Economic resilience during COVID-19: the case of food retail businesses in Seattle, Washington

Summary

A Safe Start Washington phased reopening plan, issued by the Office of the Governor on October 7, 2020, under Governor Jay Inslee in collaboration with the Washington State Department of Health. It states that the plan establishes a data-driven approach to reopen Washington and modify physical distancing measures while minimizing the health impacts of COVID-19. It states that Washington will move through the phased reopening county by county, allowing flexibility and local control. It states that a county may apply to the secretary of Health to move to the next phase unless a freeze is in place.

Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used

Issuer: Source documents
Document type: Economic Resilience During Covid 19 The Case Of Food Retail Businesses In Seattle Washingt
Case: Economic Resilience During Covid 19 The Case Of Food Retail Businesses In Seattle Washington

Cite this document

“Economic resilience during COVID-19: the case of food retail businesses in Seattle, Washington.” Original at www.frontiersin.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/economic-resilience-during-covid-19-the-case-of-food-retail-businesses-in-seattle-washington/, SHA-256 295f9603ca5756bdc65c4a7419370d10d7c3479ed9bc056bf289fd0ddf44d731.

Full text

TYPE Original Research
PUBLISHED 28 September 2023
DOI 10.3389/fbuil.2023.1212244

Economic resilience during
OPEN ACCESS COVID-19: the case of food retail
EDITED BY
Wenjia Zhang,
Peking University, China
businesses in Seattle, Washington
REVIEWED BY
Maxim A. Dulebenets, Feiyang Sun 1†, Jan Whittington 2*†, Siman Ning 2†,
Florida Agricultural and Mechanical
University, United States
Gundula Proksch 3, Qing Shen 2 and Soﬁa Dermisi 4
Bor Tsong Teh, 1
Department of Urban Studies and Planning, University of California, San Diego, San Diego, CA,
University of Malaya, Malaysia United States, 2Department of Urban Design and Planning, University of Washington, Seattle, WA,
*CORRESPONDENCE United States, 3Department of Architecture, University of Washington, Seattle, WA, United States,
4
Jan Whittington, Department of Real Estate, University of Washington, Seattle, WA, United States
janwhit@uw.edu
†
These authors have contributed equally
to this work and share ﬁrst authorship

RECEIVED 26 April 2023
The ﬁrst year of COVID-19 tested the economic resilience of cities, calling into
ACCEPTED 18 August 2023 question the viability of density and the essential nature of certain types of services.
PUBLISHED 28 September 2023 This study examines built environment and socio-economic factors associated
CITATION with the closure of customer-facing food businesses across urban areas of Seattle,
Sun F, Whittington J, Ning S, Proksch G, Washington. The study covers 16 neighborhoods (44 census block groups), with
Shen Q and Dermisi S (2023), Economic
resilience during COVID-19: the case of two ﬁeld audits of businesses included in cross-sectional studies conducted
food retail businesses in during the peak periods of the pandemic in 2020. Variables describing
Seattle, Washington. businesses and their built environments were selected and classiﬁed using
Front. Built Environ. 9:1212244.
doi: 10.3389/fbuil.2023.1212244 regression tree methods, with relationships to business continuity estimated in
a binomial regression model, using business type and neighborhood socio-
COPYRIGHT
© 2023 Sun, Whittington, Ning, Proksch, demographic characteristics as controlled covariates. Results show that the
Shen and Dermisi. This is an open-access economic impact of the pandemic was not evenly distributed across the built
article distributed under the terms of the
environment. Compared to grocery stores, the odds of a restaurant staying open
Creative Commons Attribution License
(CC BY). The use, distribution or during May and June were 24%, only improving 10% by the end of 2020. Density
reproduction in other forums is played a role in business closure, though this role differed over time. In May and
permitted, provided the original author(s)
June, food retail businesses were 82% less likely to remain open if located within a
and the copyright owner(s) are credited
and that the original publication in this quarter-mile radius of the ofﬁce-rich areas of the city, where pre-pandemic job
journal is cited, in accordance with density was greater than 95 per acre. In November and December, food retail
accepted academic practice. No use,
businesses were 66% less likely to remain open if located in areas of residential
distribution or reproduction is permitted
which does not comply with these terms. density greater than 23.6 persons per acre. In contrast, median household income
and percentage of non-Asian persons of color were positively and signiﬁcantly
associated with business continuity. Altogether, these ﬁndings provide more
detailed and accurate proﬁles of food retail businesses and a more complete
impression of the spatial heterogeneity of urban economic resilience during the
pandemic, with implications for future urban planning and real estate
development in the post-pandemic era.

KEYWORDS

economic resilience, pandemic, urban design, neighborhood, food retail business, built
environment, COVID-19

1 Introduction
In January 2020, the US began mobilizing to combat COVID-19 (Coronavirus Infectious
Disease-2019), a disease caused by the SARS-CoV-2 virus (Severe Acute Respiratory
Syndrome, Coronavirus, Version 2) (Covid Crisis Group, 2023). The ﬁrst travel-related
case of the disease in the US was conﬁrmed by the US Centers for Disease Control and

Frontiers in Built Environment 01 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

Prevention (CDC) on January 21, in the Seattle area of the state of neighborhoods are a key indicator of economic vitality and
Washington (CDC, 2020). Declared a global pandemic on March 11, resiliency (Carré and Tilly, 2008; Sutton, 2010; Mehta, 2011), yet
COVID-19 was on track to become the deadliest pandemic the US recognized as being particularly vulnerable to crisis (Bartik et al.,
had experienced in over 100 years (WHO, 2020). As travel bans 2020; Fairlie and Fossen, 2022). As the outbreak ensued, people
failed to contain the outbreak, public health policies in the US aimed began to avoid large gatherings and crowds, and to favor business
to suppress transmission of the virus, while increasing the capacity types labeled “essential” (Goolsbee and Syverson, 2021). Supply
of the healthcare system for testing, quarantine, treatment, and the chain studies describe widespread shifts in demand for retail goods
development and dissemination of vaccines. and services that emerged due to the spread of the virus, changing
The scale of disruption was unprecedented. The Dow Jones lost consumer behavior and preferences, new stay-at-home and work-
37% of its value between February 12 and March 23, surpassing the from-home policies, as well as disruptions to supply (Pujawan and
1929 losses of the Great Depression (Frazier, 2021). By April, stay- Bah, 2021; Swanson and Santamaria, 2021; Thilmany et al., 2021;
at-home orders had been issued in 49 US states (Covid Crisis Group, Moosavi et al., 2022). Prior to COVID-19, retail businesses and
2023, chap. 6). From March to September, more than 60 million markets favored urban density, but customer preferences changed
claims for unemployment insurance were ﬁled (Cutler and rapidly during the shock of lockdown; customer trafﬁc relocated
Summers, 2020). Economic projections, assuming the disease with the work-from-home movement, and retail managers began to
would be contained by fall of 2021, forecasted losses to GDP of re-evaluate their asset management strategies (Balemi et al., 2021;
$7.6 trillion, with another $8.6 trillion in losses of productivity Nanda et al., 2021; Gupta et al., 2022). Many studies regarding the
ascribed to a forecast of 625,000 deaths and a reduced quality of life built environment have focused on analyzing the role of density in
for millions more. The ﬁrst deadly wave of disease in the US peaked the spread of COVID-19, but the impact of density is not
in April 2020, but as restrictions were relaxed and communities straightforward (Alidadi and Shariﬁ, 2022). The food retail
opened up, a second wave rose further than the ﬁrst, and peaked in industry, with its sensitivity to shifts in demand, its variegated
the early days of January 2021 (IHME, 2023). As the crisis wore on, pandemic policy environment, and its many business types,
many state policies became more focused, targeting high-contact building forms, and neighborhood settings, could offer insights
businesses, imposing occupancy restrictions, and limiting large into the complex relationship between urban density and the
gatherings (Covid Crisis Group, 2023). economic impact of the pandemic.
Though COVID-19 no longer bears the status of a public health This paper examines the effects of makeshift adaptations, the
emergency in the US, the scale of the pandemic and the nature of its staid characteristics of business and building type, and the larger
differential impact deserve scrutiny as public health problems we are scale characteristics of urban density and socio-demographic
likely to revisit in the future, and as economic problems derived in environments within and between neighborhoods, on food
part from our built environments. The scale of the crisis was greater retail business continuity as the ﬁrst two waves of the
than most projections. By the time the declaration of emergency was pandemic crossed the city of Seattle, Washington (Proksch
lifted in the US, deaths surpassed 1.1 million and hospitalizations et al., 2021). The built environment is an important, ever-
approached 6.2 million (US Department of Health and Human evolving determinant of health and wellbeing (Acuto, 2020;
Services, 2023). The federal government deployed over $5 trillion to Rice, 2020) deserving research to improve economic resilience
combat the pandemic and offer relief—an amount greater than the in harmony with pandemic health policy (Honey-Roses et al.,
entire 2019 federal budget (Covid Crisis Group, 2023). Early 2020; Megahed and Ghoneim, 2020; Asif et al., 2022). The
macroscale models framed economic impacts in relation to the operational and built environment changes to food retail
health beneﬁts of physical distancing, targeted lockdowns, and stay- establishments in the wake of the outbreak were natural
at-home orders (Acemoglu et al., 2020; Alvarez et al., 2020; Kruse experiments, observable on the streets of cities. Investigation
and Strack, 2020; Newbold et al., 2020; Thunström et al., 2020; of these changes and their effects can provide insight into the
Toxvaerd, 2020; Eichenbaum et al., 2021; Farboodi et al., 2021). multiscale characteristics of the built environment that support
Microscale economic consequences are not as well understood; the economic resilience of neighborhoods.
businesses experienced wide variations in economic impact from
the pandemic (Goolsbee and Syverson, 2021; Klein and Smith, 2021;
Monte, 2021; Autor et al., 2022). Cities were struck by the hollowing 2 Literature review
out of ofﬁce-rich downtown cores and urban ﬂight—shifts in the
populations of ofﬁce workers and residents that comprise the 2.1 Food retail resilience during COVID-19
customer base of metropolitan central business districts (Liu and
Su, 2021; Ramani and Bloom, 2021; Rosenthal et al., 2022; Hutson As the onset of COVID-19 became apparent in the US, food
and Orlando, 2023). Suddenly, the characteristics of the built retail businesses were quickly categorized in government policies as
environment, immutable due to their inordinate agglomerated essential, to maintain access to grocery stores and prepared foods
expenses and relied upon for economic performance, such as through take-out and delivery services while suppressing
building design and urban density, were up for reconsideration transmission of the virus by limiting access to in-person activities
by policymakers, academics, business owners, and consumers alike. (Goolsbee and Syverson, 2021). Whether due to fear or policy,
Some of the most visible effects of the pandemic were witnessed people reacted by adjusting their social and transportation activities
in the food retail industry, as businesses struggled, regrouped, and (e.g., Gao et al., 2021), showing a preference for essential food
adapted with makeshift efforts to repurpose the built environment businesses such as grocery stores over restaurants, taverns, and pubs
(Proksch et al., 2021). The size and types of food retail in a city’s (Goolsbee and Syverson, 2021).

Frontiers in Built Environment 02 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

