Chris Dillow's Blog, page 175

Changing banking culture

Sir Mervyn King wants "a real change in the culture of the [banking] industry. He's an optimist, I'll give him that.There are (at least) three big obstacles to such a change:

1. Bankers tend (with exceptions) to be selected for a money-chasing amorality. Most jobs in the sector are pretty dull - I'm surprised I stuck there for eight years - and "socially useless". The sort of people attracted to them are likely to be disproportionately motivated by hard cash.

2.The mere salience of money tends to make people more selfish (pdf) and more likely to cheat. It's  difficult to see how the salience of money in banking can be reduced.

3. Organizations and societies tend to display path dependence - they rarely change radically for the better quickly. To paraphrase Max Planck, culture advances one funeral at a time.

Now, I don't say this to deny the possibility that "strong leadership" might change banks' culture for the better - but it'll be a long haul.

But herein lies a problem. Talk of "culture" isn't just waffle.The trouble with banks is that they are riddled with principal-agency problems. Shareholders and regulators have only imperfect control over CEOs, CEOs have only imperfect oversight of senior managers, and senior managers have limited control over their underlings, and so on.

Now, many organizations have such problems. And in many of them, the solution lies in culture. For example, cultures of professional pride, reciprocal altruism or a work ethic ensure that workers do a good job even if they are not closely supervised. As - ahem - Bob Diamond said "the evidence of culture is how people behave when no-one is watching."

In banking, however, the culture that helps solve agency problems is lacking.

So, how to create it, given the obstacles I've described?

We could learn from pre-modern states. They imposed brutal punishments upon wrong-doers, in part because they needed to terrorize their subjects into submission simply because they had no other means of controlling them. Similarly, if shareholders or regulators cannot control bankers - and it looks like they can't, draconian punishments for wrongdoing are needed to keep them in line. As Voltaire nearly said, "it is wise to kill a banker from time to time to encourage the others."

Of course, such a policy has drawbacks. But the point is that radical change in the face of large barriers to change requires radical action.

Maximization vs morality

What is, or should be, the link between morality and narrow economic rationality? Three things raise this question:

- Barclays' attempts to manipulate Libor might well have been rational from the point of the traders involved, based upon weighing the benefits of doing so against any fines. But it's generally agreed that such behaviour was morally wrong if not perhaps illegal in the UK.

- Tax avoidance schemes of the sort used by Jimmy Carr have been called immoral, though their most of users believe them to be ways to maximize post-tax income.

- One reason why benefit "scroungers" are stigmatized is the belief that it is morally right to work, even if the personal (and social) benefits from doing so are small.

These three cases are all examples of how there's a clash between (perceived?) moral behaviour and rational maximization.

Which poses the question: should maximizing behaviour really be tempered by considerations of morality?

In many cases, the answer is clearly yes, because there is in fact no trade-off between rational maximization and moral conduct.As Deirdre McCloskey has argued, "bourgeois virtues" of prudence and temperance enrich both individuals and societies. Trustworthiness is one way of overcoming the problem of asymmetric information which can cause markets to fail to develop. 

In other cases, though,there is a sharp trade-off between morality and economic maximization.

This is most clear in the euro. The moral belief that reckless debtors and lenders must suffer ther consequences of their imprudence is perhaps the biggest obstacle to a resolution of the debt crisis.

But it's also true in other areas. For example, social solidarity and fellow-feeling can retard growth by encouraging the emergence of "Olson groups" who lobby for special favours.

And one might argue that the trade-off also exists in the cases I started with:

- The same sort of desire to follow rules that would have prevented Barclays from manipulating Libor can also inhibit innovation. One feature of genuine entrepreneurship is the desire to reject established ways of doing things. A society in which people followed unwritten rules would have fewer Libor manipulators - but also fewer good entrepreneurs.

- Greater tax morale would encourage governments to spend more, not necessarily on useful projects.

- A stronger work ethic would make the unemployed even more unhappy.

Now, I don't say all this to side with the sort of right libertarian who prizes rational maximization above all else. I do so merely to suggest that, sometimes (often?) there can be a sharp trade-off between morality and prosperity.The knee-jerk demand that economic actors behave more morally requires rather more thought.

What use is academic rigour?

