Urbanization is one of the major social changes sweeping the globe, with rapid growth of the urban population and stagnating growth of the rural population, especially in developing regions (UN-Habitat, 2020). In sub-Saharan Africa, rapid urbanization and poverty are the major fundamental development challenges that are perpetuating and deepening the crisis of food and nutrition insecurity in urban areas (Battersby, 2012). Many people living in urban areas face under-nutrition, mainly due to their lack of income rather than to a lack of capacity to produce food (Satterthwaite et al., 2010). The health and nutritional status of urban populations with very low incomes are at risk from rising prices in staple foods. This became evident with the rising hunger among urban populations after the food price rises in 2007 and the first half of 2008 (Cohen & Garrett, 2009). In South Africa, poverty, unemployment and high food prices are dominant and influence dietary change, which subsequently increases urban food insecurity and malnutrition (Battersby, 2012).

Although South Africa is food secure at a national level, large numbers of households within the country are food insecure. Statistics South Africa (2019) has reported that about 6.8 million South Africans experienced hunger (not having enough food intake) in 2017, with 1.7 million households across the country being affected. Due to low incomes, about 30–50% of the population have insufficient food or consume imbalanced diets (Hendriks, 2014). More than 43% of households are vulnerable to poverty, with almost 28% living in extreme poverty, below the food poverty line (SSA, 2017). With more than two-thirds of the South African population living in urban areas (SSA, 2014), and food prices and unemployment rates continually increasing (SSA, 2019), chronic poverty and household food insecurity become worse as the informal safety nets are weakened, especially in urban areas (Drimie et al., 2013).

According to the 2018 Global Nutrition Report (Development Initiatives, 2018), South Africa is among those countries with a triple burden of malnutrition, where high rates of childhood stunting, anemia and overweight in adult women are dominant. This is mainly because poor people tend to cope with poverty by adopting unvaried diets mainly of starch staples, such as maize-meal, with limited vegetable and fruit intake (Chakona & Shackleton, 2017, 2018; Schönfeldt et al., 2010). The triple burden of malnutrition is especially acute in urban areas, where the greater proportion of the population are more reliant on food purchases and tend to consume energy-rich and processed foods, including refined grains and foods higher in saturated fat, sugar and salt. These foods are cheap to purchase, but can lead to individuals being overweight (Drimie et al., 2013).

Poor households who depend on less formal means of securing food, such as from the informal markets, are the most vulnerable to nutrition insecurity. They mostly rely on informal social safety nets (Battersby, 2011). For example, in many cases, purchased food constitutes up to 90% of household diets, accounting for 60–80% of total household expenditure (Chakona & Shackleton, 2017). Due to reduced purchasing power, “the poor are facing higher food prices but no greater income, and they begin to starve” (Sen, 2008). In South Africa, the estimated average income of the poor is less than R524 (US$35.70) per month per person (NPC, 2012) and a quarter of the population live below the food poverty line of R561 (US$38.22) per month (SSA, 2019). This has a negative impact on a large proportion of households already vulnerable to food insecurity, thereby increasing hunger and malnutrition (Altman et al., 2009).

This situation undermines the efforts to have a country with zero hunger and good health and well-being for all by 2030 and thus has dire consequences for most of the Sustainable Development Goals (SDG) targets. For example, food insecurity has a negative influence on good health and well-being (SDG 3), as well as human development, through its effects on nutrition and as a component of overall family stress. Furthermore, with good nutrition and living healthy lives, adults may not lose employment due to illness, which may form a vicious cycle of poverty and food insecurity. With better nutrition, children may experience less sickness and therefore have more days in school, increasing their chances of attaining higher academic achievement (SDG 4). A good quality education increases their intellectual growth and knowledge; with excellent academic grades, they can get decent employment, thus increasing the nation’s economic growth (SDG 8). This in turn may reduce poverty (SDG 1), leading to reduced inequalities in the country (SDG 10), which are the major challenges South Africa is currently facing.

