https://orcid.org/0000-0001-8109-5862
Figures
Abstract
In recent years, African policymakers have embraced supermarket development as a form of food system modernization, but the impacts of the proximity of supermarkets on low-income households in a developing African city are largely unexplored. Using a sample of 680 households in seven of Accra’s poorer neighborhoods, we examine the dimensions and level of food security, household-level determinants, and what impact, if any, the arrival of supermarkets has had on local food security. Using two internationally accepted food-security metrics, (Food Consumption Score and Household Food Insecurity Access Scale), two economic measures (income equivalence and asset ownership), and satellite imagery of supermarket locations, we find that physical proximity to supermarkets is not significantly associated with household food security in low-income areas of Accra. The majority of poor households in Accra are currently food secure, but asset ownership is a much stronger predictor of food security than income. The relatively high degree of food security in Accra is likely related to credit-based relationships that households have with local food vendors. Further, given that supermarkets can compete with these so-called “informal” vendors for access to public space, these credit-based relationships will suffer with new supermarket development. Our study highlights the need for nuanced understanding regarding the role of consumer-retailer and retailer-retailer relationship in ensuring food access for low-income households in urban Africa.
Author summary
Our article examines the trend of supermarket development in Accra as a means of modernizing the food system. While development theory and literature have generally viewed the proliferation of supermarkets as a mechanism to reduce food insecurity in this context, we argue that supermarkets might exacerbate the issue, especially in low-income neighborhoods. Our study, based on 680 households in poorer areas of Accra, indicates that assets, not just income, are crucial for food security. Currently, food security in these areas is maintained through credit-based relationships with local vendors. However, the rise of supermarkets, competing for public space, might disrupt these relationships, potentially leading to food insecurity. Despite the lack of important relationship between supermarket development and food security, our article suggests that as supermarkets expand and contest public space, they could inadvertently increase food insecurity for impoverished households. We emphasize the urgent need for legislative adjustments to safeguard the current food access for the poor.
Citation: Fobi D, Waldman KB, Dwyer MB, Robeson SM, Blekking JP (2024) Negligible impact of supermarkets on food security (so far) in low-income neighborhoods of Accra, Ghana. PLOS Sustain Transform 3(10): e0000133. https://doi.org/10.1371/journal.pstr.0000133
Editor: Prajal Pradhan, University of Groningen: Rijksuniversiteit Groningen, NETHERLANDS, KINGDOM OF THE
Received: October 23, 2023; Accepted: September 19, 2024; Published: October 28, 2024
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: This work was supported by the Tobias Center for Innovation in International Development, Indiana University Bloomington (DF) and the Sustainable Food Systems Science, Indiana University Bloomington (DF). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Food insecurity is a long-standing problem worldwide and a major concern of governments and international development organizations. Despite advances in agricultural technology and production capacity, including modernized farm machines, postharvest handling, cooling and packing technologies, storage, and transport [1], about 2.4 billion people around the world experienced moderate or severe food insecurity [2]. These advances have both benefited from, and enhanced the establishment of supermarkets and mega-stores in the global South [3], which are an increasingly common feature of urban landscapes. However, there remains a mismatch between advances in agricultural technologies and general food security, given the latter’s dependence on several social and economic factors that facilitate food access, including income, assets, proximity, and social networks, among other factors [4,5]. For instance, sub-Saharan Africa (SSA) regularly exceeds adequate food availability, yet more than one-fourth of the population over 15 years old is considered food insecure [6], often the result of limited food accessibility.
Economic access for low-income households in urban Africa features high levels of “informal” work, whereby income can be seasonal and legal protections are limited. More than 75% of the urban workforce in African cities is engaged in informal employment—almost double the global average [7]. Ghana is not dissimilar, with an informal sector around 80% of its total workforce [8]. Income variability is also common among small business owners who experience fluctuation in sales [8]. Due to the seasonality of their income, poorer urban residents often rely on multiple food retailers, such as informal vendors (e.g., local street vendors and some open-air markets) from whom they can purchase food in small quantities frequently, as well as negotiate and obtain food reliably on informal credit arrangements [9].
Despite their relevance in achieving urban food security, informal vendors tend to face pressure from local authorities to exit public spaces in order to modernize SSA cities. Enforcing laws against street vending and, to a lesser but not insignificant extent, against open-air markets is common across SSA and reflects a debate about both who owns public space [10] and who wields political power [11]. In the last two decades, for instance, local authorities in Ghana’s capital city, Accra, have implemented policies based on the premise that economic growth and investment are the fundamental goal of urban governance [12]. Through zoning laws and other incentives aimed at “modernizing” urban space, city planners have begun to allow local and international developers to build infrastructure such as supermarkets that cater to higher- and middle-income consumers [13]. At the same time, authorities in Accra regularly evict informal vendors using spatial-control regulations that prohibit them from operating in downtown Accra and its immediate surroundings [14]. In recent years, these evictions have escalated in terms of force, violence, and frequency [14]. Informal vendors in Accra are not only forced out of their vending spots, but often their wares and vending stands are confiscated under the guise of modernization of the urban environment development [12].
Most of the literature on supermarketization, informal vendors and food insecurity have focused on shopping preferences and behaviors among low-income consumers [15,16], relationship between informal vendors and supermarkets [17,9] and the vulnerability and resilience of informal vendors and how they support low-income consumers in poor SSA countries [18–21]. There is however a gap in assessing the influence of the longevity and distance of supermarkets on low-income consumers in large African cities like Accra, and the effectiveness of income versus asset ownership for maintaining food security among low-income urban residents. This knowledge gap obscures our understanding of the extent to which an established supermarket can increase or reduce food security in a low-income neighborhood, but also how poor urban residents strive to maintain food security in the face of inconsistency in jobs and income.
