Improving food security in the rural areas of KwaZulu-Natal province, South Africa: Too little, too slow

Abstract

Food insecurity is still remarkably high in the poorest areas of the KwaZulu-Natal province of South Africa. Many rural households struggle to have sufficient access to the food they need or prefer. This article explores the extent of food access insecurity and assesses the perceived impact on food security of an Empowerment for Food Security Programme that was launched in the Province in 2007. One of the programme aims was to improve agricultural practices in community gardens, home gardens and broiler production. Data were collected among 390 beneficiary households involved in these agricultural projects in 2010. Findings confirmed that experience-based food insecurity levels were still high, despite the agricultural support programme and the government income transfers. Nevertheless, respondents attribute an improved dietary diversity and better access to resources to the programme.

Keywords:

• food security
• case study
• HFIAS categories
• dietary diversity
• hunger index

1. Introduction

Undernutrition and obesity, two problems on the extremes of the food intake scale, are frequently reported in South Africa (Faber et al., 2011; Labadarios et al., 2011). Many studies have concluded that a large number of South African households are food insecure (e.g. Hendriks, 2005). Based on the 1995 Income and Expenditure Survey, Rose & Charlton (2002) calculated that 43% of households were subjected to food poverty; the National Food Consumption Survey of 2005 showed that 52% of households were experiencing hunger (Labadarios et al., 2008); and based on the General Household Survey of 2007 it was estimated that 10% of adults and 12% of children were sometimes or always hungry (Altman et al., 2009). Based on a comparison of two National Food Consumption Surveys (2000 and 2005) and the 2008 SA Social Attitude surveys, Labadarios et al. (2011) conclude that the national prevalence levels of food insecurity decreased dramatically since 1999. The authors estimate that 26% of all South African households were food insecure in 2008. While these studies each used their own definition of food insecurity, food poverty, or hunger (refer to the individual studies for details on the methodologies used⁹), they all show that South Africa still faces high rates of food insecurity and hunger in certain areas of the country (see also Hendriks, 2005). The South African National Health and Nutrition Examination Survey (SANHANES-1) established by the Human Sciences Research Council provides a broad and comprehensive platform to study the health and nutritional status of South Africa. According to the 2012 SANHANES-1 survey, 45.6% of the population were food secure (meaning access by all members of a household at all times to enough food for an active, healthy life), 28.3% were at risk of hunger, and 26.0% were experiencing hunger (SANHANES- 1, 2013).

In this paper we attempt to contribute to the debate on rural food insecurity in South Africa with a case study from two districts in KwaZulu-Natal (KZN), where an agricultural support programme – the Empowerment for Food Security Programme (EFSP) – was launched three years prior to the survey. The aim of this study is twofold, namely: to characterise the self-reported food access insecurity among rural households in the case study area; and to gauge whether involvement in the EFSP has made a difference to the participants' food security status.

Our approach to measuring food insecurity is based on the self-reported experience of food insecurity outcomes caused by insufficient access to food such as feeling hungry, cutting down on the number of meals, eating food that is less liked or having a less diverse diet. This is measured by a Household Food Insecurity Access Scale (HFIAS), which has been widely used (e.g. in Shisanya & Hendriks, 2011; Swindale & Bilinsky, 2006b; Becquey et al., 2010; Knueppel et al., 2010). Furthermore, the Household Diet Diversity Scores (HDDS), expenditure on food and the number of hungry months in a year were calculated.

There is an expectation that food security levels would improve because of the different support measures implemented by the government, including the EFSP and financial support in the form of social grants (largely child support grants and pensions). Yet self-reported food access insecurity levels are still high. The EFSP makes a difference to self-reported livelihoods, but it does not seem to lift households out of the food-insecurity risk zone. One obvious explanation is that food security cannot be seen in isolation from other development questions such as income sources, rural and urban development, household demographics, and access to different types of resources, including access to water, land, credit, technology, and markets (Altman et al., 2009). As elsewhere, food insecurity in South Africa is strongly associated with poverty and lack of income generation. South Africa has one of the most unequal income distributions in the world (Gini score of 0.63 in 2009; World Bank, 2012). While the income inequality has well-known historical roots, this distribution has also continued to dominate the post-apartheid period. It is worth mentioning that recent studies indicate that inequality is no longer driven by differences between population groups, but it is also manifested within these groups (Leibbrandt et al., 2007; Pauw, 2007). This partly explains why South Africa is characterised by high levels of poverty when compared with other middle-income countries. Depending on what poverty line is used, estimates show that between 45% and 57% of the population is considered to be poor (Rose & Charlton, 2002; Jacobs, 2009).

