Tax ratios derived using aggregate data - also known as implicit tax rates - have attracted increased attention from policymakers and analysts as a possible approach to measuring average effective tax rates on labour, capital, households, corporations and consumption. This study reports on conceptual and practical difficulties encountered in the measurement of average tax rates using aggregate data (e.g., National Accounts and Revenue Statistics). To examine the robustness of previous studies on tax ratios, the existing indicators are re-calculated for an expanded sample of countries and over an extended time period, and the results are compared with those derived under an alternative suggested methodology. The work finds that most tax ratios reported in the literature suffer from a number of flaws, and highlights measurement problems that are much broader than discussed in the literature. The identification of substantial shortcomings with these measures is useful, given the interest in their use for policy purposes. The message of this study is that policymakers should be aware of the measurement problems underlying average tax rates based on aggregate data, should they be fielded to shape public policy debates.