Migration and the changing social economy of Windhoek, Namibia

Abstract

This study focuses on the economic strategies employed by poor urban households in Windhoek, Namibia. It is based on the findings of a household questionnaire survey conducted in Windhoek and qualitative case studies collected in both Windhoek and the rural northern regions of Namibia. The central argument of this article is that rural–urban migration in Namibia is not unilinear but involves a complex relationship between rural and urban households that is fostered by high levels of personal mobility between the rural and urban settings. These close and complex social linkages between the rural and urban sectors make it possible for people to withstand the economic difficulties associated with limited employment in the formal urban economy. Further, this study shows that a key survival factor for urban households is in fact food that is produced in the rural areas.

In order to survive, we must be able to move–this is the way it is now.

(Migrant in Windhoek, 2000)

1. INTRODUCTION, CONTEXT AND RATIONALE OF THE RESEARCH

1.1 Introduction and context

The population of Namibia's capital city, Windhoek, has been growing at an annual rate of 5.4 per cent in recent years, the largest annual growth rate in its history. The 2001 population of Windhoek was about 224 000, which represents almost half of all urban residents in the country (Municipality of Windhoek, 2001). While not unique to Namibia, rapid urbanisation in the context of slow economic growth poses makes it difficult for urban managers to cater adequately for a growing population's economic and infrastructure needs. It is within this context that this research considers the welfare of relatively poor migrants to Windhoek and how they survive under difficult conditions.

Most of the population growth is taking place in Katutura, a large area to the northwest of the city, previously designated an African township, where about 60 per cent of the city's population live on about 20 per cent of its land (Pendleton, 1998). It is estimated that the population of Windhoek will double between 2000 and 2015 as a result of both natural population growth and rural–urban migration (Municipality of Windhoek, 1996b, Vol.1:20; Frayne & Pendleton, 2001, |Frayne & Pendleton, |2003). Of total migration to the city between 1990 and 2000, more than two-thirds has been to Katutura (Municipality of Windhoek, 2001, vol.1:62). Moreover, some 67 per cent of migrants in the sample in this study had moved to Windhoek since Independence in 1990, with 42 per cent having arrived in the period 1996–2000. The majority were born in the rural north of Namibia, with 79 per cent of the first generation migrants in the sample coming from the central rural north (Oshiwambo-speaking regions), which is more than double the percentage of non-migrants with the same mother tongue.

Windhoek is by no means an exception, as urbanisation and increasingly complex forms of mobility are on the rise throughout southern and eastern Africa. This article is based on a study conducted by the author in Katutura and rural districts of Namibia in 2000 (Frayne, 2001). The study is important for understanding the changing social economy of Namibia, and the findings resonate with the challenges urban managers face within the region.

1.2 Rationale for the research

It has been argued elsewhere that prior to Namibian Independence in 1990 both stabilised residents and contract migrants faced enormous economic pressures (Frayne, 2001). Recent studies show that the contemporary situation for urban residents is not much improved (Simon, 1991; Pendleton, 1991, 1996, 1998; Peyroux & Graefe, 1995; Pomuti & Tvedten, 1998; Frayne & Pendleton, 2001). Although employment opportunities have broadened with Independence, the sheer volume of urban growth appears to negate the potential benefits for the urban poor (Pendleton, 1998:72; Hansohm, 2000). A survey undertaken in the informal areas of the city reported an unemployment rate of 46 per cent among household heads (Peyroux & Graefe, 1995). The growth in the informal economy is largely in response to the real constraints on employment in the formal sector (Norval & Namoya, 1992; Pendleton, 1996).

This tension between migration, urbanisation and urban poverty has been variously described as an ‘urban crisis’ and conceptualised as a transfer of rural poverty to the urban context (Pile & Mooney, 1999:1; Pomuti & Tvedten, 1998; Tvedten & Nangulah, 1999). Moreover, vulnerability and deprivation are increasingly viewed as an urban problem, which is more severe than the situation in the rural areas (Pomuti & Tvedten, 1998:122). Devereux et al. (1995:41) make the following observations with regard to urban poverty:

Competition for employment is fierce, wages are low, and many [unskilled shanty dwellers] are forced to eke out a subsistence in the urban informal sector. The informal sector itself is underdeveloped, with an overemphasis on petty-commodity trading, which, in many quarters, has reached saturation point. Urban poverty is thus a growing phenomenon in Namibia and the situation is likely to deteriorate further if employment opportunities are not created in both rural and urban areas.

