In which I write the piece on thinktanks Dsquared promised

Here’s a story that throws light on a lot of things that are wrong with thinktanks, even the ones that have content beyond just wanktanking. Neil O’Brien has advice for the Tories. O’Brien observes that only rich people, and specifically rich people from southern England, want to vote Conservative, and that this is not enough for a majority.

Well, yes. If you were to ask me “what does a Tory do all day?”, I’d tell you that they try to resolve this problem and convince people who aren’t rich and from southern England to vote for them.

Anyway, O’Brien pitches various concepts (what, pray, is “blue-collar modernisation” meant to mean, in the context that it has to be something that wouldn’t make a Tory run a mile to get away?), but his policy recommendations turn out to be summed up as “more means testing”.

“What about middle-class benefits? Do we really need to give child benefit to households that are better off than average? Post-election we should stop giving free TV licences, winter fuel payments and bus passes to millionaires.”

It takes a substantial degree of cheek and personal entrepreneurship to try to sell means-tested child benefit to the Tories as a new idea in the autumn of 2012, after the fiasco last year. But then, thinktanking rewards the chancer ethos.

O’Brien used to run something called Open Europe, which argued that the European Union wasn’t neoliberal enough and was letting the French impose social democracy on us. This reminds me a little of the once-fashionable Shoreditch nightclub “The Last Days of Decadence”, which shut its doors a couple of years back. I like to think that this was because its name had become a sick joke, and I feel the same way about Open Europe. As with nightlife, so with thinktanks – at these moments, the indestructible, indeed irrational, sense of self-confidence and comfort with the most jarring self-reinvention that is common to enterpreneurs is at a premium.

As a result, O’Brien has binned his aspirations to getting the key to the European Commission’s international delicatessen, junked the reports on the need for further deregulation of the financial sector, and pulled on the ruddy tweeds of the professional English Tory. Policy Exchange has always cultivated a faintly hip and international style, by Tory standards, and in many ways serves to reintegrate people who have strayed ideologically, like Anthony Browne and Nicholas Boles.

There is nothing fundamentally wrong with this, although it does lead to a certain shallowness and tumbleweed mentality. But the problem is shown up by O’Brien’s policy proposal. When the Tories last mooted this, they had a conversation with the Revenue that went like this.

Tory: Why don’t we means-test child benefit? I mean, even my wife gets it and we’re LOADED!

Sir Humphrey: That is a..brave decision, Minister.

Tory: You’re looking at me like that again.

Sir Humphrey: I take it you are aware of the reasons the last three ministers had for not taking any further action on Starter 328?

Tory: Naturally.

Sir Humphrey: Oh, you aren’t? Well, it would involve assessing recipients’ taxable income on a household basis, not an individual basis. Child benefit is paid to the mother, so this would involve treating wives as part of their husband for tax purposes, which might be illegal, as well as going against broad principles of public policy, and risking the return of Barbara Castle from beyond the grave to eat our brains. There are bound to be all sorts of complexities….

Tory: Mnfffs, the fucking Human Rights Act!

Sir Humphrey: Well, the European Convention, and the Equality Act 2010, and the legislation it replaced, as well. Anyway, as I was saying, there are bound to be all sorts of complexities regarding couples who are living apart but aren’t divorced, couples who are living together but aren’t married, couples who are divorced but are living together, civil partnerships, men or indeed women who deliberately manage their affairs to stay under the threshold…

Tory: Well, the gay couples are surely only a tiny minority.

Sir Humphrey: It’s easy for you to say that, but I have to tell the civil servants in Longbenton what to do when they all ring up. And you will probably have to explain it in court when they sue.

Tory: Where there’s a will there’s a way.

Sir Humphrey: It’s my job to find a way to collect the Government revenue from 30 million taxpayers every month, and to pay tax credits back to quite a lot of them. At the moment, we do this on an individual basis, and our forms, operational processes, staff training, database architecture, and software are all designed to do just that. If we had to change, it possibly might not work, it would probably be a terrible lot of trouble, and it would certainly take years and cost a fortune.

I believe the Prime Minister thinks reducing the cost of Government is his highest priority.

Tory: So what do you suggest?

Sir Humphrey: I suggest we make a substantial saving of public monies by not doing it, and hold a policy review that won’t report until after the next election, to look like we’re doing something.

Tory: Thank you.

Now, O’Brien has either forgotten this whole mildly comic story, or he’s pretending it didn’t happen. And this is the problem with thinktanks.

There are things you can do effectively with a group of bright generalists with laptops and no particular access to the machinery of the state. One of these is to assess policy – to review evidence, derive metrics, and draw provisional conclusions. This is a useful function, especially as the most controversial policies and the biggest cockups tend to be the things the government tries not to assess.

Another one is to suggest improvements. This is much harder. One good thing about thinktanks, though, is that the output of suggestion can be fed into the input of assessment. Opinion journalists, politicians, and bloggers do not normally make any effort at all to do this. Also, the process of assessment itself gives rise to suggestions for improvement.