Both poverty reduction and achieving food security are high on the agenda of the South African government. The government adopted an Integrated Food Security Strategy in 2002 that explicitly aims to eradicate hunger and nutrition deficits among poor households (Jacobs, 2009). One of the projects is the EFSP in the KZN province, which was launched in 2007. This programme seeks to improve agricultural practices, improve institutional capacity at all layers of government, increase agricultural focus in education and awareness programmes, and increase collaboration with health and social services functions of provincial and national governments. Within this programme, support is given to the establishment of communal food gardens, homestead gardens and broiler projects in several communities of the Province.

In 2010 we interviewed 390 beneficiary households of one of these EFSP interventions. The results point to high levels of experience-based food access insecurity, which contrasts with the fact that most households are positive about the support they receive through the EFSP and how the programme has improved their diet diversity. However, the results confirm the challenges policy-makers face in intervening to reverse the food insecurity of ‘creeping’ vulnerability (Misselhorn, 2009), or in other words to break the vicious poverty–food insecurity–livelihood failure cycle that causes the chronic malnutrition described in Faber & Wenhold (2007). Hence interventions such as the EFSP may improve food security, although the impact could be smaller and slower than expected.

The next section of this paper presents a short literature overview on factors explaining rural food insecurity in South Africa. This is followed by a description of the methodology used to collect and analyse data. Subsequently the results are presented and discussed, followed by a conclusion.

2. Rural food (in)security in South Africa

Poor households are at risk of food insecurity, both in rural and urban settings. This risk increases with a lack of availability and access to resources, which includes money, land, transport, and intellectual capital (Van Averbeke & Khosa, 2007; Mkandawire & Aguda, 2009; Faber et al., 2011); these factors all influence the households' ability to access food, either by having the means (land, labour, money for seeds and inputs, entrepreneurship) to produce food themselves or to buy food on the market (money and transport).

Rose & Charlton (2002) suggest that the incidence of household food poverty in South Africa increases with decreasing income (less money to buy food), increasing household size (higher demand for food), female-headed households (generally poorer and limited in access to resources) and being located in rural areas. Jacobs (2009) adds that, apart from income (poverty), food insecurity is determined by households' wealth/asset status and livelihood strategies. The recent rises in food prices (attributed to factors such as domestic electricity supply constraints, rising oil prices, demand for biofuels, increased speculation in the commodity markets and the changing powers within agro-food chains) have impoverished food insecure households even more (Altman et al., 2009; Faber et al., 2011).

On the issue of income, it should be noted that in many poor South African rural areas, households depend strongly on income and cash obtained through government transfers. Also in the study population, grants (old age pensions, disability grants, foster care grants and child support grants) are important contributors to the households' purchasing power (as also reported by Labadarios et al. [2011] amongst others).

While food insecurity is strongly correlated with rurality (Jacobs, 2009), one should note that hunger and nutrient deficiencies also emerge in other parts of the country, including in urban and peri-urban informal settlements across South Africa. Oldewage-Theron & Slabbert (2008) and Oldewage-Theron & Kruger (2011) found high levels of food insecurity in the poverty-stricken informal settlements of the Vaal Region of South Africa. With an estimated 13.5% of all South African households living in informal settlements, this group of food insecure poses particular challenges because of their poor shelter conditions, difficult access to water and to food (Oldewage-Theron & Slabbert, 2008), and strikingly high poverty levels. Food security of the urban poor and of the rural non-farm communities is highly dependent on their ability to buy food, and therefore on their ability to earn cash income, and on the price of food (Ruel et al., 1998).

Poor rural landed households supplement food bought on the market by their own production. It is estimated that there are some four million smallholder subsistence farmers in South Africa. Among them, only a small percentage can generate enough income from farming alone (Pauw, 2007). The mainly black small-scale farming sector contributes little to the total South African agricultural gross domestic product, but it has an important role to play in improving the food security of rural communities. Two distinct, widely recognised, nutritional benefits are: the extra food for own consumption; and ‘income saving’, as money is saved that can be spent on even more nutritious foods (i.e. energy-dense foods such as fats, oils, and meat) (Kirsten et al., 1998, 2003; Aliber & Hart, 2009). Dovie et al. (2003) estimated the direct value of the small-scale farm production for a case study in the Limpopo province. They estimate the direct-use value of crops at $443 per household per year. The production output seems to be relatively high compared with what could be expected from low input use, but lack of support services and uncertainty about access, entitlement and control over farmland are limiting factors (Dovie et al., 2003). Van Averbeke & Khosa (2007) confirm that irrigated vegetable production contributed to vitamin A and C availability, but production did not alleviate the lack of proteins and iron in the diets. Hendriks & Msaki (2009) show that dietary diversity and nutrition intakes increased with engagement in certified organic farming in the Embo community of the KZN Province.

