Measuring the effective tax progressivity of personal income tax over the whole income distribution, and consistently defined over time, is an important tool with which to evaluate tax reforms. While survey data provides the representative income distribution, it measures the tax liabilities imprecisely and lacks representativity at the top. Administrative tax data overcomes these disadvantages, but does not reflect the overall income distribution, as it excludes non-taxable incomes and those who do not pay the tax. In this paper, I provide a matching, easy-to-implement strategy that combines these two types of data to exploit their advantages. This approach can be applied to different years of repeated cross-sections, which results in a representative and time-consistent database for the distributional analysis of income and taxes. As a result, the effective tax rate of the top one percent of the income distribution is found to be 30.1 percent in 2019, similar to 30.3 percent in 1998 despite cuts in the marginal tax rate for the top earners during the time span.