The Bactra Review: Occasional and eclectic book reviews by Cosma Shalizi   155

A Call for Judgment

Sensible Finance for a Dynamic Economy

by Amar Bhidé

Oxford: Oxford University Press, 2010

Hayek contra Chicago0

The first part of the book is a summary of Bhidé's previous work on innovation in the non-financial sectors of modern capitalism, and the importance of hard-to-articulate local or personal knowledge, of wide-spread and decentralized opportunities for participating in new ventures, and of enduring, high-stakes relationships both as ways of gathering valuable information about businesses and their potentialities and as ways of governing businesses. (This is in part an honest Hayekian's attempt to answer the question of why, despite all of the beautiful things Uncle Friedrich said about the price mechanism, we see so little of our economy actually coordinated through market prices.) Bhidé's great bugaboos in this part of the book are centralization, and the replacement of informed judgment in under genuine uncertainty with "mechanistic", quantitative, actuarial rules.

The second part of the book looks at the development of the theory and practice of modern finance. What Bhidé wants to see is markets that coordinate the dispersed judgment of many individuals with real, if personal and qualitative, information about specific businesses. Instead the whole trend has been towards either direct centralization in the hands of large firms, or effective centralization through people all applying the same (or very similar) formulas to the same data, and that data being very thin and abstract. The paradigmatic examples are the Black-Scholes formula for option pricing, and Markowitz optimal portfolio theory. Both presume a weird sort of near-omniscience on the part of the user: the true distributional form of asset price changes is known, the true volatility is known (for Markowitz, the covariance between all assets is also known), none of these things ever change, and all that is unknown is exactly what random values the fixed stochastic process will realize. Actual substantive knowledge of the conditions and prospects of firms is irrelevant, and every participant in the market should have the same knowledge and the same opinions. (At most they differ in their tastes for different sorts of well-characterized stochastic risks.) Despite the demonstrable "impossibility of informationally efficient markets", assumptions like these have been built into the practices of modern financial markets (Bhidé soundly and extensively relies on MacKenzie on this point), and, perhaps even more importantly, into modern financial regulation.

Bhidé, naturally, thinks this is all misguided. The economy keeps changing its structure, so we face genuine uncertainty rather than merely actuarial risk. The pricing models lying behind modern financial engineering, and especially behind the more complex instruments like, say, collaterized debt obligations, are accordingly nonsense. Evidence that those instruments have actually done anything to help capital allocation is conspicuously lacking. The point of most financial engineering is rather different. Because we need a reliable system of payment and credit, we as a society subsidize the banking system. This limits how much banks can lose when they gamble, giving them an incentive to make big, risky, quick, highly leveraged bets if they can. (The risks to bankers are further capped by turning banks into limited-liability corporations.) The New Deal essentially made the banks an offer they couldn't refuse: banking would be supported (through measures like deposit insurance), but risk-taking and leverage would be highly curtailed. The government even went so far as to create new, safe forms of lending for banks to engage in, most prominently the multi-year, fixed-rate, continually-amortized residential mortgage. Financial engineering becomes profitable because it creates ways of taking risky, leveraged bets that evade regulatory limits. (Cf. Crotty.) Take CDOs: an investor who genuinely wanted exposure to a bank's portfolio of loans could always have bought stock in the bank, or opened a savings account. But the bank had to retain capital reserves against those loans, lowering its leverage. The CDO is a way around this, so as to bet more with the same reserves.

Even if a bank was very cautious and wanted to avoid such risks, if it competes for deposits on rates (with other banks, money market accounts, etc.), then it pretty much has to roll the dice with the rest, or lose its customers. This does not, however, help serve the purposes which are supposed to justify a financial system. Similarly, Bhidé argues, having non-financial firms orient themselves towards maximizing returns on their stock, held by anonymous, transient, deliberately uninformed people and institutions, is a recipe for bad corporate governance. (I don't remember him quoting Keynes's line about "When the capital development of a country becomes a by-product of the activities of a casino, the job is likely to be ill-done", but I doubt he'd disagree.) He wisely does not say that precisely our current financial crisis and aftermath was inevitable, but we had arranged everything to make some such crisis very likely, and are on track to keep just those features of our institutions which led to the crack-up.

