Winter 2014 Rinku Murgai, Martin Ravallion, and Dominique van de Walle
With participants’ forgone earnings factored in, workfare may be less cost-effective against poverty than other options
India’s Mahatma Gandhi National Rural Employment Guarantee Scheme, established through a 2005 law, is explicitly aimed at reducing rural poverty. It promises 100 days of work a year per rural household to those willing to do unskilled manual labor at the statutory minimum wage announced for the program. The work requirement is meant to ensure propoor targeting. By attracting only those in genuine need, and encouraging a return to the regular workforce when help is no longer needed, the program aims to incentivize behaviors that solve the problem of knowing who is genuinely “poor” and who is not.
The available evidence suggests that the program is quite well targeted to poor rural households. But even excellent targeting matters little if the net gains to participants are small. While work requirements can ensure good targeting for workfare programs, this probably comes at a cost to poor participants—notably in forgone earnings. To minimize such costs, workfare tends to be advocated in places and times with high unemployment, such as in lean agricultural seasons and during macroeconomic crises. Advocates often assume that workers would be idle in the absence of the program, and conclude that the net gain is the workfare wage.
Yet even with reasonably high unemployment overall, the forgone income of workfare participants is unlikely to be zero—and its magnitude clearly matters in whether workfare is a cost-effective policy choice. Little is known about these hidden costs. The few available estimates of mean forgone income for participants range from 25 percent of workfare earnings (in the Indian state of Maharashtra) to 50 percent (in Argentina). To arrive at estimates of forgone earnings, standard evaluation methods rely on comparing means for program participants with those for a selected comparison group of nonparticipants.
A recent paper by Murgai, Ravallion, and van de Walle revisits these questions for the Mahatma Gandhi National Rural Employment Guarantee Scheme in Bihar. The state is among the poorest in India, with a rural poverty rate of 55 percent in 2009/10 and a rural unemployment rate (18 percent) that is twice the national average. Its rate of underemployment is likely to be even higher. This is the type of poor labor-surplus economy in which workfare has been seen as having much promise for fighting poverty.
The authors ask whether the selftargeting aspect of the workfare program is sufficiently pro-poor to justify it as an efficient means of transferring money to poor people. Could it be that the information constraints are so severe and unemployment so high that the self-targeting mechanism using work requirements tilts the balance in favor of workfare even if the work were to produce nothing of value? Or are the latent forgone incomes too large even in this poor labor-surplus economy? The authors employ a novel survey-based method to measure these costs at the individual level and use these cost measures in estimating the poverty effects of the workfare program relative to those of various counterfactuals.
The survey asked sampled participants about their expected employment and earnings had they not been taking part in the program, providing individual-specific information on forgone opportunities. While acknowledging that counterfactual survey questions can be difficult, the authors found that response rates were high and the answers make sense.
The results indicate that workfare participants in Bihar, though poor, are not drawn solely from the pool of the unemployed. Many report forgone earnings, but these vary greatly, ranging from roughly zero to close to the wage rate for casual market labor. On average, forgone earnings amount to about a third of the workfare wage rate.
Factoring in these householdspecific opportunity costs, the authors find that the extra earnings from the workfare program in Bihar had less impact on poverty than would the same gross expenditure used to provide either a uniform transfer to everyone (whether poor or not) or a uniform transfer to all those holding a government-issued ration card intended for poor families. This also holds when the authors allow for 10 percent leakage in the transfer schemes, though if leakage turned out to be much larger than that, workfare would have a greater effect on poverty.
The authors conclude that even in this poor labor-surplus rural economy, the much-vaunted self-targeting mechanism that is achieved by imposing work requirements does not tilt the balance in favor of unproductive workfare over options using cash transfers with little or no targeting. For workfare to be more cost-effective than other options, it would have to work better. Reforms would need to reduce the substantial unmet demand for work, close the gap between stipulated wages and wages received, and ensure that workfare is productive—that the assets created are of value to poor people. And cost-effectiveness would need to be reassessed at the higher levels of funding that these changes would imply.
Rinku Murgai, Martin Ravallion, and Dominique van de Walle. 2013. “Is Workfare Cost-Effective against Poverty in a Poor Labor-Surplus Economy?” Policy Research Working Paper 6673, World Bank, Washington, DC.