The US food retail industry suffered signiﬁcant losses in business attenuated demand from customers and impacts to supplies due to
continuity. After healthcare, food services were poised to be the lockdowns and barriers to logistics. Notably, supply chain studies
sector most threatened by the pandemic (Dua et al., 2020). In an have forecasted and drawn attention to spikes in demand such as
early nationwide survey of small businesses, Bartik et al. (2020) panic buying (Yuan et al., 2021), and spikes in demand for
found 45% of the sample either temporarily or permanently closed healthcare services (Govindan et al., 2020) as well as declines,
on 1 April 2020. Balla-Elliott et al. (2020, p. 17) found 34% of small disruptions, and plunges in demand in other sectors, including
businesses surveyed to be either temporarily or permanently closed, food services (Akseer et al., 2020). The bibliometric literature
with another 34% operating at less than full capacity (i.e., “partially review from Swanson and Santamaria (2021) on the topics of
open”). Two-thirds of closed businesses involved workers operating “pandemic” and “supply chain” reaches back to 1997, pulling
in close proximity to one another, as found in food retail businesses together analyses of inﬂuenza as well as COVID. Authors
such as restaurants and cafes. A study of US points of interest found emphasize the fundamental differences between the disruptions
food services, retail and trade to be signiﬁcantly more likely to be posed by pandemics in comparison to other types of events and
closed than recreational and entertainment businesses (Zhai and suggest how limited current knowledge is of these “black swan”
Yue, 2022). Rates of closure for food establishments in New York events, while offering advice on further research and modeling.
City’s Chinatown were high enough to prompt researchers to shift Recommendations from these studies to improve resilience include
their focus to questions of food security (Russo et al., 2022). ﬂexibility and redundancy (Moosavi et al., 2022), and elevating the
Researchers suggest that customers were making safety their ﬁrst importance of safety, digitalization, and supply-base localization
priority (Gursoy and Chi, 2020; Yost and Cheng, 2021). Surveys of (Pujawan and Bah, 2021). Many of the same themes exist in
customer concerns in the US ranked sit-down restaurants the most empirical studies focused on food retail performance in the US.
concerning type of retail business for COVID-19 transmission Though many establishments closed, many also responded in a
(McFadden et al., 2021). People’s impressions in the US of the variety of ways to the restrictions and to the concerns of customers,
safety of food establishments have been described as differing in which persisted even as restrictions were lifted (Goolsbee and
relation to media reports, concern about the crowdedness of Syverson, 2021). Grocery stores marked sidewalks, controlled
establishments, and according to their own risk factors for occupancy to store interiors, made disinfectant available, and
COVID (Byrd et al., 2021; Wang et al., 2021). It should be noted invested in contactless point-of-sale equipment; some also
that research intended to explain perceptions of risk can be quite reconﬁgured uses of space to promote drop off and pick up
complex, and institutional environments may differ in ways that service (Proksch et al., 2021). Restaurateurs and cafe managers
strain comparison within and between markets (e.g., Siegrist et al., adopted highly visible, customer-facing hygiene and health
2021; Yuan et al., 2021). Public impressions are likely to differ, for protection measures, changed operating hours and employee
example, in relation to the effectiveness of educational campaigns roles and responsibilities, offered new or limited menu items, and
about public health requirements and strategies (Khorram-Manesh expanded communications with their customers through signage
et al., 2021). and online sites (Norris et al., 2021; Proksch et al., 2021; Russo et al.,
Inﬂuenced at least in part by regulatory requirements, public 2022; Yenerall et al., 2022; Bonfanti et al., 2023; Mandhan and
preferences in the US have been shown to contribute to a shift in foot Gregg, 2023). Within their existing building spaces and public
trafﬁc from in-person dining to grocery stores (Goolsbee and rights-of-way, some businesses expanded or re-conﬁgured their
Syverson, 2021). Data from the US Department of Agriculture’s use of space for indoor storage, exterior-facing contactless
Economic Research Service showed inﬂation-adjusted expenditures payment and take out, drive-thru services, and outdoor dining
at grocery stores, supercenters, convenience stores, and other experiences (Proksch et al., 2021; Mandhan and Gregg, 2023).
retailers (food at home) up 19% in March 2020 compared to the These makeshift adaptations to the built environment were
same period in 2019, and food expenditures from restaurants and accompanied by online and ﬁnancial efforts. Digitalization
other eating-out places down by 29% during the same period (Dong expanded and took many forms, such as web sites, online
and Zeballos, 2021). During the second quarter of 2020, the decline ordering, options for customer pick up, store-run delivery
in expenditures at restaurants and similar establishments continued, services, and third-party apps for ordering and/or delivery
at 38% below expenditures in 2019. In other words, the impacts of (Brewer and Sebby, 2021; Norris et al., 2021; Proksch et al.,
COVID-19 were not equally distributed across the types of 2021). Many also sought ﬁnancial support, through the Paycheck
businesses comprising the food retail industry, nor would they Protection Program and a variety of other instruments, such as the
likely be equally distributed within a given urban area. COVID-19 Economic Injury Disaster Loan Program (Autor et al.,
Studies on disruptions to supply chains often include 2022; Barraza et al., 2020; U.S. Small Business Administration,
examination of impacts to retail demand and can add depth to 2023).
interpretations of the impact of the pandemic on business continuity
(Pujawan and Bah, 2021; Swanson and Santamaria, 2021; Moosavi
et al., 2022). Moosavi and colleagues (2022), in a literature review, 2.2 Questioning the role of density in
bibliometric analysis, and citation analysis on supply chain COVID-19
management and COVID-19, found “food security” and “food
supply chain” to be the fourth and ﬁfth top keywords in search There have been many studies suggesting a positive relationship
results. The most highly cited of the articles, by Ivanov (2020) and between population density and the incidence of COVID
Ivanov and Das (2020), featured a simulation of the ripple effects of (Ganasegeran et al., 2021; Ilardi et al., 2021; Lee et al., 2021;
COVID-19 in supply chains, including the simultaneous impacts of Qeadan et al., 2021; Sarmadi et al., 2021; Tchicaya et al., 2021;

Frontiers in Built Environment 03 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

TABLE 1 Density and COVID risk: summary of literature.

Perspective Author Dependent Density measure Research methodology, Setting
outcome relationships, and key
conclusionsa
Economics/business Balla-Elliott operational status, small logarithm transformed Regression Analysis US nationwide survey of small
management et al. (2020) business reopen population density at businesses
decision county level - fully open status (+)

- temporarily closed status (−)

- expected months to reopen (−)

- estimated months until restrictions
are lifted (−)

Economics/business Sedov (2022) restaurant closures establishment density, Regression Analysis 128,285 restaurants in 42 major US
management number of cities from Yelp and Safegraph
establishments in 500 m - establishment density (+)
buffer
- distance to city center (+)

Economics/business Wang et al. Restaurant dine-in visit logarithm transformed Two-stage regression analysis Visit counts Jul-December
management (2022a) population density at zip 2020 January 2021, and May 2021
code level - population density and social index
(+), but not signiﬁcant

- pre-vaccine: social index, dine-
in (−)

- post-vaccine: social index,
vaccination rate, dine-in (+)

- restaurants located in city centers,
high social indices initially (−), post-
vaccine (+)

Economics/business Jun et al., 2022 business survival until population density Regression analysis 340,000 businesses in Seoul from
management June 2021 2017 to 2021
- population density (+)

- total number of employees at the
dong level (similar scale to a
municipality in the United States of
America) (−)

- the number of shops provide same
amenity (−)

- number of other stores that provide
related amenities in the same
cluster (+)

Real estate Allan et al. log (rent) commercial population density Regression analysis market rents and capital values on
(2021) capital values 38 cities in 12 countries
- commercial real estate rent (+, ns) Asia–Paciﬁc region

- commercial capital values (+)

Real estate Liu and Su real estate inventory, population density Regression analysis ZIP code, 25 large MSAs in the U.S.
(2021) home price index, rent
- real estate inventory during
COVID (−)

- home price index during COVID
(-, ns)

- rent (−)

Real estate Rosenthal et al. ln (commercial rent per employment density Regression analysis ofﬁce and retail leases signed
(2022) square foot on an between January 2019 and October
annual basis) - Ofﬁce rent (−) 2020 in 89 distinct U.S. urban areas

- Retail rent (-, ns)

- New tenant lease (−)

- Renewal lease (−)
a
s means signiﬁcance, ns means no signiﬁcance in the regression results.

Frontiers in Built Environment 04 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

Wong and Li, 2020; Jamal et al., 2022), and many studies showing a place by government when contemplating the prospects for
negative or insigniﬁcant correlation (Hamidi et al., 2020; reopening business, while restrictions, in turn, correlated with the
Klompmaker et al., 2021; Perone, 2021; Basellini and Camarda, prevalence of COVID-19 in the population, as well as density at the
2022). These differences in ﬁndings exist on a variety of scales, from county level. In a study speciﬁc to restaurant closures in major US
the state (Ganasegeran et al., 2021; Ilardi et al., 2021; Perone, 2021; cities, Sedov (2022) found a positive relationship between closure
Sarmadi et al., 2021; Tchicaya et al., 2021; Basellini and Camarda, and the density of establishments within a 500-m radius, as well as
2022), to county (Hamidi et al., 2020; Wong and Li, 2020; proximity to the city center. However, the study relied on “Yelp”
Klompmaker et al., 2021; Lee et al., 2021, p. 202; Qeadan et al., records for indications of business closure. Wang et al. (2022b), in an
2021; Jamal et al., 2022), to more local levels (Tribby and Hartmann, intriguing study of data regarding visits to the 100 largest restaurant
2021; Almagro and Orane-Hutchinson, 2022; Laefer and chains in the US from July 2020 to May 2021, found a 1% increase in
Protopapas, 2022; Sun et al., 2022). Findings suggest the effects COVID cases associated with a 2.5% decrease in restaurant visits in
of density have differed over time as well (Lee et al., 2021; Carozzi the period prior to the availability of vaccines. Stepping outside of
et al., 2022; Wu et al., 2022; Tsou et al., 2023). Some differences in the US, it is important to note the paper by Jun et al. (2022),
outcomes may be traced to differences in research methods, such as examining effects at the neighborhood level on business closure
cluster analysis (Fan et al., 2022; Wu et al., 2022), the application of across the city of Seoul, Korea, including the number of businesses,
indexes built from multiple variables (Rahman et al., 2023), and the employment density, population density, and other demographic
use of advanced statistical methods to control for endogeneity of and economic variables, such as income. Business survival was
density measures (Carozzi et al., 2022; Hong and Chakrabarti, 2022). negatively associated with the number of businesses and
Other differences may result from different control variables used in employment density. Furthermore, the study clustered businesses
models (Teller, 2021; Zhang et al., 2022), and different cultural, according to their amenities, ﬁnding that though the replication of
policy and institutional contexts (Hasnat, 2022). That said, socio- amenities across stores in proximity to one another was a common
economic factors (e.g., income, ratio of minority group) are strategy prior to the pandemic, this feature worked against stores
commonly considered across many studies, with beneﬁts for their during the pandemic.
ﬁndings and interpretation (Benita et al., 2022; Carella et al., 2022; Studies focused on business closure and business characteristics
Choi and Denice, 2022; Huang et al., 2020; Kang et al., 2020; Loomba run the risk of missing the effect of greater movements of the
et al., 2021; Paul et al., 2021; Rahman et al., 2023; Schmiege et al., population deﬁned or indicated by shifts in markets for housing,
2023; Tang et al., 2021; Tsou et al., 2023; Verma et al., 2021; Wang ofﬁce, and retail real estate. Note, in this context, Ramani and
et al., 2021). Bloom’s (2021) description of the hollowing-out of the
Population is not the only meaningful measure of density for population from the downtown core, either to work-from-home
understanding the impact of COVID-19. Studies that use less in residential suburbs or to take ﬂight to more distanced residence.
common sources of data, such as mobility traces (Liu et al., The advent of work-from-home policies and urban ﬂight—in
2021) or records of COVID transmission at scales smaller than addition to previously mentioned public concerns, policies, and
the county, provide the basis for more promising hypotheses about shifting preferences—change the foundations for understanding the
the role of density in COVID transmission. These studies suggest effects of the pandemic on urban markets for retail, including the
that connectivity (Jo et al., 2021; Almagro and Orane-Hutchinson, survival and closure of food retail stores and services. COVID-19, for
2022; Carozzi et al., 2022), customer density (Laefer and Protopapas, example, did lead to a strong shift in housing demand in the US,
2022), crowding (Lee et al., 2021; Almagro and Orane-Hutchinson, especially within densely populated neighborhoods and central cities
2022; Schmiege et al., 2023) or contact density (Verma et al., 2021) (Liu and Su, 2021). Research suggests that work-from-home policies
might be more meaningful indicators. opened the possibility for people to relocate outside of densely
The most common question asked in these studies is one that populated areas; a possibility incentivized by the relatively high cost
can be reasonably inferred from the airborne transmission of the of housing in dense urban areas in addition to concern for exposure
virus, which is the effect of population density, alone or in to COVID. Relatedly, an examination of the changing value of
connection with socio-spatial connectivity and socio- commercial rent across 89 US urban areas found steeper declines,
demographics, on rates of transmission, incidence of disease, and compared to pre-COVID times, in the value of commercial real
mortality (Dietz et al., 2020; Liu et al., 2021). This same question estate in metropolitan city centers that are characterized by enough
drives urban planners, architects, and engineers to draw normative density to support vibrant transit systems, such as subways,
conclusions for improving urban design, by integrating the aims of commuter rail, and bus rapid transit (Rosenthal et al., 2022).
design with health (Megahed and Ghoneim, 2020). While COVID
transmission is important to understand and predict for its spatial
distribution, research on COVID transmission should not be 2.3 Research question and contribution
equated to research on business continuity. Fewer studies attempt
to distinguish between density, the incidence of COVID-19, and The purpose of this research is to help distinguish the role of the
business continuity, though all three factors may be related. built environment in mediating pandemic disturbance to business
To understand the relationship between business continuity and continuity. As a disruption in otherwise slowly shifting urban
the built environment during COVID-19, reviews of the literature economic environments, the pandemic occurred in waves,
should take a broader perspective, reaching business, economic, and coupled with public health policies designed to suppress
real estate literature (Table 1). Balla-Elliott et al. (2020), for example, transmission of the virus, prior to the dissemination of vaccines.
ﬁnds that small businesses took cues from the restrictions put in Policies were intended to be applied uniformly across urban settings,

