Defying the Food Desert, Food Swamp, and Supermarket Redlining Stereotypes in Detroit: Comparing the Distribution of Food Outlets in 2013 and 2023

Abstract

Despite the numerous food studies conducted in Detroit, none have assessed changes in the food landscape over a decade. No previous study has systematically analyzed food store closures in the city either. We will address these oversights by examining the distribution of food outlets in the city ten years apart. This paper probes the following questions: (1) How has the distribution of Detroit’s food outlets changed in the decade between 2013 and 2023? (2) Does Detroit fit the definition of a food desert in 2013 or 2023? (3) Does Detroit fit the definition of a food swamp in 2013 or 2023? (4) Has supermarket redlining occurred in Detroit in 2013 or 2023? (5) How is population decline related to food outlet distribution? (6) How do food store closures impact food store distribution? We conducted exhaustive searches to collect information on thousands of food outlets from Data Axle, Google, and Bing. The data were analyzed and mapped in SPSS 28 and ArcGIS 10.8. We compared 3499 food outlets identified in 2013 with 2884 identified in 2023. We expanded our search for food outlets in 2023 and found an additional 611 food outlets in categories not studied in 2013. The study’s findings are significant as they unearth evidence of extensive population decline—driven by Black flight—and a vanishing food infrastructure. Detroit lost more than 600 food outlets between 2013 and 2023, a staggering number that underscores the severity of the issue. Moreover, in 2023, we documented food store closures and found 1305 non-operational or closed food outlets in the city. Regardless of the neighborhood’s racial composition, the household median income, or the educational attainment of residents, food store closures were widespread in 2023; 27.3% of the food outlets identified that year were defunct. Despite the massive food store closures, Detroit did not fit the description of a food desert; each of the city’s 54 neighborhoods had between 7 and 300 food outlets. The food swamp thesis did not accurately describe the city either, as supermarkets/large grocery stores were intermingled with convenience and corner stores in both study periods. The data did not find evidence of supermarket redlining, as supermarkets/large grocery stores were found in formerly redlined neighborhoods alongside dollar stores and variety stores in both study periods.

1. Introduction

Food insecurity is a ubiquitous problem affecting people globally. In 2022, about 258 million people in 58 countries were acutely food insecure [1]. Reducing hunger and food insecurity is a core component of efforts to attain global sustainability. Consequently, the second aim of the United Nations’ Sustainable Development Goals (SDGs) is to end hunger and ensure that all people—especially the poor and vulnerable—have access to safe, nutritious, and adequate food [2]. Global comparisons show that the U.S. is not immune to the problems of food insecurity and inequitable access. The 2022 Global Food Security Index shows that the U.S. ranked 13th among the 113 countries studied. Though the U.S. had an overall score of 78.0, its score for food availability was 65.1 [3]. The food availability score reflects the fact that in 2021, an estimated 33.8 million people in the U.S. lived in food-insecure households [4]. Given the pervasiveness and urgency of the problem, this paper examines food access in Detroit, Michigan, a major post-industrial American city wrestling with many of the challenges urban areas face.

Michigan is a significant producer of agricultural products, yet food insecurity and inequitable access to food are widespread in the state. Consequently, numerous studies on food insecurity and disparities in access to food have been conducted therein. Most of these were conducted in Detroit, Flint, and Lansing [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19].

2. The Detroit Context

This study focuses on Detroit, Michigan’s largest and most famous city. Detroit has a lengthy history of food access advocacy and policymaking. It is also an interesting city to study because of its historic innovations in sustainable food systems, food insecurity, and food access. For instance, Detroit opened one of the first farmers’ markets in the U.S. in the 1840s in Cadillac Square. That market merged with another to become Eastern Market in 1891 [20,21]. Civic leaders tackled food insecurity during the economic crisis that enveloped the country during the 1890s. In 1894, the then mayor Hazen Pingree allowed residents to farm on 430 acres of the city’s vacant land for free to alleviate food shortages. Thousands of families participated in the program, which lasted until 1901. Other cities, such as New York, Boston, Chicago, Minneapolis, Seattle, Duluth, and Denver, developed similar programs [22,23].

Detroit commands attention because, in 2013, it became the largest municipality in U.S. history to file for bankruptcy. The USD 18 billion filing prompted the governor to put the city under emergency management [24,25]. Detroit is 139 square miles, with around 19 square miles of vacant land. Though the amount of vacant land in Detroit is decreasing, in 2020, the U.S. Postal Service reported 74,313 vacant residential lots, or 19.3% of the city’s total [1,26,27]. The city’s financial woes, dwindling population, and large swaths of derelict land create challenges and opportunities for food acquisition.

3. The Study

This paper presents the results of first-of-a-kind research on Detroit’s changing food landscape. It examines the following questions: (1) How has the distribution of Detroit’s food outlets changed in the decade between 2013 and 2023? (2) Does Detroit fit the definition of a food desert in 2013 or 2023? (3) Does Detroit fit the definition of a food swamp in 2013 or 2023? (4) Has supermarket redlining occurred in Detroit in 2013 or 2023? (5) How does population decline relate to food outlet distribution? (6) How do food store closures impact food store distribution?

The paper examines three popular theses that scholars, policymakers, and food advocates use to guide their studies or explain urban food access: food deserts, food swamps, and supermarket redlining. This study demonstrates why none describe Detroit accurately or help us understand the city’s food landscape well. Instead, we argue that population decline, extensive food store closures, and the intermingling of different types of food outlets help explain the contemporary distribution of food outlets in Detroit.

The study is significant because of the many firsts it accomplishes. This is the first study to analyze a comprehensive set of food establishments in a city in two different periods. This is also the first study to examine the relationship between prolonged and extensive population decline and a city’s food infrastructure. It is the first food-access study to examine the relationship between Black flight or outmigration from a city and the distribution of food outlets. This study is the first to examine arguments implied in the food swamp and supermarket redlining theses in Detroit and propose the concept of intermingling to describe the distribution of food outlets. Moreover, it is one of a handful of studies that examine citywide food store closures and their impact on food access. It is the first to assess food store closures in Detroit comprehensively.

4. Literature Review

4.1. Conceptualizing Detroit’s Food Environment

4.1.1. The Food Desert Label

The concept of a food desert originated in the United Kingdom in the 1990s to describe suburban developments that lacked food stores and other community amenities [28,29,30]. The term food desert is now commonly used to describe communities where residents lack ready access to fresh, healthy, and affordable foods. In the United States, poor urban neighborhoods with high percentages of People of Color are often described as food deserts [31,32,33,34,35,36,37].

Research describing Detroit as a food desert is plentiful [31,38,39,40,41,42,43,44,45]. Though scholars criticize this characterization of the city [46,47,48,49,50,51,52,53,54], Detroit has difficulty shedding the image of a vast urban wasteland bereft of places to acquire healthy, nutritious, and affordable food.

Analyses of Detroit’s food landscape began appearing in the early 2000s. For instance, Pothukuchi published an article in 2005 on efforts to get Detroit’s corner stores to sell healthy and affordable foods [55]. However, food desert research has received the greatest attention and publicity. Among the earliest researchers, Zenk et al. [31] found that residents of the poorest neighborhoods lived farther from supermarkets than residents of wealthier neighborhoods. They also found that residents of high-percentage Black neighborhoods resided farther from supermarkets than those in low-percentage Black neighborhoods. Zenk and colleagues [56] broadened their sample to include chain grocery stores, large independent groceries, “mom-and-pop” stores, and liquor stores. Zenk et al. [57] later added fast-food restaurants to the food outlets analyzed in their Detroit food study.

Another food landscape study found that Detroiters travel twice as far to reach a “mainstream” grocery store as they do to reach a fringe food establishment [32]. Santarossa et al. [18] studied the proximity to healthy grocery stores in Detroit, while LeDoux and Vojnovic [17] analyzed the consequences of supermarket decentralization in the city.

Jang et al. [54] studied 2635 food stores in Detroit and neighboring suburbs, but concentrated on only three food outlet types—supermarkets, grocery stores, and convenience stores. However, Taylor and Ard [5] examined an extensive list of food venues in the city, studying the distribution of 3499 food establishments classified into 34 categories. Eckert and Vojnovic [11] examined restaurant choice and travel distance on the city’s Lower Eastside, while Vojnovic et al. [15] studied food shopping behavior and travel patterns of Detroit’s Lower Eastside residents. Scholars have also studied Detroit’s urban agriculture and community gardening sectors [13,16,50,51,52,58,59].

Despite the prevalence of the food desert label, scholars critical of that framing argue that the food desert concept is misleading. Not only does it evoke images of spaces devoid of venues to purchase healthy foods, but many studies do not recognize the variety of small food outlets that carry the healthy foods that urban residents want [1,5,7,8,60,61,62]. Skeptics also claim that the food desert narrative fails to recognize the agency that neighborhood residents employ when they adopt food sovereignty approaches, grow their food, or find ways to obtain healthy and affordable food [5,50,52].

Food desert critics also contend that emphasizing supermarkets and large grocery stores as the lone or prime indicator of healthy food access warps our perceptions of local food environments and may underestimate food availability [5,7,8,46,61,63,64,65,66,67]. The emphasis on supermarkets and large grocery stores also overshadows the critical roles that independent grocers and small ethnic grocery stores play in providing residents with affordable and culturally desirable foods [6,62,68].

Though parts of Detroit are underserved by food retailers, portraying the whole city as a food desert is problematic. When researchers from Data Driven Detroit analyzed National Establishment Time Series data from 2010, they found 115 grocery stores in the city. The researchers contended that the city had small pockets lacking easy access to grocery stores [49]. Using the United States Department of Agriculture’s food desert definition, the researchers estimated that only about 10% of the city containing about 90,000 people matched the definition of food desert [69].

Wang [70] used Southeast Michigan Council of Governments (SEMCOG) data to analyze the distribution of grocery stores in metropolitan Detroit and found that Detroit had a higher density of grocery stores than other parts of the Tri-County area. While there were 0.134 grocery stores per 1000 residents in Detroit, the density was 0.113 grocery stores per 1000 residents in Macomb, Oakland, and Wayne (the suburbs outside of Detroit) Counties.

4.1.2. Food Swamp

Researchers also use the term food swamp to describe low-income, urban neighborhoods saturated with fast-food restaurants, convenience stores, mini-marts, gas stations, and liquor stores selling unhealthy foods. As the argument goes, establishments selling energy-dense foods in food swamps overwhelm those selling healthy food options. Consequently, supermarkets and large grocery stores are scarce or nonexistent in food swamps [53,71,72,73,74,75,76]. The USDA’s research scientists also use this concept [77,78,79].

4.1.3. Supermarket Redlining

Increasingly, scholars and activists have used the term supermarket redlining to describe the distribution of supermarkets and large grocery stores in cities [70,80,81,82,83]. Proponents of the term suggest that supermarkets and large grocery stores cease operations and close in low-income, inner-city neighborhoods and relocate to high-income or suburban areas. They argue that such economic decisions are racialized and contribute to the inequitable distribution of supermarkets and large grocery stores [79,80,84,85,86,87,88,89].

Proponents of this thesis contend that supermarket redlining is an outgrowth of the residential segregation and neighborhood redlining initiatives promoted by the Home Owners’ Loan Corporation (HOLC) during the 1930s and 1940s [17,80,90]. Taylor [91] discusses redlining, blockbusting, government-sanctioned residential segregation, and White flight extensively. HOLC developed maps of 239 U.S. cities that labeled neighborhoods with large Black populations as “hazardous” and enclosed them in red lines. All-White middle- and upper-income neighborhoods were delineated in green. Today, many formerly redlined neighborhoods are poor, disinvested, and have high percentages of People of Color as residents.

Zhang and Ghosh [89] developed a supermarket redlining index and a redlining impact model, showing that closing supermarkets in poor People of Color neighborhoods made it challenging for residents to find healthy foods. Shannon [92] found that a few supermarkets and superstores in Atlanta were in formerly redlined neighborhoods. Shannon [88] also associates supermarket redlining with the proliferation of dollar stores in Communities of Color. Several other investigators have also focused on supermarket redlining [87,93,94,95,96,97,98,99,100].

4.2. New Conceptualizations of Detroit’s Food Environment

4.2.1. Population Size, Decline, and Demographic Shifts

We argue that the extensive and prolonged outmigration of people from Detroit is a significant driver of which kinds of food outlets are located in the city. We contend that population decline is associated with the neighborhood loss of supermarkets, large grocery stores, and other food outlets. Population size is an essential determinant of supermarkets and large grocery stores’ location. Neighborhoods with few residents (or not attracting many visitors) or neighborhoods experiencing significant population declines are unlikely to attract or retain supermarkets and large grocery stores.

Once a dominant force in automobile design and manufacturing, Detroit has experienced significant industrial and retail decline, massive White flight, and extensive loss of housing infrastructure. This paper will also document extensive Black flight characterized by pronounced outmigration from most of the city’s neighborhoods. Detroit had a population of 1849,568 in 1950, but the population had fallen to 639,111 in 2020 [101,102]. The population decline continues; by 2022, the city’s population was down to 620,376 [103].

Table 1. Comparison of population characteristics of Michigan and the City of Detroit.

Population Characteristics	Michigan					City of Detroit
	2010		2020			2010		2020
	Population	Percent	Population	Percent	Percent Change	Population	Percent	Population	Percent	Percent Change
Total Population	9,883,640	100.00	10,077,331	100.00	2.0	713,777	100.00	639,111	100.00	−10.46
White alone (not Latinx or Hispanic)	7,569,939	76.59	7,295,651	72.40	−3.6	55,604	7.79	60,770	9.51	9.29
Black alone (not Latinx or Hispanic)	1,383,756	14.00	1,358,458	13.48	−1.8	586,573	82.18	493,212	77.17	−15.92
Latinx or Hispanic	436,358	4.41	564,422	5.60	29.3	48,679	6.82	51,269	8.02	5.32
Native American or Alaska Native	54,665	0.55	47,406	0.47	−13.3	1927	0.27	1399	0.22	−27.40
Asian	236,490	2.39	332,288	3.30	40.5	7436	1.04	10,085	1.58	35.62
Native Hawaiian or Pacific Islander	2170	0.02	2603	0.03	20.0	82	0.01	111	0.02	35.37
Other	9866	0.10	37,183	0.37	276.9	994	0.14	3066	0.48	208.45
Two or more races	190,396	1.93	439,320	4.36	130.7	12,482	1.75	19,199	3.00	53.81

Compiled from [102,103,104,105].

shows how the city’s population changed between 2010 and 2020 [26,102,104]. While Michigan’s population grew by 2% during the intercensal period, Detroit’s population declined by 10.5%. Despite the steep drop in the number of Black residents, Detroit remains a predominantly Black city—77.2% of the residents are Black. Whites comprise the second-largest racial group, at 9.3%. Latinx/Hispanic residents make up 8% of the population. The poverty rate is 31.8%, and the median household income is USD 34,762. According to the census, 82.6% of the city’s residents graduated from high school, and 16.2% have a bachelor’s degree [26,102].

