Residential land prices prior to development

Journal of Real Estate Research, The, 1997 by Karl L Guntermann

Subdivision activity in an area is an obvious sign that the development of any specific parcel in the area is a possibility. However, it is unknown whether the level of development activity per se affects the value of surrounding undeveloped land. A second test relating to development expectations uses the level of residential activity at various distances from each parcel as a basis for forming expectations. It is hypothesized that expectations about development are formed based on the level of subdivision activity at certain, critical distances around each parcel. The expectations hypothesis will be tested for residential activity (number of active subdivisions) located from one to three miles around each parcel.

Model and Data

Based upon this discussion and previous research, the various factors that are hypothesized to explain variations in the price per acre of vacant land can be organized into the following general categories:

1) Parcel characteristics or attributes;

2) Local expected growth rates;

3) Micro-development expectations;

4) Other (macro) economic factors.

Hedonic models are estimated to test hypotheses related to categories two and three. Standard parcel characteristic variables are included to ensure adequate specification of the models. Parcel characteristics are measured using variables for size, and accessibility (distance to development and distance to the freeway) and dummy variables for the availability of off-site infrastructure and two residential zoning categories (high and low density). Variables are included to test for other economic factors or events, such as changes in tax laws or interest rates, that might also affect land values. The dependent variable in all cases is real price per acre and each model is estimated in both a semi-log and log-linear form. A more complete description of all variables is presented in Exhibit 1.

Population, net in-migration and employment series for the Phoenix metropolitan area are readily available to test the Capozza-Helsley growth hypothesis.4 Both total wage and salary employment and a construction employment series are tested in alternative models. While wage and salary employment would be a broad-based measure of growth, construction employment is a narrower measure of growth that may correspond more closely to activity in the land market. Several financial series on funds availability, interest rates and mortgage commitments for commercial and residential loans, etc. are tested as macro-type variables. Dummy variables for important tax law changes during the period of the land transactions data (1983-91) also are tested. All variables are tested in both a concurrent and lagged form with respect to the dependent variable, real price per acre.

The population and employment forecast data used to test one of the micro development expectations hypotheses are from reports prepared by the Maricopa Association of Government (MAG), which is a quasi-official planning organization in Maricopa County (Phoenix MSA).5 Each report contains forecasts for five-year intervals (1985, 1990, etc.) for municipal planning areas (MPAs), which typically cover nine square miles. Individual land transactions are matched to the closest forecast for their MPA from the report currently available as of the transaction year.6 Data on subdivision activity are from the Greater Phoenix Housing Study (GPHS), which tracks subdivision activity and publishes data quarterly on permits, starts and closes for each subdivision. Aerial photographs published in connection with the GPHS data were used to locate and identify active subdivisions within a three-mile radius of each parcel. Separate variables were created for each mile based on the count of subdivision within that distance.