This study proposes a structural change with threshold approach to re-evaluate the empirical validity of Okun's law using data from Canada. Based on the Hodrick–Prescott and band-pass filtered data, we find strong support of structural change as well as threshold nonlinearity. This suggests that the use of purely linear specifications for analyzing Okun's law may lead to misleading results. The implications of the empirical results for macroeconomic policy are also briefly discussed.
