Article
Outlook on Agriculture
2018, Vol. 47(3) 233–243
The transformation of urban food ª The Author(s) 2018
Article reuse guidelines:
systems in Ghana: Findings from sagepub.com/journals-permissionsDOI: 10.1177/0030727018785918
journals.sagepub.com/home/oag
inventories of processed products
Kwaw S Andam1, David Tschirley2, Seth B Asante1,
Ramatu M Al-Hassan3 and Xinshen Diao4
Abstract
Urban food systems in Ghana are changing, along with rapid urbanization and growth in household incomes. Using data
from retail inventories of packaged products carried out in eight cities in 2015 and 2016, we find that the interplay of
urbanization, imports and domestic processing and packaging has led to some surprising outcomes. Imports are dominant,
especially for milled rice and tomato paste, and the shares are higher in smaller cities than in Accra. Imported products are
more prevalent in traditional retail outlets than in modern retail outlets. Moreover, imported products come mainly from
East Asia; excluding South Africa, which accounts for 6% of imports, less than 3% of imported products were from other
African countries.
Keywords
Supermarkets, food processing, imports, packaging, Ghana, sub-Saharan Africa
Introduction have been excluded from supply chains as consumers and
retailers have shifted to processed foods. Market data on
Developing countries are going through significant changes
processed foods can provide a starting point for policy
in their food systems (Reardon et al., 2009; Reardon, 2015;
analysis on this topic.
Regmi and Meade, 2013; Tschirley et al., 2015a). These
In many developing countries, however, such data typi-
changes are generating new market opportunities and pol-
cally are not available. Ministries of industry and parastatal
icy questions. Diet shifts and growing demand for pro-
agencies may track sales of some processed foods by
cessed foods can create opportunities for adding value to
selected state enterprises, and ministries of agriculture may
primary agricultural products. However, a transforming
collect information on the performance of development
food system also can lead to higher food import bills and
projects that support some aspects of food processing.
possible macroeconomic imbalances, and changing
However, data on processed food imports, domestic pro-
diets can create new public health risks such as non-
duction and retailing of processed food products typically
communicable diseases caused by over-nutrition (Rischke
are not recorded in official statistics. This lack of data on
et al., 2015). A new food system paradigm has further
the food processing sector is particularly noticeable for
implications for education and employment policies, as
African countries, while data have been available for
labour markets shift to respond to changing production
research on Asia (Reardon, 2015).
systems (Tschirley et al., 2015b).
As the food system changes, policymakers can benefit
from reliable data to help determine how to enable devel- 1 Development Strategy and Governance Division, International Food
oping economies to take advantage of new opportunities Policy Research Institute, PMB CT112, Cantonments, Accra, Ghana
2
and address new challenges posed to farmers, processors Department of Food, Agricultural, and Resource Economics, Michigan
and consumers (Timmer, 1997). For example, one of the State University, East Lansing, MI, USA3 Department of Agricultural Economics & Agribusiness, University of
policy issues raised by these shifts is the impact on small- Ghana, Accra, Ghana
holder farmers of a consumer shift towards processed 4 Development Strategy and Governance Division, International Food
foods. There is some evidence from China that some farm- Policy Research Institute, Washington, DC, USA
ers can benefit from such a transition (Wang et al., 2009).
Corresponding author:
In contrast, some smallholder farmers in Kenya (Neven
Kwaw S Andam, International Food Policy Research Institute, PMB
et al., 2009), South Africa (D’Haese and Van Huylen- CT112, Cantonments, Accra, Ghana.
broeck, 2005) and Vietnam (Mergenthaler et al., 2009) Email: k.andam@cgiar.org
234 Outlook on Agriculture 47(3)
Research on food system transformation in Africa has staples decreases (Minten and Reardon, 2008; Reardon
focused on eastern and southern Africa (D’Haese and Van et al., 2009, 2010, 2015). Households, especially in urban
Huylenbroeck, 2005; Hassen et al., 2016; Ijumba et al., areas, often demand more diverse and exotic foods, which
2015; Minten and Reardon, 2008; Minten et al., 2013; food processors typically supply through supermarkets
Minten et al., 2016; Tschirley et al., 2015a). Until recently, (Reardon et al., 2010). As a result, modern retail systems
relatively little attention has been paid to West Africa (Hol- and supermarkets expand as part of the transformation of
linger and Staatz, 2015; Liverpool Tasie et al., 2016; Ther- the food system (Minten et al., 2013; Reardon et al., 2010;
iault et al., 2018). West Africa is a suitable setting for this Timmer, 2004,2009). Supermarkets serve as middle agents,
type of research because it is experiencing important trends collecting processed food products from around the world
such as rising food imports (Zhou and Staatz, 2016) and and making them available and accessible to urban consu-
rapid urbanization; it is home to highly integrated cross- mers. Supermarkets also can as act as vehicles for increas-
border markets for agricultural inputs and food staples; and ing consumer welfare by providing reduced prices, broader
it has a broader diversity of staple diets compared to the rest sources of diet diversity (introducing consumers to new
of Africa. products) and better food safety (Reardon et al., 2010).
