Investment and the long swings of unemployment

Abstract

We estimate the medium-term relationship between investment and unemployment over the period 1960–2017 in 20 OECD countries, a period that includes the years of the Great Recession. While neoclassical growth theory typically assumes full employment—with no effect of investment on unemployment—and New-Keynesian models assume that such a relationship only exists in the short term at business cycle frequencies, we find that over our sample period covering more than 5 decades, a statistically significant negative relationship does exist: in decades when investment fell, unemployment increased. Panel estimation, using a measure of stock market volatility as an instrument for investment, confirms the negative relationship between investment and unemployment. High stock market volatility and low investment are associated with high unemployment, other things equal. The evidence supports both Keynes' ideas about an unemployment equilibrium as well as more recent investment-based theories of the natural rate of unemployment.