Discussion: (0 comments)
There are no comments available.
A public policy blog from AEI
The latest example of poverty measurement problems is on display in a new report by the Institute for Research on Poverty (IRP) at the University of Wisconsin-Madison. Similar to the supplemental poverty measure produced by the US Census Bureau, the Wisconsin measure raises confusion about who meets the definition of poor and how government programs serve them. Highlighting these problems is important because the supplemental measure is often presented as the best way to assess poverty.
Contrary to almost all other measures of the Wisconsin economy, the report concluded that poverty increased in Wisconsin from 2015 to 2016. The US Census Bureau’s official and supplemental poverty rates for Wisconsin also contradicted what was in the report. This illustrates well how problems with the relative threshold and the data used to calculate the supplemental poverty measure lead to inconsistent and difficult to interpret poverty trends.
The threshold used for the supplemental poverty measure reflects a percentage of how much everyone spends on food, clothing, housing, and utilities, averaged over a five year period. Survey data is used to estimate these expenditures at the national level, and in the case of Wisconsin, researchers adjust it to reflect cost of living differences in the state. It is considered a relative threshold because it is adjusted based on how much people consume relative to everyone else — if people consume more the threshold goes up, if they consume less it goes down. This means that in a booming economy poverty reduction is harder (because people consume more and the threshold goes up), while it’s easier when the economy is struggling. Neither of which makes sense.
Researchers Bruce Meyer of the University of Chicago (and AEI) and James Sullivan of Notre Dame agree, describing the supplemental poverty measure as “a complex and convoluted way of determining changes in poverty over time that we argue makes it difficult to interpret.” An analysis of New York City’s supplemental poverty measure illustrated the policy implications, finding that an inflation-adjusted threshold resulted in poverty declines nearing 5 percentage points from 2005–2014, with the relative threshold resulting in no change.
Sign up for The Ledger newsletter
Adding to this problem are the data used in calculating the supplemental measure. At the state and local level, survey data can be problematic (either because of small sample sizes or missing/underreported data), and researchers must make adjustments. The Wisconsin Poverty Measure uses the US Census Bureau’s American Community Survey, but survey respondents don’t report how much they get in food benefits. Because this is a major source of income for low-income households, researchers must use other data sources to statistically predict how much respondents receive and then add it to the survey.
Similar methods must be used to estimate child care expenses, housing costs, medical expenses, and transportation costs. Self-reported data (like earnings) is combined with imputed data (like food and housing benefits) to determine who is poor. It’s impossible to identify the degree of imprecision introduced by these methods, let alone how they combine to affect the overall poverty rate.
All of this makes it difficult, if not impossible, to use the supplemental poverty measure to assess policy, especially when trends conflict with other known data. The authors of the Wisconsin report claimed that their results “underscore the importance of the safety net that is now doing less in Wisconsin than a few years ago to enhance low earnings for families with children, put food on the table, and encourage self-reliance.” Yet, it’s nearly impossible to say what the safety net has or has not done considering the shortcomings of their measure.
Ultimately, the supplemental poverty measure was introduced as a superior measure to the official poverty measure. Yet, far from making policy assessment easier, it presents obstacles of its own and brings us no closer to certainty about the effects of the social safety net than the flawed official measure.
There are no comments available.
1789 Massachusetts Avenue, NW, Washington, DC 20036
© 2018 American Enterprise Institute