According to the World Bank report entitled India Development Update: Towards a Higher Growth Path, published in April 2015 (please click here to access):
• With over 50 million beneficiary households, and expenditures between 0.5 and 1% of GDP, MGNREGA is amongst the largest anti-poverty programmes in the world.
• Analysis of household survey data from Bihar shows that under ideal conditions, the rural poverty rate of 50% at the time of the survey could come down by at least 14 percentage points. However, compared to a potential reduction in poverty by 14 percentage points, actual impact on rural poverty is only about 1 percentage point.
• At one end of the spectrum, the NREGS is shown to have delivered significant positive impacts on a range of outcomes -- from consumption and nutrition, to quality assets and productivity improvements, particularly for the poorest – in Andhra Pradesh. At the other end, impacts of the scheme in Bihar fall far short of potential.
• The World Bank study has highlighted a number of reasons why the potential impact of MGNREGS may not be realized in practice: the supply side is too slow to respond to the demand for work on the scheme; workers are not paid the full scheme wage; delays in wage payment; and awareness of how to demand work is limited.
• In Bihar, more than two-thirds – about 10 percentage points – of the gap between potential and actual impacts is attributable to the ways in which the NREGS is not fulfilling the provisions of the Act. The rest is due to the foregone income, i.e., an opportunity cost to the worker from giving up alternate employment in order to take up MGNREGS work. Forgone incomes are hard to avoid, but are important to the calculus when evaluating the net income gains from MGNREGS participation and its cost effectiveness against other alternatives.
• At the national level, 2009-10 National Sample Survey (NSS) data show a great deal of unmet demand. Almost 46% of households report that one or more members of their household would have liked to work on the scheme; only 25% secured any work over the course of the year.
• Participation rates in the scheme are not, as a rule, any higher in poorer states. There is greater demand for work on the scheme in poorer states, but also a lower capacity to meet that demand. As a result, the degree of unmet demand or rationing, i.e., the fraction of people who wanted work but did not get it, is bigger in the poorer states. In Bihar, unmet demand alone accounts for nearly three-fifths of the gap between potential and realized poverty impacts.
• Discrepancies between the scheme’s stipulated wage rates and actual wages received by workers also contribute to the gap between potential and realized impacts. In Bihar, on average, workers on the scheme received about 10% lower wages than the stipulated rate. This gap is not due to payment delays, as the gaps are estimated on total wages owed to the individual, not the amount actually received by the time of the survey.
• More recently, payment delays have emerged as a major bottleneck in program administration, and are a strong disincentive to participating in the program. National surveys highlight variation amongst states – in 2009/10 a survey found nearly 68% of MGNREGS beneficiaries received wages within 15 days in Andhra Pradesh. The same estimate for Rajasthan and MP were 10% and 23%.
• Despite government’s efforts to promote transparency and encourage monitoring from administration and civil society (social audits), MGNREGS is plagued by leakages. Leakage can take multiple forms, such as inflating the number of days worked per person, or registering fake persons to siphon off funds. Comparison of MGNREGS aggregate levels of employment reported in official data with independent measures based on household surveys show that the gap exists, but the range of available estimates is too wide to be conclusive. NSS-based estimates for 2007-08 can confirm only a quarter to half of MGNREGS employment recorded in official reports, depending on the method used.
• Specialized surveys with more detailed questions on participation and remuneration in the scheme suggest substantially lower estimates of leakage. For instance, the Bihar survey of MGNREGS implies a 20% gap between survey-based estimates of employment and official data in 2008-09 compared to the 2007-08 NSS-based estimate of 70% gap in the state.
• Even though MGNREGS is better targeted than a cash transfer scheme would be, wage earnings alone are not sufficient to make it more cost-effective at reducing poverty. But, it could be, if the assets created under it are of sufficient value to the poor. Other non-pecuniary benefits – like empowerment for socially vulnerable groups including women -- would further strengthen the case for workfare.