Similarly, Faber et al. (2011) studied the beneficial effects of participating in food garden projects. In the three different projects they studied, vitamin A intake and diet variety were higher among the participants in the projects, and participation seemed to decrease the incidence of diarrhoea. However, Webb (2000) came to a different conclusion and he is more critical about the contribution of garden projects to the nutritional status of the participants. He critiques previous studies on their methodological flaws and argues that production from these gardens is low and that the emphasis is on cash income, and not on nutrition. As a result, a significant group of farmers are net deficit food producers (Drimie et al., 2009) and many smallholder farmers fail to achieve food security (also reported by Oni et al. [2010] and Shisanya & Hendriks [2011]). Moreover, Drimie et al. (2009) argue that this results not only from insufficient access to food, but also from limited dietary diversity. Shisanya & Hendriks (2011) also found high food insecurity levels among participants in community food gardens in KZN. Nevertheless, they show that community food gardens have a positive effect on increasing productivity.¹⁰

Despite the more critical viewpoints, there is consensus that supporting farm production may reduce vulnerability to food insecurity and improve livelihoods (Baiphethi & Jacobs, 2009). As mentioned by Baiphethi & Jacobs (2009) and Drimie et al. (2009), to name two, the main problems to be tackled include the lack of access to production technologies, inputs, water and markets in a context of high environmental pressure on the already marginal lands. The success of food gardens tends to be limited by high start-up costs, drought, difficult access to markets, lack of fencing, and inadequate land for production (Baipethi & Jacobs, 2009). Aliber & Hart (2009) report on the failure of food gardens in the Limpopo Province for similar reasons. Dovie et al. (2003) stress the uncertainty that farmers face with regard to access to farm land. Government and donor-funded projects seem to fail to impact on food security because they tend to focus on the transfer of high-input technologies, often being centred on ‘exotic’ crops. Valente (2009) finds that land reform beneficiaries do not appear to have experienced lower food insecurity, which she attributes to the lack of complementary production factors.

Against this background, this paper explores experience-based food insecurity levels of four communities in the KZN Province. The household's food access insecurity experience is measured on three dimensions (adapted from Hart, 2009), namely: an intensity dimension proxied as the relative level in self-reported experience of having insecure access to food (using an assessment of the self-reported outcome of food insecurity problems in a Household Food Insecurity Scale); a temporal dimension (by calculating a hunger index that reflects the length of the lean season); and a health dimension (by calculating a household dietary diversity score). The experience of the intervention on the people's livelihoods is assessed by asking the beneficiaries whether they perceived changes, and by comparing the self-reported food security indicators of households who joined the project since the start with those who joined more recently.

3. Methodology

3.1 Research setting

The EFSP is coordinated by the Department of Agriculture and Environmental Affairs of the province in cooperation with the Flemish Development Cooperation. Programme activities cover a large area of KZN. Two districts in the northern region and two in the southern region were selected for farm interventions based on the prevalence of poverty, unemployment rates and HIV/AIDS. In each district, local municipalities were chosen based on these indicators and other factors including institutional aspects and the presence of other development programmes. Direct support to households is given by providing inputs, fencing and training on community gardening, home gardening and broiler projects.

A survey was organised in July and August 2010 in two districts in the south and two districts in the north. Within these districts, nine municipalities were selected (Edumbe, Pongola, Big 5 False Bay, Umhlabuyalingana, Umgeni, Umsunduze, Richmond, Vulamehlo and Umziwabantu). These nine municipalities are amongst the 29 poorest performing municipalities in the province. A sample of 390 participating households was randomly selected from amongst households that benefited from the 38 EFSP projects in these nine municipalities. Most households interviewed (92%) were involved in community gardens, 42% were involved in a home gardening project, and 36% in a broiler project. About one-half of the households were involved in only one project, 34% in two projects, and 13% in three projects. Households that recently joined the project served as the control group in this study.

The questionnaire contained seven sections dealing with different aspects at household level: household demographics; characteristics of the households; food availability, consumption and dietary diversity in the household; food production in the household; income and expenditure of the household; stresses, shocks, and coping and intervention strategies affecting the household; and project outcome evaluation.

3.2 Food insecurity indicators

As mentioned above, the HFIAS is a self-reported food insecurity measure, based on a methodology developed by USAID in the Food and Nutrition Technical Assistance (FANTA) Project. FANTA (Frongillo & Nanama, 2006; Coates et al., 2007) to categorise households into different food security types. The classification is based on the answers to nine questions related to different dimensions of access to food. The nine items were designed to capture experiences associated with varying levels of food insecurity severity and to reflect three domains perceived as central to the experience of food insecurity: anxiety about household food supply; insufficient quality, which includes variety, preferences and social acceptability; and insufficient food supply and intake and the physical consequences.

The HFIAS method differs from more quantitative measure of food security (e.g. 24-hour recalls). Its main advantage is that data collection and analysis is easier compared with food intake recalls or anthropometric measures (FAO, 2008). The disadvantage is that it mainly measures the perception of anxiety about insufficient access to food. Some studies have tried to assess the external validity of HFIAS measures, and its validity and reliability have been confirmed by Becquey et al. (2010), Knueppel et al. (2009), and FAO (2008). The HFIAS measurement alone will not give a full picture of the food security and nutrition status of households and individuals, and it is prone to measurement errors. For this reason, additional measures, including dietary diversity, the number of months people felt hungry in a year, expenditure on food, and production diversity, have been reported here.

