Chapter 15: A Forecaster's Nightmare: Predicting Job Creation in the early 21st Century
One of the more puzzling aspects of the U.S. economy since 2001 has been the seeming difficulty in creating new jobs. In the late 1990's, economists were becoming concerned about chronic labor shortages as a smaller cohort of younger workers was entering the market, and the retirement of the large "baby boom" cohort was approaching. The collapse of the dot.com bubble, taking the NASDAQ index from over 5,000 to under 2,000 wiped out enough paper wealth in 401k plans, IRA's and pensions so that a lot of those retirements will now come later than might have been anticipated in 1999. And, of course, the economy proved that the business cycle was not dead, by throwing in a relatively mild recession in 2001-2002.
But what's happened since then has been a surprise to virtually all of the experts. While the national economic recovery has been neither as steady nor as robust as the recovery from the previous recession in 1991-92, it has lagged behind almost every recovery on record in the creation of new jobs. Democrats, unsuccessfully, tried to blame this on President George W. Bush throughout the 2004 campaign, arguing that he would be the first president since Herbert Hoover to end his first term in office with fewer Americans working than when he started. The victory in Ohio that was crucial to the President's re-election came despite significant job losses in the Buckeye state.
While policy makers continued to struggle with the issue of job creation well into 2005, so too did economic analysts, who ran up a rather miserable record in predicting the amount of job growth. While making accurate predictions about the job growth numbers on a monthly basis certainly can be extremely difficult, the margin of error seems to be expanding, and perhaps even more troubling, the direction of error has become more consistent in over-predicting.
A series of articles from The Wall Street Journal discussed a variety of factors that perhaps suggest changes in the nature of unemployment, along with other factors that have kept job creation under wraps, and job creation forecasters wishing their forecasts had been kept under wraps.
If there is a silver lining in the dark cloud of unemployment, it has been the growth in productivity. The long-run annual average growth in productivity has been between 2 and 2.5 percent in the post-World War II period, but over the past few years, growth has been more in the 4 percent range, with some quarterly rates of double or even triple the long-run average. Higher productivity per worker means employers don't need as many workers to produce their output, and helps explain the reluctance of employers to hire new employees, and the tendency of forecasters relying on historical averages to predict that they will. While higher productivity should result in lower costs and higher profits for firms and ultimately higher wages for workers, softening the blow of less job creation, so far is has seeming only been helping employers.
Another factor impacting domestic job creation has been the increased trend towards outsourcing of jobs. This also entered the political realm in the 2004 campaign, with some calling for restrictions on the right of firms to ship jobs overseas. Though not much has happened on that front, the combination of outsourcing, both domestically and abroad, with other structural changes in the economy is hampering domestic job creation and lengthening the average duration of unemployment. Historically, many workers who were laid off during an economic downturn or a slow period at their firm would eventually return to work at their old jobs. That continuity is growing less and less common, and job loss for many now means a change in career as well, extending the duration of their unemployment.
A third factor hampering job growth has been the increase in medical costs, which make employers more reluctant to take on new employees for whom they will need to provide insurance. Although wage pressures have generally been mild, an employer's total outlay for new employees is the relevant consideration (total marginal outlay, that is), and with the recovery starting and slowing several times, many employers have been using temporary workers instead of taking on new permanent employees they aren't sure they will be able to retain if business slows again.
The focus of two of The Wall Street Journal articles was more specifically with the job creation estimates than job creation itself. Aaron Luchetti took a user's point of view, discussing the difficulties bad estimates were causing bond traders. When the March 2004 job growth fell over 100,000 short of the consensus estimate, it marked the ninth out of eleven forecast "misses" that had overestimated job growth since 2000. Jon Hilsenrath looked at the estimates from the estimators' point of view. One possible source of the forecasting errors lies in a common claim that the reported payroll numbers themselves are understating the actual amount of job creation. But a bigger factor in Hilsenrath's report was the possible bias in forecasting models that stem from surveys of executives on plans for new hiring. Surveys, such as those by the Business Roundtable, ask employers to indicate whether they expect to increase or decrease employment in the next quarter, six months, or year. But many don't ask by how much they expect employment to increase, and data used in the models may be accurately representing widespread optimism, but less accurately capturing generally modest intentions.
Sources:
"Payroll Slump Has Economists Rethinking Ideas on Job Creation," by Jon E. Hilsenrath, The Wall Street Journal, September 8, 2003 (p. A2)
"U.S. Employment Climbs Slowly, The Fed Reports," by James R. Hagerty, The Wall Street Journal, March 4, 2004 (p. A2)
"Job Growth Falls Short of Estimates," by James R. Hagerty and Jon E. Hilsenrath, The Wall Street Journal, March 8, 2004 (p. A2)
"Weak Jobs Data Surprise Bond Arena," by Aaron Luchetti, The Wall Street Journal, March 8, 2004 (pp. C1, C6)
"Bets On Jobs Report Proved Perilous," by Justin Lahart, The Wall Street Journal, March 8, 2004 (p. C3)
"Good Intentions May Skew Hiring Data," by Jon E. Hilsenrath, The Wall Street Journal, March 15, 2004 (p. A2)
Despite the slow growth in new jobs, the unemployment rate declined from January 2003 to early 2005. With significant job losses going on during that same time, how is this possible?
Although they were not explicitly discussed in the case, what kinds of unemployment did you see evidence of-frictional, structural, or demand- deficient/cyclical?
In 2003 and 2004, the U.S. ran record federal budget deficits and had some of the lowest interest rates in history. Were there policy options available to help combat cyclical, or demand-deficient employment? How about structural unemployment? Describe them, and indicate their strengths and weaknesses.