Food insecurity has often misleadingly been seen as an issue that affects only rural populations, especially in South Africa, where studies have commonly noted the highest levels of food insecurity in these settings (De Cock et al., 2013). There has been a strong bias in promoting food security programs among rural populations (Crush & Frayne, 2010, 2011), including support of smallholder agriculture, and neglecting urban populations. Yet, due to increasing urbanization and rising food prices, food insecurity is increasingly prominent in urban settings, especially within informal peri-urban areas (Chakona & Shackleton, 2017, 2018; Crush & Frayne, 2010). With two-thirds of the South African population living in urban areas and the majority being dependent on cash income to purchase food, increases in food prices directly affect their food security status. In South Africa, food prices have been on the rise since 2015, forcing poor urban dwellers to consume lower quantities of food and adopt diets that are cheap, affordable and of low nutritional quality (Jonah & May, 2020). Despite high levels of poverty and rising food prices, urban agricultural practices are limited in poor urban households. Many households rely on social grants from the government (Chakona & Shackleton, 2019), while limited access to land, labour and inputs constrain the contribution of agriculture to household food security (Crush et al., 2011). As a result, malnutrition remains one of the major nutritional disorders in the country (Chakona & Shackleton, 2018), which prevents the country from achieving the 2030 Agenda for Sustainable Development Goals.

The prime causes of household food insecurity in South Africa are widespread chronic poverty, rising food prices and unemployment, which also weaken informal safety nets, especially in urban areas (Drimie et al., 2013). In South Africa, most households are female-headed, and urban food insecurity is a gendered problem (Caesar & Riley, 2018) related to higher poverty levels (57.2%) among women than their male counterparts (53.7%) (SSA, 2017). Women are generally more vulnerable to food insecurity, in part because the responsibility to make sure the entire household is fed is typically assigned to women. When faced with inadequate income, women who are household food managers trade off food quality for quantity in their own diets to prevent household members, especially children, from feeling persistently hungry (Kuku et al., 2011). Monotonous diets, based mainly on grains and lacking vegetables, fruits and animal-source foods, dominate their diets. These diets are closely associated with food insecurity and malnutrition (Kennedy, 2009). Some women may rely on less expensive foods that are energy-rich but nutrient-poor (Chakona & Shackleton, 2017), thus jeopardizing their health.

While there has been some research in larger urban centres in South Africa (Caesar & Riley, 2018; Drimie et al., 2013; Oldewage-Theron et al., 2006), information is scarce on dietary patterns and food security status among female-headed households in peri-urban settlements surrounding South Africa’s smaller urban settlements. This chapter analyzes household food insecurity and dietary diversity of households residing in the peri-urban areas of small urban centers in South Africa, where higher poverty levels limit households’ ability to access food. The study examined and compared the determinants of food security among male- and female-headed households in these settlements, focusing on household employment and income status, food expenditure, geographical location, household size, educational level and access to food. The novelty of this study is in using a large sample size to carry out comparative analyses on food security of households in peri-urban settlements in four urban centers across three South African provinces.

Study Area

The study was carried out in Richards Bay, Dundee, Harrismith and Balfour. Richards Bay in Kwa-Zulu Natal Province (population 57,000) is characterized by highly developed urban areas surrounded by poor and undeveloped peri-urban and rural areas. The municipality is faced with high levels of unemployment (40%), mainly due to lack of skills and widespread poverty, especially in peri-urban and rural areas (uMhlathuze Municipality, 2012). Average annual household income for the municipality is R121,177 (US$8,243.67) and 15.2% of people do not have any source of cash income (SSA, 2012). The majority of the population are poorly educated: 52.4% of adults do not have secondary qualifications and only 8.5% have higher education. Most households are male-headed (57.7%) and 42.3% are female-headed (SSA, 2012).

Dundee in Kwa-Zulu Natal Province (population 35,000) has a well-developed urban area supporting the agricultural, tourism and manufacturing economic base of the district (uMzinyathi District Municipality, 2012). The municipality is predominantly urban with only 16.8% of the population living in non-urban areas. Dundee has significantly higher educational levels, with about 31.9% of people having completed secondary education and only 7% with no education. However, there are high levels of unemployment in the municipality (26.4%), with the unemployment rate among economically active youth (15–34 years) being 36.2%. Average annual household income for the Endumeni municipality is R87,430 (US$5,848.30) (SSA, 2012) and about 12.4% of households do not have any cash income (SSA, 2012). More than 60% of the population are living below the poverty line. Most households in the municipality (59.3%) are male-headed and only 40.7% are female-headed (SSA, 2012).

Harrismith in the Free State Province (population 28,000) has a central location in relation to the large national cities of Johannesburg, Durban and Bloemfontein. Its economy is mainly influenced by various forms of movement flows, transport and related industries (Maluti-a-Phofung Municipality, 2014). Harrismith is in the most poverty-stricken area in the Free State, with approximately 60% of households earning below R1,650 (US$112.33) monthly and 40% earning an average monthly income of R600 (US$40.85) (Maluti-a-Phofung Municipality, 2014). Approximately 13.5% of households have no cash income (SSA, 2012). Education levels are low: Only 26.8% of people have secondary qualifications; 7.9% have higher education; and 9% have no education. The unemployment rate is high (41.8%), and among the economically active youth (15–34 years) the rate is 53% (SSA, 2012). The majority of the population depends on subsistence farming and backyard gardens for their food supply and livelihood. Half of the households in the municipality are female-headed (50.9%).