The HDDS was also developed by the FANTA project (Swindale & Bilinsky, 2006a, b). Tests of these methods can be found in Onyango et al. (1998), Ruel (2003), Bernal & Lorenzana (2005) and Savy (2006). Different food groups are distinguished in the survey. If a household reported that it consumed one or more food items from a food group in the last seven days, that food group scores a one; if not, a zero. To obtain the HDDS index, the food group scores are summed. The maximum HDDS value is 12 with the minimum value equal to zero (following Swindale & Bilinsky, 2006a).

The hunger index reflects the number of months in which the household reported experiencing food security problems. This index ranges between zero and 12.

To estimate the importance of farm production, a livestock index was calculated as an aggregated variable that reflects livestock ownership. Different types of livestock are converted into large livestock units and added to obtain an aggregate value. The crop index is a simple reflection of the number of crops cultivated by the household.

To measure poverty levels, the $1.25 and $2 per day poverty line was used. Total household income is converted from dollar to Rand and into income per household member per day. This variable is used to create two categories of households depending on whether household members acquire more or less than $1.25 per day. The same procedure can be applied for the $2 per day poverty measure.

Similar household and food insecurity indicators were used in the baseline survey of the project. Data reported in Table 1 were collected in 2007 (LIMA, 2008).

Table 1: Overview of food insecurity and poverty measurement: 2007 baseline study

	2007 (n = 646)
Food secure households (%)	6.5
Mildly food insecure households (%)	6.2
Moderately food insecure households (%)	41.3
Severely food insecure households	46
Less than $1 per day (%)	51
Less than $2 per day (%)	85
HDDS	9.08
Hunger index (number of hungry months)	3.75
Livestock index (TLU)	50.19
Crop index	4.31

Source: LIMA (2008).

3.3 Project impact analysis

The project evaluation is based on two approaches. The first consisted of recording the perception of the participants on the impact on their livelihoods. The participants were asked how they felt the project had (or had not) improved or changed their lives and livelihoods.

A second approach was to estimate how the project impacts on the household's food security levels. We use data from the recently joined households (who had joined within the last year at the time of the survey) as a control group.

4. Results

4.1 General socio-economic characteristics

Generally the baseline report (LIMA, 2008) concluded that high levels of food insecurity and poverty were prevalent throughout the entire KZN study area. Almost one-half of the respondents were food insecure and had to live off less than $1 per day. Conditions of physical infrastructure at household level were found to vary considerably between municipalities. Furthermore, it was shown that participation in local trading of agricultural products including vegetables, fruits, meat and poultry, eggs, nuts, roots and tubers and legumes had the potential to stimulate growth in income. Therefore the report recommended that these crops and livestock products should be prioritised in EFSP activities. Finally, the report emphasised the need for an enabling macroeconomic environment consisting of sound institutional arrangements and land tenure structures.

Table 2 reports the general characteristics of households covered in the 2010 survey. On average, a household in the sample consists of seven people. Household sizes tend to be a little bigger in the northern region, although little difference can be found. The average age of household heads in the north was 54, while in the south it was 56. When looking at the gender of the household heads in our sample, we see that 55.1% of the interviewed households are male headed while 45% are female headed. The table shows the low level of education in the area. One-third to one-half of all household heads did not enjoy any schooling at all. Analysis shows that in the southern area 25% of household heads obtained some secondary education, while in the northern region only 18% did.

Table 2: Household characteristics of the 2010 follow-up survey

			North KZN		South KZN		Tests^a
	North	South	Zululand	Mkhanyakude	Ugu	Umgungundlovu
n	257	133	106	140	83	29
Demographics							3.124***
Household size	7 (2.8)	6 (3.2)					–1.875*
Age of household head	54 (13)	56 (12)					2.472
Male household head (%)	58	49.6
Schooling (% completion)
No schooling			33.0	51.4	37.3	41.4	19.241
Grade 0 to 2			28.3	20.0	21.7	13.8
Grade 5 to 7			17.9	12.1	16.9	17.2
Grade 8 to 11			14.2	11.4	20.5	17.2
Matric			6.6	3.6	3.6	10.3
Courses or certificates			0.0	0.7	0.0	0.0
Diploma or degree			0.0	0.7	0.0	0.0
Crop production (% producing)
Spinach			95	90	88	91	2.912
Cabbages			97	91	77	88	22.715***
Onions			89	87	78	58	24.254***
Beetroot			84	80	69	72	8.317**
Carrot			53	61	72	67	8.199**
Maize			39	51	57	54	6.730*
Tomato			39	56	31	35	17.557**
Purchased inputs (% reported use)
Fertiliser			96	92	97	91	4.820
Pesticide			86	70	58	52	25.617**
Constraints (% reported occurrence)
Lack of seeds			100.0	95.2	100.0	100.0
Lack of fertilisers			100.0	92.9	100.0	100.0
Lack of water			100.0	96.7	100.0	100.0
Lack of labour			85.7	94.1	100.0	100.0
Too old/young/weak			87.5	90.0	100.0	100.0
Pest			100.0	83.3	100.0	92.3
Little money			0.0	100.0	100.0	0.0
Other reasons			95.7	96.0	100.0	100.0