At face value, this line of argument appears to be supported generally by the data. For example, in 1991, some 67 per cent of migrants reported the lack of employment as a ‘serious problem’ they faced in Windhoek (Pendleton, 1991). In the same survey, 70 per cent of the sample reported food shortages as a ‘serious problem’. However, although consistently high unemployment rates are reported among households in Katutura, and migrants face the highest levels of unemployment in the city (Municipality of Windhoek, 1996a, 2001; Pendleton, 1996; Pomuti & Tvedten, 1998), data also show that on aggregate poor urban residents are not as vulnerable as they were prior to Independence. In a comparison of household data collected from Katutura between 1991 and 1996, Pendleton (1998) reports a decrease in the proportion of households that consider food to be a ‘serious problem’, from 70 per cent to 30 per cent. In addition, similar decreases have been reported in the problems associated with debt and health. Using the same indicators and sample population, this study found a further decrease, with only 11 per cent of respondents reporting food as a ‘serious problem’.

How is this apparent contradiction to be explained? Unemployment is high, rural–urban migration continues, and yet respondents themselves report lower levels of hunger than was the case only ten years ago. The stated improvement in the food situation from 1991 to 2000 leads one to the hypothesis that this hidden ‘income’ is most probably in the form of food. In addition, the primary source of migrants is the rural north, and Owamboland in particular, where land continues to be used productively. Food transfers from rural households to migrants during the colonial era were not documented, but this study and Pendleton's (1991, 1996) recent work confirm that this is a new factor among both migrant and non-migrant households. In addition, migrants have, since Independence, become a highly mobile group, making frequent visits to the rural areas each year (Pendleton, 1996; Frayne & Pendleton, 2001, |Frayne & Pendleton, |2003). Social linkages between the rural and urban areas thus appear to make for greater mobility and potential for transporting goods between households. Therefore, the apparent contradiction might be best explained by the increasing fluidity of rural–urban linkages in Namibia. This, in turn, has been made possible by the deregulation of the labour market, and the freedom of movement now possible under a new and independent government.

Section 2 that follows situates the research within its scholarly context and discusses the issues in the context of the literature on migration and urban livelihoods in Namibia. Section 3 describes the methodology of the research, Section 4 examines the transfers of food between rural and urban households, section 5 discusses urban agriculture in Windhoek, section 6 focuses on household level demographic coping strategies and section 7 looks at circular and reciprocal migration trends. The concluding section summarises key research findings and reflects on their implications for development in Namibia and the region, and for future research.

2. LITERATURE AND CONCEPTUAL FRAMEWORK

Previous scholarship has emphasised the importance of urban wages for the survival of rural households in Namibia, and more broadly within southern Africa (Wilson, 1972; Beinart, 1980; Murray, 1981; Moorsom, 1995; Potts, 2000; Crush et al., 2006). Linked to this concept of rural dependency on urban incomes is the conventional wisdom that unilinear rural–urban migration and urbanisation, as part of the development process, are increasingly replacing cyclical labour migration (Ferguson, 1996; Pomuti & Tvedten, 1998; Smit, 1998; Crush & Soutter, 1999). However, cities in Southern Africa have recently been described as being ‘in crisis’ (Baker & Aina, 1995; Potts, 1995, 1995; Potts, 2000; Moser, 1996, 1998; Todaro, 1997; Pomuti & Tvedten, 1998; Koc et al., 1999; Pile & Mooney, 1999). Limited industrial growth, rising unemployment and urban poverty demonstrate the inconsistencies of development practice, which aims to relieve rural and urban poverty through urbanisation and economic growth strategies (Todaro, 1969, 1995, 1997; Lipton, 1977, 1982). Yet despite high rates of unemployment rural–urban migration persists, as does urbanisation (Ellis & Harris, 2004).

Although still largely rural, more than 50 per cent of Southern Africa's population is expected to be living in urban areas by 2030 (UN-HABITAT, 2004–2005). At present, at least one-third or more of the urban population in much of southern and eastern Africa live below the poverty line (Kessides, 2005; Drimie et al., 2006). The combination of rapid urbanisation and poor economic performance has led to the proliferation of slums throughout the region. Nearly three-quarters of urban residents in sub-Saharan Africa live in slums (UN-HABITAT, 2004–2005).

As a result of these deteriorating conditions–and within the constraints of economies limited by declining per capita agricultural yields, structural adjustment and trade liberalisation (including export-oriented agricultural policies, reductions in wage employment and reductions in welfare policies), environmental stress, war and natural disasters–the challenge for national and municipal governments to meet even basic needs is immense (IDRC, 1997; Potter & Lloyd-Evans, 1998; Crush et al., 2006). Equally, the pressure on the poor to fill this resource gap themselves is extreme within the context of growing urban poverty. Yet migration continues apace. How then do urban residents, and in particular migrants, survive in increasingly hostile urban environments?