Our paper fills this gap by examining this relationship empirically, using two internationally accepted metrics to examine the state of household-level food security and the influence of supermarket growth in and around seven poor neighborhoods of Accra metro. Using a combination of geospatial data on supermarket growth in Accra and household data from 680 households in seven of Accra’s poorer neighborhoods that have largely remained outside the zone of new supermarket development, we examine (1) the dimensions and level of food security in these neighborhoods; (2) the factors that best explain our household-level findings; and (3) what impact, if any, the arrival of supermarkets in the broader area has had on local levels of food security. Our paper demonstrates the urgency of policy change to protect existing means of food access for the urban poor in SSA. In the process we discuss the risk of applying the concept of “food deserts” uncritically to large cities in SSA. Food deserts are generally conceptualized as areas with limited local access to food, as typically measured by factors like income, transportation, time, and proximity to grocery stores. Building on the work of previous research we highlight the inappropriateness of the food desert concept in this context, and show how supermarkets negligibly impact food security in areas that, despite their poverty, are relatively food secure.
2. Materials and methods
2.1 Study site
Accra is one of the fastest growing cities in SSA [22]. As of 2020, Accra had an estimated population of about 2.5 million people [23] excluding people who commute daily to the city for various socio-economic activities. It is the seat of government, and many businesses are concentrated in the city, pulling many people to move there or commute daily from other regions. Close to 10% of the economically active population in the Accra Metropolitan Area are unemployed [24]
Seven low-income neighborhoods within the Accra Metropolitan Area (AMA) are studied (Fig 1). James Town and Ussher Town are located along the coast in Accra central–near the central business district of Accra. These areas have well-defined boundaries, but the unplanned nature of buildings in these residential areas has resulted in the development of slums. Kaneshie, Bubuashie, Nima, and Maamobi are located north of Accra’s downtown. These residential areas are generally surrounded by neighborhoods inhabited by academics, government officials, and foreign ambassadors. Abossey Okai, also located north of downtown Accra, is surrounded by neighborhoods such as Dansoman which is predominantly a working-class community and considered to be the largest neighborhood in Accra [25]. Other surrounding low-income neighborhoods such as Mataheko, New Town, Russia and Darkoman were included in the study (these were referred to as “other” in Fig 1). These were not the focus of the study but some of the respondents lived in these neighborhoods and thus, were included in the study. The AMA classifies James Town, Ussher Town, Kaneshie, Nima, and Maamboi as high poverty neighborhoods, while Bubuashie and Abossey Okai are classified as moderate poverty neighborhoods [26]. Poverty criteria outlined by CHF International [26] were used to develop a poverty index for each neighborhood. CHF uses qualitative and quantitative measures to calculate aggregate poverty. Qualitative poverty measures include solid waste disposal, water supply, room occupancy, toilet facilities and housing type. Quantitative measures assigned index to income levels, housing density and population density. Poverty neighborhoods were established because of their undesirable qualitative measures, that is, a very high percentage of their population lacked desirable living conditions, but also they lived on less than $1 a day, their housing density was above 1500 per kilometer square and their population density was above 30,000 people per kilometer square [26].
Other neighborhoods include Mataheko, New Town, Darkoman and Russia in AMA. Basemap source: Esri (2023).
2.2 Data
Data from the Accra Urban Food Security Survey (AUFSS) and supermarket location information from Google Earth Pro were used for this analysis. AUFSS data was used to calculate two standard measures of food security: (a) the Food Consumption Score (FCS), an indicator of dietary diversity; and (b) the Household Food Insecurity Access Scale (HFIAS), an indicator of food access. In the 2017 AUFSS, which was conducted in Accra over a three-week period in July/August, 680 households were surveyed. The sample method used was structured-area sampling [27]. The selection of residential areas was based on San Diego State University vernacular neighborhood maps that highlight finer-scale (census enumeration area) socioeconomic characteristics [28,29]. To completely cover each residential area, local enumerators conducted household surveys at roughly even spatial intervals based on the density of homes. In addition to FCS and HFIAS questions, the survey asked questions about market preferences, housing, assets, income, and socioeconomic characteristics of households. One willing adult per family who was aware of the household budget and food purchases was interviewed by enumerators. However, enumerators gathered up to five household members’ labor characteristics as well as demographic data for the entire home. On iPad tablets, data was gathered using Qualtrics’ mobile data gathering platform in Abossey Okai, James Town, Maamobi, Ussher Town, Kaneshie, Bubuashie and Nima neighborhoods.
To evaluate spatial patterns of food security in low-income neighborhoods in Accra, we map responses to the FCS and the HFIAS. The FCS was established in 1996 by the United Nations World Food Program (WFP) to estimate rural family food consumption rates in southern Africa [30]. The FCS is a composite score depending on how often a variety of meals are consumed [30]. A score of 0–7 is assigned to each household food item consumed depending on how many days in the preceding 7 it was consumed by the household. After that, the foods are categorized (e.g staples, vegetables, meats). Food groups that are consumed more than seven times per week are re-coded as seven. The caloric density of each food group–staple (2), pulse (3), vegetable (1), fruit (1), mean and fish (4), milk (4), sugar (0.5) oil (0.5) and condiments (0)–determines its weight [30]. Each household’s final FCS is calculated by multiplying the number of days of consuming a specific food group by the food weight, and then finally adding the weighted scores as shown in the equation below:
Where Fi is the food group, n is the number of food groups and wi is the standard weight of each food group that reflects the energy content of the food group [30].
