Alison Wolf

The French practice of seconding research experts to spending ministries is one we would do well to copy.

The Government's mantra of "evidence-based policy" shows a touching determination to play to the country's weaknesses. In a nation that still (just) has serious universities and excellent research, our Civil Service is extraordinarily bad with evidence: bad at collating it, bad at understanding it and very bad indeed at procuring research that could enlighten policy.

I have worked closely with only three British ministries; but the problems are systemic. They are manifest in the tide of initiatives that pours into any public service sector: railways, workplace training, medical general practice, university funding. If you work in an area for a while, as academics do, ideas start to seem familiar. Policies that were tried and trashed a while back resurface, gleaming and newly hatched. Yet not only the ministers but also the civil servants are completely unaware of this fact.

Many Civil Service departments have no institutional memory. Those responsible for turning ideas into detailed policies are often young and new to their area. Indeed, if they do not move on fast, they start to worry. Rapid changes in responsibilities and ministries are key to a successful career. Moreover, their predecessors leave nothing behind from which they can learn. Ministers behave much the same way.

You can scour a government office in vain for copies of previous White Papers, never mind think-tank reports from more than a year ago, or a single academic book about the office's policy remit. In the workforce training area, I meet people with policy responsibilities who do not know the names of recently abolished programmes for which they are designing successors. These people are not indolent, incompetent or uncaring. It is a system fault, compounded by the generalist tradition. This hires graduates and gives no further in-depth training in research design, statistics or economics. Instead, academics or big survey companies all too often present results to a roomful of policy-makers of whom only one or two can understand, let alone engage critically with, the arguments.

Departments do commission some properly constituted work, including some by research centres that have been praised by the Organisation for Economic Cooperation and Development. But ministers are always in a hurry. They pose unanswerable questions, or questions that can be answered only by trying things out, or tracking events for two to three years. Civil servants want juicy research outcomes now so they can promise ministers and colleagues what they asked for, within the annual budget cycle. So in comes yet another invitation to tender for another completely impracticable, expensive and pointless study with a six-month reporting deadline.

Things can be done better. In the US, career civil servants work in a single bureau or division and achieve genuine policy expertise. If I have a question about Job Corps, the US's training programme for disadvantaged adults, I can talk to a civil servant who knows its history, helped design sophisticated evaluations and can defend the findings in the face of informed but hostile questioning on Capitol Hill.

France far excels the UK in its ability to plan and implement public-sector policies coherently, notably in health, transport and education. The reasons lie deep in the two countries' different institutions and economies, but there is one key policy that we could and should copy. Civil servants, like the rest of us, promote their own careers. However, the French have created a cadre whose careers depend not on pleasing the junior minister (directly or at two removes) but on carrying out good research.

The highly respected French National Institute for Statistics and Economic Studies (Insee) lends staff to, for example, the education ministry's evaluation office. The careers of these staff are made within Insee, not within the host ministry, they remain Insee employees, and so they have a strong incentive to demonstrate their expertise in designing good evaluations. That is what will serve them well when they go "home".

Britain's Office for National Statistics is also highly regarded. Its professionals similarly depend, for their career success, on their technical reputations within the ONS and with the wider academic community.

Seconding ONS professionals to other ministries, where they could give expert and critical opinions without risking their careers, would not transform the Government's approach to evidence, but it could help improve it. Whether it ever happens, however, will depend, like everything else, on the Treasury. Does it want evidence-based policy? Or, like the spending departments, does policy-based evidence suit it a great deal more?

Alison Wolf is professor of management at King's College London.