In contrast to the household level studies that have been done in rural areas on economic entitlements, urban food security measures and strategies have generally been considered at the city level (Dando, 1980; Sen, 1981; Rotberg & Rabb, 1983; Watts, 1983, 1983; Watts, 1987, 1991; Currey & Hugo, 1984; Bowbrick, 1986; Glantz, 1987; De Waal, 1990; Devereux, 1993, 1999; Devereux & Næraa, 1996; Young, 1996; Potts, 2000). However, it is well documented that urban poverty is often most acutely felt at the household level (Moser, 1996, 1998; UNICEF, 1998, 2000; Devereux, 1999; Tvedten & Nangulah, 1999; Barrett & Carter 2000; Mougeot, 2005). Moreover, the most direct and possibly most threatening consequence of poverty is limited or threatened food security and consequent hunger, despite adequate levels of food security being reported at the city scale. Urban poverty reduction strategies generally aim at increasing productivity within the manufacturing and retail sectors (that is, increasing employment opportunities). Yet with persistently high levels of urbanisation and limited economic opportunity, vulnerability to hunger and its associated problems is not adequately addressed in the majority of the urban centres of the Third World (Drakakis-Smith, 1990, 1991, 1995, 1997; Moser, 1996, 1998; Todaro, 1997; UNICEF 1998; Potts & Mutambirwa, 1998; Koc et al., 1999; Mougeot, 2005). Recognising the failure of the formal economic and urban sectors to provide the levels of service and employment required to address increasing poverty in much of the Third World, the international and local development and research communities have drawn into their ambit the question of how urban populations feed themselves under constrained and difficult conditions (Mougeot, 2005).

The conceptual framework for this research is the centrality of mobility as the mechanism underpinning urban livelihoods. In this framework both urban and rural households share in the livelihood process, with resource flows moving from urban to rural and from rural to urban contexts. This study examines the pathways used by individuals and households to navigate the increasingly complex social and economic terrain which is contemporary Namibia and, more broadly, southern Africa. The interactions between rural and urban systems are continuous within this framework, suggesting the evolution of a more integrated rural–urban social and economic system that moves beyond the traditional bounds of the rural–urban dichotomy. This is what the author has termed urban–rural reciprocity, a process which increasingly underwrites the new social economy of migration (Frayne, 2001, 2005; Crush et al., 2006).

In the analysis of urban livelihoods, therefore, this research is situated at the intersection of three different bodies of scholarship: urbanisation and survival, migration, and economic entitlement. It contributes to an emerging theory of urban entitlement, and connects with the growing body of theoretical and empirical work on migration and survival (Watts & Bohle, 1993; Drakasis-Smith, 1997; IDRC, 1997; Moser, 1998; Devereux, 1999; Potts, 2000; Crush et al., 2006) (Figure 1).

Figure 1: Conceptual framework – reciprocal migration and livelihoods

As described by Frayne & Pendleton (2001), the migration process–which is fundamental to the livelihoods of rural and urban households–occurs through a complex interplay between the macro and micro contexts of any given migrant's geography. Although individual and/or household agency usually prevails in decisions to migrate (except in cases of forced migration due to disaster, including conflict), historical, economic and political/institutional contexts all contribute to the decision-making process. Similarly, micro factors such as family dynamics and needs, household welfare, health and education influence the migration process. This process is dynamic and iterative, and increasingly dramatic in southern Africa. Indeed, studies of both internal and cross-border migration show clearly that all forms of mobility are on the rise in southern Africa, and that migration is an increasingly important dimension of households' livelihoods in the region [for a full list of studies and related outputs, refer to the Southern African Migration Project (SAMP) at Queen's University, Canada (www.queens.ca/samp)].

3. METHODOLOGY

The methodology of this study used two data collection techniques, the standardised questionnaire survey and in-depth, semi-structured case study interviews. These two data collection methods complement each other by providing generalised information through the survey and more fine-grained information through the in-depth interviews. Given the importance of the sending areas in the research, the methodology was extended beyond the original proposal to include in-depth interviews in the rural areas. Using this approach to track rural–urban migrants between sending and destination points made it possible to answer the research questions posed by this study:

1. How does rural–urban migration contribute to the household's ability to diversify economic risk and opportunity?

2. Do social relations between the rural and urban areas promote urban food security and, if so, in what ways and by what mechanisms (economic entitlement)?

3. Are rural–urban links part of an established cyclical migration pattern, or do they represent a demographic and social transition from rural to urban?

