Journalists and commentators parsing the U.S. Census Bureau’s new Supplemental Poverty Measure (SPM) when it debuted about a month ago missed one surprising result of what happens when the new measure is applied. As most news stories correctly pointed out, the SPM counts as income certain public benefits that the official measure didn’t. Chief among them are the Earned Income Tax Credit (EITC), food stamps (now Supplemental Nutrition Assistance Program-SNAP), and low-income housing assistance.  Under the official measure, the roughly $170 billion spent on these programs was totally under the radar, even though these three benefits amount to about $3,700 per low-income person. (That’s over $15,000 annually for a family of four).   Under the SPM, these and some other benefits are counted as income, though not fully because people report less in benefits than the government has paid out.

Several headlines highlighted the higher estimated poverty rate yielded by the SPM than by the official measure.  The Washington Post’s Michael Fletcher, for instance, claims that the new Census measure “…painted a more dismal picture of the nation’s economic landscape than the official measure from September.”

So how can poverty go up if the new measure raises collective incomes by over $170 billion?  As some journalists and experts noted, children fared better when public benefits are counted. But why should extra spending on kids result in higher poverty among the elderly and other groups?

Let me oversimplify a bit to explain.  The official measure was set up as an absolute measure—the income needed to achieve a specific unchanging living standard. The threshold set was three times the cost of an economy food budget. Over time, that threshold has risen only to keep pace with inflation, not with rising living standards.  In contrast, the SPM threshold is a relative concept.  It equals what a family with two children at the 33^rd percentile of spending devote to food, clothing, housing, and utilities plus another 20% of this amount.  This threshold is then adjusted for family size and local housing costs.

The percentile used to calculate the SPM threshold is somewhat arbitrary.  Choose a relatively high percentile (say 50%) and you get a high poverty threshold and higher level of measured poverty. Choose a lower one (say, 25%) and both drop.  By selecting 32-34%, the Census Bureau raised the income threshold so now it’s about 10% higher than under the official measure. That statistical move doesn’t mean that the poor’s living standards have dropped, so the SPM doesn’t really “…paint a more dismal picture ” so much as it creates a new benchmark based on a higher standard of living.

A second conceptual shift is that the SPM deducts from income what the Census Bureau terms “necessary expenses.”  These include taxes, work expenses, and the amount of child support individuals pay, all of which lower net income.  Also deducted are spending on child care and out-of-pocket health expenses.  Child care is usually a work expense, but people still have discretion over what quality they buy. Health spending is clearly consumption and differs from income or expenses necessary to generate income.  Health services are valuable—sensible uses of income—and more spending presumably raises an individual’s living standards.

Like the official measure, the SPM doesn’t count government-paid health services, even though they can greatly enhance living standards and life itself.  So, yes, older Americans spend more out-of-pocket on healthcare, which pushes up their poverty rate.  But, the presumed improvements in living standards financed by significant Medicare and Medicaid benefits go uncounted.  This approach raises the same concern that led to the SPM—the distorted picture you get when you don’t count government benefits aimed at alleviating hardship.

Certainly, health spending poses a quandary for counting poverty.  For an individual, paying more health expenses may reflect poorer health.  When unhealthy people must spend more of their own money to achieve the same health status as healthier individuals, they have fewer dollars to spend on everything else and thus have (non-health) living standards as low as individuals who have income levels below the poverty line.  On the other hand, when rising Medicaid and Medicare spending makes a population better off over time, they are surely enhancing living standards and should be counted.  Moreover, if the improvements from added health spending are not worth the costs, then policymakers should shift spending toward cash or other supports that would benefit recipients more.