Notes: Data presented as mean (standard deviation). ^at-tests are reported for continuous variables, chi-squared test for categorical variables; no chi-squared tests are reported for constraints because of the many farmers concerned. *Significant at the 10% level; **significant at the 5% level; ***significant at the 10% level.

Different types of crops are grown in the study area, with spinach, cabbages, onions and beetroot the most popular. These crops are each cultivated by more than 75% of the households. Crop diversification is evaluated using a crop index that reflects the number of crops cultivated by a household. An average crop index of 7.7 (standard deviation [SD]: 3.4) was recorded, implying that on average each respondent cultivates eight different crops.

Northern KZN seems to have the largest share of respondents who utilise pesticides, while the use of fertilisers is equally high in the two regions. Furthermore, some 80% of the respondents sell some of their agricultural production (data not shown), which is surprisingly high. About one-fifth of the respondents produce crops solely for own consumption, with the highest proportion found in Umgungundlovu (27%) and the lowest in Zululand (14%).

Most farmers indicated that they are not using all of their accessible arable land. The most commonly cited reasons were lack of seeds, lack of fertilisers, lack of access to water and high incidence of pests (Table 2). In Ugu, households indicated that they suffered from all the reasons put forward.

4.2 Household food production

Figure 1 shows access to different types of resources, including access to information, water, farm technology, credit and markets. In general, access to information is high (95% of respondents) and access to credit more problematic (27% of respondents). There were few significant differences between the northern and southern regions.

Figure 1: Access to resources in two regions (share of households)

Most households keep chickens, followed by cattle and goats (Figure 2). Goat and chicken holding tends to be higher in the northern districts. Cattle owners own on average two head of cattle, but as with other livestock types the number varies greatly (with a SD of 5) while goat owners have on average two goats (SD: 4.2). This reflects the very limited presence of livestock in this area. The only exception is poultry, where a household owns on average 10 chickens (SD: 11.6); this is probably due to the broiler projects that have been started in the project area. The total livestock index, calculated in livestock units (cattle: 10; goats and sheep: 5; chickens: 1) was also calculated for the different districts. It is the highest for households in Zululand (average: 63; SD: 79) and lowest for Umgungundlovu (average: 22; SD: 40).

Figure 2: Livestock ownership

4.3 Household food consumption

Diversity in food consumption was measured by the frequency of consumption of different types of food over a period of 7 days. Figure 3 describes the share of households that indicated that at least one person in the household had consumed the specific type of food at least once in the last seven days. Note that vitamin A-rich fruits and vegetables are yellow/orange coloured (such as butternut and carrot) or dark-green leafy vegetables (such as spinach and beetroot leaves). Other vegetables include beetroot, cabbage, onions and tomato. Other fruits include peaches and oranges.

Figure 3: Consumption in the last seven days (share of households)

Maize was consumed in most households at least once in the last seven days prior to the interview, while consumption of fruits and protein sources such as eggs and dairy is quite low. For the latter products only 60% of households indicate that they have consumed the product in the last seven days. Consumption of maize, other cereals and vegetables is highest.

When looking at the frequency of consumption the same pattern can be found. Maize is the most popular food product as it is consumed on average five times a week. Vegetables and other cereals are consumed on average three to four times a week. The lowest consumption rates are found for meat, fruits and eggs; these are consumed on average twice a week. Households across the province mainly purchase their food (Table 3). Over 70% of the participating households indicate that they purchase most of their food. Vegetables tend to be an exception as the main source was own production (for over 65%).

Table 3: Importance of different sources of food products (% of households)

	Purchased	Own production	Other sources
Maize	88	10	2
Other cereals	94	2	4
Roots and tubers	70	24	6
Vitamin A-rich fruits and vegetables	22	74	4
Other vegetables	30	68	2
Other fruits	77	10	13
Legumes and nuts	78	17	5

Another indicator of the food consumption status is the number of meals taken per day: 15% of the respondents indicated that the adults in the household took only one meal the day before; and 17% that children in the households only took two meals the day before.

Table 4 presents the average monthly food expenditures per capita. Expenditure on food averaged some 60% of total expenditure. Figure 4 illustrates how household food expenditure is distributed. On average the largest share of the food budget is spent on cereals and bread, followed by animal products.