This methodology was innovative in two ways. First, by combining quantitative and qualitative approaches to the question of migration, it created the methodological synergy needed to uncover the multidimensional nature of ‘the household’ as a unit of analysis (an important consideration where households are fluid and may extend across time and space). Second, the rural homes (places of origin) of migrants interviewed in the urban centres were identified and rural household members were selected for in-depth interviews in the rural areas. This helped to provide data and information on the migration and reciprocity process from both the urban and rural perspectives. In order to evaluate possible changes over time, use was made of secondary survey data from a variety of sources available in Windhoek.

Given that Windhoek is almost ten times the size of any other urban place in Namibia, and that as the capital city it provides a destination for all sectors of Namibian society, it was selected as the research locale for the quantitative survey and urban household case studies. Because most of the growth in the city occurs within the former African township of Katutura, the urban component of the research was undertaken there. Katutura is the primary destination of migrants to the city, and appears to have the strongest urban–rural linkages in Windhoek. Furthermore, it is home to more than half of the city's entire population, and represents the poorest (and most vulnerable) sectors of society. It should be noted that the name Katutura is used to refer to both the formal area of the township and the informal areas to the northwest of the city.

Within each of the various residential areas of Katutura selected for the survey, the number of housing sites was counted. The number of surveys allocated to a particular area was then divided into the number of dwelling sites, to arrive at a sample interval. An arbitrary point within each residential section was selected as the starting point. While the head of the household is considered to be the primary decision maker, to gain insights into intricate dynamics within and between urban households and their rural components it was necessary to interview other members of the household as well. For example, migration in recent years has achieved closer gender parity than was the case under the male dominated contract labour system (Frayne, 2001). This change, together with the fact that social, political and economic conditions vary according to gender in Namibia, made it crucial that the methodology was designed to make a gender analysis of the data possible (Iipinge & LeBeau, 1997). It was therefore decided to select adult respondents (those aged 18 years and older) within each household systematically.

There were 305 interviews conducted through a standardised survey which included 95, mostly closed-ended, questions. A systematic random sampling technique was used for sample selection. Questions were designed to collect information at the household and the individual levels, and were divided into five parts to collect information on demography and socio-economic characteristics, migration and household arrangements, food/commodity transfers and remittances between rural and urban households, social linkages and urban agriculture. Data were collected through local interviewers fluent in local languages.

Besides these, semi-structured in-depth interviews were also conducted with 31 urban and ten rural respondents. Convenience sampling, derived from introductions through the survey and community connections, was used to select the interviewees. Local interviewers and languages were used to facilitate communication. Questions were grouped into five sections, covering socio-economic and demographic information, migration history, rural and urban assets, food security and commodity transfers.

The research resulted in three major findings that help to explain the urban food security conundrum described in the introduction to this article, and which help answer the research questions. Despite the conclusion that urban poverty is increasing in Windhoek, the evidence from this research suggests that urban households are increasingly reliant on transfers of food from the rural areas to supplement their urban food budgets. The results also demonstrated that it is impossible to understand this household-level coping mechanism in the urban areas without also understanding the complexity of rural–urban linkages and the high degree of social reciprocity that makes them possible.

4. RURAL–URBAN FOOD TRANSFER AND SURVIVAL

The primary argument of this study is that migrants survive in the urban areas in part because of food they receive from the rural areas. Some 62 per cent of the Windhoek households sampled had received food from relatives in the rural areas, and a further 4 per cent from friends, over the previous year (Table 1). This represents two-thirds of all the households surveyed, including those that did not have first generation migrants living in them (i.e. all members born in Windhoek, which is 14 per cent of the sample).

Table 1: Incidence of household receiving food from relatives and friends in the rural areas over the past year (2000)

	Receive food from relatives		Receive food from friends
	N	%	N	%
Yes	190	62	11	4
No	115	38	294	96
Total	305	100	305	100

Even more significant is the fact that about 58 per cent of the households reported being sent food between twice and six times per year. Respondents also said they had received a wide range of products, some of which are seasonal, including cultivated and wild foods, some meat, poultry and fish (Table 2).

Table 2: Type of food people report receiving from the rural areas (2000)

Food type	N	%
Pearl millet (mahangu)	150	42
Other cereals	13	4
Meat and fish	31	9
Commercial foods	14	4
Wild foods	147	41
Total	355∗	100
∗Note that multiple respondents were possible, hence N = 355.

Pearl millet (mahangu) is the staple cereal crop in Namibia, and when asked what quantity of particular food items were sent to the household the last time they received food, the amount of millet reported was significant. Table 3 shows that 48 per cent of the respondents claimed that they received between five and 19 kg of millet the last time it was sent to them, with about one-quarter of the sample receiving between 20 and 50 kg.

