August 11, 2006
Good and not-so-good reasons to disagree with Bernanke
Some of the reasons people have given for why the Fed should keep raising interest rates make sense to me, and some don't.
I have to say that the line of reasoning from the Wall Street Journal takes the cake:
The Fed's Open Market Committee decided not to raise interest rates again -- not because inflation is contained but because it says the economy is slowing. Uh, oh. Here we go again, back to the era of the Phillips curve, the economic theory that postulates a trade-off between inflation and unemployment. We thought Paul Volcker and Alan Greenspan had buried that notion years ago. But apparently it lives on like Arthur Burns's ghost in the attic of the Fed, ready to inhabit a new Chairman who has inherited an inflation and is afraid that breaking it will send the economy into recession.
To me, there is a basic question here that can be settled without appeal to ideology or consultation with ghosts. Surely the relevant question for an objective observer is the following: if the unemployment rate goes up or the rate of growth of real GDP slows down, should that cause a rational person to anticipate a lower rate of inflation than you would have predicted in the absence of those changes?
Harvard Professor James Stock and Princeton Professor Mark Watson, two of the nation's most careful and respected economic researchers, conducted a very thorough investigation of how well different models succeeded for purposes of forecasting inflation in an article published in the Journal of Monetary Economics in 1999. They compared Phillips-Curve specifications based on measures of the level of real activity such as the unemployment rate or the growth rate of industrial production with alternatives that included 19 different interest rate measures, 12 different measures of the money supply, 21 different price or wage indexes, and a number of other variables. Here was their conclusion:
The major conclusion of this study is that the Phillips curve, interpreted broadly as a relation between current real economic activity and future inflation, produced the most reliable and accurate short-run forecasts of US price inflation across all of the models that we considered over the 1970-1996 period. This conclusion will come as no surprise to applied macroeconomic forecasters in business and government, where the Phillips curve plays a central role in short-run economic forecasting. The conclusion is also consistent with the recent academic literature on short-run economic forecasting.
Mark Thoma raises another important issue for the WSJ to consider. Basically, the Fed had just one choice here, namely, whether to raise the fed funds rate or hold it constant. We all agree that pausing is likely to mean a higher inflation rate over the near term than would a raise. But if inflation is the only concern, then why not raise the fed funds rate arbitrarily high to get an even lower inflation rate? Surely the reason we don't want to raise interest rates arbitrarily high is that there are other consequences of higher interest rates, namely slower economic growth or a possible recession, that we worry about as well as being concerned about taming inflation.
A rational evaluation of the current situation calls for weighing the risk of higher inflation against the risk of an economic downturn. The Wall Street Journal may have a different evaluation of the size of these risks or place a different weight on the importance of avoiding one outcome or the other relative to Bernanke. But the fact that Bernanke calls attention to an economic slowdown that is already underway is surely relevant for this discussion, no matter what your intellectual paradigm.
Posted by James Hamilton at August 11, 2006 02:04 PMdigg this | reddit
Ah, apparently you have made the error of taking the editorial page of the WSJ seriously. I don't know any rational person who does. They have proven over and over again that they are lying ideologists who will say anything to support their political agenda no matter how far fetched. The quote you cite is just one more example. On the other hand, the news side of the WSJ has some of the best reporters in the business and fortunately there is a brick wall between them. Unfortunately the editorial side takes the wall so seriously that they don't even read their own front page when it contradicts their editorial views with objective facts. You are wasting your talents bothering to respond to them.
Posted by: Joseph at August 11, 2006 05:33 PM
At the risk of splitting hairs econo-barber style, I have to suggest that there's a non-trivial difference between (a) the classic Phillips curve as developed by Phillips and evolved by Samuelson and Solow, and the relationship between current economic activity (broadly defined) and future inflation.
From my perspective, Phillips' initial analysis of unemployment versus wage inflation in the U.K. was a straightforward labor-to-labor comparison. So far so good. But the later work by Samuelson-Solow that compared U.S. unemployment to CPI wasn't so straightforward nor as accurate or insightful as Phillips' initial work. Seems to me that unemployment versus CPI trade-off is one of those relationships that might hold very precisely for some time under a certain macro regime or other but eventually will break down and not work at all.
As an aside, Anarchus once worked briefly in the same small firm as a Dr. James Kichline, an economic expert who'd run the Federal Reserve Board's research division for a decade or more. Based on extensive Fed research under his direction (through the late 1980s), Dr. Kichline understood that there was a tight and unyielding relationship between capacity utilization in the U.S. and U.S. inflation. Unhappily for Kichline's interest rate forecasts in the private sector, however, an invasion of exogenous macro-factors in the early-to-mid 1990s wrecked the "Kichline Curve" and the tight relationship between capacity utilization and inflation was no more.
Anyway, I'm not about to defend the WSJ's conjuring up of the ghost of Arthur Burns (or if you want a more fearful ghost, think of G. William Miller! in full cry), but I do think that (a) the original Phillips' Curve was neat stuff, (b) the Samuelson-Solow extension was NOT neat stuff and is what's appropriately given the so called Phillips' Curve its bad name, and (c) that it's quite obvious that stronger economic activity today produces stronger prices tomorrow and vice versa.
Posted by: Anarchus at August 11, 2006 06:09 PM
While we're on the topic of bad reasons to disapprove of the Fed, many people seem to evaluate the Fed based on how the stock market is doing. More than one person has told me that Greenspan was a disaster, because his (the talker's) portfolio has not recovered to year 2000 levels.
