Behavioral changes during the COVID-19 pandemic decreased income diversity of urban encounters

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

TL;DR

While urban mobility volume has largely returned to pre-pandemic levels, the diversity of our social mixing has not. This study of over one million mobile users reveals that the income diversity of urban encounters dropped by 15–30% during COVID-19 and remained significantly depressed through late 2021, driven primarily by a persistent decline in our willingness to explore new and socially diverse places.

What Problem Does This Paper Address?

Cities are often described as "engines of innovation," a reputation built on the dense, diverse social connections formed through physical encounters. High "income diversity"—the frequency with which people from different socioeconomic backgrounds cross paths—is a known predictor of economic growth and community resilience.

When COVID-19 hit, mobility restrictions didn't just keep us home; they tore the social fabric of these encounters. While many studies tracked how much we moved, this paper addresses the quality of that movement. The authors ask: Did the pandemic permanently change who we encounter in the city? And if mobility volume has recovered, why hasn't the diversity of our social mixing followed suit?

Methods

The researchers employed a multi-step analytical framework to quantify "experienced income diversity":

  1. Socioeconomic Proxying: Using home location data at the Census Block Group (CBG) level, they assigned users to one of four income quartiles based on median household income.
  2. Diversity Metrics: They calculated diversity ($D$) for both specific locations (POIs) and individual users. This metric measures the "evenness" of time spent with people from different income groups. A value of 1 represents a perfect 25/25/25/25 split, while 0 represents total homogeneity.
  3. Counterfactual Simulations: To isolate why diversity dropped, the team built "what-if" models. They simulated scenarios where they only changed total activity levels, only changed travel distances, or only changed "microscopic" preferences (like exploration behavior) to see which factor most closely matched the real-world decline.
  4. Social-EPR Model: They used a "Social Exploration and Preferential Return" model to quantify the probability ($\sigma_s$) of an individual visiting a new place where their own income group is the minority.

Datasets

  • Spectus Mobility Data: Anonymized, high-resolution GPS pings from over 1 million mobile phone users across four major US metropolitan areas (Boston, Dallas, Los Angeles, and Seattle) from 2019 to late 2021.
  • Foursquare API: A dataset of 433,000 verified Points-of-Interest (POIs) used to match GPS pings to specific venues (restaurants, museums, etc.).
  • American Community Survey (ACS) 2016–2020: Used to estimate the socioeconomic status of users based on their home census blocks.
  • OxCGRT (Oxford COVID-19 Government Response Tracker): Used to correlate policy stringency with changes in encounter diversity.

Key Insights and Ideas

  • The "Diversity Gap": By late 2021, metrics like the number of daily visits and unique POIs visited had returned to 2019 levels. However, income diversity remained ~10% lower. This suggests a "new normal" where we move as much as before, but within more homogeneous "echo chambers."
  • Exploration vs. Routine: The counterfactual analysis revealed that the initial drop in diversity was due to people staying home. However, the persistent loss of diversity is driven by a change in behavior: we are less willing to explore new places and are sticking to "routine" locations (like grocery stores) that tend to be more socioeconomically segregated.
  • Category Sensitivity: Not all places lost diversity equally. "Museums," "Leisure," and "Coffee" shops saw the most dramatic and persistent drops in income mixing. "Grocery" stores were the most resilient, likely because they remain a non-discretionary necessity for all income groups.
  • The Policy Trade-off: There is a stark, negative correlation ($\rho$ between -0.73 and -0.90) between the stringency of COVID-19 policies and encounter diversity. While restrictions saved lives, they had a measurable "side effect" of increasing urban segregation.

Key Results

  • Quantitative Drop: Experienced income diversity decreased by 15% to 30% during the height of the pandemic.
  • Persistent Segregation: Even as cities reopened, diversity stayed roughly 10% lower than pre-pandemic baselines through the end of the study period (late 2021).
  • Spatial Heterogeneity: Areas with higher population density, higher reliance on public transit, and a higher proportion of working-age populations (25–64) experienced the largest decreases in income diversity.
  • Behavioral Shift: The "social exploration" parameter ($\sigma_s$)—the likelihood of visiting a place where one is in the socioeconomic minority—dropped significantly and stayed low, explaining roughly 45% of the diversity loss that couldn't be accounted for by simple activity reduction.

Limitations and Future Work

The authors acknowledge that mobile pings are a proxy for "co-location," not necessarily "interaction." Being in the same coffee shop doesn't guarantee a meaningful social exchange. Furthermore, the study focuses exclusively on income diversity; future work is needed to explore how the pandemic affected racial or ethnic diversity in urban encounters.

The researchers suggest that urban planners should now focus on "re-diversifying" cities. This could include policies like fare-free transit or the development of public spaces specifically designed to draw people out of their socioeconomic bubbles and back into the "socially exploratory" behavior that makes cities vibrant.

Summary updated

Abstract

Diversity of physical encounters in urban environments is known to spur economic productivity while also fostering social capital. However, mobility restrictions during the pandemic have forced people to reduce urban encounters, raising questions about the social implications of behavioral changes. In this paper, we study how individual income diversity of urban encounters changed during the pandemic, using a large-scale, privacy-enhanced mobility dataset of more than one million anonymized mobile phone users in Boston, Dallas, Los Angeles, and Seattle, across three years spanning before and during the pandemic. We find that the diversity of urban encounters has 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 behavioral changes including lower willingness to explore new places further decreased the diversity of encounters in the long term. Our findings 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 COVID-19 pandemic decreased income diversity of urban encounters}},
  booktitle = {Nature Communications},
  year = {2023},
  doi = {10.1038/s41467-023-37913-y},
}
Figure