4.2.2. Vanishing Food Infrastructure and Store Closures

We contend that mass store closures and a vanishing food infrastructure exist in Detroit. This is evident across all food sectors and neighborhoods, regardless of racial or socio-economic characteristics. Researchers use the term vanishing food infrastructure to describe phenomena observed in other Michigan cities, such as Flint [61], Lansing and East Lansing [8], and Saginaw [106], where they observe extensive closures of food venues. Studies examining this phenomenon have found numerous closed restaurants, small groceries, corner stores, pharmacies, dollar stores, variety stores, urban farms and community gardens, food pantries, and soup kitchens. Scholars have also investigated and found extensive grocery store closures in rural communities [107,108].

Despite the limited attention paid to food store closures, American cities have been experiencing this phenomenon since the 1960s [61,85,89,109,110,111,112,113,114]. The trend accelerated as food store closures spiked during the Coronavirus-19 (COVID-19) pandemic and continued in the post-pandemic era.

4.2.3. The Intermingling of Food Outlet Types

To date, the food access literature has not identified configurations of food outlets that reflect what is commonplace in Detroit. The city presents numerous examples of neighborhoods where supermarkets and large grocery stores are intermingled with dollar stores, variety stores, corner and convenience stores, and other food outlets. There is ample evidence that rather than being driven out of neighborhoods or commercial strips by small grocery stores, dollar stores, variety stores, or fast-food restaurants, many of Detroit’s supermarkets and large grocery stores co-mingle with them.

5. Methods

5.1. Defining the Food Environment

We collected data on various food venues in 2013 and 2023 to find out where Detroiters obtain food and how the food landscape changed over a decade (see Appendix A). Researchers studying Atlanta examined how residents’ proximity to food stores changed over five years. Those investigators focused on six census tracts and stores that accepted government food assistance program benefits cards [92,115]. Our study differs from the Atlanta study in that ours is a citywide study that examines thousands of food outlets. The food outlets studied are defined in Appendix A; the appendix also contains the sources used to help construct the definitions [5,7,8,60,67,116,117].

5.2. Data Collection and Sources

5.2.1. 2013 Data Collection

Between 2011 and 2013, we collected data from multiple sources and merged them to create a dataset. We use a multi-method approach because no existing repository contains all the food venues in a given city. Hence, reliance on one source can lead to severe undercounts of outlets. Other researchers use similar techniques [117,118,119].

We collected data on food venues from two international databases: ReferenceUSA and Orbis. Other food access studies relying on ReferenceUSA include Lisabeth et al. [117], Liese et al. [118], Raja et al. [60], Taylor and Ard [5], and Rybarczyk et al. [6]. We identified food outlets using the Standard Industrial Classification (SIC) division codes. Other food access researchers have used SIC codes to identify venues [33,60,117]. In addition to the division codes, we used the major group, industry group, and industry codes to identify food retailers and businesses such as mass merchandisers, supercenters, and variety stores that sell substantial amounts of food, even though food is not the primary business.

We also collected data from the Michigan Department of Agriculture, the Michigan Department of Human Services, the Michigan Farmers’ Markets Association, Local Harvest, FoodPantries.org, the Detroit Public Schools website, the Detroit Yellow Pages business directory, Eastern Market’s vendor directory, and local nonprofits’ listings of community gardens. Finally, we also used Google Street View, Bing, emails, and telephone calls to businesses to identify additional food outlets, get accurate addresses, and obtain the latitude and longitude of food outlets that were not in any databases. Fact-checking allowed us to identify duplicates, incorrect addresses, and defunct, inaccurately classified, and non-food-related businesses. Duplicated entries and defunct or non-food-related outlets were removed from the database.

We organized the food venues into 34 categories. Ten categories of food outlets were identified and classified using the Food Marketing Institute’s (FMI) typology [120]. Block et al. use the same definitions for full-service and fast-food restaurants [121]. Our team identified and classified the remaining food outlet types. Appendix A identifies all the food categories studied in 2013 and additional categories studied in 2023.

5.2.2. 2023 Data Collection

We expanded the types of food outlets we searched for when we gathered data on Detroit’s food outlets from 2021 to 2023 (see Appendix A). We used Data Axle (formerly ReferenceUSA) as the primary source of information on food outlets. We followed the abovementioned techniques to find food outlets. We used supplementary sources, such as the Census Geocoder [122], Geocode by Awesome Table [123], and social media sites (such as Facebook and Instagram) to help identify new food venues and verify addresses. The operating food outlets in our database were classified into 57 categories. We used the FMI typology [5,7,120,121] to guide the classification (Appendix A).

5.2.3. Mapping

We used ArcGIS Pro 10.8.1 to plot the exact coordinates of the food outlets on the maps. We used the U.S. Census Bureau’s Open Data mapping tool to obtain Detroit’s boundaries. We obtained the Detroit neighborhood shapefile from the City of Detroit Open Data Portal [124]. Redlining maps of Detroit were obtained from the University of Richmond’s Mapping Inequality project site [125]. We used the intersect tool to combine the Detroit neighborhood shapefile with the Michigan census tract shapefiles. We combined neighborhood information and 2020 census tract data, a technique used by other investigators [6,7,8,31,61,116,126,127]. The food retailer information was added as a comma-separated value file.

We also employed interpolation techniques when necessary. We did this for four census tracts partially contained in a neighborhood. We snipped the census tracts to match the neighborhood or city boundary and used proportional allocation to estimate the population size and racial/ethnic composition for the partial census tracts. Lastly, we joined the census tract shapefile and the food retailers’ shapefile to generate maps depicting the location of each food outlet within the census tracts. Other researchers rely on this technique [7,61,128,129,130].

5.2.4. Statistical Analyses

We analyzed the relationship between the dependent variables (number of food outlets and type of food venues in the census tracts and neighborhoods) and four independent variables (race/ethnicity, educational attainment, median household income, and population density). The 2020 census data were used to create a categorical variable to analyze the percentage of Black residents the neighborhoods contained: 0–40%, 41–70%, 71–90%, and over 90%. We also analyzed two other racial/ethnic groups—Whites and Latinx/Hispanics. However, because less than 40% of the residents of most neighborhoods were either White or Latinx/Hispanic, we analyzed percent White and percent Latinx/Hispanic as continuous variables.

Because the data on food outlets collected in each study period had a non-normal distribution, Mann–Whitney U tests were performed to determine if the number of food outlets varied significantly from 2013 to 2023. A similar analysis was conducted to determine whether there were significant differences in the distribution of neighborhood food establishments over the study period. The Mann–Whitney U test statistic is represented in the following equations as U. It is the smaller value of U₁ and U₂ below. R₁ is the sum for group 1, and R₂ is the sum for group 2 [131].

$$U_1 = n_1{n}_2 + {\displaystyle \frac{n_1\left(n_1 + 1\right)}{2}} - R_1 \mathrm{and} U_2 = n_1{n}_2 + {\displaystyle \frac{n_1\left(n_1+1\right)}{2}} - R_2$$

We used Poisson and negative binomial regression analyses to study the relationship between the dependent (number of food outlets in a census tract) and independent variables (racial composition of the census tract, population density (pop/km²), median household income, and percentage of population with a high school education). The first independent variable was categorical, while the last three were continuous. The dispersion of the dependent variable determined whether we ran a Poisson or negative binomial regression. Poisson regression models were used for food outlet types that were not overdispersed, while negative binomial regression was used to analyze overdispersed food vendors. Several scholars have used this approach [7,35,127,132,133,134].

The negative binomial (Poisson–gamma) regression model for observation i is:

$$\mathrm{P}\mathrm{r}\left(\mathrm{Y}=y_{i }∣{\mu }_i,\alpha \right)={\displaystyle \frac{\Gamma \left(y_i + {\alpha }^{-1}\right)}{\Gamma \left(y_i+1\right)\Gamma \left({\alpha }^{-1}\right)}} {\left({\displaystyle \frac{{\alpha }^{-1}}{{\alpha }^{-1} + {\mu }_i}}\right)}^{{\alpha }^{-1}} {\left({\displaystyle \frac{ {\mu }_i}{{\alpha }^{-1} + {\mu }_i}}\right)}^{y_{i }}$$

where

$${\mu }_i = t_i\mu$$

$$\alpha = {\displaystyle \frac{1}{\nu }}$$

The parameter µ is the mean incident rate of y for every unit examined [135].

Two models were created: Model 1 (the crude model) included only tract racial composition, and Model 2 adjusted for educational attainment, median household income, and population density (pop/km²). The full model also included interaction terms.

In the models, F is the type of food outlet, YDC represents the data collection year, CTR is the census tract racial composition, I is the median household income in USD 1000, E is the percentage of the population with a high school or equivalent degree, and PD is population density.

Model I:

log (F) = β₀ + β₁ × YDC

Model II:

log (F) = β₀ + β₁ × YDC ×+ β₂ × CTR + β₃ × I + β₄ × E + β₅ × PD

The variance inflation factor (VIF) was used to check for multicollinearity. The paper uses a VIF ≥ 2.5; Johnston et al. [136] apply this threshold in their research. The analyses were performed using IBM SPSS Statistics Version 28.

6. Results

6.1. Changes in the Overall Food Landscape—2013 and 2023 Comparisons

The study shows a pronounced change in Detroit’s food landscape over the decade. It compared eight major categories of food outlets: supermarkets and large grocery stores; small groceries and convenience stores; pharmacies, dollar and variety stores; specialty stores and vendors; restaurants and other food services; farms, community gardens, farmers’ markets, and produce vendors; emergency food assistance; and supply chain. An assessment of the food outlets studied in 2013 and 2023 shows a substantial shrinkage in the number of food establishments. We identified 3499 food venues in these categories in 2013 and 2884 in 2023. This means that there were 615 or 17.6% fewer food establishments in 2023. The number of food venues declined in most of the categories studied (see

Table 2. A comparison of the food outlets found in Detroit in 2013 and 2023.

Food Outlet Type	2013 Food Outlets		2023 Food Outlets		Percent Change
	Frequency	Percent	Frequency	Percent	Frequency	Percent
All food venues:	3499	100.0	2884	100.0	−615	−17.6
Supermarkets and large grocery stores:	96	2.7	74	2.1	−22	−22.9
Traditional supermarkets and large groceries	63	1.8	42	1.2	−21	−33.3
Limited-assortment stores	26	0.7	14	0.4	−12	−46.2
Mass merchandisers	1	0.0	1	0.0	0	0.0
Supercenters	1	0.0	1	0.0	0	0.0
Fresh-format supermarkets	1	0.0	10	0.3	9	900.0
Super warehouses	1	0.0	4	0.1	3	300.0
Superstores	3	0.1	2	0.1	−1	−33.3
Small groceries and convenience stores:	1110	31.7	726	20.7	−384	−34.6
Gas stations with food	371	10.6	283	8.1	−88	−23.7
Liquor stores and party stores	460	13.1	301	8.6	−159	−34.6
Small groceries, convenience, and corner stores	279	8.0	142	4.1	−137	−49.1
Pharmacies and dollar and variety stores:	306	8.7	292	8.3	−14	−4.6
Pharmacies or drug stores	183	5.2	167	4.8	−16	−8.7
Dollar stores and variety stores	123	3.5	125	3.6	2	1.6
Specialty food stores and vendors:	279	8.0	135	3.9	−144	−51.6
Bakeries	76	2.2	51	1.5	−25	−32.9
Ice cream parlors	23	0.7	21	0.6	−2	−8.7
Health food and nutrition supplements	31	0.9	12	0.3	−19	−61.3
Meat markets and delicatessens	116	3.3	40	1.1	−76	−65.5
Food cooperatives	2	0.1	0	0.0	−2	−100.0
Confectionaries	31	0.9	11	0.3	−20	−64.5
Restaurants and other food service:	1245	35.6	1244	35.6	−1	−0.1
Full-service restaurants	618	17.7	484	13.8	−134	−21.7
Fast-food restaurants	338	9.7	354	10.1	16	4.7
Coffee, tea, and juice shops	40	1.1	92	2.6	52	130.0
Bars and clubs	185	5.3	292	8.3	107	57.8
Caterers	64	1.8	23	0.7	−41	−64.1
Farms, gardens, farmers’ markets, and produce vendors:	206	5.9	214	6.1	8	3.9
Urban farms and community gardens	92	2.6	99	2.8	7	7.6
Farmers’ markets and produce markets	61	1.7	31	0.9	−30	−49.2
Community-supported agriculture (CSA)	4	0.1	5	0.1	1	25.0
Dairies	7	0.2	3	0.1	−4	−57.1
School gardens	42	1.2	76	2.2	34	81.0
Emergency food assistance:	100	2.9	68	1.9	−32	−32.0
Food pantries or soup kitchens	98	2.8	55	1.6	−43	−43.9
Food banks/distribution	2	0.1	13	0.4	11	550.0
Supply chain:	157	4.5	131	3.7	−26	−16.6
Wholesalers	97	2.8	81	2.3	−16	−16.5
Manufacturers, processors	31	0.9	28	0.8	−3	−9.7
Distributors	29	0.8	22	0.6	−7	−24.1
Outlets studied only in 2023:			Frequency	Percent
Additional food venues—total:			611	17.5
Supermarkets and large grocery stores:			3	0.1
Wholesale clubs			3	0.1
Specialty food stores and vendors:			2	0.1
Condiments and spices			2	0.1
Restaurants and other food service:			202	5.8
Takeout establishments			139	4.0
Banquet halls and hotels			63	1.8
Farms, gardens, farmers’ markets, and produce vendors:			33	0.9
Market prepared food			8	0.2
Market stores			16	0.5
Market produce vendors			9	0.3
Mobile food sources:			17	0.5
Mobile produce distributor			1	0.0
Food trucks			9	0.3
Mobile food distribution			7	0.2
Attractions and amusement parks:			12	0.3
Attractions and amusement parks			8	0.2
Casinos			4	0.1
Social, religious, educational, and community services:			316	9.0
School cafeterias			124	3.5
Retirement communities and homes			7	0.2
Childcare			103	2.9
Religious institutions			30	0.9
Community centers			25	0.7
Emergency Shelter			1	0.0
Youth organizations and centers			11	0.3
College and university food venues			4	0.1
University bakery			1	0.0
University cafés and coffee shops			3	0.1
University fast-food restaurant			3	0.1
University food pantry or soup kitchen			1	0.0
University convenience stores			2	0.1
University takeout establishment			1	0.0
Gyms and health centers:			16	0.5
Fitness centers, gyms, and health centers			11	0.3
Hospitals and medical centers			5	0.1
Internet, online purchase, and delivery:			10	0.3
E-commerce, online			10	0.3

and Figure 1a–c).

