The authors construct a broader definition of poverty and factor in programs like Social Security, food stamps, and unemployment insurance. Based on that data, the fraction of Americans with incomes below the poverty line has dropped from 26 percent in 1967 to 16 percent today.
But the paper offers another way to look at this, too. The country has been less successful in helping people break free from the need for safety-net programs in the first place.* That is, if you don't factor in all these programs, then the number of Americans with incomes below the poverty line has actually grown, from 26 percent in 1967 to 29 percent in 2012.
The new study attempts to create a better, more accurate index of poverty that takes into account the full range of household expenses and incorporates government taxes and transfers. It's based on a new method of calculating poverty introduced by the Census back in 2010 — this paper extends that index back to 1967.
Goldfarb's story is worth reading in full, but here are a few other findings — and key charts — from the Columbia study, written by Christopher Wimer, Liana Fox, Irv Garfinkel, Neeraj Kaushal, and Jane Waldfogel:
1) When you factor in safety-net programs, the portion of Americans in poverty fell from 26 percent in 1967 to 16 percent in 2012. That's the red line here:
In the chart above, the blue line shows the old "official" poverty rate, which, the authors suggest, didn't do as good a job of incorporating both the expenses of families or the full range of incomes flowing into households (including safety-net programs). By that older index, poverty has barely budged since 1967.
The red line is their preferred index, which tries to take into account everything from health-care and child-care costs to programs like Social Security and the Earned Income Tax Credit.
The new index shows that people were a lot poorer in 1967 than we thought (although this depends on what inflation index you use). And many of the programs enacted in the 1970s and 1980s — from the Earned Income Tax Credit to low-income heating assistance to the expansion of food stamps — really do appear to have helped reduce poverty.
Note, however, that progress against poverty essentially stalled after 2000, and it's ticked up a bit in recent years.
2) The biggest gains, by far, have come in poverty among the elderly, which fell from 47 percent in 1967 to 15 percent today:
Why such a drastic change? In part because, the authors note, the new index takes into account medical expenses, something the old index didn't account for. That makes health-care programs for the elderly, particularly Medicare and Medicaid, look more valuable.
3) Child poverty has also fallen significantly, from 29 percent to 18 percent. There was a particularly big drop in the 1990s:
4) On the other hand, poverty for working-age adults has barely budged from 15 percent since the 1970s:
Tax and transfer programs don't seem to have had nearly as much effect on working-age adults. (Credit to Kevin Drum for first pointing this out.)
5) The safety net was particularly effective at stopping poverty rates from soaring in the last recession:
In the chart above, the green line calculates the poverty rate before taxes and transfers. The blue line shows poverty after the effects of various safety-net programs.
There are a couple of ways to read this chart. One, if you don't include safety-net programs, then poverty has actually risen from 26 percent in 1967 to 29 percent in 2012. There are more people dependent on safety-net programs to stay out of poverty than ever before.
Or here's another view: The green line shows that poverty rates would have soared during the most recent recession if there were no safety-net programs in place. But as the blue line shows, the poverty rate actually stayed fairly constant. The expansion of food stamps, unemployment insurance, and the Earned Income Tax Credit blunted a lot of misery. (See this recent paper (pdf) by Robert Moffitt of Johns Hopkins University for more detail on that.)
* Rephrased this sentence very slightly in response to Sean Fremsted's criticisms here, whose full post is worth reading.
** Tweaked the bit on the Child Tax Credit thanks to a pointer by Len Burman on Twitter.
Further reading: Here's an older post explaining why the "official" poverty rate is inadequate. The Census tried to improve on the measure in 2010. This latest study tries to tweak the metric further and extend it back to 1967.