How Long does it take to Gain Expertise?

Dominic Cummings, Boris Johnson’s right-hand man, currently is said to be set on shaking up the Civil Service. The three elements that are rumoured to be on the agenda are:

• Better training in data science, systems thinking and ‘super-forecasting;
• Staff will spend longer in a given post than the current expectation of 18 months; and
• Civil servants will be ‘reoriented to the public’;

I am quoting here from what has been put into the public domain and of course can’t judge how accurate the information is. The last point should perhaps be read in the context of ‘Yes Minister’ episodes. Quite honestly, I don’t know how to read it. In principle there is nothing wrong with hoping that civil servants are able to express their views in clear language that the public will ‘get’, but somehow I suspect that isn’t really the goal. However, it is a long time since I read George Orwell, so I will move on to consider the first two points.

The first one is a reflection of the extreme paucity of civil servants who enter via the ‘Fast Stream’ route with a STEM qualification. The figure quoted was 17%. Just like the lack of MPs who have a firm grasp of science, technology or innovation issues based on their qualifications and experience, it is a matter of concern so few civil servants are well-informed about much of what should be central to their role. Julian Huppert, the former Cambridge MP, proposed upon election in 2010 (as the only MP trained in scientific research at the time) that MPs should be provided with mandatory training in relevant scientific matters. It sounds as if Cummings is taking the same line with the administrators. I cannot think this is a bad idea, although I would add into his list statistics (though perhaps he includes that in the term data science), reading graphs and some basic science facts so that they don’t make endless gaffes or misinterpretations. This might not stop their political masters making sweeping statements based on pure ignorance and/or wishful thinking.  I am thinking here of the daftness, not to mention sheer impossibility of the ‘Star Wars’ shield that Ronald Reagan was so enthusiastic about (no doubt readers can think of other jaw-dropping examples). Reading Sharon Weinberger’s book Imagineers of War about (D)ARPA, as I said I would in my last post, I was struck by her description of the reaction to Reagan’s announcement of his plans:

‘DARPA’s director, Cooper, and other senior officials – including Secretary of Defense Weinberger – sat slack-jawed as they tried to digest the President’s address. The president had just made one of the most significant military technology decisions of the past few decades without consulting the key people in the Pentagon responsible for that technology.’

These days, with fact being even less in evidence in some political statements, civil servants, however well-informed personally, may find themselves in a similar position. Nevertheless, making sure they are technologically and numerically well-informed themselves does sound like a good thing to do.

However, it is the second bullet point I find most interesting. How long should one stay in a given post? I had not appreciated that 18 months was regarded as a suitable time scale for those in Whitehall, with the implicit message that if you’re not moving onward and upward into a different role by that benchmark, you’ve failed. It strikes me as deeply troubling, possibly even reflecting a concern that if you spend longer you might start to know what you’re talking about. It’s those damned experts again that Michael Gove had such a dislike for. Yet expertise and experience is exactly what is needed in so many situations, in government or outside.

My experience of different parts of what is now UKRI highlights this. These observations may not be current but certainly in the past EPSRC had a habit of swiftly moving employees on from one programme to another. Just when you’d got to know one of the team, so you felt able to ring them up to discuss a call for proposals or a financial question, they’d disappear to somewhere else. One of my colleagues once referred to a visit by EPSRC to the department as a ‘bunch of schoolchildren’ they looked so young and fresh-faced (and this remark was made by someone a good 15 years younger than me). Of course that doesn’t necessarily mean they were not experts let alone competent, but there was the worrying sense they simply did not have enough experience to be able to answer the questions they were about to be bombarded with.

This habit could be contrasted with BBSRC, where people tended to stay in a particular post much longer. They would have plenty of experience of sitting through committees and hearing discussions, have had many phone calls with irate or anxious PIs and applicants, and would know what rules could or could not be bent. I feel that level of experience tends to lead to better outcomes. Swift rotations between programmes stops anyone ‘going native’, but also may restrict their ability to make good decisions, or indeed any decisions at all without referring up.

My most recent interactions along these lines were with UKRI in the summer of 2018 just after they had come into formal existence, when I went to discuss funding for interdisciplinary (aka multidisciplinary) work in the new structures. It was, I’m afraid to say, a dispiriting experience precisely because those I talked to did not have prior knowledge of any of the research councils and their modus operandi. One was fresh out of Oxford and admitted to being on a 3-month rotation, the others were – I deduced – ex-BEIS and may well have been expert in science but not in what goes wrong for interdisciplinary research at standard research council responsive mode panels. They diligently took copious notes (I was reminded of the jury in the courtroom scene in Alice in Wonderland) but were in no position to comment on how the brave new world of UKRI might manage things better. Indeed, I am still waiting to hear more details of the review of peer review of multidisciplinary research promised in the UKRI Delivery Plan of summer 2019.

So, I am all in favour of civil servants (and UKRI employees) being able to stay in a post for more than 18 months without that being taken as a sign of failure. It is an interesting question how long is ‘optimum’ to stay in a given role. Too short, and you simply don’t know enough to be particularly effective, too long and there is the danger of becoming fossilised and of losing all enthusiasm for change or innovation. As a head of a Cambridge college, I am aware that this is a question asked of us by our constituents. I feel, after five years, I have some awareness of where there may be skeletons. There seem to be remarkably few in Churchill, as far as I’ve discovered; nevertheless at the outset I had to chair committees where everyone else knew more about history of a given topic, what worked and what did not and should never be attempted again, and that makes things decidedly tough. But by 10 years – my statutory limit in post; every college has different rules – maybe I will be bored, brusque and impatient and everyone will be counting down the days till I leave so they can do things differently. Of course it is possible they are doing that already, but I hope not.

So I am quite enthusiastic about what appears to being proposed. Whether Dominic Cummings should be allowed to shake up the civil service given his less-than-formalised role, is of course a totally different question.