The New Labour government made a great song and dance about “evidence based policy”, which was generally observed more in rhetoric than in reality. But to be fair to them, their period in office did see a big increase in knowledge about “what works” and “what performs” in government, even if they didn’t always (often?) act on it.
New Labour look positively civilised beside the current lot. Just take a little inventory:
– big cuts in governments own spending on research and evaluation, including closing down projects that are well underway or even near completed – see this is the Guardian
– big cuts in University research budgets
– closing down of the National School of Government (at great expense, see David Walker on this). The NSG – the rebranded Civil Service College – was never up to much in research terms. I had a student doing a comparative study on such institutions in several countries and what stood out about the NSG was the almost complete lack of anyone with a PhD and of any self-generated research. But crucially it was one place where at least civil servants came into contact with research and evidence, which its training staff were good at recycling into their programmes.
– the Audit Commission’s work on “best practice” across a range of subjects may have sometimes been flawed methodologically, but at least it was there and freely available for scholars and practitioners and even, sometimes, policy-makers to pore over. Soon too, all that will be gone. (Although interestingly local government is still going to be required to provide The Ministry with reams of data about their performance – as yet unspecified. It is just that now there is no guarantee it will ever see the light of day, except via FOIs, or that there will be any pretence to independent analysis.)
– and of course the Strategy Unit and the Delivery Unit, which both in their own ways contributed to analysis and evidence about what works, have also gone
All of this adds up to a reversal of much of the progress made under New Labour towards introducing at least a modicum of “learning in government”. Both data and analytical capacity are being stripped away, and from a Whitehall which is already far from the (slightly rose tinted) days when the old bureaucracy was fairly good at memory retention (see Christopher Pollitt’s excellent discussion of this).
Of course, the government will claim they have this massive “transparency” agenda. And it is true we will know a lot more about what £500 was spent on what – but we’ll have much less idea of why, or to what effect, it was spent. This is a government on a mission, and missionaries don’t need evidence, analysis, doubt and debate – they just need faith.
STATECRAFT - improving government 
True, although of course the requirement for evidence along with the target culture pushed us hard towards valuing the easy things to measure, which were often not the most important or the ones people cared most about. Certainly deep cuts in the voluntary sector at the same time as trying to bring about a massive expansion of voluntarism make little sense, and it is disingenuous to imagine or pretend that deep cuts in statutory budgets (not just local authorities – quite a lot of police and a lot of PCT money goes to the VCS, for example) would not lead to deep cuts in VCS funding. Moreover, making the deep cuts this year, instead of, say, a one-year hike in income tax, inevitably resulted in old-fashioned cuts in many places, whereas I’m convinced that there are real opportunities for achieving outcomes better and cheaper by doing things diffferently (sometimes through the VCS), but researching, planning and implementing these takes time.
In fairness, though, I suspect a lot of the illogic is characteristic of a new government trying to make massive changes quickly. New Labout in 1997, by contrast, made relatively few fundamental changes. In these circumstances, contradictory decisions are made while the Jim Hackers have not yet learnt to tell when the Sir Humphreys have their own agenda.