24 February 2006
The government needs to put its house in order over how it approaches its much-publicised efficiency drive or it may find that the result is to make the public sector even less effective than it was before
Is the government achieving its efficiency targets, as set out in Sir Peter Gershon's grand plan, which was launched alongside the July 2004 Spending Review?
The National Audit Office, in its first report on Gershon, says it's not entirely sure, but the government's claims should be treated as 'provisional' and 'subject to further verification'.
I think they mean that the figures are 'provisional' in the sense that an estate agent's description of a house is 'provisional', and that 'further verification' means 'get the surveyors in, that looks like dry rot to me'.
The NAO's doubts about government figures have been widely, and rightly, aired. No baselines, poor information systems, inconsistent counting, not counting costs as well as gains, insufficient audit — the list is almost endless.
If this were a financial audit the NAO would not be signing the government's claimed savings off as a fair and accurate record. But this might not bother the government too much — after all, it is only anoraks like me and political opponents who care about the details.
The NAO's doubts about the programme are rather more audacious. Remember, it is constrained by law not to criticise policy. In describing the whole Gershon programme as 'high-risk', and giving some considerable detail to just how high risk it is, the watchdog has come close to criticising the policy design of the efficiency drive itself.
The programme is highly reliant on just a few departments and projects for huge savings, which inevitably makes it risky. But more importantly, in the absence of proper monitoring there is a real risk, says the NAO, of 'unintended falls in the quality of service delivery'. And it goes on to criticise the programme's management.
All in all, even in the usual diplomatic language of the NAO, this was a pretty damning report. But does it go far enough? I'd suggest not, although I don't blame the NAO for this because of the various constraints it works under.
The NAO's study focuses on the detailed projects that make up the efficiency programme — it chose to study 20, £6bn worth, out of some 300 projects overall.
While this sort of close scrutiny has the advantage of getting down to concrete details, it does raise a rather big question. Would it be possible for these 20 projects, or indeed all 300 projects, to be successful and the public sector to be less efficient than it was to start with?
There are several reasons to think this is possible.
A project-based approach means that it is inevitably partial. Achieving efficiency improvements in one part of a complex system does not inevitably mean increased efficiency for the system as a whole.
A small example suffices: in the mid-1980s, NHS hospitals, under pressure from efficiency targets (nothing's new) decided largely to stop dispensing drugs to patients who were being discharged. The result: instant cost saving for each hospital, and an apparent efficiency gain (same output of patients for less inputs).
But this localised gain proved to be an expensive one for the NHS, because patients still had prescriptions that had to be delivered by the independent pharmacists whose cost structure was far higher than hospitals. The actual outcome was a decline in NHS efficiency (same outputs for greater inputs). This is the 'unintended consequences' law applied to project-based efficiency drives.
And then there is the displacement effect — focusing management attention on one part of a complex system may mean that other areas get neglected, rather like putting in CCTV cameras may simply drive crime elsewhere.
The next problem is another paradox — project-based efficiency gains may actually be 'Hawthorne effects' — that is, they result from the current attention being paid to an area of activity, rather than being a sustainable change. This is well-known in pilot projects, which are often highly successful but prove much less so when 'mainstreamed'.
The only way of knowing if overall efficiency is growing or declining in a public service is to measure it overall. This might seem rather obvious, but it seems to have escaped most of those involved in the Gershon debate.
The problem, of course, is that we do not have reliable information — especially about outputs and the quality changes to outputs — that would enable any such analysis to take place.
Although Sir Tony Atkinson and the Office for National Statistics are trying to move in this direction, it is a painfully slow process. And the figures produced so far — °mainly for health, as it so happens — have tended to show a decline in productivity over the past few years.
This rather makes the point that localised savings don't necessarily add up to systemic efficiency gains. Or put another way, just because the estate agent's description of the kitchen as 'fully modernised' is accurate, it doesn't mean the place isn't falling down.
Colin Talbot is professor of public policy and management at Manchester University's Centre for Public Policy and Management
PFfeb2006