During the study period, the city lost a third (or 21 of 63) of its traditional supermarkets and large grocery stores. It also lost 46.2% (or 14 of 26) of its limited-assortment food stores. While mass merchandisers and supercenters remained constant, the city gained nine new fresh-format supermarkets and three super warehouse stores (Figure 2a,b).

The small grocery and convenience store sector was devastated even more than the supermarket sector. Over the decade, the city lost 34.6% (384) of its small grocery stores, mini-marts, convenience stores, and corner stores. In other words, Detroit lost 88 (or 23.7%) of its gas stations that sold food, 159 (or 34.6%) of the liquor stores or party stores that sold food, and 137 (or 49.1%) of the small groceries, corner stores, convenience stores, and mini-marts (Figure 3a,b).

There were 144 (or 51.6%) fewer specialty stores and vendors in 2023 than in 2013. Bakeries declined by a third, and the number of health food and nutritional supplement vendors, meat markets, delicatessens, and confectionaries dropped by about 60% (Figure 4a,b).

On the surface, the total number of restaurants seems to have remained stable over the decade. However, the sector was turbulent. While there was a 21.7% (134) drop in the number of full-service restaurants, fast-food restaurants increased by 4.7% (16). The number of coffee, tea, and juice shops increased by 130% (52), and the number of bars and clubs by 57.8% (107). The number of caterers decreased by 64.1% (41) (Figure 5a,b).

There was a slight decline in the number of pharmacies and dollar and variety stores; there were 4.6% (14) fewer of these retailers in 2023 than in 2013. While 16 (8.7%) fewer pharmacies and drug stores were identified in 2023 than in 2013, the 2023 data collection found two more (1.6%) dollar stores and variety stores than were found in 2013 (Figure 6a,b).

There was a dramatic decline in emergency food assistance organizations; 32% (32) fewer of these institutions existed in 2023 than in 2013. Despite rising food insecurity, the number of food pantries and soup kitchens plummeted from 98 in 2013 to 55 in 2023, marking a decline of 43.9%. However, there were only two food banks/distribution centers in 2013, but there were 13 in 2023. This is a 650% increase (Figure 7a,b). The number of supply chain venues dropped by 16.7% from 2013 to 2023 (Figure 8a,b). However, the number of urban farms and community gardens increased by 7.6% (7) from 2013 to 2023. The number of school gardens identified also increased significantly—from 42 in 2013 to 76 in 2023—an 81% increase (Figure 9a,b).

6.2. Additional Food Outlets Studied in 2023

The 2023 data-gathering efforts collected information on several categories and types of food outlets that had not been studied a decade earlier. These new outlets added 611 venues (or 17.5%) to the total food outlets identified in 2023. In all, we identified 3495 food establishments operating in Detroit in 2023 (see Table 2 and Figure 10).

We identified 202 (5.8%) restaurants in two new categories. We found 139 takeout establishments, constituting 4% of the food outlets. In addition, 63 banquet halls and hotels were identified; these constituted 1.8% of the venues.

The 2023 data-gathering efforts also uncovered 33 food outlets in the farms, gardens, farmers’ markets, and produce vendors categories. These include 8 market-prepared-food vendors, 16 market stores, and 9 farmers’ market produce vendors. It should be noted that Eastern Market has numerous market vendors, but many are established food businesses that sell at the market. Such businesses are logged and mapped under their business name and address. The market vendors recorded in this section are small food entrepreneurs who do not operate freestanding food stores elsewhere.

There were 17 mobile food sources, most of which were food trucks (9) and mobile food vans (7). The 316 food venues in social, religious, educational, and community service organizations accounted for 9% of the food landscape. The most common food venue in this category was school cafeterias. The 124 identified constituted 3.5% of the food landscape. Religious institutions providing food were also numerous; 103 comprised 2.9% of the food environment.

6.3. Multivariate Analyses of Temporal Changes in the Citywide Food Environment

Tests of Significance and Regression Analyses

We conducted a Mann–Whitney U test to determine if there were statistically significant differences in the number of food outlets in each major food category studied in 2013 and 2023.

Table 3. Mann–Whitney U test comparing food outlets in 2013 and 2023.

Food Outlet Categories Compared in the Two Study Periods	Total Food Outlets in the City			Racial Composition of Neighborhoods
				0–40% Black Residents			41–70% Black Residents			71–90% Black Residents			91% or More Black Residents
	U	z	p-Value a	U	z	p-Value a	U	z	p-Value	U	z	p-Value	U	z	p-Value
Total	1134.500	−1.988	0.047 *	12.000	−0.893	0.432	20.000	−1.085	0.313	186.500	−0.809	0.419	133.500	−1.044	0.301
Supermarkets and large grocery stores	1203.000	−1.627	0.104	9.500	−1.345	0.202	28.000	−0.240	0.875	192.000	−0.702	0.483	152.000	−0.500	0.643
Small groceries and convenience stores	905.500	−3.397	0.001 ***	10.000	−1.218	0.268	9.500	−2.227	0.022 *	179.500	−0.987	0.324	114.000	−1.636	0.106
Pharmacies and dollar and variety stores	1382.000	−0.469	0.639	15.500	−0.332	0.755	25.500	−0.494	0.635	184.000	−0.876	0.381	158.000	−0.304	0.777
Specialty food stores and vendors	840.500	−3.837	0.000 ***	9.000	−1.418	0.202	10.500	−2.158	0.031 *	124.000	−2.437	0.015 *	122.000	−1.413	0.171
Restaurants and other food service	1384.500	−0.452	0.651	14.500	−0.488	0.639	25.500	−0.489	0.635	212.500	−0.152	0.879	167.500	−0.015	0.988
Urban farms, community gardens, farmers’ markets, and produce vendors	1283.000	−1.086	0.278	13.500	−0.654	0.530	29.500	−0.054	0.958	213.000	−0.140	0.889	116.500	−1.602	0.120
Emergency food assistance	1195.500	−1.668	0.095	13.500	−0.671	0.530	29.500	−0.059	0.958	201.000	−0.455	0.649	92.500	−2.404	0.021 *
Supply chain	1318.500	−0.887	0.375	7.500	−1.659	0.106	27.500	−0.279	0.792	185.000	−0.884	0.377	158.500	−0.303	0.777

Notes: ^a Significance levels: * p-values α ≤ 0.05; *** p-values α ≤ 0.001.

shows that three of the observed differences were significant. First, the difference between the total number of food outlets in 2013 and those in 2023 was statistically significant (U = 1134.50, p = 0.047). There was also a significant difference in the number of small groceries and convenience stores (U = 905.50, p = 0.001) and in the number of specialty food stores and vendors (U = 840.50, p = 0.000).

Furthermore, we examined whether the differences in the distribution of the food outlets were related to the racial composition of the census tracts in the two time periods. Table 3 shows no relationship between the racial composition of the census tracts and the total number of food outlets in each category.

The table also shows that the differences in the number of each food outlet in 2013 and 2023 in census tracts containing 0–40% of Black residents were insignificant. There were only a few instances where significant differences were observed for specific food outlet types. For census tracts classified as 41–70% Black in 2013 and 2023, there was a statistically significant difference in the number of small groceries and convenience stores (U = 9.50, p = 0.022) and specialty food stores and vendors (U = 10.50, p = 0.031) in the two time periods. Census tracts with 71–90% Black in both years showed a statistically significant difference in the number of specialty food stores and vendors (U = 124.00, p = 0.015). Lastly, in census tracts that were more than 90% Black in both years, there was a statistically significant difference in the number of emergency food assistance vendors (U = 92.50, p = 0.021) in 2013 and 2023.

6.4. Neighborhood Population Changes

Detroit lost more than 10% of its population over a decade, going from 713,776 residents in 2010 to 639,111 in 2020 [26,102,103,104,105]. The paper is based on the population counts and other demographic information from the 2010 and 2020 censuses, as these are the two closest census years to our study periods of 2013 and 2023.

Forty (74.1%) of Detroit’s 54 neighborhoods studied lost population between 2010 and 2020, while 14 (25.9%) gained population. The fastest-growing neighborhood was Lower Woodward; it gained over 2300 new residents. Indian Village and Lower East Central each added around 1300 residents, and Near East Riverfront added about 1100 people. The five neighborhoods experiencing the steepest population declines each lost more than 4000 residents. Connor lost almost 5200 people, Mt. Olivet lost about 4800 residents, Durfee lost around 4600 people, and Nolan’s and Tireman’s populations dropped by roughly 4400 (

Table 4. Changes in neighborhood racial/ethnic composition in Detroit, 2010–2020.

