The Mortgage Crisis and GIS

Was the FGDC Mortgage Crisis Meeting a Silver Lining on a Huge Cloud?

By Art Kalinski, GISP

On May 7, I attended a special meeting addressing the mortgage crisis hosted by the Federal Geographic Data Committee (FGDC) and the International Association of Assessing Officers (IAAO). The purpose of the day long meeting was to discuss the use of land parcel data in managing mortgage related issues. The meeting focused on how GIS can monitor and manage these issues and to get feedback from leaders on which specific data elements are important to the mortgage industry and its financial oversight. The meeting was held at the American Institute of Architects Building in Washington, D.C.

The participants included representatives from banking, finance, credit, and mortgage firms as well as members of federal regulatory agencies. Also present were GIS professionals, private sector solution providers, and land parcel data producers. The meeting was opened by Dr. David Moyer, the meeting facilitator; Mike Howell, OMB and vice chair of the FGDC Steering Committee; Bob Ader, co-chair of the FGDC Cadastral Committee; and Dr. Josephine Lim, president of the IAAO.

Key presenters in the morning sessions were Anne Hale Miglarese of Booz Allen Hamilton, the chair of the National Geospatial Advisory Committee (NGAC); Dr. Dave Cowen, a leading GIS educator; and Dr. Nancy von Meyer who has worked on cadastral issues for over 20 years, including her current work for the U.S. Bureau of Land Management (BLM).

According to the presenters, some analysts and GIS professionals saw the mortgage crisis coming but there was also a lot of “whistling past the grave yard,” as people hoped that the trend would not turn into a crisis. The rise in distressed mortgages, foreclosures, and decreasing real estate values were visible long before the crisis became a crisis. Key to monitoring the issues was good parcel level data. Unfortunately, even though we have high tech tools such as GIS, there are institutional barriers that limited the effectiveness of those tools. There was a lack of consistency and interoperability of key datasets at a national level.

I personally experienced two such frustrations related to the post office and to the census. When I first joined the Atlanta Regional Commission (ARC) as the GIS Manager in 1993, I learned that ARC published zip code maps as a service to the regions’ economic development offices. My response was “why don’t they get them from the post office?” After all, during my graduate work at the University of North Carolina we constantly used zip code marketing data for site selection and trade area analysis. I assumed that the zip code polygons came from the source, the U.S. Postal Service. I was wrong.

In the pre-GIS age, when zip codes were created, the purpose was to build delivery routes not defined polygons. As a result, the zip code boundary files we use are very fuzzy and the polygons only approximate boundaries. I was so incredulous, that I personally drove up one major road in Cobb County to conduct my own survey. What I saw were businesses in one long block on the same side of the street that had alternating zip codes as one progressed up the street. I was told that this happened because the business where given a choice of having a Marietta city address or a Kennesaw city address. This, of course, is one example of one street in one county, but my understanding is that this is a system wide issue. Even today the Postal Service doesn’t publish zip code maps but instead refers users to commercial vendors who create zip code maps. Of course these commercial maps come with disclaimers regarding the validity and positional accuracy of the data.

The second frustration was helping with the U.S. Census 2000 Local Update of Street Addresses (LUCA). Under this census program, local talent was used to update TIGER street data and addresses. The key limitation was that we all had to sign confidentiality agreements that any data collected and any data sets that were enhanced with this collection effort would also be confidential. That meant that if I sent staff to a new neighborhood and identified new streets and news addresses, not even houses, just address locations, that information was confidential and if I used the information to update our massive ARC street centerline database that file also became confidential.

I could then send the same people out on a different day to the same location to collect the same data but if the data collection was not in the name of LUCA, it was not confidential and we could use it. Yes, I fully appreciate privacy but the mere existence of a street address tells you nothing about the parcel or the building, if there is one, or even the inhabitants, if there are any. In fact, most of this data in even greater detail is already available from commercial sources. These are just two examples that hurt data collection and interoperability, but there are more when it comes to parcel level data. Now I’m about as conservative as one gets, but this silliness in the name of privacy defies common sense and is an example of how we twist ourselves into knots to our own detriment.

The afternoon sessions were facilitated by David Stage of the FGDC Cadastral Subcommittee, Susan Marlow of Smart Data Strategies, and Roger Clark of IAAO. The meeting focused on obtaining feedback from users in the mortgage industry as to specific data they need to do their job effectively. The list of data elements included factors that one would expect. There was a large list related to the property descriptions, such as lot and building size, construction — several attendees highlighted the need for street level or oblique imagery of each property, and annotation of the condition of improvements. George Donattelo of IAAO indicated that these were especially needed if IAAO AVM and Mass Appraisal Standards are used. Financial factors included the type of mortgage, equity, and terms of the mortgage. Geospatial factors such as nearness to water, flood plains, neighborhood comparable appraisals, demographics, and regional economic factors such as employment levels added even more complexity to the analysis.

The discussions all pointed to the need for timely analysis using comprehensive data. We know that spatial analysis of all factors, perhaps with regression analysis, could help with future forecasting, but analysts need data that is detailed, accurate, and most importantly, consistent. There was a general consensus that just a small portion of the rescue funds should be devoted to creation of comprehensive data collection efforts. All agreed that this should be the message from this meeting to our leaders.

I’m hoping that this economic crisis will open everyone’s eyes to the need for consistent and comprehensive data if government is going to insert itself into the market. Homer Simpson one stated: “Beer, the cause of and solution to life’s problems.” … Dare I substitute “government” for “beer?” Maybe with superb data and competent analysis we can look back on this crisis with open eyes and not repeat our mistakes. Perhaps the realization that we need good data and GIS analysis will be the silver lining.

For additional information, please contact Don Buhler, Department of Interior, Bureau of Land Management, don.buhler@blm.gov, (202) 452-7781. The meeting details are posted on the Cadastral Subcommittee’s website at www.nationalcad.org.