Conference Paper
Do Automatic Enrollment Schemes Cannibalize Voluntary Retirement Saving? Evidence from a Nationwide Quasi-Experiment
No.
ERF32AC_142
Publisher
ERF
Date
May, 2026
Topic
H3. Fiscal Policies and Behavior of Economic Agents
H2. Taxation, Subsidies, and Revenue
Using the nationwide automatic enrollment (AE) policy implemented in Türkiye in 2017, this paper examines whether automatic enrollment into retirement plans generates new saving or mainly reshuffles existing saving. Using administrative data spanning universe of IPS and AES contracts, we track voluntary and automatic contributions at the individual level. A difference-in-differences design comparing individuals in a narrow ±2-year bandwidth around the age threshold isolates the impact of AES on voluntary IPS contributions and on total contributions (voluntary plus automatic), allowing us to quantify crowding out and net saving. Our findings reveal that AES reduces voluntary IPS contributions by about 85₺ per participant and roughly 30 percent of the new automatic saving is offset by lower voluntary contributions. Net of this partial crowding out, total retirement saving increases by about 195₺ per participant. Among persistent savers, the reduction in voluntary contributions is larger, about 178₺, and the implied crowding-out rate is close to 60 percent; even for this group, AES yields a net gain in retirement saving of roughly 140₺ per individual. These findings show that default-based enrollment policies change the level and composition of retirement saving and that their net impact depends on prior saving behavior.