HOW WOULD you appraise a vicar's performance? By the number, length and quality of sermons? Attendance at church? Out of wedlock births? Ratio of marriages to divorce? Doctrinal purity?
This intriguing question was raised by proposals put forward last week by the Church of England's General Synod to make incompetent vicars easier to sack, and to subject them to the kind of performance measures that apply to other workers.
Don't laugh: even our box (see right) may be less satirical than you think. In one study, a Norwegian hospital chaplain had performance measures that counted not only bedside visits, but also the number of last rites he performed. In fact, the church's measurement problem illustrates with blinding clarity the tensions inherent in all performance management.
The first point it demonstrates is that in all organisations the pressure to measure is inescapable. Some of the pressure is external, from regulators, inspectors and shareholders. But as in the church, it is also generated internally by the ubiquitous requirement to do more with less. By definition you can't improve performance unless you know what the current level is. So you can't not measure.
But although measurement is inevitable, it is also problematic. The church's bottom line, to put it in vulgar accounting terms, is presumably saved souls. Unfortunately as an outcome salvation is even more ineffable than some public-sector targets such as improving education or health. Even in the private sector, the final goals of economic activity - happiness, well-being - are ultimately unmeasurable.
Which puts a premium on finding a proxy that's adequate both technically and politically. Technically, the problem is that a single measure is highly likely to oversimplify, while a portfolio of measures becomes self-contradictory and impossible to prioritise. Do good sermons outweigh a poor marriage performance?
Short-term effectiveness too may conflict with long-term goals. As the trade union Amicus, which represents about 2,000 clergy, noted last week, priests' present system of tenure gives them the job security 'to build valuable long-term relationships in their parish to the advantage of their community'.
In this case, an effective performance measure needs to focus attention on the whole soul-saving system, not just a part of it - electrifying sermons aren't much good if babies aren't being baptised and the supply of believers is drying up.
Otherwise it will fall into the political trap expressed by the truism, 'what gets measured, gets managed'. It sure does, even when the consequences are clearly counterproductive. Perverse incentives are rife throughout badly designed organisations - and not only in the public sector.
For example, the purpose of having traffic wardens is to keep traffic moving smoothly so that law-abiding people can go about their business. But measuring them on numbers of tickets, as some councils do, has the opposite effect. Milkmen, mayors, taxi drivers, postmen, emergency plumbers, undertakers, firemen and bus drivers have all been ticketed by attendants keener to make their numbers than to achieve their proper aim.
In many, perhaps most, companies, individual performance-measures implicitly encourage empire-building and competition for scarce resources at the expense of wider organisational aims. Meanwhile, largely at the behest of ministers (the political kind), some sections of the public sector have raised performance-management abuse to an art form.
For instance, policemen are encouraged to clear up as many cases as possible. This has the same effect as with parking attendants - creative arrests to make up the numbers. But wait. The next part of the system, the Crown Prosecution Service, is judged on successful prosecutions. So it takes on only the ones that it can be sure will succeed, ie as few as possible.
Again, the government announces a crackdown on car crime, makes that this year's performance-management priority and a year later proudly proclaims that car-crime figures have plummeted. But that's at least in part because the villains, helpfully tipped off about police priorities, have turned their attention to other petty crime where life is quieter. Is that laughing I hear, or crying?
Performance management is the classic example of both the first law of management - if it looks simple, it ain't - and also the second - but if it's complicated, it's wrong. Moreover, the problems of specific perverse incentives aren't the worst: the most devilish, as we might say, is the general one.
To improve performance, organisa tions set stricter performance measures. But the tighter the measures, the more damaging they are of commitment, initiative and trust - the things that exceptional performance depends on. Hence 'the supervisor's dilemma', a vicious circle in which surveillance, monitoring and authority lead to increasing distrust and underperformance - and the perceived need for more surveillance and monitoring. Performance management destroys performance.
If performance management is both inevitable and impossible, what's to be done? The answer is to disconnect measurement from control - to reconceptualise it, as professor Andy Neely of the Advanced Institute for Management Research has put it, as a system of learning rather than control.
For that, it has to be connected to the work so that it throws insight on what works and what doesn't and stimulates those who do it to find better methods of getting to the ultimate aim: saving souls. That can't be decreed from above. To make it happen, you have to let go: come to think of it, who better than the church to teach business a thing or two by making such a performance-measurement system work? As that old cynic Lou Reed helpfully put it, you do indeed need 'a busload of faith to get by'.
The Observer, 20 February 2005