Whilst I was away (don't ask: suffice to say I ended up envying Aung San Suu Kyi her house arrest) two of the main developments were Michael Gove's call to "inject greater rigour" into education and Ed Miliband's apology for Labour's immigration policy. The two are related.

I say this because rigorous academic studies - summarized by Jonathan - show that immigration from A8 countries had little effect upon natives' wages. Miliband's claim that there "was a direct effect on wages" is at best overstated and at worst plain wrong.

His attitude to immigration, then, is not greatly informed by academic rigour. In this, he is not alone. Tory attitudes to top taxes, housing benefit, "troubled families" and tax cuts are similarly uninformed.

Gove wants teachers to introduce pupils to "the best that has been thought and written". In politics, however, the "best" is ignored in favour of a pandering to the concerns of the mob.

This raises the question: if academic rigour is not a basis for policy-making, what use is it?

The benign possibility is that it is valuable in itself. It's a good thing that someone does social research even if it has no political utility - just as a study of medieval literature or Horace's odes are intrinsically good. Such a view rejects the Browne report's belief that higher education funding be directed towards "priority subjects" which have narrow and predictable economic utility.

Curiously, though, Gove hasn't said that his philosphy of education flatly contradicts Browne's.Which raises the suspicion that he wants "rigour" for another reason.Perhaps "rigour" might be just an attempt to legitimate a sharper class divide, with rich kids getting the chance to pursue academic qualifications whilst poor ones are abandoned.

And even the few poor kids who do get a chance face a cost here. Miliband says, correctly, that Labour became "disconnected from the concerns of working people." This is not just a political problem but an individual one for those of use who jumped through the Govean hoops of "rigour": we become socially isolated, geeks, weirdos and nerds. Academic success has big drawbacks.

It could, then, be that the costs of rigour outweigh the benefits.

Tax-dodging: the left's problem

Lefties love talking about tax dodging, perhaps because it's an easy way of claiming moral superiority. But in fact, it raises an embarrassing question for them, namely: why do legal tax loopholes persist?

There's nothing new about such schemes. Back in 1996, Peter Mandelson wrote:

The present tax system...permits a proliferation of tax shelters and avoidance loopholes that are not in practice available to the hard-working majority. Tax reform is about guaranteeing a fair deal for all types of saver and doing away with the vast array of fiscal privileges that in practice are accessible only to a minority. (The Blair Revolution p23, 108)

Mandelson is not renowned for his left extremism, and yet the last Labour government failed to meet even his demands for the closing of legal fiddles.

There are two possible reasons why it didn't.

One is a lack of expertise. The state just lacks the ability to draft legislation that permanently closes such wheezes. As Roy Lyness, the manager of Jimmy Carr's tax dodge said:

It's a game of cat and mouse...The Revenue closes one scheme, we find another way round it...That's all we do with tax avoidance. The revenue puts a block in, we just go round the block.

The other possibility is a lack of desire. The lobbying power of the rich is so great that the loopholes stay open; the demand that the government encourage the film industry is a favourite cover for such special pleading.

Now, whichever of these explanations is correct, there's a depressing message for the left. It's that the state lacks either the will or the ability to achieve even basic principles of justice. Marx was, to a large extent, right; the state is the instrument for advancing the interests of the rich, not the cause of justice. This in turn implies that merely having a large parliamentary majority is nothing like sufficient for achieving leftist ideals.

So, what are the alternatives?

The Marxian answer is that the left must control not only parliament but the means of production as well, which would put the rich under the control of the majority.

Another possiblity, suggested by David Aaronovitch in the Times, is to create a stronger social norm in favour of paying taxes - to increase tax morale.

The latter might (just, and arguably) be more desireable than the former - but I'm not at all sure whether it's much more feasible.

Bosses as robber barons

News that the pay of FTSE 100 chief executives rose by 12% last year reminds us that the bosses' pay con-trick is still working well.

A couple of recent papers corroborate my view that it is indeed a con.

First, Thomas Lambert and Eundak Kwon show that movements in top US incomes since 1929 are correlated with both the political environment - "neoloiberalism" favouring high pay - and with the rate of surplus value, defined as the ratio of non-wage incomes to the wages of "productive" workers.They say:

The Marxian concept of the rate of exploitation appears to have some statistical validity.

