The government has refused to publish the risk assessment of its NHS reforms, claiming that doing so "could harm the quality of future advice from civil servants." Andrew Lansley says:

There...needs to be safe space where officials are able to give ministers full and frank advice in developing policies and programmes.

This makes little sense to me. Think of this in terms of incentives: how would openness reduce civil servants' incentives to give full and frank advice? Surely, the knowledge that your thinking will be publicized increases your incentives to do better, simply because it imposes an extra cost - public derision - upon bad thinking.

The counter-argument, I suppose, is that some "full and frank" advice consists of devil's advocate-type arguments - outrageous thoughts intended to clarify issues. But such arguments could be flagged as such in the minutes.

There's some quasi-experimental evidence which I think favours transparency. There's one area of policy-making which, pretty much around the world, has become more transparent in recent years - monetary policy. And whilst lots of people have complaints about central bankers, nobody AFAIK blames their failings upon excessive transparency. If openness is good enough for monetary policy, why shouldn't it be good elsewhere?

In fact, there must be a presumption of openness, simply because it is voters who pay for policy. As Joe Stiglitz said in this marvellous lecture (pdf):

Information gathered by public officials at public expense is owned by the public.

What's more, there are at least four arguments against secrecy:

1. As US Senators Moynihan and Wyden said, it "breeds inefficiency by providing cover for the influence of special interests and opening the door to corruption and bribery."

2. It allows bad thinking to escape scrutiny.

3. It breeds distrust. The public will ask: "what have you got to hide?" At a time when politicians are already deeply distrusted, this danger is especially acute.

4. It is anti-democratic. It supposes that there is some information which is only safe in the hands of an elite, and with which the public cannot be trusted.

On this last point, though, perhaps Mr Lansley is onto something. Maybe the quality of public thinking and of democratic debate is so bad, so debased by cognitive biases or by media manipulation, that the risk assessment would be misinterpreted. And maybe the government's rhetorical and political skills are so scant that it cannot correct such misinterpretations. But this poses the question: do we really want to be governed by people who are both sceptical of the value of democracy and lacking in political skills?

There is, of course, another possibility. Maybe he really has got something to hide.