Thursday, August 27, 2009

Taylor on the Taylor Rule (amv)

Below you find John Taylor's recent comment at bloomberg.com on recent claims to modify the Taylor rule by adjusting the relative weights between inflation gap and output gap (Taylor summarizes the Rule below, so I do not). His major point is that central banks should return to the rule that at least the Fed has left in 2001 and that at the moment the Fed has indeed done so. He further argues that there is no rationale for adjusting the rule to the fed's actual performance since 2001 (which was too expansionary according to the original formulation of the rule) because (1) it favors policy that has led to the crisis and (2) because adjusting rules often and unprincipled indeed dispenses any rule-based policy. I think Taylor is right and his advice should be taken seriously by policy-makers and policy-advising economists. There is only one point that I want to stress. The Taylor rule is perheps the first-best practical norm available in an uncertain world with pure theory still unequipped to provide for better models. But it is no substitute for theory. Whereas central banks should return to the rule, because there is nothing better around, economist should be more ambitious and stop focusing solely on reduced-form systems and rather focus on the lessons to be learned by more stringent general equilibrium analysis (including global stability problems which describe disequilibrium adjustments). Indeed, a major shortcoming of modern macro is that it ignores the major lessons of general equilibrium analysis because models are derived with eyes only on data and prediction (that does not imply, of course, that theorist do not at the same time derive much of their confidence from their general equilibrium backing). Since data seldomly informs which assumptions to adjust if predictions fail (Duhem-Quine), the profession has become used to crude ad-hocery, thereby misapplying occam's razor (which claims that simple theories and models are preferred only if all other relevant criteria for theory choice are equal).

Taylor Rule Change Will Hurt Fed’s Inflation Fight

The Taylor Rule -- the guideline for central bank interest-rate decisions -- has been the subject of a heated debate among Fed watchers this summer, popping up in market newsletters, blogs and Bloomberg screens.

The perennial late summer monetary conference held last week in Jackson Hole, Wyoming, was no exception. Just as central bankers from around the world started arriving in the mountain resort, former Federal Reserve governor Laurence Meyer sent around his latest missive, “Dueling Taylor Rules,” suggesting ways to alter the guideline’s formula.
The rule has served us well, and there is no compelling reason to change it.

All this attention now on the Taylor Rule may seem surprising because the Fed’s target interest rate has been resting at zero all summer, and the central bank has been preoccupied with unorthodox decisions such as whether to enlarge its enormous portfolio of mortgages and other private securities.

There are three reasons people are asking about the Taylor Rule now. First, it helps determine when the Fed will start to move its interest-rate target above zero. Second, it’s a gauge for how many unorthodox actions the Fed should take. Third, a big deviation from the Taylor Rule is viewed by many, including me, as a primary cause of the financial crisis and people want to avoid another one.

‘Bad Old Days’

First some background. I started developing what other people would later call the Taylor Rule about two decades ago. I was looking for a simple guideline for the Fed and other central banks to set interest rates to prevent a return to the bad old days of frequent recessions and high inflation like in the 1970s.

Using the latest in economic thinking, I came up with a mathematical formula. It is not as accurate as E=mc-squared but it turned out to be useful, and it can be stated like this: the interest rate should be one and a half times the inflation rate, plus a half times the GDP gap, plus one. (The GDP gap measures how much the gross domestic product departs from its normal trend level.)

This simple rule was meant to be a guideline, not something to be followed mechanically. Nevertheless, it turned out to be an accurate description of Fed decisions during many years of good economic performance, at least until the period leading up to the recent crisis.
Debate Rages

What does the Taylor Rule say now? Rounding off to the nearest whole number, the inflation rate is about 2 percent and the GDP gap is about 8 percent below normal (minus 8 percent), as GDP has declined sharply in the recession. Thus, the Taylor Rule says the interest rate should be 1.5 times 2, plus 0.5 times minus 8, plus 1, which is 3 - 4 + 1, or 0, about where the Fed’s target rate is right now. This suggests that if inflation rises or the GDP gap narrows, then the Fed will have to increase the federal funds rate, perhaps early next year.

So why the debate? One reason is that people have started modifying the Taylor Rule. For example, Meyer argues that the phrase in the rule “a half times the GDP gap” should be amended to read “the full GDP gap.” That might not seem like much but it makes a huge difference now: If the GDP gap is minus 8 percent, then the amendment lowers the interest rate guideline from zero to 4 percent below zero.

If the interest rate guideline is 4 percentage points below zero, then inflation can rise a lot and the economy can grow strongly and the Fed will not have to increase interest rates for quite a while. That is why Meyer is forecasting a zero interest rate until the end of 2011.
Fed Intervention

These large subzero rates are also used to justify continuing the Fed’s massive intervention into mortgage markets or long-term Treasury markets. The rationale is that since the modified Taylor Rule’s call for a subzero interest rate can’t be carried out in the short-term federal funds market, then the Fed should try to carry it out in other, longer-term markets.

Others have suggested their own modifications. Economists at Goldman Sachs Group Inc., for example, have effectively suggested amending the words “the inflation rate” in the rule with “the forecast of the inflation rate.” That can also lead to a big change in the interest-rate target. If you forecast deflation rather than inflation, then the target interest rate will be well below zero.

There is nothing wrong with changing rules if there is a good rationale for it. After all, times change and we learn lessons from history.

Poor Reasoning

The problem in this case is that the rationale isn’t good. It is based on the empirical observations that the Fed set lower interest rates in recent cycles, including the period leading up to the financial crisis, when its forecasts for inflation were too optimistic in retrospect. Thus, the modifications can give a better description of Fed behavior over some recent periods.

But this doesn’t make the change a good idea. In fact, one of those low interest-rate periods may have helped bring on the crisis. So amending the Taylor Rule in this way can perpetuate mistakes. Moreover, other economists, such as Robert Hall of Stanford University, suggest amendments in the reverse direction based on optimal policy considerations.

One of the advantages of rules is that they reduce arbitrary discretion and add predictability to monetary policy decisions. Predictability is a major factor in favor of rules, but if one changes the rules too much or too frequently, it creates instability.

So I say stick with the rule that worked. If we are fortunate and the economy recovers smartly, or if the inflation rate picks up, expect the Fed to raise the interest rate appropriately. That will keep the expansion going and help avoid another crisis.

(John Taylor is a professor of economics at Stanford University and a former Treasury undersecretary.)