Research

Summer Youth Employment Programs are known to have significant impacts on youth outcomes based on lotteries from oversubscribed programs. But most cities cannot use a lottery design due to heterogeneity across youth and jobs. How can programs achieve efficiency and equity under alternative assignment mechanisms? Using hiring platform data, we study youth application and employer selection behavior to explore these design challenges. We find large mismatches between the distribution of youth versus jobs leaving 10% to 25% of positions unfilled. Moreover, employers were nearly twice as likely to select white youth relative to their representation in the applicant pool. This disparity persisted when controlling for other demographics, the number and timing of applications, and job readiness. Our findings reveal that workforce development programs may perpetuate in equities in the absence of simple random assignment. Using a job matching algorithm, we show that placing just 30% of positions by lottery can improve both equity and efficiency.

We provide the first causal evidence that congestion pricing induces asymmetric temporal substitution exclusively at specific pricing boundaries. Using New York City’s January 2025 implementation and high-frequency establishment-level mobility data, we employ a Difference-in-Discontinuities design testing multiple potential thresholds. We find a statistically significant 5.75% temporal shift only at the 5AM weekday boundary (p<0.05), with precisely estimated null effects at all other transitions including both weekend boundaries. This sharp concentration at a single threshold robust to bandwidth selection and alternative control cities reveals that temporal substitution requires precise alignment between pricing boundaries and latent schedule flexibility. Despite limited behavioral change at only one of four pricing boundaries, non-linear traffic flow relationships translate this 5.75% shift into substantial congestion relief. These findings challenge standard congestion pricing models that assume symmetric schedule delay costs and suggest that program effectiveness depends primarily on trip elimination and modal substitution rather than temporal redistribution.

This paper estimates short-run employment effects of California’s sectoral minimum wage for fast-food restaurants. On April 1, 2024, covered limited-service chains faced a $20 hourly floor while the statewide minimum wage was $16 (up from $15.50 in 2023). Using establishment-level mobility data and a difference-in-differences design restricted to “clean” counties without local wage ordinances, I proxy on-site staffing with weekly long-duration device visits (>4 hours). Baseline estimates indicate an average 8% decline in on-site employment at treated outlets relative to comparable full-service and retail establishments. An event-study shows no pre-trends and a gradual, persistent post-announcement decline, consistent with anticipatory staffing adjustments. Placebo estimates for exempt “enclosed” venues are null, and effects are similar across urban and rural locations. The results imply that a sectoral wage floor induces meaningful reductions in labor utilization on intensive margins (hours, shift density) rather than immediate large-scale job loss.

This paper examines how California’s 2024 fast-food minimum wage ($20 under AB-1228) affected consumer behavior. Using high-frequency transaction data from ten national chains and a difference-in-differences design, I find that customer visits and transactions fell by 6–7 percent, while revenue declined by about 5 percent, implying higher spending per visit and partial price pass-through. Income-resolved data show sharp redistribution: higher-income consumers spend 2–3 percent more per visit despite visiting less, whereas low-income customers show little change. Distinguishing composite from income-specific effects reveals that consumer-side incidence is heterogeneous and borne disproportionately by higher-income households.

Last decade has seen noteworthy local policy decisions, especially a trend in decentralisation of wage determination. Considering local policy changes are aimed at the local areas where boundaries are porous, there is a need for detailed and accurate geographic and time information. Using the establishment location and mobile-device location data by SafeGraph, this study explores how the labor market responds to local minimum wage ordinances. I use the difference-in-differences approach to estimate the effect of increase in minimum wage on the variation in duration of visits at a location which can be used as a proxy to employment hours. I find a decrease in employment hours when there is a proportionate increase in local minimum wage and an increase in distance travelled from home with an increase in minimum wage. The study further demonstrates that the local labor market, especially in the non-tradeable sector, is more responsive to the changes in local minimum wage than the state binded minimum wage changes.

The prior literature finds no negative relationship between minimum wage and employment in the U.S., especially for the nontradable industry. The argument hinges on the use of the contiguous region to study the minimum wage variation by controlling for economic shock which might be correlated with the minimum wage changes. I use mobile-device location data to study cross-area movement for labor market zones when local minimum wage changes. I use the spatial and temporal differences in the minimum wages to find the decrease in visits at the establishment located that experienced an increase in minimum wages controlling for the home destination of the visitors. Moreover, this decrease in visits depends on the distance between the destination and the home location of the visitor. Further, I use home Census Block Group (CBG) characteristics to predict the low-wage visitor at an establishment.