Neighborhoods	Neighborhood Demographic Changes e
	Total Population d			Number of Whites (Not Latinx or Hispanic)			Number of Latinx or Hispanics			Number of Blacks (Not Latinx or Hispanic)			Change in Percentage of White Residents		Change in Percentage of Latinx Residents		Change in Percentage of Black Residents
	2010	2020	Percent Change	2010	2020	Percent Change	2010	2020	Percent Change	2010	2020	Percent Change	2010	2020	2010	2020	2010	2020
Detroit total	713,766	639,060	−10.5	55,604	60,775	9.3	48,679	51,271	5.3	586,573	493,151	−15.9	7.8	9.5	6.8	8.0	82.2	77.2
Cerveny/Grandmont	32,769	31,225	−4.7	452	498	10.2	243	346	42.4	31,474	29,290	−6.9	1.4	1.6	0.7	1.1	96.0	93.8
Mackenzie	26,660	22,913	−14.1	370	490	32.4	235	481	104.7	25,613	21,128	−17.5	1.4	2.1	0.9	2.1	96.1	92.2
Finney	26,031	23,396	−10.1	2847	2197	−22.8	275	410	49.1	22,166	19,907	−10.2	10.9	9.4	1.1	1.8	85.2	85.1
Evergreen	25,277	24,205	−4.2	588	441	−25.0	164	275	67.7	24,029	22,594	−6.0	2.3	1.8	0.6	1.1	95.1	93.3
Harmony Village	24,209	21,749	−10.2	265	300	13.2	195	227	16.4	23,236	20,496	−11.8	1.1	1.4	0.8	1.0	96.0	94.2
Brooks	24,195	24,740	2.3	3741	4622	23.5	642	1201	87.1	19,173	17,741	−7.5	15.5	18.7	2.7	4.9	79.2	71.7
Mt. Olivet	23,390	18,554	−20.7	929	645	−30.6	159	227	42.8	21,338	16,843	−21.1	4.0	3.5	0.7	1.2	91.2	90.8
Rouge	21,841	22,145	1.4	4789	4002	−16.4	1411	2559	81.4	15,005	14,528	−3.2	21.9	18.1	6.5	11.6	68.7	65.6
Greenfield	21,627	20,952	−3.1	245	238	−2.9	147	202	37.4	20,809	19,801	−4.8	1.1	1.1	0.7	1.0	96.2	94.5
Chadsey	21,121	19,768	−6.4	5850	5431	−7.2	12,335	12,549	1.7	2172	1063	−51.1	27.7	27.5	58.4	63.5	10.3	5.4
Denby	20,135	18,266	−9.3	950	501	−47.3	158	191	20.9	18,631	16,971	−8.9	4.7	2.7	0.8	1.0	92.5	92.9
Conner	18,950	13,760	−27.4	389	280	−28.0	105	176	67.6	18,045	12,816	−29.0	2.1	2.0	0.6	1.3	95.2	93.1
Durfee	18,207	13,599	−25.3	373	734	96.8	160	211	31.9	17,303	12,052	−30.3	2.0	5.4	0.9	1.6	95.0	88.6
Redford	18,182	18,435	1.4	2615	1730	−33.8	236	362	53.4	14,779	15,426	4.4	14.4	9.4	1.3	2.0	81.3	83.7
Pembroke	18,017	18,227	1.2	157	304	93.6	135	189	40.0	17,375	17,126	−1.4	0.9	1.7	0.7	1.0	96.4	94.0
Burbank	17,959	14,132	−21.3	1035	500	−51.7	143	226	58.0	16,306	12,943	−20.6	5.8	3.5	0.8	1.6	90.8	91.6
Pershing	17,356	15,902	−8.4	486	380	−21.8	189	246	30.2	16,313	14,755	−9.6	2.8	2.4	1.1	1.5	94.0	92.8
Bagley	16,912	17,472	3.3	85	390	358.8	125	241	92.8	16,400	16,196	−1.2	0.5	2.2	0.7	1.4	97.0	92.7
Vernor/Junction	16,126	13,452	−16.6	2854	2066	−27.6	11,150	9665	−13.3	1760	1324	−24.8	17.7	15.4	69.1	71.8	10.9	9.8
Rosedale	16,121	14,951	−7.3	785	776	−1.2	127	209	64.6	14,849	13,475	−9.3	4.9	5.2	0.8	1.4	92.1	90.1
Rosa Parks	15,984	12,072	−24.5	353	745	111.0	225	253	12.4	15,003	10,480	−30.1	2.2	6.2	1.4	2.1	93.9	86.8
Cody	15,008	14,830	−1.2	647	382	−41.0	162	220	35.8	13,864	13,671	−1.4	4.3	2.6	1.1	1.5	92.4	92.2
Nolan	14,724	10,263	−30.3	401	206	−48.6	128	144	12.5	13,932	9509	−31.7	2.7	2.0	0.9	1.4	94.6	92.7
Springwells	14,703	13,284	−9.7	3058	1866	−39.0	10,584	10,284	−2.8	761	766	0.7	20.8	14.0	72.0	77.4	5.2	5.8
Lower Woodward c	14,550	16,921	16.3	3488	5633	61.5	310	451	45.5	9117	8457	−7.2	24.0	33.3	2.1	2.7	62.7	50.0
Davison	14,510	13,135	−9.5	2838	3021	6.4	252	128	−49.2	7307	3820	−47.7	19.6	23.0	1.7	1.0	50.4	29.1
Tireman	13,538	9166	−32.3	139	199	43.2	519	1148	121.2	12,557	7445	−40.7	1.0	2.2	3.8	12.5	92.8	81.2
Winterhalter	13,234	10,416	−21.3	122	211	73.0	121	177	46.3	12,718	9586	−24.6	0.9	2.0	0.9	1.7	96.1	92.0
Brightmoor	12,836	12,302	−4.2	1175	746	−36.5	205	290	41.5	11,046	10,709	−3.1	9.2	6.1	1.6	2.4	86.1	87.1
Middle Woodward	12,476	11,839	−5.1	596	2006	236.6	128	329	157.0	11,342	8733	−23.0	4.8	16.9	1.0	2.8	90.9	73.8
Lower East Central	11,484	12,805	11.5	840	2104	150.5	117	281	140.2	10,133	9628	−5.0	7.3	16.4	1.0	2.2	88.2	75.2
Kettering	10,345	6938	−32.9	164	321	96.0	54	85	56.6	9956	6256	−37.2	1.6	4.6	0.5	1.2	96.2	90.2
Grant	10,334	8437	−18.4	512	360	−29.7	85	126	48.2	9499	7697	−19.0	5.0	4.3	0.8	1.5	91.9	91.2
Palmer Park	9463	9600	1.4	811	1456	79.5	112	205	83.0	8272	7411	−10.4	8.6	15.2	1.2	2.1	87.4	77.2
McNichols	9107	7223	−20.7	551	692	25.6	121	126	4.1	8112	6105	−24.7	6.1	9.6	1.3	1.7	89.1	84.5
Airport	8221	6273	−23.7	1376	1743	26.7	74	65	−12.2	6416	4140	−35.5	16.7	27.8	0.9	1.0	78.0	66.0
Boynton	8210	6437	−21.6	467	205	−56.1	1019	773	−24.1	6569	5171	−21.3	5.7	3.2	12.4	12.0	80.0	80.3
Chandler Park	8011	5680	−29.1	125	171	36.8	60	108	80.0	7638	5197	−32.0	1.6	3.0	0.7	1.9	95.3	91.5
Jeffries b	7905	7552	−4.5	1166	1332	14.2	259	259	0.0	6226	5537	−11.1	14.8	17.6	3.3	3.4	78.8	73.3
East Riverside	7399	6457	−12.7	507	657	29.6	48	122	154.2	6640	5399	−18.7	6.9	10.2	0.6	1.9	89.7	83.6
Condon	7140	5591	−21.7	688	474	−31.1	2598	2717	4.6	3715	2214	−40.4	9.6	8.5	36.4	48.6	52.0	39.6
Butzel	7134	5754	−19.4	448	1144	155.3	63	171	172.1	6469	4089	−36.8	6.3	19.9	0.9	3.0	90.7	71.1
St. Jean	6561	3896	−40.6	115	102	−11.0	46	37	−18.9	6295	3628	−42.4	1.8	2.6	0.7	1.0	95.9	93.1
Central Business District a	5292	6151	16.2	1399	3190	128.0	167	402	140.7	3342	1859	−44.4	26.4	51.9	3.2	6.5	63.2	30.2
Middle East Central	5286	4249	−19.6	369	390	5.7	67	98	46.3	4698	3578	−23.8	7.0	9.2	1.3	2.3	88.9	84.2
Foch	5090	3925	−22.9	121	207	70.8	30	52	72.2	4837	3499	−27.7	2.4	5.3	0.6	1.3	95.0	89.1
Indian Village	4639	5976	28.8	806	1455	80.5	55	141	156.4	3591	4047	12.7	17.4	24.3	1.2	2.4	77.4	67.7
State Fair	4315	4344	0.7	586	473	−19.3	89	110	23.6	3481	3488	0.2	13.6	10.9	2.1	2.5	80.7	80.3
Jefferson/Mack	3592	2415	−32.8	151	149	−1.3	39	64	64.1	3323	2126	−36.0	4.2	6.2	1.1	2.7	92.5	88.0
West Riverfront	2783	1405	−49.5	721	371	−48.5	1304	732	−43.9	684	228	−66.7	25.9	26.4	46.9	52.1	24.6	16.2
Hubbard Richard	2080	1544	−25.8	335	339	1.2	1067	712	−33.3	602	418	−30.6	16.1	22.0	51.3	46.1	28.9	27.1
Near East Riverfront	1404	2560	82.3	243	1026	322.2	30	91	203.3	1061	1224	15.4	17.3	40.1	2.1	3.6	75.6	47.8
Corktown	1200	1623	35.3	450	811	80.2	206	173	−16.0	492	482	−2.0	37.5	50.0	17.2	10.7	41.0	29.7
Upper East Central	123	154	25.2	26	63	142.3	1	4	300.0	94	79	−16.0	21.1	40.9	0.8	2.6	76.4	51.3

^a Listed as Downtown-CBD in [4] ^b Listed as Woodbridge in [4]; ^c Listed as Midtown in [4]. ^d Population totals differ slightly from 2020 census figures because of rounding and interpolation. ^e Neighborhood population change—population increase, no change in population, or population decrease. Compiled from [5,103].

).

Detroit, which had a declining White population for several decades, saw a reversal of that trend and a resultant increase during the intercensal period. The number of White residents increased by 9.3%, or an estimated 5171; the increase occurred in 30 neighborhoods. More specifically, the White population increased by more than 2100 people in Lower Woodward, by almost 1800 in the Central Business District (CBD), by more than 1400 in Middle Woodward, and by more than 1200 in Lower East Central. Still, some neighborhoods experienced significant declines in the White population. The most extreme are Springwells, which lost almost 1200 Whites; Redford, where the White population fell by 885; and Vernor/Junction and Rouge, which lost almost 800 Whites.

Detroit’s Latinx/Hispanic population increased by 5.3%, or roughly 2592. The Latinx/Hispanic population rose in 45 neighborhoods. The neighborhoods with the most significant drops in the Latinx/Hispanic population were Vernor/Junction, which had almost 1500 fewer; West Riverfront, which had around 580 fewer; and Hubbard Richard, which had about 350 fewer of these residents. The Latinx/Hispanic population grew by more than 1100 in the Rouge, more than 600 in Tireman, and 550 in Brooks.

Detroit experienced significant Black flight during the intercensal period, resulting in a precipitous decline in the number of Blacks. Hence, the city’s Black population fell by 15.9%, or about 93,422. It declined in 49 neighborhoods and increased in only five. Conner, Durfee, and Tireman each lost more than 5000 Black residents, while Rosa Parks, Mt. Olivet, Mackenzie, and Nolan each lost between 4400 and 4500 Black residents. The Black population grew by about 650 in Redford and roughly 450 in Indian Village.

6.5. Changes in the Distribution of Food Outlets in Detroit’s Neighborhoods

6.5.1. Total Number of Food Outlets Lost in Various Neighborhoods

Table 5. Comparing the distribution of food outlets in Detroit’s neighborhoods in 2013 and 2023.

Neighborhoods	Supermarkets and Large Grocery Stores			Small Groceries and Convenience Stores			Specialty Food Store and Vendors			Pharmacies and Dollar and Variety Stores			Restaurants and Other Food Service			Supply Chain			Farms, Gardens, Farmers’ Markets, Produce Vendors			Emergency Food Assistance			Total Number of Food Outlets
	2013	2023	Diff.	2013	2023	Diff.	2013	2023	Diff.	2013	2023	Diff.	2013	2023	Diff.	2013	2023	Diff.	2013	2023	Diff.	2013	2023	Diff.	2013	2023	Diff.
Detroit total	96	74	−22	1110	726	−384	279	135	−144	306	292	−14	1245	1244	−1	157	131	−26	206	214	8	100	68	−32	3499	2884	−615
Finney	5	1	−4	26	22	−4	7	7	0	13	16	3	41	45	4	1	2	1	3	8	5	2	2	0	98	103	5
Brooks	4	5	1	45	28	−17	18	12	−6	11	8	−3	45	34	−11	7	2	−5	5	3	−2	2	1	−1	137	93	−44
Conner	4	2	−2	39	25	−14	2	1	−1	6	5	−1	17	14	−3	1	0	−1	1	2	1	1	0	−1	71	49	−22
Kettering	4	1	−3	18	13	−5	3	0	−3	4	5	1	9	7	−2	1	2	1	0	1	1	2	2	0	41	31	−10
Lower Woodward c	4	4	0	20	11	−9	9	6	−3	11	14	3	79	109	30	3	3	0	13	12	−1	2	3	1	141	162	21
Mackenzie	4	2	−2	44	29	−15	7	2	−5	12	9	−3	26	20	−6	2	2	0	1	4	3	6	3	−3	102	71	−31
Mt. Olivet	4	2	−2	25	20	−5	5	2	−3	13	12	−1	24	17	−7	1	1	0	3	3	0	2	1	−1	77	58	−19
Cerveny/Grandmont	3	3	0	41	32	−9	16	5	−11	13	12	−1	39	35	−4	3	1	−2	8	9	1	1	1	0	124	98	−26
Cody	3	1	−2	21	19	−2	1	0	−1	6	7	1	17	10	−7	3	2	−1	6	4	−2	0	0	0	57	43	−14
Greenfield	3	3	0	32	20	−12	10	5	−5	13	15	2	29	28	−1	0	0	0	5	4	−1	2	0	−2	94	75	−19
Harmony Village	3	1	−2	34	28	−6	7	3	−4	10	12	2	40	29	−11	3	2	−1	3	3	0	0	1	1	100	79	−21
Lower East Central	3	2	−1	12	4	−8	1	1	0	4	5	1	7	17	10	4	2	−2	11	8	−3	1	1	0	43	40	−3
Springwells	3	2	−1	28	19	−9	6	3	−3	7	5	−2	21	23	2	1	1	0	3	2	−1	0	0	0	69	55	−14
St. Jean	3	2	−1	9	6	−3	2	2	0	5	5	0	10	7	−3	1	2	1	2	3	1	2	1	−1	34	28	−6
Vernor/Junction	3	2	−1	29	17	−12	6	6	0	8	11	3	29	39	10	6	2	−4	9	11	2	3	3	0	93	91	−2
Airport	2	1	−1	22	13	−9	4	0	−4	6	4	−2	12	14	2	2	0	−2	7	6	−1	2	1	−1	57	39	−18
Bagley	2	1	−1	17	8	−9	10	1	−9	5	8	3	21	22	1	2	0	−2	1	2	1	3	1	−2	61	43	−18
Chadsey	2	0	−2	31	20	−11	6	4	−2	10	5	−5	32	31	−1	6	2	−4	7	4	−3	1	2	1	95	68	−27
East Riverside	2	1	−1	7	5	−2	1	0	−1	2	3	1	13	19	6	1	2	1	3	0	−3	2	0	−2	31	30	−1
Evergreen	2	4	2	42	28	−14	9	4	−5	9	9	0	34	30	−4	0	0	0	2	2	0	1	1	0	99	78	−21
Grant	2	2	0	17	13	−4	5	3	−2	8	7	−1	18	21	3	3	4	1	2	1	−1	1	0	−1	56	51	−5
Jeffries b	2	1	−1	13	11	−2	1	0	−1	3	2	−1	9	14	5	1	2	1	8	7	−1	1	0	−1	38	37	−1
Nolan	2	2	0	23	14	−9	2	2	0	4	3	−1	14	15	1	0	0	0	2	4	2	0	0	0	47	40	−7
Pershing	2	1	−1	20	12	−8	5	2	−3	7	4	−3	25	17	−8	2	1	−1	1	1	0	2	1	−1	64	39	−25
Redford	2	0	−2	27	20	−7	5	3	−2	6	8	2	35	24	−11	0	0	0	1	4	3	2	2	0	78	61	−17
Rosedale	2	2	0	12	10	−2	5	2	−3	6	9	3	22	23	1	0	1	1	0	2	2	1	0	−1	48	49	1
Rouge	2	1	−1	28	18	−10	6	0	−6	7	7	0	30	17	−13	1	0	−1	4	7	3	0	0	0	78	50	−28
Boynton	1	0	−1	12	9	−3	2	1	−1	0	2	2	10	11	1	2	1	−1	1	0	−1	2	2	0	30	26	−4
Burbank	1	2	1	23	15	−8	7	3	−4	6	4	−2	24	18	−6	0	2	2	0	3	3	0	0	0	61	47	−14
Condon	1	1	0	16	8	−8	2	0	−2	2	4	2	8	12	4	2	2	0	1	3	2	1	1	0	33	31	−2
Davison	1	1	0	26	18	−8	5	0	−5	6	7	1	3	12	9	0	1	1	1	5	4	1	0	−1	43	44	1
Denby	1	1	0	18	14	−4	2	2	0	2	1	−1	15	11	−4	1	0	−1	1	3	2	1	0	−1	41	32	−9
Durfee	1	2	1	30	14	−16	4	1	−3	7	5	−2	16	13	−3	1	0	−1	2	5	3	8	4	−4	69	44	−25
Hubbard Richard	1	1	0	3	2	−1	2	2	0	0	1	1	12	11	−1	3	0	−3	2	0	−2	1	1	0	24	18	−6
Indian Village	1	1	0	1	0	−1	1	1	0	0	2	2	3	3	0	0	0	0	2	2	0	1	1	0	9	10	1
Jefferson/Mack	1	1	0	8	2	−6	0	0	0	6	5	−1	6	8	2	0	0	0	2	2	0	1	0	−1	24	18	−6
McNichols	1	1	0	16	12	−4	7	0	−7	0	1	1	8	4	−4	3	2	−1	2	3	1	3	2	−1	40	25	−15
Middle East Central	1	3	2	18	17	−1	31	23	−8	3	1	−2	32	43	11	61	61	0	16	9	−7	3	1	−2	165	158	−7
Middle Woodward	1	1	0	30	16	−14	5	2	−3	10	5	−5	40	31	−9	0	1	1	12	12	0	3	2	−1	101	70	−31
Near East Riverfront	1	2	1	6	5	−1	2	1	−1	3	4	1	18	12	−6	2	0	−2	0	7	7	0	1	1	32	32	0
Palmer Park	1	1	0	9	7	−2	2	1	−1	3	3	0	18	12	−6	0	0	0	2	3	1	1	1	0	36	28	−8
Pembroke	1	3	2	18	11	−7	5	2	−3	5	5	0	24	22	−2	0	1	1	2	1	−1	2	3	1	57	48	−9
Rosa Parks	1	1	0	22	9	−13	0	0	0	13	9	−4	19	13	−6	0	0	0	1	3	2	3	2	−1	59	37	−22
State Fair	1	1	0	10	3	−7	5	6	1	2	0	−2	12	9	−3	2	1	−1	4	4	0	0	0	0	36	24	−12
Tireman	1	0	−1	22	14	−8	3	1	−2	4	1	−3	9	2	−7	1	0	−1	3	2	−1	1	0	−1	44	20	−24
Winterhalter	1	0	−1	24	20	−4	1	0	−1	9	8	−1	22	16	−6	0	1	1	1	1	0	6	2	−4	64	48	−16
Brightmoor	0	0	0	19	14	−5	6	1	−5	2	1	−1	11	9	−2	0	0	0	7	5	−2	7	4	−3	52	34	−18
Butzel	0	0	0	12	6	−6	1	2	1	6	6	0	10	20	10	0	1	1	4	7	3	6	4	−2	39	46	7
Central Business District a	0	2	2	39	12	−27	21	9	−12	4	5	1	181	212	31	1	4	3	11	5	−6	1	2	1	258	251	−7
Chandler Park	0	0	0	18	15	−3	2	0	−2	1	1	0	10	7	−3	0	1	1	2	2	0	1	4	3	34	30	−4
Corktown	0	0	0	4	5	1	1	0	−1	2	0	−2	22	36	14	2	1	−1	3	4	1	1	3	2	35	49	14
Foch	0	0	0	11	6	−5	1	0	−1	1	1	0	5	8	3	1	0	−1	5	3	−2	0	0	0	24	18	−6
Upper East Central	0	0	0	3	1	−2	1	1	0	0	0	0	2	0	−2	4	3	−1	3	0	−3	0	1	1	13	6	−7
West Riverfront	0	0	0	10	6	−4	1	0	−1	0	0	0	7	8	1	16	12	−4	5	3	−2	4	0	−4	43	29	−14

