This article explores subjects in optimal income taxation characterized by recent research interest, practical importance in light of concerns about inequality, potential for misunderstanding, and prospects for advancement. Throughout the analysis, paths for further investigation are highlighted. Areas of focus include multidimensional abilities and endogenous wages; asymmetric information and the income of founders; production and consumption externalities from labor effort; market power and rents; behavioral phenomena relating to perceptions of the income tax schedule, myopic labor supply, and the interactions of savings, savings policies, and labor supply; optimal income transfers; the relationship between optimal income taxation and the use of other instruments; and issues relating to the social welfare function and utility functions, including non-welfarist objectives, welfare weights, heterogeneous preferences, and taxation of the family. (JEL D63, D82, D91, H21, H24, H53, J22)

The optimal stringency of the burden of proof is characterized in a model in which relaxing the proof burden enhances deterrence but also chills desirable behavior. The result are strikingly different from those in prior work that uses a simpler model in which individuals only choose whether to commit a harmful act (so only deterrence is at stake). Moreover, the qualitative differences between the optimal rule and the familiar preponderance of the evidence rule—and related rules that look to Bayesian posteriors—are great, much more so than revealed by prior work.

The burden of proof is a central feature of adjudication, and analogues exist in many other settings. It constitutes an important but largely unappreciated policy instrument that interacts with the level of enforcement effort and magnitude of sanctions in controlling harmful activity. Models are examined in which the prospect of sanctions affects not only harmful acts but also benign ones, on account of the prospect of mistaken application of sanctions. Accordingly, determination of the optimal strength of the burden of proof, as well as optimal enforcement effort and sanctions, involves trading off deterrence and the chilling of desirable behavior, the latter being absent in previous work. The character of the optimum differs markedly from prior results and from conventional understandings of proof burdens, which can be understood as involving Bayesian posterior probabilities. Additionally, there are important divergences across models in which enforcement involves monitoring (posting officials to be on the lookout for harmful acts), investigation (inquiry triggered by the costless observation of particular harmful acts), and auditing (scrutiny of a random selection of acts). A number of extensions are analyzed, in one instance nullifying key results in prior work.

Self-reporting--the reporting by parties of their own behavior to an enforcement authority--is a commonly observed aspect of law enforcement, such as in the context of environmental and safety regulation. We add self-reporting to the model of the control of harmful externalities through probabilistic law enforcement, and we characterize the optimal scheme. Self-reporting offers two advantages over schemes without self-reporting: enforcement resources are saved because individuals who report their harmful acts need not be detected, and risk is reduced because individuals who report their behavior bear certain rather than uncertain sanctions.

How should moral sanctions and moral rewards—the moral sentiments involving feelings of guilt and of virtue—be employed to govern individuals’ behavior if the objective is to maximize social welfare? In the model that we examine, guilt is a disincentive to act and virtue is an incentive because we assume that they are negative and positive sources of utility. We also suppose that guilt and virtue are costly to inculcate and are subject to certain constraints on their use. We show that the moral sentiments should be used chiefly to control externalities and further that guilt is best to employ when most harmful acts can successfully be deterred whereas virtue is best when only a few individuals can be induced to behave well. We also contrast the optimal use of guilt and virtue to optimal Pigouvian taxation and discuss extensions of our analysis.

The public at large, many policymakers, and a number of economists hold views of social welfare that are non‐welfarist. That is, they attach some importance to factors other than the effects of policies on individuals’ utilities. We show, however, that any non‐welfarist method of policy assessment violates the Pareto principle.