 22.5% and 24.3%. We conclude that the figure of 23% reduction in obesity prevalence is robust to these variations up to 1%.

It is possible for a retailer to utilise the mechanisms at their disposal (reformulation, new product development, promotions, stock availability, placement and advertising) to shift consumer purchasing to baskets which are of equal value but lower in calories purchased in line with the target. For example, we have created (as an example from the data) a weekly shopping basket for an individual living with obesity:

White bread (800g), cinnamon swirl (2), potatoes (1kg), carrots (1kg), broccoli (375g), chocolate squares cereal (330g), angel cake (240g), McCain frozen chips (1kg), Rich Tea biscuits, veggie burgers (400g), crisps multipack, pork sausages, eggs, berry yoghurt pots.

With just a few changes in the basket, nudged by the retailer rather than conscious choices by the consumer, their basket could change to the following:

Wholemeal bread (800g), almond croissant (2), sweet potatoes (1kg), carrots (1kg), broccoli (375g), chocolate crisp cereal (400g), cherry cake (240g), McCain frozen chips (1kg), Rich Tea biscuits (reduced sugar) 300g, spaghetti bolognese (400g), crisps multipack, chicken sausages, eggs, reduced fat yoghurt pots.

This new basket (which is an example of a weekly shop for one person) is exactly the same price but contains the equivalent of around 80 fewer calories per day (the calorie reduction we would expect to see from retailer action to meet their target). Note this example is illustrative only.

While we recommend that this current policy apply to large retailers, targets could also be extended to the out-of-home (OOH) sector. The OOH sector also plays a significant role in the health of the nation, contributing approximately 20 to 25% of our consumed calories with major chains contributing meals with significantly greater calories than products consumed in the home. The next phase of the research will investigate options for improving the overall healthiness of the out of home sector via a similar targets-based policy.

To appraise the broader economic costs of targets to retailers and consumers, we commissioned an economic assessment of the policy. This assessment found that the proposed target should not have a significant impact on either the costs to retailers or the costs of consumers’ shopping baskets. Consumer costs are unlikely to rise for two reasons: 1) the policy will carry very limited costs to businesses and 2) the highly competitive nature of grocery retail means businesses would not be able to pass on any costs that are incurred to their consumers. Business costs will be limited because retailers have the flexibility to adapt their existing practices and operations most cost-effectively, such as making strategic stocking and marketing decisions, reformulating unhealthier products, or altering their store layouts. While the policy will incentivise businesses to modify these practices to prioritise health, with a suitable transition period it is not generally expected to increase their overall cost or subsequent costs for consumers. Furthermore, due to the highly competitive nature of the grocery retailer sector where many retailers are already close to compliance with the target, it is unlikely that a business would pass on any costs that are incurred to the consumer as they would face a very significant risk of losing profits and market share.

A further analysis also shows that there is no correlation between current retailer healthiness and their food costs indicating that healthier supermarkets are not necessarily the most expensive. This means that there is considerable potential for most retailers at least to improve their overall nutritional score without increasing costs to consumers.

We specifically designed this policy to only apply to the largest grocery retailers as they sell a diverse but relatively similar range of products and account for most of the population-wide calorie consumption. This means that the degree of healthiness across stores is relatively consistent as demonstrated by Figure 1. The healthiness scores across in-scope retailers only range from approximately 63 to 68.4 with nine of the 11 supermarkets already within two points of the target.

We commissioned an economic assessment of the policy to explore its cost impact on businesses and consumers. This assessment evaluated whether higher health scores (as defined by the nutrient profiling model scores) were associated with higher prices by comparing the average price per kilogramme of over 70,000 food products across the 11 largest UK grocery retailers to their health scores. This assessment found that there was no correlation between retailers average food prices and their health score suggesting that it is currently no more costly for consumers to shop in retailers with higher nutritional scores than in others. For more information, see The Economic Assessment of Health Targets for Retailers.