Figure 4: Food expenditure pattern

Table 4: Food expenditure and share of food expenditure (% food on total expenditure)

	North KZN		South KZN		Total sample
	Zululand	Mkhanyakude	Ugu	Umgungundlovu
Monthly food expenditure (R)	508.68 (279.71)	461.84 (267.08)	530.0 (252.23)	563.55 (342.12)	501.93 (277.04)
Share of total expenditure on food purchases (%)	0.58	0.63	0.57	0.46	0.58 (0.178)

Note: Data presented as mean (standard deviation). F-stat for monthly food expenditure (1.905), χ² for share of food purchases (9.074), significant at the 1% level.

4.4 Household income and its sources

Household income was measured using different income categories, from having less than R500 to having more than R3000 per month. Figure 5 illustrates the distribution of income categories in every district. In Zululand and Umgungundlovu, relatively more households belong to higher income categories compared with other districts.

Figure 5: Distribution of households in different income categories

The average monthly income per capita was calculated using these income categories. Average monthly income per capita is highest in Umgungundlovu (R320; SD: 429), followed by Ugu (R275; SD: 260). The southern districts both have an average income per capita per month of R190 (SD: 122).

An income analysis for the different household members shows that the head of the household obtained an income by working in only one-fifth of the households; 22.5% and 19.8% of household heads in the south and north respectively. At district level, Mkhanyakude and Ugu had the highest percentage of household heads working for cash or in-kind income, namely 24.3% and 23.9% respectively. Meanwhile, in Zululand and Umgungundlovu districts, 13.8% and 19.5% of household heads were working for cash/in-kind income respectively. Common reasons why some household heads were not working include old age, inability to find a job and preference to be a homemaker.

When looking at the main income sources in the area, it is clear that grants such as pension and child support and gifts from family and neighbours are the most important income sources (Table 5) – between 72 and 88% of households partially depended on this type of income, and it is the main source for 66% of households (data not shown). Farming contributes to the household income of 20% of the households, while it is the main income source for only 3% of respondents. Income shortages are mainly experienced during the months of January, February, June, and July.

Table 5: Income sources (%) of households

	North KZN		South KZN
	Zululand	Mkhanyakude	Ugu	Umgungundlovu
Grant/gift	88	86	84	72
Farm income	28	23	24	21
Formal salary	14	33	26	33
Remittances	28	13	9	19
Skilled labour and business	6	9	10	2
Unskilled labour	2	9	1	5

4.5 Food insecurity and poverty status

Across the full sample only 5.6% of the population were labelled food secure; about 55.4% of the population are severely food insecure, 30.5% are moderately food insecure and 6.9% are faced with mild food insecurity. The food security status at district level is summarised in Table 6. The most affected district is Umgungundlovu, where about 70% of the respondents are faced with food insecurity according to the HFIAS categories.

Table 6: Household food security status (share of household)

	North KZN		South KZN
	Zululand	Mkhanyakude	Ugu	Umgungundlovu
Food secure	7.3	3.4	9.1	2.4
Mildly food insecure	11.9	2.8	6.8	9.5
Moderately food insecure	37.6	28.3	34.1	16.7
Severely food insecure	43.1	65.5	50.0	71.4
Total	100	100	100	100

A second indicator of food security is the hunger index, which is highest for Mkhanyakude and Umgungundlovu districts (3; SD: 3) and lowest for Ugu district (2; SD: 2).

Table 7 presents a comparison of household characteristics by HFIAS category. The groups do not differ much in general household characteristics, yet they differ with respect to income, expenditure, expenditure on food and share of this expenditure in total expenditure. Clearly the severely food insecure groups are on average the poorest group in terms of income and expenditure. Poverty levels (as defined by an income below $1.25 and $2 per person per day) are the highest among the severely food insecure group.

Table 7: Household characteristics by food security status

	Food secure	Mildly food insecure	Moderately food insecure	Severely food insecure	F-stat
n	20	26	119	226
Age	52.14 (11.95)	58.41 (14.99)	55.22 (12.38)	53.28 (12.49)	1.780
Total household size	5.00 (2.70)	5.48 (3.35)	5.88 (2.59)	5.85 (2.97)	0.733
Total income earners	1.09 (1.11)	0.78 (0.85)	0.76 (0.94)	0.57 (0.76)	5.948***
Time joined to the EFSP (months)	25.48 (10.33)	24.54 (10.25)	26.66 (8.73)	24.95 (10.39)	0.817
Monthly income per capita (R)	474.47 (450.47)	395.09 (263.35)	225.39 (256.40)	178.17 (132.85)	18.871***
Daily income per capita (US$)	2.15 (2.04)	1.79 (1.19)	1.02 (1.16)	0.82 (0.60)	18.871***
Share on less than $1.25 per capita per day	45.5	29.6	74.8	80.1	χ^2 39.864***
Share on less than 2$ per capita per day	63.6	81.5	91.6	96.3	χ^2 33.270***
Monthly expenditure (R)	1364 (1021)	1576 (1029)	948 (811)	899 (740)	6.869***
Monthly expenditure on food (R)	509 (202)	724 (354)	509 (242)	472 (279)	6.679***
Share expenditure on food	0.48 (0.21)	0.53 (0.17)	0.60 (0.15)	0.59 (0.18)	3.449***
Crop index	8.00 (3.90)	7.89 (3.62)	7.76 (3.45)	7.57 (3.62)	0.174
Livestock index	51.05 (54.65)	52.11 (59.77)	45.76 (61.09)	34.49 (59.89)	1.588
Hunger index	0.54 (1.01)	1.00 (1.11)	2.04 (2.09)	3.15 (2.85)	14.176***
HDDS	11.18 (1.14)	11.00 (2.13)	10.55 (1.68)	9.62 (2.17)	10.245***