The smallest possible score is 0 and the maximum possible score is 112. Households with composite scores below 21 have "poor" dietary diversity consumption, those between 21 and 35 have "borderline" dietary diversity consumption, and those over 35 have "acceptable" dietary diversity consumption [30]. FCS positively correlates with caloric intake and has been used to measure food security levels in two SSA countries–Burundi and Malawi–indicating its reliability to measure food security in Ghana [30]. FCS also correlates with other food security indicators such as Household Dietary Diversity Score (HDDS) [31].
HFIAS is an experience-based metric that captures household-level responses to food shortages within the last 30 days [32]. HFIAS was developed by the USAID-funded Food and Nutrition Technical Assistance II project (FANTA) to measure the incidence of household food insecurity and changes in food insecurity over time. Respondents were asked to identify if a particular action occurred, and if it did, how frequently that action occurred (“rarely”, “sometimes”, or “often”). The response of each of the nine questions in the questionnaire was transformed into a continuous indicator of food security and scored from 0 to 3, with 3 representing the highest frequency of occurrence (“often”). The range of possible HFIAS scores is 0–27. Typically, households with lower HFIAS scores experienced less difficulty accessing food than those with higher HFIAS scores [31].
2.3 Economic measures
Income equivalence of households was calculated using self-reported income from salaries, rents, remittances, and gifts. We summed the monthly total income of all household members and divided by the square root of the number of household members in each household as shown in the equation below:
Where n = number of income types, Ii = type of income, x = number of household members.
Dividing by the square root of the number of household members allows a fair income distribution across the households so that income will not be disproportionately affected by unequal number of household members [33], especially if there are more nonworking than working household members. Income equivalence for the month was used instead of just total monthly income because household total monthly income does not consider how many members within the household are depending on that income and thus, it may be inappropriate to correlate with household level food security metrics. Income equivalence in Ghana Cedis was ranked into five indices with 5 being the largest income equivalence.
An asset index was calculated from a series of questions asking respondents whether they possessed certain assets in working condition in their house. The asset index includes refrigerators, mobile phones, vehicles, televisions, radios, computers, bicycles and motorbikes. Each asset is scored “1” if the household owns it and “0” if they do not. We applied principal component analysis (PCA) to produce a wealth index that assigns a larger weight to assets that are uncommon across the household. Assets that are common in all households are given a weight of zero [34]. The results were categorized into quintiles, with 5 being the largest and indicative of larger wealth. Using assets as a proxy for wealth is useful because it depicts a household’s economic status in the long run and is less responsive to short-term fluctuations from economic shocks or prosperity [35]. Asset was used as an economic measure because it is a proxy for income in Ghana and can be used as an exchange for goods and services but also serve as a guarantee for acquiring goods on credit [36]. We use Spearman’s correlation coefficient to evaluate the association between the two economic proxies (income equivalence and household assets) and the two food-security indices (FCS and HFIAS).
2.4 Supermarket locations
In this study, we define the term “supermarket” as a foreign or local grocery store that has four or more locations within the Accra Metropolitan Area (AMA). Focusing on this definition enabled us to clearly distinguish large grocery store chains from smaller ‘mom-and-pop’ food retail shops in Accra. Companies’ websites, business directories, and Facebook pages were used to compile the initial list of supermarkets. We used Google Earth Pro to visually validate foreign and local supermarket chains based on the list and recorded their geographic coordinates from Google Earth Pro. A total of 43 supermarkets owned by 5 companies were included in the study from 2000 to 2017: Koala (3), Shoprite (5), MaxMart (5), Spar (10), and Melcom (20). We then located each supermarket in the most recent images from Google Earth Pro, working back through the archival images to determine when they were constructed (by determining when the structure was erected) and opened (by assessing when vehicles appeared in the parking lot). One limitation of using the Google Earth Pro was that some images of supermarkets in AMA had cloud cover, making it somewhat difficult to identify them and their exact geographic coordinates. Also, some supermarkets have renovated existing buildings. The unclear Google images made it hard to identify when these buildings started to be used for the supermarket. To ensure correctness about the Google Earth Pro images, we complemented them with online local news reports and official store websites to validate the year the supermarkets started operation. We then used this information to create spatial data layers of supermarkets and their year of establishment using ArcGIS Pro. Maps throughout this paper were created using ArcGIS® software by Esri. ArcGIS® and ArcMap™ are the intellectual property of Esri and are used herein under license.
To identify the spatial relationship between households and their nearest supermarket, we calculated the distance between households and the nearest supermarket using the Near tool in ArcGIS Pro. From this, we were able to identify the closest supermarket to each household, as well as the approximate age of the nearest supermarket relative to the 2017 date of the survey collection data. For example, if the nearest supermarket was identified as Koala supermarket and it began operation in 2014, then the age of that supermarket is 3 years. We perform two sets of analysis. First, we examine the correlation between the distance from the household to the nearest supermarket and the food security metrics (FCS and HFIAS). Second, we examine the correlation between the age of the nearest supermarket, dietary diversity (FCS), and food access (HFIAS) as shown in section 3.4. An alpha level of 5% was used to determine the significance of Spearman’s correlation between supermarket age and distance, and the food security metrics.
