Persistently problematic productivity

A key part of the answer to all three questions is unusually weak labor productivity (output per unit of labor input). This trend is apt to continue, since it is a legacy of years of sub-trend capital formation and R&D in the US. However, one possible cause of any surprisingly better labor productivity growth could be pressure emanating from an unusually tight labor market.

Soft productivity growth has been evident in both key measures of productivity—labor productivity and multifactor productivity. Labor productivity growth and business output expansion are positively correlated; however, labor productivity growth over the past year has been unusually weak versus its normal relationship to business output growth. Multifactor productivity growth has also been softer than normal compared to the second half of the 1990s, as well as during the business expansion before the Great Recession.

Disappointing capital formation has been a major factor limiting labor productivity growth. The latter is the sum of contributions from capital intensity, labor input composition and multifactor productivity growth. In 2013—the latest year for which detailed multifactor productivity data are available—0.9% labor productivity growth was 120 basis points less than the 2.1% trend between 1987 and 2013. Three-fourths of that slowing—90 basis points—was due to no growth during 2013 in capital intensity (ie the capital/output ratio). In turn, capital intensity partly reflects business equipment spending as a percentage of GDP. It has been subnormal in the current economic expansion, especially for information processing.

A flat trend in the ratio of private R&D spending to GDP also helps explain weaker productivity growth. When that stops growing, there is subsequent slowing in labor productivity growth. Notably, we compare recent productivity growth with a four-year moving average of R&D/GDP. R&D’s effects on productivity take a number of years to materialize; in Figure 6, we illustrate a relationship between the annualized productivity growth over a two-year period and the change in the four-year moving average of R&D/GDP lagged by two years.