Frontiers in Built Environment 05 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

yet disruptions to business continuity (i.e., risk to economic food retail businesses. The unit of analysis is the business (per
resilience) were not uniformly distributed. What explains spatial location, not per organization), inclusive of its surrounding
variation in business continuity? This study is designed to improve environment. It is a repeated cross-sectional study of the full
understanding of the role of the built environment in the spatial population of food retail businesses (n = 921) in a purposive
heterogeneity of business continuity during the ﬁrst year of the sample of 16 neighborhoods (44 census block groups) across the
pandemic, using the food retail sector as a proxy for the urban metropolitan city of Seattle. As a matter of interest to policymakers,
economy. this research design is selected to provide for external as much as
The target construct for this study views COVID-19 and internal validity (e.g., minimizing individual unit and contextual
subsequent policies and behaviors as one bundled source of selection bias). The sampling of neighborhoods (i.e., collections of
disruption in an otherwise functioning urban economy. The census block groups) is designed to encompass the business districts
built environment mediates such disruptions and contributes to of neighborhoods (containing the full complement of food retail
the spatial heterogeneity of food retail business continuity in business types and building types within neighborhoods), to
several plausible ways. The built environment is implicated at maintain representativeness with the city at large for controlled
the scale of the business location because policies and covariates (e.g., socio-demographics and economics), to maximize
behavioral preferences differed in terms of type of business the variance of independent variables (e.g., population density,
operations and in terms of the uses of indoor and outdoor space employment density, land use mix), and to rule out alternative
(e.g., Gursoy and Chi, 2020; Goolsbee and Syverson, 2021; Yost plausible explanations of business closure at the scale of the census
and Cheng, 2021). Built environments, at a more aggregate block group and/or neighborhood (e.g., protests, educational pivots
scale, differed for urban density and demographic compositions to remote learning). The data representing many variables pre-dates
of the customer base of businesses, and for qualities that the pandemic, giving research design the characteristics of a pre- and
manifested through the adoption of stay-at-home and work- post-test, with post-tests carried out during each of the ﬁrst two
from-home policies, and behaviors such as urban ﬂight (e.g., waves of pandemic (prior to the dissemination of vaccines).
Swanson and Santamaria, 2021; Thilmany et al., 2021). The aim As illustrated in Figure 1, this study examines factors
of this research is to untangle conﬂicting explanations of the hypothesized at the business and neighborhood level to
relationship between density and business continuity in the inﬂuence pandemic food retail business continuity, applying
literature, and to understand the effects of actions the new data and methodologies to examine effects that remain
businesses took themselves to maintain continuity, which either mixed or unclear in the literature. This study
include makeshift changes to the operation and the site of contributes to current literature in several ways. First, as
businesses. desirable as ﬁeld research might have been, we are unaware of
Besides the direct mediating effects of built environments, any other studies that were able to provide in-person
spatially heterogeneous disruptions can also be observed due to documentation of ground-level details of built environment
associations between built environments and other mediators, changes during the ﬁrst wave of COVID-19 (Megahed and
with ﬁnancial relief (loans, grants, rent relief) and the Ghoneim, 2020; e.g.; Russo et al., 2022; Liu et al., 2021). Thus,
digitalization of services (websites for ordering, payment, this study adds to the existing body of literature by generating
pickup and delivery services) being the most commonly more detailed and accurate proﬁles of food retail businesses in
discussed (Barraza et al., 2020; Brewer and Sebby, 2021; situ and in relation to business continuity during this critical time
Norris et al., 2021; Autor et al., 2022). Though these policies period (Proksch et al., 2021). Secondly, this study compares
and approaches were designed to be applied uniformly across different types of urban density measures, quantifying
urban settings, they could have had heterogeneous spatial effects thresholds of urban population and employment density, plus
as they were adopted by different businesses and in different land use mix in relation to business continuity. By examining
neighborhoods. Due to the difﬁculty of collecting business-level density measures besides population density in a repeated post-
ﬁnancial relief and digitalization information at the city scale, test, the study helps provide clarity to conﬂicting interpretations
this study design does not allow us to distinguish the direct effect of the role of urban density found in the literature.
of built environments and the indirect effect through associations This study also implements a novel methodological approach
with these other mediators. However, this study discerns the that combines the advantage of variable selection and cluster
latent vulnerabilities—either vulnerability as a lack of resilience detection of a classiﬁcation and regression tree analysis with the
to disruptions or vulnerability as ineffectiveness in adopting convenience of interpretation of a regression model. This is a useful
countermeasures—in urban built environments, and the methodology for studies that attempt to model correlated urban
capacity to adapt at the pace of widescale waves of disruption. design and built environment factors. The methodology, combined
Follow-on research could, for example, combine the independent with domain knowledge, is intended to provide a causal discovery
variables in this study with evidence of ﬁnancial relief and approach to model nonrandomly implemented urban design and
digitalization, apply the same independent variables in an built environment interventions collected from observational data
examination of business continuity in urban areas that were (Leist et al., 2022). However, it is essential to acknowledge that the
subject to different sets of public health policies than Seattle, set of variables used in this study is not exhaustive, leaving room for
or seek to generalize results in suburban and rural settings. potential confounding effects. As a result, while our results
As an evaluation of a natural experiment, this research is demonstrate statistical associations, we cannot deﬁnitively
designed to identify associations between selected independent conﬁrm causation due to the inherent limitations of
variables and the dependent variable of open or closed status of observational data.

Frontiers in Built Environment 06 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

FIGURE 1
Conceptual framework. The business level linkages describe makeshift business-side changes and factors such as business type and building type.
The neighborhood level linkages describe consumer-side changes in the built environment, and with factors that include population density,
employment density, and land use mixture, as well as neighborhood demographics. The highlighted boxes are the main interests of this study.

3 Data and methodology On March 23, Governor Inslee announced the “Stay Home, Stay
Healthy” order (Washington Governor, 2020b), requiring
3.1 Study setting Washingtonians to stay home except for essential activities,
asking people to stay at least six feet from one another when
The city of Seattle is the center of a metropolitan area within the outside, and closing all businesses except those deemed essential,
Puget Sound region of the state of Washington. The city contains the such as grocery stores. This was followed by a series of
largest urban population in the state and is home to several proclamations deﬁning a phased, scientiﬁc approach to reopening
noteworthy companies in information technology, biotechnology, (Washington Governor, 2020c). As of June 1, measures such as
and health sciences. The greater Seattle area of King County has trends in hospitalization were to inform decisions for each County to
been the subject of several studies of the effects of COVID-19 (e.g., move through phases of a reopening plan. Physical distancing and
Liu et al., 2021). Its well-coordinated corporate and civic culture and the use of personal protective equipment were to be maintained
high proportion of the workforce in technology are credited with the (among other requirements), while occupancy rates could be
city having 50% of the population work from home during the gradually increased in each phase of reopening. On June 3, with
pandemic (Balk, 2022). The ﬁrst case of a US patient testing positive known infection rates in decline, King County, including Seattle,
for the SARS-CoV-2 virus occurred on 21 January 2020, in a suburb placed an application for a “modiﬁed” phase 1 opening which, as
of Seattle, Washington (Gallagher, 2020). The following section approved on June 5, allowed retail stores to open at 30% or less of
describes the policy environment that took shape as the pandemic maximum occupancy and for restaurants and bars to open for 25%
arrived and this study began. of indoor and 50% of outdoor occupancy (King County, 2020b; King
On 29 February 2020, Governor Inslee issued a State of County, 2020c). By June 19, King County advanced to phase 2 of the
Emergency for Washington State and King County, where Seattle State’s reopening plan, allowing 50% indoor occupancy of
is located (Washington Governor, 2020a). Public activities were restaurants, bars, and similar establishments (King County, 2020d).
gradually curtailed until March 15, when King County ordered all Witnessing a resurgence of the virus, on 16 Nov 2020, Governor
gatherings of 50 or more persons to cease and, particular to the food Inslee once again ordered all restaurants, cafes, and such to eliminate
retail industry, declared (King County, 2020a). in-person dining, resorting to take out and delivery services for a
period of 4 weeks (Washington Governor, 2020d). After a surge of
• Restaurants, bars, dance halls, clubs, theaters, health and cases over the holiday season, Governor Inslee announced a new
ﬁtness clubs, and other similar indoor social or recreational “Healthy Washington” plan on January 5th, to bring the state
venues must cease operations until 31 March 2020 toward reopening in 2021 (Washington Governor, 2021).
• Restaurants and food service establishments may remain open
only for drive-through, delivery, and pick-up only, until
31 March 2020 3.2 Data collection
• All other retail such as groceries, pharmacies, banks, gas
stations, hardware stores, shopping centers, etc. may From May to June 2020, the research team conducted a three-
remain open provided they meet Public Health directives in part protocol to collect and validate data on the population of food
the previous Local Health Order retail establishments in a selection of 44 census block groups,

Frontiers in Built Environment 07 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

FIGURE 2
Operation Status of Food Retail Businesses in the Sampled Neighborhood Areas in Seattle, Washington. The neighborhoods were selected to cover
the majority of Urban Centers and Hub Urban Villages identiﬁed in Seattle’s comprehensive plan as key urban growth centers. Field observations, followed
by phone calls, identiﬁed and conﬁrmed the operation status of food retail establishments (n = 921) across 16 neighborhoods (44 census block groups)
over two peak pre-vaccination time periods of COVID-19.

comprising 16 neighborhoods, geographically distributed validation of the number, identiﬁcation, location, visible indoor and
throughout the city of Seattle (Figure 2) (Proksch et al., 2021). outdoor built features and makeshift changes, and status of the
The neighborhoods were purposely selected to cover Seattle’s Urban businesses in the selected block groups. To alleviate any doubt about
Centers and Hub Urban Villages, identiﬁed in the comprehensive open or closed status, phone calls were made to all businesses. There
plan as centers driving local and regional growth. The sampled were 921 food retail businesses found in the sample of
neighborhoods were demographically similar to the average neighborhoods. This is the total population of the 44 census
characteristics of the city, while also indicative of the variety of block groups selected for study, which amounts to 17% of all
socio-economic and demographic characteristics of its food retail businesses in the city of Seattle. The research team
neighborhoods. Parts of downtown Seattle and the Capitol Hill conducted a follow-up ﬁeld and phone survey to obtain updated
neighborhood (northeast to downtown Seattle) were excluded to information on business operations and closure between November
avoid the introduction of confounding variables caused by protests and December 2020. To maintain the consistency of ﬁeld
and riots during that time. The area in the vicinity of the University observation in the two time periods, the team maintained the
of Washington campus was also excluded from the sample, as the Google form to guide the collection of information and were
shift to remote instruction and physical absence of the student body instructed to follow the same protocol to provide uniform photo
created its own signiﬁcant impacts to food retail establishments. and phone documentation of each visited business.
Data collection began with King County food permit data from The research team also conducted structured interviews with a
2018, joined in May 2020, with Google Knowledge Panel data to limited set of food businesses about their ﬁnancial situation and
append to and update the listing of businesses, along with census digitalization activities from November to December 2020, as a
data and zoning information. Then, from May to mid-June of 2020, supplement to the ﬁeld surveys. In total there were 82 food
the team conducted ﬁeld surveys administered with a Google Form businesses interviewed, distributed among 27 census block
and a guide for photographic documentation to provide in-person groups. The interview asked about business ownership, rent

Frontiers in Built Environment 08 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

relief, and ﬁnancial relief, and changes in comparison to 2019 in Supplementary Table S1, were processed using QGIS 3.16.16 and
hours of operation, usage of online delivery, percentage of pick-up R 4.2.3.
orders, levels of employment, labor costs, food costs and revenues.

3.4 Methodology
3.3 Data processing
The modeling ﬂow is illustrated in Figure 3. The study uses
The outcome variable of interest is open or closed status per classiﬁcation and regression tree (CART) analysis to ﬁrst select and
business in each time period of study (May through June and identify possible combinations and thresholds that are most
November through December 2020). Open or closed status has important to business continuity. We then combine the outputs
been used as an indicator of economic resilience in recent literature with a negative binomial regression model to estimate their effect
(Jun et al., 2022; Karniouchina et al., 2022; Li et al., 2022). sizes. The CART analysis is a tree-based recursive binary splitting
Neighborhood characteristics were ﬁrst obtained at either census algorithm that selects, at each step, the variable and cutoff threshold
block group level or parcel level and grouped into two domains: with the strongest association with the outcome (i.e., business open
density measures and socio-demographic characteristics. or closed status) and performs a binary split of the data according to
We used three density measures: population density, that variable (Hothorn et al., 2006). Further partitioning occurs on
employment density, and land use mix entropy. Both population top of prior partitions, maximally sorting observations into groups
density (Alidadi and Shariﬁ, 2022; Carozzi et al., 2022; Schmiege or combinations based on the cutoff value at each splitting variable.
et al., 2023; Diao et al., 2021; Ganasegeran et al., 2021; Hong and Its main advantage over regression models is that CART models
Chakrabarti, 2022; Huang et al., 2020; Kadi and Khelfaoui, 2020; Lee allow the existence of correlations among variables without biasing
et al., 2021; Lima et al., 2021; Paul et al., 2021; Rahman et al., 2023; model performance. It has been used in prior studies with similar
Sun et al., 2020; Tsou et al., 2023) and employment density (Allan urban contexts to address correlated design and locational
et al., 2021; Jo et al., 2021; Liu and Su, 2021; Rosenthal et al., 2022; predictors, such as bus stop amenity changes affecting ridership
Rahman et al., 2023) are popular measurements of urban density in (Shi et al., 2021), and store locations inﬂuencing customer shopping
recent COVID-19 related articles. Hamidi et al. (2020) proposed the choices (Hounwanou et al., 2018).
use of activity density (population and employment per square Makeshift changes were implemented during and in response
mile); we contend that separating employment from population to the pandemic and related policy. As suggested in literature
density helps to disentangle the impact of work-from-home orders (e.g., Saelens and Handy, 2008), business-level changes are not
from population migration during COVID. We also controlled for selected at random but in combination, with several types of
the land use mix entropy, which is a traditional measure of urban strategies implemented in accordance to the physical constraints
form from the economic agglomeration literature (Cervero and of a business’s building type (e.g., storefront, individual building,
Kockelman, 1997; Cervero and Duncan, 2003), and has been food court). Similarly, urban density measures are spatially
used in other similar studies as well (Haddad et al., 2023; clustered and can have a nonlinear relationship with retail
Nguyen et al., 2020; Wu et al., 2022). performance. Therefore, CART analysis was used in our study
For each neighborhood-level variable, an average value was to identify the combinations of makeshift changes and building
calculated for each business with all census blocks or parcels that types and thresholds of urban densities most relevant to business
overlap within a one-quarter-mile radius buffer surrounding the continuity.
business. The choice of one-quarter-mile radius is based on the We produced two sets of CART models. The ﬁrst set identiﬁed
established walkability literature (Ewing et al., 2011; Pivo and Fisher, combinations of the most important makeshift changes
2011; Ameli et al., 2015; Ewing et al., 2016; Park et al., 2019). (i.e., distance marks and customer counters) and associated
Additional business characteristics obtained from the ﬁeld survey building types (i.e., walkable storefront and food court) for
were grouped into two domains: general business characteristics businesses. The second set of models selected the most important
(including immutable built environment conditions, such as density measures and formed threshold measures of density relevant
available parking onsite and offsite) and indoor and outdoor to business continuity at the neighborhood level. Business open or
makeshift changes. The idea that the built environment contains closed status in each period of survey (May to June and November to
makeshift or temporary characteristics pre-dates the pandemic, December 2020) were used as the outcome variables in the CART
whether as rhetorical ﬂourish in the examination of the history analysis.
of cities (Rybczynski, 2010) or as empirical classiﬁcation separating In applying CART analysis to the two time periods, the analysis
pop-up from more permanent retail establishments (Hietanen et al., identiﬁes the most statistically signiﬁcant factors in each period and
2016). Pandemic research (e.g., Ali et al., 2021) and casual shows how these factors change between the two periods of study
observation of new options for outdoor dining (e.g., Gold, 2020) (May to June and November to December 2020). This is particularly
also use the term. In Makeshift Metropolis: Ideas About Cities, useful to understand how estimations of the effect of urban density
Rybcynski (2010), 12), deﬁnes the “makeshift metropolis” as an are inﬂuenced by the type of measurement and the timing of the
“unplanned, almost anarchical area for individual enterprise.” The study period. It may also be potentially illuminating for unresolved
“makeshift” changes in the built environment observed during the questions in the literature about the role of density in the economic
early phase of the pandemic can be seen as anarchical advances of impacts of COVID-19.
individual businesses to improve the potential for business success in As CART analysis does not provide direct estimations of the size
an urban environment in transition. All variables, also described in of the effects of generated classiﬁcations, negative binomial

