Abstract
Background. Implementing food stores in deprived neighborhoods to improve access to healthy food is a debated topic. Aims. To uncover important contributors to the closure of grocery store interventions in urban food deserts. Method. We systematically reviewed both peer-reviewed and gray literature for publications on the failure of grocery store interventions. Results. We found nine articles on six different failed food stores. The reasons stated for closure included low sales, a lack of food retail experience, poor marketing, and difficulty in attracting and retaining a high volume of consumers from the local market. Discussion. Current literature on the topic of the closure of inner-city grocery stores does not have a concise rationale to explain why inner-city grocery store interventions were not successful. Conclusion. We must consider the most appropriate interventions to improve food environments in food deserts using local and national policies to address the social determinants of health.
Opening grocery stores in neighborhoods with few or no full-service grocery stores (stores that contain a full range of foods such as produce, dairy and frozen foods), commonly called food deserts, appears a sound business decision that is beneficial for communities. Access to grocery stores has been shown to improve dietary quality and health (Abeykoon, 2016). Literature on grocery store interventions in food desert neighborhoods is sparse and rarely discusses why the stores may not successfully continue operations. The purpose of this article is to uncover factors identified as having resulted in closure of grocery store interventions in food desert neighborhoods.
Method
We conducted a review to investigate grocery stores that opened in neighborhoods with poor healthy food access, yet were ultimately unsuccessful. Peer-reviewed and gray literature were systematically searched. “Gray literature” is used to describe materials not published commercially or indexed by major databases and may be of questionable relevance or quality but can contribute where the peer-reviewed literature is sparse such as was expected for this search (Mahood, Van Eerd, & Irvin, 2014).
The search methods were guided by procedures described by Effective Public Health Practice Project (EPHPP) and Mahood et al. (2014). The EPHPP provides a standardized means to assess study quality and develop recommendations. The steps for the EPHPP Quality Assessment Method that include (1) question formulation, (2) literature retrieval, (3) developing relevance criteria, (4) assessing studies for relevance and then for methodological quality, (5) data extraction and narrative synthesis, (6) peer review of the report, and (7) dissemination. For gray literature, a well-organized checklist and search strategy maintained focus and direction, and the same key words and strategy were used. Rigorous inclusion criteria ensured only the most relevant studies were included in the final analysis.
The research question was as follows: What is the current state of evidence about failed grocery store interventions in neighborhoods with few or no grocery stores? The search strategy (shown in Table 1) was developed using a PICO approach, which stands for population, intervention, comparison, and outcome. The population represented urban food deserts or neighborhoods reported as having few or no grocery stores. A grocery store opening was considered to be an intervention. This search did not specify a comparison group. The outcome of interest was closure, operationally defined as stores that closed after 5 years or fewer. Inclusion criteria were (1) failure or closure of a newly opened grocery store in an urban previous food desert or neighborhood described as having few or no grocery stores; (2) peer-reviewed English-language scholarly articles and gray literature published from 1950 to 2018. We chose this lengthy period because we expected to find few publications, and we wanted to review as much information as possible, however, no publications prior to 2004 were uncovered.
Search Strategy and Search Sources.
Note. Hierarchical list of peer-reviewed and gray literature databases searched for evidence brief (Mahood, Van Eerd, & Irvin, 2014).
Eight electronic databases were searched: Ovid MEDLINE, EMBASE, Cumulative Index to Nursing and Allied Health Literature (CINAHL), ProQuest Public Health, Web of Science, Scopus, PsycINFO, and Cochrane Library. The search strategy, initially developed on MEDLINE (see Table 1), was adapted to the other seven databases. The key words that guided the search strategy were developed based on three concepts: (1) food/nutrition environment (e.g., grocer*, supermarket*, food retail*, etc.), (2) intervention (e.g., implement*, develop*, establish*, etc.), and (3) health-related outcomes (e.g., obesity, body mass index [BMI], food habits, etc). Articles were relevant if the key words appeared in the title or the article abstract. Published and gray literature sources searched are in Table 1.
