What’s the Best Way to Measure Poverty: Income or Consumption?


Yesterday we learned that 15.1% of Americans were living in poverty in 2010, the highest level since 1993, and up nearly 1 percentage point from 2009, when it was 14.3%. That data is based on an income measurement which shows that in 2010, 46.2 million Americans were living below the poverty line, defined as $22,314 a year for a family of four.

But income is just one way to measure poverty, and a particularly tricky (and narrow) way at that – so says Notre Dame economist and National Poverty Center research affiliate, James Sullivan, who believes that to measure poverty strictly by income fails to accurately reflect people’s true economic circumstances. Income alone ignores the effects of things like the Earned Income Tax Credit, Medicaid, food stamps, and housing subsidies. From a Notre Dame press release on Sullivan’s recent poverty research:

“Income received from food stamps, for example, grew by more than $14 billion in 2009. By excluding these benefits in measuring poverty, the Census figures fail to recognize that the food stamps program lifts many people out of actual poverty,” Sullivan says. “If these programs are cut back in the future, actual poverty will rise even more.”

Using income-based numbers only also overlooks the struggles of many Americans who are tightening their belts – those who are worried about losing their jobs or facing foreclosure, or those who devote a large chunk of their paychecks to paying off medical bills. The standard of living for these people is lower than their income would suggest.

In a recent paper, Sullivan and co-author Bruce D. Meyer of the University of Chicago, argue that consumption offers a more robust measurement of poverty than income. Their key point is that poverty, when measured correctly, has declined over time, which is contrary to official measurements. Here’s the full version of their paper. From the abstract:

This paper examines changes in the extent of material deprivation in the United States from the early 1960s to 2009. We investigate how both income and consumption based poverty have changed over time and explore how these trends differ across family types. Estimates of changes in poverty over the past five decades are very sensitive to how resources are measured. A poverty measure that incorporates taxes falls noticeably more than a pre-tax income measure. Sharp differences are also evident between the patterns for income and consumption based poverty. Income poverty falls more sharply than consumption poverty during the 1960s. The reverse is true for the 2000s, although in 2009 consumption poverty rises more than income poverty… Income based poverty gaps have been rising over the last two decades while consumption based gaps have fallen. We show that how poverty is measured affects the composition of the poor, and that the consumption poor appear to be worse off than the income poor.

Some quick highlights:

  • Income and consumption measures of the poverty gap have generally moved in opposite directions in the last two decades, with income based poverty gaps rising, but consumption based poverty gaps falling.
  • Sullivan and Meyer show that upward bias in the Consumer Price Index (CPI-U) has a large effect on changes in poverty over long periods of time. For example, between the early 1960s and 2009, an income poverty measure that corrects for this bias declines by 13.5 percentage points more than a comparable measure based on the CPI-U.
  • Compared to the income poor, the consumption poor are less educated, less likely to own a home, more likely to live in married parent families, and much less likely to be single individuals or elderly. The fraction of the consumption poor living in married parent families is 80% higher than the fraction of the income poor living in such families in recent years.

Here are three graphs from their paper demonstrating the differences between income-based measurements of poverty, and consumption based measurements over time:



Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.



View All Comments »
  1. KPres says:

    Hidden due to low comment rating. Click here to see.

    Disliked! Like or Dislike: Thumb up 11 Thumb down 16
    • Owen says:

      First, Poverty in America is an absolute number. It is determined by the consumer price index which is a measurement of the cost of living determined by the most basic goods (think “milk”). Look at a European country like Germany if you are serious about poverty being relative.

      Second, I disagree. Poverty should be measured relatively.

      Thumb up 6 Thumb down 4
      • Nathan says:

        I also agree that poverty should be measured relatively (at least in America), for the following reason: Money directly correlates with political and economic power. Thus, in a society where $1M/yr is the median income, an individual who earns $600k/yr would need a friend just to match the political power of a median incomed individual and would need dozens if not hundreds (or thousands in a nation with as wide an income gap as America’s) of friends to match the political and economic power of somebody in the 95th or 99th percentile.

        Thumb up 3 Thumb down 2
      • Enter your name... says:

        Except that relative measures result in occasional nonsense. According to relative poverty standards, one of my relatives is “poor”, because her income is at the national median (~$50,000), but currently lives in a town with a lot of millionaires. Conversely, a family of four in Detroit whose income is below the federal poverty threshold is NOT “poor” according to relative poverty standards, because so many Detroit households have such low incomes that this is normal.

