Understanding Land Values: What Can We Learn from Teardowns and Vacant Lot Sales in New York City?

Abstract

This study aims to improve our understanding of the determinants of land values in large cities. We use two methods to recover land values. First, we rely on so-called teardown sales, which occur when a buyer purchases a property intending to demolish the existing structure and rebuild. As Rosenthal and Helsley (1994) and Dye and McMillen (2007) show, teardown sales provide a unique opportunity to measure land values in dense urban areas; the value of the land can be estimated as the purchase price of teardown properties plus the demolition costs. We are able to identify more than 4,000 teardown sales occurring in New York between 1994 and 2006 by matching a unique data set of all property sales to a full listing of demolition permits. For each teardown sale, we estimate the value of land as the purchase price plus the estimated demolition costs: the value of land per square foot is calculated as this amount divided by the lot area. The second method for recovering land values is from sales prices of vacant parcels. Although sales of vacant land are generally rare in dense cities, this is not the case in New York, where thousands of vacant lot sales have occurred in recent decades. We analyze more than 10,000 vacant land sales. One concern with these sales is that parcels that remain vacant for long periods of time in rapidly-appreciating cities like New York may not be representative of the typical lot, and therefore may not produce an accurate measure of land values. The analysis proceeds in two steps. First, we explore the extent to which vacant land sales produce land value estimates that are similar to the teardowns method, after controlling for important locational and neighborhood characteristics of the parcel. For the second phase of analysis, we use the variation in estimated land values across sites to explain the determinants of land values in New York City, running separate regression models for land value estimates derived from vacant land sales versus teardown sales. In particular, we focus on the following factors: proximity to the central business district; proximity to public amenities (subway stations, parks); neighborhood demographics; the quality of local public services (schools, crime); and zoning restrictions (maximum allowable development, location in a designated historic district). Findings from this study will aid researchers and local public officials in developing more sophisticated estimates of land values in dense urban areas.

JEL Classification: R3