Read

Problems with Productivity ... or Not

Last week the Bureau of Labor Statistics announced that its measure of worker productivity fell by 0.9% in the first quarter of 2012. Spinning this...

Last week the Bureau of Labor Statistics announced that its measure of worker productivity fell by 0.9% in the first quarter of 2012. Spinning this as positive news — which several politicians and economic commentators did — makes as much sense as asking Mrs. Lincoln if she enjoyed the play. Pretending that this drop might have some salutary effect on the economy suggests a fundamental misunderstanding not only of what productivity is, but also of how it is measured. Labor productivity is a measure of how much a worker produces in a given hour of work. In the long run it is the most important determinant of our standard of living. If workers aren’t producing more goods and services, then it’s impossible for the economy to expand in a sustained way. From World War II until the early 1970s, productivity grew at about 2.5% a year, a rate at which our standard of living doubles every generation or so. Productivity then mysteriously slowed to about half that rate for the next quarter century. While it’s not entirely clear what caused this diminution, economists think that it had something to do with the oil crisis and the massive influx of new workers into the economy as the baby boom generation reached adulthood and women entered the labor force in large numbers. High oil prices and a surplus of labor led firms to spend precious R&D dollars economizing on energy while worrying less about labor productivity. In the mid 1990s productivity growth again surpassed 2.5% a year, likely due to the fruits of the I.T. revolution, prodded by the retail sector’s embrace of these technological gains. More recently, productivity has fallen, and a few economists have suggested that our rapid gains of the previous 15 years may have dissipated. However, we aren’t that good at measuring productivity, so our current picture isn’t terribly clear. For instance, there is evidence that some of the measured gains in productivity have been the result of our statistical agencies conflating gains resulting from trade with gains resulting from actual advances in worker productivity. Other economists suggest that the outsized contribution of massive increases in computing power to our measure of productivity growth overstates the actual gains to our economy. Yet another problem is that measuring productivity is tricky in the short run because we do not measure hours worked very well. While people who work on an assembly line and are paid by the hour can accurately report how many hours they worked last week, most people no longer have such jobs. If asked, they will likely spend a few seconds contemplating the previous week and then make a good-faith guesstimate, with a few hours added to make them look industrious. Here’s the problem this causes: When the economy begins to slow, companies respond to a decreased demand for their services by first reducing hours worked. This is because companies don’t want to jettison skilled labor that was costly to hire and train in the first place, and the savings from any short-term layoff tends to be slight. (We call this phenomenon labor hoarding). The problem — at least for the BLS — is that since they don’t measure hours very well, there is a tendency to overstate the declines in productivity. For example, suppose a firm cuts back production and hours by 5% each. The data gathered by the BLS will indicate output falling by 5% (which is accurate), but it is likely reported hours will have fallen only, say, 2%. As a result, it looks like the average worker somehow became less productive when in reality the data are just incorrect. This likely explains why our measured productivity fell last quarter. Absent some sort of gigantic disaster befalling our country, actual productivity never falls. When it is reported to have fallen, it usually just means that the economy is slowing down and companies have begun cutting back hours. If the slack in demand persists, companies will start cutting workers instead of hours; when this happens, the economy is in real trouble.