Frontiers in Built Environment 09 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

FIGURE 3
Modeling Flow Diagram. Classiﬁcation and regression tree analysis (CART) was applied to identify combinations and thresholds of variables
statistically signiﬁcant to the business continuity of food retail establishments within and across 16 neighborhoods over two time periods. CART results
were used as predictors in negative binomial regression analysis to estimate the size of the effect or association of the variables on operation status.

regression models were developed following CART analysis. The facilities). Based on our ﬁeld survey, 20% of businesses had
makeshift combinations and urban density thresholds from the onsite parking, though most businesses were located within
CART models were introduced as predictors in the negative 100 feet of curbside parking.
binomial regression models, with business open or closed status Employment density averaged 68 jobs per acre. The average
from May to June and November to December 2020, as the residential population density was slightly fewer than 16 persons
dependent variables. Additional controlled covariates included per acre, which is to be expected in a city with nearly half of the
business primary type, neighborhood median household income, land area zoned for single family residential use. Other
and percentage of non-Asian people of color. The choice of negative characteristics obtained from the American Community
binomial regression models was based on the presence of over- Survey included household income, racial composition, and
dispersion in the outcome variables. age range. Mean aggregate household income within walking
distance of the sampled retail businesses (about a quarter-mile)
was USD 83.5 thousand. The median age of persons living within
4 Results and ﬁndings walking distance of the businesses was within the 18–44 age
cohort.
4.1 Summary statistics Table 3 summarizes the makeshift changes found in use by
food retail businesses. These are changes made to both outdoor and
Table 2 provides summary statistics. The ﬁeld survey during indoor areas of establishments, documented by type of business in
May and June 2020, found 25% of food retail businesses closed. The May and early June 2020. Signage, posted on the store window,
repeated survey in November and December found a slight wall, or sidewalk, is the most common outdoor change adopted,
recovery, with about 14% of businesses closed. Figure 2 shows possibly due to its low cost of implementation and to its obvious
the spatial distribution of business operation status in the two use to signal that the site is open for business. Other changes were
periods. The majority of businesses in the sampled neighborhood observed with much less frequency. These include customer
census block groups were restaurants, though all forms of counters or outdoor pickup tables, to provide distance between
customer-facing food businesses were included in the sample, customers and employees and facilitate customer pick-up of
including grocery stores, convenience stores, supermarkets, prepared food, often changing the point of sale. They also
quick service and/or fast-food stores, cafes, bakeries, coffee include indicators for customers to use to reinforce social
shops, taverns and pubs. Note that business types aggregate distancing, and materials such as hand sanitizer. Note
small groceries and convenience stores with supermarkets to differences in the frequency of strategies observed between
form a food retail category. Similarly, coffee shops are merged business types. Compared with other business types, restaurants
with bakeries, pubs, and taverns, and other miscellaneous, yet have a higher percent adoption rate of signage and a lower percent
customer-facing, food service establishments. adoption rate of customer counters.
Businesses were situated in a wide variety of urban and
architectural settings. A majority of businesses in the sample
were situated along walkable storefronts. Only about 3% of the 4.2 Classiﬁcation and regression tree of
surveyed food businesses were located in a food court, as would be building type and makeshift changes
found inside of an indoor shopping mall or ofﬁce complex. Many
were housed in individual buildings (e.g., standalone buildings, Figure 4; Table 4 show the combination of makeshift changes
drive-through facilities, and former single family residential and building storefront types that emerge from the CART

Frontiers in Built Environment 10 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

TABLE 2 Summary statistics: open/close, neighborhood characteristics, business general characteristics.

Overall; (N = 921)
Outcome of Interest Open in May/Jun — —

0 231 (25.1%) —

1 690 (74.9%) —

Open in Nov/Dec —

0 132 (14.3%) —

1 789 (85.7%) —

General Business Characteristics Business Type —

Coffee/Tea/etc. 130 (14.1%) —

Fast Food 60 (6.5%) —

Food Retail 157 (17.0%) —

Restaurant 574 (62.3%) —

Price Range —

$ 281 (30.5%) —

$$ 534 (58.0%) —

$$$ 3 (0.3%) —

Not Available 103 (11.2%) —

Building Type

Walkable storefront 716 (77.7%) —

Food court 30 (3.3%) —

Individual Building and Other 175 (19.0%) —

Onsite Parking —

0 745 (80.9%) —

1 176 (19.1%) —

Curb Parking —

0 32 (3.5%) —

1 889 (96.5%) —

Neighborhood Density Measures — Mean (SD) Median [Min, Max]

Job Density (jobs/acre) 68.0 (83.5) 26.4 [0.660, 533]

Population Density (people/acre) 15.8 (6.49) 13.8 [7.93, 45.2]

Entropy Index 0.571 (0.0992) 0.596 [0.221, 0.730]

Neighborhood Demographic Characteristics — Mean (SD) Median [Min, Max]

Median Household Income (dollar) 83,700 (24,300) 82,500 [33,100, 148,000]

Percent Female 46.4% (5.1%) 47.1% [30.7%, 57.3%]

Percent non-Asian People of Color 17.2% (8.9%) 17.0% [3.2%, 48.5%]

Percent Age 18 to 44 54.8% (11.6%) 52.7% [28.9%, 78.1%]
a
Variables in italics were excluded from the negative binomial regression model due to multicollinearity.

analysis. They form seven classiﬁed combinations in the ﬁrst exhaustive classiﬁcation of the storefront types and makeshift
survey period (May to June) and two combinations in the second changes that best predict variations in business open status in the
survey period (November to December) of 2020. The respective period. Food court, walkable storefront, and individual
combinations provide a mutually exclusive and collectively buildings, with signs posted on store windows and/or on the

Frontiers in Built Environment 11 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

TABLE 3 Summary statistics: makeshift changes by primary business type (May/June 2020).

Example Coffee/Tea Fast food Food retail Restaurant Overall
(N = 130) (N = 60) (N = 157) (N = 574) (N = 921)
Outdoor: Sign post on store — — — — —
window or wall

0 63 (48.5%) 26 (43.3%) 69 (43.9%) 221 (38.5%) 379 (41.2%)

1 67 (51.5%) 34 (56.7%) 88 (56.1%) 353 (61.5%) 542 (58.8%)

Outdoor: Sign post on — — — — —
sidewalk

0 113 (86.9%) 51 (85.0%) 141 (89.8%) 478 (83.3%) 783 (85.0%)

1 17 (13.1%) 9 (15.0%) 16 (10.2%) 96 (16.7%) 138 (15.0%)

Outdoor: Pickup sign in — — — — —
parking zone

0 121 (93.1%) 56 (93.3%) 149 (94.9%) 526 (91.6%) 852 (92.5%)

1 9 (6.9%) 4 (6.7%) 8 (5.1%) 48 (8.4%) 69 (7.5%)

Outdoor: Pickup point (e.g., — — — — —
table)

0 129 (99.2%) 58 (96.7%) 153 (97.5%) 564 (98.3%) 904 (98.2%)

1 1 (0.8%) 2 (3.3%) 4 (2.5%) 10 (1.7%) 17 (1.8%)

Outdoor: Social distance — — — — —
mark

0 122 (93.8%) 58 (96.7%) 150 (95.5%) 560 (97.6%) 890 (96.6%)

1 8 (6.2%) 2 (3.3%) 7 (4.5%) 14 (2.4%) 31 (3.4%)

Indoor: Customer counter — — — — —

0 112 (86.2%) 50 (83.3%) 136 (86.6%) 527 (91.8%) 825 (89.6%)

1 18 (13.8%) 10 (16.7%) 21 (13.4%) 47 (8.2%) 96 (10.4%)

Indoor: Hand sanitizer — — — — —

0 129 (99.2%) 60 (100%) 156 (99.4%) 568 (99.0%) 913 (99.1%)

1 1 (0.8%) 0 (0%) 1 (0.6%) 6 (1.0%) 8 (0.9%)

(Continued on following page)

Frontiers in Built Environment 12 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

TABLE 3 (Continued) Summary statistics: makeshift changes by primary business type (May/June 2020).

Example Coffee/Tea Fast food Food retail Restaurant Overall
(N = 130) (N = 60) (N = 157) (N = 574) (N = 921)
Indoor: Customer waiting — — — — —
mark

0 124 (95.4%) 58 (96.7%) 147 (93.6%) 569 (99.1%) 898 (97.5%)

1 6 (4.6%) 2 (3.3%) 10 (6.4%) 5 (0.9%) 23 (2.5%)

FIGURE 4
Classiﬁcation and regression tree of makeshift changes and retail storefront types predicting open versus closed. Other makeshift changes and retail
building types were not selected by the CART models due to low associations with the outcome variable in each period; (A) Seven Combinations were
classiﬁed in May and June 2020; (B) Two Combinations were classiﬁed in November and December 2020.

sidewalk, as well as a customer counter, were identiﬁed as the 4.3 Classiﬁcation and regression tree of
most important predictors of open businesses in May and June. urban density measures
Signs posted on store windows were estimated to be the only
predictor for open businesses in November and December 2020. Figure 5; Table 5 show the results of density classiﬁcation from the
In the ﬁrst period, the combination of signs posted on store CART analysis with three measures of urban density—employment
windows and adjusted locations for customer counters density, residential population density, and entropy index of land use
were observed at the highest percentage of open businesses, mixture. Employment density was estimated to be the only signiﬁcant
with the next highest percentage of open businesses predictor of open or closed status in May and June, with 95 jobs per acre
having signs posted on both store windows and sidewalks. as the threshold that separated the classiﬁcation of businesses by their
By November and December, the repositioning neighborhood location. In contrast, residential density was the only
of customer counters no longer played a role in business predictor in November and December, with 23.6 persons per acre as the
continuity. threshold for separating classiﬁcations of businesses by their location.

Frontiers in Built Environment 13 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

TABLE 4 Combinations of building types and makeshift changes.

Combinations of building types and strategies Number of businesses Number of open (percentage)

May/June 2020

Storefront (Y) sign posted on store window 271 161 (59.4%)

Storefront (Y) sign posted on store window 290 212 (73.1%)

(N) sign posted on sidewalk

(N) customer counter

Food court/Individual building (Y) sign posted on store window 76 67 (88.2%)

(N) sign posted on sidewalk

(N) customer counter

Food court (N) sign posted on store window 17 2 (11.8%)

Individual building (N) sign posted on store window 111 86 (77.5%)

Storefront/Food court/Individual building (Y) sign posted on store window 125 124 (99.2%)

(Y) sign posted on sidewalk

Storefront/Food court/Individual building (Y) sign posted on store window 51 51 (100%)

(N) sign posted on sidewalk

(Y) customer counter

Nov/December 2020

Storefront/Food court/Individual building (Y) sign posted on store window 542 495 (91.3%)

Storefront/Food court/Individual building (N) sign posted on store window 399 314 (78.7%)
a
(Y) with this design strategy (N) without this design strategy.

4.4 Negative binomial regression results For the variables indicating makeshift adaptations, storefront
businesses without any sign are used as the baseline reference in both
Negative binomial regression estimates of the effect sizes of study periods. Compared with the baseline, businesses with a sign on
variables on open or closed status, in May through June and their window and those with signs on both the window and sidewalk
November through December 2020, are shown in Table 6. The were found to have higher odds of being open in May and June. Food
table includes the effects of makeshift change combinations and court establishments with no signs on windows were negatively
urban density thresholds produced through the CART analysis. All associated with being open. While the relocation of customer
coefﬁcients were converted into odds ratios. A ratio less than one counters by food establishments was observed with a 100% open
indicates that the presence of the corresponding variable decreases rate in Table 4, after controlling for other factors this strategy was no
the likelihood of business continuity, while a ratio greater than one longer found to be statistically signiﬁcant. In the November and
indicates that the variable increases the chance of businesses being December model, businesses with signage were also found to have
open. The further away an odds ratio from one, the stronger the higher odds of remaining open compared to those without.
association between the independent variable and the dependent Measures of employment and residential density were
variable. reclassiﬁed into binary variables based on the thresholds from
For the business type variable, food retail (e.g., grocery stores) is CART. Overall, density measures were found to be negatively
used as the baseline reference because this type of business was associated with the odds of being open. In the May and June
labeled “essential,” and demand was unlikely to decline. Therefore, it model, businesses were 82% less likely to be open if the quarter-
provides a relatively stable baseline reference to understand how all mile area around them averaged more than 95 pre-pandemic jobs
other types of businesses, by comparison, are expected to per acre. In the November and December model, businesses were
underperform during these peak pandemic periods. Compared to 66% less likely to be open if the quarter-mile area around them
food retail, restaurants were 76% less likely to be open in May and averaged more than 23.6 persons per acre.
June of 2020, and 66% less likely to be open by the end of the year. For other neighborhood characteristics, median household
Similarly, shops selling coffee and tea, including bakeries and other income and percentage of non-Asian persons of color were both
establishments with specialized products, were 69% less likely than positively and signiﬁcantly associated with business continuity in
food retail to remain open during the May to June period of 2020, May and June, as well as November and December of 2020. In
improving the odds of reopening by only 2% by the end of the year. particular, the magnitude of the estimated effect of percentage of

Frontiers in Built Environment 14 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

FIGURE 5
Classiﬁcation and regression tree of urban density thresholds predicting open versus closed. Other urban density measures were not selected by the
CART model due to low associations with the outcome variable in each period. (A) Employment density was the only signiﬁcant predictor in May and June
2020; (B) Residential population density was the only signiﬁcant predictor in November and December 2020.

