What works in government – lessons from the other Washington

As part of its investigation of a possible “What works in social policy” institute for the UK, the Cabinet Office invited Steve Aos, director of the Washington State Institute for Public Policy (WSIPP) to a roundtable at IFG last week. To many WSIPP is a blueprint for this type of body. It has been around for a relatively long time – set up in the 1980s – and with Steve as its first and so far only director. Where other US states have institutes based in the Governor’s office, WSIPP serves the strong Washington State legislature and is set up with cross-party governance to ensure impartiality and credibility. It’s budget is small - $1.5m which pays for 11 analysts. For the rest it brings in the people it needs to do the work commissioned by legislators. Steve described the WSIPP’s as an “investment advisor”, informing government about the best uses of its money in core areas such as education, child welfare and criminal justice. But like any investment adviser, it had to recognise that its advice might or might not be taken. A baseball batting average – where batting above 0.3 is good – was a reasonable hit rate. Steve set out WSIPP’s 'design principles': • being clear about the outcomes that one wants to achieve; •generating evidence above agency silos, •pulling together all rigorous research that meets its evidentiary standards including evidence of ”what does not work” as well as what does ; and knowing what weight to attribute to what type of studies •making clear where there is insufficient evidence •presenting estimates of both the expected effects of individual policies as well as of the risks of failure. If there was not a reasonable chance a policy would break even, it was not good value for the taxpayer. All policies were looked at from the perspective of the benefit to the beneficiary, the taxpayer and the wider public benefit. Standardisation was important – so policy makers knew they were comparing like with like. WSIPP used standard values and presented their results in a standard format, a benefit of a single institute with one long-term director able to spend half his time number crunching and the other half taking that evidence into the policy process. But the big challenge for such institutes is being listened to. The WSIPP itself gets real feedback on its usefulness through the extent to which it was commissioned by policy makers. And in the last four years of reducing budgets in the state, it has seen its workload increase. So could a 'What works in social policy' institute work in the UK? Washington state is small – the size of Scotland rather than the UK. Its range of policy responsibilities is much narrower than the UK government. The strong legislature/weak governor model offers the potential for influencing policy in its formative stages – in a way that would be unusual in the UK. Bipartisan governance is logical in the US system for a body advising the legislature. That sort of governance would be unusual for an arm’s-length body. WSIPP has built its influence gradually and over time – not something our system is good at. Interestingly though NICE – the model for much UK thinking about extending the ‘what works’ model – has had the same chief executive and chair since its foundation in 1999. And would we be prepared to regard a 30% influencing rate as a sign of success – or as a 70% failure rate? It is clear though that independent evidence institutions can affect political policy making. This is an issue we will explore further in our final Making Policy Better seminar on 23 May.