Behavioral Insights into Tax Time Savings from CSD’s Tax Filing Experiment

As I mentioned yesterday on the blog, the findings shared in Krista Holub’s presentation at the National Community Tax Coalition’s 2013 conference have the potential to reshape the way we think about tax-time savings. While many pro-savings initiatives like SaveUSA and the existing system of IRAs and 401(k)s include financial incentives for building savings, the Center for Social Development’s “Refund to Savings,” or R2S, experiment relies entirely on behavioral-psychological insights to increase saving rates.

The experiment exists (and will continue to exist, as the group is planning a similar trial for tax season 2014) wholly within Intuit’s TurboTax filing software, which is made available for free to low-income filers. The team exposed each participant to a different intervention expected to have an effect on saving behavior. At the session, Holub explained that the R2S experiment was made up of three sequential steps: the intention survey, the intervention, and the household financial survey. For the purposes of asset-building practitioners, the last two are the most relevant. The household financial survey was a follow-up survey after tax-filing season that aimed to discover the medium-term effects of the filer’s tax-time saving decision. The “intervention” step included three conceptual elements that informed the types of treatment participants were exposed to: decision prompting, choice architecture, and anchoring. Each of these concepts was expected to play a role in changing tax filers’ saving behavior.

The first concept, decision prompting, is employed through a series of motivational prompts that appear at the decision point at which the tax filer can elect to save a portion of her refund. One prompt may say something like, “Think about yourself at retirement. Where do you want to live? What do you want to be able to do? Try to form a vivid image of your life at retirement.” Or another may offer: “In 2011, the average American household accumulated about $5,500 in savings. Think about the amount you saved last year.” They are designed to encourage the filer to reflect on the importance of saving.

The second concept, choice architecture, is simpler, and perhaps more effective. At its worst, it can be seen as a gimmick or a trick to get a consumer to do something he otherwise would not. But just as prize-linked savings harness the usually harmful human impulse to gamble for the purposes of encouraging good behavior, choice architecture is used for good in the R2S experiment. The choice to save is made easier than the choice to not save. It’s the same principle that makes opt-out retirement savings plans much more effective than opt-in systems at developing savings. In the case of the R2S experiment, participants are given several prominent option boxes for them to decide the amount to save, as if that were the default option, while an inconspicuous opt-out option lies below.

Finally, the well-known behavioral-psychological concept of anchoring is put to good use in increasing the amount saved by those who choose to save. While other savings programs may simply allow participants to input the amount of money they want to save, the experiment’s designers instead provided set options, or “anchors,” for various dollar amounts. This simple procedure was found to significantly increase the amount saved for most participants.

The design and results of the experiment are especially useful in broader efforts to search for promising ways of building savings at tax time. A synthesis of promising tax-time saving methods is the topic of a forthcoming Asset Building Program paper, which will be released within the next few weeks. It will collect the most recent evidence on tax-time savings initiatives, including Krista Holub’s (and the rest of the CSD team’s) R2S results; the work on D2D’s prize-linked savings; and the recent findings from MDRC’s SaveUSA evaluation.

The R2S experiment also has implications for the design of incentive-based tax-time savings programs in that the results could be applied to make incentive-based programs even more effective. The behavioral perspective represented by CSD’s R2S experiment and the incentive structure represented in programs like the Financial Security Credit and D2D’s SaveYourRefund are essentially two very different methods for achieving the same goal. Incentive-based programs assume an economic world-view in which people of all backgrounds respond to incentives, while the behavioral approach assumes a human tendency to follow the easiest path. If employed in tandem, these approaches could make for powerful savings policies at both the federal and local levels.