Undercounting the Poor

In 1995, a blue-ribbon panel of poverty experts selected by the National Academy of the Sciences (NAS) told us that the “current U.S. measure of poverty is demonstrably flawed judged by today’s knowledge; it needs to be replaced.” Critics have long pointed out shortcomings including the failure to adequately account for the effects of “safety net” programs and insensitivity to differences in the cost of living between different places.

The Census Bureau, the federal agency charged with publishing the official poverty numbers, has yet to replace the poverty line. However, in the last couple years it has published an alternative, the Supplemental Poverty Measure (SPM). The SPM is the product of over two decades of work to fix problems in the federal poverty line (FPL).

This new measure takes us one step forward, two steps back. On the one hand, it has some genuine improvements: The new measure makes clearer how the social safety net protects people from economic destitution. It adds basic living costs missing from the old measure. On the other hand, it does little to address the most important criticism of the poverty line: it is just too damned low. The fact that the poverty line has only now been subject to revision—50 years after the release of the first official poverty statistic—likely means that the SPM has effectively entrenched this major weakness of the official measure for another 50 years.

The 2011 official poverty rate is 15.1%. The new poverty measure presented—and missed by a wide margin—the opportunity to bring into public view how widespread the problem of poverty is for American families. If what we mean by poverty is the inability to meet one’s basic needs a more reasonable poverty line would tell us that 34% of Americans—more than one in three—are poor.

What’s in a Number?

The unemployment rate illustrates the power of official statistics. In the depths of the Great Recession, a new official statistic—the rate of underemployment, counting people working part time who want full-time work and those who have just given up on looking for work—became part of every conversation about the economy. One in six workers (17%) counted as underemployed in December 2009, a much higher number than the 9.6% unemployment rate. The public had not been confronted with an employment shortage that large in recent memory; it made political leaders stand up and pay attention.

The supplemental poverty measure had the potential to do the same: a more reasonable poverty line—the bottom line level of income a household needs to avoid poverty—would uncover how endemic the problem of economic deprivation is here in the United States. That could shake up policymakers and get them to prioritize anti-poverty policies in their political agendas. Just as important, a more accurate count of the poor would acknowledge the experience of those struggling mightily to put food on the table or to keep the lights on. No one wants to be treated like “just a number,” but not being counted at all is surely worse.

With a couple of years of data now available, the SPM has begun to enter into anti-poverty policy debates. Now is a good time to take a closer look at what this measure is all about. The supplemental measure makes three major improvements to the official poverty line. It accounts for differences in the cost of living between different regions. It changes the way it calculates the standard of living necessary to avoid poverty. And it accounts more fully for benefits from safety net programs.

Different Poverty Lines for Cost-of-Living Differences

Everyone knows that $10,000 in a small city like Utica, New York, can stretch a lot farther than in New York City. In Utica, the typical monthly cost of rent for a two-bedroom apartment, including utilities, was about $650 during 2008-2011. The figure for New York City? Nearly double that at $1,100. Despite this, the official poverty line has been the same regardless of geographic location.The supplemental poverty measure adjusts the poverty income threshold by differences in housing costs in metropolitan and rural areas in each state—a step entirely missing in the old measure.

We can see how these adjustments make a real difference by simply comparing the official poverty and SPM rates by region. In 2011, according to the official poverty line, the Northeast had the lowest poverty rate (13.2%), the South had the highest (16.1%), and the Midwest and the West fell in between (14.1% and 15.9%, respectively). With cost-of-living differences factored in, the regions shuffled ranks. The SPM poverty rates of the Northeast and South look a lot more alike (15.0% and 16.0%, respectively). The Midwest’s cheaper living expenses pushed its SPM rate to the lowest among the four regions (12.8%). The West, on the other hand, had an SPM rate of 20.0%, making it the highest-poverty region.

Updating Today’s Living Costs

Obviously, household expenses have changed a lot over the last half-century. The original formula used to construct the official poverty line used a straightforward rule-of-thumb calculation: minimal food expenses time three. It’s been well-documented since then that food makes up a much smaller proportion of households’ budgets, something closer to one-fifth, as new living expenses have been added (e.g., childcare, as women entered the paid workforce in droves) and the costs of other expenses ballooned (e.g., transportation and medical care).

