The robustness of Okun's law: Evidence from OECD countries

This paper evaluates the robustness of the Okun relationship based on postwar data for 16 OECD countries. While Okun's law is statistically valid for most countries, the quantitative as opposed to qualitative estimates are far from uniform. Estimates are also sensitive to the choice of models, including the first-difference and the "gap" specifications, the relevant data for the latter are constructed alternatively from the HP filter, the Beveridge-Nelson decomposition procedure, and the Kalman filter based on the NAIRU framework. There is mixed evidence of asymmetric behavior, but strong evidence of structural breaks occurring around the 1970s, after which time most countries began to experience a smaller output loss associated with higher unemployment.