TABLE 5 Classiﬁcations of urban density measures.

Classiﬁcations of urban density characteristics Number of businesses Number of open (percentage)

May/June 2020

Job density >95.05 jobs/acre 256 135 (52.7%)

Job density ≤ 95.05 jobs/acre 685 568 (82.9%)

Nov/December 2020

Residential population density >23.56 people/acre 105 76 (72.4%)

Residential population density ≤ 23.56 people/acre 836 733 (87.7%)

non-Asian persons of color increased greatly between the ﬁrst and reassess how they might be able to weather the crisis (Bartik
second period of analysis. et al., 2020). Existing literature clearly anticipates and describes
impacts to customer-facing food businesses from reductions in
demand (Bartik et al., 2020; Goolsbee and Syverson, 2021; Balla-
5 Discussion Elliott et al., 2022). However, most studies have described such
impacts at more aggregated scales, either at the county level or
Despite the designation of essential businesses (Goolsbee submarket level, failing to identify the market segments and
and Syverson, 2020), efforts to pivot to online communications individual businesses that require assistance during the
and delivery (Yenerall et al., 2022; Norris et al., 2021; Russo pandemic. With data collected in the ﬁeld within the ﬁrst
et al., 2022; Bonfanti et al., 2023), and ﬁnancial assistance few months of the pandemic and the use of CART analysis
programs, such as the Paycheck Protection Program (Autor and regression models, this study adds to the existing body of
et al., 2022), businesses carrying relatively small amounts of literature by generating more detailed and accurate proﬁles of
cash on hand found drops in demand to be highly disruptive, food retail businesses in relation to business continuity during
leading many to lay off workers, temporarily shut down, and these critical time periods.

Frontiers in Built Environment 15 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

TABLE 6 Impact factors of food and retail business continuity.

— GLM - May/June 2020 GLM - Nov/December 2020

(Open = 1, close = 0) (Open = 1, close = 0)

(1) (2)

Odds ratio (S.E.) p-value Odds ratio (S.E.) p-value

Business Type

Coffee/Tea/etc. (ref = “Food Retail”) 0.306 (1.43) 0.005 0.325 (1.509) 0.006

Fast Food (ref = “Food Retail”) 0.445 (1.61) 0.089 0.601 (1.736) 0.356

Restaurant (ref = “Food Retail”) 0.238 (1.34) <0.001 0.341 (1.425) 0.002

Combinations of Building Types and Makeshift Changes in May/June 2020

Storefront (N) sign posted on store window References — — —

Storefront (Y) sign posted on store window 2.31 (1.23) <0.001 — —

(N) sign posted on sidewalk

(N) customer counter

Food court/Individual Building (Y) sign posted on store window 3.75 (1.51) 0.001 — —

(N) sign posted on sidewalk

(N) customer counter

Food court (N) sign posted on store window 0.195 (2.22) 0.040 — —

Individual building (N) sign posted on store window 1.353 (1.38) 0.349 — —

Storefront/Food court/Individual Building (Y) sign posted on store window 8.244 (2.77) <0.001 — —

(Y) sign posted on sidewalk

Storefront/Food court/Individual Building (Y) sign posted on store window 4.481 e+07 (1.81 e+225) 0.973 — —

(N) sign posted on sidewalk

(Y) customer counter

Combinations of Building Types and Makeshift Changes in Nov/December 2020

Storefront/Food court/Individual Building (N) sign posted on store window — — References

Storefront/Food court/Individual Building (Y) sign posted on store window — — 3.300 (1.230) <0.001

Classiﬁcation of Urban Density Measures in May/June 2020

Job density ≤ 95.05 jobs/acre References — — —

Job density >95.05 jobs/acre 0.179 (1.23) <0.001 — —

Classiﬁcation of Urban Density Measures in Nov/December 2020

Residential population density ≤ 23.56 people/acre — — References —

Residential population density >23.56 people/acre — — 0.339 (1.320) <0.001

Neighborhood Characteristics

Median Household Income (thousand dollars) 1.012 (1.00) 0.013 1.023 (1.006) <0.001

Percent non-Asian people of color 3.910 (4.74) 0.019 153.701 (5.351) 0.003

Constant 1.589 (1.908) 0.463 0.664 (2.129) 0.587

Observations 921 — 921 —

Log Likelihood −387.98 — −341.34 —

Akaike Inf. Crit 801.97 — 698.68 —
a
(Y) with this design strategy (N) without this design strategy.

Frontiers in Built Environment 16 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

FIGURE 6
(A) Business clusters under each employment density classiﬁcation in May through June 2020; Businesses in Group 1, of higher employment density,
have a lower opening rate (52.7%) than those in Group 2 (82.9%) (B) Business clusters under each residential population density classiﬁcations in
November through December 2020; Businesses in Group 1, of higher population density, have a lower opening rate (72.4%) than those in Group 2 (87.7%).

In May and June 2020, at the business level, fast food sites were supporting programs, such as the issuance of special parking
the businesses most likely to continue operations. At the permits to facilitate take-out and food delivery, and permits to
neighborhood level, businesses found in neighborhoods with job allow outdoor dining in the public right-of-way, proﬁles can assist in
density of less than 95 per acre were more likely to remain open. evaluating the effectiveness of those public programs in reaching the
Median income within a walkable area of a business was positively businesses most in need, and to identify potential service gaps. These
correlated with business continuity. Yet these beneﬁts of business proﬁles identify successful (and unsuccessful) cases for further in-
continuity also accrued to locations within a walkable area that depth study to understand the causal mechanisms of how urban
includes comparatively higher proportions of non-Asian persons of designs at the street-level and the neighborhood-level can facilitate a
color, which are comparatively lower income neighborhoods. vibrant economy, despite the pandemic.
Conversely, businesses most vulnerable to closure were found in Besides the proﬁles, the ﬁndings also add to the understanding of
job-dense neighborhoods, especially in the areas around or in underlying factors that affect business continuity. Despite the
downtown, Seattle. Food court closures, inside retail malls and presence of signiﬁcant investments in indoor and outdoor dining
similar establishments, made restaurants in these settings facilities in our study areas within Seattle, the statistically signiﬁcant
particularly vulnerable to closure. Business closure also was more factor observed to have the strongest impact on business continuity
likely if the business was located in a less-wealthy area, or in a less was the posting of signs either in windows or afﬁxed to sandwich
diverse area. These proﬁles, however, changed within 6 months. By boards on the sidewalk. There are several reasons for signage to
November and December, the business types most vulnerable to matter in the immediate wake of pandemic restrictions on business
closure were coffee and tea shops. Vulnerability to closure was also operations. During our ﬁeld audits, we found that information
found in residential neighborhoods with pre-pandemic residential online was often inaccurate, lagging behind the times in regard
density greater than of 23.6 persons/acre, or lower median income, to operating status and open hours. Signage, as simple as it may be,
or lower diversity. provides a quick and cost-effective communication channel with
These proﬁles can serve multiple urban planning purposes. In customers about new hours, menu items, and safety guidelines.
the short term, the proﬁles, especially the business and building Business continuity, perhaps emphasized in signage, may also
types and the neighborhood level characteristics, can help public reﬂect unobserved variables, such as ﬁnancial conditions and
programs to better target market segments and business groups for digitalization. The results from structured phone interviews of a
more efﬁcient resource allocation. By comparing the proﬁles with subsample (n = 82) of businesses (Supplementary Table S2), found
guidelines and criteria used for the qualiﬁcation of business that businesses with posted signs were on average accruing more

Frontiers in Built Environment 17 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

revenue than in pre-pandemic times (i.e., 2019), and were beneﬁting was also true, however, for other neighborhoods in the city. Studies
from higher percentages of ﬁnancial assistance (e.g., PPP, EMIC) have found racial and ethnic minority groups less likely to leave the
and rent relief. In terms of operations, they had more full-time city and more dependent on work required to be conducted on-
employees and longer operating hours in May and June of 2020 than premises during the pandemic (Coven and Gupta, 2020; Fu and
their counterparts. A signiﬁcantly higher percentage were advancing Zhai, 2021; Zhai and Yue, 2022), which corresponds to the ﬁnding
digitization, adding to their existing online delivery services during that businesses in neighborhoods with higher percentages of non-
COVID-19, and/or witnessing higher percentages of pickup orders Asian persons of color are more likely to stay open.
in May and June of 2020. The small sample size of this follow-up Placed in the context of the real estate literature, the ﬁndings
phone survey cannot show statistical association. Still, the encourage mixed-use development in the city core and the
descriptive results do suggest the possibility of relationships redistribution of restaurants and similar food service businesses
between makeshift adaptations in the built environment, ﬁnancial to diverse residential neighborhoods, to maintain close ties between
relief, and the efforts of businesses to elevate take-out and delivery local businesses and their customer base. Traditionally, the real
throughput. estate market has valued the centrality and density of urban areas.
Perhaps the most interesting results of this study examine how Rosenthal et al. (2022) found that pandemic employment density
density in the built environment and socio-economic factors relate had a negative impact on both ofﬁce rent and the number of new
to the continued operation or closing of food retail businesses. Pre- tenant leases and renewals. Other recent studies also document a
pandemic employment density is negatively and signiﬁcantly shift of demand in the housing market away from densely populated
associated with customer-facing food business continuity, as areas (Liu and Su, 2021; Ramani and Bloom, 2021). This study offers
measured in the May to June period of 2020. Just 6 months later, a retrospective explanation of these changes, beginning with work-
however, this variable was replaced by pre-pandemic residential from-home and remote-work vacancies of high-density ofﬁce
population density, which is also negatively and signiﬁcantly buildings (and the closure of associated local food retail
associated with food retail business continuity. businesses) and followed, 6 months later, by the ﬂight of
These results may be better explored through further residents from high-density residential buildings on the outskirts
understanding of the spatial distribution of density in Seattle. of downtown (and the closure of associated local food retail
Figure 6A shows the spatial distribution of businesses in the businesses).
study, according to their placement above or below the CART
threshold of 95 jobs per acre within a quarter-mile radius.
Businesses clustered in and around the downtown area of Seattle, 6 Conclusion
which is dominated by ofﬁce buildings, were 82% less likely to be
open from May to June of 2020. As downtown ofﬁce workers shifted The economic impacts of COVID-19 were not equally
to work-from-home practices, the ﬂoor fell out from underneath the distributed across the businesses comprising the food retail
businesses dependent on this customer base. industry, nor were they equally distributed across the urban areas
Results in November to December 2020, show food retail of cities. This study sought to improve understanding of role of the
businesses 66% less likely to stay open if the area within a built environment in mediating urban economic decline and
quarter mile radius of the establishment had a pre-pandemic resiliency during the ﬁrst year of the pandemic, by investigating
residential population density greater than 23.6 per acre. To try the effects of business-level factors (e.g., business type, building type,
to understand why pre-pandemic residential density could worsen and makeshift adaptations made by the business to improve
customer-facing food business continuity, we again examine the outcomes), and neighborhood-level factors, (e.g., employment
spatial and built environment context for our results, as shown in density, residential population density, and socio-demographics),
Figure 6B. What began as a collection of business closures associated on the business continuity of food retail establishments in the city of
with a loss of employment density in the central business district, Seattle. Results show that closure was much more likely for
Pike’s Place Market and Pioneer Square, grew outward over the businesses such as restaurants, cafes, and pubs, than for their
ensuing months to the South Lake Union area, associated with a loss counterparts in food retail, such as grocery stores and
of residential population density. supermarkets. Results also demonstrate a meaningful role for
Placing these results in the context of the literature, Ramani and density in the risk of business closure, though that role changed
Bloom (2021) describe a “donut effect” of the pandemic on cities, a over time. The ﬁrst wave of food retail business closure coincided
hollowing out of city centers, with business demand shifting to the with the ﬁrst wave of the pandemic and the immediate and
suburban and outer rings of large US metropolitan areas, based on widespread shift of workers in ofﬁce-rich downtown areas to
data from the US Postal Service and Zillow. Coven et al. (2022) work-from-home practices. Six months later, this was followed
suggest that parts of downtown Seattle had, like other US urban by a separate wave of food retail business closures, possibly in
centers, also experienced the mass relocation of people from urban connection with the ﬂight of people away from high density
to more suburban and exurban areas, based on smartphone GPS and residential areas, on the outskirts of downtown. Both dynamics
ZIP code information. Our research suggests that the driving forces contrast with observations in the socio-demographically diverse
of the hollowing out of cities likely changed over time, attributable areas of the city, where food retail businesses exhibited stability
ﬁrst to work-from-home policies and followed by the movement of and resilience. Understanding the effects of such factors on business
residents to locations that met latent preferences for residential closure offers opportunities to provide more efﬁcient and targeted
living—preferences that were masked by the desire to be ﬁnancial support (Autor et al., 2022), should a similar crisis appear
conveniently located close to one’s place of work. The converse in the future.

