Will We Have Another “Jobless” Recovery?
After the last business cycle peak—still officially March 2001—labor productivity remained strong, but employment took longer than usual to recover. Some of the labor productivity gains had come from the high-tech capital boom of the late 1990s, which in turn was partially fueled by the need to address potential y2k problems. As we continue through what is at least another soft economic patch, an important question is whether labor productivity will repeat the same relatively strong performance observed after the last business cycle peak. While this would be good for real wages and living standards in the long run, it could again lead to a slow recovery in employment growth in the short run.
To answer this question, we look first at past trends. The first chart below plots growth in nonfarm labor productivity since the first quarter of 2001. As the chart shows, labor productivity is a fairly volatile series—even when viewed as year-over-year figures. Consequently, it is easier to tease out the underlying patterns if the raw data are statistically smoothed. The second chart plots the smoothed annualized quarterly growth rate for nonfarm labor productivity, output, and hours for a typical postwar expansion.
In a typical expansion, labor productivity, output, and hours all fall sharply after the peak but by different amounts. More significantly, they recover at different rates. Output growth plunges the most but then recovers quickly. It reaches its peak about 10 quarters after the previous peak, after which it tends to decline over the rest of the cycle. Growth in hours worked follows a similar path but does not tank quite as much. It recovers more slowly, about 14 quarters out. This slow recovery is evidence of a phenomenon called labor hoarding: Firms are reluctant to let skilled workers go during a temporary downturn because they may find other employment and thus not be available once conditions recover. The net result is that labor productivity (output per hour) recovers quickly at first (as output growth outpaces hours growth), but tends to slow once the recovery is well under way (as hours growth catches up with output growth).
The next chart illustrates the unusual trend of the current cycle. A smoothed version of labor productivity growth in the current cycle is plotted, along with the average of the previous postwar cycles and a corresponding 95 percent confidence interval. Productivity growth through the first 15 quarters after the peak was abnormally strong, just above the upper end of the range of the 95 percent confidence interval. The latest available value for year-over-year growth in nonfarm business labor productivity (2008:Q1) is 3.3 percent, the largest since the 3.8 percent reported for the 2003:Q2–2004:Q2 period.
Note that this analysis treats the current period as part of an ongoing expansion. This period would be even more unusual if it were treated as the beginning of the next period, as labor productivity has not dropped off as it usually does after a business cycle peak. Even though employment has declined for the past six months, real GDP has not fallen in any quarter, so it will be interesting to see how the NBER dating committee treats this period once all the data revisions are in.
How did this unusually strong labor productivity growth come about? First, output growth did not fall as much after the peak as it normally does, but over the cycle it has tracked at the lower bound of the 95 percent confidence interval. In the most recent observations it has even dropped below.
This weak output performance has been accompanied by even weaker growth in hours worked than normal. Some of this weaker-than-average growth is due to demographics: The labor force is growing more slowly this cycle than in previous postwar cycles—a trend that is likely to continue. Since the 1950s, labor force growth has averaged 1.6 percent, but it is widely expected to continue to slow to no more than 0.7 or 0.8 percent by the middle of the next decade.
Going forward, labor productivity’s growth rate will depend, in part, on whether the economy continues to expand or whether it enters a recession. It will also depend on which sectors experience relatively more growth. Currently, exports, of which manufactured goods comprise a large share, are doing relatively well given the continued growth overseas and the weak dollar. As manufacturing’s gains in labor productivity have tended to be stronger than the service sector’s, this should result in at least a modest boost to overall labor productivity. With the economy in transition, adjusting to higher energy prices and lower housing prices, overall output demand is likely to be weak. Combined with robust labor productivity growth, this could lead to weak employment growth as well.