2024-09-29-美联储-社会保障与高频劳动力供给_来自优步司机的证据（英）_62页_3mb

Summary of "Social Security and High-Frequency Labor Supply: Evidence from Uber Drivers"

1. Key Findings

• On average, older Uber drivers reduce their labor supply by 2–5% during the week of Social Security benefit receipt, primarily through reduced workdays (extensive margin), but these effects are statistically precise but economically small.
• Individual analysis reveals substantial heterogeneity: Approximately 1% of drivers reduce labor supply by over 40% (20 percentage points in probability of working), while the majority show minimal response.
• Effects are observed both anticipatorily and post-receipt, lasting up to one week before and after the benefit date. Drivers using Uber's Instant Pay (faster access to earnings) are more likely to respond.

2. Methodology

• A staggered treatment design leveraged the exogenous variation in Social Security disbursement timing (based on birth date) to isolate causal effects.
• Two econometric approaches were used:
  • Paired-event study: Compared cohorts within 7 days of treatment to estimate anticipation and post-receipt effects.
  • Two-stage difference-in-differences (DiD): Scrutinized identifying assumptions, confirming effects are limited to the two-week window around benefits.
• Individual-level treatment effects analysis showed that heterogeneity cannot be fully explained by observed covariates, suggesting unobserved preferences (e.g., target-earning behavior) may drive responses in a small subset of drivers.

3. Policy Implications

• The results highlight the tension between average and individual effects in labor supply models. While most drivers behave as predicted by neoclassical theory, a small group exhibits significant deviations, undermining the assumption of uniform responsiveness.
• Policies addressing labor supply distortions may need to target specific individuals (e.g., through behavioral nudges) rather than relying on broad-based programs. The arbitrary timing of Social Security benefits may impose welfare costs on responsive workers.