When the latest data on state-level poverty rates was released last week (September 10), I was asked to comment for a story in the Arkansas Democrat-Gazette. I pointed out that poverty rates are not adjusted for differences in the cost of living.
My point was the comparisons across states and regions can be misleading. For example, Arkansas poverty rate in 2007 ranked it as the 4th highest in the nation, yet Arkansas also ranked as the 6th lowest in the nation for median gross rental rates for housing. When it comes to the affordability of the basic necessities of life, a uniform nationwide poverty threshold doesn’t adequately capture differences across states and regions.
This has long been a recognized shortcoming of the existing poverty rate measures. In 1995, studies from the General Accounting Office (GAO) and the National Academy of Science (NAS) independently reached the same conclusion: Cost of living differences matter, but a lack of data prevents the adjustment of poverty rates for regional cost disparities. Since the NAS study, researchers have been working on recommendations for improving the quality of poverty rate statistics.
A recent flurry of research activity at government statistical agencies has aimed to address this shortcoming. Just last month, Trudi Renwick—a statistician/economist at the Census Bureau—presented some recent findings on this topic at the Annual Meeting of the American Statistical Association. In addition to providing a thorough review of the literature on the topic, Renwick reported comparisons of three different methods for adjusting the official statistics. The first method is the Census Bureau’s experimental adjustment for Fair Market Rent (FMR) using data from the Department of Housing and Urban Development. The second method—from an April 2009 report by Census Bureau researcher Alemayehu Bishaw—uses median gross rental costs derived from the American Community Survey (ACS). The third adjustment technique— reported in November 2008 by Betina Aten and Roger D’Souza—represents collaborative research by the Bureau of Economic analysis and the Bureau of Labor Statistics to use hedonic regression methods for estimating regional price parity (RPP) differences incorporating a more comprehensive measure of cost-of-living.
The table below reports various measures of the poverty rate for Arkansas, comparing some of the alternatives to the current official standard. The first set of rates (from Bishaw, 2009) uses the official poverty definition, comparing the unadjusted measure with two of the techniques for housing cost differences. After accounting for relatively inexpensive rents in Arkansas, the poverty rate falls by two percentage points, and Arkansas moves down in the rankings from 4th highest to 8th highest poverty rate in the nation.
The second set of estimates (from Renwick, 2009) uses the NAS alternative definition for the poverty threshold (which incorporates different expenditure and income definitions from the official definition). The unadjusted alternative measure shows a slightly lower poverty rate for Arkansas, with a correspondingly lower state ranking. When adjustment for regional price disparities is included in the analysis, the results are dramatically different: For two of the three adjustment techniques, Arkansas does not even rank above the median for poverty rates. Indeed, the most comprehensive adjustment—the Regional Price Parity approach—Arkansas ranks as having one of the lowest poverty rates in the nation.
The cost of living matters when it comes to poverty rates. After accounting for differences in prices from state to state, the poverty rate in Arkansas is not nearly so high as the official statistics suggest.
– Michael Pakko
Alternative Poverty Rates and Rankings for Arkansas
Aten, Betina, and Roger D’Souza. November 2008. “Regional Price Parities: Comparing Price Level Differences Across Geographic Areas,” Survey of Current Business 88:11, pp. 64-74.
Bishaw, Alemayehu. April 2009. “Adjusting Poverty Thresholds Based on Differences in Housing Costs: Applications in the American Community Survey, “ poster presentation prepared for the Population Association of America Annual Conference.
Citro, Constance F., and Robert T. Michael (eds). 1995. Measuring Poverty: A New Approach. Washington, D.C.: National Academy Press. (National Academy of Science recommendations).
Renwick, Trudi. August 2009. “Alternative Geographic Adjustments of U.S. Poverty Thresholds: Impact on State Poverty Rates,” Paper presented at the Annual Meeting of the American Statistical Association Section on Social Statistics, Washington, DC.