3. Results
3.1 Household characteristics
Household size ranged from 2 to 12 members with a mean of 4. The mean monthly household income equivalence was GHC606.20 ($121.24) and a median of GHC246.70 ($49.34). Approximately 83.4% of households surveyed purchased food from open air markets, 0.3% from supermarkets, 11.9% from hawkers and street vendors, and 4.4% purchased cooked food from the streets. More than half (61%) of the sampled households walked to purchase food, 31% used motorized transport to purchase food and 8% were not sure about the mode of transport their household used to purchase food. The household demographics have been summarized in Table 1 below:
3.2 Food Consumption Score (FCS) and Household Food Insecurity Access Score (HFIAS)
A majority of the households (97%) scored ‘acceptable’ values for dietary diversity (>35), which suggests that the majority of households eat several varieties of food. Most of the households sampled did not have difficulty accessing adequate amounts of food in a typical month for consumption as most of them (63%) score below 7 on the HFIAS. The spatial distribution of the FCS and HFIAS in Accra’s poorest neighborhoods show that most of the sampled households in low-income neighborhoods in Accra consume acceptable levels of both dietary diversity (Fig 2A), but also generally do not experience much difficulty in accessing food (Fig 2B). There are a few households that have poor dietary diversity (FCS) and food inaccessibility (HFIAS); however, the majority of households in our sample have acceptable levels of food diversity and / or food accessibility.
The mean household value for FCS is 65.92. Majority of poor households have acceptable levels of dietary diversity. Basemap source: Esri (2023). Fig 2B: Household Food Insecurity Access Scale (HFIAS) scores of respondents of low-income neighborhoods in Accra, 2017. The mean value for HFIAS is 5.5. Majority of poor households in Accra rarely experience problems of food inaccess. Basemap source: Esri (2023).
3.3 Supermarket influx in Accra
The number of supermarkets increased steadily from 5 in 2005, to 16 in 2010, and to 43 in 2017 (Fig 3 below). Melcom Shopping center is the oldest foreign supermarket chain in the country, with its first official outlet in Ghana opened in Accra Central in 2000. MaxMart is the oldest local supermarket chain in Ghana with its first outlet at 37 Liberation Road, Accra, and started in 2001. Shoprite and Koala Shopping chains are relatively recent, both beginning operation in Accra in 2008. Maxmart and Koala Shopping center are Ghanaian-owned companies. Supermarkets in this area, and throughout AMA, are usually located along major roads to improve access for consumers.
The number of supermarkets increased steadily from 2005 through to 2017. Basemap source: Esri (2023).
3.4. Correlation among economic measures, food security and supermarkets
The income equivalence index is slightly negatively correlated with HFIAS scores (Fig 4A), which measures behavior such as reducing food quantities consumed by the household due to inadequate resources. There is a stronger connection between FCS and household income equivalence (Fig 4B). The highest income equivalence index has a lower median food consumption score than all but the lowest category. Household assets strongly correlate negatively with HFIAS (Fig 5A). FCS does not vary much across the household asset indices (Fig 5B), indicating that households with greater assets may not typically consume more diverse foods.
Household Food Insecurity Access Scale (HFIAS) decreases with income equivalence (4A) suggesting that as income equivalence increases, the difficulty in accessing food reduces. Food Consumption Score (FCS) increases as household income equivalence increases (4B).
Household Food Insecurity Access Scale (HFIAS) consistently decreases across asset categories (5A), but Food Consumption Score (FCS) does not vary much across the household asset categories (5B), indicating that households with greater assets may not typically consume more diverse foods.
Table 2 below displays the correlation between economic measures and FCS and HFIAS. An alpha level of 5% was used to determine the significance of Spearman’s rank correlation coefficient. Income equivalence has a weak positive linear relationship with FCS, and it is statistically significant (rs = 0.16, p<0.01). The asset index has a moderate negative linear relationship with HFIAS (rs = -0.33, p<0.01). It is statistically significant.
The two food security metrics (FCS and HFIAS) were evaluated with respect to nearest supermarket distance and age of the supermarket as shown in Fig 6 below.
FCS has a weak distance-decay relationship, while HFIAS appears to have no relationship to supermarket distance. Bottom horizontal scatter plots show the relationship between food metrics–FCS and HFIAS–and supermarket age. There appears to be no apparent relationship between the two metrics and supermarket age.
The Spearman’s correlation coefficient showed there is no significant relationship between supermarket distance and FCS (p = 0.520 > 0.05) and HFIAS (p = 0.638 > 0.05). Similarly, age of nearest supermarkets does not significantly correlate with FCS (p = 0.280> 0.05) and HFIAS (p = 0.577> 0.05)
4. Discussion
4.1 Food security among low-income residents in Accra
Our results suggest that the majority of households in the seven and other surveyed low-income neighborhoods in Accra have “acceptable dietary diversity”–they eat a wide range of food from staples and vegetables to meat and fish and this translates into an acceptable caloric intake [37]. This is evident in the spatial visualization of the FCS and HFIAS. It is not surprising that acceptable levels of dietary diversity are widespread even in low-income neighborhoods because fresh staples such as yam, plantain and maize, and vegetables are not out of reach of the urban poor in SSA cities [38]. Low-income neighborhoods largely depend on several informal procurement sources such as street vendors. These outlets sell fresh fruit and vegetables at relatively lower prices, are easily accessible and allow for informal credit arrangements [39–41]. Additionally, transportation does not pose a barrier to acquiring groceries as the majority of low-income households walk to buy food. Although the majority of low-income neighborhoods appear to reach acceptable caloric levels due to the ubiquity of informal vendors [42], some of them sometimes experience food accessibility problems. We interpret this to be influenced by factors such as the household’s social networks outside of the poor neighborhood, the health of family household members or food preferences [43]. Our results however do not generalize the idea that every SSA city has acceptable levels of food security. In some cases, low-income neighborhoods in conflict areas may not necessarily achieve food security even in the presence of informal vendors [44].