The search resulted in 214 records, which were exported into a RefWorks and EndNote folder by two research assistants. These records underwent deduplication, which removed 21 records. The remaining 195 records were screened based on the title and abstract. Figure 1 describes the search strategy and decision-making pathway. We retained 30 peer-reviewed and gray literature articles for full-text review, the rest did not meet our inclusion criteria based on the abstract and title review. An additional 21 papers were removed during full-text screening because they did not meet inclusion criteria.
Search strategy flow.
Results
The nine remaining articles reported on six different failed grocery stores and are listed along with their stated reasons for closure in Table 2. Examples were found from Brooklyn, New York (Morland, 2010); Flint, Michigan (Sadler, Gilliland, & Arku, 2013); Hartford, Connecticut (Haar, 2011); Richmond, Virginia (Gilligan, 2004); Syracuse, New York (Weaver, 2017, 2018); and Saskatoon, Saskatchewan (Engler-Stringer et al., 2016; Fitzpatrick, 2016).
Articles Meeting Inclusion Criteria.
All six stores were located in central neighborhoods within mid-sized cities ranging in size from about 100,000 to about 250,000 (with the exception of Brooklyn, New York, which was about 2.5 million) that previously had few or no full-service grocery stores and were described in the extracted articles as having areas of high rates of poverty. The stores could be grouped into two groups with slightly different objectives. The first aimed at offering a wide-range of basic grocery items such as vegetables and fruit, meats, dairy products, dried, canned, and frozen foods (The Good Food Junction in Saskatoon, Live to Eat Food Cooperative in Syracuse, East New York Food Cooperative in Brooklyn, and Community Pride Store in Richmond) (Fitzpatrick, 2016; Gilligan, 2004; Morland, 2010; Weaver, 2018). In addition to the range of items listed in the first group, the second group included an additional emphasis on deli and ready-made items (The Market at Hartford 21 and Witherbee’s Market). This second group were located in neighbourhoods that were reported as mixed income but still had high rates of poverty (Harr, 2011; Sadler et al., 2013; Turner, 2011; Young, 2013).
Three of the stores were privately owned (Gilligan, 2004; Harr, 2011; Sadler et al., 2013; Turner, 2011; Young, 2013), two were cooperatives (Fitzpatrick, 2016; Morland, 2010) and one was funded through various community and private donations (Weaver, 2017, 2018). Four of the stores were under 5,000 square feet (Fitzpatrick, 2016; Harr, 2011; Morland, 2010; Weaver, 2017, 2018) one was 8,000 (Sadler et al., 2013; Turner, 2011; Young, 2013), and one was 15,000 square feet (Gilligan, 2004). The stores were open for as little as just over 6 months to almost 5 years.
The Community Pride grocery was the only chain and was included because one of its stores opened and then closed soon thereafter. The stores in the chain were opened in predominantly African American communities and soon after the last store in this chain opened, the owner noticed houses in the neighborhoods were deteriorating and storefronts were vacant as a result of a residential and commercial exodus to the suburbs (Gilligan, 2004).
The reasons for closure put forward in the publications (Table 2) focused on business and marketing factors primarily. For example, the article by Morland (2010) suggests that appropriate business type, size of space, food retail experience, strong branding and marketing, committed and continuous leadership, and sufficient resources are crucial to generate the income for success. Others stated that poor sales and high cost of business operation were contributing factors, with monthly expenses typically far outpacing income (Fitzpatrick, 2016; Haar, 2011; Morland, 2010; Weaver 2017, 2018). The owner of the Witherbee’s Market in Flint, Michigan, spoke to the high cost of business operations “The 10,000 square-foot store needs to generate more than $1 million in sales a year to meet its needs. . . . Its utility bills are $8,000 a month in the winter—an astronomical cost that’s almost impossible to keep up with” (Sadler et al., 2013; Turner, 2011; Young, 2013). The Community Pride owner commented that the supermarket industry is a low-margin business and that the low profit margin can be a particular challenge when typical customer purchases are smaller compared to the larger stores located in the suburban neighborhoods (Gillian, 2004); a finding that was consistent with the closure of the Good Food Junction (Fitzpatrick, 2016).
