The productivity debate

The good news, according to the New York Times:

Productivity of U.S. companies rocketed at a 9.4 percent annual rate in the third quarter, the best showing in 20 years, offering an encouraging sign that the economic resurgence will be lasting. The increase in productivity — the amount an employee produces per hour of work — reported by the Labor Department on Wednesday was even stronger than the 8.1 percent pace initially estimated for the July-to-September quarter a month ago and was up from a 7 percent growth rate posted in the second quarter of this year. The third-quarter’s productivity gain, based on more complete data, was better than the 9.2 percent growth rate economists were forecasting and marked the strongest performance since the second quarter of 1983, when productivity grew at a blistering 9.7 percent rate. The report raised new hopes that businesses may be more confident than before that the economic rebound is genuine.

The bad news, according to Stephen S. Roach writing in last Sunday’s New York Times — the way productivity is being measured leads to a likely overestimation of that figure:

productivity measurement is more art than science — especially in America’s vast services sector, which employs fully 80 percent of the nation’s private work force, according to the United States Bureau of Labor Statistics. Productivity is calculated as the ratio of output per unit of work time. How do we measure value added in the amorphous services sector? Very poorly, is the answer. The numerator of the productivity equation, output, is hopelessly vague for services. For many years, government statisticians have used worker compensation to approximate output in many service industries, which makes little or no intuitive sense. The denominator of the productivity equation — units of work time — is even more spurious. Government data on work schedules are woefully out of touch with reality — especially in America’s largest occupational group, the professional and managerial segments, which together account for 35 percent of the total work force. For example, in financial services, the Labor Department tells us that the average workweek has been unchanged, at 35.5 hours, since 1988. That’s patently absurd. Courtesy of a profusion of portable information appliances (laptops, cell phones, personal digital assistants, etc.), along with near ubiquitous connectivity (hard-wired and now increasingly wireless), most information workers can toil around the clock. The official data don’t come close to capturing this cultural shift. As a result, we are woefully underestimating the time actually spent on the job. It follows, therefore, that we are equally guilty of overestimating white-collar productivity. Productivity is not about working longer. It’s about getting more value from each unit of work time. The official productivity numbers are, in effect, mistaking work time for leisure time.

A quick perusal of Roach’s writings reveal him to have replaced Henry Kaufman as the Dr. Doom of the U.S. economy. That said, he’s raising a fair point about measurement issues here. UPDATE: On the other hand, Tyler Cowen has a post suggesting that productivity gains have been underestimated.