4.2 Factors explaining household-level food security
The relationship between the two economic measures we use–income equivalence and asset index–and food security metrics suggest that while both economic measures have statistically significant relationships with FCS and HFIAS, the asset index has the strongest statistical significance with the food metrics. An increase in the asset index is associated with a decrease in HFIAS and an increase in FCS. We interpret this finding to be facilitated by flexibility among informal vendors to make credit arrangements possible based on household assets and personal relationships with consumers [41]. Income equivalence also shows a strong association with both HFIAS and FCS. However, the relationship between income equivalence and HFIAS does not vary much, especially, the third, fourth and fifth median are almost the same. This is interpreted that higher income equivalence may enable households to achieve acceptable FCS scores but there may not be much difference in food access (HFIAS) between different income-level households.
Inadequate income reduces urban families’ ability to purchase food [37]; hence, it follows logically that households with greater purchasing power can consume more varieties and quantities of food than those with smaller income equivalence. In the results, however, the highest income equivalence had low FCS. This deviates from previous studies that household dietary diversity is strongly correlated with household per capita income [45]. We attribute our finding to the novel income measure we used. Our method essentially considered the number of household members in relation to total income of household, ie. income equivalence. This suggests that higher income does not necessarily translate into higher dietary diversity especially if there are many household members who do not contribute to the household income. In such applications, using income equivalence suggests that providing adequate food for all household members may reduce household caloric intake. We interpret our finding to mean that some low-income households may have many household members or other pressing needs to attend to, including paying utility bills, school fees, among others. Research by [5] showed that food expenditure is related to the welfare of households such that an increase in food budget decreases the share of household income to provide other essential non-food resources that promote a decent standard of living. Thus, our result suggests that households may have higher income equivalence but if there are other relevant pressing issues, they may cut back on their dietary diversity. We interpret the relatively weaker correlation between income equivalence and HFIAS as the result of inconsistency of income related to employment “informality”, which can contribute to chronic food insecurity levels [46]. Income equivalence is significantly correlated with FCS because FCS evaluates consumption in terms of caloric intake and dietary diversity, and it is relatively cheaper to eat a ‘diverse’ diet in many African cities.
The relatively strong relationship between HFIAS and household assets suggests that the gap created by unstable income throughout the year is moderated by assets [35]. Asset indices are often used as proxies for wealth in rural and some urban settings for this reason. It is common for households to procure food on credit from informal vendors because the vendors know they own assets that could be traded to pay back [36]. Therefore, the assumption that food security in urban areas is solely dependent on an individual or a household’s income may be misleading [47]. Instead, our study suggests that urban household food security could be understood better by examining both income and asset ownership, especially among poor households.
4.3 Supermarket development in Accra
Supermarkets are a relatively recent development in Accra (began operation in 2000) compared to other African cities such as Pretoria and Harare, where SPAR supermarkets were established in the 1960s [48]. The gradual but steady establishment of supermarkets in Accra illustrates the continued proliferation of the supermarket industry over the last two decades and highlights the role city planners have played in facilitating their entry into the urban food system of Africa. Previous research from Zambia and other developing countries show that national and local governments allow supermarkets into cities because it provides them with more revenue through import tariffs but also adds to the aesthetics of the urban landscape [49,50]. There are mixed findings about whether supermarkets contribute to alleviating food insecurity, and how supermarkets relate to informal vendors. For instance, [51] find that supermarket purchases are positively associated with children’s height-to-age measurements; however, in some cases, supermarkets are associated with obesity in adults [52]. Some studies find that supermarkets are likely to crowd out local food sources and perpetuate food insecurity [18]. Others argue that informal vendors are resilient and competitive, and they can often have a symbiotic relationship with supermarkets [17], where informal vendors source non-traditional food from supermarkets at wholesale prices. Still other scholars suggest that supermarket locations do not matter–that food insecurity and poor diets are more a function of consumer preferences than consumer access to healthy food [19]. In general, it seems that focusing on supermarkets in urban planning efforts without consideration of the role other actors, like informal vendors, play in an urban food system may not assuage food insecurity. This is because other retailers are particularly important to the food purchasing behavior of vulnerable households [20].
4.4 Supermarket arrival and local food security in Accra
The correlation between the food metrics, and supermarket distance and age show no significant results. While it appears that supermarket distance correlates with FCS, this relationship is not significant. We interpret this to mean that the absence or lack of proximity to supermarkets do not significantly reduce or increase local food security. Similarly, the length of time since the establishment of supermarkets in Accra does not significantly correlate with FCS and HFIAS. More seasoned supermarkets like Melcom that are relatively older within the Accra metro do not impact household-level food security levels. With the negligible impact of both supermarket distance and age, our findings suggest that supermarkets do not increase or decrease food security for low-income households in Accra. Our findings provide the pragmatic spatial and statistical evidence that buttresses the assertion that low-income neighborhoods in Accra have acceptable levels of food security despite households infrequently using supermarkets. This is consistent with [53] who found that informal vendors are the main sources of food for the urban poor in Accra. The results also support [9] assertion that the majority of poor urban households in SSA broadly and Ghana specifically, rely on informal retail outlets, despite the massive penetration of supermarkets.