This article contributes to the emerging literature on Supermarkets can lead to higher consumption of processed
food system transformation in Africa by describing pro- foods but, adversely, they may be associated with higher
cessed food retail in urban Ghana. It examines the pace at rates of adult obesity (Rischke et al., 2015).
which supermarkets are opening in Ghana’s cities, mea- Policymakers in a transforming food system often face
sures the penetration of imports and compares domestic questions specific to a country’s food needs. For example,
and imported food products in terms of their availability in the diversity of a country’s traditional diet may affect the
different types of retail outlets and their product packa- pace of the transformation and the associated policy chal-
ging and labelling. Although the transformation of Gha- lenges. Minten et al. (2013) and Hassen et al. (2016) have
na’s urban food system has been recognized (Hollinger documented how changes in Ethiopia’s food system were
and Staatz, 2015), the product inventory approach to driven in part by public investments to develop modern
assessing food availability is novel in the Ghanaian con- varieties of teff, the main staple cereal. However, Ethio-
text. As a rapidly urbanizing, lower middle-income coun- pia’s experience may not be easily transferrable when it
try, Ghana has conditions that are favourable for demand comes to setting Ghana’s public research priorities in crop
growth for processed foods. In recent years, the country’s technologies. Instead of one dominant staple in rural and
urban population has grown to 54% of the total population urban diets, Ghana’s traditional diets are based on a variety
(United Nations, 2014), and incomes have been rising of staple crops including cassava, maize, plantain, rice and
steadily. However, to date, the implications of these yams. Ghanaian diets also are shifting across these staples
changes for food retail in Ghana’s cities have not be as people move to cities (Hollinger and Staatz, 2015). To
assessed. grapple with the new challenges of the changing food sys-
To obtain data on the marketing channels and the types tem, specific research is needed on each country’s food
of food products in the cities, we collected information on system, pre-existing diet patterns and the capacity of its
retail outlets and processed foods for sale in eight cities, food processing sector.
following an inventory approach used recently in Mali
(Theriault et al., 2018), Nigeria (Liverpool Tasie et al.,
2016) and Tanzania (Ijumba et al., 2015; Snyder et al., The Ghanaian context: A quest for a
2015). The inventories focused on specific products under competitive domestic food processing
the broad categories of processed meats; fruits and vege-
sector
tables; grains; and starches. As in most African countries,
Ghana’s governments have sought to build a food process- Even before the country’s recent trends in urbanization,
ing sector to add value to raw outputs and to compete economic policy debates in Ghana included concerns about
against imports. Accordingly, as described in detail in the processed food imports. In the 1960s, Ghana’s post-
methods section below, the products selected for data col- independence industrialization program established state-
lection have high demand among urban consumers and are run enterprises to promote domestic food processing of
produced using crops and livestock that have been focus of selected commodities (Ackah et al., 2014). These policies
recent policy strategies to improve the competitiveness of are considered to have failed, owing to inefficiencies with
Ghana’s agriculture. state intervention, and recent policies recognize the need
for private-sector involvement.
Transformation of urban food systems The policy approach has not changed entirely since the
earlier programs were abandoned, however, and the state
in developing countries
continues to support specific industries such as tomato and
Some common trends have been associated with transform- cassava processing. Nevertheless, the range of processed
ing food systems in developing countries, such as increas- foods has expanded, and several private processors cur-
ing availability of processed foods, diet shifts and the rently produce maize, rice, cassava, yam, plantain, toma-
realization of Bennett’s Law, which holds that as household toes, oil palm, groundnut, cocoa, fruits and fish products. In
incomes increase, the share of food expenditure on food spite of these changes, the agricultural sector still struggles
Andam et al. 235
Cape Coast, Ho, Koforidua, Kumasi, Sekondi-Takoradi,
Sunyani and Tamale from May to July 2016 (Figure 1).
These study sites include the four most populous urban
areas in Ghana and represent different aspects of the eco-
nomic geography of the country (Table 1). Accra, Ghana’s
administrative and commercial capital, is one of Africa’s
emerging megacities and Kumasi is the commercial centre
of the country’s middle belt. Sekondi-Takoradi is the sec-
ond major coastal city and the centre of Ghana’s new oil
industry, whereas Tamale is the urban centre of northern
Ghana. The other four cities are regional capitals. The
populations in the surveyed cities represent 44% of Gha-
na’s total urban population, as defined by the Ghana Sta-
tistical Service (2012, 2014) and 78% of the country’s
urban population in large, medium-sized towns with popu-
lations greater than 50,000 (Table 1).
Retail inventory
We identified six types of processed food retail outlets: (1)
Figure 1. Map of Ghana showing study sites. Source: Adminis- stalls in open-air markets; (2) street-side vendors, operating
trative boundaries based on map from Wikimedia Commons in wood or metal sheds outside markets; (3) traditional,
(https://upload.wikimedia.org/wikipedia/commons/7/7b/Ghana_ non-self-service shops in concrete structures; (4) self-
location_map.svg).
service grocery stores with one register; (5) single-
enterprise, self-service supermarkets with two or more
to achieve the productivity levels required to provide pro- registers; and (6) self-service supermarkets with two or
cessors with adequate, reliable supplies of raw material more registers that are part of a store chain (labelled ‘chain
(see, e.g. the case of tomato processing in a study by Robin- supermarkets’ in tables below). These categories include
son and Kolavalli (2010)). all types of food traders operating from fixed structures in
Although there are few reliable official statistics on the Ghanaian cities. We excluded shops whose primary busi-
performance of the food processing sector, there is evi- ness is in non-food products and transient street traders. We
dence that the broader manufacturing sector has declined identified and listed retail outlets through pre-inventory
over the years. The contribution of industry to Ghana’s exploratory visits, using city maps and key informants in
economy has increased, but the contribution of manufac- each city.
