In an article in support of Mr. Modi’s decision to limit MNREGA to only 200 poorest districts, Bhagwati and Panagariya argue the following:
To appreciate fully how inefficient NREGA is at transferring income to the poor, consider the following. Existing data show that on average 30% of NREGA expenditures are incurred on materials and 70% on wages. Assuming daily NREGA wage to be Rs 130, this requires an expenditure of Rs 186 to employ one worker per day.
But not all Rs 130 in wages amount to transfer. When accepting NREGA employment, the worker forgoes the opportunity to work elsewhere. Even assuming daily market wage to be a low Rs 80, net transfer under NREGA is only Rs 50. So we spend a solid Rs 186 to transfer a mere Rs 50.
Even though the opportunity cost argument might sound logical prima facie it may not hold much water when confronted with reality. In absence of productive work, the opportunity cost for the rural labor could be in fact zero. In that case the transfer is the total Rs. 130. So it really boils down to the number one chooses or the wage rate and probability of finding work the actual worker faces!
Secondly, as argued by Kotwal and Sen, the wage rate offered by MNREGA may have served as a reference rate for determining the market wage rate as demanders and suppliers of agricultural labor internalize this opportunity cost. So most likely, the opportunity cost logic works from MNREGA to the market for agricultural labor and not other way round. Dili Abreu et. al provide a more balanced critic of MGNREGA that is worth reading in this context.