Measuring Local Individual Housing Returns from a Large Transaction Database

Housing returns are an important element in households' housing tenure choices (buy/rent) and investors' portfolio compositions. Usually in the absence of individual data, practitioners resort to the ratio of the average of rents to transaction prices over large areas to proxy these returns, but this measure may not capture the different responses to shocks of each variable varying with the heterogenous characteristics of the dwellings, and it does not take into account the vacancy risk for the investors. Working with a panel of 40,000 rented flats or houses surveyed on a yearly basis by the OLAP and with the database of the registration by notaries of all the housing transactions between 1996 and 2007 in the greater Paris area, we develop econometric models of prices, rents, occupation, and vacancy spells, and develop statistical procedures to produce local measures of rental returns, capital gains, and the associated risks depending on physical characteristics. This allows us to illustrate a significant and persistent heterogeneity in rental returns and the impact of vacancy depending on the type of dwellings. Areas with initial high rental returns seem to experience higher capital gains which negatively impact the rental returns afterwards. Finally, the variance of returns varies with location and these differences are persistent. All these features should be taken into account when investing in housing.