turing to total industrial gross domestic product (GDP) fell The inventory included packaged foods with any level
from 70% in the 1980s to less than 10% in the mid-2000s. of post-harvest processing. Thus, the range of products
These trends, combined with rising food imports across extended from minimally processed items, such as milled
West Africa (Hollinger and Staatz, 2015; Zhou and Staatz, grain, to products with more sophisticated processing and
2016), have led to continued attention from Ghanaian pol- packaging, such as canned tomatoes. We excluded pre-
icymakers seeking ways to improve the domestic food pro- pared meals consumed away from home. The inventoried
cessing sector to generate employment and to serve as a food products were selected from three processed food
market for agricultural produce. groups within the United Nations Inventory of Classifica-
Policymakers are likely to maintain their keen interest tions (ISIC Revision 4, Code 10, http://unstats.un.org/unsd/
in the performance of the sector, given the ongoing demo- cr/registry): processing and preserving of meat; processing
graphic changes that will have implications for the coun- and preserving of fruits and vegetables; and manufacture of
try’s food system. From 2000 to 2015, the urban grain mill products, starches and starch products.
population growth rate averaged around 3.9%. The urban Within each food group, we selected products for inven-
population surpassed the rural population in 2009 and now tory based on Ghana’s policy and market context (Table 2).
stands at 54% of the total population. Ghana’s population For example, packaged chicken meat and eggs are included
also has become wealthier – per capita nominal GDP because of the keen attention paid to the poultry industry’s
increased from about US$1100 in 2009 to more than performance against frozen meat imports, and processed
US$1400 in 2014. tomatoes are included because tomato processing has been
a focus of Ghana’s industrial policy. The milled grains,
Data and methods roots and tubers category includes milled rice, maize flour
and other product types that are specific to West Africa
Study sites
such as gari and fufu flours (Table 2).
This study conducted inventories of processed foods in all It is important to distinguish between product type and
the major urban centres in Ghana. The inventories were product. We use the definitions proposed by Tschirley et al.
completed in Accra in 2015 (Andam et al., 2015) and in (2016), in which product type (e.g. chopped tomatoes) is
236 Outlook on Agriculture 47(3)
Table 1. City populations (2010, 2016), population growth rates (1970–1984, 2000–2010) and poverty levels (2012/2013).
Population Estimated Average population Average population Poverty level
City (2010) population (2016)a growth rate (1970–1984) growth rate (2000–2010) (2012/2013)b
Accra 2,591,970 2,982,107 3.0 2.2 2.6
Cape Coast 108,374 188,099 1.1 7.5 2.6
Ho 104,532 120,124 3.2 5.4 21.9
Koforidua 122,300 139,138 1.7 3.3 4.6
Kumasi 1,730,249 1,956,767 2.6 5.7 5.3
Sekondi-Takoradi 559,548 679,901 1.8 6.5 12.9
Sunyani 75,366 85,473 3.6 1.8 19.4
Tamale 233,252 268,937 4.9 6.3 24.6
Total population of study sites 5,525,591 6,420,546 – – –
Total national urban populationc 12,545,229 – – – –
National population 24,658,823 28,308,301 2.6 2.5 24.2
Source: Derived from Ghana Statistical Service (2012, 2014, 2015).
aAuthors’ estimates based on population projections for administrative regions.
bPoverty level is defined as share of population living below US$1.83 per day.
cThe Ghana Statistical Service classifies any towns with populations greater than 5000 as urban localities.
Table 2. Product types included in inventory. For Accra, we conducted a city-level inventory (Tschirley
et al., 2016) to collect information on all food products
Processed foods
regardless of the retail outlet type in which the product was
selected for
ISICs category inventory Examples of product types sold. For the other seven cities, we conducted a city-outlet
inventory (Tschirley et al., 2016) to make exhaustive lists of
Group 101: Packaged chicken  Frozen chicken, cut products within each product type that were available for
Processing and meat and eggs parts
 sale within each type of retail outlet, to allow for compar-preserving of Frozen chicken, whole
meat  Packaged eggs isons across outlet types. For example, for milled rice, we
Group 103: Processed  Chopped tomatoes recorded all the milled rice products (brands) found in open-
Processing and tomatoes  Tomato paste air stalls and all the rice products found in traditional shops –
preserving of  Tomato puree even if some of these products were the same as the products
fruits and inventoried in open-air stalls – and so on for the other types
vegetables
 of retail outlets. Owing to the different inventory approachGroup 106: Milled grains, Milled rice
Manufacture of roots and  Fufua flour used in Accra versus the other cities, we exclude the Accra
grain mill tubers  Garib data from the comparisons of retail outlets below.
products,  Maize cereal The city-outlet inventories started from one type of
starches and  Plantain chips retail outlet and continued within that retail outlet type
starch products until few or no new products could be found within that
Source: Processed food categories based on United Nations Inventory of outlet. For packaged chicken meat and eggs, the enu-
Classifications (ISIC Revision 4, Code 10, http://unstats.un.org/unsd/cr/ merators could not identify any new products in any of
registry) and processed foods and product types based on authors’ the cities by the end of the survey. For processed toma-
selection.
toes, at the close of the inventory enumerators could
ICS: inventory of classification.
aFufu is a starchy staple made from boiled and pounded tubers such as identify less than one new product in every three retail
cassava, plantain and yam. Processed fufu flour consists of smooth outlets surveyed. For milled grains, roots and tubers, the
milled tubers that can be boiled and used to prepare fufu without broadest food category, enumerators encountered on
pounding.
bGari is a ready-to-eat meal made from grated, fermented and fried cassava. average less than one new product per store visit by the
end of the inventory.
Inventories started from commercial centres, where
derived from the intersection of the processed food group food retail activities are concentrated, and moved on to
(processing and preserving of fruits and vegetables), the other parts of the city when few new products were encoun-
raw materials used (tomatoes) and the type of processing tered, usually after 2 or 3 days of inventory. As the cities
(chopping), whereas product refers to a particular brand of had relatively few supermarkets, the survey included all
a product type manufactured by a company (following the food supermarkets. The team visited 98 retail outlets in
example above, company X’s chopped tomatoes). In each Accra in 2015. In 2016, the enumerators visited 545 retail
retail outlet, enumerators, with assistance from shop atten- outlets in the other seven cities, including 196 open-air
dants, identified the range of products; took pictures of stalls, 49 street-side vendors, 206 traditional non-
each product, taking care to capture labelling information self-service shops, 66 self-service grocery stores, 17
and packaging; and then entered information from the pic- single-enterprise supermarkets and 11 chain supermarkets
tures into a database. (Table 3). Ten retail outlets declined the survey request.
