Monday 12 October 2015

At last, the truth about targets is beginning to dawn

Embedded image permalink
I've spent much of the past fifteen years being misunderstood about what I was saying about the corrosive effect of numbers in policy. I remember the blank look a member of Tony Blair's cabinet gave me when I told him about the contents of my book The Tyranny of Numbers.

By then the the corrosion had really taken hold.  It wasn't so much a problem with counting itself; it was the corrosion of effectiveness when we start muddling up the numbers with the real thing.  It was, and is, extremely hard to convince policy people that there is a real problem here.

I'm not even sure when this particular round of numerical corruption set in. Was it Robert Macnamara's period in control of the American war effort in the Vietnam War, when he introduced kill quotas for each unit. The result was huge numbers of dead but American defeat.

Or was it Harold Macmillan's much-praised 300,000 housebuild targets from 1954, when huge numbers of homes were built but, to quote Gerald Ratner, and not to put too fine a point on it, they were total crap - a new generation of slums, this time without mutual community ties of support.

It has taken a long time, but I detect the beginning of a shift.

For one thing, the situation has become extreme. There is now almost no official measure of public life that we can trust - the Libor rate, the emissions data for new cars, the success of individual schools and hospitals.

Something else is required, though - like so much else about recent years, it isn't clear quite what. Unfortunately, this is one of the challenges that the coalition failed to rise to - worse they replaced the few targets they abandoned with payment-by-results contracts, which just compounded the problem.

Then last week, there was a fascinating article in the Financial Times about the struggle to achieve, not just lots of apprenticeships, but apprenticeships worth having - a distinction that tends to get lost in the virtual world.

Professor Alison Wolf, of King’s College London, is author of the Wolf Review of Vocational Qualifications, and this is what she said:

“If we hadn’t been chasing targets and numbers for the last 10 years we could have had far more good apprenticeships (and fewer pointless so-called apprenticeships) than we have had. One hand of government knows this. The other hand is inflicting a ridiculous and unattainable 3m target on us. The risk is that the target is all-destructive — pile them high, cheap and pointless."

The real, and now urgent, question is what should we do to provide some measure of accountability, without the sheer waste of resources and the ineffectiveness that creeps in when you control using numerical targets. Answers on a postcard addressed to Michael Barber and Tony Blair...

AND! My ebook Operation Primrose is on special offer for 99p this week. There is also a conventional print version here

Subscribe to this blog on email; send me a message with the word blogsubscribe to When you want to stop, you can email me the word unsubscribe


Unknown said...

For the same reasons we have thousands of graduates leaving universities each year with degrees that literally are not worth the paper they're printed on.

Unknown said...

For the same reasons we have thousands of graduates leaving universities each year with degrees that literally are not worth the paper they're printed on. - Non said...

Dream on
The tyranny of numbers is here to stay, it is now called Data Analytics.

You won’t need targets when a Data scientist explains what the big data means and what you have to do about it.

Joe Otten said...

This is a bit strong: "almost no official measure of public life that we can trust"

Should we abolish the ONS? That would reduce the deficit, if indeed there is a deficit.

Should we replace the Monetary Policy Committee with some dice?

Is unemployment falling? Have wages and productivity been flat lately? Was there a recession post 2008? How can we possibly tell? There is no (such thing as) anecdotal evidence.

David Boyle said...

Joe, the problem isn't measurement so ONS can stay (if it likes). The problem is using numerical measures as a means of control, which then falls foul of Goodhart's Law...

Joe Otten said... no target means anything - or no measurement that a target is defined in terms of means anything.

I suspect this is still a little strong. The MPC has an inflation target - does this make inflation meaningless? Or is there a difference between macroeconomic statistics and management statistics?

I think most politically inspired government targets are meaningless for a different reason. To be any use they have to be challenging, but challenging targets risk being missed even with decent performance, and the opposition will be talking about it for years.

I think I largely agree with your main point here - and I did ask one of our parliamentarians about a year ago why the coalition hadn't done more to unpick the box-ticking target culture of public service management. The answer seemed to be that it was needed to challenge poor performance and low ambition. So the question then becomes - is there a way of challenging poor performance and low ambition that actually works?

David Boyle said...

Exactly. We need a way of challenging low performance that works - and which also, as a by-product, doesn't hollow out the service so that they are actually transforming services into one-dimensional ways of chasing the target numbers.

Stephen Tall said...

Your last para is key, David. There is a need for the kinds of things - data analysis, benchmarks, value-for-money - that ensure accountability.

Too often though we measure only the outputs and forget about the outcomes. Apprenticeships are a case in point. A similar arms race is happening in rival parties' promises on early years free hours. Never mind the quality feel the width!

But we still do need to be able to measure as accurately as we can how lives are improved (or not) as a consequence of policies.

I'm more of a fan of measuring the kinds of inputs which are likely to lead to the outcomes you want to see - eg, we know (more or less) what high-quality apprenticeships / child-care provision needs to work. Let's focus on getting those inputs in place, then track the outcomes (carefully, not obsessively).

David Boyle said...

Stephen, thanks so much for joining in. I don't really agree though. The distinction between outputs and outcomes has been hyped as an excuse for carrying on measuring, but they are actually extremes on a continuum. I think the problem here is centralisation, so that the only kind of accountability has to be measured. I think from CQC onwards, this is really failing. We need something else, at least to provide equal weight to the measures. Otherwise we get stuck in Blair-Brown compliance hell!

Phil Wainewright said...

Compliance has become a proxy for trust, but how can you trust that compliance will not be gamed? Ah, perhaps you still need to rely on people's trust, integrity and probity after all, instead of assuming the system can be made inherently perfect. But how to inject such virtues without relying on people to behave for the greater good of society? Hmmm.