Mexico's Oportunidades: The Challenges of Self-Selection for Targeted Poverty Programs
César Martinelli, Professor of Economics, ITAM, Mexico City and Woodrow Wilson Center/Comexi Public Policy Scholar
Susan W. Parker, Professor of Economics, CIDE, Mexico City
Moderator: Andrew Selee, Woodrow Wilson Center
As Mexico's poverty alleviation program, Oportunidades (formerly known as Progresa) nears its ten year mark, it is seen as a potential model for other countries facing similar challenges of transitioning their economies and people out of poverty. The program, which was first implemented in rural areas of Mexico in 1997 and later extended to urban areas in 2001, operates by creating incentives with a cash transfer that is conditional on regular health clinic visits and regular school attendance by the recipient's children. The cash incentive amounts to an average of about $30 dollars per month – a seemingly low amount that however represents approximately 30% of a typical recipient's monthly expenditures. This money is intended to cover the cost that parents may face by taking their child out of the labor market and placing them in school. The program attempts to reach out mostly to the female heads of households with the belief that they will take more responsibility for their children's welfare. Also, based on research showing that girls in Mexico tend to drop out of school earlier, the program also slightly favors girls in the allocation of funds. Around five million families—almost a quarter of Mexico's population—receives cash transfers through the Oportunidades program, making it the major pillar of Mexico's social policy.
César Martinelli pointed out that poverty alleviation programs in developing countries face various challenges when trying to correctly identify the poor populations to target. Because the poor often do not have transparent and quantifiable incomes, social programs must rely on participants to self-report their own information – opening up room for misreporting. Martinelli commented that until recently there had been little, if no, research done on misreporting of information by participants when applying for social programs.
The Oportunidades program is of interest to researchers for several reasons, including the fact that in the rural phase of the project a random assignment of towns was used with a control group to better measure the outcomes. In their study however, Martinelli and Susan W. Parker address the component that involves self-selection of participants in the urban area phase of the program. They look at how eligibility for Oportunidades is determined and the problems associated with misreporting.
As Martinelli noted, the selection process in the administration of Oportunidades in Mexico's urban areas requires that potential participants fill out a questionnaire that is used to determine eligibility. Once eligible, another questionnaire is administered as well as a follow-up home visit to verify the information. Both questionnaires use indicators that serve as proxies for measuring levels of poverty. With the data from the home visits serving as a tool to verify the information reported in the questionnaires, Mexico's Oportunidades program offers researchers the valuable opportunity to assess levels of misreporting and hypothesize about the reasons for this behavior.
The poverty indicators on the questionnaires that Martinelli and Parker looked at when measuring levels of misreporting included, among others, family size per number of rooms in the house, the number of children below 11 years of age, whether the head of household had no schooling, whether the family had a vehicle, paved floor, gas boiler, refrigerator, toilet, and running water.
Martinelli and Parker reported that they found that both underreporting and overreporting had occurred. Approximately 80% of participants who owned vehicles underreported having one. Approximately 73% reported to not having a gas boiler, a phone or a TV while in actuality they did own these goods. This finding, however, was expected: applicants for social programs have a great deal to gain by underreporting so that they can ensure they qualify. Nonetheless, a particularly surprising finding was the overreporting of goods. Of the participants who did not have a toilet, nearly 40% reported having one. Furthermore, about 30% who were verified in the home visit not to have a gas stove or tap water had reported that they did have these goods. Similarly, 25% overreported having a concrete floor. Martinelli pointed out that the four items that were consistently overreported were items that most people in low-income urban communities possess. Martinelli and Parker also found that men were less likely to overreport and more likely to underreport than women. They also found, surprisingly, that as the participants' education level increased, the more likely they were to overreport.
The findings of the study lead Martinelli and Parker to hypothesize that underreporting may have been motivated by participants' desire to maximize their eligibility for benefits under the program. The overreporting, however, points to a possible cultural or social stigma that participants may anticipate experiencing if they report not having basic household items.
Martinelli pointed out that their study's findings highlight the problems that lie in both using certain variables to indicate poverty and also in using self-reporting methods when determining eligibility for social programs. Overreporting, in the case of a program run like Oportunidades, creates much concern because eligible and needy participants run the risk of being eliminated from the program's benefits. Finally, Parker and Martinelli noted that there is an important need to revisit the design of eligibility determination procedures in social programs, especially those that target the poor and attempt to reduce poverty levels.