Yet a third is to introduce new ideas. Coming up with genuine innovations is incredibly hard, the percentage of them which are actually good ideas is low, and the percentage of this sub-group which can be implemented in practice is also pretty low. Further, there is a source of new ideas which is almost guaranteed to be full of nonsense – the previously mentioned politicians, pundits, and random bullshitters – and one which is certain to be full of nonsense – the lobbying industry.

So, we have two functions which are analytical in nature, and one which is propagandistic. The confusion between them can be either deliberate or accidental, but is dangerous in either case. And the propaganda-function is especially dangerous, because its new ideas tend to be big and eye-catchingly radical, in short, to belong to the strategic level of analysis rather than the tactical.

Thinktanks are similar to private-sector analytical and consulting firms in various ways, but most importantly because they share many of the same functions and methods, they recruit the same kind of people, and they have some of the same failure modes.

They differ from them in that they generally derive revenue from customers, rather than sponsors and grantmongers, and that sometimes they take responsibility for implementation. Management consultants are always criticised for not taking enough responsibility for implementation, but at least they do take some and do actually descend from the aircraft and step out on the tarmac.

Thinktanks very rarely have any responsibility for implementation, or even any contact with implementers. Their relationship with the implementing party is not a relationship between a professional firm and its client, or between a supplier and a customer. The customer, for the thinktank, is the funder, a third party whose interests are not necessarily aligned with either those of the implementer or even of the thinktank. In fact, the funder’s interests may be irrelevant to either, perhaps because the funder is a disinterested philanthropic entity or an eccentric person with money, or perhaps because the funder’s aims are indirect.

(For example, they might fund research into building an airport in the sea, not because they are interested in infrastructure planning or airports as such, but because they want to promote any other option than an airport near their home.)

As a result, there is next to no discipline on their thinking as to whether it is in any degree practical. Bad thinking, like bad money, drives out good. When implementation doesn’t matter, the propaganda function of thinktanks comes to dominate the analytical function. Further, the number of analytical projects one can undertake is practically limited, but this is not true of the volume of propaganda that can be produced. And it is easier to talk nonsense than it is to speak the truth.

Producing propaganda makes you a supplier to the wealthy and very active lobbying industry. It is very hard to quantify the value of any particular burst of drivel. When a customer with pots of cash meets a supplier whose product is difficult to price, you know who’s going to win. Rather as SMS messages were historically priced much higher than the equivalent amount of generic Internet traffic, the profit margin is potentially enormous. And the confusion between the analytical and the propaganda function tends to improve the quality of the propaganda. Hence, the wanktank concept.

Although thinktanks are usually organised as non-profit entities, this doesn’t mean for a moment that profit is not a motivation. There are more ways than dividends to get money out of a company. In this case, the relevant ones are executive salaries and the expense account. New Labour thinktanks were famous for entertaining copiously.

Interestingly, all these points are also very true of private sector consulting firms, and therefore may represent a deeper truth. (Certainly, the one I work for has produced rather better analytical work lately after it started to pick up more consultancy business. Even if our parties, frankly, suck.)

What are my recommendations? Well, the first would be tell Neil O’Brien to fuck off, because he’s talking out of his arse.

The second would be that the funding and formation of thinktanks should be more strongly regulated. Specifically, a strong distinction should be drawn between lobbying and policy-analysing functions, and there should be rules governing access to public servants and information, as the ability to let some of them but not others talk to Sir Humphrey seems to have played a major role in several Coalition disasters.

The third would be that thinktanks ought to care much more about implementation. This, oddly, has been tried a bit – anyone remember the Do Tank? – and deserves to be treated with caution after some coalition fuck-ups, even if A4e is more of a half-arsed and dodgy implementer turned lobby group than vice versa. But I think Crapita and EDS could do with more competition.

The fourth? There may or may not be a general skill of management, but if there is, it is embedded in social context. What works for the club trade probably won’t at the Department of Work and Pensions. (If this blog had an editor, at this point I would be yelling I TOLD YOU I’D LINK THAT.)


  1. Phil

    Well, it would involve assessing recipients’ taxable income on a household basis, not an individual basis. Child benefit is paid to the mother, so this would involve treating wives as part of their husband for tax purposes, which might be illegal

    Not sure it’s that straightforwardly a bad idea – Working Families Tax Credits are assessed on just this (household) basis, which is why they’re such a pain in the bum to claim. (Why, oh why, are you asking me to tell you all about my salary and my savings, given that you’re the Revenue and as such already know about both of them?)

    Or rather, it is a straightforwardly bad idea, but so is WFTC, and it didn’t stop that going through.

  2. Pingback: What You Can Get Away With (Nick Barlow's blog) » Blog Archive » Worth Reading 70: After the fall

Post a comment

You may use the following HTML:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>