Notes: ***Significant at the 1% level.

Table 8 presents an overview of the most important shocks and stresses experienced in the research area. The most important are increases in food costs and, related to that, increases in food production costs. Drought and a sudden loss of livestock are the most prevalent stresses. Food insecure households were relatively more fearful of suffering from serious injuries or illnesses, whereas food secure households indicated relatively less stress due to droughts.

Table 8: Common shocks and stresses (share of households)

	Food secure	Mildly food insecure	Moderately food insecure	Severely food insecure	χ^2 stat
Food cost or food price increase	77.3	70.4	79.0	76.9	0.941
Increase in food production cost	31.8	51.9	52.7	52.3	3.454
Serious injury or chronic illness	36.4	25.9	43.7	50.0	6.830*
Drought	31.8	33.3	32.8	49.1	10.369**
Increase in household size	31.8	33.3	31.9	34.7	0.302
Death of livestock	18.2	22.3	36.1	27.3	5.070

Notes: **Significant at the 5% level; ***significant at the 1% level.

Borrowing from friends and relatives seems to be the most important coping strategy in the area, followed by a reduction in food consumption and spending, followed by sales of assets. Only one-fifth of the respondents are able to use savings when stresses or shocks occur. Furthermore, only one-fifth look for additional income sources through taking up additional work. The importance of the negative coping strategies such as lending and selling of assets clearly reflects the severe negative effects that stresses and shocks might have. Such strategies result in debt and asset depletion, which leads to negative poverty spirals and poverty traps.

4.6 Impact of the EFSP on food security

An overview is now given of the perceptions of the project participants on the project activities and outcomes. Next the participants' perceptions are compared over the three project components (community gardens, homestead gardens, and the broiler projects).

4.6.1 Perception of project participants on project outcome

The outcome of these projects is measured in terms of different aspects of the livelihoods of the beneficiaries, namely changes in diet, access to resources, the introduction of new crops and household income levels. Table 9 presents an overview of the perceived project outcomes as indicated by the respondents.

Table 9: General project outcomes

	Share of households indicating changes (% yes)
Diet
Changes in daily consumption (n = 386)	81
More traditional foods (n = 387)	80
More diet diversity (n = 389)	86
More healthy diet (n = 386)	88
More processed food (n = 370)	28
Crop diversity
Introduction of new crops	35
Access to resources
Access to land (n = 272)	70
Access to water (n = 219)	55
Access to farm technology (n = 289)	76
Access to credit (n = 150)	15
Access to information (n = 320)	88
Access to markets	45
Income
Increase in income (n = 370)	46

Respondents seem to be positive about the project's outcome. When looking at changes in diet, more than 80% of the respondents stated that their daily diet has changed since participating in the project. Most respondents perceive their diet to be healthier and more diverse than before.

Beneficiaries also experience positive effects on access to resources, especially access to information, farm technology and land. Yet less than one-half of the respondents feel that their access to markets has improved, while only 15% experienced an improvement in access to credit (Figure 6). However, these overwhelmingly positive perceptions should be put in perspective as fewer than one-half of the respondents find a positive effect on their household income. Furthermore, project involvement has actually led to the cultivation of a new type of crop for only 35% of the respondents. This means that most beneficiaries must already have been involved in crop production and project involvement has not really increased their production or income.

Figure 6: Perceived impact on access to resources

4.6.2 Perceptions of project participants by project component

Community gardening is the most common activity in all of the projects. Of all respondents, 360 (92%) are involved in community gardening; 90% of them indicating that they have received training in vegetable production while 5% indicated that they have received training in fruit production. Table 10 presents an overview of the share of beneficiaries experiencing a positive outcome. It is clear that a large share of beneficiaries do experience positive outcomes from the community garden project. Increases in social capital, consumption, production, and dietary diversity are experienced by the largest share of the beneficiaries. Table 10 also shows that one-third of all interviewed beneficiaries indicated that they received support for homestead garden production.