Andam et al. 237
Table 3. Number of surveyed retail outlets, by city.
Retail outlet typea
Stalls in
open-air Street-side Traditional Self-service Single-enterprise Chain
City markets vendors non-self-service shops grocery stores supermarkets supermarkets Total
Cape Coast 24 9 15 10 1 1 60
Ho 25 6 34 11 2 2 80
Koforiduab 19 2 13 3 0 1 38
Kumasi 72 15 51 8 6 3 155
Sekondi-Takoradi 39 8 25 14 5 2 93
Sunyanib 13 5 23 5 0 1 47
Tamale 4 4 45 15 3 1 72
Totalc 196 49 206 66 17 11 545
Share (%) of outlets surveyed 36 9 38 12 3 2 100
Source: Inventory data (2016).
aThe Accra retail inventory, which was conducted a year earlier than the inventories in the other seven cities, used the following classifications for the
retail outlet types (Andam et al., 2015): (1) open-air market stalls; (2) kiosks (street-side vendors in wood structures); (3) container shops (street-side
vendors in recycled shipping containers); (4) small shops, corresponding to ‘traditional shops’ in this artilce’s categories; (5) local supermarket chains
(single-enterprise supermarkets); and (6) international supermarket chains. Since the categories for the Accra data are only roughly comparable to
those used in 2016 for the other cities, this paper avoids direct comparisons of the data from the two inventories.
bIn Koforidua and Sunyani, there were no single-enterprise supermarkets.
cTen retail outlets declined the survey request.
Results and discussion Table 4. Number of inventoried products, by city.
Pace of food retail modernization Number of unique
City Number of products products
Contrary to the expectation that a supermarket boom will
accompany rapid urbanization and per capita income Accra 1779 539
growth, the team found modest supermarket growth in all Cape Coast 328 172
eight cities, with around one supermarket for every 60,000 Ho 384 208
residents. The inventory data included the year in which Koforidua 208 122
Kumasi 607 379
each retail outlet started operations. As noted earlier, all the
Sekondi-Takoradi 538 282
surveys included all supermarkets operating in the cities, Sunyani 280 175
and therefore the data allow for a description of the pene- Tamale 298 173
tration of modern retail in each city.
Accra has seen the most growth. Before 2005, the city Source: Inventory data (2015, 2016).
Column 2 records all products found in all retail outlet types and
had only three supermarkets; three more opened there therefore includes duplicates in cases where products were encountered
from 2005 to 2010, and an additional 10 have started in more than one type of outlet. Column 3 excludes the duplicates.
operating in the city since 2010. These include the well-
known South African supermarket chains Game and
Shoprite and local supermarket chains such as Koala, column 2). As described earlier, these inventories recorded
MaxMart and Shop and Save. In Kumasi, the first 10 all products within each type of retail outlet. Therefore,
supermarkets opened in the past 20 years. Sekondi- some products may have been counted more than once if
Takoradi has seven supermarkets. Tamale experienced a found in more than one retail outlet. The second count
brief spurt in modern retailing between 2001 and 2006, eliminates these duplicates in order to calculate the number
when four supermarkets were established, but there have of unique products in each city. The availability of pro-
been no supermarket openings since then. The four cessed foods corresponded with the size of city. The enu-
smaller cities have had eight supermarkets opening in the merators recorded the highest number of unique products,
past 15 years. Instead of a rapid spread of supermarkets, 539, in Accra, followed by 379 in Kumasi and 282 in
the shift to processed foods is currently occurring largely Sekondi-Takoradi (Table 4, column 3).
through traditional retail outlets. In the tables below, we use the two different product
counts for different purposes. For example, we use the
unique product count to compare shares of domestic and
Availability of processed food products imported products in each city and to compare types of
The shops have a wide range of processed products. Table 4 product packaging. This is the appropriate measure because
presents two different counts of product availability. The the comparisons focus on the characteristics of the unique
first counts all products recorded during the inventory. By product, such as whether it was manufactured domestically
this count, there were 1779 products in Accra alone and or imported. In contrast, when we compare findings across
hundreds of products in each of the other cities (Table 4, retail outlet types (for instance, availability of imported
238 Outlook on Agriculture 47(3)
100
90
80
70 86
77 80
60 71 76 7371
50 63
40
30
20
10
0
2.98 million 1.96 million 0.68 million 0.27 million 0.19 million 0.14 million 0.12 million 0.09 million
CITY AND POPULATION
Figure 2. Shares (%) of imported products, by city population. Source: Import shares derived from inventory data (2015, 2016) and city
populations based on Ghana Statistical Service (2014) data.
products across retail outlet types), we maintain the overall Table 5. Share and number of imported products, by food
count of products within each outlet type. product category and by city.
