« 'The Evolution of Scale Economies in U.S. Banking' | Main | Links for 09-03-15 »

Wednesday, September 02, 2015

'The Triumph of Backward-Looking Economics'

In response to Paul Krugman's recent post, "The Triumph of Backward-Looking Economics" -- no surprise here -- there is  disagreement from Steve Williamson. So let me offer this from Blanchard and Johnson's intermediate macroeconomics text discussing this issue. But first, a brief review of the controversy. Krugman says:

... What did orthodox salt-water macroeconomists believe about disinflation on the eve of the Volcker contraction? As it happens, we have an excellent source document: James Tobin’s “Stabilization Policy Ten Years After,” presented at Brookings in early 1980. Among other things, Tobin laid out a hypothetical disinflation scenario based on the kind of Keynesian model people like him were using at the time (which was also the model laid out in the Dornbusch-Fischer and Gordon textbooks). These models included an expectations-augmented Phillips curve, with no long-run tradeoff between inflation and unemployment — but expectations were assumed to adjust gradually based on experience, rather than changing rapidly via forward-looking assessments of Fed policy.
This was, of course, the kind of model the Chicago School dismissed scathingly as worthy of nothing but ridicule, and which was more or less driven out of the academic literature, even as it continued to be the basis of a lot of policy analysis. ...

Krugman goes on to claim that the Chicago School was wrong. Williamson says:

... Tobin is using what he thinks is a conventional macroeconometric model. It's got adaptive expectations and a Phillips curve with what people then would have called a "high sacrifice ratio." You have to suffer a lot of unemployment to get a small reduction in inflation. Of course, Tobin's simulation looks nothing like what happened. ...

He is focused on whether a particular set of assertions by Tobin are correct, and he claims they do not fully fit the evidence. But there is a larger issue here that goes far beyond the particulars of Tobin's paper (and a broad view of the evidence is supportive of Krugman's view of the Volcker disinflation). Blanchard and Johnson:

... Fighting inflation implied tightening monetary policy, decreasing output growth, and thus accepting higher unemployment for some time. The question arose of how much unemployment, and for how long, would likely be needed to achieve a lower level of inflation, say 4%-which is the rate Volcker wanted to achieve. Some economists argued that such a disinflation would likely be very costly. .....
The natural conclusion was that it would make sense to go slowly, so as not to increase unemployment by too much in a given year. Some economists argued that disinflation might in fact be much less costly. In what has become known as the Lucas critique, Lucas pointed out that when trying to predict the effects of a major policy change-like the change considered by the Fed at the time-it could be very misleading to take as given the relations estimated from past data. In the case of the Phillips curve, taking equation (8.10) [note: this is the backward looking Phillips curve specified below] as given was equivalent to assuming that wage setters would keep expecting inflation in the future to be the same as it was in the past, that the way wage setters formed their expectations would not change in response to the change in policy. This was an unwarranted assumption, Lucas argued: Why shouldn't wage setters take policy changes directly into account? If wage setters believed that the Fed was committed to lower inflation, they might well expect inflation to be lower in the future than in the past. ...
Lucas did not believe that disinflation could really take place without some increase in unemployment. But Thomas Sargent, looking at the historical evidence on the end of several very high inflations, concluded that the increase in unemployment could be small. The essential ingredient of successful disinflation, he argued, was credibility of monetary policy-the belief by wage setters that the central bank was truly committed to reducing inflation. Only credibility would cause wage setters to change the way they formed their expectations. Furthermore, he argued, a clear and quick disinflation program was more likely to be credible than a protracted one that offered plenty of opportunities for reversal and political infighting along the way.
Who turned out to be right? In September 1979, Paul Volcker started increasing the interest rate so as to slow down the economy and reduce inflation. From 9% in 1979, the three-month Treasury bill rate was increased to 15% in August 1981. The effects on inflation, output growth, and unemployment are shown in Table 8-1. The table makes clear that there was no credibility miracle: Disinflation was associated with a sharp recession, with negative growth in both 1980 and 1982, and with a large and long-lasting increase in unemployment.
Does this settle the issue of how much credibility matters? Not really. Those who argued before the fact that credibility would help argued after the fact that Volcker had not been fully credible. After increasing the interest rate from September 1979 to April 1980 and inducing a sharp decrease in growth, he appeared to have second thoughts, reversing course and sharply decreasing the interest rate from April to September, only to increase it again in 1981. This lack of consistency, some argued, reduced his credibility and increased the unemployment cost of the disinflation. A larger lesson still stands: The behavior of inflation depends very much on how people and firms form expectations. The Lucas critique still stands: The past relation between unemployment and inflation may be a poor guide to what happens when policy changes. ...

They are saying that a Phillips curve of the form π = πe - α(U - Un), where πe  at time t equals πt-1 is a good model of the Volcker era (that is, expectations were backward looking), and that was reasonable at the time. However, as expectations formation changes with experience this may no longer hold true -- the tradeoff could change as expectations are more responsive to Fed announcements. Krugman is arguing that recent experience lends credence to the idea that a backward looking model (or a model with a linear combination of past and rationally expected inflation, i.e. a model with persistence in unemployment in response to policy changes) continues to hold.

    Posted by on Wednesday, September 2, 2015 at 12:07 PM Permalink  Comments (28)


    Feed You can follow this conversation by subscribing to the comment feed for this post.