Is Monetary Policy More Powerful in Recessions or Expansions? Evidence from Canada

Abstract

Using the local projection strategy of Jordà (2005) along the smooth transition regression of Granger and Teräsvirta (1993), this paper investigates whether the impact of monetary policy varies according to the state of the business cycle. To estimate the model, the paper uses a novel Canadian monetary policy shock series identified by Champagne and Sekkel (2017) and a “state variable” that captures the state of the business cycle. The paper finds that monetary policy tends to be more powerful in expansions than in recessions, particularly due to a higher sensitivity of business and housing investment and durable consumption to changes in the policy rate. While prices appear to be fairly “sticky” in expansions following a contractionary shock, the opposite is true in recessions when prices respond more vividly. The differences observed across expansionary and contractionary regimes emerge whether the “state variable” used in the empirical estimation captures the growth rate of economic activity (i.e. growth rate of GDP) or the level of resource utilization (i.e. output gap).