Introducing a Relative Poverty Calculator

Previously, I created a poverty calculator for the Official Poverty Metric and Supplemental Poverty Metric that lets users select which incomes (and income reducers) to include in a poverty calculation (see here). The main goal of that calculator was to teach people how the poverty metrics work and determine how many people were kept out of poverty by transfer incomes.

To supplement that calculator, I have now created a relative poverty calculator (see here). It does not use the official or supplemental poverty metric, but instead allows users to define poverty as a percentage of median income, which is consistent with how many other countries define it.

Right now, the poverty calculator only has data from 2009 to 2013, which is also the years the first poverty calculators use. If I can get the data, I may extend it back to 1967 at some point, which I think would make it a lot more interesting. Even if I don't do that, it should still be useful in the years to come (and as a teaching tool).

The way the relative poverty calculator works is largely explained in the poverty calculator itself. But there are still some details missing:

  1. Income Concept. The calculator uses disposable income. You can see the complete definition of this concept by clicking on this link (here) and looking at the checked boxes. It is market income plus transfer income (not counting health benefits) minus taxes minus child support paid.
  2. Equivalization. To equivalize incomes (as defined by (1)) across different household units, I take each household's income and divide it by the square root of household members. So a one-person household with $10,000 of income has an equivalized income of $10,000 ($10,000 divided by the square root of 1, i.e. 10,000/1). A four-person household with a $10,000 income has an equivalized income of $5,000 ($10,000 divided by the square root of 4, i.e. 10,000/2). Dividing by the square root of the number of household members is a common equivalization method. It is done in order to simulate (in some way) the economies of scale larger households enjoy. Each person in a household is assigned the equivalized income of their household for purposes of calculating the overall median income and calculating how many people are below the poverty threshold.
  3. Household Unit. All the incomes flowing into a household are used to determine the household's income for purposes of (1) and (2). Thus, it is necessary to determine how to group people into households. For this, I use the SPM unit, which is defined as: "All related individuals who live at the same address, and any coresident children who are cared for by the family (such as foster children) and any cohabiters [i.e. cohabitating partners] and their relatives." See discussion of this household unit here and here.

On all three of these points, other methodological choices are possible. You can define income somewhat differently (e.g. to exclude in-kind benefits). You can use a different equivalization method (e.g. by dividing by the number of household members, or assigning household members different weights depending on whether they are adults or children). And you can pick a different household unit (e.g. by including everyone living at an address, even if they aren't necessarily in a unit likely to share their income resources, such as roommates in an apartment). Depending on what choices you make, you will get somewhat different poverty rates. For this reason, change over time using a consistent measurement is the more interesting question to probe, which is why I created the relative poverty calculator to display data from all of the years at once.