Monday, August 30, 2010
Robert Barro, Ivy League Retardation Demo Case
Robert J. Barro is Paul M. Warburg Professor of Economics at Harvard University, a senior fellow of the Hoover Institution of Stanford University, and a research associate of the National Bureau of Economic Research. He has a Ph.D. in economics from Harvard University and a B.S. in physics from Caltech. Barro is co-editor of Harvard’s Quarterly Journal of Economics and was recently President of the Western Economic Association and Vice President of the American Economic Association. He is honorary dean of the China Economics & Management Academy, Central University of Beijing. He was a viewpoint columnist for Business Week from 1998 to 2006 and a contributing editor of The Wall Street Journal from 1991 to 1998. He has written extensively on macroeconomics and economic growth. Noteworthy research includes empirical determinants of economic growth, economic effects of public debt and budget deficits, and the formation of monetary policy. Recent books include Macroeconomics: A Modern Approach from Thompson/Southwestern, Economic Growth (2nd edition, written with Xavier Sala-i-Martin), Nothing Is Sacred: Economic Ideas for the New Millennium, Determinants of Economic Growth, and Getting It Right: Markets and Choices in a Free Society, all from MIT Press. Current research focuses on two very different topics: the interplay between religion and political economy and the impact of rare disasters on asset markets.A very good resume, right? The physics degree had me whimpering and groaning. A physicist ought to be able to do better.
So how do I, the redneck cracker, dare to contradict the esteemed Barro? Well, read on and judge for yourself.
In the article, Herr Professor Doktor Physiker Barro compares the current recession to 1982, and contends that the cause of our current high unemployment is extended unemployment benefits. I'm not making this up:
To begin with a historical perspective, in the 1982 recession the peak unemployment rate of 10.8% in November-December 1982 corresponded to a mean duration of unemployment of 17.6 weeks and a share of long-term unemployment (those unemployed more than 26 weeks) of 20.4%. Long-term unemployment peaked later, in July 1983, when the unemployment rate had fallen to 9.4%. At that point, the mean duration of unemployment reached 21.2 weeks and the share of long-term unemployment was 24.5%. ... The peak unemployment rate of 10.1% in October 2009 corresponded to a mean duration of unemployment of 27.2 weeks and a share of long-term unemployment of 36%. The duration of unemployment peaked (thus far) at 35.2 weeks in June 2010, when the share of long-term unemployment in the total reached a remarkable 46.2%. These numbers are way above the ceilings of 21 weeks and 25% share applicable to previous post-World War II recessions. The dramatic expansion of unemployment-insurance eligibility to 99 weeks is almost surely the culprit. To get a rough quantitative estimate of the implications for the unemployment rate, suppose that the expansion of unemployment-insurance coverage to 99 weeks had not occurred and—I assume—the share of long-term unemployment had equaled the peak value of 24.5% observed in July 1983. Then, if the number of unemployed 26 weeks or less in June 2010 had still equaled the observed value of 7.9 million, the total number of unemployed would have been 10.4 million rather than 14.6 million. If the labor force still equaled the observed value (153.7 million), the unemployment rate would have been 6.8% rather than 9.5%.All of the quoted figures seem to be correct or close to correct, but there is one aspect of this situation that Barro seems to have overlooked from his ivory tower. To explain, I will use BLS data from the household survey (get it here by selecting seasonally adjusted Employed ):
This is 1975-1985 total employed. Note that the impact of the 80/81 recessions is very evident. Total employment was quite flat from 1979 through the first half of 1983. The population was not. Hence, sorrow.
NBER has the 1981 recession ending in November of 1982. Employment is a lagging indicator, but it did begin to rebound very strongly in the spring of 1983.
This is 2000-2010 (so far) total employed. That is quite a camel. Note that the current number of jobs is most comparable to the level in the first half of 2004. Hence, sorrow.
We're only one year away from the proverbial seven years of bad luck, kids. Six years of no net new jobs is a lot worse than four years.
Now most people not entitled to introduce themselves as Herr Professor Doktor Physiker (insert name here) would look at this and think, Gee, we have very high long term unemployment because we lost a staggering number of jobs, not because people are refusing to go out and get jobs.