Notes: ^a Listed as Downtown-CBD in [4]; ^b Listed as Woodbridge in [4]; ^c Listed as Midtown in [4].

reveals that the total number of food outlets increased in relatively few Detroit neighborhoods throughout the study period. Only eight neighborhoods had more food outlets in 2023 than in 2013. The Lower Woodward neighborhood led the way; it had a net of 21 food outlets. Fourteen food outlets were added in Corktown, seven in Butzel, and five in Finney. In contrast, the overall number of food outlets declined in 46 neighborhoods. This was most evident in Brooks, whose food outlets decreased by 44. The findings show that 27 neighborhoods lost 10 or more food outlets during the decade.

6.5.2. Supermarkets and Large Grocery Stores

Detroit’s neighborhoods had anywhere from zero to five supermarkets/large grocery stores in their boundaries. Brooks, the city’s second most populous neighborhood, had five supermarkets/large grocery stores in 2023, while Lower Woodward and Evergreen each had four. Four neighborhoods, including Detroit’s largest—Cerveny/Grandmont—each had three supermarkets/large grocery stores.

Neighborhoods without Supermarkets or Large Grocery Stores

Though Detroit added 11 supermarkets and large grocery stores during the study period, the city lost 33 of these food establishments, resulting in a net loss of 22. Table 4 and Table 5 show that the neighborhoods with four or five supermarkets and large grocery stores had 10,000 or more residents in 2010. Four of those neighborhoods had populations that exceeded 20,000 in 2020.

Eight neighborhoods had no supermarkets or large grocery stores in 2013. These neighborhoods tended to have small populations; all except Brightmoor had fewer than 10,000 residents. The CBD was the only neighborhood that did not have a supermarket or large grocery store in 2013 that had such food outlets in 2023. The revitalized CBD had two supermarkets/large grocery stores in 2023. Seven of the 12 neighborhoods without supermarkets/large grocery stores in 2023 had fewer than 10,000 residents.

Neighborhoods That Lost Supermarkets or Large Grocery Stores

Some neighborhoods lost one or more supermarkets or large grocery stores. The most dramatic losses occurred in Finney and Kettering. Finney, a neighborhood of almost 24,000 residents, lost roughly 10% of its population and four of its five supermarkets or large grocery stores. Kettering lost a third of its population and three of its four supermarkets or large grocery stores. Kettering had just over 10,000 residents in 2010 and under 7000 in 2020.

Seven neighborhoods lost two supermarkets or large grocery stores over the study period. All seven had more than 15,000 residents in 2010, and the population dropped in each. However, six of these neighborhoods still had populations that exceeded 15,000 in 2020.

Neighborhoods That Gained Supermarkets and Large Grocery Stores

Two new supermarkets/large grocery stores opened in the CBD, Pembroke, Middle East Central, and Evergreen. The CBD and Pembroke grew, while Evergreen and Middle East Central lost population. Two neighborhoods (Pembroke and Evergreen) also had more than 18,000 residents, while the CBD and Middle East Central had fewer than 10,000 residents. Though the CBD has a small population, it grew by about 10%. The CBD also went from being a majority Black neighborhood in 2010 to a majority White one in 2020.

6.5.3. Small Groceries and Convenience Stores

Detroit lost over a third of its small grocery and convenience stores during the study period, occurring in all but one neighborhood. While Corktown gained one additional small grocery/convenience store, 12 neighborhoods lost 10 or more of these food outlets. The neighborhoods that lost the most of these food venues were the CBD, which lost 27; Brooks, which lost 17; Durfee, which lost 16; and Mackenzie, which lost 15. Three neighborhoods lost 14 outlets—Conner, Evergreen, and Middle Woodward.

6.5.4. Specialty Food Stores and Vendors

The most significant change in this sector occurred in the CBD, which lost 12 of its specialty stores and vendors. Cerveny/Grandmont lost 11 outlets, Bagley lost nine, and Middle East Central lost eight. Cerveny/Grandmont had 31,225 residents in 2020 despite losing 4.7% of its population. The Black population declined by 6.9%, the White population grew by 10.2%, and the Latinx/Hispanic population declined by 42.4%.

6.5.5. Pharmacies and Dollar and Variety Stores

Two neighborhoods—Chadsey and Middle Woodward—each lost five pharmacies, dollar stores, and variety stores. On the flip side, several neighborhoods added three pharmacies and dollar and variety stores: Bagley, Finney, Rosedale, Lower Woodward, and Vernor/Junction. While the populations of Bagley and Lower Woodward increased, the remaining neighborhoods had declining populations.

6.5.6. Restaurants and Other Food Service Providers

At first glance, this sector appears stable, recording almost identical numbers of restaurants in each study period. However, a closer look reveals noticeable increases and decreases in the number of restaurants and other food service providers in various neighborhoods. For instance, 31 additional restaurants operated in the CBD in 2023 compared to 2013, while 30 were added in the Lower Woodward neighborhood. Seven neighborhoods added ten or more new restaurants or other food service venues.

The Rouge neighborhood lost the most restaurants/food service providers—it lost 13. Other neighborhoods, such as Brooks, Harmony Village, and Redford, each lost 11 restaurants or food service providers. Overall, 13 neighborhoods lost 5 or more restaurants or other food service providers during the period studied. Harmony Village was the only one of these four neighborhoods that lost population.

6.5.7. Other Food Outlets

Twenty neighborhoods had fewer urban farms, community gardens, farmers’ markets, and produce vendors in 2023 than in 2013. The number of these food venues in 11 neighborhoods remained the same, while this type of food outlet increased in the remaining 23 neighborhoods. The number of emergency food assistance venues decreased in 26 neighborhoods, remained the same in 20 neighborhoods, and increased in only 8 neighborhoods.

6.6. Neighborhood-Level Multivariate Analyses

The study employed multivariate techniques to further analyze the distribution of food outlets in Detroit’s neighborhoods in 2013 and 2023. We conducted Poisson and negative binomial regressions to determine the likelihood of having an additional food outlet in a given neighborhood. When the Black population was analyzed, the reference category was 0–40% Black. Whites and Latinx/Hispanics were treated as continuous variables because the percentage of each was low in most neighborhoods. Each model controlled for the percentage of the neighborhood residents over 25 with at least a high school education, the neighborhood’s median income, and the neighborhood’s population density.

6.6.1. Neighborhood Analyses for 2013

One significant relationship was found between the distribution of food outlets and the racial/ethnic composition of the neighborhood in 2013 (

Table 6. Incidence rate ratios for full model showing Detroit neighborhood racial characteristics.

Major Food Categories b,c	Neighborhood Racial Characteristics
	41–70% Black Residents				71–90% Black Residents				91% or More Black Residents				Percent White				Percent Hispanic
	Incident Rate Ratio (IRR) a	95% Confidence Interval (CI)		p-Value f	Incident Rate Ratio (IRR)	95% Confidence Interval (CI)		p-Value	Incident Rate Ratio (IRR)	95% Confidence Interval (CI)		p-Value	Incident Rate Ratio (IRR)	95% Confidence Interval (CI)		p-Value	Incident Rate Ratio (IRR)	95% Confidence Interval (CI)		p-Value
		Lower	Upper			Lower	Upper			Lower	Upper			Lower	Upper			Lower	Upper
2013
e Total number of food outlets	0.617	0.141	2.695	0.521	0.594	0.089	3.982	0.591	0.604	0.073	4.995	0.640	1.019	0.981	1.058	0.329	0.994	0.963	1.026	0.718
d Supermarkets and large grocery stores	1.786	0.100	31.921	0.693	3.205	0.092	112.102	0.521	3.790	0.080	178.805	0.498	1.024	0.962	1.091	0.458	1.001	0.948	1.058	0.963
e Small groceries and convenience stores	1.244	0.284	5.453	0.772	1.142	0.175	7.433	0.890	1.305	0.163	10.473	0.802	1.007	0.970	1.046	0.706	0.994	0.965	1.024	0.691
e Pharmacies and dollar and variety stores	3.272	0.054	199.587	0.572	6.171	0.040	952.307	0.479	9.751	0.045	2120.416	0.407	1.045	0.957	1.140	0.330	1.004	0.932	1.082	0.913
e Specialty food stores and vendors	0.546	0.030	10.077	0.684	0.488	0.013	17.811	0.696	0.267	0.005	14.027	0.513	0.995	0.929	1.065	0.883	0.988	0.931	1.047	0.677
e Restaurants and other food service	1.206	0.202	7.213	0.837	1.535	0.153	15.402	0.716	1.885	0.149	23.807	0.624	1.049	1.004	1.096	0.033 *	1.020	0.980	1.061	0.325
e Urban farms, gardens, farmers’ markets, and produce vendors	0.288	0.031	2.697	0.275	0.453	0.029	7.175	0.574	0.275	0.013	5.691	0.404	1.028	0.974	1.086	0.315	0.991	0.944	1.040	0.701
e Emergency food assistance	0.103	0.005	2.004	0.133	0.094	0.003	3.509	0.200	0.053	0.001	2.926	0.151	0.949	0.880	1.024	0.177	0.961	0.902	1.023	0.211
e Supply chain	0.166	0.005	5.146	0.306	0.480	0.004	58.274	0.765	0.203	0.001	45.989	0.564	1.001	0.907	1.105	0.978	1.008	0.923	1.100	0.866
2023
e Total number of food outlets	0.525	0.231	1.191	0.123	1.011	0.341	2.998	0.985	1.100	0.305	3.964	0.884	1.023	0.997	1.051	0.089	0.998	0.980	1.016	0.791
d Supermarkets and large grocery stores	1.225	0.288	5.205	0.784	1.337	0.203	8.825	0.763	1.731	0.187	16.009	0.629	1.006	0.962	1.052	0.784	1.005	0.972	1.038	0.776
e Small groceries and convenience stores	0.667	0.307	1.449	0.306	0.979	0.370	2.587	0.965	1.117	0.358	3.485	0.848	0.990	0.965	1.015	0.413	0.992	0.976	1.007	0.290
e Pharmacies and dollar and variety stores	0.694	0.262	1.840	0.463	0.564	0.173	1.839	0.342	0.476	0.121	1.866	0.287	0.978	0.948	1.008	0.150	0.986	0.967	1.006	0.168
e Specialty food stores and vendors	2.031	0.189	21.801	0.558	31.320	1.038	945.030	0.048 *	37.730	0.655	2172.152	0.079	1.082	1.003	1.167	0.041 *	1.046	0.986	1.109	0.133
e Restaurants and other food service	0.428	0.164	1.120	0.084	0.934	0.264	3.308	0.915	1.050	0.233	4.731	0.950	1.039	1.008	1.071	0.013 **	1.000	0.979	1.022	0.988
e Urban farms, gardens, farmers’ markets, and produce vendors	1.235	0.509	2.994	0.640	1.656	0.526	5.213	0.389	1.139	0.289	4.490	0.853	1.024	0.994	1.053	0.113	1.000	0.981	1.019	0.997
d Emergency food assistance	0.894	0.224	3.564	0.873	2.039	0.340	12.241	0.436	2.053	0.236	17.823	0.514	1.035	0.994	1.079	0.098	1.011	0.979	1.045	0.496
e Supply chain	0.792	0.096	6.499	0.828	5.654	0.393	81.334	0.203	4.813	0.205	112.730	0.329	1.052	0.983	1.127	0.144	1.026	0.981	1.073	0.266

Notes: ^a For percent Black residents: Incident rate ratio values of less than 1 indicate that there are fewer stores than in the reference group. The reference group is neighborhoods with 0–40% Black residents. ^b Each row represents a separate model that adjusted for population density, median household income per USD 1000, and the percentage of the population over 25 that has at least a high school education. ^c For 2013: Model for the following food outlets—supermarkets and large grocery stores, pharmacies, dollar and variety stores, specialty food stores, and vendors—and emergency food assistance was not significant; p-values reported as needed. For 2023: Model for the following food outlets—supermarkets and large grocery stores, specialty food stores and vendors, urban farms, gardens, farmers’ markets, and produce vendors—and emergency food assistance was not significant; p-values reported as needed. ^d Indicates models that followed a Poisson regression. ^e Indicates models that followed a non-binomial regression. ^f Significance levels: * p-values α ≤ 0.05; ** p-values α ≤ 0.01.