Table 10: Perceived outcomes of crop support activities

	Share of households indicating changes (% yes)
	Community garden (n = 360)	Homestead garden (n = 127)
Increased production	92	97.5
Increased income from crop production	75	82
Increased consumption of fruits and vegetables	93	96
Increased crop diversity	84	86
Increase diet diversity	90	90
Better soil management	89	91
More time for children education	75	69
Stimulated for other economic activity	65	63
Increased social capital	95	88

Similarly, support for the homestead gardens has highly positive outcomes on most livelihood aspects of the beneficiaries. The perceived outcomes were not different over the study regions and districts.

The final activity discussed is the involvement in broiler production (Table 11). From all respondents participating in broiler production, 97% indicated that they have received training. On average members spent R36 per month (SD: 56) on the broiler unit while they report to have received R164 per month (SD: 154) in income.

Table 11: Perceived outcomes of the broiler production activity (n = 69)

	Share of households indicating changes (% yes)
Increased number of chickens	86
Increased profit from chicken production	77
Increased meat consumption	83
Better soil management	87
Stimulated for other economic activity	76.5
More time for children education	80
Increased social capital	96

4.6.3 Comparison of food security status among beneficiaries

When a participant joined the EFSP programme did not seem to be related to the HFIAS category. Table 12 compares the household characteristics and food insecurity indicators by time of participation in the ESFP programme.

Table 12: Comparison of household characteristics and food security indicators by time of participation in the EFSP

	One year or less	Between one and two years	Between two and three years	F-stat
N	46	146	159	351
Age	50.80 (11.82)	55.49 (12.49)	54.16 (13.16)	2.504*
Total household size	5.25 (2.59)	5.99 (2.89)	5.79 (2.92)	1.351
Total income earners	0.55 (0.75)	0.64 (0.86)	0.60 (0.88)	0.232
Monthly expenditure (R)	1051.39 (690.71)	914.15 (635.28)	1001.01 (891.42)	0.772
Monthly expenditure on food (R)	507.63 (261.34)	496.78 (251.81)	499.40 (283.80)	0.029
Share expenditure on food	0.53 (0.19)	0.61 (0.16)	0.58 (0.19)	3.235**
Crop index	7.33 (3.64)	7.68 (3.48)	7.66 (3.60)	0.198
Livestock index	31.53 (61.57)	41.43 (66.27)	43.20 (55.77)	0.746
Hunger index	2.67 (2.66)	2.62 (2.70)	2.45 (2.55)	0.248
HDDS	9.17 (2.17)	10.53 (1.63)	10.01 (2.14)	10.029***

Notes: *Significant at the 10% level; **significant at the 5% level; ***significant at the 1% level.

Households that recently joined the EFSP were younger, but no significant differences were found in other characteristics such as household size, number of income earners, estimates of monthly expenditure and monthly expenditure on food. Furthermore, no indications were found that number of crops produced (as measured by the crop index) or the number of livestock (measured in a livestock index) changed with time of involvement in EFSP. The only measurable effect was increased diet diversity the longer that households were involved in the programme.

5. Conclusions

More than one-half of the beneficiary households interviewed can be labelled as severely food insecure while many are living on less than $1.25 per person per day. Severely food insecure households have lower income and expenditure levels, which points (again) to the importance of income to buy food. Households in the northern region seem relatively more food secure than households in the southern region. Average diets consist mainly of cereals, and some animal products.

Food production is rather diverse, with vegetables, cereals and tubers being the most important crops produced. Production difficulties related to access to water and lack of inputs are reported by many respondents. The most important sources of income are social grants and gifts, followed by a formal salary. Unemployment is high because only one-fifth of the household heads obtain an income through working for cash or in kind. Borrowing was the most important coping strategy when shocks or stresses occur.

The outcomes of the projects are perceived as very positive by beneficiaries, and household diets and access to resources seem to have improved with project involvement for most of the beneficiaries. Beneficiaries reported improved access to information and farm technology, but not to credit and markets.

A clear impact of the projects on food security could not be established when comparing new entrants with established participants. Yet some improvement was shown in terms of diet diversity. Hence, the EFSP demonstrates slow improvement in the diet but no clear improvement of food security levels could be established (yet). This observation might be due to different constraints mentioned in the previous paragraph or due to the relative short duration of the project.

Acknowledgements

The authors would like to thank and acknowledge Stellenbosch University Food Security Initiative HOPE project for funding that enabled the research to be undertaken. The assistance of the Stellenbosch Institute for Advanced Studies towards the research is also acknowledged, as is the cooperation of the Department of Agriculture: KwaZulu-Natal.

Notes

⁹Aliber & Hart (2009) and Jacobs (2009) point out that this variation in the obtained results is because each survey focuses on different dimensions of food security (food expenditure, hunger, household food production). The use of different methodologies and the relatively long time period between surveys makes it difficult to compare across the different surveys let alone monitor the food security status of households over a long period of time (Jacobs, 2009).

¹⁰We refer to the extensive literature review on community gardens in the paper by Shisanya & Hendriks (2011) for more references on this subject.