The new poverty measure takes these other critical expenses into account by doing the following. First, the SPM income threshold tallies up necessary spending on food, clothing, shelter and utilities. The other necessary expenses like work-related child care and medical bills are deducted from a household’s resources to meet the SPM income threshold. A household is then called poor if its resources fall below the threshold.

These non-discretionary expenses clearly take a real bite out of family budgets. For example, the “costs of working” cause the SPM poverty rate to rise to nearly doubles that of the official poverty rate among full-time year-round workers from less than 3% to over 5%. Bringing the Social Safety Net into Focus

Today’s largest national anti-poverty programs operate in the blind spot of the official poverty line. These include programs like Supplemental Nutrition Assistance Program (SNAP) and the Earned Income Tax credit (EITC). The supplemental measure does us a major service by showing in no uncertain terms how our current social safety net protects people from economic destitution. The reason for this is that the official poverty measure only counts cash income and pre-tax cash benefits (e.g., Social Security, Unemployment Insurance, and Temporary Assistance to Needy Families (TANF)) towards a household’s resources to get over the poverty line. The supplemental poverty measure, on the other hand, adds to a household’s resources near-cash government subsidies—programs that help families cover their expenditures on food (e.g. SNAP and the National School Lunch program), shelter (housing assistance from HUD) and utilities (Low Income Home Energy Assistance Program (LIHEAP))—as well as after-tax income subsidies (e.g., EITC). This update is long overdue since the 1996 Personal Responsibility and Work Opportunity Reconciliation Act (a.k.a., the Welfare Reform Act) largely replaced the traditional cash assistance program AFDC with after-tax and in-kind assistance.

Here are some figures for 2011 that illustrate the impact of each of twelve different economic assistance programs. Social Security, refundable tax credits (largely EITC but also the Child Tax Credit (CTC)), and SNAP benefits do the most to reduce poverty. In the absence of Social Security, the supplemental poverty rate would be 8.3 percentage points higher, shooting up from 16.1% to over 23.8%. Without refundable tax credits, the supplemental poverty rate would rise 2.8 percentage points, up to nearly 19%, with much of the difference being in child poverty. Finally, SNAP benefits prevent poverty across households from rising 1.5 percentage points. The SPM gives us the statistical ruler by which to measure the impact of the major anti-poverty programs of the day. This is crucial information for current political feuds about falling over fiscal cliffs and hitting debt ceilings.

A Meager Supplement

Unfortunately, the new poverty measure adds all these important details to a fundamentally flawed picture of poverty.

In November 2012, the Census Bureau published, for only the second time, a national poverty rate based on the Supplemental Poverty Measure: it stood at 16.1% (for 2011), just one percentage point higher than the official poverty rate of 15.1%. Why such a small difference? The fundamental problem is that the supplementary poverty measure, in defining the poverty line, builds from basically the same level of extreme economic deprivation as the old measure.

In an apples-to-apples comparison (see sidebar), the new supplemental measure effectively represents a poverty line roughly 30% higher than the official poverty income threshold for a family of four. For 2011, the official four-person poverty line was $22,800, an adjusted SPM income threshold—one that can be directly compared to the FPL—is about $30,500. Unfortunately, the NAS panel of poverty experts appears to have taken an arbitrarily conservative approach to setting poverty income threshold. Reasonably enough, NAS panel uses as their starting point how much households spend on the four essential items: food, clothing, shelter, and utilities. A self-proclaimed “judgment call,” they choose what they call a “reasonable range” of expenditures to mark poverty. What’s odd is that their judgment leans back toward the official poverty line – the measure they referred to as “demonstrably flawed.”

To justify this amount they show how their spending levels fall within the range of two other “expert budgets” (i.e., poverty income thresholds) in the poverty research. What they do not explain is why, among the ten alternative income thresholds they review in detail, they focus on two of the lower ones. In fact, one of these two income thresholds they describe as an “outlier at the low end.” The range of the ten thresholds actually spans between 9% and 53% more than the official poverty line; their recommended range for the threshold falls between 14% and 33% above the official poverty line.

