An Experiment Examining the Effects of Income Tax Deferral on Retirement Account Preferences and Savings Strategies

by Shane Stinson and Marcus M. Doxey, The University of Alabama

Americans typically make retirement planning decisions long before the fruits of prudent savings may be utilized. These decisions often take the form of periodic elections to contribute to personal (e.g., IRA) and/or employer-sponsored (e.g., 401(k), 403(b)) retirement plans. While the federal government has offered several tax incentives in recent years to encourage greater plan participation, prior research shows that many individuals still save too little and/or engage in naïve investment strategies that impede long-term capital appreciation. These trends help produce waves of eligible retirees who must continue their employment beyond their desired retirement age, rely on government assistance, and/or curtail their postretirement spending in ways that cannot be fully explained by expected declines in consumption.

All else equal, common retirement savings plans targeted by government incentives offer economically equivalent future outcomes if individual tax rates remain stable over time. However, differences in the funding methods and tax structures used to administer various account types alter the salience of the immediate sacrifice embodied in periodic retirement contributions. In this study, we posit that such contextual differences affect individuals’ willingness to save for retirement. We further seek to examine the underlying motivations for this phenomenon and determine whether certain tax mechanisms may be more effective than others at encouraging greater participation in retirement savings plans.

We propose an experiment investigating individual savings preferences under conditions simulating annual retirement elections. We will assign participants (employed U.S. citizens, ages 19-59), recruited from the Amazon Mechanical Turk and the Rand American Life Panel, a taxable income stream and reward them based on their total after-tax holdings after a simulated investment period. We will manipulate the timing of retirement contribution elections to simulate conditions for personal and employer-sponsored plans and offer retirement accounts with tax structures mirroring those of common traditional and Roth savings plans. We will further manipulate the participant’s expected tax position (refund or amount due) before any retirement contributions are made. By comparing asset allocations and tax-adjusted performance across our proposed experimental conditions, we hope to provide evidence that the nature and timing of taxes on retirement income affect the choices made by plan participants. Depending on the context, these characteristics of common retirement savings plans may encourage or discourage the assumption of equity risk that generally fuels the long-term growth of investment capital and may improve one’s chances of maintaining financial independence after retirement.

Back to 2016 Sandell Grant Program Recipients