Share

Working Paper No. 2010-32, Paris School of Economics, Paris, Lehman, Christian, 2010

A randomized control evaluation in Mexico finds no effect of government transfer programs on reducing inequality in rural villages, yet greater household spending by the poorest villagers raises the welfare of all villagers through linkages in financial and commodity markets.

The persistence of rural poverty in developing countries is shown, in the literature, to be accompanied by a significant level of inequality.  Government transfer programs are intended to alleviate poverty and may include conditional cash transfers, unemployment insurance, and old-age pension. Consider the distribution of household incomes in a rural village, and define the ‘lower tail’ as the poorest households and the ‘upper tail’ as those households whose welfare is highest and who are ineligible for government transfers. For inequality to be reduced, transfers must target the lower tail specifically (which they do) and not hold any benefit for the upper tail. Lehman argues that the lower and upper tails do not “behave in isolation from each other” and part of the government transfer is redistributed to the upper tail, thus hindering a reduction of inequality.

The Mexican government’s Progresa program provides the setting for a randomized control evaluation of transfers on rural inequality. Under the program, monetary grants valued at approximately 20% of household income were provided to lower tail households in rural villages. Data are obtained from a baseline and three follow-up studies conducted between 1997 and 2000. Village characteristics are consistent across the program; they have 45 households on average and are predominantly agriculture-based, with statistically significant levels of income inequality. Drawing conclusions on the impact of transfers on village inequality requires knowing how that village would have fared in the absence of a transfer program. An empirical solution is found in the random selection of 320 villages to participate in the first phase of the program in 1997, with the remaining 186 villages serving as a ‘counterfactual’, or control group. The empirical model uses two common measures of inequality, the Gini coefficient and the Coefficient of Variation.

The results find no statistically significant reduction in inequality in participating villages, whether the impact of monetary transfer is estimated using the change in household income or the change in household welfare (proxied as food consumption). Moreover, income and welfare of the upper tail is found to improve during Progresa, suggesting a spillover of program benefits. Evidence of the spillover effect is found using the control group data and the observed impact on transfer recipients – if program benefits were not somehow redistributed to the upper tail, villages would experience a reduction in inequality. The author describes two types of linkages by which government transfers are redistributed from the lower tail to the upper tail.

First, linkages exist through informal credit and insurance markets. Agricultural production is subject to numerous and complex risks, creating a demand for credit and insurance products to ensure a consistent stream of household income. Inspired by a culture of reciprocity, gift-giving has become a form of informal lending that often replaces deficient formal financial markets. In this way, the lower tail redistributes their income from government transfer to the upper tail. As per the literature, the Progresa program found every household along the village income distribution participating in gift giving.

The second linkage occurs through commodity markets, specifically from the heightened demand for agricultural and non-agricultural imports. The demand stems from consumption levels that surpass the village’s own productive capacity once the government transfer increases household income. However, since the costs of importing remain prohibitively high for most households, those with the capacity to do so create small enterprises and engage in retail import activity to meet village demand. The upper tail, therefore, accrues income from the lower tail by creating import activity and collecting a mark-up on retail goods sold.

The findings suggest that government transfer programs are less effective in reducing inequality than previous thought. There appears to be a redistributive, or spillover, effect with gains accruing to the program-ineligible upper tail. Considering that the ‘upper tail’ in rural villages are typically not wealthy, but rather less poor than the lower tail, the redistribution of transfers will effectively improve the overall poverty level. As a matter of public policy, evaluations that do not consider the linkages will underestimate the benefit of transfer programs.