SUBSTITUTION OF URBAN LAND FOR OTHER INPUTS

Existing estimates of the elasticity of substitution between land and nonland rely on data from the housing sector. These studies are replicated with a sample of high-rise office buildings; the replication supports the previous estimates of an elasticity in the range of 0.5 to 0.75. But there are some sources of downward bias in these estimates. The most serious is errors in the measurement of land value. Engineering approaches are used to correct for bias in a subsample of 59 observations. The engineering estimates support a unitary substitution elasticity. The importance of the Cobb-Douglas function for urban modeling suggests the need to use engineering approaches on larger samples and other types of property. © 1979.