The rate of decline in employment moderated substantially in May, according to the BLS figures released June 5, to about half the monthly rate of job loss recorded over the preceding six months (345,000 vs. 642,000). The news was received in a variety of ways.
First, the cynics. They tend to wax sarcastic at the idea of “things are not getting worse quite as fast as they were” as a good-news proposition. But a wide variety of recent data indicate that the economy is no longer in the state of free-fall that it entered last September, and this is indeed good news. To begin to level off is the first step toward the start of the recovery.
Second, the academics note (correctly) that there is little information in each individual monthly statistical fluctuation that is measured, because the data are inevitably noisy. Still, the public wants to know, in real time, what is the best we can glean from the information we have.
Third, the financial press, in particular, had been asking whether this quarter could turn out to be the bottom of the recession. The May employment report encouraged enthusiastic speculation that the answer was “yes.” The stock market reacted positively.
The members of the NBER Business Cycle Dating Committee (of which I am one) will be responsible for calling the trough when the time is right. We have a range of views regarding the proper place of employment numbers in such deliberations. But one can say, on the one hand, that a decline in economic activity is a decline in economic activity, and therefore still a state of recession, even if the rate of decline has moderated a lot. One can also say, on the other hand, that employment is usually a lagging indicator of economic activity. (For example, the economy continued to lose jobs long after the ends of the 1991 and 2001 recessions. Hence the “jobless recoveries.”)
Speaking entirely for myself, I like to look at the rate of change of total hours worked in the economy. Total hours worked is equal to the total number of workers employed multiplied by the average length of the workweek for the average worker. The length of the workweek tends to respond at turning points faster than does the number of jobs. When demand is slowing, firms tend to cut back on overtime, and then switch to part-time workers or in some cases cut workers back to partial workweeks, before they lay them off. Conversely, when demand is rising, firms tend to end furloughs, and if necessary ask workers to work overtime, before they hire new workers. (The hours worked measure improved in April 1991 and November 2001 which on other grounds were eventually declared to mark the ends of their respective recessions.) The phenomenon is called “labor hoarding” and it is attributable to the costs of finding, hiring and training new workers and the costs in terms of severance pay and morale when firing workers.
Unfortunately, pursuing this logic leads to second thoughts about whether the most recent BLS announcement was really good news after all. Forbes picked up on the point. The length of the average work week fell to its lowest since 1964 ! The graph below shows that, not only did total hours worked decline in May, but the rate of decline (0.7%) was very much in line with the rate of contraction that workers have experienced since September. Hours worked suggests that the hope-inspiring May moderation in the job loss series may have been a monthly aberration. If firms were really gearing up to start hiring workers once again, why would they now be cutting back as strongly as ever on the hours that they ask their existing employees to work? If one factors in falling wages, to compute total weekly earnings, the picture looks still worse. My bottom line: the labor market does not quite yet suggest that the economy has hit bottom.