Mr. Barro is quite correct that people will stay on unemployment longer if unemployment is offering them a higher level of income than any job they can find. And he is correct in observing that some European countries have produced a high level of structural unemployment by offering very generous unemployment benefits over a long term.
However it does not necessarily follow that in this case, extended unemployment benefits are causing long term unemployment. It rather looks as if long term unemployment is causing extended benefits.
To settle the argument experimentally, we would have to find a US population that gets no unemployment benefits, wouldn't we? And then we could compare unemployment in that population to unemployment in the population that gets extended unemployment benefits and see if people are behaving differently in the two populations.
Who'd 'a thunk? Apparently not the Ivy League! We have such a population in youngsters. There are young people graduating from high school, colleges and universities. Very few of the qualify for unemployment benefits, because even if they have been working since they were fourteen, their jobs are almost solely classed as temporary. You get this information from Table A-10 of the household survey, unless you are firmly ignoring any evidence which contradicts your hypothesis:
Unemployment rates 16-19 years:
Unemployment rates 20-24 years:
Unemployment rates 25-54 years
Unemployment rates 55-> years:
This is the reciprocal of what we would expect to see if Mr. Barro's hypothesis were true. By far the highest percentage of households with fixed income streams from non-employment fall in the 55 and over bracket. The persons in our economy who are least likely to have other income streams are teens and early twenties. Yet as you move up the age/assured income ladder, unemployment rates fall dramatically. We have over 25% unemployment among teens dropping to 7% unemployment in the 55 and over group.
The reason, of course, is that more experienced workers are pushing less experienced but more desperate workers out of jobs. This trend will continue, and in large part the growth of teen and early twenties unemployment is because persons with some but inadequate retirement income have been supplementing their incomes by taking jobs that do not pay a living wage away from younger people. This trend was clearly evident even before the recession, and it is due in part to the change in the CPI calculation instituted in the late 1990s by the Clinton administration. This has slowly reduced Social Security benefits and the accumulated effect will persist even if the US economy were to rebound sharply.
It is worth closing this by reminding everyone of several elementss:
A) The household survey uses a very expansive definition of "employment". It's anything you do for pay. In the household survey, you can be counted as "employed" if you mowed someone's lawn for money once in the survey week, or if you spent a day collecting tin cans for sale to a salvage point. Such marginal "employment" is probably a larger percentage of those "employed" now than it was during the 1980s.
B) The demographics of the US population are very different than they were during the 80s. Retirements have cut the unemployment rate in the US during this cycle compared to the 80s. Unemployment would have peaked near 12% in this cycle if it were not for retirements.
C) The structural change involved in older workers pushing younger workers out of the employment pool will increase over the next two decades.
D) I am genuinely shocked that an individual with such resources who has done highly estimable work would go into print with such a shoddily researched claim.
In conclusion, I will post the same series of graphs for the years 1975-2010, so you can see that the differences pool at the ends of the age spectrum:
16-19 years; 1975 to 2010. Note the much higher peak of youth unemployment.
20-24 Young adult 1975-2010. Very similar, although slightly higher at peak, in this cycle so far. Few in this age bracket qualify for unemployment benefits.
20-54 years. 1975-2010. Almost the same over the two bad recessions. In this bracket, most qualify for unemployment benefits, although over time the incidence of contract workers and self-employed who do not qualify has risen somewhat.
The unemployed geezers aged 55 and over from 1975 to 2010. Here, as in teens, you see a very significant difference. In part this is because far more geezers are seeking employment. A BLS write up on the older workforce from April of this year. Nice graphs. Too bad Barro didn't read it.
But we're not going to get anything like that. We're going to have our choice between leftist fundamentalists and libertarian fundamentalists. And so it goes. "There is a tide in the affairs of man..."
I mean, this truly must be an embarassing moment for the critiqued author. Nonetheless, his peer group will probably dismiss your argument as 'white trash' in spite of its irrovocable evidence.
Have you been following the ruckus caused by Minn. Fed president Narayan Kocherlakota?
Another genius that misses the economic wood for the mathematical tree.
But M_O_M, this is masterly, demonstrating with numbers what we all knew by instinct. It takes a lot of schooling and hard work to be ignorant enough to be an economist.
Links to this post: