Nassim Nicholas Taleb wants to end bonuses at systemically important major financial institutions:

Any person who works for a company that, regardless of its current financial health, would require a taxpayer-financed bailout if it failed should not get a bonus, ever. In fact, all pay at systemically important financial institutions — big banks, but also some insurance companies and even huge hedge funds — should be strictly regulated.

Taleb believes that this will yield a safer, more stable financial system:

Banning bonuses addresses the principal-agent problem in economics: the separation between an agent’s interests and those of the client, or principal, he is supposed to represent. The potency of my solution lies in the idea that people do not consciously wish to harm themselves; I feel much safer on a plane because the pilot, and not a drone, is at the controls. Similarly, cooks should taste their own cooking; engineers should stand under the bridges they have designed when the bridges are tested; the captain should be the last to leave the ship. The Romans even figured out how to deter cowardice that causes the death of others with the technique called decimation: If a legion lost a battle and there was suspicion of cowardice, 10 percent of the soldiers and commanders — usually chosen at random — were put to death.

No such pain faces bailed-out, bonus-taking bankers. The period from 2000 to 2008 saw a very large accumulation of hidden exposures in the financial system. And yet the year 2010 brought the largest bank compensation in history. It has become clear that merely “clawing back” past bonuses after the fact is not enough. Supervision, regulation and other forms of monitoring are necessary, but insufficient — consider that the Federal Reserve insisted, as late as 2007, that the rapidly escalating subprime mortgage crisis was likely to be “contained.”

What would banking look like if bonuses were eliminated? It would not be too different from what it was like when I was a bank intern in the 1980s, before the wave of deregulation that culminated in the 1999 repeal of the Glass-Steagall Act, the Depression-era law that had separated investment and commercial banking. Before then, bankers and lenders were boring “lifers.” Banking was bland and predictable; the chairman’s income was less than that of today’s junior trader. Investment banks, which paid bonuses and weren’t allowed to lend, were partnerships with skin in the game, not gamblers playing with other people’s money. [Emphasis added]

My guess is that this op-ed will make Taleb an even more popular man than he already is — banning bonuses is likely to be very popular, particularly with near-rich people who compete for positional goods with bankers. And Taleb makes a number of good points, e.g., regarding the asymmetric nature of the bonus (a carrot if you do well but a lack of a corresponding stick). My problem with Taleb’s proposal is that it attacks the problem of financial instability at the wrong level. Later he writes:

I believe that “less is more” — simple heuristics are necessary for complex problems. So instead of thousands of pages of regulation, we should enforce a basic principle: Bonuses and bailouts should never mix.

First of all, there is no way that this will prove a simple measure. In lieu of a bonus, might I give a banker a “raise” in her base pay the following year? Or can I give her a “promotion” that comes with an increase in the value of her health benefits? Lavish corporate facilities might be built to cater to employees who achieve certain performance targets. There are many ways to get around Taleb’s simple rule. 

Others favor asset limits on banks. I’ve made favorable remarks about this approach in the past. The problem is that asset size limits might encourage financial entities to conceal assets from regulators. 

If we want to discourage excessive risk-taking, why don’t we simply eliminate or, failing that, pare back deposit insurance? Raghuram Rajan has made the case effectively:

Deposit insurance does help keep small, undiversified banks in business. To the extent that these small banks are important in making loans in the local community—to the local bakery or toy shop—they have some economic and social value. One possibility is to retain deposit insurance for small and medium-sized banks in return for their paying a fair insurance premium, but to reduce it progressively for larger banks until it is eliminated.

Clearly, if banks are seen as too big to fail, eliminating deposit insurance is moot, as the bank will be bailed out anyway. The United Kingdom deposit insurance system, which was partial, did not prevent Northern Rock from getting into trouble or the government from coming to the rescue. The point of eliminating deposit insurance, however, is to make depositors think before they make a bank too big. Unlike depositors in the United Kingdom (where all bank deposits were partially insured, and therefore depositing in a large bank was significantly safer), depositors in large banks under my proposal would have the choice between being fully insured in a small bank and largely uninsured in a large bank. Such a measure would place some constraints on the growth of seriously mismanaged larger banks while also leveling the playing field.

One of the central appeals of this approach is that it encourages large depositors to take a more active role. There will be a flight from risky banks to less-risky banks, a market discipline that might lead banks to abandon large bonuses on their own. 

End deposit insurance: a simple heuristic for a complex problem. Taleb of all people should appreciate that we need firms to have the freedom to pursue a variety of different organizational strategies to survive and flourish. If the problem is excessive risk-taking, and I think it may well be, deposit insurance pours fuel on the fire. Eliminate it for large financial institutions and the competition won’t be over which can make the biggest bets, but rather which can attract the most depositors by being as stodgy and risk-averse as possible. The bonus “problem” will solve itself.