Our result supports previous research [54,55] that finds supermarkets do not have a substantial, positive relationship with food security in low-income African cities. Importantly, our finding buttresses the assertion that supermarkets are not economically accessible to poor SSA households because they are relatively more expensive than other traditional outlets, they do not offer flexible credit arrangements, and groceries are often in bulk quantities unlike informal vendor outlets [56]. Despite the rapid proliferation of supermarkets, open-air markets in Accra are generally larger, more numerous, and more accessible than supermarkets. FAO [57] found that the high density of street vendors in Accra places them within easy walking distance of nearly every household. Although it is not clear how many open-air markets exist in Accra, Google Earth images show about 50 open-air markets within Accra. Informal vendors provide poor urban consumers with more opportunities for price matching and convenience [58]. To maximize patronage and profits, supermarkets are usually strategically located in wealthier neighborhoods in urban Africa [56]. However, this is changing in some countries such as Angola where Shoprite now has lower-end subsidiaries that are being developed to target lower-income households [59]. When supermarkets are located in poorer neighborhoods, they typically provide affordable, but lower quality food products [54,47], thus even lower-end subsidiaries contribute to unequal food access despite being spatially available.
4.5 Implications and limitation of study
Our findings from this research imply two things for the city of Accra. First, that low-income neighborhoods in Accra are not necessarily areas of low food access, but the continuous expansion of urban beautification through supermarketization and at the same time removal of informal food vendors may risk creating low food access areas. As stated earlier, our results suggest that despite decades of supermarket development there appears to be no positive or negative association of supermarkets with the food security of low-income neighborhoods in Accra. This may be an indication that the evictions and crackdowns that take place in downtown Accra have not been systematic enough in the surrounding neighborhoods that they had a measurable correlation on the food security indicators we use. Second, that income only or asset only can be used to achieve some level of food security but that the combination of income and asset may be more efficient to explain why low-income neighborhoods in Accra are not areas of limited food accessibility. This implication is linked with the existence and operation of informal vendors and implies that ensuring high food quality and fair prices among informal vendors should be a priority of the local authorities in Accra. This could be achieved by encouraging informal vendors to sell fresh food and vegetables further away from open gutters and dusty environments, but also using a common measurement such as weight (kilograms or pounds) to ensure portion control and fair pricing.
Our research data is limited by the fact that it does not include informal vendor interviews or survey data about their credit arrangement with poor urban residents. Future research on this will provide clarity on why and how informal vendors make such flexible and valuable arrangements to contribute to food security in poor neighborhoods in Accra. The study is also limited in the use of FCS as a food security metric and means of evaluating dietary diversity. Urban areas typically do not suffer from food unavailability, specifically for the food categories FCS asks about, yet FCS assumes some level of food unavailability [46]. In terms of dietary diversity, FCS does not consider portion size, nor does it provide insights into the nutritional contribution of the foods consumed. Because FCS is based on calorie consumption, it assigns higher ratings to high protein diets such as meat, dairy, and other animal products, and lower weights to fresh fruits and vegetables [32], even though fruits and vegetables provide important micronutrients. Despite these weaknesses, we use FCS to understand the usual everyday caloric intake of the sample population [60]. Future research should utilize metrics that are more sensitive to staples consumed in Africa to appreciate the consumption levels among households.
5. Conclusion
Over the last two decades, supermarkets have steadily penetrated many SSA cities, and Accra is no exception. City planners have collaborated with private sector actors to modernize Accra—often with the intention of beautifying cities, increasing government revenue through tax payment, and providing other urban amenities, including supermarkets.
Our research finds no significant association between household food security and the time a supermarket has been in operation or the physical proximity of a supermarket. We also find a stronger relationship between food security, measured in terms of access and dietary diversity, and asset ownership than we do with income equivalence. Because supermarkets have no measurable impact on food security, this study highlights the importance of informal vendors on ensuring food security, particularly for low-income households. These findings also suggest that neither of the traditional food desert tenets–income nor supermarket proximity–associates strongly with food security in low-income neighborhoods in Accra. Our research found that more than 90% of households surveyed purchased food from informal vendors. Based on our results, we assert that the embeddedness in local communities of food vendors that are widely framed (and periodically criminalized) as “informal” is precisely what allows poor neighborhoods in Accra to translate local households’ assets into the credit mechanisms on which local food security often depends.
We contribute to the supermarket-household consumption literature by demonstrating that supermarkets have not yet compromised food access and dietary diversity in low-income areas of Accra. Our results suggest that policies that support modernizing SSA cities through developments such as large grocery chains without concomitant support for informal vendors may be counterproductive. The findings suggest that city authorities’ continual granting of public space to modernization and local politicians’ inconsistent support to informal vendors may change the urban African food system. This may truncate domestic rural to urban food flows and block the most essential pathway through which many poor urban households access food. We suggest that city authorities take a broader approach to food system development to consider the many ways that different types of food retailers interact to reach households across the income spectrum.
Supporting information
S1 Data. Data underlying results for household analysis.
https://doi.org/10.1371/journal.pstr.0000133.s001
(CSV)
S2 Data. Data underlying results for supermarket analysis.
https://doi.org/10.1371/journal.pstr.0000133.s002
(XLSX)
References
- 1. Norton R. Agricultural value chains: A game changer for small holders. Retrieved March. 2014 Jul;20:2017.