Secondly, Olivier Fournout describes how the description of the CEO as hero in many management books resembles the depiction of heroes in many films. In both, our hero assumes a role, is reasonable, sensitive and listening, and unorthodox and creative - and takes risks. He writes:

The figure of the hero promoted by management literature and the American film industry is—at a structural level—the same.

This, of course, shows how the position of the boss is sustained by an ideological construct; the pretence that the boss is like a Hollywood hero serves to legitimate his role and his huge salary, just as stories of medieval chivalry helped to legitmate robber barons.

You might object here by pointing to work by Brian Bell and John van Reenan, which finds a strong correlation in recent years between UK CEO pay and performance:

Senior management appear to have pay that is strongly associated with various measures of firm performance...A 10% increase in firm value is associated with an increase of 3% in CEO pay but only 0.2% in average workers’ pay. Falls in firm performance are also followed by CEO pay cuts and significantly more CEO firings.

However, this finding does not reject the possibility that bosses, in aggregate, are exploiters.

To see what I mean, imagine a feudal society in which lords exploit peasants but claim, in exchange, to offer them protection. In this society, the lord who fails to offer protection might well suffer badly, as rival lords attack his land and rob both him and his peasants, whilst the successful lord who protects his peasants would see his wealth grow as he usurps other lords. There will then be a strong correlation between lords' performance and reward. But the lord-peasant relation will be exploitative.

A link between individual ability and performance is entirely consistent with an aggregate relationship which is unjust. 

Abusing the poor

Jonathan Portes says that the government's plan to tackle 120,000 "troubled families"  is based upon an abuse of data. He points out that the official definition of a "troubled family" is based upon poverty and ill-health, not upon criminal or anti-social behaviour.He says:

The "troubled families" in the Prime Minister's speech [in December] are not necessarily "neighbours from hell" at all.  They are poor.

Now, Jonathan and Tim pointed out the flaws in the data weeks ago. So why is the government persisting with the error?

Jonathan thinks it is because it is because they are too embarrassed to admit their error. I fear, though, that some other motives might be present.

One is simple arrogance. As I've said, it could be that the Tories' belief that they are entitled to govern has created a lazy complacency that leads to avoidable error.

Secondly, there's the pursuit of spurious precision. It's plausible that criminal behaviour follows some kind of power-law distribution; a few people commit a large proportion of crime. But leaving it at this is not good enough for politicians. A precise number grabs press attention, gives the impression that ministers know what they are doing, and emboldens them to think they can do something. Keynes once said that it is better to be roughly right than precisely wrong. Not in government, it isn't.

I fear, though, that there's worse. Eric Pickles says:  

These folks are troubled: They’re troubling themselves, they’re troubling their neighbourhood.

But as Jonathan has said, the latter does not necessarily follow from the former. Sure, some troubled families cause trouble for others; anything is true of someone. But that 120,000 number does not measure this. So, why conflate being troubled with causing trouble? An obvious possibility is that this government thinks that being poor is itself a moral failure, and so exaggerates the correlation between being criminal and being poor. This is, of course, an ancient impluse. Here's C.B. Macpherson describing 17th century attitudes:

The Puritan doctrine of the poor, treating poverty as a mark of moral shortcoming, added moral obloquy to the disregard in which the poor had always been held. (The Political Theory of Possessive Individualism, p226)

In this sense, the coalition's attitude to troubled families is part of the same mindset that gives us its latest attack on the families of immigrants. Both are motivated by a suspicion that the less well-off are morally defective. The government isn't just abusing the data. It is abusing the poor themselves.

What's wrong with economics?

My defence of economics has posed the question: what, then, is wrong with economics?

My answer would be that the overly-strong split between micro and macro has distracted us from the fact that micro failures can have huge macro effects.

I say this because the crisis was, to a large extent, due to the failure of a handful of organizations. It was/is a banking crisis, not a financial crisis - and, I'd add, a failure of ownership structures rather than of markets.

Conventional macro forces alone don't suffice to explain the crisis.

If high debt were the main culprit, you'd expect defaults and bad debts to have risen before the recession. They did, but not by much. In late 2007, the delinquency rate on US household mortgages was less than a percentage point above its mid-90s level. The biggest rise came after Lehmans collapsed.

Nor is it obvious that the bursting of the housing bubble was to blame. US prices fell 10% between April 2006 and late 2007 whilst the economy continued to grow.

