housing_markets_in_a_pandemic.pdf
(390.39 KB, PDF)
… and regular sales. For our analysis, we use the repeat- sales pairs identified in Korevaar (2020). For rents, we use the existing index of Eichholtz et al. (2019). For Paris, we use data from the sommier foncier. This government register served to check the veracity of property tax and estate tax payments and contains information on the universe of sale prices in Paris between 1809–1943, as well as data on the rent prices of these properties. Rent prices were either obtained from new rental … related to the plague. The dummy variable Plague is equal to 1 when in the six months prior to the transaction date, a plague epidemic has started. The 6 and 12 months lagged variables are denoted by Plague.L6M and Plague.L12M.9 As a robustness check, we use a hedonic price model (Rosen, 1974), given by lnPi,t = α + µt + x′i,tβ + εi,t, (3) 9Conditional on the variances (σ2 ε , σ 2 δ ) the time-weighted repeat sales model (2) can be estimated by generalized least squares. The variance … . As we do not have detailed property information, we include street fixed effects (1521 streets). We have information on transaction type (four different types) and some crude descriptions of the property, like the presence of a building, a garden, a shop, etcetera. In total, we have 25 property related dummy variables. We use identical variables for the plague as in the repeat sales model. Table 2 presents estimation results from the repeat sales and hedonic price model. Our sample covers seven …