Packaged Milled grains,
Import penetration poultry meat Processed roots and
and eggs tomatoes tubers
Imports dominate the processed foods sector in all eight
cities. Import shares range from 63% to 86% of inventoried Share Share Share
products, depending on the city (Figure 2). In terms of man- City (%) Number (%) Number (%) Number
ufacturer location, 71–83% of processing firms are located Accra 53 10 87 75 58 252
outside Ghana. Import shares do not appear to be higher in Cape Coast 100 6 95 42 69 84
larger cities. Surprisingly, although Accra, as the principal Ho 100 2 97 60 62 89
and most-populous city, would be expected to have higher Koforidua 0 0 96 44 80 61
shares of imports than the other cities, we find the opposite Kumasi 71 5 97 97 61 166
result. In Accra, 63% of the products are imported, whereas Sekondi- 89 8 94 75 61 117
Takoradi
Koforidua, one of the smallest cities, has the highest share of
Sunyani 0 0 91 42 77 98
imported products at 86%. Nor does import penetration seem Tamale 100 1 98 47 68 84
to vary by distance from ports. In Tamale, the city farthest
from coastal ports, 76% of the products are imported, Source: Inventory data (2015, 2016).
whereas the coastal cities Cape Coast and Sekondi-
Takoradi have import shares of 77% and 71%, respectively.
There could be two possible explanations for this unex-
Table 6. Main sources of imported milled rice and tomato paste.
pected finding. First, domestic food manufacturers, in the
initial stages of competition against imports, may be target- Processed food Number of Share of
ing the larger market in Accra, leading to lower shares of group Source of imports products imports (%)
imported products there. Second, retailers in smaller cities
Milled rice Vietnam 280 19.1
may be selling domestic processed foods in unpackaged
Thailand 225 15.4
form while retailers in larger cities carry packaged, domes- India 46 3.1
tic products. We found anecdotal evidence of the latter United Arab Emirates 17 1.2
during the surveys for processed grains such as milled rice United States 12 0.8
and maize. Future research on unpackaged processed Tomato paste China 223 15.2
foods, an area beyond the scope of this study’s inventories, Italy 45 3.3
would provide further clarity on this issue. United States 29 2.0
Hong Kong 7 0.5
Import dominance is clearly visible in the poultry and
United Arab Emirates 5 0.3
processed tomato categories (Table 5). For example, for
tomato pastes, which make up 76% of processed tomato Source: Inventory data (2015, 2016).
IMPORT SHARES OF FOOD PRODUCTS (%)
Accra
Kumasi
Sekondi-Takoradi
Tamale
Cape Coast
Koforidua
Ho
Sunyani
Andam et al. 239
Table 7. Manufacturer location for products manufactured in Ghana, by city.
Percentage share and number (in parenthesis) of products
Manufacturer location Accra Cape Coast Ho Koforidua Kumasi Sekondi-Takoradi Sunyani Tamale
Accra/Tema 95(191) 91(58) 63(60) 87(26) 50(74) 54(65) 73(43) 53(28)
Cape Coast 1(1) 0 0 0 0 0 0 0
Ho 1(1) 0 10(9) 0 0 0 0 0
Koforidua 0 0 0 3(1) 0 0 0 0
Kumasi 1(2) 2(1) 6(6) 0 36(53) 8(9) 5(3) 4(2)
Sekondi- Takoradi 0 0 0 0 7(11) 23(28) 2(1) 2(1)
Sunyani 0 2(1) 0 0 1(1) 0 12(7) 0
Tamale 0 0 0 0 0 0 0 26(14)
Other locations
Southern Ghana 0 2(2) 0 0 4(5) 3(4) 2(1) 4(2)
Middle Ghana 1(1) 2(2) 6(6) 0 3(4) 0 0 2(1)
Northern Ghana 1(1) 0 0 0 0 1(1) 0 0
Source: Inventory data (2015, 2016).
Table 8. Share (%) and number of imported products, by city and by retail outlet type.
Stalls in open-air Street-side Traditional non-self Self-service Single-enterprise Chain
markets vendors service shops grocery stores supermarkets supermarkets
City % No. % No. % No. % No. % No. % No.
Cape Coast 81 58 84 28 85 51 78 68 80 41 72 18
Ho 83 71 75 9 79 92 74 55 61 31 71 32
Koforiduaa 93 53 75 3 84 65 92 37 NA NA 67 20
Kumasi 90 162 85 35 90 104 56 25 54 70 66 63
Sekondi-Takoradi 88 91 83 29 81 84 75 83 74 100 62 31
Sunyania 90 53 82 9 80 90 70 39 NA NA 71 30
Tamale 100 19 100 15 90 79 72 60 78 50 76 22
Average 89 72 83 18 84 81 74 52 69 58 69 31
Source: Inventory data (2015, 2016).
a In Koforidua and Sunyani, there were no single-enterprise supermarkets.
products, import shares range from 83% to 98% (not shown In each city, around 50% of domestic products in the milled
in Table 5). Milled grains, roots and tubers – the products grains, roots and tubers category come from Accra, and
with the least value added after the farm – show the lowest about 25% come from the city in which the product was
import shares from 58% in Accra to 80% in Koforidua. For found (Table 7). Processing firms in Accra sell their prod-
milled rice, which accounts for about 61% of the product in ucts in other cities but there seems to be limited transport of
this category, import shares range from 79% to 98%. For processed food products manufactured outside Accra to
non-rice products in the milled grains category, import other cities in Ghana.
shares are much lower (35% for Accra and less than 10%
for Kumasi, Sekondi-Takoradi and Tamale).
Imports come mainly from East Asia (Table 6). China, Are imports more prevalent in modern retail outlets?
