"GIS in Real Estate Valuation"

The following text is Gil Castle's final draft of the conference paper appearing in the GIS in Business '94 Conference Proceedings, June 1994

Copyright © 1994 GIS World, Inc.

Abstract

Placing a value on land and improvements underlies virtually every real estate activity. If one were to select just one facet of the real estate industry for supplying or using geographic software, hardware, data or services, property valuation would be a logical choice. Moreover, all valuation methods benefit from a GIS approach.

Properties have traditionally been valued using three methods: comparable sales, income-based, and replacement cost. In recent years a fourth method has been gaining acceptance: statistical extrapolation or, more generally, pattern recognition. For each of these four methods, GIS can be used to efficiently locate and analyze market data for inclusion in standard valuation models. For each, GIS can also be used to optimally present findings and conclusions in the final valuation document.

This paper examines the applicability of GIS to each of the four valuation methods. The emphasis is on "how to," and encompasses both computer tools and data sources. The overall objective is to underscore how GIS can significantly enhance the productivity and credibility of valuation professionals, to the benefit of those professionals and others who rely on their work.

_______________________

My real estate column in the November/December 1993 issue of Business Geographics provides an overview of using Geographic Information Systems (GIS) to value property. This paper focuses on the details, but with a caveat. As with many business applications of GIS, the potential benefits are extensive but are only beginning to be tapped by real estate organizations.


Comparable Sales Approach

This is arguably the valuation approach that can benefit the most from GIS technology. A property to be valued ("the subject") is usually similar to other properties that have recently been sold ("comparable sales" or simply "comps"). If we know the selling prices of the comps, and if we can adjust those selling prices upwards or downwards to reflect differences between the comps and our subject, in theory we can accurately estimate the value of our subject. The specific steps are as follows.

1. Display Candidate Comps

Using a GIS, first we display the location of the subject and its surrounding "submarket" (an area as small as a few city blocks or as large as many square miles, and defined by homogeneous real estate market conditions). We then display the locations of known comps within that submarket. The source of comps might be our own proprietary data base developed over time, local government records, or a vendor of comps information—some of which now offer on-line GIS displays (1). Finally, we might display other data layers for evaluating our comps, such as zoning, land use, traffic counts, public facilities (e.g., schools, transit lines), and special amenities (e.g., view sheds).

2. Select Relevant Comps

With GIS tools for creating buffers or simply measuring straight line distances, we can quickly determine which comps are within a reasonable proximity of our subject; usually, the closer the proximity, the more relevant the comp. With GIS tools for displaying attributes, we can "click on" different comps to investigate property type, age, condition, size, selling price, and other comparison variables. With GIS query tools, we can quickly locate all comps fulfilling certain criteria, such as "all office buildings greater than 100,000 square feet that have sold within the last year, were built after 1980, and have interior parking." With GIS tools for thematic mapping, we can differentiate the comps by assigning various display colors, symbols and sizes, on the basis of each comp's age, size, selling price, etc. In short, we can efficiently learn a considerable amount about all the comps on the GIS display screen and then select the comps most like our subject—much more quickly and accurately than by laboriously thumbing through textual reports in conjunction with hard copy maps.

3. Develop a Comps Adjustment Matrix

The subject and comps comprise the columns of this matrix; property characteristics (property type, age, size, and so on) constitute the rows. In a judgmental process that is as much art as science, we compare the property characteristics of the subject to those of the first comp. To the extent that the subject is superior to the comp for a given characteristic (say, more parking spaces, higher quality construction, larger size) we adjust the comp's selling price upward; if the comp is superior to the subject, we adjust the selling price downward. We also make an adjustment based on how long ago the transaction occurred. We then repeat the process for all the other comps, comparing each in turn to the subject, concluding with an adjusted sales price for each. The final step is to reconcile all the adjusted selling prices of the comps into a single estimated market value of our subject.

A comps adjustment matrix can easily be created with a standard spreadsheet program, data base management system, or even sophisticated word processing package. Depending on our hardware and software, we might be able to run our GIS in one window on the computer screen and the comps adjustment matrix in another window, and move back and forth between the two windows when making judgments on the magnitude of each adjustment. By doing so we can check the GIS for supplemental, ad hoc information not contained in the adjustment matrix, such as proximity to the subject. Similarly, if we decide we want to add or delete comps, we can use the GIS to select a new or modified sample in the same manner as under Step 2 above.

