Here is exactly how the Census poverty calculation is derived

Confusion abounds about how the annual poverty figure is produced. I sought to put to rest my own uncertainty on this matter once and for all by replicating the poverty figures from the microdata. What follows is an explanation of how the official rate is computed, including how families are constructed, what income is included, and so on.

The poverty figure is computed on the family level. A person is in poverty if they belong to a family that is in poverty. In the Census report, there are three types of families.

  1. Families
  2. Unrelated Subfamilies
  3. Unrelated Individuals

In the microdata, there are actually five types of families.

  1. Primary Families
  2. Related Subfamilies
  3. Unrelated Subfamilies
  4. Nonfamily Householders
  5. Secondary Individuals

A primary family is like Marge, Homer, Lisa, Bart, and Maggie from the Simpsons. A related subfamily lives in the same household as a primary family, and is related to them. So imagine Bart married someone and they, as a couple, remained in the house on Evergreen Terrace. Bart and his spouse would be a related subfamily. For poverty purposes, related subfamilies are collapsed into their primary family. So all of the people are treated as if they were one big family, all of their income goes towards the overall family income, which ultimately determines whether all of them are above or below the poverty line. In the Census report, the families category includes all primary families and related subfamilies.

Unrelated subfamilies are pretty self-explanatory. Imagine Marge and Homer took in Ned and his wife (before she died). They would be an unrelated subfamily living in Marge and Homer’s household. They are treated as a distinct family unit. Their income is not bunched in with the primary family for poverty calculation purposes.

Nonfamily householders are single people who live in their own housing unit or in a housing unit made up of purely unrelated people. If Lisa moved into her own apartment, she would be a nonfamily householder. If Lisa moved into an apartment with entirely unrelated friends, they would all be individual nonfamily householders. Secondary individuals are roomers, boarders, or resident employees with no relatives in the household. A live-in butler like Geoffrey from the Fresh Prince is a secondary individual. Some unrelated roomer boarding in a house is also a secondary individual. Nonfamily householders and secondary individuals are combined into the unrelated individuals category for the Census report. Unrelated individuals under the age of 15 are excluded from poverty calculations altogether.

Poverty Threshold
For a family (and all its members) to be in poverty, their total family income must be below the poverty threshold. There are 48 different poverty thresholds. A given family’s poverty threshold is determined by how old the householder is, how many people there are in the family, and how many of those people are kids.

The numbers themselves were first established in 1963, and originally represented triple what a given family would have needed to afford a minimal food diet. Since then, the figures have just been adjusted for inflation.

The following types of income count towards a family’s income. Recall, if all of the income from all of the family members from all of these sources is below the poverty threshold, the whole family and all its members are in poverty.

  1. Wages and salary
  2. Non-farm self-employment income
  3. Farm self-employment income
  4. Unemployment compensation
  5. Worker’s compensation
  6. Social Security (includes regular SS and SSDI)
  7. Supplemental Security Income
  8. Public assistance or welfare (e.g. TANF)
  9. Veterans’ benefits
  10. Survivor’s income (i.e. for widows)
  11. Disability insurance (other than SSDI)
  12. Retirement income (e.g. pension)
  13. Interest
  14. Dividends
  15. Rents
  16. Education benefits (e.g. grants)
  17. Child support
  18. Alimony
  19. Financial assistance
  20. Other income (catch-all for other pre-tax, cash income that is not capital gains)

So in sum, it counts pre-tax market income, plus cash benefits, but not non-cash benefits, not tax credits, and not capital gains. Add up all the income from all of the family members from the above list of sources, and if the sum is lower than the poverty threshold, the family and all its members are in poverty.

Notes Unrelated To Measuring Official Poverty
So for poverty purposes, notably absent from the income adding side are, among other things:

  1. SNAP (i.e. food stamps)
  2. Section 8 Housing Vouchers
  3. Child Tax Credit
  4. Earned Income Tax Credit
  5. Medicaid
  6. Medicare

And notably absent from the income subtracting side are, among other things:

  1. FICA Taxes
  2. Income Taxes
  3. State and Local Taxes

But also, the poverty threshold is problematic. Inflation-adjusting the food budget from 1963 may not capture poverty well nearly 40 years later. There are also geographic differences in costs. And so on. The Census has a supplemental measure which it has not yet released for this year that tries to take account of many of these issues.

In addition to thresholds and what income to count, there is of course the crowd that says you should look at consumption. I wont go to deeply into all of that here though, as I just want to explain the official measurement.