Table 3: Amount of millet received the last time by people in household (2000)

Amount	N	%
1–4 kg	51	27
5–9 kg	26	14
10–19 kg	64	34
20–50 kg	48	26
Total	189	100

Besides millet, fish is important, as are wild foods. Commercial foods, which would be purchased in a store, are of very little significance (N values range from 0 to 4). Given the generally low levels of rural cash income, and widespread poverty, it is not surprising that the food the rural household has access to is available through farming or collection from the bush. The most common wild food sent to urban households is spinach, which provides micronutrients and is also culturally important. Seasonal fruit is also sent, both fresh and dried. The following testimony illustrates this reliance on the productive capacity of the land for the type and volume of food sent to urban relatives:

We send mahangu flour, beans and dried spinach, when someone is visiting them from here [Owamboland], or if they visit us. We don't send money, and the food is only sent a few times per year, perhaps five or six. This is partly because we have to rely on visitors to transport the goods to Windhoek, which is not a regular thing, and also because food is not as abundant here as we would like.

These rural–urban transfers of food further indicate the central role of migration and urban–rural links in levels of urban food security at the household level. The survey shows that the latest amount of millet received lasted nearly half of all households about one month, and a further 16 per cent of households between two to six months. Moreover, 81 per cent of respondents rated the food they receive from the rural areas as ‘important’ or ‘very important’ to the household, with a further 11 per cent reporting that the food they received was ‘critical to their survival’ (Table 4).

Table 4: Importance of food sent from the rural areas to urban households (2000)

How important is this food to the household?	N	%
Not important at all	1	1
Somewhat important	15	8
Important	50	26
Very important	104	55
Critical to our survival	20	11
Total	190	100

The importance of the food to survival is further validated by the responses given when asked why food was sent to the household from the rural areas. Of the rural respondents, 79 per cent said it is to help the members of the urban households feed themselves. Approximately 91 per cent of the urban respondents reported consuming the food they receive exclusively, with only 6 per cent using some of it for business purposes (the remaining 3 per cent was given away to family and friends). Without these food transfers, food insecurity and malnutrition among migrant households would undoubtedly be significantly higher than current levels.

5. URBAN AGRICULTURE IN WINDHOEK

It is well established that households in many cities in the world engage in urban agriculture as a means of improving food security (Koc et al., 1999; Mougeot, 2005). Mass urbanisation and a rise in urban poverty are factors central to the development of urban agriculture. It is therefore important to quantify the extent to which urban agriculture is used by households in Windhoek to contribute to their level of food security.

Urban agriculture does occur in Windhoek, despite the climatic constraints. Five per cent of the sample was involved in some form of urban agriculture, and a further 4 per cent knew of someone else in the city growing some food themselves (a total of 9 per cent). General observation in the area confirms that, although it is limited in incidence and scale, urban agriculture exists in Windhoek. On average, urban agricultural produce saves households approximately NA$60 per month in groceries that they would otherwise have to purchase from a retail outlet. In Namibia this is a significant amount of money, and indicates the potential importance of improved conditions for urban agriculture in Namibia for cutting the cost of food for poorer households.

Although less than one per cent of respondents reported keeping urban livestock (within a ten-kilometre radius of the city), it is possible that this activity is under-reported. When asked whether or not they knew of any neighbours or friends who kept livestock in the city, a little more than three per cent said that they did. Also, goats were observed on numerous occasions within the city limits. However, the fact that the municipal and health bylaws are strict about this might explain people's reluctance to say they have urban livestock. None the less, the numbers are small and these figures appear to correctly indicate the very limited extent to which people are engaged in livestock production within the urban area of Windhoek. Even with more proactive policies, the strained water supply of the city and the limited biomass available suggest that keeping livestock in the city and its environs has limited potential in the future under current systems of land use and water allocation.

Despite the potential that might exist for the expansion of this sector, the current low levels of production again point to the importance of food sources beyond the urban boundary in feeding poorer urban households. The quantity of food being produced in the city suggests that urban agriculture does not play a significant role in ameliorating urban food insecurity at the household level in Windhoek.

6. DEMOGRAPHIC COPING STRATEGIES AND SURVIVAL

Sen (1981) argued that when entitlement to food is threatened as a result of economic shock or stress, rural households employ various coping strategies to ameliorate those entitlement failures. One such strategy is to increase sharing between households. Sen's conclusions are supported by this study, which recorded limited intra-household reciprocity within the boundaries of the Windhoek urban area (less than five per cent of the sample reported borrowing food with any regularity), in contrast to the rural situation.