Balfour is a town situated along the Kat River Valley in Raymond Mhlaba municipality in the Eastern Cape Province. The area is home to about 50,000 people, mainly IsiXhosa (84%) (SSA, 2012). Education levels are poor, with approximately 9.7% of the population having completed secondary education (SSA, 2012). The area is characterized by extremely low formal employment rates, with 90% of households receiving one or more monthly old age pension or disability grants of R1,570 (US$105) or monthly child support grants of R480 (US$32.50) (SSA, 2012). There are high unemployment rates and government social grants are the most common sources of cash income. The majority of the households are female-headed and many households rely on food purchasing.

Research Methods

Data were obtained through administering questionnaires to randomly selected households within different neighborhoods in each town. A total of 78 households in Balfour, 76 in Richards Bay; 75 in Dundee and 85 in Harrismith participated in the surveys. All interviews were conducted with the person who prepares most of the meals in their preferred language of IsiXhosa (Balfour), IsiZulu (Richards Bay and Dundee) and Sesotho (Harrismith) or English. Women were targeted because in most African households, women are more knowledgeable about household food dynamics and the individual dietary intake occurring within their households.

The questionnaire measured food insecurity in a South African context using household dietary diversity and food insecurity measures reflecting household food availability and food access (Coates et al., 2007; FAO, 2011). Household Dietary Diversity Scores (HDDS) were determined from the respondents’ recall of all food and drink items consumed by the household members during the previous 24 hours. The HDDS tool reflects household food availability and food access, on the premise that households consume a variety of foods when they have the means to acquire them. The tool uses 12 different food groups for analysis and the sum of the food groups consumed by household members from the total of 12 makes the HDDS. Depending on the number of food groups consumed, households were classified into classes of low dietary diversity (≤5 food groups), medium dietary diversity (6–7 food groups) and high dietary diversity (≥8 food groups) (FAO, 2008).

The Household Food Insecurity Access Scale (HFIAS) forms a continuous measure of the degree of household food insecurity, mostly related to access in the past 30 days (Coates et al., 2007). A standardized HFIAS questionnaire composed of nine specific questions with three themes of food insecurity was used (Coates et al., 2007). The nine questions represent a generally increasing level of severity of food insecurity and nine “frequency-of-occurrence” questions were asked as a follow-up to each occurrence question to determine how often the condition would have occurred in the past 30 days. For each frequency-of-occurrence question, a score was assigned to each household: (0) if the answer was “never”; (1) if the response was rarely (condition having happened once or twice in the past 30 days); (2) if it occurred sometimes (three to ten times in the past 30 days) or (3) if the answer was often (occurred more than 10 times in the past 30 days) (Coates et al., 2007). A score was assigned at the end of the nine questions for each household and this ranged from 0 (if answer was “never” for all households) to a maximum of 27 (if answer was “often” for all nine questions). Households were then grouped into different levels of food insecurity, with a high HFIAS score indicating a household’s poor access to food and significant household food insecurity (Chakona & Shackleton, 2018). The food insecurity classes were: food secure (HFIAS 0–1), mildly food insecure (HFIAS 2–7), moderately food insecure (HFIAS 8–11) and severely food insecure (HFIAS > 11).

Some explanatory variables, including household characteristics such as income, ages of interviewee and household head, gender of household head, education status, household size, number of children and adults within the household, employment status of household head and other members, sources of food and household’s monthly food expenditure, were captured in the interviews.

Data were entered and cleaned using Microsoft Excel, and all statistical analyses were performed using Statistica version 13 (StatSoft Inc.). All descriptive data are presented as means and SDs (mean ± sd) and percentages, and this was presented per study site and overall. The differences in HDDS and HFIAS between study sites and gender of household head were tested using two-way ANOVA with Bonferroni’s correction for multiple comparisons. The relationship between HDDS and HFIAS with food expenditure, gender of household head and study site were examined through Spearman correlation tests, and HDDS and HFIAS were used as response variables analyzed as a function of food expenditure, gender of household head and study site. The association between HDDS and HFIAS was also tested using Spearman correlation. Statistical significance was set at p < 0.05.

