Behavioral changes during the pandemic worsened income diversity of urban encounters

Takahiro Yabe, Bernardo García Bulle Bueno, Xiaowen Dong, Alex Pentland, Esteban Moro

TL;DR

While urban mobility volume in the US has largely returned to pre-pandemic levels, the income diversity of our physical encounters has not. This study reveals that the pandemic triggered a persistent 15% to 30% drop in the diversity of urban interactions, driven not just by staying at home, but by a fundamental shift in behavior: we have become less likely to explore new neighborhoods and more likely to stick to "socially familiar" bubbles.

What Problem Does This Paper Address?

Cities thrive on "social friction"—the random encounters between people of different backgrounds that spur innovation, economic productivity, and social resilience. When the COVID-19 pandemic hit, lockdowns and social distancing abruptly halted this mixing.

The authors sought to answer a critical question: As the world "reopened," did the social fabric of our cities mend itself? Specifically, they looked at income diversity, measuring whether people were still visiting places where they would encounter individuals from different socioeconomic strata. If diversity remains low even as mobility returns, it suggests a long-term "scarring" effect that could lead to increased urban segregation and diminished social capital.

Methods

The researchers employed a multi-layered analytical approach to disentangle why diversity changed:

  1. Diversity Metrics: They calculated two primary metrics: $D_\alpha$ (the income diversity of a specific Point of Interest) and $D_i$ (the diversity of encounters experienced by an individual). These measure the "evenness" of time spent by people from four different income quantiles.
  2. Counterfactual Simulations: To isolate the cause of the decline, they created "synthetic" 2019 datasets. By randomly down-sampling pre-pandemic data to match the reduced activity levels of 2020 and 2021, they could determine if the diversity loss was simply due to people going out less (volume) or because they were choosing different places to go (behavior).
  3. Social-EPR Model: They used a "Social Exploration and Preferential Return" model to quantify the probability of an individual visiting a new place where their income group is not the majority.
  4. Regression Analysis: They modeled the spatial heterogeneity of these changes against census-based variables (like age, race, and transit usage) to see which communities were most affected.

Datasets

The study relies on a massive, high-resolution longitudinal dataset:

  • Mobility Data: Privacy-enhanced GPS pings from over 1 million anonymized mobile phone users across four major US metropolitan areas (Boston, Seattle, Los Angeles, and Dallas) spanning 2019 to late 2021. Provided by Spectus.
  • Points-of-Interest (POIs): 433,000 verified places obtained via the Foursquare API, categorized into types like "Grocery," "Museums," and "Leisure."
  • Socioeconomic Data: US Census Bureau data at the Census Block Group (CBG) level was used to assign income proxies to users based on their home locations.
  • Validation Data: ReferenceUSA Business Historical Data was used to ensure the results were robust regardless of the POI dataset chosen.

Key Insights and Ideas

  • The Recovery Paradox: By late 2021, the number of daily visits and time spent at POIs had returned to 2019 levels. However, the diversity of those visits remained significantly lower. We are moving as much as we used to, but in more segregated patterns.
  • The "Exploration Deficit": The most striking finding from the Social-EPR model was a significant decrease in the social exploration parameter ($\sigma_s$). People became less willing to visit new places where they would be the "socioeconomic minority."
  • Volume vs. Behavior: In the early stages of the pandemic (April 2020), about 55% of the diversity loss was explained by the simple fact that people were staying home. By late 2021, however, behavioral changes (where we choose to go) became the dominant factor in the persistent diversity gap.
  • Category Sensitivity: Not all places are equal. Museums, leisure spots, and coffee shops saw the most drastic and persistent drops in diversity. Grocery stores were the most resilient, likely because they represent essential, localized utility rather than elective social exploration.

Key Results

  • Quantitative Drop: Income diversity of urban encounters decreased by 15% to 30% during the height of the pandemic.
  • Persistence: Even two years after the initial outbreak, diversity remained roughly 10% lower than pre-pandemic baselines, despite the recovery of aggregate mobility.
  • Demographic Impact: Areas with higher population density, a higher proportion of working-age populations (25–64), and a higher reliance on public transit experienced the largest decreases in encounter diversity.
  • Policy Correlation: There is a strong negative correlation ($\rho$ between -0.73 and -0.90) between the stringency of COVID-19 policies and urban diversity.

Limitations and Future Work

The authors acknowledge several limitations:

  • Proxy for Interaction: GPS co-location is a proxy for "meaningful encounters." Two people in the same grocery store might not speak, but the authors argue this still represents a "bound" for potential social mixing.
  • Demographic Scope: The study focuses primarily on income. Future work could apply these methods to racial, ethnic, or political diversity to see if similar "behavioral scarring" exists.
  • Purpose of Visit: The data doesn't capture the intent of the visit (e.g., a cocktail party vs. a quick errand), which influences the quality of the encounter.

Future Work: The authors suggest that urban policy should move beyond just restoring mobility volume. Interventions like fare-free transit or the development of inclusive public spaces should specifically target the recovery of "social exploration" to prevent long-term urban fragmentation.

Summary updated

Abstract

Diversity of physical encounters and social interactions in urban environments are known to spur economic productivity and innovation in cities, while also to foster social capital and resilience of communities. However, mobility restrictions during the pandemic have forced people to substantially reduce urban physical encounters, raising questions on the social implications of such behavioral changes. In this paper, we study how the income diversity of urban encounters have changed during different periods throughout the pandemic, using a large-scale, privacy-enhanced mobility dataset of more than one million anonymized mobile phone users in four large US cities, collected across three years spanning before and during the pandemic. We find that the diversity of urban encounters have substantially decreased (by 15% to 30%) during the pandemic and has persisted through late 2021, even though aggregated mobility metrics have recovered to pre-pandemic levels. Counterfactual analyses show that while the reduction of outside activities (higher rates of staying at home) was a major factor that contributed to decreased diversity in the early stages of the pandemic, behavioral changes including lower willingness to explore new places and changes in visitation preferences further worsened the long-term diversity of encounters. Our findings suggest that the pandemic could have long-lasting negative effects on urban income diversity, and provide implications for managing the trade-off between the stringency of COVID-19 policies and the diversity of urban encounters as we move beyond the pandemic.
BibTeX
@inproceedings{yabe2023behavioral,
  author = {Takahiro Yabe and Bernardo García Bulle Bueno and Xiaowen Dong and Alex Pentland and Esteban Moro},
  title = {{Behavioral changes during the pandemic worsened income diversity of urban encounters}},
  booktitle = {Nature Communications},
  year = {2023},
  doi = {10.1038/s41467-023-37913-y},
}

Loading related content...

Figure