Gis System

Abstract

The author was recently asked to make a presentation on the use of geographic information systems, commonly referred to as GIS, to a local chapter of the Appraisal Institute. The purpose was to show how an appraiser or appraisal reviewer could use GIS to find cases of white collar real estate scam. The items discussed at that presentation are covered in further detail in this paper.

Appraisers need to show the reasoning behind their value opinions by discussing important spatial relationships and their likely effect on value. Geographic information systems (GIS) can be used to analyze these relationships and to show why a client should select an appraiser who has this level of information.

Gilbert Castle has noted that real estate is essentially a game of information arbitrage. The likely winner of the game is the person that takes advantage of computerized analyses. Castle explains that GIS is an attention-getting way of showing what you know.(n1) Of course, larger data sets are used for GIS analysis, not just the minimum "three comps."

The visual aids that GIS can generate could also be very useful in litigation, to help explain complex issues to a jury that is relatively unfamiliar with real estate valuation. Clear communication of complex technical issues is the basis of forensic consulting, an emerging field that is expected to grow more rapidly in the future. The need for forensic consulting has been created by rapid changes in technology.

The Arden-Guthrie Problem

Arden-Guthrie is a neighborhood in San Bernardino, California. A number of fraudulent transactions in that neighborhood inflated the ostensible value of local quadruple properties. The question is, How could a reviewer have used GIS to find the problems caused by the fraudulent sales?

Many of the properties in question are located within the block group outlined in red in Figure 1. Other problem properties are located in a block group just south of the outlined area. The larger red area at the top of the map is part of a color-coding system that shows median rents by census block groups. As we can see, renters in this area one-half mile to the north were paying from $913 to $1,001 per month at the time of the 1990 census. This represents the highest rent category for San Bernardino County.

Rental data from the 2000 census will be available soon. A reviewer could print out such a map and use it to check quickly for inconsistencies. One obvious inconsistency would be an appraisal that concludes that rents in the highest bracket are indicated for a property that is located in a low-rent area. Census data is relatively inexpensive. Data for the entire country was available for less than $500 more than five years ago.

The appraiser's first step is to define the proper market area to search. Once this is done, sales data can be downloaded from a time-sharing data service. The data retrieved can be imported into a GIS system and then mapped to show the range in prices.

One common method is to search within a one-mile radius of the subject property, but an appraiser using this method would retrieve data from the high-income area. Another method is to search by zip codes, in this case 92404. The heavy blue line on the map in Figure 2 shows the resultant boundary. As we can see, searching by zip code also results in too broad of a search. Searching by block groups would provide much more detail, but this type of search is not available using data from typical real estate data services.

Searching by census tract provides the best results. Census tracts are indicated by the heavier black boundaries. Zip code and block group boundary files typically have to be purchased separately. Updated, high-resolution zip code boundary files for California cost approximately $500.

It costs only a few dollars to export the data, but you have to pay an annual fee to use the system. Basic GIS mapping systems for PCs typically cost more than $1,000, but newer, more limited products cost much less. You can display the data with the less expensive GIS programs, but you won't be able to do the analyses shown later in this article. A tremendous amount of other data can also be mapped. Some is available for free via the Internet; other data sets can cost thousands of dollars per year for the database and a similar amount for each of the annual updates.

The data has to be geocoded after it has been downloaded. This is the process of adding latitude and longitude coordinates to each of the records so that the data can be displayed on electronic maps. This is the most time-consuming and expensive part of the analysis. Geocoding the addresses yourself requires street files, and street files can range in price from $500 to $10,000. Only the most expensive street files have the most recent streets, so finding properties in newer subdivisions has been problematic. This situation may change in the near future.

Data services have latitudes and longitudes for each property, but many will not sell them. This may be because they want to entice you to pay an extra $250 per county each year to create simplistic maps with their mapping system. It will typically cost an extra 10% of your subscription price to obtain lat/ long coordinates that are available for other property types from other data services.

Once the data has been imported into the GIS, geocoded, and mapped, the inconsistencies are readily apparent. The ranges in sale prices are labeled in the legend in the upper right corner of the map in Figure 3. The subject neighborhood looks like a rainbow, with the properties shown in red selling at the highest prices, other properties selling for half as much, and still others selling for only one-fifth of the highest-priced ones. How often does an appraiser find such a range in prices? Within one block, some properties are selling at prices 25 times higher than other properties.

Another effective method of conveying the diversity in prices is by graphing. Most advanced GIS packages allow the analyst to graph the data that has been mapped. Quadruplexes in the area come in four sizes, as shown on the horizontal axis in Figure 4. As the graph shows, there is a phenomenal range of prices for units of the same size.

The typical relationship is for increases in sale prices to taper off as the size of the home increases. Figure 5 shows that this neighborhood is not typical. This graph compares the prices of resale housing (on the left) to the sale prices of new housing (on the right) located in the same block group.