WELL, I was right about targets and foreign criminals. According to a Panorama special aired last week, the reason so many villains with exotic accents have vanished unhindered into the countryside at the end of their prison sentences is that, until last month, officials at the Immigration and Nationality Directorate weren't answering prison officers' phone calls asking what to do with them - they were too busy working out how to fulfil the Prime Minister's party conference pledge to deport more failed asylum seekers than were applying to stay. 'Losing' prisoners in this way, of course, had the useful added advantage of heading people off from becoming asylum seekers in the first place.
Einstein said that doing the same thing over and over and expecting a different result was a definition of insanity. That's what the obsession with targets is. Whether in business or public service, misuse of targets is the single most important reason for public cynicism, rock-bottom employee morale and failed improvement efforts. Targets wreck systems, driving up costs and making things worse. Can we dispose of them once and for all? Let's try.
What's wrong with targets? Yes, everyone has them - but whether in the public or the private sector, they have the same perverse results. The attraction of targets is their simplicity. But it's a fatal one. As part of the misguided managerial obsession with quantification, they misapply partial, linear measures to a complex, shifting world. As Blair's undertaking to magic away asylum seekers shows, they are basically a wish. Bearing no relation to the ability of the system to deliver it, they are arbitrary - they might as well be plucked from the air.
None the less, the pressure to meet them is real enough. As the Panorama interviewees confirmed, after Blair's conference pledge, panicked managers in the immigration directorate ordered immigration officers to stop everything else and pursue failed asylum seekers. This is a good example of the way arbitrary targets become the de facto purpose of the system to the detriment of its wider goals. There are countless examples of this phenomenon: just last week a critical report on maths achievement noted that students were learning to pass the tests (thus meeting teachers' targets) but not to do the calculations - they knew the answers but couldn't add up.
Some targets do work- and that's one of their biggest problems. Because they are products of one world view applied to another - reductive mechanical measures applied to non-mechanical systems - targets have unpredictable and quickly ramifying consequences. To cut waiting lists hospitals do easier, rather than more urgent, operations to meet exam pass rates schools exclude difficult students or encourage them into easier subjects and to hit City earnings targets companies overstate profits or cut advertising or R&D budgets. Enron was the most target-driven company on earth, and to meet its targets it tore itself apart. The reply to ministers' repeated refrain that 'the private sector has targets' is: look at Enron.
Where they are accompanied by strong incentives, another side effect of targets is to undermine the integrity of the figures they are meant to apply to. An unfortunate primary school head was jailed for altering pupils' exam marks, so he said, to make them reflect candidates' real abilities in an under-resourced school. He got caught many others don't. Everyone in the public services with experience of the way the figures are gathered knows they are a fiction, 'managed' or finessed as the only thing under the control of people caught in a dishonest and unjust system.
So isn't the answer to set fewer, better targets? That's what ministers always say. But 'fewer, better' is an oxymoron. The fewer the targets, the broader they have to be. The broader they are, the wider the range of unintended consequences and the more attempts to exert control multiply. Targets manage the remarkable feat of simultaneously imposing oppressive controls on people and losing control of the system. As the statistician W Edwards Deming observed tartly: 'Targets achieve nothing. Wrong. Their achievement is negative.' Since centrally set targets are the problem, the solution isn't fewer of them: it's to get rid of them altogether.
There is an alternative. Forget centrally set targets (which always wildly underestimate possibilities for improvement) and start at the other end - the customer or citizen. First, find out what demand is and how well it is being met (these basic figures are usually unavailable and invariably shocking when they are - see asylum seekers) second, redesign the system to remove all the things that don't contribute to meeting that demand (lots) and improve those that do then measure the results and start again. The object for everyone engaged in the process is unchanging and relative: to make it ever easier for customers to 'pull' what's of value to them and make it ever smoother for the system to deliver it. Since, in the case of asylum, the customer is the government, wouldn't this be a good place to start?
The Observer, 21 May 2006