Unpacking the Empirics of Targeting in Low- and Middle-Income Countries | 111
Evidence Base for the Costs of Poverty Targeting There has been a gradual accretion of evidence on the costs of poverty targeting. There is an increasingly firm body of knowledge on labor disincentives in the impact evaluation literature. There is much less in quantity and rigor on the other topics because the main data sources are process evaluations, which tend to be buried in the archives of program documents rather than the stuff of journal articles, but the available evidence is fairly consistent.
Labor Disincentives Theory and intuition supports the notion that targeting that reduces benefits as a household’s or individual’s earnings rise could decrease work effort. In richer countries with programs that tend to use means testing, benefit differentiation, and sometimes offer significant levels of benefits, incentive issues are a noticeable feature of concern in the literature and policy debates. This is especially the case where families may be eligible for multiple programs, each with its own means test or sliding scale of benefits. Moffitt (2015) compiles evidence on the largest programs in the United States. The marginal tax rates across individual programs vary and also across income levels. For example, the Supplemental Nutrition Assistance Program (commonly known as food stamps) has a nominal 30 percent marginal tax rate, but it is effectively 24 percent because of earnings exclusion provisions. The Earned Income Tax Credit generates a marginal tax rate as high as −45 percent at the bottom of the scale, but it is 21 percent in the phaseout range. Cumulative marginal tax rates for families in the Supplemental Nutrition Assistance Program and facing federal and state income and payroll taxes, which implicitly include the Earned Income Tax Credit and child tax credits, show a range depending on family composition, earnings, number of workers, and so forth. For families with earnings below 50 percent of the poverty line, the marginal tax rate varies from −3 to 35 percent, with a median of 13 percent. For families with earnings between 150 and 200 percent of the poverty line, the marginal tax rates range from 22 to 51 percent, with a mean of 31 percent. The empirical evidence on impacts on work effort shows no significant effects overall but some for single-mother households. So far in developing countries, few programs have a combination of features that would trigger a high level of concern about work effects and/or there are countervailing features at work in labor decisions. Many programs do not determine eligibility based only or principally on current earnings. Few adjust benefits at all as income rises, and those that do tend to do so with one or two steps or with earnings disregards. Few programs