The Swedish historical data base project will put together and make publicly available highly disaggregated data on roughly a yearly basis for about 2500 Swedish administrative districts over the period 1749-1952. The finished data set will consist of comprehensive and detailed information on economic activity, political characteristics, vital statistics, occupational structure, education, social and agriculture statistics and infrastructure investments (e.g., railway construction). The comprehensiveness and complete coverage of historical data at the local administrative level is what makes this project unique from an international perspective. Since Sweden has the longest continuous and reliable data series on population and vital statistics in the world,starting as early as 1749, makes it possible to construct a comprehensive panel data set over all 2,500 Swedish local administrative units covering a 200 year period. Consequently, the total number of observations for each variable can be as large as 0.5 million (N=2500×T=200). With this type of rich and disaggregated historical data it become possible to get a better understanding of economic growth, structural transformation and economic development. Also, within-country variation allows for more satisfying empirical identification strategies such as instrumental variables, regression discontinuities or difference-in-differences estimation. As a case in point, I have demonstrated the potential usefulness of the Swedish historical data by addressing the question of whether redistribution of resources towards the poor differs between types of democracy after democratization. The identification strategy is based on a regression-discontinuity design where the type of democracy partly is a function of population size. This paper is currently “revise and resubmit” 2nd round at Econometrica. After collecting the new data, we intend to studying a number of questions related to economic development and growth.