Productivity, Model Uncertainty, and the New Home Sales Price in the U.S.
86 Pages Posted: 27 Apr 2020
Date Written: 2017
In this paper, I study the recent swing in the new home sales price in a dynamic stochastic general equilibrium model that features sector-specific total factor productivity (TFP) shocks, agents who concern about model uncertainty (in the Knightian sense), indivisible labor in households’ preferences, and adjustment costs of factor inputs. My paper dedicates to answer the question that whether a one-time temporary change in the model uncertainty level, along with the productivity shocks, can account for a significant amount of the boom and bust observed in the recent new home sales price in U.S. I construct the TFP shocks for the construction and non-construction sectors separately using aggregate macroeconomic data. The benchmark model, with only TFP shocks and no model uncertainty, can account for 25 percent of the surge and more than 60 percent of the slump in the new home sales price. Model uncertainty is specified in the Knightian sense. Agents who fear about model misspecification are not fully confident that the model precisely defines the true data generating process, thus they form expectations in an extremely pessimistic way and guard themselves against the worst-case scenario in decision-making. Quantitative analysis shows that with a one-time temporary change in the model uncertainty level the benchmark model is able to capture 35 percent of the increase in the new home sales price. I simulate the distributions of TFP shocks under the worst-case model. The results show that the worst-case state transition law, compared to the transition law under the approximating model, acts more like a negative mean shift rather than a variance dispersion. Besides, a rise in model uncertainty does not increase the correlation between TFP shocks under the worst-case model.
Keywords: New home sales price, Model uncertainty, Sector-specific productivity, Indivisible labor, Adjustment costs
JEL Classification: C61, D81, G11, G12, R21, R31
Suggested Citation: Suggested Citation