Essays in Public Finance and Taxation

Abstract

This dissertation uses administrative tax data to study behavioral responses to taxation and tax enforcement. Chapter I focuses on the two tax penalties associated with Individual Retirement Accounts (IRAs). Chapter II considers the impact of a temporary suspension of tax collection efforts on future tax compliance and income. Chapter III investigates the use of professional tax preparation services by the top 1% of the income distribution.

Chapter I, which is co-authored with Victoria Bryant, focuses on tax-benefited retirement savings accounts. These accounts have features designed to encourage retirement savings, including a penalty for withdrawing before age 59.5. Account holders also face a penalty for failing to take required minimum withdrawals after age 72. Using a bunching analysis, we estimate that these penalties cause more than 17% of traditional IRA holders to change their withdrawal timing each year, shifting close to $60 billion of distributions annually. We estimate a dynamic life-cycle model and run counterfactual policy analysis to analyze the effect of changing these penalties. For both penalties, we find alternative combinations of age threshold and penalty rate that lead to increased average welfare and lifetime tax remittances: increasing the age threshold for penalty-free withdrawals while simultaneously lowering the penalty rate, and increasing the age threshold for required withdrawals while leaving the penalty rate unchanged.

In chapter II, which is co-authored with William C. Boning, Joel Slemrod, and Alex Turk, we ask whether a temporary suspension of efforts to collect outstanding tax debt ultimately lead to lower or higher tax compliance and income. When economic hardship prevents a tax debtor from paying basic living expenses, the Internal Revenue Service puts debt collection efforts on hold and designates the debt currently not collectible (CNC). This paper uses the quasi-random assignment of IRS Revenue Officers to tax debtors' cases as an instrumental variable to identify the causal effects of suspending debt collection on tax compliance and future income. In contrast to uninstrumented estimates, we find no evidence that putting off attempts to collect debt reduces compliance with future tax obligations or future reported income. Among marginal hardship cases, pausing collection instead increases future income, specifically wages earned by the taxpayer's spouse.

Chapter III, which is co-authored with Giacomo Brusco, Yeliz Kacamak, and Mark Payne, considers a recent addition to the conversation about income inequality and the "top 1%:" the extent to which this population avoids and evades its tax liability. As the vast majority of the top 1% of the income distribution use a paid tax preparer, understanding the role of professional tax preparation services among this population is critically important to understanding their tax outcomes. We find that, among the bottom 99% of the income distribution, self-prepared returns have smaller corrections from a random audit relative to paid-prepared returns, but among the top 1%, paid-prepared returns experience smaller corrections relative to self-prepared returns. We then turn to the question of why someone in the top 1% wouldn't use paid tax preparation. We conclude that individuals in the top 1% who are either an accountant or financial advisor, or who have a higher proportion of their income from third-party reported sources such as wages, are less likely to use paid tax preparation services.