Jobless Recovery: Fasten Your Seatbelts

by: Brad DeLong

20090717.xlsClick to enlarge

[20090717.xls]Sheet1 Chart 1Click to enlarge

As of this writing, it looks as though the average unemployment rate in 2009 is going to average at least 1.5 percentage points above where last December the incoming Obama administration thought that it was likely to be. Instead of the 7.8% forecast last December, year-2009 unemployment looks to average 9.3% or higher. Year-2009 real GDP also looks to be lower than the income Obama administration was forecasting last December: $11.40 rather than $11.53 trillion. The macroeconomic news has been bad. The financial crisis that gathered force from the summer of 2007 through the summer of 2008 and then exploded after the collapse of Lehman brothers did more damage to the economy than the consensus of forecasters had imagined.

Back in the 1960s one of President Johnson's economic advisers, Brookings Institution economist Arthur Okun, set out a rule of thumb others quickly named "Okun's Law": if production and incomes--GDP--rises or falls 2% because of the business cycle, the unemployment rate will fall or rise by 1% along with it: the magnitude of swings in the unemployment rate will be half or a little less than half the magnitude of swings in GDP. Why? For four reasons: (a) businesses will tend to "hoard labor" in recessions, keeping useful workers around and on the payroll even if there is temporarily nothing for them to do; (b) businesses will cut back hours when unemployment rises, and so output will fall more than proportionately because total hours worked will fall by more than total bodies employed; (c) plant and equipment will run less efficiently when hours are artificially shortened because of the recession; and (d) some workers who lose their jobs won't show up in the unemployment statistics but will instead retire or drop out of the labor force. For all four of these reasons, whatever rise in the unemployment rate we see in a recession is supposed to be a fraction of the fall we see in GDP relative to trend.

But this time we are not following this rule. This time Okun's Law is being broken. The unexpected 1.2% extra decline in real GDP in 2009 should have been accompanied by a 0.5 or 0.6 percentage-point rise in the unemployment rate, not by the 1.5 percentage point rise in the unemployment rate we are now seeing. I confess that the fact that this is happening comes as a surprise to me. But when I think back we have seen this before. In 1993--two full years after the National Bureau of Economic Research had called the end of the 1990-1991 recession--the unemployment rate was still higher and the employment-to-population ratio lower than it had been at the recession trough. And we saw the same "jobless recovery" after the recession of 2001: it took 55 months after the formal end of the recession in November 2001 before a greater share of Americans had jobs than had had them in November of 2001.

It is likely to be a recovery. The central tendency forecast right now is that real GDP contracted at a rate of 1% per year or less between the first and second quarters of 2009, and will grow between the second and third quarters at a rate of 2% per year or so. When the NBER Business Cycle Dating Committee gets around to it, it is most likely to call the end of the recession for June 2009, second most likely to call its end in April, and a recession-end date later than June 2009 is a less likely possibility. One reason that we are likely to see a recovery starting... right now... is the stimulus package. It probably boosted the real GDP annual growth rate relative to what otherwise would have been the case by about 1.0 percentage point in the second quarter, and is going to boost the annual GDP growth by about 2.0 percentage points between now and the summer of 2010--after which its effects tail off.

But it will not feel much like a recovery. After the 1982 recession the turnaround in employment lagged the turnaround in GDP by only six months. Thereafter employment growth was very strong: in the eighteen months up until the end of 1984, growth in work hours averaged 4.8% per year. It took only 7 months after the 1982 recession trough for the employment-to-population ratio to rise above its trough level (1980: 2 months. 1975: 5 months. 1970: 18 months. 1961: 13 months. 1958: 4 months. 1954: 8 months.) By contrast, it took 29 months after the 1991 recession trough for the employment-to-population ratio to exceed its trough level, and 55 months after the 2001 recession trough for the employment-to-population ratio to do so. Productivity growth in the immediate aftermath of the end of the 1991 and 2001 recessions was surprisingly rapid: rapid enough to eat up all of real demand growth and more as businesses decided to take advantage of the economic downturn to slim down their labor forces and become more efficient.

Today--unless we get much faster real GDP growth than currently looks to be in the cards--we are headed for a jobless recovery. The answer to the economic question--was the stimulus sufficient to rapidly return the economy to something like normal unemployment?--is likely to be: "h--- no, it was much too small..."