Rather, the findings demonstrate that there are strong links between rural and urban households, and that coping strategies are based on these. Thus, while it is not possible to transpose Sen's (1981) analysis of intra-household sharing onto the urban context in Windhoek, it is plausible to argue that urban households reduce their food gap by invoking rural entitlements that are theirs by virtue of their social links. This is an unexpected application of the entitlement approach in regard to household sharing, but is valid nonetheless and emphasises the dominance of urban–rural links in the urban survival equation.

Limited intra-household reciprocity in urban areas is therefore closely related to the substantial connection and flow of food between urban and rural households. These very networks provide the ‘social infrastructure’ that promotes links between urban migrants and their places of origin. This social infrastructure (capital) is also responsible for a number of demographic activities that help promote food security in urban areas (and vice versa in that rural areas also benefit from the transfer of money and other commodities from the urban areas).

In cases where urban households, both migrant and non-migrant, experience economic stress and strained food availability, sending children and adults to live with rural relatives is a common practice. This strategy is referred to as ‘distress migration’ in the entitlement and coping literature (Devereux, 1999:11), in describing the process of migration by rural households to urban areas in search of employment as a result of failing sources of entitlement in the rural context. However, in this study it is equally valid to talk about distress migration of both children and adults in the urban area. Seventy per cent of the sample indicated that they sent their children to stay with relatives elsewhere. The primary reason given by respondents (35 per cent) was that the children are sent away because there is not enough money to support them in Windhoek (Table 5). Of those children sent away, at least 90 per cent live with relatives in the rural areas of Namibia, and about 70 per cent stay away from their home in Windhoek for more than a year at a time.

Table 5: Reasons why children are sent to live with relatives (2000)

Reasons	N	%∗
Not enough money to provide for them	81	35
No schools in Windhoek for them	29	12
Not enough time to look after them properly	44	19
∗Column percentage not equal to 100 as multiple responses possible.

Another demographic adjustment (distress migration) strategy of the struggling urban migrants and/or households is to send adult members to the rural areas, either as returnees or, in some instances, as new migrants to the rural homestead. The burden of providing the daily food requirements for an extended household where few are employed in any significant fashion is immense. This strategy of returning adults to the rural areas therefore helps to minimise the need to ration food, sell off assets, borrow food or money, or engage in crime as a means of survival (all of which are coping and survival strategies invoked in response to failing entitlements).

Vulnerability to hunger in the rural context is determined in part by the variability and reliability of food sources (Sen, 1981; Devereux, 1993, 1999). The results of this research show that a loss of entitlement sources increases vulnerability and hunger in Windhoek. The most marginal and vulnerable urban households were those that had poor or non-existent relationships with kin in the rural areas, and few social or economic urban resources. In other words, the limitation of their social infrastructure directly curtailed their entitlement to both urban and rural resources. Typically these households comprised young, single males and were often involved in piecework and crime, or female-headed households that relied on the informal sector for their income (see Frayne, 2004, for a detailed discussion of the gender aspects of vulnerability and urban food security). Although these cases represent a small proportion of the sample, they are significant because they illustrate the importance of urban–rural links in the survival equation, while also helping to identify the most vulnerable members of society. In most cases, however, returning to the rural areas permanently was not perceived as a solution to the household's current circumstances of deprivation, and the urban area remains the destination of first choice.

7. CIRCULAR AND RECIPROCAL URBANISATION

Social networks have been described as important in understanding the demographic strategies employed by rural and urban households to ameliorate economic stress and improve food security. In this regard, the survey found that 85 per cent of the respondents are migrants to Windhoek, and only five (1.6 per cent) of the households sampled reported having no relatives in the rural areas of Namibia. The fact that more than 98 per cent of the sample have relatives in the rural areas, and 86 per cent of respondents say they visit their relatives and friends in the rural areas at least once a year, confirms the strength of social ties between urban and rural households in the country. These strong and dynamic social connections underpin the reciprocal characteristics of the Namibian social economy.

Remittances by migrants living in urban centres back to their rural households are a well-established tradition in Namibia (Pendleton, 1994; Moorsom, 1995; Frayne & Pendleton, 1998, 2003; Pomuti & Tvedten, 1998). The historical pattern of forced migration to meet the colonial labour needs of Namibia has given way to voluntary migration, precipitated by the increasing reliance of rural households on non-farm incomes (UNICEF/NISER, 1991; Pendleton, 1994). The recent decline in the formal urban economy in Namibia has reduced urban migrants' income earning opportunities, thereby diminishing their potential to send money back to rural households. Yet studies show the importance of urban earnings to the survival of rural households, especially during times of economic hardship (Pendleton, 1994; Devereux et al., 1995; Moorsom, 1995).

