NAHB Study Finds Housing Options Particularly Limited for Retail Sales Workers
Recent reports on housing affordability have made it clear that a growing number of working families are finding it difficult to find housing that meets their needs. Among these working families who spend a disproportionate amount of their incomes on housing or endure long commutes in order to find housing they can afford, retail sales workers are particularly hard hit, according a recent study by NAHB.

NAHB’s study, completed in late 2004, found that households supported by retail jobs not only have insufficient incomes to buy homes in large metropolitan areas, but they are also infrequently able to rent homes without spending disproportionate shares of their incomes.

In order to demonstrate the rental housing plight of retail workers, NAHB looked  at the 25 largest metro areas in the United States, using data from the 2000 Census.¹ Individual household records were used to identify people in the top 25 metro areas who work in any one of the following Census-defined categories: cashiers, counter and rental clerks, parts salespersons, and retail salespersons. Households are then categorized as supported by a retail sales job if the primary earner (either the head of household or spouse, whoever earns more) works in one of these job categories. NAHB then used the median earnings of these retail sales primary earners to calculate, for each of the top 25 metros, a Rental Opportunity Index (ROI). (NAHB does a similar index tracking the affordability of for-sale homes called the NAHB-Wells Fargo Housing Opportunity Index.)
 
The ROI measures the share of rental units in a metro area that the median household supported by a retail sales job in that area can afford. We classify a unit as affordable to a particular household if the unit’s gross rent consumes no more than 30 percent of that household’s income.²  The 30% threshold is a standard often applied in government housing programs.³  A high ROI indicates that a large share of the rental units are affordable. A low ROI indicates that a small share of the rental units are affordable. In short, a high ROI is good; a low one is bad.

The ROI is then applied to  the census tracts, which are areas designed by the Census Bureau to capture a group of roughly 5,000 people who are relatively homogenous economically and socially. This allows us to identify each metro area as “affordable” or “unaffordable” to retail sales workers. We classify a tract as affordable if the ROI is above 50 (that is, over half of the rental units are affordable to the median income of a household supported by a job in retail sales). Otherwise, the tract is unaffordable. 

In the top 25 metros, there are almost 21,000 census tracts, yet only about 1,000 of these tracts can be classified as affordable to households supported by retail sales jobs. That means that of the 13.4 million rental units contained in those top 25 metros, only 0.6 million rental units are in tracts that are affordable to retail workers(Figure 1).

At 26%, Kansas City had the highest share of affordable tracts among the 25 largest individual metro areas that NAHB studied. Although that share represents only a quarter of the housing in Kansas City, it is a dramatically higher share than in other large cities—San Diego, for example, registered a 0.0% share. In fact, few of the metros approached Kansas City’s affordability share. Next to Kansas City, the largest shares of affordable tracts are found in St. Louis (21%), Dallas (19%), Minneapolis (16%), Atlanta (14%), and Chicago (8%). 

Moreover, the difference in ROIs between the unaffordable tracts and those classified as affordable is significant. In the affordable tracts, taken as a whole, the ROI is 63.1% (indicating that a little over 63% of the rental units are affordable). In the unaffordable tracts, the ROI is only 10.1%, so the unaffordable tracts are truly unaffordable (Table 1).

Census data contains relatively little detail on housing characteristics, but we can use what information there is to further characterize the affordable and unaffordable tracts. For instance, in the absence of other information, we can use age of the rental stock as a rough proxy for quality, under the assumption that substandard units are concentrated in the older properties.

The rental stock tends to be older in the few tracts classified as affordable. Nearly half—49%—of the rented housing stock in these tracts was built before 1960, compared to 41% in the affordable tracts. The spread tends to be much greater than that in the metros that have an appreciable number of affordable tracts (Kansas City, St. Louis, Dallas, Minneapolis, Atlanta, and Chicago). So for these metros in particular, the rental stock in the affordable tracts tends to be much older than in the unaffordable tracts (Table 2).

The affordable tracts in these metros also have comparatively high rental vacancy rates. Although not conclusive, the combination of high vacancy rates and low rents is a further indication that there may be a problem with housing quality in metros that have an appreciable share of affordable tracts.

 In short, NAHB’s study finds that America’s 25 largest metro areas contain relatively few neighborhoods where retail sales workers can afford to live. Whether we are willing to stipulate that these workers should rent rather than buy is irrelevant in this context, because the rents, as well as the house prices, are beyond their means. In the few neighborhoods where retail sales workers can afford to rent, the housing is older and more likely to be vacant, suggesting that there may be problems with its quality.
 
Taken together, these statistics contradict the perception that government housing programs are needed only to serve the unemployed. In many places, rents are too high for the people who are providing a valuable service to the community by working in retail sales. In these places at least, an increased supply of below-market rate rental housing is needed to serve that segment of the work force.

---------------------------------------

¹The definitions of metro areas used (Metropolitan Statistical Areas and Primary Metropolitan Statistical Areas) are the ones in effect at the time of the Census. Census data come in two forms—summary tables and files of individual household records—and we use both types of data to investigate rental affordability. For confidentiality reasons, much of the geographic detail is suppressed on the Census data that include detailed information on occupation. This limits the number of areas we can investigate.

² Gross rent includes contract rent plus payments for all utilities except the telephone. Most academic housing cost studies are based on gross rent, due to uncertainties generated by differences in the way properties sometimes do and do not include utilities in the contract rent. 

³ In the Low-Income Housing Tax Credit Program, for example, gross rents are limited to at most 30% of a target household income. [ return to top ]