Miscarriage of justice

Next month the Institute for Government will be publishing the results of a year-long look at the implementation of four social justice policies which we have carried out with the support of the Joseph Rowntree Foundation. The report will spell out the lessons both from these case studies – and draw out lessons with much wider applicability. All implementation matters – done poorly it wastes time, money and other resources and damages the credibility of the government of the day, the reputation of the politicians promising change and the public servants carrying it out on their behalf. But the consequences of poor implementation are even more acute when they affect the most vulnerable members of society – in the case of the introduction of the personal independence payment, the disabled and the terminally ill. The PAC report highlights some crucial problems that could have been foreseen – with a more rigorous process of stress testing the implementation of the policy in advance of full roll-out. The particular issue it identifies are two mistaken assumptions on which its implementation was based:

• That only 75% of assessments would be face-to-face
• And that assessments would take an average of 75 minutes.

The problem was that the 75/75 assumption turned out in practice to be 97/120 – throwing out all of the underlying calculations on speed of processing. This meant backlogs built up and people were forced to wait unacceptable lengths of time for assessment and then payment. The second problem identified by the PAC was that the service standards of one contractor fell short of those offered in their bid, meaning in particular that far more people have had to travel longish distances for their assessments. The PAC report criticises DWP for failing to test those assumptions. A third problem is that there has been less cooperation than might have been expected in getting information to support claims from third parties – GPs, social workers and physiotherapists. Underpinning these specific concerns is a worry that they are symptoms of rushed implementation. A failure to get the basic behavioural assumptions right is often a source of failure in government policies. As the PAC says, this could have been avoided if the department had piloted the scheme in advance – that would have shown up that the assumptions were too optimistic. A starting point for good implementation is to test the robustness of critical behavioural assumptions – and where they cannot be tested in advance, ensure there are very short feedback loops to be able to adjust the basis for implementation as new information emerges. DWP’s failure to do this meant a significant underestimate of the capacity needed to deliver the service. In our implementation research, the ability to understand the capacity of the system to deliver, including assessing provider capacity – existing and needed – emerges as an important factor in successful implementation. The inability to rely on immediate cooperation from other parts of the public sector also emerged as a theme in our case studies. Implementers need to make sure that there are sufficient incentives in the system to ensure that cooperation is forthcoming, otherwise new policies tend to be crowded out by pre-existing or more immediate priorities. And all this is compounded by the problems of effective public service contracting which we have highlighted in earlier work. The PIP case shows why implementation needs to be thought through rigorously as policy is developed.