Friday, 18 October 2013

It isn't just bankers' bonuses - it's ALL bonuses


The European Union's financial services commissioner Michel Barnier has warned of the UK's isolation on the issue of bankers' bonuses.  It is true that, of all the aspects of Europe that annoys the Conservatives, it is this issue which has taken George Osborne to the European Court of Justice.

But the real problem isn't really capping bonuses - though that may have been the most possible of all the political alternatives - but of paying bonuses at all.  And not just bonuses to bankers, but bonuses to pubic service managers and everyone else.

I say this partly because of my scepticism about reducing complex tasks to simple numbers, but also because of reading Michael Lewis' classic The Big Short.

This book is destined to become the non-fiction classic of the sub-prime crisis, following the handful of traders and financiers on Wall Street who could see what was about to happen.

At every level, in the disaster that destroyed the world's banks, the behaviour of staff was dominated by bonuses related to narrow targets.

The mortgage sales teams were only interested in how many mortgages they could sell, not whether they could ever be repaid. The bond departments were only interested in packaging up new bonds, packed with mortgage debt, rather than whether or not the debts were sound.

Even the ratings agencies were dominated by targets, by how much they could earn from the bond departments, rather than whether the bonds were accurately rated.

And here's the problem: bonuses over-simplify jobs, sometimes disastrously.  Just like targets, bonuses persuade people to focus on reaching simplified numerical targets which can't possibly sum up the complexity of the broad objectives they really need to strive for.

They are also easy to game.  Like targets, bonuses narrow complex objectives down to impoverished output figures.  They sacrifice broad improvement for narrow outputs. You might as well replace highly paid human beings with extremely expensive machines (hence the picture above).

They also fall foul of Goodhart's Law (when numbers are used for control purposes those figures will always be inaccurate).

So, yes I would cap them.  It would reduce house price inflation in London, which is exacerbated by bank bonuses.  But I would go further: I would tax all bonuses at 90 per cent.

All bonuses boulderise and subvert.  Everybody's.  Our companies would be better managed as a result and so would our services.

9 comments:

Anonymous said...

Bonuses are just performance-related pay. I assume you're against all forms of performance-related pay, then (as from what you write you think that measuring performance is impossible)?

David Boyle said...

That's right. Performance related pay is unfair at the bottom and simply distracting and, at worse, subverting, at the top.

Anonymous said...

Okay. So how do you stop someone who is better than their colleagues leaving for a higher-paid job at some other company, if you're not allowed to increase their pay to reflect their better performance?

David Boyle said...

I don't see any problem with having different pay grades. The problem is how you evaluate who should be in them. If it is based on numerical approximations then you will skew the result.

Anonymous said...

But 'different pay grades' are just performance-related pay, which you said you are against.

So are you fine with performance-related pay (and, by extension, bonuses) as long as it's based on some sort of fuzzy assessment by a manager of 'how good someone is', instead of numbers?

Is your real issue not with bonuses, but with using numbers for measurement, on the grounds that any numerical value can be gamed (which is true)?

Anonymous said...

Oh, and what about commission? Is that to be ruled out too?

Mark Treveil said...

That exchange sums up the problem. How do you incentivise staff and reward performance, while avoiding distorting the business for the worse.

Soviet communism demonstrated the dangers of removing all incentives. Bankers bonuses demonstrated the other extreme.

Simply paying by job description is neither effective nor fair, just as paying teachers for being old rather than good is an outrageous anachronism the current coalition might finally end.

All judgements in the real world benefit from a subjective element, if only as a sanity check. Bonus formulae typically lack that.

Even if you do have a subjective element built into to your business decisions, if those in a position to make those judgements benefit from going along with the game, all is lost.

The systemic problem of the UK banking system in the 21st century is surely shortism. When the motivation for short term profits goes all the way back to the shareholders, it is hard to it changing substantially in the near future.

David Boyle said...

Compromiser!

Mark Treveil said...

...or simply the science of the negative feedback loop.

You can't buck the 'math' [sic].