Frontiers in Built Environment 18 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

Other authors have expressed concern that there may be Author contributions
strategic misalignments between business assets and the shifting
environmental demands of novel shocks or crises, such as FS and SN organized the data, cleaned the data, and performed
COVID-19, that place industries at risk (Amankwah-Amoah statistical analyses. FS contributed the draft of the introduction, data
et al., 2021). Our study demonstrates that the built and methodology, results and ﬁndings, discussion, and conclusion
environment itself, with its form, density, and mix of land use, of the manuscript and contributed to the literature review. SN
is one such asset misaligned with the aim of resilience against contributed the draft of the data and methodology and literature
pandemic shock. The hollowing out of the ofﬁce-rich downtown review sections of the manuscript. JW, GP, and FS conceptualized
core of Seattle may have been swift, but its effects have continued the research design. JW acquired the funding, was the principal
as vacancy rates remain higher than pre-pandemic levels, and as a investigator for the research, developed and oversaw the research
substantial subset of workers reconsider their preferred place of design and methodology, contributed to the literature review, and
residence with work-from-home options in mind. The adaptive substantially revised the draft to produce the ﬁnal manuscript. At the
capacity and therefore the resilience of the food retail industry are time of this research FS and SN were doctoral students in the lab
affected by the dynamic interplay of characteristics of the built founded by JW. GP oversaw and managed the ﬁeld audit and phone
environment with these pandemic-driven economic impacts. The surveys, and created and revised the ﬁgures. All authors contributed
food retail industry, often dependent upon a local customer base, to the article and approved the submitted version.
may be a bellwether for both swift and lasting change in the urban
economy.
There are four types of limitations to this study. First, this Funding
study focused on the built environment, seeking a balance
between internal validity and external validity. Though this This research was made possible by the generous and timely
research includes the full population of food retail businesses in support of University of Washington Population Health Initiative,
sampled neighborhoods, the sample was not randomly selected with a COVID-19 Rapid Response Grant and a COVID-
and there were no controls. Additional explanatory variables, 19 Economic Recovery Research Grant, both administered in 2020.
such as ﬁnancial condition and digitalization, have the potential
to change the outcomes of this study. And apart from a temporal
relationship between the ﬁrst two waves of COVID-19 Acknowledgments
incidence and the implementation of the study, the research
design did not provide explicit linkages with variables The authors would like to thank the members of the Urban
representing policy and behavior. Perceived risk, for Infrastructure Lab and the Circular City and Living Systems Lab at
example, as a controlled variable, has the potential to change the University of Washington for their contributions.
the sign of density, and may differ among groups in society
(Chauhan et al., 2021; Carella et al., 2022; Joiner et al., 2022).
Second, though our selection of neighborhoods for study Conﬂict of interest
excluded prominent areas of unrest, it is important to note
that the “Black Lives Matter” protests occurred during these The authors declare that the research was conducted in the
periods of observation, which may have contributed to business absence of any commercial or ﬁnancial relationships that could be
closure (Motoyama, 2022). Third, there were limits to the construed as a potential conﬂict of interest.
frequency of observations recorded in the ﬁeld for this study.
More recent public health and real estate research papers are
beginning to ﬁll this gap by examining impacts over longer Publisher’s note
periods of time (Lee et al., 2021; Carozzi et al., 2022; Guo et al.,
2023; Kiaghadi et al., 2022; Motoyama, 2022; Wang et al., 2022a; All claims expressed in this article are solely those of the authors
Wu et al., 2022; Tsou et al., 2023). Fourth, as a study of urban and do not necessarily represent those of their afﬁliated
neighborhoods, zoned for growth, in a US city, at the center of a organizations, or those of the publisher, the editors and the
metropolitan area, there are limits to the generalizability of this reviewers. Any product that may be evaluated in this article, or
work and its ﬁndings. This study suggests that Seattle claim that may be made by its manufacturer, is not guaranteed or
increasingly resembles the transit-rich metropolitan centers endorsed by the publisher.
described by Rosenthal et al. (2022).

Supplementary material
Data availability statement
The Supplementary Material for this article can be found online
The raw data supporting the conclusion of this article will be at: https://www.frontiersin.org/articles/10.3389/fbuil.2023.1212244/
made available by the authors, without undue reservation. full#supplementary-material

Frontiers in Built Environment 19 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

References
Acemoglu, D., Chernozhukov, V., Werning, I., and Whinston, M. D. (2020). A Carozzi, F., Provenzano, S., and Roth, S. (2022). Urban density and COVID-19:
multi-risk SIR model with optimally targeted lockdown. MA: National Bureau of understanding the US experience. Ann. Reg. Sci., 1–32. doi:10.1007/s00168-022-01193-z
Economic Research Cambridge.
Carré, F., and Tilly, C. (2008). America’s biggest low-wage industry: continuity and
Acuto, M. (2020). COVID-19: lessons for an urban(izing) world. One Earth 2, change in retail jobs. Center for Social Policy Publications 22.
317–319. doi:10.1016/j.oneear.2020.04.004
Cervero, R., and Duncan, M. (2003). Walking, bicycling, and urban landscapes:
Akseer, N., Kandru, G., Keats, E. C., and Bhutta, Z. A. (2020). COVID-19 pandemic Evidence from the San Francisco bay area. Am. J. Public Health 93 (9), 1478–1483.
and mitigation strategies: lmplications for maternal and child health and nutrition. Am. doi:10.2105/ajph.93.9.1478
J. Clin. Nutr. 112, 251–256. doi:10.1093/ajcn/nqaa171
Cervero, R., and Kockelman, K. (1997). Travel demand and the 3Ds: Density,
Ali, S. H., Imbruce, V. M., Russo, R. G., Kaplan, S., Stevenson, K., Mezzacca, T. A., diversity, and design. Transp. Res. Part D Transp. Environ. 2 (3), 199–219. doi:10.
et al. (2021). Evaluating closures of fresh fruit and vegetable vendors during the 1016/S1361-9209(97)00009-6
COVID-19 pandemic: methodology and preliminary results using omnidirectional
CDC (2020). Coronavirus disease 2019. Centers for Disease Control and Prevention.
street view imagery. JMIR Form. Res. 5, e23870. doi:10.2196/23870
URL https://www.cdc.gov/media/releases/2020/p0121-novel-coronavirus-travel-case.
Alidadi, M., and Shariﬁ, A. (2022). Effects of the built environment and human factors html (accessed July 7, 23).
on the spread of COVID-19: a systematic literature review. Sci. Total Environ. 850,
Chauhan, R. S., da Silva, D. C., Salon, D., Shamshiripour, A., Rahimi, E., Sutradhar, U.,
158056. doi:10.1016/j.scitotenv.2022.158056
et al. (2021). COVID-19 related attitudes and risk perceptions across urban, rural, and
Allan, R., Liusman, E., Lu, T., and Tsang, D. (2021). The COVID-19 pandemic and suburban areas in the United States. Findings. doi:10.32866/001c.23714
commercial property rent dynamics. JRFM 14, 360. doi:10.3390/jrfm14080360
Choi, K. H., and Denice, P. (2022). Socioeconomic variation in the relationship
Almagro, M., and Orane-Hutchinson, A. (2022). JUE Insight: the determinants of the between neighbourhoods’ built environments and the spread of COVID-19 in Toronto,
differential exposure to COVID-19 in New York city and their evolution over time. Canada. Can. Stud. Popul. 49, 149–181. doi:10.1007/s42650-022-00070-6
J. Urban Econ. 127, 103293. doi:10.1016/j.jue.2020.103293
Coven, J., and Gupta, A. (2020). Disparities in mobility responses to COVID-19. New
Alvarez, F. E., Argente, D., and Lippi, F. (2020). A simple planning problem for covid- York: New York University.
19 lockdown. Cambridge: National Bureau of Economic Research.
Coven, J., Gupta, A., and Yao, I. (2022). JUE Insight: urban ﬂight seeded the COVID-
Amankwah-Amoah, J., Khan, Z., Wood, G., and Knight, G. (2021). COVID-19 and 19 pandemic across the United States. J. Urban Econ. 133, 103489. doi:10.1016/j.jue.
digitalization: the great acceleration. J. Bus. Res. 136, 602–611. doi:10.1016/j.jbusres. 2022.103489
2021.08.011
Covid Crisis Group (2023). Lessons from the covid war: An investigative report. Public
Ameli, S. H., Hamidi, S., Garﬁnkel-Castro, A., and Ewing, R. (2015). Do better urban Affairs.
design qualities lead to more walking in Salt Lake City, Utah? J. Urban Des. 20 (3),
Cutler, D. M., and Summers, L. H. (2020). The COVID-19 pandemic and the
393–410. doi:10.1080/13574809.2015.1041894
$16 trillion virus. JAMA 324, 1495–1496. doi:10.1001/jama.2020.19759
Asif, Z., Chen, Z., Stranges, S., Zhao, X., Sadiq, R., Olea-Popelka, F., et al. (2022).
Diao, Y., Kodera, S., Anzai, D., Gomez-Tames, J., Rashed, E. A., and Hirata, A. (2021).
Dynamics of SARS-CoV-2 spreading under the inﬂuence of environmental factors and
Inﬂuence of population density, temperature, and absolute humidity on spread and
strategies to tackle the pandemic: a systematic review. Sustain. Cities Soc. 81, 103840.
decay durations of COVID-19: a comparative study of scenarios in China, england,
doi:10.1016/j.scs.2022.103840
Germany, and Japan. One Health 12, 100203. doi:10.1016/j.onehlt.2020.100203
Autor, D., Cho, D., Crane, L. D., Goldar, M., Lutz, B., Montes, J., et al. (2022). The
Dietz, L., Horve, P. F., Coil, D. A., Fretz, M., Eisen, J. A., and Wymelenberg, K. V. D.
$800 billion paycheck protection program: where did the money go and why did it go
(2020). 2019 Novel Coronavirus (COVID-19) pandemic: built environment
there? J. Econ. Perspect. 36, 55–80. doi:10.1257/jep.36.2.55
considerations to reduce transmission. mSystems 5, e00245-20. doi:10.1128/
Balemi, N., Füss, R., and Weigand, A. (2021). COVID-19’s impact on real estate mSystems.00245-20
markets: review and outlook. Financ. Mark. Portf. Manag. 35, 495–513. doi:10.1007/
Dong, X., and Zeballos, E. (2021). COVID-19 working paper: The effects of COVID-
s11408-021-00384-6
19 on food sales (No. AP-088). Washington: USDA Economic Research Service.
Balk (2022). Only one major city ranks higher than Seattle for remote work. [WWW
Dua, A., Ellingrud, K., Mahajan, D., and Silberg, J., 2020. Which small businesses are
Document]. The Seattle Times. URL https://www.seattletimes.com/seattle-news/only-
most vulnerable to COVID-19—and when [WWW Document]. McKinsey and
one-major-city-ranks-higher-than-seattle-for-remote-work/ (accessed July 7, 23).
Company. URL https://www.mckinsey.com/featured-insights/americas/which-small-
Balla-Elliott, D., Cullen, Z. B., Glaeser, E. L., Luca, M., and Stanton, C. (2022). businesses-are-most-vulnerable-to-covid-19-and-when (accessed July.21.23).
Determinants of small business reopening decisions after covid restrictions were lifted.
Eichenbaum, M. S., Rebelo, S., and Trabandt, M. (2021). The macroeconomics of
J. Policy Analysis Manag. 41, 278–317. doi:10.1002/pam.22355
epidemics. Rev. Financial Stud. 34, 5149–5187. doi:10.1093/rfs/hhab040
Balla-Elliott, D., Cullen, Z. B., Glaeser, E. L., Luca, M., and Stanton, C. T. (2020).
Ewing, R., Greenwald, M., Zhang, M., Walters, J., Feldman, M., Cervero, R., et al.
Business reopening decisions and demand forecasts during the COVID-19 pandemic.
(2011). Trafﬁc generated by mixed-use developments—six-region study using
Barraza, S., Rossi, M., and Yeager, T. J. (2020). The short-term effect of the paycheck consistent built environmental measures. J. Urban Plan. Dev. 137 (3), 248–261.
protection program on unemployment. SSRN J. doi:10.2139/ssrn.3667431 doi:10.1061/(ASCE)UP.1943-5444.0000068
Bartik, A. W., Bertrand, M., Cullen, Z., Glaeser, E. L., Luca, M., and Stanton, C. (2020). Ewing, R., Hajrasouliha, A., Neckerman, K. M., Purciel-Hill, M., and Greene, W.
The impact of COVID-19 on small business outcomes and expectations. Proc. Natl. (2016). Streetscape features related to pedestrian activity. J. Plan. Educ. Res. 36 (1), 5–15.
Acad. Sci. 117, 17656–17666. doi:10.1073/pnas.2006991117 doi:10.1177/0739456X15591585
Basellini, U., and Camarda, C. G. (2022). Explaining regional differences in mortality Fairlie, R., and Fossen, F. M. (2022). The early impacts of the COVID-19
during the ﬁrst wave of Covid-19 in Italy. Popul. Stud. 76, 99–118. doi:10.1080/ pandemic on business sales. Small Bus. Econ. 58, 1853–1864. doi:10.1007/
00324728.2021.1984551 s11187-021-00479-4
Benita, F., Rebollar-Ruelas, L., and Gaytán-Alfaro, E. D. (2022). What have we learned Fan, P., Chen, J., and Sarker, T. (2022). Roles of economic development level and
about socioeconomic inequalities in the spread of COVID-19? A systematic review. other human system factors in COVID-19 spread in the early stage of the pandemic.
Sustain Cities Soc. 86, 104158. doi:10.1016/j.scs.2022.104158 Sustainability 14, 2342. doi:10.3390/su14042342
Bonfanti, A., Vigolo, V., Yfantidou, G., and Gutuleac, R. (2023). Customer experience Farboodi, M., Jarosch, G., and Shimer, R. (2021). Internal and external effects of social
management strategies in upscale restaurants: lessons from the Covid-19 pandemic. Int. distancing in a pandemic. J. Econ. Theory 196, 105293. doi:10.1016/j.jet.2021.105293
J. Hosp. Manag. 109, 103416. doi:10.1016/j.ijhm.2022.103416
Frazier, L. (2021). The coronavirus crash of 2020, and the investing lesson it taught us.
Brewer, P., and Sebby, A. G. (2021). The effect of online restaurant menus on New Jersey: Forbes.
consumers’ purchase intentions during the COVID-19 pandemic. Int. J. Hosp. Manag.
Fu, X., and Zhai, W. (2021). Examining the spatial and temporal relationship between
94, 102777. doi:10.1016/j.ijhm.2020.102777
social vulnerability and stay-at-home behaviors in New York City during the COVID-
Byrd, K., Her, E., Fan, A., Almanza, B., and Liu, Y.Lei US Department of Health and 19 pandemic. Sustain. Cities Soc. 67, 102757. doi:10.1016/j.scs.2021.102757
Human Services tch, S. (2021). Restaurants and COVID-19: what are consumers’ risk
Gallagher, G. M. (2020). Coronavirus case conﬁrmed in Washington state. Infect.
perceptions about restaurant food and its packaging during the pandemic? Int.
Control Today.
J. Hosp. Manag. 94, 102821. doi:10.1016/j.ijhm.2020.102821
Ganasegeran, K., Jamil, M. F. A., Ch’ng, A. S. H., Looi, I., and Peariasamy, K. M.
Carella, G., Pérez Trufero, J., Álvarez, M., and Mateu, J. (2022). A bayesian spatial
(2021). Inﬂuence of population density for COVID-19 spread in Malaysia: an
analysis of the heterogeneity in human mobility changes during the ﬁrst wave of the
ecological study. Int. J. Environ. Res. Public Health 18, 9866. doi:10.3390/
COVID-19 epidemic in the United States. Am. Statistician 76, 64–72. doi:10.1080/
ijerph18189866
00031305.2021.1965657

