Policy evaluation via a statistical control: A non-parametric evaluation of the 'Want2Work' active labour market policy

- We evaluate a policy to help those on health related benefits get back into work.
- We establish a statistical counterfactual using the Quarterly Labour Force Survey.
- Participants were more likely to find employment than the control group.
- The benefits were slightly weaker when focusing on full-time employment.
- There is no evidence that participants accept lower wage jobs than the control group.

Active labour market policies are popular tools used by governments to help get unemployed people back into work. It is a common problem that interventions of this type are often established with no equivalent control group on which to base an evaluation of programme effectiveness. This study uses propensity score matching to evaluate the success of an active labour market policy (the Welsh Assembly's 'Want2Work' programme) that cannot otherwise be evaluated using standard parametric techniques. Using a range of matching and estimation methods, sub-samples and types of job, the scheme was successful. Our most conservative estimate indicates that participants were 7 percentage points more likely to find employment than a control group of non-treated job-seekers. The method adopted here is a useful addition to the real world policy evaluation tool kit, where programme effectiveness needs to be judged in the absence of an established control group. Our results provide evidence that even those who are currently out of the labour market and on health related benefits can find work with the help of appropriately designed support.