Even with the help of financial support from the community, stores struggled to reach a point of self-sufficiency (Sadler et al., 2013; Turner, 2011; Young, 2013). The Eat to Live grocery outlet opened with funds from Syracuse University and private investors but was only open for 2 months before closing for 2 years to reevaluate management and internal operations (Weaver, 2017). The grocery outlet got a $400,000 grant from the state of New York and was able to reopen, but ultimately closed after another year (Weaver, 2018). The cooperative received more than $1 million dollars in grants total but was not able to remain profitable and open (Weaver, 2018).
Discussion
The stated reasons stated for closure of the grocery store interventions were few and not described in very much detail in most cases. The peer-reviewed literature discussed a lack of food retail experience, poor marketing, and difficulty in attracting and retaining consumers of high-volume purchases from the local market (Cummins et al., 2008; Morland, 2010). We supplemented the academic literature with gray literature (especially media articles) that also primarily discussed poor sales as a primary reason for closure.
The food environment in a neighborhood is dynamic and may reflect both local and national economic trends (Filomena, Scanlin, & Morland, 2013). Food purchasing behavior has changed over the past 2 decades (Cummins et al., 2008; Cummins, Flint, & Matthews, 2014). Consumers, particularly low-income consumers, may use a complex combination of resources for food access, including e-commerce, food banks, fast and convenience food outlets, and grocery stores. As Donald (2013) describes “the food economy mirrors the larger U.S. economy in that industry depends on low wages and slim margins. Consumers have come to depend on low-cost food products just to make ends meet” (p. 235).
While not the main focus of any of the reviewed publications, the high rates of poverty in the target neighborhoods may be an additional explanation for why the food store interventions were not successful. Research on the Good Food Junction found that the average purchase from the Good Food Junction totalled less than $10 (Fuller, Engler-Stringer, & Muhajarine, 2015). No marketing, food retail experience, or accounting will help a store remain open if the people in the area, the intended consumers, simply cannot afford to buy the food in the store.
Grocery shopping is affected by factors much more complex than the convenient location of a grocery store. Many studies focus on geographic proximity to grocery stores, validating a policy approach that emphasizes supply-oriented challenges rather than demand-oriented ones of affordability and need.
These supply-side reports evoke the adage “if you built it they will come” but there is still little empirical evidence that if you build food retail outlets, people will buy healthy food. Good quality, fresh healthy food is generally more expensive than junk food and more people either simply cannot afford it or choose not to. (Donald, 2013, p. 232).
City planning and policies that encourage more grocery stores to be located in food deserts will not be able to provide an adequate service to the surrounding community until the institutional and systemic issues affecting food choice are addressed.
Limitations and Implications
The public perceptions of the stores were not discussed and could have had an important impact on sales. The literature on the topic of failed grocery store interventions in food desert neighborhoods is limited and as such conclusions should be taken with caution. However, food environment research must move beyond its tendency toward description toward a greater focus on intervention.
Conclusions
The challenges that inner-city grocery stores face are significant: low profit margins and high overhead costs, for example. This is combined with consumers who have less time and knowledge for home meal preparation, living in an environment with abundant cheap, processed convenience foods, which may lead them away from traditional grocery store shopping.
We must consider the most appropriate interventions to improve food environments in low-income neighborhoods. If physical access is the primary issue, as may be the case with people with low mobility, then improving access may be appropriate. But increased access is not sufficient if low socioeconomic status is the primary problem. Poverty reduction could make food retail more viable in all areas of a city.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this research was provided by a grant from the Canadian Institutes for Health Research (No. 127084).