Frontiers in Built Environment 20 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

Gao, X., Fan, C., Yang, Y., Lee, S., Li, Q., Maron, M., et al. (2021). Early indicators of Jun, B., Jara-Figueroa, C., and Yu, D. (2022). The economic resilience of a city: the
human activity during COVID-19 period Using digital trace data of population effect of relatedness on the survival of amenity shops during the COVID-19 pandemic.
activities. Front. Built Environ. 6, 607961. doi:10.3389/fbuil.2020.607961 Camb. J. Regions, Econ. Soc. 15, 551–573. doi:10.1093/cjres/rsac029
Gao, X., Shi, X., Guo, H., and Liu, Y. (2020). To buy or not buy food online: the impact Kadi, N., and Khelfaoui, M. (2020). Population density, a factor in the spread of
of the COVID-19 epidemic on the adoption of e-commerce in China. PLOS ONE 15, COVID-19 in Algeria: statistic study. Bull. Natl. Res. Centre 44, 138. doi:10.1186/
e0237900. doi:10.1371/journal.pone.0237900 s42269-020-00393-x
Gold, M. (2020). Indoor dining returns to New York city, but will customers? The New Kang, M., Choi, Y., Kim, J., Lee, K. O., Lee, S., Park, I. K., et al. (2020). COVID-19
York Times. Available at: https://www.nytimes.com/2020/09/30/nyregion/indoor- impact on city and region: what’s next after lockdown? Int. J. Urban Sci. 24, 297–315.
dining-nyc.html. doi:10.1080/12265934.2020.1803107
Goolsbee, A., and Syverson, C. (2021). Fear, lockdown, and diversion: comparing Karniouchina, K., Sarangee, K., Theokary, C., and Kübler, R. (2022). The impact of the
drivers of pandemic economic decline 2020. J. public Econ. 193, 104311. doi:10.1016/j. COVID-19 pandemic on restaurant resilience: lessons, generalizations, and ideas for
jpubeco.2020.104311 future research. Serv. Sci. 14, 121–138. doi:10.1287/serv.2021.0293
Govindan, K., Mina, H., and Alavi, B. (2020). A decision support system for demand Khorram-Manesh, A., Dulebenets, M. A., and Goniewicz, K. (2021). Implementing
management in healthcare supply chains considering the epidemic outbreaks: a case public health strategies—the need for educational initiatives: a systematic review.
study of coronavirus disease 2019 (COVID-19). Transp. Res. Part E Logist. Transp. Rev. IJERPH 18, 5888. doi:10.3390/ijerph18115888
138, 101967. doi:10.1016/j.tre.2020.101967
Kiaghadi, A., Adepoju, O. E., Rifai, H. S., Liaw, W., and Woodard, L. D. (2022).
Guo, Z., Zhang, F., and Ji, Y. (2023). Investigating the nonlinear relationship between Examining longitudinal disparities in COVID-19 prevalence in the U.S. a county level
takeout order demand and built environment under different periods of COVID-19. J. growth rate perspective. Ann. Med. 54, 1277–1286. doi:10.1080/07853890.2022.2069852
Adv. Transp. doi:10.1155/2023/5248888
King County (2020a). Executive constantine and king county health ofﬁcer announce
Gupta, A., Mittal, V., and Van Nieuwerburgh, S. (2022). Work from home and the new orders to limit spread of COVID-19 - king county. King County.
ofﬁce real estate apocalypse (No. w30526). Cambridge, MA: National Bureau of
King County (2020b). King county safe start variance application.
Economic Research. doi:10.3386/w30526
King County (2020c). Modiﬁed phase 1 guidance for king county restaurants.
Gursoy, D., and Chi, C. G. (2020). Effects of COVID-19 pandemic on hospitality
industry: review of the current situations and a research agenda. J. Hosp. Mark. Manag. King County (2020d). King county enters phase 2 of safe start recovery plan. Available
29, 527–529. doi:10.1080/19368623.2020.1788231 at: https://kingcounty.gov/elected/executive/constantine/news/release/2020/June/19-
phase-2-reopening.aspx.
Haddad, J. A., Mondal, A., and Bhat, C. B. (2023). Eat-in or eat-out? A joint model to
analyze the new landscape of dinner meal preferences. Transp. Res. Part C Emerg. Klein, A., and Smith, E. (2021). Explaining the economic impact of COVID-19: Core
Technol. 147. doi:10.1016/j.trc.2023.104016 industries and the Hispanic workforce.
Hamidi, S., Sabouri, S., and Ewing, R. (2020). Does density aggravate the COVID-19 Klompmaker, J. O., Hart, J. E., Holland, I., Sabath, M. B., Wu, X., Laden, F., et al.
Pandemic? J. Am. Plan. Assoc. 86, 495–509. doi:10.1080/01944363.2020.1777891 (2021). County-level exposures to greenness and associations with COVID-19
incidence and mortality in the United States. Environ. Res. 199, 111331. doi:10.
Hasnat, B. (2022). Culture, institution, and policy response to COVID-19. J. Econ.
1016/j.envres.2021.111331
Issues 56, 624–631. doi:10.1080/00213624.2022.2066446
Kruse, T., and Strack, P. (2020). Optimal control of an epidemic through social
Hietanen, J., Mattila, P., Schouten, J. W., Sihvonen, A., and Toyoki, S. (2016).
distancing.
Reimagining society through retail practice. J. Retail. 92, 411–425. doi:10.1016/j.
jretai.2016.05.005 Laefer, D. F., and Protopapas, D. (2022). Essential business visits and social
vulnerability during New York city’s initial COVID-19 outbreak. Epidemiologia 3,
Honey-Roses, J., Anguelovski, I., Bohigas, J., Chireh, V., Daher, C., Konijnendijk, C.,
518–532. doi:10.3390/epidemiologia3040039
et al. (2020). The impact of COVID-19 on public space: A review of the emerging
questions (preprint). Open Sci. Framew. doi:10.31219/osf.io/rf7xa Lee, W., Kim, H., Choi, H. M., Heo, S., Fong, K. C., Yang, J., et al. (2021). Urban
environments and COVID-19 in three Eastern states of the United States. Sci. Total
Hong, A., and Chakrabarti, S. (2022). Compact living or policy inaction? Effects of
Environ. 779, 146334. doi:10.1016/j.scitotenv.2021.146334
urban density and lockdown on the COVID-19 outbreak in the US. Urban Stud. 60,
1588–1609. doi:10.1177/00420980221127401 Leist, A. K., Klee, M., Kim, J. H., Rehkopf, D. H., Bordas, S. P. A., Muniz-Terrera, G.,
et al. (2022). Mapping of machine learning approaches for description, prediction, and
Hothorn, T., Hornik, K., and Zeileis, A. (2006). Unbiased recursive partitioning: a
causal inference in the social and health sciences. Sci. Adv. 8, eabk1942. doi:10.1126/
conditional inference framework. J. Comput. Graph. Statistics 15, 651–674. doi:10.1198/
sciadv.abk1942
106186006X133933
Li, Y., Sungur, E., Jimenez, A., and Brail, S. (2022). Covid-19, urban economic
Hounwanou, S., Comi, A., Gonzalez-Feliu, J., and Gondran, N. (2018). Inner city
resilience and the pandemic pivot: toronto’s restaurant scene. Regional Stud. Regional
versus urban periphery retailing: store relocation and shopping trip behaviours.
Sci. 9, 1–4. doi:10.1080/21681376.2021.2013732
Indications from saint-etienne. Transp. Res. Procedia 30, 363–372. doi:10.1016/j.
trpro.2018.09.039 Lima, F. T., Brown, N. C., and Duarte, J. P. (2021). Understanding the impact of
walkability, population density, and population size on COVID-19 spread: a pilot study
Huang, J., Kwan, M.-P., Kan, Z., Wong, M. S., Kwok, C. Y. T., and Yu, X. (2020).
of the early contagion in the United States. Entropy 23, 1512. doi:10.3390/e23111512
Investigating the relationship between the built environment and relative risk of
COVID-19 in Hong Kong. ISPRS Int. J. Geo-Information 9, 624. doi:10.3390/ Liu, C., Liu, Z., and Guan, C. (2021). The impacts of the built environment on the
ijgi9110624 incidence rate of COVID-19: A case study of king county, Washington. Sustain. Cities
Soc. 74, 103144. doi:10.1016/j.scs.2021.103144
Hutson, N. M., and Orlando, A. W. (2023). The effects of COVID-19 on downtown
land use: evidence from four U.S. cities. J. Urban Aff. 0, 1–19. doi:10.1080/07352166. Liu, S., and Su, Y. (2021). The impact of the COVID-19 pandemic on the demand for
2023.2195665 density: evidence from the U.S. housing market. Econ. Lett. 207, 110010. doi:10.1016/j.
econlet.2021.110010
IHME (2023). COVID-19 projections. Institute for Health Metrics and Evaluation.
URL https://covid19.healthdata.org/ (accessed May 7, 23). Loomba, R. S., Aggarwal, G., Aggarwal, S., Flores, S., Villarreal, E. G., Farias, J. S., et al.
(2021). Disparities in case frequency and mortality of coronavirus disease 2019
Ilardi, A., Chiefﬁ, S., Iavarone, A., and Ilardi, C. R. (2021). SARS-CoV-2 in Italy:
(COVID-19) among various states in the United States. Ann. Med. 53, 151–159.
population density correlates with morbidity and mortality. Jpn. J. Infect. Dis. 74, 61–64.
doi:10.1080/07853890.2020.1840620
doi:10.7883/yoken.JJID.2020.200
Mandhan, S., and Gregg, K. (2023). Managing the curb – public space and use of
Ivanov, D., and Das, A. (2020). Coronavirus (COVID-19/SARS-CoV-2) and supply
curbside cafes during the Coronavirus pandemic. Cities 132, 104070. doi:10.1016/j.
chain resilience: a research note. IJISM 13, 90. doi:10.1504/IJISM.2020.107780
cities.2022.104070
Ivanov, D. (2020). Predicting the impacts of epidemic outbreaks on global supply
McFadden, B. R., Malone, T., Kecinski, M., and Messer, K. D. (2021). COVID-19
chains: a simulation-based analysis on the coronavirus outbreak (COVID-19/SARS-CoV-
induced stigma in U.S. consumers: evidence and implications. Am. J. Agric. Econ. 103,
2) case - ScienceDirect.
486–497. doi:10.1111/ajae.12188
Jamal, Y., Gangwar, M., Usmani, M., Adams, A. E., Wu, C.-Y., Nguyen, T. H., et al.
Megahed, N. A., and Ghoneim, E. M. (2020). Antivirus-built environment: lessons
(2022). Identiﬁcation of thresholds on population density for understanding
learned from Covid-19 pandemic. Sustain. Cities Soc. 61, 102350. doi:10.1016/j.scs.2020.
transmission of COVID-19. Geohealth 6, e2021GH000449. doi:10.1029/2021GH000449
102350
Jo, Y., Hong, A., and Sung, H. (2021). Density or connectivity: what are the main
Mehta, V. (2011). Small businesses and vitality of main street. J. Archit. Plan. Res. 28
causes of the spatial proliferation of COVID-19 in Korea? Int. J. Environ. Res. Public
(4), 271–291.
Health 18, 5084. doi:10.3390/ijerph18105084
Monte, L. M., 2021. Historical look at unemployment, sectors shows magnitude of
Joiner, A., McFarlane, C., Rella, L., Uriarte-Ruiz, M., Grover, J., Godfrey, A., et al.
COVID-19 impact on economy [WWW Document]. Census.gov. URL https://www.
(2022). Problematising density: cOVID-19, the crowd, and urban life. Soc. Cult. Geogr.
census.gov/library/stories/2021/03/putting-economic-impact-of-pandemic-in-context.
0, 1–7. doi:10.1080/10903127.2022.2137744
html (accessed May.2.22).

