The case for universal infant free school meals

In 2014, the Conservative – Liberal Democrat coalition government announced a policy of Universal Infant Free School Meals (UIFSM), providing a free lunch to all children in the first three years of primary school from September 2015 onwards. This policy has survived intact, despite competing manifesto pledges ahead of the 2017 election, with the Conservatives pledging to scrap universal free lunches in favour of a free breakfast offer, and Labour proposing to extend universal Free School Meal entitlement to all primary school children.

The Department for Education’s stated aims for the policy were to improve children’s educational attainment; to help families with the cost of living; and to ensure children have access to a healthy meal a day and develop long-term healthy eating habits. It costs £437 per child per year, and over £15m was spent in the first year on improving school kitchens to meet the increased demand. With another election looming and the fate of this costly policy potentially at stake, is important to know whether it has delivered on its aims.

In Nuffield Foundation funded research at ISER at the University of Essex, my colleague Birgitta Rabe and I evaluated the effect of UIFSM on the bodyweight outcomes of English children in their first year of school (aged 4-5). Specifically we looked the probability that children are of healthy weight, overweight or obese, and their body mass index (BMI). We used school-level data from the National Child Measurement Programme (NCMP) from the 2007/08 to 2017/18 academic years. The data come from trained nurses who visit each primary school in England, once per year, to measure children’s heights and weights.

We found that even before UIFSM was introduced, the bodyweight outcomes of children measured later in the school year tended to be healthier than those measured earlier. For example, the prevalence of obesity among those measured in June and July was around 1 percentage point lower than those measured in September and October, and the proportion at a ‘healthy weight’ accounting for their age and sex around 3 percentage points higher. A similar improvement could still be seen after controlling for other characteristics of the pupils and the schools. In other words, a combination of seasonal effects and the school environment appears to be beneficial for children’s bodyweight outcomes even without UIFSM.

Our research also showed that children exposed to UIFSM but measured in the first half-term of the school year had very similar bodyweight outcomes to those who never received UIFSM, other things being equal. This was expected, as they will have eaten few Free School Meals by that time, and any daily difference in calorie intake would not have had time to accumulate and make a noticeable difference to BMI. However, those measured later in the school year did show significantly improved bodyweight outcomes compared with those measured at the same time of the school year but who never received UIFSM. For example, the “treatment effect” of a whole academic year of exposure to UIFSM (i.e. for a child measured in June or July) was a 1 percentage point increase in their probability of being a healthy weight, and 0.5 percentage point decrease in probability of being obese.

These effects are large compared with other school-based interventions to improve bodyweight outcomes, delivered either in the classroom (education-based) or playground (physical activity-based), but so are the comparative costs of UIFSM. The results suggest that UIFSM is unlikely to be cost-effective solely for improving this measure of child health, but we are continuing to research the effects on school performance, attendance and absences.

Read the blog on the New Statesman website here.


Holford A, Rabe B (2019) ‘The impact of Universal Infant Free School Meals on child body weight outcomes’, Conference Paper, Special Session on Health Behaviours, Royal Economic Society, Warwick, April 2019.

This project has been funded by the Nuffield Foundation, but the views expressed are those of the authors and not necessarily the Foundation. Visit