Socio-Demographic Influences on Worker Departure Times across the U.S.

Abstract: This study explores the influence of socio-demographic factors on the distribution of workers’ departure times across census tracts in the DMV (Washington D.C., Maryland, and Virginia) area using data from the 2022 American Community Survey (ACS) 5-year estimates. Departure times are a crucial factor in transportation planning and congestion management, as they directly influence peak demand on transportation networks. By modeling how socio-demographic factors like household income affect departure times, planners can gain insights into peak load distribution and develop strategies to manage traffic congestion more effectively.
The data for this study was extracted from the ACS 2022 5-year estimates, focusing on worker departure times across 1,335 census tracts in the DMV area. The primary variables of interest were the number of workers departing at different time intervals (e.g., 12:00 AM to 4:59 AM, 5:00 AM to 5:29 AM, etc.) and median household income for each tract. The dataset was reshaped from wide format to long format to facilitate time series cross-sectional (TSCS) analysis, where each row represented a unique combination of census tract and departure time interval. This structure allowed for analysis of temporal patterns within tracts over different time intervals.
We estimated a random effects model using the plm package in R, with the model accounting for both the fixed effects of median household income and departure time intervals, and random effects capturing census tract-level variability. This specification enabled us to analyze how socio-demographic characteristics affect worker departure patterns while accounting for the unobserved heterogeneity across census tracts. Model results indicated that median household income did not significantly affect the number of workers departing at different times. However, there were significant differences in worker departure patterns across the different time intervals. For instance, compared to the 12:00 AM to 4:59 AM interval (the reference category), the model estimated that, on average, 1,884 more workers departed between 6:00 AM and 6:29 AM, and 100 more workers departed between 5:30 AM and 5:59 AM. Other time intervals, such as the 7:00 AM to 7:29 AM period, were not statistically different from the reference category, suggesting a lower or more stable number of workers departing during these times.
The findings have important implications for transportation planning and congestion management in metropolitan areas. For instance, the sharp increase in departures during the 6:00 AM to 6:29 AM interval suggests that targeted congestion management strategies, such as variable tolling, staggered work hours, or enhanced public transportation services during early morning hours, could alleviate peak load congestion. The insignificance of income in influencing departure times suggests that departure behavior may be more uniformly distributed across income levels, meaning that strategies aimed at reducing congestion should focus more on time-based demand management rather than socio-economic targeting.In future versions of this model, we aim to incorporate additional factors that may impact worker departure times, such as vehicle availability, commute distance, and access or proximity to public transit. We also plan to expand our analysis to include other metropolitan areas across the country.