Before the election, I prepared an election briefing for the CEP on regional and urban policy. It touched on some of the difficulties in evaluating RDAs and concluded:"In short, there is no compelling evidence as to whether the RDAs are a good or bad thing. Labour is committed to them; the Conservatives and Liberal Democrats are (probably) committed to abolishing them. It should be clear that these positions cannot truly reflect evidence-based positions on RDAs’ effectiveness." Revisiting the evaluation evidence over the last few days really hasn't done anything to change my opinion on this.
Even if you look at the overall growth performance it's still impossible to make an evidence-based judgement. Individual region growth rates and the gap between the growth rates in the North and South are essentially unchanged in the periods before and after the introduction of RDAs. If you think the underlying trends (net of the effect of government policy) would have been the same in the two periods, then RDAs were more effective than the previous arrangements if they spent less money (and vice-versa). What if you went to the data and the RDAs had spent more? This looks bad for the RDAs unless you think that things would have got progressively worse for the Northern regions in the absence of intervention, so we had to spend more to stand still. In short different assumptions on the counterfactual (what would have happened in the absence of government intervention) allow you to reach different conclusions, but as the counterfactual is unobserved the aggregate data can't help us much either.
In the end, what we are left with are the broad arguments around costs and benefits of different arrangements. My feeling is that, on balance, the somewhat arbitrary regional structure makes less sense than something based around groupings of Local Authorities. The latter have democratic legitimacy. Such groupings are also more likely to end up covering "functional" economic areas (i.e. sub-national areas in which intra-area economic interactions important). Although, the evidence on whether this makes much difference is surprisingly limited.
In the end, given the evidence we have, I think that we are better placed to answer questions about what policy should do, rather than how it should be delivered. More on this next week.