Principal investigator: Adriano Valladão Pires Ribeiro
Article title: GOVERNMENT TRANSFERS AND POLITICAL SUPPORT
Article authors: Marco Manacorda, Edward Miguel and Andrea Vigorito
Location of the intervention: Uruguay
Sample size: 2232 Families
Main theme: Economic Policy and Governance
Type of Intervention: Money transfer
Primary variable of interest: Support for the government
Evaluation method: Discontinuous regression
Policy Problem
Questions about how government support is impacted by implemented public policies are of broad interest and difficult to measure. Given its design, an anti-poverty transfer program in Uruguay allowed for progress on the topic. It was possible to estimate the effect of transfers on government support among program beneficiaries when compared to non-beneficiaries.
Evaluation Context
Uruguay faced a severe economic crisis between 2001 and 2002, with a drop in per capita income and an increase in poverty and unemployment. Even after showing signs of recovery in 2004, the party that had been in power since 1995 failed to win re-election, giving way to the left-wing party. wide front taking office in March 2005. In this context of recovery, the new government, already in April 2005, implemented the largest anti-poverty program in the country up to that point, the National Social Emergency Care Plan (PANES).
Intervention Details
PANES was a temporary aid program that lasted until December 2007 and had two objectives. First, to provide assistance to families affected by the 2001-2002 crisis and, second, to strengthen the human and social capital of the poorest so that they could eventually escape this condition. The main element of the program consisted of monthly cash transfers that were independent of family size. The transfers were significant and amounted to more than half of the average income reported by program beneficiaries before the start of the aid. Families with children and pregnant women would also receive a food card with limited values depending on the number of children and pregnant women. About 70% of beneficiary families received the card.
The Ministry of Social Development was responsible for collecting information on the families, housing, income, work, and education of the nearly 190 applicants to the program. More than 102 families became beneficiaries, representing about 10% of all families and 14% of the Uruguayan population. To determine who would be eligible for PANES, a scoring system was used considering only the socioeconomic characteristics of the families. In order to avoid any kind of manipulation, neither the data collectors nor the families knew that eligibility would be based on a scoring system, nor what variables and weights would be used. Families below a certain score (eligibility threshold) would be eligible for the program.
After the initial data collection, two further surveys were conducted. The first, between October 2006 and March 2007, collected information not only on housing, family composition, durable goods, work, income, education, medical care, knowledge of rights, and participation in social groups, but also on economic satisfaction, opinions on the program's design, and the respondent's political orientation, including support for the government in power, the variable of interest in the study. The second, between February and March 2008, after the program ended in December 2007, was quite similar to the previous survey, but included additional questions about social and political positions.
Exclusive
To assess the political support of PANES for the party that implemented it, a sample of 2.232 families enrolled in the program whose scores were close to the eligibility threshold was used. The idea is to compare eligible and ineligible families who, being close to the threshold, would have very similar socioeconomic characteristics, and thus the only major difference would be whether or not they received the cash transfer. Therefore, differences in the average questionnaire responses regarding the political position of the groups could be inferred as caused by the government's financial assistance.
Several points regarding the quality of the program's design, aimed at preventing potential manipulation and fraud and thus ensuring the accuracy of the research results, deserve highlighting. First, the eligibility threshold was found to be effective; that is, families below the threshold received the transfer, while those above did not. Second, the scoring rule for each family was created after data collection, preventing manipulation of responses from respondents or administrators. Third, the scoring rule was only made public after the program's conclusion. Finally, precautions regarding the political nature of the questionnaire were taken to prevent respondents from acting strategically in their answers for fear of losing government assistance.
Results
The impact of PANES on government support was measured by responses to the following question: “Compared to the previous government, do you believe the current government is worse (0), the same (1/2), or better (1)?”. This question was part of the questionnaire applied both throughout the program and after its termination. The average of the responses is restricted to the range between 0 and 1, with values close to 1 representing greater support for the incumbent government. As mentioned earlier, the difference in the average responses between the groups eligible and ineligible for transfers near the eligibility limit is what allows for a comparison of the program's impact on government support.
Impact reported during the program: In the survey conducted between the end of 2006 and the beginning of 2007, support for the government among families eligible for PANES was 0.90, while among ineligible families it was 0.77. Therefore, eligibility for the program is associated with a 13 percentage point increase.
Impact reported after the program: Regarding the research conducted in 2008, after the program ended, the results are similar, although smaller. The increase in support for the government among PANES beneficiaries is estimated at between 8 and 12 percentage points.
It is therefore observed that there is a significant increase in support for the government that implemented the transfer program, even after the program has ended.
Lessons in Politics
We can learn from the results and the program design. First, the results of this study indicate that government transfer policies can have large and lasting effects on the political views of beneficiaries. Second, there are several potential problems with this type of measurement, notably the possibility of confusing the causality of events (in this case, the program beneficiaries would be families that already supported the government) and manipulation of questionnaire responses. However, a good program design and implementation show that not only can such problems be avoided, but the results generated are reliable.
Reference
Manacorda, Marco; Miguel, Edward; Vigorito, Andrea. “Government Transfers and Political Support.” American Economic Journal: Applied Economics, 3(3): 1-28. 2011.
