Evidence for Sticky Consumption Growth in Canada

Abstract

Consumption stickiness occurs when current consumption can be predicted by past consumption. The presence of consumption stickiness has important implications on our understanding of consumption dynamics and how they respond to shocks in income. I use the instrument-based framework by Carroll, Slacalek, and Sommer (2011) for the case of Canada over the period 1980Q1 to 2020Q1. I show strong evidence for the presence of weak instruments in the data and therefore use weak-instrument robust inference techniques: the conditional likelihood ratio test (Moreira, 2003) and the Anderson Rubin test (Anderson & Rubin, 1949). I find strong evidence of consumption stickiness in the case of Canada for all examined subsamples.