Both the Financial Times and the Wall Street Journal have been full of articles and blog posts warning of the ill-effects of QE3. In my previous post, I discussed the most substantial of the recent anti-QE discussions. I was going to do a survey of some of the others that I have seen, but today all I can manage is a comment on one of them.
In the Wall Street Journal, Benn Steil, director of international economics at the Council of Foreign Relations, winner of the 2010 Hayek Book Award for his book Money, Markets, and Sovereignty (co-authored with Manuel Hinds), and Dinah Walker, an analyst at the CFR, complain that since 2000, the Fed has stopped following the Taylor Rule, to which it supposedly adhered from 1987 to 1999 during a period of exceptional monetary stability, and, from 2000 to the present, the Fed supposedly abandoned the rule. This is a familiar argument endlessly repeated by none other than John Taylor, himself. But as I recently pointed out, Taylor has implicitly at least, conceded that the supposedly non-discretionary, uncertainty-minimizing, Taylor rule comes in multiple versions, and, notwithstanding Taylor’s current claim that he prefers the version that he originally proposed in 1993, he is unable to provide any compelling reason – other than his own exercise of discretion — why that version is entitled to any greater deference than alternative versions of the rule.
Despite the inability of the Taylor rule to specify a unique value, or even a narrow range of values, of the target for the Fed Funds rate, Steil and Walker, presumably taking Taylor’s preferred version as canonical, make the following assertion about the difference between how the Fed Funds rate was set in the 1987-99 period compared how it was set in the 2000-08 period.
Between 1987, when Alan Greenspan became Fed chairman, and 1999 a neat approximation of how the Fed responded to market signals was captured by the Taylor Rule. Named for John Taylor, the Stanford economist who introduced the rule in 1993, it stipulated that the fed-funds rate, which banks use to set interest rates, should be nudged up or down proportionally to changes in inflation and economic output. By our calculations, the Taylor Rule explained 69% of the variation in the fed-funds rate over that period. (In the language of statistics, the relationship between the rule and the rate had an R-squared of .69.)
Then came a dramatic change. Between 2000 and 2008, when the Fed cut the fed-funds target rate to near zero, the R-squared collapsed to .35. The Taylor Rule was clearly no longer guiding U.S. monetary policy.
This is a pretty extravagant claim. The 1987-99 period was marked by a single recession, a recession triggered largely by a tightening of monetary policy when inflation was rising above the 3.5 to 4 percent range that was considered acceptable after the Volcker disinflation in the early 1980s. So the 1992 recession was triggered by the application of Taylor rule, and the recession triggered a response that was consistent with the Taylor rule. The 2000-08 period was marked by two recessions, both of which were triggered by financial stresses, not rising inflation. To say that the Fed abandoned a rule that it was following in the earlier period is simply to say that circumstances that the Fed did not have to face in the 1987-99 period confronted the Fed in the 2000-08 period. The difference in the R-squared found by Steil and Watson may indicate no more than the more variable economic environment in the latter period than the former.
As I pointed out in my recent post (hyper-linked above) on the multiple Taylor rules, following the Taylor rule in 2008 would have meant targeting the Fed Funds rate for most of 2008 at an even higher level than the disastrously high rate that the Fed was targeting in 2008 while the economy was already in recession and entering, even before the Lehman debacle, one of the sharpest contractions since World War II. Indeed, Taylor’s preferred version implied that the Fed should have increased (!) the Fed Funds rate in the spring of 2008.
Steil and Watkins attribute the Fed’s deviation from the Taylor rule to an implicit strategy of targeting asset prices.
In a now-famous speech invoking the analogy of a “helicopter drop of money,” [Bernanke] argued that monetary interventions that boosted asset values could help combat deflation risk by lowering the cost of capital and improving the balance sheets of potential borrowers.
Mr. Bernanke has since repeatedly highlighted asset-price movements as a measure of policy success. In 2003 he argued that “unanticipated changes in monetary policy affect stock prices . . . by affecting the perceived riskiness of stocks,” suggesting an explicit reason for using monetary policy to affect the public’s appetite for stocks. And this past February he noted that “equity prices [had] risen significantly” since the Fed began reinvesting maturing securities.
This is a tendentious misreading of Bernanke’s statements. He is not targeting stock prices, but he is arguing that movements in stock prices are correlated with expectations about the future performance of the economy, so that rising stock prices in response to a policy decision of the Fed provide some evidence that the policy has improved economic conditions. Why should that be controversial?
Steil and Watkins then offer a strange statistical “test” of their theory that the Fed is targeting stock prices.
Between 2000 and 2008, the level of household risk aversion—which we define as the ratio of household currency holdings, bank deposits and money-market funds to total household financial assets—explained a remarkable 77% of the variation in the fed-funds rate (an R-squared of .77). In other words, the Fed was behaving as if it were targeting “risk on, risk off,” moving interest rates to push investors toward or away from risky assets.
What Steil and Watkins are measuring by their “ratio of household risk aversion” is liquidity preference or the demand for money. They seem to have a problem with the Fed acting to accommodate the public’s demand for liquidity. The alternative to the Fed’s accommodating a demand for liquidity is to have that demand manifested in deflation. That’s what happened in 1929-33, when the Fed deliberately set out to combat stock-market speculation by raising interest rates until the stock market crashed, and only then reduced rates to 2%, which, in an environment of rapidly falling prices, was still a ferociously tight monetary policy. The .77 R-squared that Steil and Watkins find reflects the fact, for which we can all be thankful, that the Fed has at least prevented a deflationary catastrophe from overtaking the US economy.
The fact is that what mainly governs the level of stock prices is expectations about the future performance of the economy. If the Fed takes seriously its dual mandate, then it is necessarily affecting the level of stock prices. That is something very different from the creation of a “Bernanke put” in which the Fed is committed to “ease monetary policy whenever there is a stock market correction.” I don’t know why some people have a problem understanding the difference. But they do, or at least act as if they do.