- 2.
UNicef. The state of food security and nutrition in the world 2023: urbanization, agrifood systems, transformation and healthy diets across the rural-urban continuum.
- 3. Atkin D, Faber B, Gonzalez-Navarro M. Retail globalization and household welfare: Evidence from Mexico. Journal of Political Economy. 2018 Feb 1;126(1):1–73.
- 4.
Sen A. Poverty and famines: an essay on entitlement and deprivation. Oxford university press; 1982.
- 5. Donkoh SA, Alhassan H, Nkegbe PK. Food expenditure and household welfare in Ghana.
- 6.
FAO E ECA. Regional overview of food security and nutrition. addressing the threat from climate variability and extremes for food security and nutrition. Food and Agriculture Organization of the United Nations: Accra. 2018.
- 7.
Sow, M. Figures of the week: Informal employment in African cities. Africa in Focus. 2018 https://www.brookings.edu/blog/africa-in-focus/2018/06/01/figures-of-the-week-informal-employment-in-african-cities/#:~:text=Former%20Research%20Analyst%20%2D%20Africa%20Growth%20Initiative&text=The%20report%20states%20that%20most,world%20average%20of%2044%20percent.
- 8.
Osei-Boateng C, Ampratwum E. The informal sector in Ghana. Accra: Friedrich-Ebert-Stiftung, Ghana Office; 2011 Oct.
- 9. Anku EK, Ahorbo GK. Conflict between supermarkets and wet-markets in Ghana: Early warning signals and preventive policy recommendations. International Journal of Business and Social Research. 2017;7(10):1–7.
- 10. Brown A. Legal paradigms and the informal economy: Pluralism, empowerment, rights or governance?. InRebel Streets and the Informal Economy 2017 Feb 24 (pp. 19–36). Routledge.
- 11. Resnick D. The politics of crackdowns on Africa’s informal vendors. Comparative Politics. 2019 Oct 1;52(1):21–51.
- 12. Spire A, Choplin A. Street vendors facing urban beautification in Accra (Ghana): eviction, relocation and formalization. Articulo-Journal of Urban Research. 2018 Jan 1(17–18).
- 13. Josse G. Planning and managing Africa’s cities: what place for technology innovations?. Field Actions Science Reports. The journal of field actions. 2020 Dec 23(Special Issue 22):52–7.
- 14. Blot J, Spire A. Evictions and conflicts around urban legitimacy in the cities of the South. The Political Space. Online journal of political geography and geopolitics. 2014 Mar 18(22).
- 15. Ma X., Liese A. D., Hibbert J., Bell B. A., Wilcox S., & Sharpe P. A. (2017). The association between food security and store-specific and overall food shopping behaviors. Journal of the Academy of Nutrition and Dietetics, 117(12), 1931–1940. pmid:28366811
- 16. Lenk KM, Winkler MR, Caspi CE, Laska MN. Food shopping, home food availability, and food insecurity among customers in small food stores: an exploratory study. Translational behavioral medicine. 2020 Dec 1;10(6):1358–66. pmid:33421081
- 17. Oteng-Ababio M, Arthur IK. (Dis) continuities in scale, scope and complexities of the space economy: The shopping mall experience in Ghana. In Urban Forum 2015 Jun (Vol. 26, pp. 151–169). Springer Netherlands
- 18. Reardon T., & Hopkins R. (2006). The supermarket revolution in developing countries: Policies to address emerging tensions among supermarkets, suppliers and traditional retailers. The European journal of development research, 18, 522–545.
- 19. Allcott H., Diamond R., Dubé J. P., Handbury J., Rahkovsky I., & Schnell M. (2019). Food deserts and the causes of nutritional inequality. The Quarterly Journal of Economics, 134(4), 1793–1844.
- 20. Wanyama R, Gödecke T, Chege CG, Qaim M. How important are supermarkets for the diets of the urban poor in Africa?. Food Security. 2019 Dec;11(6):1339–53.
- 21. Giroux S, Blekking J, Waldman K, Resnick D, Fobi D. Informal vendors and food systems planning in an emerging African city. Food Policy. 2021 Aug 1;103:101997.
- 22.
Oluwole, V. Fastest growing cities in Africa 2021. Business Insider Africa. 2021. https://africa.businessinsider.com/local/lifestyle/fastest-growing-cities-in-africa-2021/b97e271
- 23.
World Population Review. 2021 https://worldpopulationreview.com/world-cities/accra-population.
- 24.
Ghana. Statistical Service. 2010 Population & housing census report: Disability in Ghana. Ghana Statistical Service; 2014.
- 25.
Gambrah A. Regulation of housing and services for the Urban Poor: a Case Study of Accra, Ghana (Doctoral dissertation, University of British Columbia).
- 26.
Cooperative House Foundation International (CHF). Accra poverty map: a guide to urban poverty reduction in Accra. 2010
- 27. Montello DR, Sutton P. An introduction to scientific research methods in geography and environmental studies.
- 28. Weeks JR, Hill A, Stow D, Getis A, Fugate D. Can we spot a neighborhood from the air? Defining neighborhood structure in Accra, Ghana. GeoJournal. 2007 Jun;69:9–22. pmid:19478993
- 29. Jankowska MM, Weeks JR, Engstrom R. Do the most vulnerable people live in the worst slums? A spatial analysis of Accra, Ghana. Annals of GIS. 2011 Dec 1;17(4):221–35.
