The strange math behind US poverty numbers

What is the best measure of economic means?
What is the best measure of economic means?
Image: Reuters/Joshua Lott
We may earn a commission from links on this page.

The US Census reported today that 11.8% of Americans were living in poverty in 2018. This is down from 12.3% in 2017. These widely reported numbers, based on an antiquated set of thresholds, deserve a healthy dose of skepticism. Even the Census recognizes that.

For an adult living alone, the official definition of poverty in 2018 was that the person’s pre-tax income be less than about $13,100. If they lived with another adult, their total income must be below $16,800. If the two of them had a child, it is $20,200. The numbers are not doubled because it is assumed that people within a household can share resources.

These standards are based on calculations developed by economist Mollie Orshansky in the early 1960s (pdf). Working for the Social Security Administration, Orshansky was tasked with developing a poverty threshold when, in her words, there was “no generally accepted standard of adequacy for essentials of living except food.” Since the amount of spending needed for a nutritious diet had been calculated by other researchers, Orshansky based her method on that number. She also knew that the average family spent about one-third of their money on food at the time. She decided that a reasonable poverty number would be to take the minimum amount of money needed for food and triple it.

That is the basis of the poverty threshold used to this day. The same number is used across the country, regardless of whether you live in an expensive city or cheaper rural area.

Orshanky’s method was reasonable at the time, with so little known about the income and spending of Americans. Today, it doesn’t look so good.

Starting in 2011, the Census recognized that, and started calculating the “Supplemental Poverty Measure” (SPM). This new poverty-rate method goes beyond tripling the amount needed for food. It is based on an assessment of the income needed for a minimum amount of food, clothing, shelter and utilities. It also has a geographical adjustment, for differences in regional costs. In addition, the income measure was adjusted to include non-cash benefits from the government, like the Supplemental Nutrition Assistance Program (sometimes referred to as “food stamps”). By this measure, the poverty level is set at at 13.1%, slightly higher than it is under the classic poverty rate.

Although an improvement, the SPM also has its critics. Both poverty statistics reported by the government are based on income, but many economists argue that deprivation is not only about how much you make, but what you are able to buy. As a result, they prefer to measure poverty in terms of consumption. By that measure, poverty has fallen dramatically since the 1980s. A 2017 study by economists at the University of Chicago and Notre Dame found that relative to what the average person could consume at the poverty level in 1980, only about 3% of Americans consumed less in 2016.

The list of problems with the government’s poverty statistics doesn’t end there. Many researchers believe the Census uses a bad measure of inflation, and that income is underestimated because survey respondents tend to forget to include a lot of government benefits.

Still, the classic US poverty measure remains one of the most-frequently cited markers of US economic progress. It probably shouldn’t be.