This paper presents revised tax ratios based on more realistic assumptions than those used in a previous study applying the same approach (based on tax revenue statistics and national accounts data) to measuring the effective tax burden. Although the levels of the revised tax ratios are sometimes quite different from those previously found, the two data sets are generally highly correlated. The paper also presents a sensitivity analysis of relaxing some remaining unrealistic assumptions for countries and periods where that is possible. It is found that this often has a large effect on the tax ratios, especially for capital, and the two data sets are sometimes no longer highly correlated. This highlights the need to use these ratios in conjunction with other indicators, such as average effective tax rates, to corroborate the story they tell.