John Kay writes:

The contrast between the work of the [Roskill Commission in 1968-71] and the dismal quality of the material supporting the proposed new high-speed rail link to Birmingham is a measure of how far standards of evidence in policy making have declined.

Good judges share his lamentation about policy formation. And this is not a partisan issue; look at Labour's drugs policy, for example.

However, on the other hand, most of us would agree that, in important respects, economic policy is better now than it was at the time of the Roskill Commission, when we had prices and incomes policies, fixed exchange rates and an industrial policy that obsessed about British Leyland.

This raises a paradox. How can policy-making be both better and worse than it was 40 years ago? Here's a suggestion (only that). It's that the "neoliberal" turn in politics has two adverse effects:

1. If you believe markets know best and that centralized information-gathering is bound to be a deeply flawed process, then you'll invest less effort in it, or be sceptical of the product of doing so. Cost-benefit analyses will then be founded upon flimsier evidence, or won't carry much weight even if it is.

2. The increased belief in consumer sovereignty and decline in faith that "the man in Whitehall knows best" (to which "Nudge" economics is the counter-reaction) has devalued expertise. If politics is about giving voters what they want, you don't need experts and evidence, but just pollsters and market researchers.

In these senses, "neoliberalism" has had some performative effects upon policy. It doesn't just describe the world (whether well or not is another issue), but helps to create it.