House Prices and Consumption

I study the consumption responses of heterogeneous households following changes in both house prices and interest rates. I show the common assumption that household period utility is separable in housing and consumption can be consistent with the observed co-movement between these two series only in the absence of housing transaction costs. When these costs are introduced into dynamic stochastic general equilibrium (DSGE) models characterized by separable preferences, consumption no longer increases after a rise in house prices. As it is well known, transaction costs are an important ingredient in house sales. I address this issue by developing a model that allows for non-separable preferences in housing and consumption alongside housing transaction costs. The results of my model closely match the aggregate data. Furthermore, it predicts that credit-constrained households will be substantially more responsive to changes in both house prices and interest rates than unconstrained households. Following a rise in house prices, consumption among constrained households increases by far more than the consumption of unconstrained households. Following a rise in interest rates, constrained households' consumption falls by more than that of unconstrained households. I trace this differing responsiveness in consumption to the house loan-to-value ratio of credit-constrained households. Higher loan-to-value ratios imply larger differences in their elasticity of response relative to unconstrained households. I also find that these differences widen with the degree of complementarity between housing and consumption. These predictions of my model are confirmed by household data from the Consumer Expenditure Surveys.