Thailand and Vietnam account for two-thirds of imports. One would expect to find higher shares of imported prod-
Remarkably, we do not find much evidence of imports ucts in modern retail outlets than in traditional outlets
from neighbouring West African countries or indeed from because supermarkets tend to cater towards wealthier cus-
other African countries. In Accra, excluding imports from tomers (Traill, 2006) and because it tends to be difficult for
South Africa (representing 6% of imports), only 3% of domestic processors to enter into modern supply chains
imports come from African countries. Products imported (D’Haese and Van Huylenbroeck, 2005). However, in our
from other countries in Africa are nearly non-existent in the inventories, this is not the case. On average, 83–89% of
other seven cities (less than 3% of imported products in products in traditional outlets are imported, compared with
Kumasi, Sekondi-Takoradi and Tamale were from other an average of 69–74% of products in modern retail outlets
sub-Saharan African countries). In contrast, a recent inven- (Table 8). In making this assessment, we categorize tradi-
tory in Tanzania recorded high shares of imports from other tional and modern retail outlets based on whether the outlet
sub-Saharan African countries (Snyder et al., 2015). is self-service or not. Accordingly, we classify open-air
Domestic processing is done primarily in Accra, fol- markets, street-side vendors and traditional shops as tradi-
lowed by the city in which the product is sold (Table 7). tional retail outlets and the other three categories – grocery
240 Outlook on Agriculture 47(3)
Table 9. Shares of imported products in modern and traditional retail outlets, by city.
Imported products in modern retail outlets Imported products in traditional retail outlets
Difference in shares
City Number Share of products (%) Number Share of products (%) (percentage points)
Cape Coast 127 78 137 83 5
Ho 118 69 172 80 11**
Koforidua 57 81 121 88 6
Kumasi 158 59 301 89 30***
Sekondi-Takoradi 241 72 204 84 12***
Sunyani 69 70 152 83 13**
Tamale 132 75 113 93 18***
Source: Inventory data (2015, 2016).
Last column reports difference in share (%) of imported products in modern outlets and share (%) of imported products in traditional outlets for each
city and significance levels of the difference using two-sample tests of proportions. Significance levels are denoted as follows:
*p < 0.10.
**p < 0.05.
***p < 0.01.
Table 10. Share (%) and number of imported milled grains, roots and tubers, by city and by retail outlet type.
Stalls in open-air Street-side Traditional non–self Self-service Single-enterprise Chain
markets vendors service shops grocery stores supermarkets supermarkets
City % No. % No. % No. % No. % No. % No.
Cape Coast 68 28 80 17 85 41 73 45 74 26 67 12
Ho 64 23 40 2 71 58 66 35 54 22 64 21
Koforiduaa 91 21 0 0 81 42 90 28 NA NA 55 11
Kumasi 87 112 83 25 86 62 46 17 41 41 50 28
Sekondi-Takoradi 87 55 81 17 73 47 64 47 63 54 53 17
Sunyania 94 29 100 3 75 61 65 30 NA NA 71 22
Tamale 100 5 100 4 85 45 64 41 70 32 63 12
Average 84 39 69 10 79 51 67 35 60 35 60 18
Source: Inventory data (2015, 2016).
a In Koforidua and Sunyani, there were no single-enterprise supermarkets.
stores, single-enterprise supermarkets and chain supermar- attacked by pests during transportation, handling and stor-
kets – as modern retail outlets. age. We classify hard and soft packaging, where hard packa-
We combine the product counts for retail outlet types ging comprises products in tin cans and glass or plastic
within each of the two categories and use simple two- bottles and jars, whereas soft packaging includes all others
sample tests of proportions to determine whether imported such as plastic bags, polypropylene sacks, aluminium foils,
products are more commonly found in modern retail outlets jute sacks, paper bags and paper boxes. Second, we compare
than in traditional outlets in the cities (excluding Accra for products in terms of the presence of a bar code for tracking.
reasons explained earlier). In the seven cities, the shares of Third, we compare products in terms of the presence of
imported products in modern outlets are higher than the nutritional information. Consumer surveys in Accra have
shares of imported products in traditional outlets, and the suggested that consumers care about packaging as a sign
differences are significant in five of the seven cities (Table of product quality (Hollinger and Staatz, 2015). We assess
9). The situation is even starker when we compare shares of the extent to which food manufacturers are responding to the
domestic and imported products in the milled grains, roots demand for high-quality packaging. In Table 11, we report
and tubers category alone. For products in the latter cate- the shares of imported and domestic products with each of
gory, import shares in the traditional outlets are 69–84% the packaging attributes listed above and compare the dif-
compared with import shares of 60–67% in the modern ferences using a simple two-sample test of proportions.
outlets (Table 10). We find that domestic and imported products have similar
shares of hard packaging. In two cities, Koforidua and
Tamale, the share of domestic products with hard packaging
Do imported products have better packaging? is higher than the share of imported products with hard packa-
We compare domestic and imported products in terms of ging. In the other five cities, there are no significant different
three packaging and labelling attributes. First, we compare differences between domestic and imported products in terms
the type of packaging, which determines whether a packaged of packaging type. In four of the seven cities, there are no
product could be opened (intentionally or accidentally) or significant differences in the shares of domestic and imported
Andam et al. 241
Table 11. Shares of products with attributes: hard packaging, supermarket growth in Ghana’s cities has been modest to
presence of bar code and presence of nutrition information date. Second, imports dominate the processed foods sector.
by city. Notably, the highest shares of imported products are found
Share (%) of products with in processed tomatoes, a sector where the government has
packaging attribute promoted domestic processing for decades (Robinson and
Kolavalli, 2010). Third, few products within the food cate-
Hard packaging (Tin can/glass/hard plastic) gories in this study were imported from other African coun-
Imported Domestic Significant tries. For policymakers and development agencies
City products products difference? promoting intra-regional trade, this may be a cause for
concern. Fourth, surprisingly high percentages of products
Cape Coast 28 25 No in traditional retail outlets were imported, compared to the
Ho 40 33 No shares of imported products in modern retail outlets.