Description of Households

The sample consisted of 314 households with the overall mean age of the interviewee and household head being 32.3 ± 10.8 and 50.8 ± 15.6 years, respectively (Table 11.1). Overall, 63% of the households were female-headed. The mean household size for the study sites ranged from 4.9 ± 2.4 to 7.7 ± 4.7 persons, with the overall mean of 6.4 ± 4.1 persons. Most interviewees had finished secondary education, with an insignificant percentage having a post-secondary qualification of a certificate, diploma or degree. Balfour had the lowest education level, with only about a quarter of interviewees having matriculated. Overall, food expenditure was low, with households in Richards Bay spending less cash per month on food than those in other locations, and households in Balfour spending more money on food. Food expenditure was generally lower in female-headed households than in male-headed households. All households were dependent on food purchasing and about a third farmed some of their food. Households in Richards Bay were the most likely to use means other than purchasing to acquire food. Most households were relying on income from different social grants and pensions, while a few had members with part-time or full-time employment. Only Harrismith (close to 50%) and Richards Bay (almost 40%) had a higher percentage of employed household heads.

Table 11.1 Sample distribution by household characteristics

Household Dietary Diversity

Household diets were similar across all four sites, with more than 50% of households consuming food groups that included: cereals; spices, beverages, condiments (mostly salt and curry powder); oils and fats; sweets (mostly sugar and sweetened juice); and vegetables (mostly cabbage and onion). Meat (mostly chicken) and milk and milk products were consumed by more than 50% only in Dundee, Harrismith and Richards Bay. Furthermore, white roots and tubers (mostly potatoes) and legumes, nuts and seeds (mostly beans) were consumed by more than 50% of households only in Balfour and Richards Bay, respectively (Fig. 11.1). Other food groups such as fruits, eggs, and fish and seafoods were hardly consumed by households in any of the towns, with the lowest percentage observed in Balfour.

Fig. 11.1
figure 1

Percentage of households consuming different food groups

The HDDS were generally low in all towns, with an overall mean of 7.4 ± 1.7 (Table 11.1). Richards Bay had the highest mean HDDS (8.3 ± 1.61) and Balfour had the lowest (6.2 ± 1.3). There was a significant difference in HDDS between the four locations (F6,309 = 2.45, p < 0.005) which was higher in Richards Bay than in the other towns. Significant differences were also observed in HDDS between Balfour and the other two towns, which were not significantly different from one another (Table 11.1). Most households in these locations were classified in the high-HDDS group, except for Balfour, which had half the households classified as medium and 29% as low-HDDS groups. No significant differences were observed in HDDS between gender of household head (F2,312 = 2.96, p > 0.05), although HDDS were generally higher in female-headed than male-headed households.

Household Food Access and Food Security

The HFIAS was generally low in Richards Bay and Harrismith and moderately high in Balfour and Dundee, with a mean of 9.0 ± 7.0 for the sample (Table 11.2). Overall, 23% of the households were food secure, 22% were mildly food insecure, 31% were moderately food insecure and 24% were severely food insecure. Households in Dundee and Balfour had the highest HFIAS scores, of 11.4 ± 7.1 and 10.7 ± 6.2, respectively, showing low food access, while Richards Bay had the highest level of food access (HFIAS score of 5.5 ± 6.8). There was a significant difference in HFIAS between these peri-urban locations (F6,309 = 44.9, p < 0.001). Significant differences were observed between Richards Bay and the other three towns and between Dundee and Harrismith; no significant differences were observed between Balfour and these two towns. The prevalence of household food insecurity was much higher in Balfour, with almost 50% of the households being severely food insecure (Table 11.2). About 50% of households in Richards Bay were food secure, with only 8% being severely food insecure. No significant differences in HFIAS were observed between gender of household head (F2,312 = 49.7, p > 0.05), although food insecurity was generally higher in male-headed than in female-headed households.

Table 11.2 Household dietary diversity scores and the percentage of households grouped in different classes

There was a significant negative correlation between HDDS and HFIAS, HDDS and town as well as HDDS and household food expenditure. Positive significant correlations were also observed between HFIAS and food expenditure and HFIAS and town (Table 11.3). No significant correlations were observed between the food security indicators and the gender of household head.

Table 11.3 Percentage of food insecure and food secure households as classified using HFIAS


South Africa has the second largest economy in Africa and has adequate food supply at the national level, but this is not translated into the food security status of many households (Shisana et al., 2014). This is also reflected in the study, as households in all four peri-urban locations had poor dietary diversity, especially in Balfour, where only one in five households had good quality diets. In the other three study sites, more than 50% of households had diets lacking good quality and diverse foods rich in all essential nutrients. Similar findings have been reported in other studies (Chakona & Shackleton, 2017, 2018; Labadarios et al., 2011). Furthermore, about 77% of the households were food insecure, with one in four households being severely food insecure (Table 11.4). It may be that most households have food on the table but are consuming cheap diets based on grains and lacking essential nutrients, as was also noted by Gina Kennedy (2009). Balfour had the highest percentage of severely food insecure households (49%), with households consuming fewer food groups, which could be attributed to high unemployment rates.

