Florence Goffette-Nagot (p), Zhejin Zhao , Loïc Bonneval

Rodrigo Perez Silva

This paper reexamines the conventional claims made by economists and policymaker concerning the effect of rent control. We pose questions about the impact of rent control on rents in the controlled and uncontrolled sector using panel data in Lyon over a 78-year period. Our study is a comprehensive empirical study of different rent control forms using multiple regressions with fixed effects as the main form of analysis. We find that the causal effect between rents and Lyon rent control is significantly negative.

Rodrigo Perez-Silva (p), Mark Partridge, and Alessandra Faggian

Florence Goffette-Nagot

In the housing market literature, the most common approach to measure the user cost is a weighted average of quality-adjusted prices. This procedure is often used to compute national average estimates, even though several spatial differentials and variations in housing rental and costs are not being captured. At the same time, there are other differentials that can arise from differing equilibrium processes in the rental and housing markets, which are both less understood and more problematic. In fact, while rental markets reflect current supply and demand processes, housing costs reflect a long-term expectation of the future price of the house: forward-looking markets capitalize long-term expectations for the growth of a city in the housing stock. Therefore, all else equal, housing prices will be higher in the most desirable or “nice” places. In the U.S., for example, the average annualized rent/housing-value ratio is .053. While Hawaii’s (a “nice” place) rent/housing-value ratio is .028, in Oklahoma (a place with fewer amenities) this ratio is .075. In both cases, applying the national average user costs would either lead to over- or under-estimations of local housing user costs. Similarly, hedonic estimates of place specific attributes such as environmental elements or amenities would be likely to be quite misleading. Using household level data, this study first computes a quality-adjusted measure of rents and housing values across the U.S., and secondly, attempts to empirically estimate the characteristics associated with differential quality-adjusted rent/housing-value ratios for several years -focusing attention on natural amenities- and their evolution over time. Consistent with our hypothesis, our results show that areas with the highest natural amenities (rank = 7) have ratios that are 42% smaller than low amenity places (rank = 1), explained by a 55% differential in values and a 13% differential in rents. Most of the differences are explained by a subset of natural amenities including hours of sun in January, temperature in July, topography, and water bodies. As expected, this relationship strengthen in non-crisis years, and almost disappears during the financial crisis of 2007-2009.