An alterative title might be my usual comment reminding Krugman’s readers of the history of “hysterisis”.
Krugman wrote a very clear even better than usual post summarizing how his views on macroeconomics have changed since 2008. It is best to click the link, but the one sentence summary is that he wasn’t surprised by the liquidity trap, but he was surprised that the Phillips curve seems to be flat at very low inflation rates so low inflation followed by a period of high unemployment doesn’t become deflation.
Off topic comment on how Krugman’s talk will be off topic.
I will exploit the flow of Thomists who came here by clicking Mark Thoma’s link here to get people to read the following semi relevant addition.
Krugman mentioned the original impetus for his post: “I’m supposed to do a presentation next week about “shifts in economic models,” “. I fear his presentation will be a bit off topic. That is I don’t think the failure of the accelerationist Phillips curve which surprised him will cause a shift in the models most macroeconomists use. It should — a horizontal Phillips curve implies that the economy isn’t self correcting even when it isn’t at the zero lower bound. But I guess most macroeconomists will continue to use models in which detrended variables converge to a unique steady state. The reason I am confident about this is the main point of this post. Such models failed utterly to fit European data from the 1980s on. But they are still used. The models were patched by making the unexplained exogenous trend very flexible and complicated.
The policy implications remained the same — take care of inflation and unemployment will take care of itself. They are now invulnerable to evidence, because when unemployment fails to converge to what had been considered a normal level, it is concluded that there must be nothing to be done, because it is assumed that it must it just must return to normal, so we have to accept 10% as the new normal if we aren’t Spanish, since they have to accept 20% as their new normal.
The key policy implication that one should focus on inflation became invulnerable to evidence in the 80s. It is an article of faith and methodological a priori. The claim that monetary policy and fiscal austerity can’t affect long run unemployment has become a no true Scotsman fallacy. When the prediction that tight money and disinflation would not be followed by high unemployment failed utterly, the prediction became that no true effect of monetary policy on unemployment could be permanent. The accelerationist Phillips curve can’t fail — it can only be failed.
Here I noted that almost universal macroeconomists’ confident claim about the long run self correction of the real economy is an a priori assumption, and not an implication of theory (that is other standard core assumptions such as rational expectations) or supported by evidence).
I wrote a semi snarky comment in which I suggested that the reason he was surprised is that the examples he always has in mind are the USA (of course) and Japan in the 90s, while he doesn’t constantly think about the European unemployment problem.
The natural unemployment rate hypothesis failed spectacularly in Europe in the 1980s. Extremely high unemployment did not lead to deflation — rather it coexisted with moderate inflation for a long time, then with low inflation.
Krugman posted a graph showing how the US graph of inflation and unemployment has changed (just click the link and look). In the past high unemployment gradually lead to lower inflation and then to lower inflation and unemployment — this is the pattern predicted by Friedman, Phelps, Tobin (and discussed already by Samuelson and Solow in 1960). But in the recent past extremely high unemployment has come with low and stable core inflation.
Things used look very different here in Italy than in the USA. Here is a graph of data from before January 2008. Extremely high unemployment was consistent with moderate and then with low inflation. The only clear shift in inflation occurred in 1996 and 1997 (which may or may not be when Italians began to think they might actually earn the wonderful reward of being allowed to adopt the Euro).
By 2008, The flat Phillips curve (the Fillipo curve?) was already very clear to anyone who read Italian newspapers.
Here are all data which are available on FRED (yes I sit in Rome and surf to St Louis for Italian data). Oddly the harmonized unemployment series is only available (at FRED) from 1983 on.
In this graph there is also very little sign of Friedman-Phelps cycles. The old pattern was a steady decline from extremely high inflation — it looks almost like an expectations unaugmented Phillips curve. But then (really from 1986 on) there was fairly stable moderate to low inflation along with extreme swings in unemployment. I stress that this is CPI inflation including food and energy not core inflation. the peak oil spike in 2007 and the collapse in 2008 are clearly visible. It is possible that the most recent observations show a slide to actual persistent deflation, but it is more likely that the recent decline in inflation is due to the collapse of the price of oil.