). The study found that for each percentage point increase in the White population, there was a 4.9% increase in the likelihood of having an additional restaurant or food service (IRR = 1.049; 95% CI = 1.004–1.096; p = 0.033).

Table 7. Incidence rate ratios for the full model showing median household income, educational attainment, and population density for 2013 and 2023.

Major Food Categories a,b	Percent of the Population with High School Education				Median Household Income per USD 1000 e				Population Density per Square Kilometer
	Incident Ratio Rate (IRR)	95% Confidence Interval (CI)		p-Value f	Incident Ratio Rate (IRR)	95% Confidence Interval (CI)		p-Value	Incident Ratio Rate (IRR)	95% Confidence Interval (CI)		p-Value
		Lower	Upper			Lower	Upper			Lower	Upper
2013
d Total number of food outlets	1.022	0.996	1.049	0.100	0.971	0.954	0.988	0.001 ***	1.000	1.000	1.001	0.001 ***
c Supermarkets and large grocery stores	0.981	0.939	1.024	0.379	0.996	0.966	1.028	0.817	1.001	1.000	1.001	0.001 ***
d Small groceries and convenience stores	0.992	0.967	1.016	0.503	0.985	0.967	1.002	0.090	1.000	1.000	1.001	0.000 ***
d Pharmacies and dollar and variety stores	0.988	0.933	1.046	0.685	0.982	0.939	1.027	0.420	1.001	1.000	1.001	0.002 **
d Specialty food stores and vendors	1.024	0.977	1.073	0.321	0.971	0.939	1.003	0.078	1.000	1.000	1.001	0.016 *
d Restaurants and other food service	1.064	1.030	1.098	0.000 ***	0.955	0.934	0.977	0.000 ***	1.000	1.000	1.001	0.000 ***
d Urban farms, gardens, farmers’ markets, and produce vendors	1.022	0.983	1.063	0.266	0.959	0.933	0.986	0.003 **	1.000	1.000	1.000	0.594
d Emergency food assistance	0.989	0.943	1.038	0.653	0.983	0.948	1.020	0.370	1.000	1.000	1.001	0.408
d Supply chain	1.020	0.936	1.110	0.655	0.965	0.921	1.011	0.137	1.000	0.999	1.000	0.048 *
2023
d Total number of food outlets	1.006	0.980	1.033	0.662	0.999	0.983	1.015	0.884	1.000	1.000	1.001	0.001 ***
c Supermarkets and large grocery stores	1.015	0.972	1.059	0.507	1.002	0.979	1.027	0.855	1.000	1.000	1.001	0.039 *
d Small groceries and convenience stores	0.977	0.953	1.002	0.071	1.006	0.990	1.022	0.476	1.000	1.000	1.001	0.004 **
d Pharmacies and dollar and variety stores	0.996	0.965	1.027	0.781	0.999	0.981	1.017	0.908	1.001	1.000	1.001	0.000 ***
d Specialty food stores and vendors	1.018	0.954	1.086	0.586	1.004	0.961	1.049	0.843	1.000	1.000	1.001	0.296
d Restaurants and other food service	1.018	0.988	1.048	0.234	0.999	0.982	1.017	0.938	1.001	1.000	1.001	0.000 ***
d Urban farms, gardens, farmers’ markets, and produce vendors	0.994	0.966	1.024	0.693	0.995	0.977	1.013	0.604	1.000	1.000	1.001	0.004 **
c Emergency food assistance	1.018	0.976	1.061	0.409	0.982	0.958	1.008	0.169	1.000	1.000	1.001	0.257
d Supply chain	1.012	0.949	1.080	0.708	0.980	0.938	1.023	0.351	0.999	0.999	1.000	0.002 **

Notes: ^a Each row represents a separate model that adjusted for population density, median income, and the percentage of the population over 25 that has at least a high school education. ^b For 2013: Model for the following food outlets—supermarkets and large grocery stores, pharmacies, dollar and variety stores, specialty food stores, and vendors—and emergency food assistance was not significant; p-values reported as needed. For 2023: Model for the following food outlets—supermarkets and large grocery stores, specialty food stores and vendors, urban farms, gardens, farmers’ markets, and produce vendors—and emergency food assistance was not significant; p-values reported as needed. ^c Indicates models that followed a Poisson regression. ^d Indicates models that followed a non-binomial regression. ^e Weighted median household income was calculated and used in the regression models. ^f Significance levels: * p-values α ≤ 0.05; ** p-values α ≤ 0.01; *** p-values α ≤ 0.001.

shows that educational attainment significantly impacted the distribution of the food outlet types studied. Hence, for every percentage increase in the population with at least a high school education, the likelihood of having an additional restaurant or other food service in the neighborhood increased by 6.4% (IRR = 1.064; 95% CI = 1.030–1.098; p = 0.000).

Regarding median household income, for every USD 1000 increase, results show a 4.1% decrease in the likelihood of having an additional urban farm, garden, farmers’ market, or produce vendor (IRR = 0.959; 95% CI = 0.9333–0.986; p = 0.003), a 4.5% decrease in the likelihood of having an additional restaurant or other food service (IRR = 0.955; 95% CI = 0.934–0.977; p = 0.000), and a 2.9% decrease in the likelihood of having an additional food outlet overall (IRR = 0.971; 95% CI = 0.954–0.988; p = 0.001). Population density was statistically significantly associated with all but two major food categories.

6.6.2. Neighborhood Analysis for 2023

One statistically significant relationship was observed when assessing the association between neighborhood racial composition and the likelihood of having additional food outlets in 2023 (see Table 6). The study found that neighborhoods that were 71–90% Black were 31.3 times more likely to have an additional specialty food store or vendor than those that were 0–40% Black (IRR = 31.320; 95% CI = 1.038–945.030; p = 0.048). It should be noted that the confidence interval is wide. Additionally, the results show that for every percentage increase in the White population, there is an 8.2% increase in the likelihood of having an additional specialty food store or vendor (IRR = 1.082; 95% CI = 1.003–1.167; p = 0.041) and a 3.9% increase in having an additional restaurant or food service (IRR = 1.039; 95% CI = 1.008–1.071; p = 0.013).

Table 7 shows the relationship between the covariates and the likelihood of having additional food outlets. Educational attainment and median household income are not statistically significantly associated with any major food categories examined. However, population density is statistically significantly associated with several major food categories.

6.6.3. Interaction Effects

The introduction of interaction terms uncovered statistically significant relationships that moderated the association between some major food categories and neighborhood racial composition. In 2013, for every USD 1000 increase in median household income, neighborhoods that are 41–70% Black are 13.5% less likely to have an additional small grocery and convenience store (IRR = 0.865; 95% CI = 0.768–0.974; p = 0.017) and 12.6% less likely to have an additional food outlet regardless of type (IRR = 0.874; 95% CI = 0.791–0.966; p = 0.008) than neighborhoods that are 0–40% Black.

Similarly, for every USD 1000 increase in median household income, neighborhoods that are 71–90% Black are 12.9% less likely to have an additional small grocery and convenience store (IRR = 0.871; 95% CI = 0.780–0.974; p = 0.015) and 10.9% less likely to have an additional food outlet regardless of type (IRR = 0.891; 95% CI = 0.812–0.977; p = 0.014) when compared to neighborhoods that are 0–40% Black. Finally, for every USD 1000 increase in median household income, neighborhoods that are 90% or more Black are 11.6% less likely to have an additional small grocery and convenience store (IRR = 0.884; 95% CI = 0.791–0.988; p = 0.030) than neighborhoods that are 0–40% Black.

There were also statistically significant interactions between restaurants/other food services, educational attainment, and the total number of food outlets. So, for every percentage point increase in the population with at least a high school education, there was a 5.5% increase in the likelihood of having an additional restaurant or other food service (IRR = 1.055; 95% CI = 1.026–1.084; p = 0.000), and a 2.5% increase in the likelihood of having an additional food outlet regardless of type (IRR = 1.025; 95% CI = 1.003–1.048; p = 0.023). The interaction effects with population density are also significant.

In 2023, the findings showed that for every USD 1000 increase in the median household income, neighborhoods that are 41–70% Black are 9.4% less likely to have an additional restaurant/other food service (IRR = 0.906; 95% CI = 0.861–0.953; p = 0.000); those that are 71–90% Black are 5.3% less likely to have an additional restaurant or other food service (IRR = 0.947; 95% CI = 0.910–0.985; p = 0.007); and those that are more than 90% Black are 6% less likely to have an additional restaurant/other food service than neighborhoods that are 0–40% Black.

There were also significant interaction effects between educational attainment, racial composition, and the likelihood of having a restaurant/food service. Hence, for every percentage point increase in the population with at least a high school education, there was a 2.7% increase in the likelihood of having an additional restaurant or other food service (IRR = 1.027; 95% CI = 1.004–1.050; p = 0.021). There were also several significant statistical associations between population density, racial composition, and several food categories.

6.7. A Question of Swamping Out and Redlining

6.7.1. Intermingled, Not Swamped Out

Figure 11a,b shows that the food swamp thesis does not accurately describe Detroit’s food landscape. In both study periods, supermarkets and large grocery stores operated close to small grocery and convenience stores, gas stations, liquor and party stores, and fast-food restaurants. The intermingling of food outlets from these different sectors was evident regardless of the racial characteristics of the neighborhoods. Supermarkets and large grocery stores are found on busy thoroughfares in Springwells, Chadsey, Condon, Vernor/Junction, CBD, and Lower Woodward, alongside large numbers of food venues reputed to create food swamps. Only a handful of supermarkets/large grocery stores were stand-alone and not interspersed with other types of food outlets.

Though a few areas such as Brightmoor do not have supermarkets or large grocery stores, but have convenience stores, gas stations, and fast-food restaurants, this configuration of neighborhood food outlets is not typical of the whole city.

6.7.2. Supermarkets in Formerly Redlined Areas

The supermarket redlining thesis is not very applicable to Detroit.

Table 8. HOLC’s neighborhood codes and the distribution of supermarkets, grocery stores, dollar stores, and variety stores in 2013 and 2023.

HOLC’s Codes	HOLC’s Color	Supermarkets and Large Grocery Stores				Dollar Stores and Variety Stores
		2013	Percent	2023	Percent	2013	Percent	2023	Percent
A	Green	3	3.1	2	2.7	5	4.1	4	3.2
B	Blue	14	14.6	9	12.2	22	17.9	25	20.0
C	Yellow	34	35.4	25	33.8	48	39.0	43	34.4
D	Red	25	26.0	21	28.4	29	23.6	28	22.4
Uncoded	No color	20	20.8	17	23.0	19	15.4	25	20.0
Total		96	100.0	74	100.0	123	100.0	125	100.0

and Figure 12a,b show that several of the city’s supermarkets or large grocery stores were in formerly redlined neighborhoods in both study periods. In 2013, 25 (or 26%) supermarkets/large grocery stores were in once-redlined areas. In 2023, 21 (or 28.4%) supermarkets/large grocery stores were in formerly redlined areas. So, though the city lost supermarkets and large grocery stores, their percentage in formerly redlined neighborhoods increased.

In both periods, most supermarkets and large grocery stores were in formerly yellow-lined neighborhoods; likewise, most dollar and variety stores were in formerly yellow-lined neighborhoods. However, contrary to what the redlining thesis predicts, the number of dollar and variety stores in formerly redlined neighborhoods decreased slightly between 2013 and 2023, a surprising finding that warrants further investigation.

One of the critical aspects we delve into is the supermarket redlining argument. The question we aim to answer is whether dollar stores and variety stores have taken the place of supermarkets and large grocery stores in once-redlined neighborhoods. However, Figure 12a,b does not provide substantial evidence to support this thesis in either study period. Despite the proliferation of dollar and variety stores, they have not displaced supermarkets or large grocery stores in formerly redlined neighborhoods. The maps reveal that in both study periods, redlined areas of the city had a mix of both types of food outlets, often nearby.

6.8. Food Store Closures: An Infrastructure in Decline

The above findings identified food outlets lost during the intercensal period, but do not say when they ceased to operate in the neighborhoods. To answer the question of how many of Detroit’s food venues were not operational at a given point in time, we collected information on closed food outlets during our 2023 data collection efforts. We found 1305 closed food outlets in Detroit.

Table 9. Number of open and closed food outlets in neighborhoods in 2023.

