Sure Start achieved its aims, then we threw it away

Good news. A government programme was rolled out on the basis of good evidence. It did what it set out to achieve. It almost certainly delivered benefits significantly greater than it cost. Even better, it created the greatest benefits for those most in need.

Bad news. The programme was first expanded and diluted and then had its funding slashed. It no longer exists in anything like its original form.

I am talking about Sure Start. Introduced in 1999, it was an early initiative of the last Labour government, designed to provide support to families of young children, with the aim of enhancing their life chances and development. In its initial incarnation, it set up 250 or so projects in areas with a very high concentration of children under five living in poverty. Each project had a remit to offer a range of services including outreach and home visits, support for play, learning and childcare, primary and community healthcare and support for children and parents with special needs.

The local projects were not told how these objectives should be met; considerable local autonomy was built in. Partially modelled on the US Head Start programme, it was introduced off the back of increasingly convincing evidence that high-quality interventions to support children and families living in poverty really could work in improving their life chances. Previous interventions had been shown to improve everything from the health of the children to their educational and labour market outcomes and to reduce the chances that they would end up in the criminal justice system.

This carefully designed, evidence-based initiative seems to have worked. As part of an equally carefully constructed evaluation, colleagues of mine at the Institute for Fiscal Studies previously have shown that it had a measurably positive long-term effect on health, particularly reducing hospitalisations among school-age children. Last week Pedro Carneiro, Sarah Cattan and Nick Ridpath showed that it also had a big positive effect on the language, communication, numeracy and social and emotional development of five-year-olds from poorer families. More importantly, these effects persisted into much-improved GCSE results at age 16. Further work will look at other long-term outcomes, including engagement with the criminal justice system. Even if no further effects are found, what we know already is enough to provide convincing evidence that the programme will effectively have paid for itself.

These findings are consistent with a slew of international evidence that high-quality, often resource-intensive interventions can be highly effective in improving the life chances of children growing up in poverty. It is great that we now have such good and robust evidence that our very own Sure Start programme was also highly effective. It is also a tragedy and a study in how good policy can be lost.

The first mistake was to roll out the programme in a way which made it very hard to evaluate. We don’t actually know which specific children used Sure Start. The strength of the effects that were detected was all the more remarkable given that they were measured using only information on who had access to the programme, not who actually used it. There once was data on who used the programme. Most of it has been lost. Where it still exists, absurd legal and bureaucratic barriers make in unobtainable. A classic example of public sector failure to collect, curate and, most importantly, make available data that could help to inform good policy.

The second error was to move from something well targeted towards poorer communities and with considerable local discretion to a much wider rolling out of the programme with much more central direction. This second phase, which one of the founders of the scheme actually described as an entirely different programme, so different was it from the original conception, appears to have been much less effective.

Herein lie two further failures of policymaking. First, not having control makes ministers and civil servants in Whitehall nervous. What if people running it locally get it wrong? Much better to tell them what to do. Over-centralisation and lack of trust is pervasive. Second, the only constant is change. If something looks good, there can never be too much of a good thing. Nothing is ever left to develop and bed down properly.

A third error was to emasculate the entire programme by cutting its budget by two thirds after 2010. Of course, it wasn’t the only service to experience cuts through that period of fiscal retrenchment, but there was more than a whiff of another failure of our political system: not invented here syndrome. If the other lot brought it in, then we’ll get rid of it.

In fact, childcare more broadly has emphatically not been a casualty of spending cuts since 2010. There has been a gradual enhancement of the universal offer of free childcare for pre-schoolers. This month’s introduction of 15 hours a week of free care for all two-year-olds with working parents was a further extension. By September next year, 30 hours a week of free childcare will be available from nine months of age for all children of working parents.

This constitutes a huge expansion of the welfare state, but in a quite different direction from that intended and envisaged 25 years ago, when Sure Start was introduced. Back then, the focus was on ensuring the best possible start for children growing up in poverty; today, we are spending far more than we ever spent on Sure Start on subsidising the childcare costs of working families.

There is nothing wrong with that, it’s just a different choice. To govern, as they say, is to choose. Choices should be much easier when you have good evidence of the effects of your choices. We know that poor children grow up with poor life chances. We know how to ameliorate that. The choice is ours.