A commenter named JTapp recently sent me a post entitled “The Fed’s Dangerous Game,” which expresses a bunch of widely held views that I think are completely off base. In this and the next post I will criticize some of those views:
The Fed has a dual mandate: to promote full employment and low inflation. For decades, the two seemed to be in conflict due to faith placed in the Phillips Curve, which held that there was a trade-off between inflation and employment. The same general idea holds today among many at the Fed, but the preferred measure is now the monthly Industrial Production and Capacity Utilization data release. In general, inflation only occurs when the economy is all full employment, or when capacity utilization is high. Under these circumstances, incremental capacity additions have the effect of “bidding up” the price of production inputs, causing a generalized rise in the price level. With capacity utilization still below its 1991 recession low and unemployment so high, the Fed sees very little chance of inflation.
This is not unreasonable. The low yields on nominal Treasury securities and continued declines in business and consumer sentiment make inflation seem unlikely at the moment. What there is less empirical support for, however, is the insistence that inflation is too low or that, as Chairman Bernanke argued last Friday, “the risks to price stability had become two-sided: With inflation close to levels consistent with price stability, central banks, for the first time in many decades, had to take seriously the possibility that inflation can be too low as well as too high.”
First of all, inflation isn’t caused by the economy being at capacity; indeed inflation was far higher in 1933-34 than 1999-2000. But more importantly, why would anyone be surprised that the Fed is trying to engineer higher inflation? If they are doing their job they should have some sort of inflation objective in mind. Because the rate of inflation has been stable for decades, we can assume that the actual rate of inflation has been close to their objective. To suggest otherwise is to accuse the Fed of extraordinary incompetence. Surely they are not so inept as to repeatedly miss their target in the same direction?
So let’s suppose they are trying to steer the economy toward 2.00% inflation. Then whenever inflation is expected to be about 1.90%, they should be trying to generate higher inflation, and whenever it is expected to be about 2.1% they should be trying to generate lower inflation. You’d expect roughly one half of the time they’d be trying to boost inflation, and one half the time they should be trying to reduce inflation. Do I have any evidence to back up this view? Yes, over recent decades they have cut rates roughly as often as they’ve raised rates. Whenever they cut rates they are trying to raise inflation, and vice versa. They may not explicitly announce that fact, but everyone in the markets understands this. Look at the response of TIPS spreads to monetary policy announcements; it’s clear that investors understand what the Fed is trying to do. So there should be no surprise that the Fed is trying to increase inflation at a time when the TIPS spreads show that inflation is likely to undershoot their implicit 2% target for years to come. The only surprise is that almost everyone finds the concept of the Fed trying to raise inflation to be somehow novel or surprising. This just shows that almost no one understands monetary policy—a point Krugman makes repeatedly.
You might respond that Ben Bernanke himself said the Fed is trying to raise inflation “for the first time in many decades,” so I must have misinterpreted Fed policy. I’m afraid the Bernanke quotation merely shows that he has a bad memory. As recently as 2002-03, Bernanke was advocating ultra-low rates as a way of boosting inflation and keeping the US out of a Japanese-style zero rate trap. How soon we forget.
The basics of price level targeting are as follows. If today’s price level is 100 and inflation is 3% per year, in five years the price level would be 116. If inflation were to fall below 3% in the early part of the period, the only way the Fed could hit its price level target would be through higher inflation later on. For example, if we were to have 1% inflation for the first 3 years, to hit its five year price level target of 116, the Fed would have to accept 6.1% annualized inflation for the next two years. But once inflation reaches those levels, bringing inflation back to 2-3% per year may not be so easy. Unmooring inflation expectations is a dangerous game and using the “price level” as a target seems especially dangerous now given that most market participants think in terms of changes to prices – inflation and returns on assets. The targeting of nominal GDP could be even more dangerous since it would deemphasize real changes in output in favor of “catch-up” based purely on increases in the price level.
This passage makes two errors. First, it is inflation targeting where expectations are unmoored. Under 2% inflation targeting, the price level 10 years from now might be 5% above current levels, or 15% higher, or 25% or 35%. There is no way of knowing. When the Fed misses its target, there is no commitment to return to the old target path. With price level targeting we know the price level will be roughly 20% higher (a bit more due to compounding.) That’s how you anchor expectations. And second, NGDP is a better target than the price level because wages are more closely linked to nominal income than to the price level. That’s why China’s workers get pay increases more than 10% higher than Japanese workers, despite having an inflation rate only about 3% higher. Chinese NGDP grows more than 10% faster than Japanese NGDP.
Of the components of GDP, none has been more reliable than consumer spending. To expect consumers to bear more of the burden for generating growth is, to hope to relive the 2003-2008 period.
The U.S. economy appears to be mired in problems that monetary policy cannot solve. And, attempting to use this blunt instrument looks like it will only make matters worse. As Bank of America chief economist Mickey Levy noted in a paper prepared for the recent e21-Shadow Open Market Committee meeting, the Fed’s “rationale for more QE has changed.” Formerly judged to be a tool to avert deflation, QE is now thought to be a mechanism to “stimulate economic growth or job creation.” Unfortunately, there is little to no data on which to base this assumption because QE is such a novel policy tool. By attempting to generate inflation to boost current expenditures, the Fed is playing a dangerous game in experimental monetary economics where the benefits are speculative but the downside risks are all too real.
In fact, the problem in 2003-08 was not consumer spending, it was investment. Too much money was going into residential investment. In any case it is not the Fed’s job to decide which sectors should expand and which should contract.
The Fed doesn’t directly target inflation or real growth, but rather AD. Changes in AD then indirectly impact both inflation and real growth. Thus it makes no sense to argue the Fed switched from a policy of diverting deflation to one of boosting growth. They are exactly the same policy.
The problem with this view is that it risks a potential monetary disaster – systemically higher inflation expectations – to correct a problem – inadequate consumer spending – that’s not necessarily evident in the data.
One man’s disaster is another man’s success. The Fed should be trying to target expectations. Right now inflation expectations are too low. Of course NGDP targeting would be better, but if they are going to insist on targeting inflation, then at a minimum they should do enough QE to raise inflation expectations to 2%, and preferably a bit higher to make up for the recent undershoot. That’s not “disaster,” that’s what happens when the Fed does its job.