We propose that the Food Standards Agency (FSA) hold the powers to enforce compliance with targets and impose penalties, leveraging its main objective to protect public health and consumer interests in relation to food. The Department of Health and Social Care should oversee the design of the policy, and then pass responsibility to FSA for enforcement. FSA would need to work with and report progress to the Department of Health and Social Care who will continue to own the wider health improvement programme. For more information, see Retailers Targets: Implementation Plan.

This policy will require legislation with three key components that will be necessary to establish:

• Mandated data collection from grocery retailers to monitor performance and progress
• Mandated targets and penalties attached to these for non-compliance
• Enforcement powers given to a suitable organisation, such as the Food Standards Agency

We consulted the legal firm, Kingsley Napley LLP, for their view on the most appropriate legislative mechanism and their advice is incorporated throughout our policy proposal. Based on the legal advice we received, it is recommended that this policy is implemented through primary legislation. This is the most effective route to achieve all of our desired policy objectives. And while it may be possible to achieve some of the components of this policy without the need for new primary legislation, we believe the weaknesses of that approach outweigh the benefits.

The concept of outcome-based regulation centred around targets has been used in the field of environmental improvement and climate change prevention. The UK has used targets to drive positive action on the environment and climate change. Through the Climate Change Act, the UK government has set a target of reducing greenhouse gas emissions by at least 100% of 1990 levels (net zero) by 2050. The Act also established the Committee on Climate Change (CCC) to ensure that emissions targets are evidence-based and independently assessed. The UK Government has also proposed a Zero Emissions Vehicles (ZEV) mandate requiring an increasing share of zero emission cars and vans, alongside a CO2 emissions regulation to ensure that new non-zero emission cars and vans do not become more polluting in future years. ZEV mandates have successfully spurred greater ZEV availability and choice for consumers and higher uptake in the United States, Canada and China. The proposed minimum ZEV target trajectory for new cars sold begins at 22% in 2024, increasing to 80% in 2030 reaching 100% in 2035.

In the context of dietary health, the best examples of mandatory targets appear internationally with mandatory salt reduction targets in South Africa, Argentina and Vietnam. Mandatory salt reduction targets aim to combat excessive salt intake and its associated health risks. These regulations set limits on the amount of salt allowed in various food products, requiring food manufacturers to reformulate their offerings with less salt. These countries employed a phased approach, with gradual reductions over time, allowing the food industry to adapt. Early evidence from South Africa indicates that the programme so far has had a significant impact in reducing population salt intake by approximately 1.2g of salt per day over a period of 5 years, including in young adults.

Smaller businesses are out of scope given the operational challenges this policy would invoke, particularly in the collection, supply and monitoring of their sales and nutrition data. However, given the role of UK grocery wholesalers as suppliers to small businesses, it should be considered whether wholesales could be in-scope. Independent retailers (supplied by wholesalers) often play the vital role of serving families in low-income areas for whom convenience is a key determinant of food consumption. Therefore, an extension of the policy to wholesalers would help to ensure greater parity across all retail food outlets, improve the healthiness of the food stocked by smaller independent stores and avoid creating or widening inequalities in the access to healthy food.

The 11 largest grocery retailers collectively hold 95% of the grocery market share in the UK. While people on lower incomes are more likely to shop at smaller convenience stores for their groceries, we found that over 90% of people on low incomes also shop at the 11 largest retailers.

Whilst we envision these targets to be a population-wide intervention that would impact the purchasing behaviours of all population groups, we have only modelled the health impact of the intervention on population groups living with excess weight for two reasons.

1. We know that the impact of food environment interventions is unlikely to be evenly distributed across the population. For example, the population living with underweight has remained stable in the past 30 years despite changes to the food environment hence we do not anticipate the proposed target will increase the prevalence of underweight in the population. Similarly, the metabolic functions in place to maintain weight would ensure a reduced calorie intake is not sustained for those with a healthy weight.
2. We also have used a static population model which does not capture the likely weight gain over time in the healthy weight population group. Therefore, to avoid an overclaim of impact in our obesity prevalence model we assumed no impact of our targets on calories purchased by underweight and healthy weight groups.