Comparing effects of general subsidies and targeted transfers on poverty : robustness analysis using data set from Tunisia

Abstract EN

This paper starts by presenting a descriptive analysis of the effects of general food subsidies on poverty in Tunisia; as revealed by the household survey data for 1990.

The analysis indicates that the poorest certainly take advantage of this system, but at the price of considerable leakages to the non-poor and a sizeable economic efficiency loss resulting from the relative price distortions.

Further, non-parametric estimations suggest that there are no commodities predominantly consumed by the poor.

This implies that a subsidy program is not an effective way to fight against poverty and so, it is unlikely to improve significantly the living standard of the less well-off members of society by restructuring the current program.

We investigate then the impact on poverty of a more targeted transfer scheme, based on proxy means tests, using an appropriate econometric technique to model it.

Simulations show that this design would be more effective in reducing poverty than general food subsidies.

Dominance tests are also used to assess the likely effects of this reform on a wide range of poverty lines and poverty measures.

The main result is that this design would first-order-dominate a food subsidies scheme within a range of poverty lines including all those estimated for Tunisia.