Comparative analysis of the effective income tax function: empirical evidence using LIS data

Abstract

The effective income tax function is a useful and practical method to analyse the relationship between income and tax amounts. It includes measures of tax progressivity, the maximum effective tax rate, and the horizontal inequity. The effective income tax function was estimated statistically using the seven countries of LIS datasets and Korean data. The estimated maximum effective tax rate is less than or very close to its maximum statutory marginal tax rate, except for Norway and Korea. This implies that estimation of the effective tax function is of great use and significance in evaluating the charateristics of income tax law. The mean squared error from the effective income tax function can be used to represent the degree of horizontal inequity as a ‘quick’ measure.

Acknowledgements

The earlier version of this paper was presented at the 60th Congress of International Institute of Public Finance at Milano, Italy in 2004. This research was financially supported by Luxembourg Income Study Fellowship.

Notes

¹ Kiefer (1984) categorized the indexes of tax progressivity into two types, which are a structural index and a distributional index. The former is the index using a function of the relationship between the amount of income and the taxes imposed on it (the tax structure), while the latter measures a function of the tax structure and of the income distribution. Duclos and Tabi (1996) also divided the tax progressivity indexes into two types; a tax share view and a redistributive approach. They suggested that the Kakwani index and the Suits index belong to the tax share view and the redistributive approach can be applicable to the Musgrave and Thin, the Pechman–Okner index, and the Reynolds–Smolensky index. In addition to these categories, there are a number of progressivity indexes, for example, Baum (1987, 1998), Aggarwal (1994), etc.

² The expression p + 1 is the constant relative risk aversion coefficient and 1/(p + 1) is the intertemporal elasticity of substitution assuming preferences take the form of additive utility functions (Gouveia and Strauss, 1999, p. 155). However, we do not deal with parameter ‘p’ in our study.

³ See Kaplow (1989), Aronson et al. (1994).

⁴ Taxable incomes used for the estimation in each country can be found in the lissification tables of LIS homepage (see www.lisproject.org).