What Do the Purified Solow Residuals Tell Us about Japan's Lost Decade?

In the 1990s, measured productivity growth decelerated dramatically in Japan. However, the standard Solow residuals may reflect factors other than changes in the rate of technological progress. This paper attempts to construct a measure of "true" aggregate technical change for the Japanese economy over the years 1973-98, controlling for increasing returns, imperfect competition, cyclical utilization of capital and labor, and reallocation effects. We find little or no evidence of a decline in the pace of technological change during the 1990s. Both cyclical utilization and reallocations of inputs have played an important role in lowering measured productivity growth relative to true technology growth. Our results thus cast doubt on the explanation of Japans "lost decade" that attributes the prolonged slump to the observed productivity slowdown.