Frontiers in Built Environment 21 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

Moosavi, J., Fathollahi-Fard, A. M., and Dulebenets, M. A. (2022). Supply chain Siegrist, M., Luchsinger, L., and Bearth, A. (2021). The impact of trust and risk
disruption during the COVID-19 pandemic: recognizing potential disruption perception on the acceptance of measures to reduce COVID-19 cases. Risk Anal. 41,
management strategies. Int. J. Disaster Risk Reduct. 75, 102983. doi:10.1016/j.ijdrr. 787–800. doi:10.1111/risa.13675
2022.102983
Sun, Y., Xie, J., and Hu, X. (2022). Detecting spatial clusters of coronavirus infection
Motoyama, Y. (2022). Is COVID-19 causing more business closures in poor and across London during the second wave. Appl. Spat. Anal. 15, 557–571. doi:10.1007/
minority neighborhoods? Econ. Dev. Q. 36, 124–133. doi:10.1177/08912424221086927 s12061-021-09413-3
Nanda, A., Xu, Y., and Zhang, F. (2021). How would the COVID-19 pandemic Sun, Z., Zhang, H., Yang, Y., Wan, H., and Wang, Y. (2020). Impacts of geographic
reshape retail real estate and high streets through acceleration of E-commerce and factors and population density on the COVID-19 spreading under the lockdown
digitalization? J. Urban Manag. 10, 110–124. doi:10.1016/j.jum.2021.04.001 policies of China. Sci. Total Environ. 746, 141347. doi:10.1016/j.scitotenv.2020.141347
Newbold, S. C., Finnoff, D., Thunström, L., Ashworth, M., and Shogren, J. F. (2020). Sutton, S. A. (2010). Rethinking commercial revitalization: A neighborhood small
Effects of physical distancing to control COVID-19 on public health, the economy, and business perspective. Econ. Dev. Q. 24, 352–371. doi:10.1177/0891242410370679
the environment. Environ. Resour. Econ. 76, 705–729. doi:10.1007/s10640-020-00440-1
Swanson, D., and Santamaria, L. (2021). Pandemic supply chain research: A
Nguyen, Q. C., Huang, Y., Kumar, A., Duan, H., Keralis, J. M., Dwivedi, P., et al. structured literature review and bibliometric network analysis. Logistics 5, 7. doi:10.
(2020). Using 164 million google street view images to derive built environment 3390/logistics5010007
predictors of COVID-19 cases. Int. J. Environ. Res. Public Health 17, 17. doi:10.
Tang, F., Feng, Y., Chiheb, H., and Fan, J. (2021). The interplay of demographic
3390/ijerph17176359
variables and social distancing scores in deep prediction of U.S. COVID-19 cases. J. Am.
Norris, C. L., Taylor, J. S., and Taylor, D. C. (2021). Pivot! How the restaurant industry Stat. Assoc. 116, 492–506. doi:10.1080/01621459.2021.1901717
adapted during COVID-19 restrictions. Int. Hosp. Rev. 35, 132–155. doi:10.1108/IHR-
Tchicaya, A., Lorentz, N., Leduc, K., and Lanchy, G. (2021). COVID-19 mortality with
09-2020-0052
regard to healthcare services availability, health risks, and socio-spatial factors at
Park, K., Ewing, R., Sabouri, S., and Larsen, J. (2019). Street life and the built department level in France: A spatial cross-sectional analysis. PLOS ONE 16,
environment in an auto-oriented US region. Cities 88, 243–251. doi:10.1016/j.cities. e0256857. doi:10.1371/journal.pone.0256857
2018.11.005
Teller, J. (2021). Urban density and covid-19: Towards an adaptive approach. Build.
Paul, A., Englert, P., and Varga, M. (2021). Socio-economic disparities and COVID- Cities 2 (1), 150–165. doi:10.5334/bc.89
19 in the USA. J. Phys. Complex 2, 035017. doi:10.1088/2632-072X/ac0fc7
Thilmany, D., Canales, E., Low, S. A., and Boys, K. (2021). Local food supply chain
Perone, G. (2021). The determinants of COVID-19 case fatality rate (CFR) in the dynamics and resilience during COVID-19. Appl. Econ. Perspect. Policy 43, 86–104.
Italian regions and provinces: an analysis of environmental, demographic, and doi:10.1002/aepp.13121
healthcare factors. Sci. Total Environ. 755, 142523. doi:10.1016/j.scitotenv.2020.
Thunström, L., Newbold, S. C., Finnoff, D., Ashworth, M., and Shogren, J. F. (2020).
142523
The beneﬁts and costs of using social distancing to ﬂatten the curve for COVID-19.
Proksch, G., Sun, F., Whittington, J., and Craddock, E. (2021). COVID-19 rapid J. Beneﬁt-Cost Analysis 11, 179–195. doi:10.1017/bca.2020.12
response: design determinants of Seattle food retail business continuity. Proc. 109th
Toxvaerd, F. M. (2020). “Equilibrium social distancing,” in Cambridge working papers
ACSA Annu. Meet.
in economics 2021 (Cambridge: Faculty of Economics, University of Cambridge).
Pujawan, N., and Bah, A. (2021). Supply chains under COVID-19 disruptions:
Tribby, C. P., and Hartmann, C. (2021). COVID-19 cases and the built environment:
literature review and research agenda. Supply Chain Forum 23, 81–95. doi:10.1080/
initial evidence from New York city. Prof. Geogr. 73, 365–376. doi:10.1080/00330124.
16258312.2021.1932568
2021.1895851
Pivo, G., and Fisher, D. J. (2011). The walkability premium in commercial real estate
Tsou, M.-H., Xu, J., Lin, C.-D., Daniels, M., Embury, J., Park, J., et al. (2023).
investments. Real Estate Econ. Am. Real Estate Urban Econ. Assoc. 39 (2), 185–219.
Analyzing spatial-temporal impacts of neighborhood socioeconomic status variables on
doi:10.1111/j.1540-6229.2010.00296.x
COVID-19 outbreaks as potential social determinants of health. Ann. Am. Assoc. Geogr.
Qeadan, F., Mensah, N. A., Tingey, B., Bern, R., Rees, T., Madden, E. F., et al. (2021). 113, 891–912. doi:10.1080/24694452.2022.2149460
The association between opioids, environmental, demographic, and socioeconomic
US Department of Heatlh and Human Services (2023). Fact sheet: End of the COVID-
indicators and COVID-19 mortality rates in the United States: an ecological study at the
19 public health emergency. HHS.gov. URL https://www.hhs.gov/about/news/2023/05/
county level. Archives Public Health 79, 101. doi:10.1186/s13690-021-00626-z
09/fact-sheet-end-of-the-covid-19-public-health-emergency.html (accessed May 7, 23).
Rahman, Md.S., Paul, K. C., Rahman, Md.M., Samuel, J., Thill, J.-C., Hossain, Md.A.,
U.S. Small Business Administration (2023). COVID-19 relief option. URL https://
et al. (2023). Pandemic vulnerability index of us cities: A hybrid knowledge-based and
www.sba.gov/funding-programs/loans/covid-19-relief-options (accessed May 7, 23).
data-driven approach. Sustain. Cities Soc. 95, 104570. doi:10.1016/j.scs.2023.104570
Verma, R., Yabe, T., and Ukkusuri, S. V. (2021). Spatiotemporal contact density
Ramani, A., and Bloom, N. (2021). The donut effect of covid-19 on cities. Work.
explains the disparity of COVID-19 spread in urban neighborhoods. Sci. Rep. 11, 10952.
Pap. Ser. doi:10.3386/w28876
doi:10.1038/s41598-021-90483-1
Rice, L. (2020). After Covid-19: urban design as spatial medicine. Urban Des. Int. 28,
Wang, J., Wu, X., Wang, R., He, D., Li, D., Yang, L., et al. (2021). Review of
97–102. doi:10.1057/s41289-020-00142-6
associations between built environment characteristics and severe Acute respiratory
Rosenthal, S. S., Strange, W. C., and Urrego, J. A. (2022). JUE insight: are city Syndrome coronavirus 2 infection risk. Int. J. Environ. Res. Public Health 18, 7561.
centers losing their appeal? Commercial real estate, urban spatial structure, and doi:10.3390/ijerph18147561
COVID-19. J. Urban Econ. JUE Insights COVID-19 Cities 127, 103381. doi:10.1016/
Wang, S., Lee, C. L., and Song, Y. (2022a). The COVID-19 sentiment and ofﬁce
j.jue.2021.103381
markets: evidence from China. Buildings 12, 2100. doi:10.3390/buildings12122100
Russo, R. G., Ali, S. H., Mezzacca, T. A., Radee, A., Chong, S., Kranick, J., et al.
Wang, B., Williams, M., Duarte, F., and Zheng, S. (2022b). Demand for social
(2022). Assessing changes in the food retail environment during the COVID-19
interactions: evidence from the restaurant industry during the COVID-19 pandemic.
pandemic: opportunities, challenges, and lessons learned. BMC Public Health 22,
Journal of Regional Science 62, 830–857. doi:10.1111/jors.12585
778. doi:10.1186/s12889-022-12890-x
Washington Governor (2020d). Inslee announces statewide restrictions for four-weeks.
Rybczynski, W. (2010). “Makeshift metropolis: ideas about cities,” in 1st Scribner
Washington Governor. URL https://www.governor.wa.gov/news-media/inslee-
hardcover ed. (New York: Scribner).
announces-statewide-restrictions-four-weeks (accessed 21 April, 23).
Saelens, B. E., and Handy, S. L. (2008). Built environment correlates of walking: A
Washington Governor (2021). Inslee announces “Healthy Washington–roadmap to
review. Med. Sci. sports Exerc. 40, S550–S566. doi:10.1249/MSS.0b013e31817c67a4
recovery. [WWW Document]. Washington Governor. URL https://www.governor.wa.
Sarmadi, M., Moghanddam, V. K., Dickerson, A. S., and Martelletti, L. (2021). gov/news-media/inslee-announces-healthy-washington%E2%80%93roadmap-
Association of COVID-19 distribution with air quality, sociodemographic factors, recovery (accessed 21 April, 23).
and comorbidities: an ecological study of US states. Air Qual. Atmos. Health 14,
Washington Governor (2020b). Inslee announces “stay home, stay Healthy” order.
455–465. doi:10.1007/s11869-020-00949-w
Washington Governor. URL https://www.governor.wa.gov/news-media/inslee-
Schmiege, D., Haselhoff, T., Ahmed, S., Anastasiou, O. E., and Moebus, S. (2023). announces-stay-home-stay-healthy%C2%A0order (accessed 21 April, 23).
Associations between built environment factors and SARS-CoV-2 infections at the
Washington Governor (2020a). Inslee issues COVID-19 emergency proclamation.
neighbourhood level in a metropolitan area in Germany. J. urban health Bull. N. Y.
Washington Governor.
Acad. Med. 100, 40–50. doi:10.1007/s11524-022-00708-5
Washington Governor (2020c). Safe start phased reopening. URL https://www.
Sedov, D. (2022). Restaurant closures during the COVID-19 pandemic: A descriptive
governor.wa.gov/sites/default/ﬁles/SafeStartPhasedReopening.pdf (accessed
analysis. Econ. Lett. 213, 110380. doi:10.1016/j.econlet.2022.110380
21 April, 23).
Shi, L., Hao, Y., Lv, S., Cipcigan, L., and Liang, J. (2021). A comprehensive charging
WHO (2020). WHO Director-General’s opening remarks at the media brieﬁng on
network planning scheme for promoting EV charging infrastructure considering the
COVID-19 - 11 March 2020. URL https://www.who.int/director-general/speeches/
Chicken-Eggs dilemma. Res. Transp. Econ. 88, 100837. doi:10.1016/j.retrec.2020.
detail/who-director-general-s-opening-remarks-at-the-media-brieﬁng-on-covid-19—
100837
11-march-2020 (accessed May 7, 23).

Frontiers in Built Environment 22 frontiersin.org
Sun et al. 10.3389/fbuil.2023.1212244

Wong, D. W. S., and Li, Y. (2020). Spreading of COVID-19: density matters. PLOS Yuan, J., Zou, H., Xie, K., and Dulebenets, M. A. (2021). An assessment of social
ONE 15, e0242398. doi:10.1371/journal.pone.0242398 distancing obedience behavior during the COVID-19 post-epidemic period in China: A
cross-sectional survey. Sustainability 13, 8091. doi:10.3390/su13148091
Wu, Y., Wei, Y. D., and Liu, M. (2022). Urban form and spatiotemporal vulnerability
of local communities to covid-19. Geogr. Rev. 0, 482–501. doi:10.1080/00167428.2022. Zhai, W., and Yue, H. (2022). Economic resilience during COVID-19: an insight from
2155519 permanent business closures. Environ. Plan. A 54, 219–221. doi:10.1177/
Yenerall, J., Jensen, K., Chen, X., and Edward Yu, T. (2022). COVID-19 risk 0308518X211055181
perception and restaurant utilization after easing in-person restrictions. Food Policy Zhang, X., Sun, Z., Ashcroft, T., Dozier, M., Ostrishko, K., Krishan, P., et al. (2022).
107, 102206. doi:10.1016/j.foodpol.2021.102206 Compact cities and the Covid-19 pandemic: Systematic review of the associations
Yost, E., and Cheng, Y. (2021). Customers’ risk perception and dine-out motivation between transmission of Covid-19 or other respiratory viruses and population density
during a pandemic: insight for the restaurant industry. Int. J. Hosp. Manag. 95, 102889. or other features of neighborhood design. Health & Place 76. doi:10.1016/j.healthplace.
doi:10.1016/j.ijhm.2021.102889 2022.102827

Frontiers in Built Environment 23 frontiersin.org

File and source

File: economic-resilience-during-covid-19-the-case-of-food-retail-businesses-295f9603.pdf
Size: 2,768,470 bytes
SHA-256: 295f9603ca5756bdc65c4a7419370d10d7c3479ed9bc056bf289fd0ddf44d731
Our copy: economic-resilience-during-covid-19-the-case-of-food-retail-businesses-295f9603.pdf
Original: www.frontiersin.org