The economics of radical uncertainty
Mervyn King's new book on the financial crisis and its aftermath is not what you might have expected from the former head of the Bank of England -- from an official, that is, who played a crucial role before, during and after the crash. "The End of Alchemy: Money, Banking, and the Future of the Global Economy" isn't a memoir. There's no blow-byblow narrative and no attempt by the author to justify what he did or failed to do. It's more ambitious and more daring than that.
The book asks deep, difficult questions about the theory and practice of finance and economics, and comes up with interesting answers every time. They're sobering answers too, in many cases, because they show how hard it will be for policy makers to avoid the next crisis. The title of the last chapter -- "The audacity of pessimism" -- is all too apt.
The central idea is "radical uncertainty," meaning the kind of uncertainty that statistical analysis can't deal with. For risks you can precisely define and measure against historical data, you can calculate probabilities. That's why the risk your house will burn down, for instance, is easily insurable.
But many possible outcomes can't even be clearly imagined, let alone tested against the record -- for example, how will the conver- gence of genetics and computer science affect the life expectancy of future generations? Confronted with questions like that, you're no longer dealing with quantifiable risk.
The distinction between risk and uncertainty isn't new. Economists have understood it since 1921, when Frank Knight wrote about it; John Maynard Keynes thought it was vital- ly important. But economics hasn't known where to go with the idea. The problem isn't just that economists have concentrated on analyzing things that can be analyzed, which was understandable; it's also that they've applied the probabilistic approach in areas where it doesn't work. In finance, where does risk end and radical uncertainty begin? There's less of the first and a lot more of the second, King says, than economists and market practitioners believe. Financial regulation leans heavily on estimates of the risks attached to different kinds of loans or other bank assets. But these risks aren't well understood. King writes that on one standard measure Northern Rock -- a big mortgage lender whose failure in 2007 tipped the U.K.'s financial system into crisis - - was a safe institution: It had the highest ratio of capital to risk-weighted assets of any major bank in the country. The error lay in inferring from decades of data that mortgages were low-risk loans, and in failing to imagine why that could turn out to be wrong.
In a way, finance only exists because of radical uncertainty. Its purpose is to make a bridge between the present and an uncertain future: Borrowers bring spending forward in time, savers push it back. The "alchemy" of banking brings the two together in such a way that supposedly safe short-term deposits are used to support risky long-term loans.
The point is, the dangers in this arrange- ment can't be readily quantified. There are too many unknown and unknowable states of the world. Risk-weighted capital requirements are therefore misleading. At the same time, their complexity and seeming sophistication are seductive, which only adds to the problem.
As King says, in the face of radical uncertainty, "it is better to be roughly right than precisely wrong." Being roughly right, he argues, requires simple rules of thumb and an open mind, rather than misleadingly precise mathematical solutions -- the approach which financial regulation (like economics as a whole) tends to rely on.
Applying this logic, he says banks should be required to finance themselves with much more capital in relation to all their assets than even the tougher post-crisis rules require, so that they can absorb bigger losses when things go wrong. In addition, to make it easier for central banks to provide liquidity against illiquid assets when financial markets seize up, banks should be made to pledge assets as collateral, in sufficient quantity to cover their deposits and other short-term liabilities. A rule to this effect could replace conventional deposit insurance (with premiums in effect collected upfront in the form of haircuts on the collateral). This plan for central banks to become "pawnbrokers for all seasons," as King puts it, is related to the "narrow banking" idea suggested by some other economists (see John Kay's excellent new book, for instance).