Instead, the problem was that the losses caused by the bursting of the US housing bubble were concentrated among a few organization who - we now know - were unable to bear them. Had the losses been more dispersed - as the tech burst's losses in 2000-03 were - the macro effects would have been much less nasty.

It's not just in the US where  organizational failure has serious macro effects. The euro crisis can be seen as a crisis of risk-bearing - too much is concentrated in a few organizations - rather than as a debt crisis.

There are two aspects of the problem here. The obvious one is that some large firms/banks are strategically significant. Also, banks are prone to a common systems (pdf) failure. This might be due to interlinkages between them - so that a "fire sale" of assets by one bank depresses prices and so worsens others' solvency. Or it might be because if all banks follow the same strategy then anything that wipes out one will wipe out all. Just as a common environmental change - such as the disappearance of a food source - can wipe out an entire species, so it can wipe out an entire set of firms.

And here, conventional macro is at fault. It has generally presumed that the economy is smooth rather than granular, and thus that specific corporate failures can be ignored. They can't be. This failure has been compounded by the tendencies: to regard macro shocks as Gaussian;to ignore discontinuity and complexity in favour of more easily tractable models; and by the presumption that firms/banks are rationally managed.

The point I'm making here is not a partisan one. Yes, complaints about mega corporations are typically associated with the left. But they needn't be. As Xavier Gabaix has pointed out, this line of thinking can help rescue real business cycle models, because difficulties suffered by a handful of large firms can be just like adverse productivity shocks.  

All this poses the question. If I'm right, how culpable are economists?

A bit of me is sympathetic. Macroeconomic thinking has to abstract from many features of the economy. A model that included all relevant features would be like Borges' map (pdf) of the Empire - perfectly accurate and perfectly useless.The problem is that in 2008, they just happened to be abstracting from important ones.

However, a bit of me is less sympathetic. I have this suspicion that the nature of macroeconomics - with its emphasis upon mathematical tractability rather than complexity - is influenced more by what can be safely inflicted upon students than by what best describes the real world.

"Economics" & rationality

One of the great irritations of our age is the tendency for non-economists to tell us what's wrong with economics. We've seen two egregious examples of this recently, with a common theme. Suzanne says:

[Economics] assumes behaviour is rational. It cannot calculate for contradiction, culture, altruism, fear, greed, love or humanity at all.

And Geoff says:

It is assumed we are coldly “rational” calculators, yet we are obviously strongly motivated by love, envy, fashion and insecurity.

Now, economists have conventionally assumed rational behaviour. There's a reason for this.Such an assumption generates testable predictions, whereas if we assume people are mad then anything goes. What's remarkable is that these predictions are quite often correct; demand curves usually do slope downwards and stock markets are sufficiently efficient that very few investors out-perform them.

But this assumption is only a starting point. And testing it against the evidence has allowed economists to discover its limits. There is a humungous and fast-growing of good research on the role of cognitive biases. Behavioural economics is mainstream - sufficiently so that one of its founders got a Nobel prize.

Suzanne and Geoff are also wrong to claim that "economics" cannot cope with altruism or love. A quick search for "altruism" or "prosocial" in the usual places will return hundreds of results.

And of course it is the purest gibber to pretend that economics has no place for greed and fear; economists from Ken Arrow through James Tobin and Harry Markowitz to Daniel Kahneman (to name only the Nobels) have made big studies of risk aversion and its implications.

It's almost as false to say that economics cannot calculate for culture. Some of the most interesting research in recent years has been precisely about how history, culture and identity shape our behaviour. And this is mainstream stuff: George Akerlof has a Nobel prize, and Nathan Nunn and Daron Acemoglu are at top economics departments.

Of course, regular readers of this blog and of my alter ego know all this, and more. So why do I labour the point?

It's because Suzanne and Geoff (and of course the many others) fundamentally misrepresent economics in two ways.

First, they misunderstand what the mainstream of economics is. Mainstream thinking is fully aware of the limits of the "max U, markets work" view that Geoff and Suzanne are attacking. I suspect Paul Krugman is closer to most economists' thinking than, say, Casey Mulligan.

Secondly, they give the impression that economics is a fossilized body of ideology. It's not. It's a dynamic and diverse field of study. Of course, there's bad economics and bad economists, just as there's bad science and bad scientists. But these mustn't be mistaken for the entire subject.

