Tax reform and social welfare in Indonesia: an empirical investigation

Kodoatie, Johanna Maria (2008) Tax reform and social welfare in Indonesia: an empirical investigation. PhD thesis, James Cook University.

PDF (Thesis) - Submitted Version
Download (1MB) | Preview


The Indonesian tax system has undergone a series of reforms since 1983. Ongoing tax reforms will need to deal with the basic tax principles of equity and efficiency, in addition to revenue adequacy. As is typical in a developing country, a significant proportion of Indonesia's taxes are collected via an indirect system, so there is a critical link between the prices of the consumer goods to which taxes are applied, the welfare of households, and government revenue. Accordingly, the major challenge facing Indonesia's tax reformers is to determine how best to collect revenue in an efficient and equitable manner.

This thesis examines whether the trade off between efficiency and equity occurs in tax reform in Indonesia. More specifically, the study aims to answer the following research questions: (1) Can Indonesia's commodity tax system be reformed in a way that maintains revenue while improving social welfare? (2) Which commodity taxes should be altered so as to increase overall tax revenues at (the) least social cost? To address these questions, the way in which demand for different commodities change in response to changes in tax rates must first be determined and the degree of inequality aversion in Indonesia must be assessed. The estimated demand elasticities and the estimated inequality aversion can then be used to answer these research questions.

Data from a survey of 64,422 households across 26 Indonesian provinces was obtained from the 2002 SUSENAS and this data was used to estimate demand equations for 13 different food groups using two different models. The basic model of tax reform analysis used in this study including the relevant variables involved is mainly derived from the approaches outlined by Olivia and Gibson (2005), by Deaton (1997) and by Ahmad and Stern (1984, 1991). The first model (Model I, Hicksian Model), where the budget share of any commodity group is assumed to be a function of total expenditure and family size was one commonly used by other researchers. The second model (Model II, Marshallian model) uses total food expenditure as a denominator of the budget share (instead of the total expenditure) of the regressand and includes an income variable to reflect quality variations and high income earner behaviour. The Unit Value Index approach is used for computing unit values for the observed commodity groups and Ordinary Least Squares (OLS) regression is used in Deaton's three stage procedure to calculate price elasticities. "Lambda" (the Benefit Cost Ratio of Ahmad and Stern) is used to estimate the relative welfare costs of raising revenues from different commodities, providing guidance on directions for tax reform. To do this, two different tax rate regimes are considered: the MIX tax rate regime (which assumes a 10% uniform tax rates across commodity groups, except for Cereals and Prepared foods and drinks); and the MAX tax rate regime (which assumes higher tax rates on some commodity groups).

This study demonstrates that Model II produced "better" outcomes compared to Model I in terms of not only the correct signs on own price elasticity estimates but also more plausible estimates of the marginal welfare impacts of tax changes. Furthermore, income elasticities and / or expenditure shares were found to be small, indicating that, essentially, the Marshallian welfare estimates associated with Model II are likely to provide reasonable approximations to compensating and / or equivalent variation.

Overall, the findings of the study suggest that a trade – off between efficiency and equity aspects exists in tax reform in Indonesia. Accordingly, this research sought to determine the level of inequality aversion evident in Indonesia. This study used the modified Atkinson Index approach to estimate a unique inequality aversion parameter of 1.5 (instead of using a series "trial and error" numbers as previous studies have mostly done).

Using Model II and an inequality aversion parameter of 1.5, the recommendations of tax reform policy that arise from this analysis are that policy makers should raise tax on Meat, Fruits, and Other consumption and lower taxes on Tubers, Beverages, and Eggs and milk. These products were generally identified as being price elastic of demand and as both the most expenditure elastic and the most income elastic commodities for the majority of Indonesians, especially for villagers and low income groups. Hence, an increase of the tax on the suggested goods, excluding Meat and Other consumption, may be relatively inefficient (since most of the goods are price elastic) but more "equitable". Similarly, lower taxes on Tubers, Beverages and Eggs and milk are likely be offsetting towards a shift of Tubers, Eggs and milk and Beverages Demand.

Recommendations derived from Model II using (a) only urban areas; (b) only rural areas are also considered in this thesis. When the analysis focuses only on urban areas, the analysis indicates that policymakers should increase tax on Cereals and Pulses and lower tax on Meat and Vegetables if considering only efficiency. However, policymakers should raise tax on Pulses and reduce tax on Vegetables if equity considerations are important. When rural areas are the focus for the tax reform policy, then the analysis suggests that policy makers should increase tax on Fruits and Prepared foods and drinks and lower tax on Tubers and Meat. More commodity groups are subject to the reform when allowing for more variation in the tax rates applied in the tax system. This regime suggests that apart from the former recommendation, the policymakers should also raise tax on Fish and reduce tax on Beverages when equity aspect is also considered.

It is important to note that the policy recommendations are likely to differ if analysing different commodity groups and / or if analysing different regions. It may therefore be important to conduct further research using smaller commodity groups as well as to analyse regional variations of consumption patterns among households. Inaddition, the unit values used in this study may not accurately represent true prices Accordingly, further research could usefully test the techniques used to estimate unit values by comparing estimates with market prices that have been collected 'in the field'. Finally, this study was unable to include non-food commodity groups within its analysis (due to an absence of quantity data; that would allow for the calculation of unit values). Further research should aim to include these commodity groups to ensure that the analysis relates to a complete demand system (using econometric estimation technique appropriate to complete systems of demand).

Item ID: 30985
Item Type: Thesis (PhD)
Keywords: Indonesia; taxation principles; tax reform; Hicksian Model; Marshallian Model; taxation policy; inequality aversion assessment; regional variables
Date Deposited: 21 Jan 2014 01:41
FoR Codes: 14 ECONOMICS > 1402 Applied Economics > 140215 Public Economics- Taxation and Revenue @ 100%
SEO Codes: 91 ECONOMIC FRAMEWORK > 9101 Macroeconomics > 910110 Taxation @ 100%
Downloads: Total: 577
Last 12 Months: 5
More Statistics

Actions (Repository Staff Only)

Item Control Page Item Control Page