The 10th anniversary of Mahatma Gandhi National Rural Employment Guarantee Act (MGNREGA) has given a fresh opportunity to the scheme’s critics to spar with its ardent supporters. There has never been a shortage of both. The scheme has been the largest social protection programme of the Government of India, and is regarded as the world’s largest workfare programme. In 2015-16, the government allocated a budget of almost 40,000 crore to mgnrega – more than all other programmes of the Ministry of Rural Development put together. However, balanced assessments of the Act based on reliable evidence have been few and far between.
Any serious assessment of MGNREGA is incomplete without addressing two separate but related questions. First: is mgnrega the best form of social protection that the government can buy for the rural poor? In other words, how has it fared on key outcomes of social protection and poverty alleviation, and how does that compare with other competing options of social protection programmes. And second: how well has mgnrega been implemented and what are the most crucial reforms to realise its full potential? While the latter question is often the subject of most debates, the former deserves greater scrutiny.
When it was launched in 2006, the primary goal of MGNREGA was to ensure social protection for the most vulnerable people living in rural India by providing 100 days of guaranteed employment at minimum wages. The most empowering aspect of the programme’s design was that work was to be made available on demand, and within 15 days of receipt of demand, failing which the State was obligated to pay an unemployment allowance. If work was provided, there was stipulation for wages to be paid within 15 days, else an extra compensation was to be paid.
A decade later, there is substantial evidence to show that the social protection promise of the anti-poverty programme has eluded the average mgnrega worker. While the scheme has been a remarkably successful experiment in self-targeting of the most vulnerable among the rural poor due to its prioritisation of manual unskilled labour work, there are significant instances of the state’s incapacity or unwillingness to provide such work on demand.
A World Bank study analysed data from 2009-10 National Sample Survey to find that across India, 44 percent of rural households that wanted work under mgnrega did not get it. The rates for some large states such as Bihar and Maharashtra were remarkably high at 79 percent and 84 percent respectively. The IHDs survey in 2011-12 finds that 70 percent of rural households did not participate in MGNREGA because “not enough work was available”. Most studies find few or negligible instances of payment of unemployment allowance.
The causes for demand rationing are varied such as irregularities in fund flows, bureaucratic incapacity, and influence of village sarpanchs, among others. But the effect is clear – barring in a few top performing states, the average mgnrega worker in most parts of India cannot turn to the scheme when in need. Indeed, in financial year (fy) 2014-15, the average annual employment per household was 40 days, and only six percent households completed a full 100 days of work.
This is compounded by the fact that mgnrega has witnessed a substantial and steady rise in the instances of wage delays. Analysis by the Accountability Initiative shows that the proportion of payments delayed beyond 15 days increased from 39 percent in fy 2012-13 to 72 percent in fy 2014-15. Moreover, only six states paid compensation for wage delays as required under mgnrega. The cumulative effect of these inadequacies is aptly summarised by two of the Act’s strongest supporters, social activists Aruna Roy and Nikhil Dey, in their recent pil to the Supreme Court, “Poor implementation of the mgnrega at the hands of the central as well as state governments has rendered the Act ineffective.”
The above facts paint a grim picture of the scheme’s potential impact on measures of poverty. However, while it is difficult to isolate the sole impact of mgnrega on rural poverty in the absence of an experimental study,the balance of evidence seems to suggest that the Act, with all its faulty implementation, has been a boon to rural India. The ihds surveys estimate that mgnrega accounted for seven percentage points reduction in poverty among participating households, and prevented 14 million people from falling into poverty. In locations with better mgnrega access and implementation, several studies have noted improvement in welfare measures related to women empowerment
But an important question still remains: are the above gains sufficient to justify the huge costs of MGNREGA ? The popularity of MGNREGA in vast tracts of rural India might indicate the fact that India really needs a social protection programme, and right now, it provides only one option. However, there is little evidence to conclude it is the most efficient one. While continuous improvements should and are being made to fix the most teething issues plaguing MGNREGA , a dispassionate look is necessary to explore the larger design question: are there other, more effective models of social protection that the government should explore?
One of the most promising options here is a universal unconditional cash transfer (UCT). It might sound too simplistic at first thought but there is increasing evidence globally about what such cash transfers can do. ucts allow poor households the choice and flexibility of allocating resources to meet the needs they find most pressing, and are less costly to administer than other options because they don’t require the monitoring and follow-up. A randomised evaluation conducted by Johannes Haushofer (Princeton) and Jeremy Shapiro (BCBE) of a UCT programme in Kenya that provided households with seven monthly instalments of $160 found that the programme had significant welfare-improving impacts, both economically and psychologically, for transfer recipients.
In India, economists Pranab Bardhan and Abhijit Banerjee have already advocated in favour of trying out ucts. Banerjee recommends attempting models that impose a small “cost” for the programme recipients to travel a certain distance to pick up cash every week or month at designated locations after giving their biometric information, thereby excluding those who might not absolutely need that small amount of money.
Another option that holds immense promise, especially for the poorest of the poor women-headed households in rural India, is one that combines cash transfers for a few months with a livelihood intervention comprising a productive asset grant, training and support, and intensive life skills coaching for 18 months. Popularly known as the Graduation or Targeting the Ultra Poor model, results from seven randomised evaluations of this model implemented by ngos in seven different countries, including in India (West Bengal), show large and lasting impacts on the standard of living of the beneficiaries. Learning from this evidence, countries in Africa and Latin America are already working to integrate this model within their social protection strategies.
Of course, none of the above means MGNREGA should immediately be replaced with either of the above or other appealing alternate ideas. As the current nda government realised in the recent spate of country wide droughts, mgnrega is making a significant difference to the life of rural poor. But that cannot be an argument against attempting other competing models of social protection programmes on a pilot basis, while simultaneously trying to fix the glaring inadequacies in MGNREGA .
An interesting scenario might be to run two of these models in parallel, and allow the people to choose (using biometric data to avoid duplication) which social protection programme they like the best on a monthly basis. After all, in a true democracy, it’s only fair that the most vulnerable among the poor have some say in deciding how the Government of India decides to spend its annual 40,000 crore on social protection.
(Views expressed are personal)