The Limits of Risk Engineering

The New York Times this morning pointed out something about the eurozone debt problem that I kind of knew, but hadn't really thought about:

Regulators bear much of the responsibility. Before 1999, when Europe forged its monetary union, regulators permitted banks to treat as risk-free the debt of any country that belonged to the Organization for Economic Cooperation and Development, a club of developed nations that includes the United States and most of Europe.

"There was encouragement from European authorities for banks to load up on more debt, because it was seen as safe," said Nicolas Véron, a senior fellow at Bruegel, a research firm in Brussels. "In hindsight, it was unwise risk management."

Some regulators realized that allowing banks to set aside no capital for sovereign defaults could be a problem and moved to address it in a 2006 accord known as Basel 2. They mandated that big, complex banks use their own models to determine if individual countries were at risk and hold some capital against them. But the European Union never enforced the stiffer regime. And amid the subprime mortgage crisis, Europe's regulators added to the problem by demanding that banks hold more safe assets, much of it sovereign debt.

The knee jerk libertarian instinct is to say "See!  Government regulation doesn't work!"  But that's not very useful.  Assuming that we are probably going to have some bank regulation that requires them to hold at least a few bucks against their deposits, what would have been a better regulation?  Requiring them to hold it in cash?

You could do that, but it would lower the return on bank assets.  That already meagre interest rate on your savings account would get even tinier.  Can you pay interest in nanoparticles?

Or you could make them invest it in "safer" sovereign debt. There are two problems with that: one of the things that defines safer debt (at least in normal times) is government debt from governments who don't issue too much of it.  Was there enough "really safe" sovereign debt to satisfy regulators once they decided that mortgage bonds were too risky?  How about after you add China into the mix?

One way to answer that question is to look around.  Why is Bank of New York demanding that large depositors pay them for the privilege of stashing money there?

I thought of this again when I saw this post from Felix Salmon:

But there's a deeper problem here, I think -- and that's related to the fact that the fund is being asked to return an "assumed actuarial rate" of return of 8% per year -- in an environment of high volatility and extremely low interest rates. The right thing to do is to say "sorry, I can't do that" -- but that's a great way to get fired. So instead, fund managers move further and further out the risk curve, in an attempt to hit their target returns. With predictable consequences.

Sometimes, the strategy works. In fact, the strategy is always going to work some of the time. The note proudly says that "the Plan outperformed the policy benchmark and the Median Plan in three of the last five 12-month periods". But this is the problem with volatility: if you overshoot on the way up and you overshoot on the way down, you end up underperforming overall.

I'm not particularly picking on Hawaii, here, I'm just using it as an example. Most public pension plans have very similar problems. They take on more risk than they should, just because they're being asked to hit unrealistic return targets. And the losers, of course, are all of us.

One possible story about the financial crisis is that an unusually rosy period of growth in the west taught us to expect--no, to need--an unsustainably high rate of low-risk return on assets.  We made a whole lot of unsustainable promises during the boom years, most of them involving permanent increases in the ratio of non-workers to workers.  Those promises required a very steady cash flow, but also (because of demographics) a very high rate of return.  Otherwise the previous compact--in which current workers had allowed the elderly and disabled to skim a portion of their rising earnings from increased productivity, in exchange for the promise that they'd get the same deal when the time came--would be broken.  Workers would have to turn over all of their increased earnings, and in fact cut into what they'd had previously, in order to keep the system going.

Note that this problem exists whether people finance disability and retirement through insurance and private savings, or through the government; they're both ultimately claims on the output of the current workforce.  And the current workforce is going to resist cutting their own consumption so that previous generations can preserve theirs.

We engineered great shifts of risk to entities that were supposedly "better" able to bear it: from workers to employers, and from employers to the government.  And to be sure, some of this is a real improvement--life insurance works!  But some of it just transforms the problem.  Companies may be less likely to see their pension funds go bust, but when they do, they take out a bunch of workers.  Governments at all level aren't very good at saving, and it's harder to impose outside watchdogs on a government because, umm, who will guard the guardians?  Plus all defined benefit programs encourage counterproductive behavior on the part of beneficiaries.  People covered by any sort of defined benefit plan tend to retire earlier, which shrinks the number of people paying into the system, and increases the number drawing out.  Social security systems seem to encourage people to have fewer children, which undermines the demographic structure those systems depend on.

We also tried to engineer a transformation of riskier assets into "safe" ones.  We all plunged money into the stock market because a well diversified portfolio offered high returns that were supposed to be volatile in the short run, but virtually guaranteed over the long term.  We bought houses, because they'd been a terrific retirement plan for generations of Americans. We glommed together dodgy mortgage securities into huge packages, and structured the losses so that the top tranches--the AAA securities--got a virtually guaranteed return well above the "risk free" rate on US treasuries, or bank accounts.  And banks, desperate for yield, bought them as if they were licenses to print money.  Which they were, for a little while.

These things were not crazy: the theory behind them is perfectly sound.  (Yes, even the dodgy mortgage securities).  Arguably, the problem is not with the strategy, but with everyone trying to deploy it at the same time.  We cannot all own rapidly appreciating real estate or portfolios that guarantee a return of 8% a year; if we try, we will just push up the price of those assets until they no longer return 8% a year.  If we try, we will eventually create a bubble that will deliver negative returns.

This is obviously simplified--I haven't mentioned other important contributors, like Asian central banks, who don't fit this description.  But I think it's an important part of what happened to us. And the most important thing is how we reacted.  When things blew up, we didn't say to ourselves, "maybe it's not possible to engineer a low-risk 6% annual return on assets" or "maybe it's not possible for everyone a demographically mature population to expect to spend as much time out of the workforce as in it."*  Instead, we search for the fault in the system: were pension fund managers too incompetent? Bankers too greedy and clever about searching for loopholes?  Regulators too lax?  Did we write the rules governing bank risk capital wrong?

Discovering that we can't all achieve higher risk-free returns by piling into mortage securities, we do not question the premise that there is some way to achieve attractive risk free returns on a mass scale; instead, we direct the banks to pile into OECD sovereign debt.  Having created a bubble in OECD sovereign debt that is about to end disastrously, bankers and regulators are undoubtedly even now searching for a "truly" risk free security that they can use to backstop their deposits.

Let us instead consider the possibility that this thing may not exist.  There may be no way for us all to enjoy steady, low-risk capital appreciation, or to exit the workforce whenever work becomes difficult.  There may be no way to engineer these possibilities into being.

That's a scary thought to live with.  On the other hand, the alternative is, so far, not actually less scary.

* Note: I'm not talking about housewives, who are simply substituting home production for market production, but who are mostly as fully employed as their husbands.  I'm talking about retirees, children, students, trust-funders, and people on various public benefits.