Abstract
In this paper, we study how heterogeneity across households affects the refinancing channel of monetary policy. We show that the correlation between two key parameters---the Likelihood of Refinancing a Mortgage (LRM) and the Marginal Propensity to Consume (MPC)--determines the strength of the refinancing channel. Using novel survey data, we uncover a small positive correlation, implying a more effective channel than previously accounted for. This positive correlation masks two opposing effects: financial (un)sophistication generates a negative relationship, while liquidity preferences generate a larger, positive relationship. These two mechanisms affect time-dependent and state-dependent refinancing inaction, respectively, affecting the size and speed of monetary policy pass-through.