Discussion

Greg Kaplan opened the discussion by inquiring about the goal of the estimation and the nature of the counterfactual exercises carried out in the paper. He expressed concern that the exercises that study the response of households to changes in prices are treating an equilibrium object—namely, house prices—as a parameter. AlisdairMcKay responded that individual households treat house prices as exogenous. Thus, it is appropriate to treat house prices as a parameter in a partial equilibrium setting and study how sensitive household decisions are to that parameter. As a following step, it is then possible to aggregate housing decisions. He agreed that a real-world counterpart of this exercise is not immediate because it would consider an average of different housing parameters across individuals. Kaplan agreed on the scope of the exercise in the context of the model, but he pointed out that we need to believe that the assumptions made in the model also hold in the real world. He also emphasized the importance of clearly stating the relevant real-world counterfactual at the core of the analysis that is being carried out. Jón Steinsson argued that it was important to distinguish clearly between two separate objects. The first is what is measured in the data, which is clearly defined by the empirical specification. The second is the interpretation of the measured object. A theoretical model can provide an interpretation for the object measured in the data. He highlighted that housing is not exogenous in the model presented in the paper, yet it corresponds to the object observed in the data. Adam Guren further pointed out that the home price shock is similar to a foreign demand shock and that the effect of such a shock on consumption is instrumental to correctly calibrating the housing wealth effect in a general equilibrium model.