The first part of this thesis covers a question currently high on the public agenda: whether and how to tax capital income. By reinterpreting the Chamley-Judd result, a well-known result in public finance which argues against taxing capital income, it shows that the steady-state assumption is more important than previously thought. Then, it studies how capital income should be taxed when returns to capital differ across individuals, for instance because capital income is positively correlated with ability, or because of returns to scale in investment. Using numerical simulations and economic theory, it concludes that the optimal tax rate on capital income is positive and economically significant. The second part of the thesis studies how public funds are actually spent, investigating possible instances of conflict of interest in the pharmaceutical procurement market. It documents a timing effect between sponsorships offered by pharmaceutical companies to doctors in public hospitals and the procurement contracts received by the companies.