- Estimates of the Benefit Incidence of Workfare
In 1996 the Argentine government, with financial and technical assistance from the World Bank, implemented a temporary public works program known as Programa Trabajar. The program was targeted toward poor and unemployed people, and a main objective was to increase the current income of recipients. Program participants received a maximum benefit of 200 pesos per month (equal to the legal minimum wage) in exchange for working in community projects. Approximately 80,000 people a year participated in the program between 1996 and 2002, representing less than one percent of total employment.
In 2001 Argentina suffered a severe political and economic crisis. The poverty rate increased from 30 percent to more than 50 percent in a year. A new government took office in 2002, and one of its principal policies was to make the public works program universal. The number of beneficiaries expanded exponentially, reaching two million people in a country of thirty-nine million. The name of the program was also changed from Trabajar to Jefes de Hogar. Figure 1 shows the monthly average number of beneficiaries of workfare programs between 1996 and 2006. The change in program coverage is clearly evident where the figure spikes in 2003.
A fundamental belief underlying this policy is the effectiveness of self-targeting. The idea is that, even if the government has a poor capacity to enforce the eligibility criteria, only those who are in real need will participate [End Page 129] in a program that imposes work requirements and provides low compensation.1 People who are not in need will choose not to participate since their opportunity cost is higher than the program benefit. Furthermore, the effect of the program on the current income of participants is expected to be positive and large (that is, close to the program benefit), since forgone income is low among those who decide to enter the program. This argument, however, assumes that the work requirement is effectively enforced. If this is not the case, the opportunity cost of participation is lower, and self-targeting weakens. Moreover, the effect of treatment on current income can be negative for some participants. If the hourly earnings during treatment are sufficiently high, people with a strong preference for leisure will prefer participation instead of employment even at the cost of a reduction in total income.2 [End Page 130]
Click for larger view
View full resolution
Analyzing the allocation of funds is not trivial, given that enforcement is usually low in Argentina and political institutions are far from transparent.3 The benefit incidence of the program is thus an empirical question. In an earlier paper, I report anecdotal evidence, mainly from newspapers, documenting cases in which political connections determined selection into the program, participants did not meet the eligibility requirements or did not comply with the work requirement, and political leaders used the resources to buy votes.4 Evidence on the magnitude of the phenomenon, however, is lacking.
The Argentine Ministry of Labor, Employment, and Social Protection (MOL) collected a sample of Trabajar participants in 1997 (hereafter, the MOL sample). Jalan and Ravallion combine this sample of participants with a national household survey (namely, Encuesta de Desarrollo Social), and find that the program was very well targeted toward people in need: 80 percent of participants were in the poorest quintile of the household per capita income distribution.5 They also estimate an average income gain during treatment between 92 and 157 pesos per month, representing between half and three-quarters of the program benefit. Good targeting and a large income gain led to an important reduction in poverty in the short run.
The international evidence on the effectiveness of public works programs is mixed. Dar and Tzannatos review about a hundred evaluations—mainly in developed countries—and find that program design and the context in which...