Subsidy Reduction and Transfer Simulation

In the second counterfactual simulation, budgetary savings of 13.7 is distributed among four poor household categories at the beginning of period 1. These categories are urban quintile 1, landless farmer Sindh, landless farmer Punjab, and landless farmer other Pakistan. A simple distribution rule is followed to make transfers to household categories. Transfer is made in proportion of share of the household categories in aggregate income during period 1 of benchmark simulation. Urban quintile 1 has the highest proportion of 54.8%, followed by 22.4% for landless farmer Sindh, 14.7% for landless farmer Punjab, and the remaining 8.1% for landless farmer other. Hence, urban quintile 1 receives 7.5, and the three landless farmer categories receive 3.1, 2.0, and 1.1 respectively.

Average annual real GDP growth declines to 2.95% in this simulation. Average infla- tion rate declines more than benchmark and subsidy reduction simulations. Budget deficit at period 8 is -6.2% of GDP which is also less than the other 2 simulations. Table 3.17 presents the macroeconomic indicators for subsidy reduction and transfer simulation. Simi- lar to benchmark simulation, average annual real income growth is positive for all household categories, except “urban other”, and “waged rural landless farmers Punjab” (Table 3.20).

Calculated utilities, compared to benchmark calculations, decline for all household cate- gories, except for the recipients of transfer payments (Table 3.21). Utilities for urban quintile 1, 3 landless farmer household categories increase in the subsidy reduction and transfer sim- ulation. This suggests that compared to broad-based energy subsidy, transfer payments improve welfare of poor households. Moreover, decrease in utility level for each household categories12 _{are smaller in subsidy reduction and transfer simulation than subsidy reduction}

12

simulation. This in a sense suggests that subsidy reduction and transfer outcomes are Pareto superior to subsidy reduction outcomes.

3.6

Conclusion

Pakistan spends substantial amount of resources in electricity subsidy, which aggravates budget deficit and crowds out priority sector spending. Moreover, non-poor households receive comparatively larger subsidy benefits than poor households. Broad-based electric- ity subsidy, therefore, may not be the right policy for serving the best purposes of poor households. The alternative policy could be generating fiscal savings by cutting subsidy, and making transfer to poor households. From the CGE analysis it appears that subsidy reduction and transfer of savings improve welfare of poor households.

One caveat of my analysis is that I consider each household category in the social account- ing matrix as one single household in the simulations. I do not know how many households are there in each category. The distribution rule for transfer of savings was based on house- hold category’s share in aggregate real income. This share can be larger if there are more households in the category. The share can also be larger if income of the households in the category are higher. Hence, larger share of transfer could be allocated to smaller number of households, which could cause upward bias in utility calculation for any of the recipient categories. Though I don’t know population distribution for SAM 2010 household categories, I have information about SAM 2008 population distribution. Aggregate population count of the four recipient categories in SAM 2008 was 16.4 million, and “urban quintile 1” cat- egory has the largest share of 52.3% of this population. The “urban quintile 1” category also receives more than 50% of the transfer share, based on the decision rule. It is unlikely that population proportions change drastically for the 2010 SAM. Hence, utility calculations for the “subsidy reduction and transfer” simulation is not affected by smaller number of

households receiving larger share of transfers. However, information on population distribu- tion will allow more accurate welfare analysis, which could be one of the extensions of this research.

International development organizations like The World Bank and IMF regularly pre- scribe developing countries to reform regressive energy subsidies, and to implement targeted transfer programs. However, implementing targeted transfer programs, particularly in de- veloping countries could be quite difficult and financially not viable. In this exercise I simply transfer the total savings from subsidy reduction to poor households without considering any cost of targeting. Welfare impacts could be much smaller or quite different if targeting and implementation costs were considered. The next step of this research would be developing a model with costs associated with targeting and implementation of transfer schemes, and to conduct welfare analysis in presence of targeting cost.

Another possible extension could be linking the CGE outcomes with household level mi- crodata, and conduct welfare analysis thorough micro-simulations. This will allow an in depth analysis of how poor and non-poor households are affected by subsidy reform. This paper provides primary results in support of regressive nature of energy subsidy; and welfare improvement for poor households from transfer of savings. However, effective policy formu- lation for subsidy reform would require further analysis of this issue. One definite finding of this paper is that cutting energy subsidies hurts both poor and non-poor households. Poor households are only better off when targeted transfers are made. If targeted transfers are not economically viable, then continuation of broad-based energy subsidy could be a better policy option. My future research agenda in this area will be to analyze and compare these policy alternatives to figure out desired policies to achieve greater social welfare.