        Also, in a nation as diverse as ours, it fails to take purchasing power into account. $1 buys more in Harlingen, Texas than in does in San Francisco.

        What we need is a Basic Income standard: Figure out what you have to have in the long run (the smallest appropriate size apartment for the number of people in the family, the cheapest healthy diet, the minimum annual expense for clothes, etc.) and add it up.

        Thumb up 0 Thumb down 0
      • James says:

        “…the smallest appropriate size apartment for the number of people in the family…”

        Though from my perspective, anyone who has to live in an apartment is poor, even if that apartment’s on Park Avenue in Manhattan.

        Thumb up 1 Thumb down 0
  2. DaveyNC says:

    I think it was Phil Gramm who said that America is the only place in the world where the poor people are fat.

    We don’t have very many poor people here. We have plenty of less-affluent, though. I think culture probably plays a role in that. I think many people reach their comfort level with things like Section 8 housing, food stamps and so on.

    Hot debate. What do you think? Thumb up 10 Thumb down 14
  3. twobeef says:

    If consumption is going up but the rate of wealth / savings for the median person is going down, what does that mean?

    Well-loved. Like or Dislike: Thumb up 7 Thumb down 2
  4. vb says:

    I wouldn’t be surprised if a large reason that consumption poverty is shrinking while income poverty is increasing is because of a growing portion of the population becoming retired. After retirement, income generally drops considerably, but consumption does not. A similar situation likely happens as people spend more time in school before entering the work force.

    Well-loved. Like or Dislike: Thumb up 9 Thumb down 1
  5. Mary says:

    Hidden due to low comment rating. Click here to see.

    Disliked! Like or Dislike: Thumb up 8 Thumb down 20
    • JohnS says:

      How can you claim that Sullivans quote is pure speculation, when your post is the most unfounded piece of nonsense ever posted on the internet?

      Thumb up 6 Thumb down 3
  6. Spencer Thomas says:

    Neither. We should measure “likelihood of starving, being forced to eat unhealthy food because that’s all you can afford, access to clean water, access to public transportation that allows you to get to work in <= 1 hour, ability to rent housing in an area with jobs and without a high level of violence and still be able to eat/pay for electricity/whatever, ability to experience an external shock like a job loss or serious medical situation and not wind up in huge amounts of debt/homeless." Forget consumption of income. Try safety.

    Well-loved. Like or Dislike: Thumb up 11 Thumb down 0
  7. Rishikesh says:

    The poverty line in India is drawn by calculating how many can afford to take in a minimum amount of calories per day. Any family that can afford to buy food worth 2100 calories per day in cities and 2400 calories per day in the villages is above the poverty line. It is around 300 Rupees or six dollars a day (at an average of 50 Rupees to a dollar).

    Thumb up 4 Thumb down 0
    • Rishikesh says:

      The poverty line in India is drawn by calculating how many can afford to take in a minimum amount of calories per day. Any family that can afford to buy food worth 2100 calories per day in cities and 2400 calories per day in the villages is above the poverty line. It is around 300 Rupees or six dollars a month (at an average of 50 Rupees to a dollar).

      Thumb up 1 Thumb down 0
  8. Nathan says:

    I’d be curious if it was even possible, but I think, in a democracy that is invariably decided to some greater or lesser extent by the amount of money raised, a much more interesting method of measuring poverty would be in measuring a household’s ability to donate money to (or otherwise exercise clout with) a political interest. This would not necessarily be a measure of the household’s proclivity to do so, but only their ability to do so.

    At the moment, there is not a single federal election in which one person equals one vote, as congressional districts differ in population size, states differ in population size (for senatorial power), and, of course, the president isn’t elected directly by the people. As such, the true measure of a person’s freedom in America would have to be, almost by definition, the extent to which he or she can influence an election.

    This would make for a simple baseline. Those who cannot afford to donate to elections are in poverty. Those who can are not. Of course, then you’d have to figure out a way to measure that ability, but at least now you’d have some kind of goal.

    Alternatively, for my Ron Paul supporting friends, it seems a simple matter to measure poverty not by ability to eat or by income, but rather by ability to collect savings. I imagine, for certain libertarians, the ability to save money is equal to, if not MORE important, than the ability to influence elections.

    As for measuring poverty by consumption, I think we’ve sighted on a really terrible baseline. In Soviet Russia, everyone got to eat. Did that make them wealthy or even middle class? Of course not. The definition of the middle class is not whether they are fat and have nice cars. It’s how much fiscal and political clout they have.

    Thumb up 2 Thumb down 1