The housing wealth effect: Quasi-experimental evidence

Empirical studies have estimated a large range of consumption responses to changes in housing prices. Using a quasi-experiment, we estimate a shock of − 19 . 4 percent to single-family house prices in the area surrounding an airport in Stockholm after its operations were unexpectedly continued as a result of political bargaining behind closed doors. Using a household data set with information on the locations of primary residences relative to the airport, we find a short-run elasticity with respect to new car purchases of 0.28, corresponding to a one-year marginal propensity for expenditure on cars (car MPX) of 0.09 cents per dollar lost in housing wealth. Households with high loan-to-value ratios and small bank deposits respond the most. A quantitative model aligns with our empirical findings but also suggests that the car MPX could be 0.31 cents when used cars are included; of this, 73 percent is explained by a collateral channel. When nondurables are accounted for, the total marginal propensity to spend is 2.1 cents. In the case of an absolute fall in housing prices, the total response is four times greater.