Does Colorado's Old Age Pension Reduce Labor Supply? Evidence from a Regression Discontinuity Design

Abstract

Colorado's Old Age Pension (OAP) program, established in the state constitution in 1937, provides cash benefits to low-income residents beginning at age 60—five years earlier than most comparable programs. This unique age threshold creates a potential regression discontinuity design for studying labor supply effects of early retirement benefits. Using American Community Survey microdata from 2015–2023, I examine whether crossing the age-60 eligibility threshold affects labor force participation among low-income Coloradans. The primary analysis finds no statistically significant discontinuity in labor force participation at age 60 among low-income individuals (effect: $-1.7$ percentage points, SE: $1.5$ pp). The weak first stage—public assistance receipt shows no discontinuous jump at age 60—suggests limited program take-up among the eligible population. Placebo tests reveal significant effects at non-threshold ages, indicating that age-related labor force exit in this population follows a smooth decline rather than a sharp discontinuity at 60. Exploratory heterogeneity analysis suggests larger effects among college-educated low-income individuals ($-7.3$ pp, significant), though this finding requires cautious interpretation given the lack of pre-registration. These results highlight the challenge of identifying program-specific effects when take-up is limited and underscore the importance of first-stage verification in RDD studies of social insurance programs.