Pension Plan Characteristics and Framing Effects in Employee Savings Behavior

David Card
Department of Economics
University of California, Berkeley
and National Bureau of Economic Research

Michael R Ransom
Department of Economics
Brigham Young University

May 2008 | Issue # 90

A growing body of research suggests that savings decisions are affected by a wide range of influences that play no role in a conventional neoclassical model of savings behavior, including framing effects, default effects, and inattention. This paper contributes to the growing evidence of behavioral decision-making effects in savings outcomes using detailed micro data on the retirement savings behavior of college and university faculty.

Many post-secondary institutions in the United States offer a defined contribution pension plan funded by the combination of an employer contribution and a mandatory employee contribution. Employees can also make tax-deferred supplemental contributions to the same asset fund. A standard lifecycle savings model predicts a dollar-for-dollar tradeoff between supplemental savings and the combined regular pension contributions made on behalf of an employee. If people compartmentalize their salaries and their employer’s pension contributions into different “mental accounts,” however, supplemental savings will be more sensitive to employee contributions (which, like supplemental savings, appear as salary deductions) than to employer contributions (which do not).

Our findings confirm that supplemental savings rates depend on how compensation is labeled. In particular, supplemental savings are significantly lower when a larger fraction of the regular pension contribution appears as a salary deduction. The discrepancy is large: we estimate that supplementary savings are reduced by 60-90 cents per dollar of employee contributions to the regular pension, but only by one-half as much per dollar of employer contributions. Consequently, two faculty members with the same total compensation and the same total contribution rates to their regular pension will reach retirement age with substantially different amounts of supplemental saving, depending on the share of regular pension premiums labeled as an employee contribution. We interpret these findings as further evidence that behavioral departures from a strict neoclassical choice framework can help to explain the observed variability in savings behavior and wealth outcomes, even among highly educated workers with predictable future income streams.

© 2014 and prior years, Teachers Insurance and Annuity Association - College Retirement Equities Fund (TIAA-CREF), New York, NY 10017