Policy often differs from the recommendations of theoretical optimal tax models in substantial and enduring ways. Such differences are sometimes surely because policy is suboptimal; however they may also be driven by alternative objectives which shape policy in practice, but which do not appear in the benchmark theoretical model. This dissertation considers three cases of such alternative objectives. The first chapter supposes that work subsidies like the Earned Income Tax Credit may be justified by corrective considerations, rather than the usual redistributive rationale for income taxation, if people are present biased and some benefits from work are delayed. The second chapter explores the role of income taxes in directing talented indiv...