On the distributional impact of a carbon tax in developing countries: the case of Indonesia

Abstract

This paper, using a computable general equilibrium model with highly disaggregated household groups, analyses the distributional impact of a carbon tax in a developing economy. Indonesia, one of the largest carbon emitters among developing countries, is utilized as a case study in this paper. The result suggests that, in contrast to most industrialised country studies, the introduction of a carbon tax in Indonesia is not necessarily regressive. The structural change and resource reallocation effect of a carbon tax is in favour of factors endowed more proportionately by rural and lower income households. In addition, the expenditure of lower income households, especially in rural areas, is less sensitive to the price of energy-related commodities. Revenue-recycling through a uniform reduction in the commodity tax rate may reduce the adverse aggregate output effect, whereas uniform lump-sum transfers may enhance progressivity.

Appendices

Appendix 1: Sensitivity analysis

In a CGE exercise, because some of the parameters are taken from other sources such as others studies, models, or literature, it is necessary to examine the reliability of the results with respect to uncertainty in the parameters. In a standard or ‘ad-hoc’ sensitivity analysis, the model is solved for one or two different sets of parameters, and then the sensitivity of the change in endogenous variables is examined. However, since there are many parameters inputted into the model, this approach is difficult or less practical to implement when we want to examine the sensitivity of the results on the independent uncertainty about the values of several parameters. In this model, for example, for Armington elasticity alone, because the model has 38 different commodities, a sensitivity analysis on each of the parameters would be computationally burdensome. This paper hence chooses the systematic sensitivity analysis (SSA) via the Gaussian Quadrature method to conduct a sensitivity analysis (Arndt 1996; Pearson and Arndt 2000). Given the distribution of M parameters, this method deals with finding the best possible choice of parameters in N simulations if we want to estimate means and standard deviations for all endogenous variables. Arndt (1996) has shown that the results using this method are surprisingly accurate, given the relatively modest number of times the model is solved.

Table 5 shows the result of systematic sensitivity analysis for carbon tax simulation (SIM 1, no-recycled revenue), assuming triangular distribution for all parameters and allowing each of the parameters to vary by as much as 50 % from its mean. In general, though some variables tend to be more sensitive than others, Table 5 suggests that the result of carbon tax simulation is robust to variation in the extraneous parameters as shown by low standard deviation of most endogenous variables.

Table 5 SSA of SIM 1: carbon tax (50 % variation in all parameters) (in %age change)

Looking at the confidence interval in real household expenditure by centiles also suggests that the distributional impact of a carbon tax is less likely to be sensitive to parameter variation. For example, it can be seen that with a 95 % confidence level, the real expenditure of the poorest (the centile 1st household group) group will rise by not less than 1.342 % and that of the richest (the centile 100th household group) will not increase (0 % rise in expenditure per capita). Therefore, the carbon tax tends to reduce inequality in rural areas.

An idea of the direction of the poverty impact can also be obtained by looking at what happens to households close to the poverty line. In urban areas, for example, it is the 13th centile household group. Since its 95 % confidence interval is between −0.386 and −0.205, with 95 % confidence it can be concluded that poverty in urban area falls following the introduction of a carbon tax.

The same robustness is also expected for the other simulations conducted in this paper.

Appendix 2: Alternative scenario

This section provides alternative scenarios in which the three scenarios are implemented with a similar carbon reduction target; i.e., a reduction of 6 % from the initial condition. In other words, in the first scenario (SIM 1A), a carbon tax is implemented, in such that the total reduction of carbon is as much as 6 % less than the initial level, without revenue recycling. In the second scenario (SIM 2A), the implementation of the carbon tax will be accompanied by a reduction in a uniform general ad-valorem sales tax rate for all commodities, such that extra government revenue disappears, while controlling the total reduction of carbon is at exactly the same level as that of the first scenario (SIM 1A). In the third scenario (SIM 3A), the implementation of the carbon tax will be accompanied by making a uniform lump-sum transfer to all households. The total reduction of carbon in this scenario (SIM 3A) is controlled to be similar to that of the first scenario (SIM 1A). The results can be seen in Tables 6 and 7. These alternative scenarios do not change either the contents of the discussion or the conclusion of this paper.

Table 6 Impact of carbon tax policies on industrial outputs (in %age change)

Table 7 Distributional effect of carbon tax policies