Given this tension between rising need and falling earning potential, it is not surprising that the percentage of households remitting money to rural households has not increased over the past ten years. Sixty-three per cent of respondents in this study said they sent no money to family elsewhere, which is the same proportion that responded to a similar question in a household survey carried out in Katutura in 1991 (Pendleton, 1991). However, the actual number of people remitting money from the urban to the rural areas has increased substantially, given that the absolute number of migrants to Windhoek has doubled over the same period. This suggests that the number of rural households that receive urban remittances continues to rise, and this source of income is therefore increasingly important. A rural respondent described the role of remittances as follows:

Hunger is not generally a problem. Even in years when there is not enough rain, we are able to get enough money to buy what we need. This comes from our relatives in town, who work and get money. Without them, there could be a serious food shortage here, especially when there is drought.

The amount of money sent to rural relatives varies, although the median category is NA$101–150 (sent ‘every few months’). Compared to the average household remittance of NA$156 reported in 1991, average amounts sent have not increased over the past decade (Pendleton, 1991). It is also significant that about 50 per cent of households remit money to rural relatives every month or at least every two or three months. The amounts of money and the frequency of remittances support the argument that rural households continue to depend on their urban counterparts for income, although the actual value of the remittances has remained static.

Placing a monetary value on food transfers from the rural areas to urban households is difficult, not least because of seasonal variation. However, it is noteworthy that during the year of the survey (2000), half of all households sampled reported that they used millet that came as food transfers from the rural areas. This food item alone is estimated to cost about NA$60 per month for a household of four, if substituted with commercial maize meal. This value compares favourably with the value of money remitted to rural households every few months. However, a greater percentage of urban households receive food regularly than remit cash to rural relatives. The economic balance appears to have shifted in favour of urban households over the past decade.

The economic values associated with demographic reciprocity are complex, but the research did record the importance of urban incomes for supporting parents and children based in the rural areas. The largest proportion of remittances (85 per cent) is sent to parents and/or children living in the rural areas. This figure highlights the importance of the social and economic ties between extended families that straddle the rural and the urban sectors. These very networks are key for our understanding of the persistence of the cyclical migration in Namibia. The increase in urban poverty and the limitations of economic opportunities available to migrants make rural–urban interdependence an integral characteristic of the Namibian social economy.

In the past, migration to urban areas was largely temporary in both legal status and practicality, serving the labour needs of the colonial system. A cyclical pattern of movement was typical (although more permanent migration with families to urban areas did occur simultaneously). Today this form of cyclical migration of labour is much less significant per se, and is being replaced by the equally mobile but more complex phenomenon of ‘reciprocal migration’.

Conventional wisdom assumes that urbanisation is both sustained and unilinear. However, this process is seen as slower in southern Africa than elsewhere, and to have been hindered by the economic slowdown of the past decade because of slow industrial growth and limited employment opportunities. While permanent urbanisation certainly continues in the region, the persistence of non-commercial agricultural production and a rise in urban poverty have set in motion a complex system of cyclical or reciprocal migration between rural and urban areas. Pressures on rural systems of production are increasing as per capita economic opportunities in urban areas are decreasing, and it appears to be this interplay of factors that is increasing the interdependency of urban and rural systems, spawning a new form of urbanisation within Namibia. This form of multilayered and complex ‘disorderly urbanisation’ is described in this research as ‘reciprocal urbanisation’ in the Namibian context. This new form of urbanisation appears to be enhanced by modern communications and transport technologies, which make the flow of information and people across geographic space easier than ever before in the country's history.

This finding challenges current theory that relies on the unilinear model of migration and urbanisation, and suggests that the urban futures of countries like Namibia are likely to be intimately tied to rural systems, and that the two will operate in a very direct symbiosis. This symbiosis will have political, social, economic and environmental dimensions.

8. CONCLUSIONS

The findings of this research are sufficiently provocative to suggest a number of important research directions. Within Namibia, a range of future research is indicated, which includes both the rural and urban sectors. At the rural end of the spectrum, the importance of rural agriculture suggests that protecting rural productivity at the household level is crucial, and this includes addressing issues of land tenure, technology and the environment.

The environmental impact of urbanisation is typically researched within the urban context, with rural environmental inquiry being carried out as an unrelated endeavour. However, migration and urbanisation might well have significant impacts on the rural environment in Namibia, with important consequences and policy implications for both rural and urban livelihoods that cut across rural and urban sectors. This research raises a potentially important question about the extent to which rural–urban migration promotes the degradation of the rural environment, and at the same time contributes to the social and economic development of the rural population.