Ideally, of course, the comps adjustment matrix will be directly linked to the GIS. We can then automatically create the comps adjustment matrix by simply clicking on the appropriate comps on the map and by selecting property characteristics from each comp's list of attributes. As of this writing, only one GIS on the market offers this capability (2).

4. Prepare the Final Report

Valuations can be reported in a short letter, a lengthy report, or something in between. Almost always, though, a critical component is a map locating the subject, comps, and reference features such as major roads. If a GIS has already been used to identify and evaluate the comps, then the map can be quickly enhanced with a title block, scale bar, etc., and then printed or plotted. Accordingly, we can use a GIS not only to increase productivity and quality when analyzing a subject property, but also when presenting the results.


Other Three Valuation Approaches

Income-Based

This approach ties the value of the property directly to the pre-tax income that the land and improvements will generate. The approach relies on annual forecasts of rents, vacancy rates, operating expenses, lease turnovers, etc., for one year, five years, or sometimes more.

A GIS can be used in a manner similar to the comparable sales approach, that is, to identify properties similar to the subject. Financial information on those other properties (again, rents, vacancy rates, etc.) can then be averaged and inserted into a pro forma for the subject. The pro forma, in turn, permits the calculation of an estimated value using several standard formulas. This approach is basically saying "The subject will be worth $X if the subject experiences the same rents, vacancy rates, etc., as the [market] average of the buildings selected with the GIS."

As with the comparables adjustment matrix, the formulas for calculating a property value typically reside in a spreadsheet program, data base management system, or other specialized application package. These programs can be run sequentially with a GIS, or possibly in a separate window from the GIS.

As always, the GIS is highly dependent on the quantity and quality of available financial data for the other properties. Such information is not as readily available as the sales data for the comparable sales approach. Brokers, property managers, appraisers, and other real estate professionals frequently treat such financial data as highly proprietary.

However, professionals with access to such information, whether from internal or external sources, can use GIS to extend that competitive edge. They can do so in the same manner as the comparable sales approach, i.e., not only to increase productivity and quality when analyzing a subject property, but also when presenting the results.

Replacement Cost Approach

This approach assumes that no one will pay more for a property than the cost of the land plus the cost of replacing existing buildings with new structures. In theory a GIS could be used in a manner similar to the two previous approaches: identify properties comparable to the subject, determine the cost of construction for those comparable buildings, and then use the same figures to estimate the construction cost of new structures on the subject's site. In practice, though, construction costs do not vary dramatically from one location to another, and thus the usefulness of a GIS is diminished. A possible exception occurs when an organization is trying to determine the least amount to pay for a particular type of building, can locate in any of several cities, and the construction costs in those cities vary significantly.

Pattern Recognition Approaches

Attention is beginning to focus on linking GIS to statistical valuation techniques, such as Computer Aided Mass Appraisal (CAMA). Given a sufficiently large and homogeneous set of property transactions, regression coefficients can be calculated relating past selling price (dependent variable) to numerous property characteristics such as size, age, construction quality, and so on (independent variables). These coefficients can then be used for estimating the values of other, similar properties. Artificial intelligence (AI) offers additional pattern recognition tools, including case-based reasoning and neural networks, which embody significant promise in automated property valuation.

The ways that GIS can be beneficial include the selection of a credible sample of transactions, the display of predicted results for visual identification of possible anomalies, and (as always) the final presentation of findings.


Conclusion

Real estate may account for as much as one-fourth of the nation's assets; consequently, estimating the value of properties is a frequent and important activity. Among the traditional and new valuation methods, the comparable sales approach is the most widely used, and the one that currently benefits the most from GIS tools. For all four approaches, though, GIS can significantly improve the productivity, credibility, and profitability of real estate valuation.

_______________________

References

1. One such firm is DataQuick in San Diego, California (619/455-6900).
2. The GIS package is "Market Monitor," designed by and available from Castle Consulting in San Francisco, California (415/955-0531).