Neighborhoods	Number of Open and Closed Food Outlets Identified	Number of Open Food Outlets Identified	Number of Closed Food Outlets Identified	Percent of Food Outlets that are Closed
Detroit	4800	3495	1305	27.2
Central Business District	409	300	109	26.7
Lower Woodward	268	210	58	21.6
Middle East Central	233	194	39	16.7
Cerveny/Grandmont	172	122	50	29.1
Finney	162	126	36	22.2
Brooks	158	110	48	30.4
Vernor/Junction	140	102	38	27.1
Harmony Village	138	99	39	28.3
Mackenzie	135	87	48	35.6
Greenfield	132	94	38	28.8
Middle Woodward	131	91	40	30.5
Chadsey	121	80	41	33.9
Evergreen	113	90	23	20.4
Rouge	99	61	38	38.4
Redford	98	71	27	27.6
Mt. Olivet	97	73	24	24.7
Springwells	89	65	24	27.0
Conner	88	55	33	37.5
Durfee	86	53	33	38.4
Burbank	81	58	23	28.4
Pershing	80	51	29	36.3
Rosedale	79	59	20	25.3
Pembroke	79	56	23	29.1
Grant	77	60	17	22.1
Bagley	72	52	20	27.8
Winterhalter	72	56	16	22.2
Cody	71	56	15	21.1
Lower East Central	69	53	16	23.2
Rosa Parks	69	43	26	37.7
Davison	67	49	18	26.9
Butzel	67	55	12	17.9
Airport	63	49	14	22.2
Corktown	62	53	9	14.5
Nolan	60	47	13	21.7
Jeffries	59	50	9	15.3
Denby	57	42	15	26.3
Condon	55	38	17	30.9
Brightmoor	54	39	15	27.8
Kettering	52	32	20	38.5
West Riverfront	49	31	18	36.7
Boynton	49	34	15	30.6
State Fair	49	29	20	40.8
East Riverside	48	33	15	31.3
McNichols	48	34	14	29.2
Palmer Park	48	37	11	22.9
Tireman	47	24	23	48.9
Near East Riverfront	45	38	7	15.6
St. Jean	44	34	10	22.7
Chandler Park	43	33	10	23.3
Foch	31	22	9	29.0
Hubbard Richard	27	24	3	11.1
Jefferson/Mack	27	20	7	25.9
Indian Village	19	14	5	26.3
Upper East Central	12	7	5	41.7

shows the number of closed outlets in each neighborhood, and Figure 13 shows the location of the closed venues.

6.8.1. Hollowed Out

In 2023, Detroit was hit hard by a significant number of food outlet closures, which had a profound and concerning impact on the city’s food infrastructure. When we considered all the food outlet types studied in 2013 and the additional venues discovered through our expanded search in 2023, we found 3495 operating food venues and 1305 non-operating ones. In other words, of the 4800 food venues identified in 2023, a staggering 27.3% of them were closed. The CBD, with 109 closed food venues, had the highest number of non-operating facilities in any neighborhood in the city. Another neighborhood, Lower Woodward, had 58 closed venues. These two adjacent neighborhoods had the highest concentration of closed food venues. In total, 28 neighborhoods had 20 or more closed food outlets.

Essentially, the food infrastructure in some neighborhoods was severely affected, with a significant percentage of food establishments closing their doors. The percentage of closed food establishments varied across the neighborhoods, ranging from 11.1% in Hubbard Richard to 48.9% in Tireman. Upper East Central and State Fair also experienced a significant loss of food stores, with 41.7% and 40.8% of their food venues closing, respectively. On the other hand, the closures were less severe in Corktown, with only 14.5% of its food venues closed, and in Jeffries and Near East Riverfront, with 15.3% and 15.6% of the food venues closed, respectively.

6.8.2. Closures and Neighborhood Racial Composition

Regardless of the neighborhood’s racial composition, food store closures were ubiquitous. But does the percentage of food store closures vary by neighborhood racial group composition? There was only a slight variation in the mean percentage of closed food establishments in neighborhoods with different racial compositions. On average, 26.1% of the stores in neighborhoods that were 40% or less Black, 27.6% of the food stores in neighborhoods comprised of 41–70% Black were closed, 28.9% of the food establishments in neighborhoods containing 71–90% Black residents, and 27.1% of the food outlets in neighborhoods where more than 90% of the residents are Black were closed (Table 9).

6.8.3. Closures and Redlining

The paper also examined the relationship between redlining and food store closures. The CBD, partly uncolored and redlined, had the most closed food outlets. Lower Woodward had the second-highest number of closed food stores and was a red- and yellow-lined neighborhood. Cerveny/Grandmont, the neighborhood with the third-highest number of closed stores (50), was mostly blue- and green-lined, but the neighborhood also had small yellow- and red-lined sections. A portion of the neighborhood was also uncolored. Brooks, a neighborhood with 48 closed food outlets, was yellow-lined and uncolored. Mackenzie also had 48 closed food outlets; it was yellow-lined, blue-lined, and uncolored. Middle Woodward had 40 closed food establishments; it was a red- and yellow-lined community.

6.8.4. Multivariate Analysis of Food Store Closures

We assessed whether there was a statistically significant difference in the number of closed outlets in the neighborhood based on racial composition. The results of the Kruskal–Wallis test show no statistically significant difference in the number of closed food outlets. Further analysis was done to determine if the racial composition of neighborhoods was a significant contributor to the likelihood of a neighborhood having an additional closed food store. As

Table 10. Incidence rate ratios for direct and interaction models showing closed food outlets identified in 2023.

Neighborhood Racial Characteristics a,b	Neighborhood Racial Characteristic				Median Household Income e				Neighborhood Racial Characteristic x Median Household Income
	Direct Model				Direct Model c				Interaction Model
	Incident Rate Ratio (IRR) a	95% Confidence Interval (CI)			Incident Rate Ratio (IRR)	95% Confidence Interval (CI)			a Incident Rate Ratio (IRR)	95% Confidence Interval (CI)
		Lower	Upper	p-Value f		Lower	Upper	p-Value		Lower	Upper	p-Value
d Percent Black Residents					0.998	0.981	1.015	0.783
41–70% Black Residents	0.513	0.270	0.976	0.042 *	---	---	---	---	0.923	0.880	0.968	0.001 ***
71–90% Black Residents	0.642	0.360	1.144	0.133	---	---	---	---	0.934	0.999	0.043	0.580
90% or more Black Residents	0.596	0.343	1.038	0.068	---	---	---	---	0.922	1.010	0.128	0.804
d Percent White Residents	1.010	0.997	1.024	0.118	0.999	0.982	1.016	0.887	1.001	1.000	1.002	0.048 *
d Percent Hispanic Residents	1.004	0.991	1.017	0.568	1.005	0.989	1.020	0.556	1.000	0.998	1.001	0.846

Notes: ^a For Percent Black Residents: Incident rate ratio values of less than 1 indicate that there are fewer stores than in the reference group. The reference group is neighborhoods with 0–40% Black residents. ^b Each row represents a separate model that adjusted for neighborhood racial categories, neighborhood racial characteristics population density, median household income, and the percentage of the population over 25 that has at least a high school education and included the interaction between neighborhood racial characteristics and median household income. ^c All models were not significant; p-values reported as needed. ^d Indicates models that followed a non-binomial regression. ^e Weighted median household income was calculated and used in the regression models. ^f Significance levels: * p-values < α = 0.05; *** p-values < α = 0.001.

shows, neighborhoods that are 41–70% Black are 48.7% less likely to have an additional closed food outlet than neighborhoods that are 0–40% Black (IRR = 0.513; 95% CI = 0.270–0.976; p = 0.042). This relationship remains significant with the inclusion of the interaction term. Thus, for every one-thousand-dollar increase in median household income, neighborhoods that are 41–70% Black are 7.7% less likely to have an additional food store closure (IRR = 0.923; 95% CI = 0.880–0.968; p = 0.001) than those that are 0–40% Black.

7. Discussion

7.1. Detroit’s Dynamic and Changing Food Landscape

The article examined the distribution of food outlets in Detroit 10 years apart and found a significantly transformed food environment in 2023. These findings lead us to highlight the key points below. There were significant differences in the total number of food outlets found in the city in 2013 and 2023. The difference in the number of small grocery stores/corner stores/convenience stores/mini-marts in the two study periods was significant. So was the difference in the number of specialty stores.

This paper assessed three popular hypotheses that food access researchers, policymakers, and activists use to explain the distribution of food outlets—food deserts, food swamps, and supermarket redlining. Though the terms are used to describe Detroit’s food landscape, none of them depict Detroit’s food environment accurately. Consequently, the paper proposes three concepts to help explain the distribution of the city’s food outlets. The article argues that long-term population decline, a vanishing food infrastructure, and the intermingling of food outlet types help account for the pattern and distribution of food outlets. These three factors influence how many food outlets are in the city, where they are located, and the likelihood that they are operational.

Detroit is undergoing significant population shifts that have impacted the food landscape. Not only are there fewer supermarkets and large grocery stores in the city in 2023 than in 2013, but there is also a significant reduction in the number of corner stores/convenience stores and mini-marts, gas stations that sell food, and other kinds of food outlets. While most food access scholars have either ignored or maligned small food stores, Detroit is offering an important lesson by showing us the need to pay more attention not just to supermarkets and large grocery stores but also to the quantity and distribution of small food stores. The immense closures of small food outlets have led to a hollowed-out or vanishing food infrastructure in some neighborhoods. We contend that some neighborhoods are at risk of losing important food infrastructure that is currently being overlooked by food advocates, policymakers, and researchers alike.

7.2. Detroit Is Not a Food Desert

The paper provides compelling evidence to support the claim that the city as a whole is not a food desert. In 2023, the number of food outlets in each neighborhood ranged from seven to 300. The inordinate focus on supermarkets and large grocery stores in the food desert literature distorts our understanding of the distribution of food outlets in Detroit and other cities. However, these types of food outlets account for less than 3% of the city’s food landscape.

Most of Detroit’s neighborhoods had at least one supermarket or large grocery store. Our study found that eight neighborhoods (14.8%) had no supermarket/large grocery store in 2013, and 12 (22.2%) had none in 2023. The study further found that seven of the neighborhoods that did not have a supermarket/large grocery store in 2013 still did not have one in 2023. Therefore, we argue that though there are neighborhoods that have been without a supermarket or large grocery store for more than a decade, it is inaccurate to conclude that the whole city is a food desert. Our data show that even when neighborhoods do not have a supermarket or large grocery store, there are other food outlets in them where residents can purchase food or obtain it for free.

Increasing access to healthy, affordable, or free food for the city’s residents is at the core of food advocacy work in Detroit. There is great enthusiasm for urban farming in the city, and that is reflected in the growing number of farms, community gardens, and schoolyard gardens identified in the study. The city and food advocates have strongly emphasized sustainability, food self-sufficiency, and food sovereignty. Hence, there has been a strong push to develop edible gardens in school yards, along community pathways, and in green spaces. The city also has a vibrant urban farming sector that is a vital source of fresh, nutritious, low- or no-cost food for residents.

Hence, our findings do not provide support for studies that frame the city as a food desert [31,38,39,40,41,42,43,44,45]. Instead, our findings are more aligned with those that refute the food desert thesis [5,7,46,47,48,49,50,51,52,53].

7.3. Detroit Is Not a Food Swamp, Either

Proponents of the food swamp thesis suggest that neighborhoods with many small grocery stores, convenience stores, corner stores, mini-marts, gas stations that sell food, party and liquor stores, and fast-food restaurants do not have supermarkets or large grocery stores because the former displace the latter. We did not find evidence to support this thesis. Our data show that supermarkets/large grocery stores, small grocery stores/convenience stores/corner stores/mini marts, and fast-food restaurants coexist with supermarkets/large grocery stores throughout the city. Instead of a swamping-out effect, we found these food outlets co-exist and are intermingled.

The food swamp thesis appears in several studies [53,71,72,73,74,75,77,78,79], but our assessment of Detroit does not corroborate these analyses. The food swamp studies do not show whether any one type of food outlet can drive another type out of a neighborhood. Though we found an intermingling of food outlet types in Detroit, we believe that the reverse may occur. So, rather than small groceries and fast-food restaurants driving supermarkets/large grocery stores out of neighborhoods, we think that mass merchandisers, supercenters, superstores, and super warehouses can potentially drive small food stores out of neighborhoods with low prices, variety, sales volume, and control over supply chains that small operators cannot compete with.

7.4. Lack of Support for the Supermarket Redlining Thesis

A group of scholars studying the impact of redlining in Detroit associate the practice with contemporary health inequities and socioeconomic disinvestment. However, their study did not examine food access or the distribution of food outlets [137]. Our findings in Detroit do not support the supermarket redlining thesis articulated by others [70,80,81,82,83,84,85,86,87,88,89,93,94,95,96,97,98,99,100,115]. Not only did we find supermarkets and large grocery stores in formerly redlined neighborhoods, but these food outlets were also intermingled with dollar stores and variety stores. We did not find evidence that the latter stores had supplanted the former in formerly redlined neighborhoods.

Though scholars use the supermarket redlining thesis to explain the distribution of food stores, few incorporate redlining maps into their spatial analysis (for example, see [89]). Shannon’s [92] study of Atlanta found evidence that redlined neighborhoods lacked supermarkets/large grocery stores. However, the study focused only on supermarkets that accepted government food assistance benefits cards in six census tracts. We urge researchers to undertake more large-scale studies to assess the complex relationship between redlining and the distribution of food outlets. Such an approach is encouraged in large urban areas..

Researchers tend to ignore yellow-lined neighborhoods, as their primary focus is on the dynamics between redlined and green-lined neighborhoods. However, we found that most of the supermarkets/large grocery stores were in yellow-lined neighborhoods. This makes sense because formerly green-lined neighborhoods were the wealthiest and were often blanketed with restrictive covenants [91]. In many cities, some of the formerly greenlined neighborhoods are elite urban residential enclaves not interested in having large supermarkets or gigantic super warehouse stores in their midst. Residents of formerly greenlined neighborhoods have personal transportation and can easily get to food stores. In contrast, the formerly yellow-lined neighborhoods were White working-class neighborhoods abutting Black neighborhoods or neighborhoods into which Blacks had moved. These neighborhoods were growing rapidly as new and long-term White immigrants vied for housing, jobs, and space with Blacks and other non-white immigrants or recent arrivals to cities.

7.5. New Approaches to Understanding Food Outlet Distribution

7.5.1. Population Size and Outmigration Matters

Detroit has hemorrhaged people for decades, and the process is ongoing. White flight is a well-documented phenomenon in Detroit, and the subject of countless studies (see, for instance, [138]). This paper shows evidence of continued outmigration of Whites from the city to the surrounding suburbs. However, Black flight—which has received scant attention—is the driving force behind population loss and outmigration. More than 93,000 Blacks moved out of the city between 2010 and 2020. They relocated to suburbs such as Southfield, Warren, Pontiac, Royal Oak Township, River Rouge, Ecorse, Inkster, Highland Park, and elsewhere [139].

