This paper uses national accounts data to adjust market and disposable Top 10% and Top 1% household survey income shares for 39 developed and developing countries that are part of the Luxembourg Income Study (LIS). An additional novelty of this study is the distinction between labor and capital income. The obtained results suggest that for most countries top income shares are significantly higher than those reported in household surveys, which mainly underestimate top capital income. While the presented results should be treated with some caution, our easy-to-implement approach seems suitable for countries for which no tax data is available.
