Housing

Reuters Rural Housing Article Draws Wrong Conclusions

“Special Report: A rural housing program city slickers just love,” published by Reuters on March 18, 2013, relies on some questionable methodology and draws dubious conclusions. The article focuses on […]

“Special Report: A rural housing program city slickers just love,” published by Reuters on March 18, 2013, relies on some questionable methodology and draws dubious conclusions.

The article focuses on USDA’s mortgage guarantee program, one of many administered by USDA. The department also makes mortgage loans itself, and assists homeowners who cannot afford to repair serious housing problems, homebuyers who help build their own homes, and tenants who cannot afford to rent decent apartments.

It is important to note the differences among these programs. In 2012, the average income of homebuyers in the direct mortgage program was $27,634, while for those in the guarantee program it was $52,577. (The article uses a different figure for the guarantee program, presumably from an earlier year.) Also in 2012, the incomes of tenants in USDA’s direct rental housing loan program averaged less than $11,400 per year. Sixty-one percent of these tenants were elderly or disabled, many of them living on fixed incomes.

Incomes and home prices vary widely, and low-income people may live near high-income people. Eligible income levels for USDA’s housing programs are based on area median incomes because the cost of living varies from place to place. The article implies that loans for properties in places where high-income people live or work must have involved ineligible people or properties, but fails to recognize that low-income people may well live in the same communities.

For example, in Ewa Beach, Hawaii – which provides some of the examples used in the article – the median income is $69,299, but more than 500 households have incomes under $15,000 per year, and another 500 have incomes between $15,000 and $35,000.

It should not be surprising that low- or moderate-income people would live in Silicon Valley or Hyannisport, or near Google’s headquarters. The people who made these places famous may earn a lot, but the people who clean their homes and bus the tables at their restaurants do not.

There are many accepted definitions of rural areas. If USDA has guaranteed mortgages within the city limits of Honolulu or Washington, D.C., it has made errors and should be held accountable for them. The Reuters article goes farther, however, and criticizes the mortgage guarantee program for making loans in places that do not fit the Census Bureau’s definition of rural. The authors seem to have confused apples and oranges here: decennial Census data help determine program eligibility, but the data can be used with any definition. More than two dozen different definitions of rural are used by USDA programs, for example, and all rely on Census data for population numbers.

The Census Bureau defines “rural” places as those with populations of under 2,500. In other words, to the Census Bureau a town of 2,510 is not rural. The definition created by Congress for USDA’s rural housing programs includes places with up to 10,000 population within metropolitan areas and up to 20,000 outside metro areas, so long as they meet other criteria. Neither of these definitions is objectively better than the other. They are simply different.

It is also not true that, as the article states, “USDA’s geographic requirements for the loans were originally based on Census Bureau definitions of urban.” USDA’s definition of rural was created by Congress in 1965. Decennial census figures have always been used to indicate what places’ populations are, but that does not mean the Census Bureau’s definition of urban has been applied.

Finally, the article does not mention necessary disclaimers about a statement regarding “grandfathered” communities – places that were eligible for rural housing aid before the 1990 Census and remain eligible today, even if their populations have grown, so long as the population remains under 25,000. The article says those grandfathered communities “represent 40 percent of the program’s current loans.” The research brief from which this estimate is drawn notes that data about current loans are not available for the specific communities in question. The estimate was based on loans made in the counties where those communities are located, so it probably overstated the production for these specific communities.

HMDA data contain errors. The article’s data analysis relies on information reported by lenders under the Home Mortgage Disclosure Act. While HMDA data are the logical choice for an analysis of this sort, they are known to contain numerous errors. Reuters reports that HMDA records showed 90 of the 51,600 loans it reviewed, less than 0.002 percent of the total, exceeded $1 million. It is entirely possible that each of these 90 loan files contains an error in the loan amount, but it is not clear whether the article’s authors understood the data limitations or took them into account.

USDA did tell the reporters that HMDA data contain errors, and the article implies that because USDA did not provide specific examples there was no need to pursue the possibility of errors. USDA’s response to the article states, however, that Reuters provided a spreadsheet identifying questionable loans but not including enough information for USDA staff to analyze properly. Reuters could have taken some additional action, either providing more data or investigating further to determine where errors might have occurred.

Balanced reporting would include positives as well as negatives. The article acknowledges that, “even with the geographic anomalies Reuters found, there is little doubt that the majority of USDA-backed loans have helped expand home ownership among the people the program was designed to serve,” but includes no further information. An example is provided by the Washington Post in a March 21 story about sequestration’s impact on the Fort Peck Indian Reservation in Montana. Isabelle Youngman, who earns $11.95 an hour working for the reservation’s school system, used a USDA-guaranteed loan to purchase a $30,000 two-bedroom house for herself and her great-nephew, who has developmental delays.

Conclusion: Reuters may have uncovered some errors made in USDA’s mortgage guarantee program, although the reporters’ methodology throws doubt on that. Furthermore, by glossing over the successes of this and other USDA housing programs, Reuters does USDA – and, by extension, rural Americans – a disservice.

(Photo credit: Ben Adamson, CC BY-NC-SA.)

Related Articles