Abstract:
We study the behavioral effects of a large-scale, repeated, and personalized reminder. Our empirical setting is Sweden's annual pension statement, which is rolled out region by region to all working-age individuals. Combining this variation with unique individual-level user data from the national pension dashboard, we find strong and immediate effects. Dashboard users' likelihood of making a pension forecast rises by 28 percentage points in the statement week-a fourfold increase-before returning to baseline within three weeks. Remarkably, similar spikes occur each year, indicating that repeated reminders consistently reactivate attention rather than losing their impact over time. Complementary regional data on actual pension claims show a 33% surge in weekly claims during the week the statement is sent out.