Whether the poor respond to economic incentives is fiercely debated and the battle lines between different groups of economists are sharply drawn. In an important contribution, T Besley and S Coate (1992) ("Workfare versus Welfare: Incentive Arguments for Work-Requirements in Poverty Alleviation", The American Economic Review, vol 82) drew attention to the disincentives of public support for the poor - a case in point being workfare or public works programme such as the NREG (National Rural Employment Guarantee) - as it makes them dependent on it, and discourages job-search and income augmenting human capital.

In another equally important contribution, J Dreze and Amartya Sen (1989) (Hunger and Public Action, Oxford: Clarendon Press) are emphatic in their endorsement of public support for the poor that performs protective and promotional roles.

The former refers to protecting the vulnerable from slipping into poverty while the latter relates to helping the poor to break out of poverty. However, huge leakages from the PDS and NREG and the fiscal burden imposed by them are likely to slow down the growth acceleration experienced in recent years, through its deleterious effects on public investment.

There is, thus, a greater urgency now whether these anti-poverty interventions are desirable and, in that case, whether they are fiscally sustainable. As an alternative, it has been argued that simple transfers to the poor are likely to be more cost-effective.

In fact, a recent comparison of the transfer benefits of the EGS with a simple transfer of the EGS funds to the poor shows the latter to be more cost-effective. While we do not wish to comment in detail on the validity and usefulness of such comparisons, it is intriguing that no attention is given to (i) the difficulties of identifying the poor (or making sure that relatively affluent and well-connected individuals do not masquerade as poor); and (ii) the potentially negative effects on labour supply.

Arguably, a one-time lump sum transfer does not affect leisure-work choices at the margin, as predicted by micro-economic theory. It is, however, not self-evident there will be no negative effects on labour supply if there is an expectation of recurring transfers. 

Although the evidence on the (dis)incentive effects of these interventions is sparse, an analysis of the EGS, based on the ICRISAT panel data, corroborated that a non-negligible subset of the poor - especially many agricultural labourers - withdrew from the EGS when better employment opportunities arose. So, the presumption that the poor tend to be dependent on public support is exaggerated, if not mistaken.

A variation on this argument is that, if the poor respond to economic incentives, food price stabilisation and higher agricultural wages may limit the 'demand' for such anti-poverty interventions. This argument is elaborated here with illustrative evidence from the 61st round of the NSS (for 2004-05).

Let us first look at the targeting of the Food-for-Work Programme (FFW) and PDS. About 37 per cent of FFW participants were poor and the remaining were non-poor. In other words, a large majority were non-poor. This is further corroborated by the fact that the average per capita expenditure of the participants (Rs 484) was well above the poverty cut-off point (Rs 358).

As shown in figure 1, the targeting was, however, better among the STs and SCs than among Others. The low coverage of the poor is not surprising as the FFW covers selected districts.

While a quarter of the beneficiaries were poor, a substantially higher fraction of the poor (about 79 per cent) were covered. So, while there was mistargeting in general, the two disadvantaged groups (that is, the SC/STs) were better targeted, as shown in figure 2.

Table 2 TARGETING ACCURACY OF PDS-RURAL INDIA, 2004-051
Poverty Status	Participants  (%)	Non-Participants  (%)	Total
Poor	79.13 (24.15)	20.87 (27.95)	100 (24.85)
Non-Poor	82.21 (75.85)	17.79 (72.05)	100 (75.15)
Total	81.44 (100)	18.56 (100)	100 (100)
1. The poverty cut-off point is Rs 358 per capita per month. For details, see Gaiha, R., G. Thapa, K. Imai, and Kulkarni  (2007), op.cit.

Our econometric analysis confirms significant expenditure-enhancing effects of the FFW and PDS, allowing for the endogeneity of participation in these two programmes. But the more surprising results relate to the determinants of participation in these programmes. After controlling for the effects of demographic, educational, and occupational characteristics, whether a household belonged to the SC/ST group, the amount of land owned, and some unobserved features of NSS regions (such as quality of infrastructure), participation in the FFW varies inversely with agricultural wage rates (including an alternative male agricultural wage rate) at the state level. This has the important implication that, if agricultural wage rates are higher (other things being equal), the demand for programmes such as the FFW would be lower, especially among the SC/STs.

Using a similar specification for the PDS, but with the variation that a food price index at the NSS regional level, and a control for unobserved state-level differences (for instance, the nature of the policy regime), are used, a significant positive effect of regional food prices on the demand for the PDS (or, goods supplied under it) is confirmed. So, the lower the food price index, the lower will be the demand for the PDS, other things being equal. 

Few would dispute the colossal waste involved in both programmes (including the more than a year-old NREG). Whatever the benefits of these to the poor, their cost-effectiveness is likely to be low. If the insights from our analysis are valid, much waste could be avoided through higher public investment in agriculture - especially in irrigation, roads and electricity - which would translate into higher yields, lower food prices and higher agricultural wage rates. Besides, while crises or an occasional market turbulence - a case in point is the avoidable import of wheat at a considerably higher price now than that in May, 2007 - cannot be eliminated, their worst effects can be minimised through better market intelligence and an enabling environment for futures trading.

In conclusion, there are other ways in which markets are capable of serving the poor better than large-scale and often unaffordable anti-poverty interventions.

Gaiha is Professor of Public Policy, Faculty of Management Studies, University of Delhi, and Kulkarni is Research Fellow, Harvard Centre for Population and Development Studies