Abstract
This paper provides comprehensive estimates of the savings effects of automatically enrolling employees in retirement plans. We use administrative U.S. tax data to measure the retirement savings of employees (and their spouses) at 745 firms. Consistent with prior findings, we estimate that automatic enrollment increases participation in the year after hire by 86 percent and retirement plan contributions by 51 percent. However, we also find employees are 33 percent more likely to take a non-rollover withdrawal, driven by employees who separate from their employer. Incorporating this offsetting behavior, we estimate net savings increase by 37 percent on average in the short run. Spouses do not alter their saving behavior. Over a longer time-horizon, the net savings effect for employees still employed by the same firm declines for high-wage employees and increases for low-wage employees. However, the net savings effect for employees who have separated from their firm declines substantially.