Do government programs encourage the poor and unemployed to stay that way? So runs the classic argument. This week there's a chart circulating that claims to demonstrate its truth. It purports to show that, as the extremely poor begin to earn more, their actual income either stays level or drops as government benefits are phased out, creating a disincentive to keep working. Is this how it actually works out? Here are the arguments. (See the chart and and analysis that started the debate here.)

Figure 1


  • Anti-poverty Programs Are 'Perverted': See Chart Clifford Thies, the creator of the chart in question, explains that "when you take into account the loss of means-tested benefits (e.g., cash assistance, food stamps, housing subsidies, and health insurance), and the taxes that people pay on earned income, the return to working is essentially zero for those in the lower two quintiles of the income distribution." The result, he claims, is "an underclass acclimated to not working; the supplement of cash and noncash benefits with income from crime and the underground economy; and the government resorting to negative incentives such as mandatory work programs."
  • Get the Government to Study This  Harvard economist Greg Mankiw, long a critic of government poverty programs, asks the Congressional Budget Office to investigate: "Producing this kind of chart correctly is not easy (and I cannot fully vouch for the accuracy of this one) because a variety of different government programs are involved, and their rules are often complex. CBO has the staff to do it right." Furthermore, Mankiw writes, a CBO report would get this problem the attention it deserves. If the chart is right, "as earned income rises from about $15,000 to $30,000, income after taxes and transfers is roughly flat. Indeed, it could even fall. The bottom line: If you are poor, the government is inadvertently ensuring that you have little incentive to try to improve your condition."
  • CBO Needs More than a Chart  The CBO, argues the Economist's Free Exchange blog, also needs to investigate whether this flat line on the left side of the graph, the so-called "poverty trap," actually "traps"; "It's not," explains the blogger, "that I don't understand the negative incentive effects at work here, it's just that those aren't the only factors being considered by individuals deciding how much to work." For example, beyond about $40,000 per year, the line on the graph "continues ... on a steady upward slope." Wouldn't that upward slope after $40,000 also be a factor, an incentive to work, even if income gains temporarily flattened out?
Generous antipoverty programmes may have a negative work incentive at the point at which they fade out, but by moving workers from a station at which additional income is unlikely to move an individual any closer to self-sufficiency to one where that is no longer the case antipoverty policies may essentially turn unproductive workers into productive workers. That's a powerful benefit to set against the potential cost of the implicit tax rate effect.
  • How to Solve This Problem Libertarian economist Arnold Kling agrees with the conclusions Thies draws from the graph, and points out that, to the problem of disincentives, "one solution is to base eligibility for means-tested benefits on total income, including other government benefits programs." That would mean the benefit phase-out wouldn't strike quite so hard, and the line wouldn't be quite so flat. "Another approach," he says, "would be to abolish a lot of specific programs and replace them with generic cash assistance."