Koforidua 41 71 Yes**
Kumasi 34 29 No Imported products hold a higher share in smaller cities,
Sekondi-Takoradi 37 41 No even those far from ports, than in larger cities. The two
Sunyani 29 29 No large coastal cities, Accra and Sekondi-Takoradi, do not
Tamale 36 51 Yes* show meaningfully higher import shares than the other
cities. Finally, local processing takes place mainly in
Bar codes
Accra, followed by the city in which the product is sold.
Imported Domestic Significant This study adds to the emerging picture of the transfor-
City products products difference? mation of the urban food system across Africa. Our findings
confirm the prevalence of imported food products in West
Cape Coast 73 63 No
Africa, as shown by similar inventories conducted recently
Ho 72 46 Yes***
Koforidua 78 88 No (Liverpool Tasie et al., 2016; Theriault et al., 2018). How-
Kumasi 71 68 No ever, unlike the Mali case, where Theriault et al. (2018)
Sekondi-Takoradi 77 63 Yes** found no international supermarket chains and only a few
Sunyani 71 66 No local supermarkets in the capital Bamako, we see penetration
Tamale 66 56 No of international and local chain supermarkets across all cities
in Ghana. Although inventories in Tanzania showed locally
Nutrition information
manufactured grain products dominating imports (Ijumba
Imported Domestic Significant et al., 2015), the opposite holds true in Ghana (this study),
City products products difference? Mali (Theriault et al., 2018) and Nigeria (Liverpool Tasie
et al., 2016). Urban food markets across sub-Saharan Africa
Cape Coast 59 63 No
Ho 45 42 No are integrating with a range of producing regions. Whereas
Koforidua 45 65 No the bulk of rice and tomato products available in Ghana’s
Kumasi 53 38 Yes*** cities are imported from East Asia, dairy and grain products
Sekondi-Takoradi 59 52 No in Mali are imported mainly from Europe (Theriault et al.,
Sunyani 58 43 No 2018) and grain imports in Tanzania are mainly from other
Tamale 44 54 No eastern and southern African countries (Ijumba et al., 2015;
Source: Inventory data (2015, 2016). Snyder et al., 2015).
In the last column, we test the null hypothesis that shares of imported and Future research on urban food systems in Africa will
domestic products with the attribute are equal using a simple two-sample require comprehensive surveys and analyses to improve
test of proportions. Significance levels are denoted as follows: *p < 0.10;
**p < 0.05; ***p < 0.01. our understanding of farmer–processor–retailer–consumer
linkages. Policymakers in Ghana and West Africa would
benefit from research on the potential for local producers to
products with the other two packaging attributes (presence of be linked to urban food systems and the drivers of consu-
bar codes and presence of nutrition information). Bar codes mers’ preferences for processed food products, which
are present on a higher share of imported products than would signal future demand trends. This study provides a
domestic products in two cities (Ho and Sekondi-Takoradi), step in this direction by describing the urban food retail
and in one city (Kumasi), a higher share of imported products sector in Ghana, and comparing the sources, availability
than domestic products have nutrition information on the and packaging of domestic and imported food products.
packages. Taken together, these results suggest that domestic The article’s findings suggest that, beyond the expected
products are fairly matched against imported products in spread of modern retail systems and changes in food
terms of packaging and labelling attributes. sources (Timmer, 2009), food system transformation will
also lead to unexpected and counter-intuitive outcomes in
urban food retail in developing countries.
Conclusion
In this article, the inventories of processed food products on Authors’ note
sale in eight cities in Ghana have highlighted the following The opinions expressed here belong to the authors and do not
four aspects of urban food system transformation. First, necessarily reflect those of PIM, IFPRI, CGIAR or USAID.
242 Outlook on Agriculture 47(3)
Acknowledgements Liverpool Tasie S, Omonona B, Ogunleye W, et al. (2016) The
The authors acknowledge financial support from United States presence of processed foods in sub-Saharan Africa: Where
Agency for International Development (USAID) through funding from and where to? Evidence from Nigeria. Department of
for the Ghana Strategy Support Program (GSSP) and the CGIAR Agricultural, Food, and Resource Economics working paper.
Research Program on Policies, Institutions and Markets (PIM), East Lansing: Michigan State University.
which is led by the International Food Policy Research Institute Mergenthaler M, Weinberger K and Qaim M (2009) The food
(IFPRI). The authors thank an anonymous referee and the editor system transformation in developing countries: a disaggregate
for helpful comments. Any errors remain those of the authors. demand analysis for fruits and vegetables in Vietnam. Food
Policy 34(5): 426–436.
Declaration of conflicting interests Minten B and Reardon T (2008) Food prices, quality, and qual-
The author(s) declared no potential conflicts of interest with ity’s pricing in supermarkets versus traditional markets in
respect to the research, authorship, and/or publication of this developing countries. Review of Agricultural Economics
article. 30(3): 480–490.
Minten B, Tamru S, Engida E, et al. (2013) Ethiopia’s value
Funding chains on the move: the case of teff. IFPRI ESSP Working
The author(s) disclosed receipt of the following financial support Paper No. 52. Addis Ababa, Ethiopia: IFPRI.
for the research, authorship, and/or publication of this article: This Minten B, Assefa TW, Abebe G, et al. (2016) Food processing,
work was financially supported by United States Agency for transformation, and job creation: the case of ethiopia’s enjera
International Development (USAID) through funding for the markets. IFPRI ESSP Working Paper No. 96. Addis Ababa,
Ghana Strategy Support Program (GSSP) and the CGIAR Ethiopia: IFPRI.