Table 11.4 Spearman correlations between HDDS and HFIAS and selected socio-economic indicators

Surprisingly, given reports that vulnerability to food insecurity is more pronounced in female-headed households (DoA, 2002); in this study, food insecurity was higher and dietary diversity lower in male-headed households, even though male-headed households spent more money on food. This could be because more female-headed households had diversified means of obtaining food other than food purchasing. Of the households who obtained some of their food from farming, collection from open spaces, and donations, 74%, 74% and 71%, respectively, were female-headed. The addition of food from these other sources acted as “women’s safety net” and could have increased their dietary diversity, thus lowering their food insecurity scores. Previous research (Chakona & Shackleton, 2017, 2019) reported that households who supplemented their food with farm produce and wild foods had a lower risk of food insecurity than those who did not, as these alternative sources would lower dependence on food purchasing. The advantage for female-headed households could also be partly due to social grants, a factor that Belinda Dodson et al. (2012) speculated accounted for better relative outcomes for low-income, female-headed households when compared to low-income nuclear households.

The study also showed strong correlations between household dietary diversity, food access and monthly food expenditure. As household dietary diversity increased, monthly food expenditure decreased, while the ability of a household to access food decreased. This was mostly evident in Balfour, where food insecurity was high, with less food diversity, although households in this study site spent more on food. This could be because households in Balfour relied mostly on food purchasing, mostly of cereals (maize-meal and rice), spices, condiments, beverages, oils and fats and sweets, which could not increase their dietary diversity and food security status. However, it has been observed that within the study sites, more female-headed households would only purchase food that is important when necessary. A greater percentage of households in the study sites consumed a number of similar food groups, irrespective of geographical location, income status or level of education. More than 50% of the households across all towns consumed similar food groups such as cereals, spices, beverages, condiments oils and fats, sweets and vegetables. Other food groups consumed by households were limited to the location of the town. For example, food groups such as meat, and milk and milk products, were consumed by more than 50% of households only in Dundee, Harrismith and Richards Bay, while roots and tubers, and legumes, nuts and seeds were consumed by more than 50% of households only in Balfour and Richards Bay. Other food groups such as fruits, eggs, and fish and seafoods were seldomly consumed, especially in Balfour, which explains the significant differences in HDDS evident across the four peri-urbans locations. Furthermore, dietary diversity was strongly negatively correlated with living in a town, also reflecting that household characteristics (e.g., food expenditure) and the food environment within a town have significant effects on dietary diversity and food security.

The levels of food insecurity were higher in the peri-urban areas characterized by low education levels and high unemployment rates, with most households relying heavily on social grants for daily cash needs, as was evident in Balfour. On average, households in Balfour consumed six food groups that lacked the essential nutrients found in fruits, vegetables and animal protein. Within the peri-urban settings, the poorest households may have limited access to food due to high levels of poverty and unemployment, possibly due to household members’ limited education, which affects their ability to get formal employment. This study concurs with Oldewage-Theron et al. (2006) and Labadarios et al. (2011), who also found that poor socio-economic status has an impact on household food insecurity in peri-urban locations in South Africa.

Food insecurity in the study sites is closely linked to poverty. Poor households coped with poverty and rising food prices by reducing the quality, quantity and number of meals consumed per day, as reported in earlier research (Chakona & Shackleton, 2018). John Hoddinott and Yisehac Yohannes (2002) also noted that higher dietary diversity is positively associated with higher socio-economic status and household food security. However, different households have different capacities to adapt and mitigate these effects, therefore umbrella intervention approaches are not enough in addressing food insecurity challenges, as some households can be left behind. Through a gender-based analysis, it is more apparent in the study that non-commercial food sources and the “informal safety net” are important factors determining household food security status. Affordability has been the major constraint to the consumption of good quality diets, in adequate amounts, in these peri-urban areas. With the global climate change impacts affecting food systems and impacting food prices, coordinated research programs that emphasize nutrition-sensitive approaches may help address these challenges. Adopting more localized and diverse food systems that are context-specific, sustainable, healthy, and available at local markets at affordable prices, is an important component in addressing poverty, food and nutrition insecurity in peri-urban settings of South Africa. This is essential for creating opportunities for increased access to locally grown foods and improved local economies through income generation and employment creation in these settings.