Since You Asked

A primer on free primary school meals

We know that cost-free hot food boosts attainment but we have no idea why, says Tim Harford

“Nick Clegg will attempt to distance himself from his Conservative coalition partners with a highly political £600m plan to give free school meals to children in their first years at primary school.” Financial Times, Sep 17

Didn’t Jamie Oliver do all this years ago?

Don’t even joke about it. The cheeky fellow did indeed descend upon the London borough of Greenwich, redesign its menus, galvanise dinner ladies and vanquish Turkey Twizzlers. A fresh-faced opposition leader named David Cameron and a fellow called Tony Blair scrambled over each other to praise Mr Oliver in 2005 for refusing to accept the slop that was being served to the nation’s children.

Hadn’t Blair been in charge for a decade by then?

Close enough. And what’s really embarrassing is that he could easily have investigated the hypothesis that healthier meals were good for the nation’s children. Instead, it was left to a celebrity chef and two economists, Michèle Belot and Jonathan James, who examined what happened in Mr Oliver’s wake. Relative to apparently similar London boroughs, pupils in Greenwich seemed healthier and achieved more at school. Perhaps that inspired a pilot study, carried out between 2009 and 2011, investigating the merits of handing out free meals to all primary schoolchildren.

I’ve heard about this – didn’t the pilot find that free school meals for all improved pupils’ academic performance?

It’s a little more subtle than that. We know more pupils had hot meals in Newham and Durham, the pilot areas.

Are we really supposed to be surprised that more children had hot meals?

It’s such an inspiring adjective, isn’t it? Hot. You can forget drizzled with truffle oil, seared and served on a bed of organic seasonal greens; as long as you know it’s hot, all the bases are covered. But it is worth checking this sort of thing. We also know attainment rose: the pupils ended up being about two months ahead of where one might expect in these pilot areas.

But the policy seems to be paying for food for children whose families could easily have afforded food anyway. It’s poorly targeted.

You have a point, but the current free school meals programme doesn’t do a great job of getting free meals to those who need feeding: some poor children are not eligible, and some eligible families do not apply. A parallel pilot programme in Wolverhampton expanded eligibility while retaining a means test. It was not a success. Middle-class families already have healthcare and education paid for by the UK taxpayer. Taxpayer-funded spaghetti bolognese is hardly the biggest of issues.

You’ve convinced me. Free school meals for all!

Not so fast. There are some odd features about the whole business. The first is that Nick Clegg’s Liberal Democrats only want to give the free meals to children up to the age of seven.

One has to start somewhere. And that is the cutest group of children.

The pilot found that the children who benefited more were aged 8-11. What is more disturbing is that the researchers who conducted the pilot have no idea why the meals are helping.

Surely hungry children are distracted, or even too sick to come to school.

That’s the obvious explanation, but attendance did not improve. Behaviour was measured only indirectly and inconclusively. And there’s another puzzle: the kids whose performance improved most were eligible for free food anyway.

So what’s going on?

Three possibilities. One: the only way to get poor children to eat free school meals is to get everyone to eat free school meals. Two: some of the less-poor children in places such as Newham and Durham had been eating badly and their behaviour improved, making classrooms more conducive to learning. Three: the entire thing is an artefact of the way the experiment was run. Kitchens were refurbished; queueing systems were redesigned; parents were informed about the pilot and teachers were trained. It’s possible that the experimenters weren’t able to disentangle the effect of the free grub from the effect of the focus on healthy food.

So what now?

In an ideal world, the government would roll out universal free school meals on a randomised basis in different parts of the country at different times, using that to find out much more about what is going on, whether the effect is real and whether it is value for money.

And in the real world?

In the real world, Nick Clegg will roll out the policy next September, declare victory and run for election.

Also published at ft.com.