Research Program on Policies, Institutions and Markets (PIM). Neven D, Odera MM, Reardon T, et al. (2009) Kenyan super-
markets, emerging middle-class horticultural farmers, and
References employment impacts on the rural poor. World Development
Ackah C, Adjasi C and Turkson F (2014) Scoping study on the 37(11): 1802–1811.
evolution of industry in Ghana. United Nations University, Reardon T (2015) The hidden middle: the quiet revolution in the
World Institute for Development Economics Research midstream of agrifood value chains in developing countries.
(UNU-WIDER) Working Paper 2014/075. Helsinki, Finland: Oxford Review of Economic Policy 31(1): 45–63.
UNU-WIDER. Reardon T, Barrett CB, Berdegué JA, et al. (2009) Agrifood
Andam KS, Al-Hassan RM, Asante SB, et al. (2015) Is Ghana industry transformation and small farmers in developing coun-
making progress in agro-processing? Evidence from an inven- tries. World Development 37(11): 1717–1727.
tory of processed food products in retail shops in Accra. Inter- Reardon T, Henson S and Gulati A (2010) Links between super-
national Food Policy Research Institute (IFPRI), Ghana markets and food prices, diet diversity and food safety in
Strategy Support Program (GSSP) Working Paper No. 41. developing countries. In: Hawkes C, Blouin C, Henson S,
Accra, Ghana: IFPRI. et al. (eds) Trade, Food, Diet and Health: Perspectives and
D’Haese M and Van Huylenbroeck G (2005) The rise of super- Policy Options. Oxford: Wiley-Blackwell, pp. 111–130.
markets and changing expenditure patterns of poor rural Reardon T, Tschirley D, Minten B, et al. (2015) Transformation of
households case study in the Transkei area, South Africa. Food African agrifood systems in the new era of rapid urbanization
Policy 30(1): 97–113. and the emergence of a middle class. In: ReSAKSS Annual
Ghana Statistical Service (2012) 2010 Population and Housing Conference, Beyond a Middle Income Africa, Trends and Out-
Census: Final Results. Accra: Ghana Statistical Service. look Report Conference Held in Addis Ababa, 1–3 September
Ghana Statistical Service (2014) 2010 Population and Housing 2015, pp. 1–16.
Census: District Analytical Report. Accra: Ghana Statistical Regmi A and Meade B (2013) Demand side drivers of global food
Service. security. Global Food Security 2(3): 166–171.
Ghana Statistical Service (2015) Ghana Poverty Mapping Report. Rischke R, Kimenju SC, Klasen S, et al. (2015) Supermarkets and
Accra: Ghana Statistical Service. food consumption patterns: the case of small towns in Kenya.
Hassen IW, Dereje M, Minten B, et al. (2016) Diet transformation Food Policy 52: 9–21.
in Africa: the case of ethiopia. IFPRI, Ethiopia Strategy Robinson EJZ and Kolavalli S (2010) The case of tomato in
Support Program (ESSP) Working Paper No. 87. Addis Ghana: Processing. IFPRI GSSP Working Paper No. 21.
Ababa, Ethiopia: IFPRI. Accra, Ghana: IFPRI.
Hollinger F and Staatz JM (2015) Agricultural Growth in West Snyder J, Ijumba C, Tschirley D, et al. (2015) Local Response to
Africa Market and Policy Drivers. African Development Bank the Rapid Rise in Demand for Processed and Perishable
(AfDB) and Food and Agriculture Organisation (FAO) Report. Foods: Results of an Inventory of Processed Food Products
Rome: AfDB, FAO. in Dar es Salaam. Innovation Lab for Food Security, Tanzania
Ijumba C, Snyder J, Tschirley D, et al. (2015) Stages of Trans- Policy Research Brief No. 2. East Lansing: Michigan State
formation in Food Processing and Marketing: Results of an University.
Initial Inventory of Processed Food Products in Dar es Sal- Theriault V, Vroegindewey R, Assima A, et al. (2018)
aam, Arusha, and Mwanza. Innovation Lab for Food Security, Retailing of processed dairy and grain products in mali:
Tanzania Policy Research Brief No. 3. East Lansing: Michigan evidence from a city retail outlet inventory. Urban Sci-
State University. ence 2(1): 24.
Andam et al. 243
Timmer CP (1997) Farmers and markets: the political economy of system transformation. Journal of International Development
new paradigms. American Journal of Agricultural Economics 27(5): 628–646.
79: 621–627. Tschirley D, Snyder J, Dolislager M, et al. (2015b) Africa’s
Timmer CP (2004) Food policy in the era of supermarkets: what’s unfolding diet transformation: implications for agrifood sys-
different? Journal of Agricultural and Development Econom- tem employment. Journal of Agribusiness in Developing and
ics 1(2): 50–67. Emerging Economies 5(2): 1–48.
Timmer CP (2009) Do supermarkets change the food policy United Nations (2014) World Urbanization Prospects: The
agenda? World Development 37(11): 1812–1819. 2014 Revision, Highlights. New York: United Nations,
Traill WB (2006) The rapid rise of supermarkets? Development Department of Economic and Social Affairs, Population
Policy Review 24(2): 163–174. Division.
Tschirley D, Ijumba C, Liverpool Tasie S, et al. (2016) Methods Wang H, Dong X, Rozelle S, et al. (2009) Producing and
for Agrifood Transformation Research: Best Practices in Con- procuring horticultural crops with Chinese characteristics:
ducting Processed Food Inventories. Feed the Future Innova- the case of northern China. World Development 37(11):
tion Lab for Food Security Policy Research Brief No. 24. East 1791–1801.
Lansing: Michigan State University. Zhou Y and Staatz J (2016) Projected demand and supply for
Tschirley D, Reardon T, Dolislager M, et al. (2015a) The rise of a various foods in West Africa: Implications for investments and
middle class in East and Southern Africa: implications for food food policy. Food Policy 61: 198–212.