Luxembourg Income Study Data Research Articles

Statistical Inference for Stochastic Dominance and for the Measurement of Poverty and Inequality

We derive the asymptotic sampling distribution of various estimators frequently used to order distributions in terms of poverty, welfare and inequality. This includes estimators of most of the poverty indices currently in use, as well as estimators of the curves used to infer stochastic dominance of any order. These curves can be used to determine whether poverty, inequality or social welfare is greater in one distribution than in another for general classes of indices. We also derive the sampling distribution of the maximal poverty lines (or income censoring thresholds) up to which we may confidently assert that poverty or social welfare is greater in one distribution than in another. The sampling distribution of convenient estimators for dual approaches to the measurement of poverty is also established. The statistical results are established for deterministic or stochastic poverty lines as well as for paired or independent samples of incomes. Our results are briefly illustrated using data for 6 countries drawn from the Luxembourg Income Study data bases.

Inequality in Five Countries in the 1980s: The Role of Demographic Shifts, Markets and Government Policies

This paper uses Luxembourg Income Study (LIS) data to examine levels of and trends in income inequality in Canada, the Netherlands, Sweden, the United Kingdom and the United States. Inequality increased in Sweden, the United Kingdom and the United States but did not increase in Canada and the Netherlands. Increased inequality of household head’s earnings and increased shares of spouse’s earnings in family income account for much of the observed increases in income inequality. The public sector can, in general, be assigned a moderating effect on these changes. Demographic shifts are not assigned any major role in inequality changes.

Targeting Welfare: A Comment*

Mitchell, Harding and Gruen (1994) survey the redistributive impact of Australia's targeted social security system using Luxembourg Income Study data. They argue that the Australian system has the highest target efficiency among ten countries, but the lowest benefit generosity (after taking account of tax clawbacks), implying that generosity is greater in universal or contributory systems. This comment identifies methodological problems with the accounting framework used by Mitchell, Harding and Gruen, although also arguing that the Australian targeted income support system is likely to be more efficient at reducing poverty than social insurance systems. An alternative methodological approach is outlined.

The Redistributive Effect of Direct Taxes: A Comparison of Six LIS Countries

The Luxembourg Income Study (LIS) is used to investigate the redistributive effects of direct taxes in six countries, Australia, Canada, Sweden, West Germany, the United States and the United Kingdom. Two periods are considered and comparisons are made among countries at different points in time and within countries across time. Lorenz curves are adapted and used to make ordinal comparisons of two very general measures of tax progressivity- so called “residual” and “liability” progression. Newly developed statistical inference procedures allow us to take the sampling variability inherent in the LIS data into account.

Income distribution and redistribution through taxation: An international comparison

Income tax progressivity is studied with Generalized Entropy measures of inequality. Luxembourg Income Study data sets for nine countries are used for international comparison and analysis. progressivity indices are generated by using the Generalized Entropy family as well as Atkinson measures. We further our understanding by examining pre-tax and post-tax measures of inequality based respectively on gross and disposable household incomes. The decomposition property is shown to be desirable for enhancing our knowledge of income inequality and the redistributive effect of income taxes. Thus decomposition based on family size and number of earners is conducted. We learn that countries vary in their emphasis regarding redistributive effects of income taxes.

International Comparisons of Welfare and Poverty: Dominance Orderings for Ten Countries

Inference-based stochastic dominance procedures are applied to Luxembourg Income Study data to rank ten western countries in terms of standards of living and poverty. First-order dominance comparisons ranks more than 50 percent of all pairwise comparisons and second-order (generalized Lorenz) dominance adds another 25 percent. Truncated dominance analysis is used to make inferences about poverty and up to 98 percent of the truncated distributions are ordered. The results differ substantially from those obtained when only relative incomes and inequality are considered.

International Comparison of Household Income Inequalities and Redistribution Effect of Taxes

The Luxembourg Income Study data sets for 10 countries are used for an international comparison of inequality among households. Inequality in household income is measured using a number of measures to gauge the robust nature of our observations. In the process, pretax and posttax inequality in income are measured in order to learn about the nature of income tax progressivity in each country. The effects of household size on the measured inequality are discussed as well.

Welfare states and welfare outcomes in the 1980s

Welfare state typologies can be a useful means of grouping together countries with similar policy regimes. Whether the policy outcomes of these regimes are similar is an empirical question which could not be adequately addressed until the advent of the Luxembourg Income Study (LIS) project. Using evidence from the LIS data on income inequality and poverty, this paper examines the extent to which regime types and outcomes coincide. The findings suggest that while there is a broad coincidence, these typological frameworks are beset by boundary problems which result in the misclassification of some countries. In addition, as individual nations change policy direction, the task of sorting ideological postures from substantive change decreases the reliability of such typologies.

International Comparisons of Income Inequality: Tests for Lorenz Dominance Across Nine Countries

This paper examines income inequality across nine countries using the Luxembourg Income Study data set. New statistical tests and comparability of data provide an exceptionally clear picture of relative income inequality. Only 4 comparisons out of a possible 108 cannot be ranked. In most cases, differences in the definition of the recipient unit make little difference in the rankings. Irrespective of recipient units, Sweden, Norway, and Germany come out at the top of the ordinal Lorenz ranking, with Australia, Canada, and the United Kingdom in the middle, and the United States and Switzerland at the bottom. Copyright 1991 by The London School of Economics and Political Science.