EXPERIMENT 6: Tax-Time Savings Interventions

Saving rates among low- and middle-income households tend to be low, but the benefits are high. In order to encourage saving, we identified tax-time (when many of these families receive a refund) as an apt opportunity to build up savings immediately and improve their financial well-being in the long run. Tax-time typically provides these families with a sizable one-time windfall of extra cash, and participants report their intention to save almost half of this lump sum.


Working with partners at Intuit, the makers of TurboTax software, we (at Duke and Washington University
in St. Louis) designed several interventions to encourage people to divert money from their refunds into savings accounts. To test which methods were most effective, we relied on two primary behavioral mechanisms: (a) motivational prompts and (b) suggested savings amounts (anchors). The motivational prompts encouraged TurboTax users to think of concrete reasons to save: an emergency, their family, or the future. This was compared with a generic prompt to save (the control). Anchors were implemented by recommending a savings level (as a percentage of the refund and/or a specified dollar amount), and pre-populating a web form with that amount. The amounts were as follows: 25%, 50%, 75%, $100, or $250.


And while the general prompt did encourage users to save more than no prompt at all (a 6.8% increase), we found that on average the targeted prompts had a higher effect on savings (7.6%). While this increase appears modest, the translation to the number of savers who would have been spenders is substantial: 4,800 participants became savers. Overall, the combined impact of these behavioral interventions increased the amount saved by roughly 6 million dollars.

View the full Journal Article