Association for Public Policy Analysis & Management

Though bank account ownership enables households to manage and track transactions, access credit, and build assets (Birkenmaier & Fu, 2015), 7% of US households are unbanked. An additional 20% have bank accounts (FDIC, 2016) but also use alternative financial services (AFS) such as payday loans, which carry high interest rates (Bertrand & Morse, 2011). Lack of bank account ownership and AFS use are associated with asset ownership (Friedline, Johnson, & Hughes, 2014; Lusardi & Scheresberg, 2013). Financial inclusion is an important federal policy objective intended to pull under-served households into the economic mainstream. The purpose of this study is to assess whether an online intervention to encourage low- and moderate-income (LMI) tax filers to save their tax refunds impacted savings account openings in the six months following tax filing. We hypothesized that intervention group participants were more likely than control group participants to add a savings account and assessed whether this treatment effect was conditional on the timing of tax filing and certain types of savings messages and prompts.

Methods:

LMI tax filers (N = 4,692) were randomly assigned to receive one of several savings messages embedded in a popular tax filing software program or to a control group during the 2013 tax season. Messages included prompts to encourage saving for different reasons (e.g., emergencies) and anchors – suggestions to save a certain amount or proportion of one's refund. Based on responses to a household financial survey, adding a savings account was coded as '1' for participants who said they did not have a savings account upon tax filing and yet said they had a savings account at six-month follow-up. We used inverse probability of treatment-weighted logistic regression with covariance control to assess intervention-control group differences in savings account status changes.

Results:

Results supported our hypothesis, as 7.4% and 5.1% of intervention and control group participants added a savings account, respectively χ²(1, N = 4,692) = 4.52, p < .05. Intervention group participants had 60% greater odds of adding a savings account compared to control group participants (p < .05). This treatment effect held among early season (1/13/14 to 2/13/14) filers (p < .05), but not among those who filed in the other two periods (2/14/14 – 3/13/14) and (3/14/14 – 4/17/14). We also found that those who received 25% or 50% anchors (encouragement to save 25% or 50% of their refunds) with or without messages encouraging saving for emergencies or for one's family were more effective in promoting savings account take-up than other types of messages.

Implications:

As savings accounts can help LMI households better cope with economic instability, encouraging LMI tax filers to save their refunds may be a low-touch, highly scalable strategy to promote financial inclusion in line with federal policy objectives. However, this may be true just among early season filers who typically receive higher refunds. Also, LMI tax filers cannot currently open a savings account when they file their taxes online – a functionality the Internal Revenue Service could consider adding through the Free File Alliance.