Abstract
In 2015, extensive marketing of state lands began as part of government programs for affordable housing. We use data from the Israel Land Authority to examine the impact of these marketing efforts on bid prices in regular land tenders for building apartments for the open market and analyze the demand factors for land for high-density residential construction in Israel. To account for the heterogeneity of the marketed lands, we apply Koenker's quantile regression adapted to a panel with fixed effects for the tender, controlled by the lasso mechanism. The lack of bids in some tenders constitutes a source of selection bias in the parameters, similar to the phenomenon described by Heckman in the labor market. To correct this selection bias, we apply the algorithm of Arrelano & Bonhomme (2017), which is based on a copula. After correcting for the selection that neutralizes the decline in the attractiveness of regular tenders in the environment of planned large government-subsidized projects, we find that bid prices in successful regular tenders increased as the marketing of lands in regular tenders decreased. Additionally, we find that proximity to localities where land is marketed for government-subsidized projects raises the bid amounts submitted in regular tenders in localities where no land was marketed within the government programs. The intensity of these effects is stronger in the periphery.
JEL Codes: C13, C14, C21, C23, D44, R30, R31, R38, R52.
Key words: residential land auctions, affordable housing, periphery, quantile regression, selection bias.
For a PDF of the full article