The Society for Social Work and Research

2013 Annual Conference

January 16-20, 2013 I Sheraton San Diego Hotel and Marina I San Diego, CA

Building Savings At Tax Time: A Large-Scale Experimental Intervention

Saturday, January 19, 2013: 11:00 AM
Executive Center 3A (Sheraton San Diego Hotel & Marina)
* noted as presenting author
Michal Grinstein-Weiss, PhD, Associate Professor and Associate Director, Washington University in Saint Louis, St. Louis, MO
Clinton Key, MA, Research Manager, University of North Carolina at Chapel Hill, Chapel Hill, NC
Dan Ariely, PhD, James B. Duke Professor of Behavioral Economics, Duke University, Durham, NC

The Refund to Savings Initiative (R2S) is a unique partnership between academic experts in asset building and behavioral economics with Intuit, the makers of TurboTax tax preparation software. R2S seeks to improve the financial health of American households by seamlessly integrating a low-touch, scalable, low-cost opportunity and motivation to save into Intuit’s TurboTax FreeFile Alliance tax preparation software. Eligibility to use this product requires that the household have an adjusted gross income of $31,000 or less ($57,000 if active military) or qualify for the Earned Income Tax Credit. Using prompts and anchors informed by behavioral economics principles and earlier R2S research, we aim to increase savings motivation and provide an easy, fast savings opportunity, allowing low-income households to convert their tax refund windfall into sustained, and sustaining, savings.


Approximately 150,000 taxpayers who file their taxes using TurboTax FreeFile Online during the experimental window are randomly assigned to one of 11 treatment conditions. Each condition varies two aspects of the tax filing process. At the end of tax preparation, the point where the taxpayer determines how to receive her refund, she is shown the prompt associated with her treatment condition (various prompts include reminders about emergency savings, savings goals, and retirement savings). She is then asked if she would like to take advantage of the opportunity to divert some of her refund directly into an existing savings account. Those who opt to split their refund are then shown a screen prepopulated with a split between savings and spending (determined by their treatment condition at 25:75 or 75:25). Data on splitting and allocation to savings are collected automatically as part of the operation of the software. An additional sub-set of respondents are identified and contacted for more detailed follow-up.


Preliminary data (from the first 75,000 cases), suggest that few taxpayers (about 1%) in either the control group or a treatment condition take advantage of the opportunity to divert their refund into separate accounts. There is a modest effect of some treatment conditions on splitting behavior. Among those who do split, those who are assigned to the higher savings anchor (75:25) save significantly more than the unanchored control group and 2 to 3 times more than those randomly assigned to the 25:75 anchor. Those assigned to the “emergency savings prompt” save significantly more, several hundred dollars on average, than those in the control condition who are shown no prompt. This is true at both anchor assignments and the prompt appears to interact with the anchor in the estimation of savings amount.


These preliminary data show the promise of small changes in financial processes and behaviors for producing benefits for low-income households. R2S also exemplifies the potential value of collaboration between industry and researchers to identify novel ways to both meet the goals of business and improve the lives of people.