Default Option Architecture and Retirement Savings Enrollment: A Large-Scale Field Experiment on Auto-Enrollment Versus Active Choice Across 84 U.S. Firms

This study investigates causal effect of auto-enrollment versus active choice default architecture on 401(k) participation, contribution rates, and investment allocation in a large-scale field experiment across 84 U.S. firms within the context of behavioral economics and retirement policy, an area of growing scientific importance given its implications for federal retirement policy auto-enrollment mandate design, employer benefit default architecture, and financial literacy program targeting for active choice regimes. Using cluster-randomized field experiment at firm level, 12-month follow-up of plan participation, contribution rates, and fund allocation; intent-to-treat and LATE estimation with firm-level randomization; heterogeneity by age, income, and financial literacy, we examine default inertia (present-focused employees accepting automatic enrollment without evaluating alternatives) producing dramatically higher participation rates in auto-enrollment arm; active choice imposing deliberation cost that suppresses enrollment particularly among lower-income and lower-literacy employees in 42 firms (8,240 new hires) in auto-enrollment arm and 42 firms (7,840 new hires) in active choice arm; 16,080 total new hires, 12-month administrative payroll data from plan administrators drawn from 84 U.S. firms across 12 industries matched by firm size, industry, and baseline 401(k) participation rate prior to randomization; data from Benefits Data Trust administrative database. Results indicate that auto-enrollment increases 12-month participation by 38.4 percentage points (68.4% vs. 30.0%, p<0.001); contribution rate conditional on enrollment not significantly different (6.2% vs. 5.8%, p=0.24); auto-enrollment effect 2.4x larger for lowest income quartile vs. highest (p < 0.001), with 38.4 pp participation increase; 68.4% vs. 30.0%; 2.4x larger effect for lowest income as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to behavioral economics and retirement policy and carry actionable implications for the design of programs and policies targeting federal retirement policy auto-enrollment mandate design, employer benefit default architecture, and financial literacy program targeting for active choice regimes.