- 30.
World Food Program. Food Consumption Score Nutritional Analysis (FCS-N) Guidelines. VAM Food security analysis. 2008 https://reliefweb.int/sites/reliefweb.int/files/resources/wfp277333.pdf
- 31. Coates J, Swindale A, Bilinsky P. Household Food Insecurity Access Scale (HFIAS) for measurement of food access: indicator guide: version 3.
- 32.
Inddex Project. Data4Diets: building blocks for diet-related food security analysis. USAID Advancing Nutrition. 2018 Jan 12.
- 33.
Organisation for Economic Co-operation and Development. Divided we stand: Why inequality keeps rising. Paris: OECD publishing; 2011 Dec 14.
- 34. McKenzie DJ. Measuring inequality with asset indicators. Journal of population economics. 2005 Jun;18:229–60.
- 35. Filmer D, Pritchett LH. Estimating wealth effects without expenditure data—or tears: an application to educational enrollments in states of India. Demography. 2001 Feb;38:115–32. pmid:11227840
- 36.
Collateral registry Ghana. 2020. https://www.collateralregistry.gov.gh/
- 37. World Food Program (WFP). Kenya Comprehensive Food Security and Vulnerability Analysis (CFSVA), June 2016
- 38. Battersby J, Watson V. Addressing food security in African cities. Nature Sustainability. 2018 Apr;1(4):153–5.
- 39.
Skinner C, Haysom G. The informal sector’s role in food security: A missing link in policy debates. Cape Town: PLAAS, UWC and Centre of Excellence in Food Security. 2016 Sep 12.
- 40. Skinner C. Challenging city imaginaries: Street traders’ struggles in Warwick Junction. Agenda. 2009 Jan 1;23(81):101–9.
- 41. Ligthelm AA. Informal retailing through home-based micro-enterprises: The role of spaza shops. Development Southern Africa. 2005 Jun 1;22(2):199–214.
- 42. Marras S, Bendech MA, Laar A. Street food vending in Accra, Ghana.
- 43. Evans A, Chow S, Jennings R, Dave J, Scoblick K, Sterba KR, Loyo J. Traditional foods and practices of Spanish-speaking Latina mothers influence the home food environment: implications for future interventions. Journal of the American Dietetic Association. 2011 Jul 1;111(7):1031–8. pmid:21703381
- 44.
Africa Center for Strategic Studies (ACSS). Conflict Remains the Dominant Driver of Africa’s Spiraling Food Crisis. 2022. https://africacenter.org/spotlight/conflict-remains-the-dominant-driver-of-africas-spiraling-food-crisis/
- 45. Hoddinott J, Yohannes Y. Dietary diversity as a food security indicator.
- 46. Blekking J, Waldman K, Tuholske C, Evans T. Formal/informal employment and urban food security in Sub-Saharan Africa. Applied Geography. 2020 Jan 1;114:102131.
- 47. Battersby J. The food desert as a concept and policy tool in African cities: an opportunity and a risk. Sustainability. 2019 Jan 16;11(2):458.
- 48.
SPAR. The history of SPAR. 2019. https://spar-international.com/aboutus/history/
- 49. Das Nair R. The internationalisation of supermarkets and the nature of competitive rivalry in retailing in southern Africa. Development Southern Africa. 2018 May 4;35(3):315–33.
- 50. Battersby J. and Muwowo F., 2018. Planning and governance of food systems in Kitwe, Zambia: A case study of food retail space. In Urban food systems governance and poverty in African cities (pp. 128–140). Routledge.
- 51. Debela BL, Demmler KM, Klasen S, Qaim M. Supermarket food purchases and child nutrition in Kenya. Global Food Security. 2020 Jun 1;25:100341.
- 52. Khonje MG, Ecker O, Qaim M. Effects of modern food retailers on adult and child diets and nutrition. Nutrients. 2020 Jun 8;12(6):1714. pmid:32521620
- 53. Tuholske C, Andam K, Blekking J, Evans T, Caylor K. Comparing measures of urban food security in Accra, Ghana. Food Security. 2020 Apr;12:417–31.
- 54. Crush J, Frayne B. Supermarket expansion and the informal food economy in Southern African cities: implications for urban food security. Journal of Southern African Studies. 2011 Dec 1;37(4):781–807.
- 55. Peyton S, Moseley W, Battersby J. Implications of supermarket expansion on urban food security in Cape Town, South Africa. African Geographical Review. 2015 Jan 2;34(1):36–54.
- 56. Berger M, van Helvoirt B. Ensuring food secure cities–Retail modernization and policy implications in Nairobi, Kenya. Food Policy. 2018 Aug 1;79:12–22.
- 57.
FAO. Street Food Vending in Accra, Ghana. Field Survey Report. 2016. www.fao.org/publications
- 58. Kwan MP. Feminist visualization: Re-envisioning GIS as a method in feminist geographic research. Annals of the association of American geographers. 2002 Dec 1;92(4):645–61.
- 59. Tshabalala S. South Africa’s Shoprite supermarket has the edge on Walmart’s Nigeria plans. 2015. https://qz.com/africa/482898/why-a-south-african-supermarket-chain-has-the-edge-on-walmart/
- 60.
Abdychev A, Alonso C, Alper ME, Desruelle MD, Kothari S, Liu Y, Perinet M, Rehman S, Schimmelpfennig MA, Sharma P. The future of work in sub-Saharan Africa. International Monetary Fund; 2018 Dec 17.