Fractional Reserve Banking and Austrian Business Cycle Theory

     

As I feared, the discussion so far has descended to the interminable debate over the legitimacy and desirability of fractional-reserve banking. Larry, George, and I line up on one side, concluding that fractional-reserve banking, if unregulated and unsubsidized, provides important monetary and economic benefits that far outweigh any potential downside. Guido embraces Murray Rothbard’s total opposition to anything but 100 percent reserves. This has spilled over into our interpretations of Mises’s texts, with Larry, George, and me in fundamental agreement (with possibly some slight disagreements), while Guido tries to demonstrate that Mises ultimately came to the same position as Rothbard. These interpretative disputes are inevitable. The reason is that Mises, like all writers and thinkers, displayed occasional ambiguity and (as Guido does point out) was not perfectly consistent and unchanging across writings that spanned over half a century.

To nudge the discussion in a more productive direction, I would like to explore how Mises’s contributions impinge on an issue of current concern to economists of all stripes: the business cycle.

At the close of my initial contribution to the discussion, I mentioned two unresolved questions about Austrian business cycle theory. The Theory of Money and Credit’s first edition introduced this theory, which since then has evolved into four slightly different variants. (1) The original variant of both Mises and Friedrich Hayek (1931, 1933, 1939), despite minor differences in emphasis, primarily focused on how credit expansion --instigated either by government or the banking system -- causes self-reversing malinvestment in the economy’s capital structure. (2) Rothbard’s variant (1963) added his blanket hostility to fractional-reserves and a concomitant enthusiasm for deflationary bank panics that cleanse the economy. (3) Roger Garrison’s variant (2001), building on his understanding of some passages in Mises, posits that malinvestment and the resulting correction drive the economy off its long-run production possibilities frontier, first outward and then inward. This permits the boom to simultaneously increase both consumption and investment. Garrison thereby implicitly incorporates the upward-sloping short-run aggregate supply curve that populates so many mainstream models of the macroeconomy. (4) George’s (1988), Larry’s, and Steve Horwitz’s (1992) variant, building instead on Hayek, grants almost equal billing to velocity shocks along with monetary shocks as a source of business cycles and couples that with Leland Yeager’s (1997) analysis of the disequilibrating effects of excess supplies and demands for money, making their version into a kind of Austrian-Keynesian-Monetarist amalgam.

Despite what they have in common, these four variants imply divergent answers to the two questions I raised in my previous comment. As I emphasized, Austrian business cycle theory “hinges on specifying two firm dividing lines: (a) between those financial instruments that constitute inside money and those comprising what Mises considered genuine manifestations of people’s savings, and (b) between those increases in the money stock, however defined, that generate a self-reversing boom and those that do not.” The Rothbard variant denies that autonomous changes in velocity play any role in the cycle, contending that they merely represent changes in people’s preferences about the demand for money. Not only does this differ from the Selgin-White-Horwitz position, which like Keynesian theory, views negative velocity shocks as a potential source of depressions. But it also differs from the orthodox Monetarist position, which denies the empirical significance of autonomous shifts in velocity. Changes in money demand, according to the Monetarist position, are primarily driven by monetary shocks. And Garrison, in some of his writings (2005, 491), has endorsed this conclusion: “An exogenous change in money demand is rarely if ever the source of a macroeconomic disruption. (Here, the Austrians fall in with the monetarists.)”

Rothbard, by eliminating any role for velocity, confines the cause of business cycles to only monetary fluctuations. This makes a precise definition of what constitutes inside money all the more critical. Yet even hardcore Rothbardians disagree about what financial instruments should be included. Should you count small time deposits (CDs), which are not negotiable like checking accounts but are redeemable de facto on demand at a penalty rate? If you do, then an increase in their quantity (without a change in any other component of the money stock) must induce malinvestment; if you do not count them, the increase becomes merely a change in time preference. Rothbard’s definition of the money stock at one point (1978) included small time deposits at their penalty rate, whereas such Rothbardians as Joseph Salerno (2010, 115-30) and Frank Shostak (2000) both exclude them. Salerno’s definition includes both savings deposits and money market deposit accounts, whereas Shostak excludes savings deposits. These may seem trivial, technical differences. You can quibble about them endlessly, and I could mention several others, or about the differences all three have with Larry’s preferred definition (1986). But on top of creating ambiguity about what 100 percent reserves would look like, these disagreements make the entire concept of “credit expansion” vague and ill-defined.

George, Larry, and Horwitz get around this problem by arguing that unregulated free banking, with a commodity money base, would tend to stabilize MV (which is money times velocity in the equation of exchange). Since MV equals nominal GDP, stabilizing MV would eliminate most business cycles. It also appears to conveniently obviate the need for a precise dividing line between inside money and what Mises considered genuine credit instruments. Any change in the quantity of something not defined as part of the money stock -- perchance, small CDs -- will necessarily be captured in velocity. But this only evades the problem. Surely credit expansions that are inconsistent with underlying time preferences are not the sole cause of changes in velocity. In developed economies with fiat money, nominal GDP is almost always rising, with fluctuations in its rate of growth. Since this means MV is also always rising, how do you precisely identify periods that represent artificial booms generating malinvestments? And how can you determine if the cause is central-bank policy or something else? To reply that all such economies are always experiencing central-bank-induced credit expansion and therefore will at some unknown date in the future suffer another recession of some unknown magnitude, is not really much of a business cycle theory. Indeed, it reduces Austrian business cycle theory to an empty tautology, untestable and irrefutable. It also leads George (2008), Larry (2008), and other Austrians to a stubborn insistence that the Federal Reserve under Alan Greenspan must have caused the housing boom that preceded the financial crisis of 2007-2008, despite the fact that the growth rates of all the monetary measures -- the base (which the Fed directly controls), M1, M2, and MZM -- were steadily declining during the period (Henderson and Hummel 2008a, 2008b).

These are a few of the several, untidy issues that require more consideration, study, and discussion in order to develop Mises’s insights into a more convincing and sophisticated understanding of the business cycle.