Regardless of the NAS panel’s intention, the Inter-agency Technical Working group (ITWG) tasked with the job of producing the new poverty measure adopted the middle point of this “reasonable range” to establish the initial threshold for the revised poverty line. This conflicts with what we know about the level of economic deprivation that households experience in the range of the federal poverty line. In a 1999 book Hardship in America, researchers Heather Boushey, Chauna Brocht, Bethney Gunderson, and Jared Bernstein examined the rates and levels of economic hardship among officially poor households (with incomes less than the poverty line), near-poor households (with incomes between the poverty line and twice the poverty line), and not poor households (with incomes more than twice the poverty line).

As expected, they found high rates of economic distress among households classified as “officially poor.” For example, in 1996, 29% of poor households experienced one or more “critical” hardships such as missing meals, not getting necessary medical care, and having their utilities disconnected. Near-poor households experienced these types of economic crises only a little less frequently (25%). Only when households achieved incomes above twice the poverty line did the incidence of these economic problems fall substantially—down to 11%. (Unfortunately, the survey data on which the study was based have been discontinued, so more up-to-date figures are unavailable.) This pattern repeats for “serious” hardships that include being worried about having enough food, using the ER for health care due to lack of alternatives, and falling behind on housing payments. So if what we mean by poverty is the inability to meet one’s basic needs, then twice the poverty line—rather than the SPM’s 1.3 times—appears to be an excellent marker.

Let’s consider what the implied new poverty income threshold of $30,500 feels like for a family of four. (This, by the way, is about what a household would take in with two full-time minimum-wage jobs.)

This annual figure comes out to $585 per week. Consider a family living in a relatively low-cost area like rural Sandusky, Michigan. Based on the basic-family-budget details provided by the Economic Policy Institute, such a family typically needs to spend about $175 on food (this assumes they have a nearby grocery store, a stove at home, and the time to cook all their meals) and another $165 on rent for a two-bedroom apartment each week. This eats up 60% of their budget, leaving only about $245 to cover all other expenses. If they need childcare to work ($180), then this plus the taxes they have to pay on their earnings ($60) pretty much wipes out the rest. In other words, they have nothing left for such basic needs as telephone service, clothes, personal care products like soap and toilet paper, school supplies, out of pocket medical expenses, and transportation they may need to get to work. Would getting above this income threshold seem like escaping poverty to you?

For many federal subsidy programs this doesn’t seem like escaping poverty either. That’s why major anti-poverty programs like that National School Lunch program, Low Income Home Energy Assistance Program (LIHEAP), State Children’s Health Insurance Program (SCHIP) step in to help families with incomes up to twice the poverty line.

If the supplementary poverty measure tackled the fundamental problem of a much-too-low poverty line then it would likely draw an income threshold closer to 200% of the official poverty line (or for an apples-to-apples comparison, about 150% of the SPM income threshold). This would shift the landscape of poverty statistics and produce a poverty rate of an astounding one in three Americans.

Now What?

The Census Bureau’s supplemental measure doesn’t do what the underemployment rate did for the unemployment rate—that is, fill in the gap between the headline number and how many of us are actually falling through the cracks.

The poverty line does a poor job of telling us how many Americans are struggling to meet their basic needs. For those of us who fall into the “not poor” category but get struck with panic from time to time that we may not be able to make ends meet—with one bad medical emergency, one unexpected car repair, one unforeseen cutback in work hours—it makes us wonder, if we’re not poor or even near poor, why are we struggling so much? The official statistics betray this experience. The fact is that so many Americans are struggling because many more of us are poor or near-poor than the official statistics lead us to believe.

The official poverty line has only been changed—supplemented, that is—once since its establishment in 1963. What can we do to turn this potentially once-in-a-century reform into something more meaningful? One possibility: we should simply rename the supplemental poverty rates as the severe poverty rate. Households with economic resources below 150% of the new poverty line then can be counted as “poor.” By doing so, politicians and government officials would start to recognize what Americans have been struggling with: one-third of us are poor.