The logic behind this question is that access to money by rural migrants in the urban areas results in return spending in the rural areas. This spending makes a range of modern goods and services affordable for rural households, which might well promote local development objectives. A good example of this is the recent rise in the number of brick houses built in the rural areas, which then reduces the reliance on local forest resources for construction. Conversely, the number of vehicles is rising in the rural areas, with significant negative effects on the environment (people drive on tracks in the bush between homesteads, and from the homesteads to town). Also, an important focus of investment for urban dwellers is rural livestock herds. Numbers of livestock rose between 1990 and 2000, and may have significant negative environmental consequences that would in turn affect urban livelihoods.

At the urban end of the spectrum, it is likely that urban agriculture will provide a significant source of urban food in the future. There is therefore a need to undertake research in Namibia into intensive, small-scale agricultural irrigation systems that are water conserving, and that can be applied at the household level in the city. In addition, ways to manage urban livestock are likely to become increasingly relevant, and proactive research in this direction is important. This work may have application to other countries in the region.

Moving to the regional scale, it seems probable that the form of reciprocal urbanisation described in this article is not unique to Namibia. Similar conditions and processes are at work elsewhere in the region, and beyond. In her most recent work on Zimbabwe, Potts (2000) suggests that migrants in the urban areas are directly dependent on their rural land, and that recent urban difficulties may have resulted in increased rural productivity. The migration and urbanisation phenomenon is sufficiently complex in Namibia to challenge modernist views of unilinear urbanisation. Potts (2000:831) concurs for Zimbabwe:

Indeed, the nexus between rural production, migration and urbanisation is too complex to comprehensively address through neoliberal assumptions about economic growth and private ownership of land.

In addition to Zimbabwe, it is likely that similar dynamics are at work in other countries in the region with similar social and economic histories. These would include Malawi, Zambia and South Africa. While South Africa is often regarded as ‘unique’ in the region, it is possible that a similar situation might be observed in certain parts of the country. Smit (1998) argues that rural ties are important in the urbanisation process for rural migrants who move to Durban.

Rural–urban linkages warrant further investigation in Namibia, given the failure of industrial growth in creating employment. If ‘reciprocal urbanisation’ is the emerging reality in the region, research is urgently needed to further uncover the dynamics at work, as these have direct implications for the way development practitioners might shape policy and practice, with poverty alleviation as a central objective.

The general issue of coping (demographic and food transfers) raises another important question in relation to these findings: what happens when sources of rural food supply are interrupted? Because of their long histories of civil war, it is likely that Angola and Mozambique offer pertinent case studies of the consequences of disrupting rural productive systems and limiting rural–urban mobility. In both Angola and Mozambique the urban boundary has been largely impermeable as a result of the sustained conflict, limiting the flow of people and food between rural and urban centres. These case studies help us to understand urban households' coping strategies when access to rural food supplies is disrupted. Urban agriculture, for example, might well play a significant role in promoting urban food security, but further investigation is required. In the case of Mozambique, now that the war is over, are people reclaiming rights to rural land and growing food there? If so, will the direction of urbanisation conform to the emerging pattern of reciprocity evident in Namibia?

The research questions posed here aim to address the degree to which the Namibian food security situation is illustrative of other countries in the region. If the findings in Namibia indicate a new form of social and economic organisation in southern Africa, as suggested by the conceptual framework in the earlier discussion in this article, development policy and planning will need to reshape its agenda accordingly. Certainly migration, and in particular the new form of reciprocal migration–the reciprocity between urban and rural households described in this study–will be central to that agenda.

Additional information

Notes on contributors

Bruce Frayne

RENEWAL Regional Coordinator/Research Fellow, International Food Policy Research Institute (IFPRI), ILRI Campus, Nairobi, Kenya. This research was funded by the International Development Research Centre (IDRC) of Canada under the Cities Feeding People Program and the Social Sciences and Humanities Research Council (SSHRC) of Canada. The author gratefully acknowledges the additional financial and in-kind contributions made by the Regional Network on AIDS, Livelihoods and Food Security (RENEWAL); the International Food Policy Research Institute (IFPRI); Queen's University (Southern African Migration Project and the Southern African Research Center); the Social Sciences Division of the Multi-Disciplinary Research and Consultancy Centre at the University of Namibia; and the international NGO Ibis (WUS-Denmark), Windhoek. The journal referees' comments are appreciated and have contributed to this article.