Posted by: Bill Conerly at August 12, 2006 08:25 PM
Is there any good reason not to let the Fed simply be Taylor's rule running on a $1000 PC?
Posted by: Eric H at August 13, 2006 06:39 PM
I think that the Fed fears by slowing the economy too much,it will lead to a hard landing in the housing sector, and could spawn a repeat of the S&L crisis of the 80's.
Posted by: james at August 14, 2006 06:15 AM
I don't understand the relevance of SW's analysis to the WSJ editorial. SW looked at reduced form forecasting equations. As such, the coefficients in their equations would incorporate average behavioral relations over the sample period, including the average conduct of monetary policy and the average formation of expectations by the public. Yet none of those behavioral relations was constant over the 1970 to 1999 sample period. Therefore an estimated PC will display a lot of instability over various subsamples.
Now, the point is that a PC is only a reduced form, and therefore is not suitable for policy analysis. In particular, if I were to estimate a PC going back to 1970, I could calculate an enormous "sacrifice ratio," that is, an estimate that it would take a lot of elevated unemployment to reduce inflation slightly. Such an estimate is worthless, as Volker and Greenspan demonstrated. They understood the importance of expectations and took actions to rein in inflation expectations, and lowered inflation with much less cost than had been generally expected.
Thus I think the WSJ may have had a point (for once). If the Bernanke Fed were to allow an inflation scare to take root, it could prove costly to bring expectations back down. That was the real risk of pausing on 8/8, and should be in the minds of policymakers going forward.
Posted by: Lowsmoke at August 14, 2006 06:38 AM
If you put the first observation together with the previous comments about the yield curve, the result makes the yield curve an important indicator of Fed policy (as the new Fed Governor Rik Mishkin has long supposed). The yield curve model from Jonathan Wright puts the odds of recession within 12 months at about 66% if the funds rate rose to 5.75% but the 10-year yield remained near 5%, and 75% if funds and T-bills both hit 6% without raising bond yields. To the extent the bill rate is higher than the funds rate, the odds of recession would be higher still. If recession really is a reliable leading indicator of lower future inflation (after, not during, recession), then today's flat yield curve suggests at least a slowdown and perhaps lower inflation ahead. Core inflation fell from 1995 to 1999, and from 2001 to 2003, so it's not impossible.
Posted by: Alan Reynolds at August 14, 2006 08:50 AM
Way to put it, Joseph! I completely agree!
As to the possibility of future increases in inflation, I notice that inflation is now the highest it's been in this current round, that energy costs factor significantly in this inflation, and that oil prices are the highest they've been ever (give or take). Go Ben go! And good luck!
Posted by: Thomas James at August 14, 2006 09:24 AM
With American average household savings in the negative, interest rate hikes will put even more load on the average American. They are already driving less to save on gas. The economy is already showing signs of slowing down. If the interests did go too high, there is a very good possibility that the economy would go into recession as the average American might end up defaulting on her/his loan. Housing prices are falling already, so the average American can't borrow much from the prices of his/her house anymore. If they lose their jobs too, the sudden decrease in spending will have serious ramifications
I guess the best way to see whether Ben made the right choice is to see what happens come start of school. The sales data from major retail stores should shed light on the condition of the economy.
Posted by: Anthony at August 14, 2006 10:42 AM
Lowsmoke, of course neither Bernanke nor anybody else is proposing that there is a long-run tradeoff between inflation and unemployment that can be exploited by policy. Bernanke's beliefs on this are as follows:
the evidence of recent decades, both from the United States and other countries, supports the conclusion that an environment of price stability promotes maximum sustainable growth in employment and output and a more stable real economy.
In other words, to the extent there is a long-run Phillips Curve, it slopes up, not down.
The WSJ is not challenging Bernanke's belief in a downward-sloping long-run Phillips Curve (which "belief" Bernanke never for a moment entertained). Instead, it is specifically challenging the appropriateness of using evidence of a slowing economy as one indicator of whether policy is on track to achieve a long-run inflation goal without causing too big a short-run loss in output. And for purposes of this question, I believe the Stock-Watson evidence is quite useful and informative.
Posted by: JDH at August 14, 2006 11:24 AM
It appears that the only tool the FED has is to economic pain. Oh, this dismal science. Wouldn't it be unique if we actually let the economy produce at the rate it feels is best and we stopped worrying about how many people are employed and just let them work?
When I see all the analysis justifying why we should put people out of work I feel ill.
We have become too smart for our own good. What happened to classical economics?
Posted by: Dick at August 14, 2006 01:59 PM
An irony of this most recent vote is that the one FOMC member voting to raise interest rates, Jeffrey Lacker, president of the Richmond Fed, is a Ph.D. out of the University of Wisconsin-Madison, with Donald Hester as his major professor. Hester was a student of Tobin's at Yale, and vigorously debated monetary policy with Milton Friedman back in the 1960s, although the Richmond Fed has long been a bastion of monetarist sentiment within the Fed system. Of course a few years back the "hardliner" of the Board of Governors was Democrat and semi-Keynesian, Lawrence Meyer. Anyway, I find this all a bit ironic.
Posted by: Barkley Rosser at August 14, 2006 02:06 PM
Is it a little strange that when the FED increases interest rates to curb inflation that inflation increases, yet when the FED pauses increasing rates that inflation begins to decrease? Hmmm!
Is it possible that FED targets of the economy actually exacerbate inflation until they harm the economy so much that things begin to fall apart?
Posted by: Dick at August 17, 2006 06:19 AM