Saturday, April 02, 2011

Expectations Vs History

This is my daily comment on Matt Yglesias. Yglesias suggests that the speed of recoveries from recessions mgiht be much influenced by the expected speed of recovery. He notes dramatic policy changes in 1982 and the dramatic inauguration of FDR.

This is an interesting hypothesis.

One issue is the extent to which expectations are self fulfilling. This is one of the many issues on which economic research as expressed in top refereed journals is very different from the vague thoughts of the man on the street. In most of those peer reviewed articles, expectations can't play an independent role. My view is that, as usual, the vague thoughts about macroeconomics of the man in the street are more nearly consistent with the data than are the academic articles. Basically, models in which expectations can have an independent role are disfavored as they imply that economists can't predict the future. Such models exist, but they are a minor sub field. This tells us about economists' preferences, but nohting about the world we study. I assure you it is a methodological a priori that good models have a unique equilibrium (so expectations can't have an independent role).

One way to try to see if expectations matter is to take advantage of errors in measurement. The BEA publishes flash GDP estimates based on limited data, then reports it's final estimates based on much more data. Thus it claims that it incorrectly promoted optimisim if the flash estimate turns out to be greater than the final estimate. It turns out that in these cases GDP growth (as eventually thoroughly measured) is higher than in cases in which the flash estimate was lower than the final measurement. See "The Macroeconomic effects of false announcments Oh and Waldman Quarterly Journal of Economomics vol 105 pp 1017-34

This paper shows, among other things, that Waldman is a brilliant economist (Waldman, Michael not Waldmann, Robert and if you think that bothers me don't ask me about the time I received a fax inviting uh someone to apply for a job at a very top place which had the salutation "dear Mike").

OK that's my contribution as an (envious) economist. More generally, I don't think the story fits the facts. I'm not a businessman, but I sure don't remember much optimism in 1982. There was a huge change in interest rates (such a change now would make them very negative). This changes things even if expected future GDP doesn't change at all.

At a key point in the current recession, there was a political event which seemed quite dramatic at the time -- the election of Obama. There was quite a lot of hope and a very large crowd at the inauguration. Notably FDR didn't grasp Keynesian reasoning (which was rather new back then) and he did not attempt a massive fiscal stimulus. The shock of his election was somewhat reduced by the fact that he was upper class and had the same last name as previous Republican president (my paternal grandparents voted for Norman Thomas because they thought FDR was just the upper class twit of the year).

In 2008-9 there were also huge dramatic policy initiatives. The ARRA
may have been too small by half (or too small by two thirds reasonable people continue to debate within that range) but it was much more resolutely Keynesian than anything Roosevelt did before Pearl Harbor. The Fed might be able to do more than it has, but it certainly did much much more than it ever had before. TARP was unpopular, but it was huge and a temporary setback (no vote in the House) terrified investors.

I am very confident that the difference between 1933 and 2009 is the difference between years and months. Roosevelt was inaugurated four months after the crash and Obama was inaugurated four months after the crash. If Obama had just been inaugurated this January, people would consider him an economic savior because of the last couple of employment reports. It is very hard to look good if you take office while the economy is falling like a rock.

1 comment:

Aidan said...

I'm confused. "Roosevelt was inaugurated four months after the crash and Obama was inaugurated four months after the crash" - shouldn't that say Roosevelt was inaugurated four years after the crash?