Despite the dramatic demographic shifts occurring in the city, food access researchers are not connecting their analysis of the location and distribution of food outlets with neighborhood and citywide population dynamics. We contend that analyses of what kinds of food outlets neighborhoods contain should be accompanied by assessments of current population size, changes in population, demographic shifts, and population density. Detroit and other cities undergoing significant population loss or gain are prime examples of why population dynamics should be integrated into assessments of the distribution of food outlets.

7.5.2. Impacts of Race, Household Income, and Educational Attainment

Though our findings did not fully support racialized theses, such as supermarket redlining, there were times when the neighborhood’s racial composition had significant direct and interaction effects. The interaction effects help demonstrate that Detroit’s Black population is not monolithic; the distribution of food outlets varies across neighborhoods of different racial compositions. Hence, in 2013, as median household incomes rose in neighborhoods that were more than 40% Black, the likelihood of having an additional food outlet and small grocery stores/convenience stores was lower than in neighborhoods that were 0–40% Black. Increased household median incomes in neighborhoods that were more than 40% Black also decreased the likelihood of having additional restaurants/food service than the reference group.

The study found that as the White population increases, so does the likelihood of having an additional restaurant in a neighborhood. Similarly with educational attainment, as the percentage of people with high school education increases, so does the likelihood of having additional restaurants.

7.5.3. Strong Evidence of Vanishing Food Infrastructure

The Opening and Closing of Supermarkets and Large Grocery Stores

Detroit’s story defies the simple explanation that supermarkets and large grocery stores closed in Black neighborhoods and relocated to White communities. Our paper shows that defunct supermarkets and large grocery stores were found in White, Latinx/Hispanic, and Black neighborhoods. These kinds of food outlets were closed in neighborhoods regardless of their racial composition.

Even though 12 new supermarkets/large grocery stores opened in Detroit between 2013 and 2023, the city had 22 fewer supermarkets/large grocery stores in the second study period than in the first. The new supermarkets/large grocery stores tended to open in neighborhoods that had few existing ones. For instance, only one new supermarket/large grocery store opened in any of the 15 neighborhoods that already had three or more supermarkets in 2013. In contrast, seven new supermarkets/grocery stores opened in neighborhoods with only one supermarket/large grocery store in 2013. In other words, 58.3% of the new supermarkets/large grocery stores went into neighborhoods, with only one of this type of food outlet in 2013. We surmise that neighborhoods with one supermarket/large grocery store demonstrate that they are viable and can support the enterprise. Such neighborhoods may also signal that they can support additional supermarkets/large grocery stores.

In Detroit, neighborhoods without supermarkets or large grocery stores have difficulty attracting new ones. The only exception is the CBD, where the city spent billions of dollars to revitalize it. That revitalization included the addition of two new supermarkets. However, the city has had limited success in spreading the revitalization efforts beyond the CBD, which may hamper the establishment of new supermarkets in communities that have none [140].

The Opening and Closing of Small Grocery Stores, Convenience Stores, and Corner Stores

This sector had devastating losses as more than a third of the small grocery stores, convenience stores, corner stores, and mini-marts closed during the study period. More specifically, almost half of the corner and convenience stores closed, and about a fourth of the gas stations selling food were closed. The findings indicate that we should pay more attention to this, as these small stores are important food access points for many residents. Detroit has a robust, healthy corner store initiative, but it is unclear how many participating stores are defunct [55,58]. No studies have tracked the closure of establishments participating in such initiatives. The extensive closure of corner and convenience stores needs immediate attention, as this may signal reduced food access for those who are unable to shop for food in locations that are not close to their homes.

Because store closures are so widespread, the city’s planning and transportation departments should ensure that there is adequate public transportation to get residents to and from food stores. Where there is not adequate public transportation, the city should develop partnerships with ride-share companies to take residents to and from food stores. Additionally, the city can help mobile food trucks or “veggie vans” selling produce from local farmers and gardeners to travel to neighborhoods that lack farmers’ markets or food stands. This can be done by establishing routes and designated stops at specific times each week.

7.5.4. Intermingling and Coexistence

We are not arguing that food deserts, food swamps, or supermarket redlining do not exist. Instead, we contend that these phenomena are not characteristic of Detroit’s food landscape. We show that the city’s food outlets are configured in ways that the current scholarship does not describe accurately. We found food outlets of different types that are intermingled and coexist with each other. Hence, the very food outlets like supermarkets, large grocery stores, small grocery stores, gas stations that sell food, liquor stores, and fast food that researchers theorize will supplant each other are located close to each other. So, it is not an either-or scenario where one set of food outlets is found instead of the other; in Detroit, they are often found alongside each other in various neighborhoods.

7.6. Food Access and Sustainability in Detroit

7.6.1. The Limited Success of Large-Scale Urban Farming

Around the time Detroit filed for bankruptcy, a local millionaire businessman, John Hantz, engineered a plan to purchase thousands of vacant lots or ones with dilapidated housing from the city’s land bank. Claiming he wanted to improve the livability and sustainability of the city, he promised to convert the lots into the world’s largest urban food farm; Hantz paid a pittance for the east-side properties. Hantz first proposed a horticultural farm that would make a profit by providing healthy food options to area residents. However, Hantz Farms became an urban tree farm by the time he purchased over 1500 parcels (more than 140 acres) in 2013. The farm would include orchards and maple trees to produce syrup. Detroit’s urban farmers and community gardeners were among those objecting to the sale of parcels to Hantz and the plan for an urban tree farm. Some of those farmers had difficulty acquiring land from the city for farming. Farmers were also afraid that a mega farm in the city could make it difficult for them to produce food and make a profit selling their products. Civic leaders who supported Hantz’s proposal saw urban farming as a way for the city to reinvent itself. Farmers opposed the mega farm but thought small urban farms could help the city feed itself and make up for the lack of supermarkets and green grocers [141,142,143,144].

To date, the city has not tracked the impact of selling off large amounts of lots to a single individual and the impact of Hantz Farms operations on east-side neighborhoods. Our data suggest fewer food outlets in the neighborhoods in 2023 than in 2013. We suggest that the Planning Department, Office of Sustainability, Food Policy Council, the Director of Urban Agriculture, the Land Bank, and the Detroit Economic Growth Corporation examine how the distribution of food stores has changed in the neighborhoods adjacent to Hantz Farms. This assessment should be done before civic leaders approve any more mass sale of lots to individuals or institutions. The entities named above should also develop clear policies about land sales and provide more equitable and transparent processes for residents to acquire lots for agricultural purposes or to add to the city’s green infrastructure.

7.6.2. The Promise of Small-Scale Urban Farming and Aquaponics

Detroit’s food advocates have developed interesting synergies for growing food sustainably, employing multipurpose land uses, and ensuring that fresh, healthy produce reaches the most vulnerable people at little or no cost. One such example is the collaboration between the Capuchin Soup Kitchen (founded in 1929), food pantry, and bakery; Gleaners’ Community Food Bank; and the Earthworks Urban Farm. Founded in 1998, the 1.25-acre Earthworks Urban Farm grows certified organic fruits and other foods. Participants adopt a food sovereignty approach of “knowing the origins of the food we eat” as the farm “strives to restore our connection to the environment and community.” Prior to the establishment of the farm, residents tended to purchase their groceries from local gas stations. The farm has an apiary, a greenhouse, and a hoop house to enable year-round production. Hundreds of volunteers—including those who also use the soup kitchen and food pantry—produce tons of food. They operate a seasonal farmers’ market, grow and sell transplants to thousands of gardens throughout the city, and have, at times, operated a mobile market and a community-supported agriculture cooperative [145].

Research has shown that small-scale urban farmers are a force for transforming the city [16]. About 20,000 Detroiters grew food in 2023. Groups such as Keep Growing Detroit track about 1645 family or home gardens, more than 100 school gardens, over 400 community, and more than 100 market gardens. They provide resources and training to gardeners, help sell products, and help promote the Grown in Detroit label. The Grown in Detroit brand identifies and markets locally grown and made products [146].

We studied community gardens, urban farms, and school gardens. However, we are aware that we have an undercount of community gardens or market gardens because we collected information on only those that were publicly identified or could be found on maps. There is secrecy around the location of many gardens, as some are guerilla gardens; they are operating without permits, or the gardeners and farmers do not want the locations to be revealed [147]. We respected the desire for privacy. It was also beyond the scope of our study to document family or home gardens.

The Detroit Public Schools have a robust edible schoolyard program that should be expanded to include more schools. The city should pay to staff the gardens year-round to ensure that students and neighborhood residents can use them as educational spaces and reliable food sources.

Detroit has urban agriculture ordinances that guide farming activities that are updated regularly. Residents can grow crops and raise animals in the city. As a result, residents combine farming with aquaponics. One entity, Central Detroit Christian (CDC) Farm & Fishery, uses a repurposed liquor store to grow herbs and microgreens, which are fed by the wastes from the tilapia vats containing 1700 fish in the basement. These are sold to area businesses. The 6000-square-foot former liquor store that serves as the aquaponics grow station was donated to the CDC. At first, the CDC wanted to convert the liquor store into a laundromat, but when those plans fell through, it turned to aquaponics instead. CDC also owns a soul food cafe and operates a garden [148,149,150].

7.6.3. Detroit’s Sustainability Action Agenda

Detroit has an Office of Sustainability that developed an action agenda in 2019 that wants all residents to thrive and steward resources in an equitable green city. The agenda’s first goal is to “increase access to healthy food, green spaces, and recreational opportunities. The sixth goal seeks to “transform vacant lots into safe, productive, sustainable spaces [151]. In 2023, Detroit appointed its first director of Urban Agriculture [146]. This appointment will help the city work on its food access sustainability goals.

7.6.4. The Green Grocer Project

The Detroit Food Policy Council recommends that there should be about 30,000 square feet of grocery space for every 10,000 residents. City officials report that Detroit is 90% of the way to realizing that goal. The Detroit Economic Growth Corporation uses the Green Grocer Project as one strategy to enhance food access and achieve the aim. The project provided more than USD 50 million to facilitate the construction of new grocery stores, expansions, and renovations of existing ones from 2010 to 2017. More than 35 stores received assistance. The program provided two years of funding of up to USD 25,000 to open new small grocery stores or assist existing ones in selling fresh, healthy, affordable foods. Detroit relaunched the program in 2024 with a budget of USD 525,000, and a goal of helping at least eight small groceries. The revived program will assist small-format specialty stores (1500–5000 square feet) that dedicate over 15% of selling space to groceries and fresh foods; mixed-market community stores (3000–15,000 square feet) that sell fresh produce, meat, dairy, specialty, organic, and alternatively sourced products; the construction of new stores selling fresh produce, meat, and dairy; and alternative format stores selling mostly fresh food [152,153,154].

Detroit should expand the Green Grocer Project to include more small stores. The city is not immune to spending large sums to attract national grocery chains. In 2011, the Michigan Economic Development Corporation provided USD 4.2 million in incentives to bring a Whole Foods store to the Midtown area. That same year, the same group provided Meijer with USD 3.3 million in tax incentives to build a new store in Detroit. Critics contend that such funding could go to independent grocers and small neighborhood food stores to help them procure and sell a greater amount of healthy foods [105,155,156].

7.6.5. Black Residents Build a Black-Owned Grocery Store and Food Cooperatives

Not all of Detroit’s Black residents are jettisoning the city for the suburbs or a reverse migration to southern states and cities. Some have adopted innovative strategies to keep food stores in their neighborhoods or add new ones. For instance, Neighborhood Grocery opened in 2023 in the Jefferson/Mack area after residents spent six years crowdfunding for it. It is a Black-owned grocery store–something unusual in the city [157].

We highlighted Brightmoor earlier as a neighborhood without a supermarket or large grocery store for over a decade. That changed in 2024 with the opening of the Abbott Resource Center Coop. The cooperative came into being through a USD 500,000 Priority Health Total Health Foundation grant. Funds were used to transform the Brightmoor Connection Food Pantry into a 6000-square-foot facility to house the pantry and cooperative [158].

Black urban farmers from D-Town Farm and the Detroit Black Community Food Sovereignty Network also opened their long-awaited food cooperative in 2024 in Middle Woodward. The USD 21 million project has taken years to realize—the co-op stocks products from over 40 local vendors. Half of the produce sold is organic, and the remainder is clean and conventional [1,159,160].

7.7. Limitations

This study has its limitations. It compares two time periods over a decade. However, annual documentation of the distribution of food outlets could provide more granular data to allow for more detailed tracking of store locations. In addition, this study did not analyze store size, sales volume, or food content in the outlets. Future studies should assess food quality, quantity, and price at different outlets. We collected data on store size and sales volume, and that data is awaiting analysis.

Our study did not examine the ethnic background of food outlet owners or managers to understand the relationship between who owns or manages the food outlets and how it is related to neighborhood demographic characteristics. More could be done to assess the availability of culturally desirable food. We collected data on ownership, racial/ethnic background, and cuisine type of the food outlet. We will examine these factors in future analyses.

The study focused on neighborhood analyses because Detroit is a city of well-defined neighborhoods that residents know and identify with. We did some census tract analysis, but there is potential for further expansion. The analysis focused on the city, but future research can focus on the metropolitan region, providing a broader context for understanding food access and urban planning. This will allow Detroit to be compared to the surrounding suburbs, offering valuable insights for policymakers and researchers.

8. Conclusions

The study identifies several dimensions of food access that need more attention and better tracking. Cities and other municipalities should assess their local food environments regularly to track the opening and closure of food outlets and use the findings to inform food policies. While much attention is paid to large food outlets, such as supermarkets and big grocery stores, small neighborhood food outlets are ignored. As the Detroit data indicate, many small stores have been closed. These closures have the potential to reduce food access for vulnerable populations. Ergo, we urge researchers, policymakers, and food activists to connect inquiries about food outlet distribution with questions related to population dynamics more explicitly. We also urge scholars to thoroughly test theses, such as food deserts, food swamps, and supermarket redlining, and to increase efforts to refine these concepts, as they are crucial tools for understanding and addressing food access issues.