And herein lies my real gripe with pieces like Suzanne's. They help to give non-economists the wholly false impression that economics is dull, ideological and useless. It's not.

Self-perpetuating inequality

Norm is amazed by Michael Lewis's story of how an arbitary inequality can create a sense of entitlement.I'm amazed he's amazed, because this is consistent with other evidence that people's behaviour can be shaped by randomly-imposed inequalities, for example:

- In the Stanford prison experiment, randomly-appointed "guards" behaved brutally to randomly-assigned "prisoners".

- The Oak School experiment shows that pupils who are randomly deemed intelligent do better than others; this is the Pygmalion effect, whereby people live up or down to expectations.

- People who are primed to conform to a stereotype of poor academic achievement (African Americans (pdf), low-caste Indians, women on maths tests) do worse in tests than people who are not so primed.

These experiments suggest that it is inequality that generates behaviour, rather than behaviour that generates inequality. In Lewis's story, the boss is greedy because s/he is the boss - she's not the boss because she's greedy. In the Stanford experiment, guards are brutal because they are guards, not because they become guards because they have a nasty nature. And people can differ in academic ability not (just) because of innate differences but because they live up and down to stereotypes.

There are other ways in which this can happen. People who achieve some success become (over-)confident, and others mistake confidence for actual ability, and so believe that the over-confident person deserves his success. And people who adopt high-power poses (pdf) - either by occupying positions of power or just randomly - become more confident and risk-loving. 

There are two big messages here. One is that people's behaviour is shaped less by their nature and more by social context than we think.

The other is that even arbitrary and undeserved inequality can very easily feed on itself. The successful quickly get a sense of entitlement and deference from others, whilst the unsuccessful lose confidence and ability. In these ways even unjust inequalities will appear to be legitimate.

Markets against markets

99 times out of 100, I warmly applaud John Kay's articles. Today's the 100th time. He criticizes Robert Shiller's proposal for macro markets:

Too often the speculative motive has tended to overwhelm the risk motive, and active markets have been a source of instability rather than security. As they were in the last financial crisis.

I'm not sure about this.

Let's imagine there had been a big, active, liquid market in house price futures and options in the 00s, as Shiller would like. What would have happened?

We know that, in a market of noise traders and rational investors, prices can overshoot because the smart money doesn't always trade against the stupid money, and might even trade with it (pdf). To this extent, house price derivatives would have been bubble-prone.

But there's another type of trader to consider - those with hedging demand.

Some of these would have gone long of the derivatives. Young people wanting to buy a house in future, or home-owners wanting to trade up, would/should have bought derivatives as a hedge against prices rising out of their reach. But other hedgers would have gone short. These would not only be home-owners planning on trading down but also mortgage lenders seeking a hedge against the falls in house prices that would increase mortgage defaults.

So, how would our market have looked in the mid-00s?

One possibility is that the weight of shorting from hedgers would have been great, and this would have held derivatives prices down. Arbitrage would then have caused physical house prices to be lower. We'd have had less of a housing bubble and hence less of a burst.

Another possibility is that the weight of noise traders - Kay's speculative motive - would have been so great that there would still have been a bubble in both physical houses and derivatives. But so what? If the derivatives had been over-priced in 2006, those mortgage lenders who did short them would have had a huge payout. The financial crisis would thus have been alleviated, as this payout offset losses on mortgages.

These two possibilities suggest that Kay is wrong and Shiller right.

However, there are two other possibilities.

One is that the bubble in derviatives might have been accompanied by counterparty risk. Those who wrongly went long might not have been able to pay the shorters.

The other is that mortgage lenders might not have used derivatives to hedge their exposure - either because they feared counterparty risk or were just over-optimistic.

In both ways, we might still have had a crisis.

I suspect there's no way of telling what would really have happened in this parallel universe. This, though, doesn't mean this thought experiment is useless.Had we had a bubble and crisis with such derivatives, it would not have been because of a market failure (a bubble) but rather because of an organizational failure - the failure of lenders to use the market to hedge or the failure of the long side of the market to manage risk.

In this sense, Shiller and Kay are both half-right. Shiller's right that more markets can mitigate market failure, if not organizational failure. But Kay's right to say that markets can't always protect us from human stupidity.