Naturally, Bhidé closes with a chapter on What Is To Be Done. He does not propose to actually ban things like credit default swaps, but to try to clearly separate utility banking from the casino. Anything which takes deposits that can be freely withdrawn and lends or invests on that basis would have its deposits guaranteed by the government, but would also be severely limited in what it could invest in: "nothing besides making loans — after old-fashioned due diligence — and simple hedging transactions", which he elaborates as loans that "can be monitored by bankers and examiners who do not have PhDs in finance". All other financial players would not be allowed to take short-term deposits from the public, or trade with or borrow from the utility banks; they would, however, be free to pursue whatever arcana of financial engineering catches their fancy.

The point of this measure — which he aptly describes as "retro", "a more stringent Glass-Steagall Act" — is not to eliminate financial speculation, but to limit the harm it can do to the real economy, by decoupling it from the machinery of day-to-day payments, credit and investment. The merits of this measure are in fact independent of those of Bhidé's somewhat eccentric Hayekian world-view.

This is perhaps just as well. Like everyone else, Bhidé is inclined to insinuate that things he already disliked helped along the crisis. But whatever the problems with modern financial theory and with financial engineering (in my view, huge), we have a very, very long history of financial crises taking place before "mechanistic" econometrics was a gleam in the Cowles Commission's eye. Whatever the problem was in 1929 (or 1907 or 1893 or...), it wasn't an excessive reliance of quantitative risk models in place of individual subjective judgment. Recurrent financial crises which bring the real economy to its knees were a persistent feature of industrial capitalism from its inception to the 1930s, and it would be strange if these were caused by theories and practices developed since the second world war.

My own, perhaps excessively simple, view of these matters is that financial markets are pretty much always as unstable as they're allowed to be, and that when they get tightly coupled to the real economy of production and distribution, they regularly cause depressions. The Great Depression was when those crashes got so bad1, and the influence of ordinary people in their polities so considerable, that something was actually done about it, beyond moral exhortation: ending the austerity-inducing gold standard, Keynesian demand management, but also regulation, what's sometimes called "financial repression". When we — the improvident, reckless heirs of the resulting and unprecedented period of growth and stability — set about dismantling those protective institutions and regulations in the name of abstract academic theories, we got back an old-fashioned financial crisis and an old-fashioned economic collapse. Financial engineering isn't really the problem; highly leveraged banks were.2

For someone with these (doubtless simplistic) views, something like Bhidé's decoupling of boring, daily finance from the casino is very attractive. It would probably lead to a much smaller financial sector, which would be bad for me personally3, but it's hard to think of evidence that it would actually be bad for the world, or even for America. Whether anything like it has even a snowball's chance in the Bahamas of happening is another question. But it would have no chance at all without efforts like this, so I welcome it.

0: Using "Chicago" as a short-hand for the views Bhidé is attacking is unfair in many ways, but that just makes it all the more attractive to me.

1: Because even commercial farmers are able to feed themselves in ways that city-dwellers can't, I suspect that 19th and early 20th century depressions caused more and more suffering as the population urbanized. But I don't have strong evidence for that.

2: Up to the political-economy glosses, this is also Robert Solow's view, which I find reassuring.

3: I teach classes for CMU's computational finance program (with very generous bonus pay), and I've done financial consulting. More broadly, modern finance has a huge demand for people who know about statistics and stochastic processes, driving up the price my skills can command. At one remove, I have very close relatives (to say nothing of friends and former students) who work, or have worked, for that industry. So if anything, my supporting proposals like this is arguing against interest, not "talking my book".

368 pp., end-notes, index

Hardback, ISBN 978-0-19-975607-0


21 January 2012; typo fixes and small wording changes, 6 February