Lodging as a migrant economic strategy in urban Zimbabwe

Abstract

The 1990s were a decade of rapid urbanisation and growth of squatter settlements in African cities. Given the severe shortage of affordable housing options, a large proportion of urban dwellers choose lodging as the only economically viable, legal shelter option. Lodging is thus an essential element of the economic survival strategy of the urban poor. Based on a longitudinal study of lodgers in a medium-sized Zimbabwean city during the mid-1990s, this article explores the economic characteristics and strategies of lodging households during a time of rapid urban growth and worsening economic conditions. The article outlines the way these changes affect lodging household composition, shelter features, economic coping mechanisms and overall vulnerability. It concludes with a brief examination of the present urban situation in Zimbabwe.

1. Introduction

Sub-Saharan Africa has the highest urbanisation rates in the world, with major acceleration occurring during the last two decades. In 1982, the overall urban population was a mere 21.8 per cent, but this increased rapidly to 36 per cent by 2003, and was expected to reach 50 per cent by 2030 (Kessides, 2005; UN-HABITAT, 2004–2005). Between 1990 and 2003, sub-Saharan Africa's urban growth rate of 4.6 per cent was almost double the population growth rate of 2.5 per cent (Kessides, 2005, based on World Development Indicators, 2005). Migration plays a central role in this rapid urbanisation, with circular migration dominant in most African countries (Grant et al., unpublished, 2006).

When migrants arrive in urban areas, they have few affordable shelter choices. This situation is worsened by the phenomenal growth of African urban slums. From 1990 to 2001 the population of these slums increased by approximately 65 million people, with an average annual growth rate of 4.49 per cent. By 2001, almost three of every four urban dwellers (72 per cent), or a total of 166.2 million people, were living in slums (UN-HABITAT, 2004–2005).

In sub-Saharan Africa, the 1990s was thus a pivotal period of urban transformation. National and local governments whose plans and policies were geared towards an ideal of home ownership even for the urban poor were overwhelmed by sheer numbers – urban growth driven by migration streams and natural increase rates – which defied fiscal abilities to meet even minimal basic human needs. Most migrants who entered urban areas had to rely on extended family members, friends and their own resourcefulness to secure shelter – with affordability and availability often limited to some type of rental accommodation. Whether in squatter settlements, peri-urban abandoned farms or established homes in every type of density, rental shelter emerged as the dominant supplier in African cities, mirroring the situation in most cities of the developing world.

Across the globe, the urban poor are severely constrained by lack of affordable shelter options, inflexible financial institutions, local governments that cannot keep pace with the housing and service needs of burgeoning cities and national housing policies that largely ignore the prevalence and significance of rental shelter. In many cities in developing countries, two-thirds or more of the housing stock is rental (Malpezzi, 1990). The types of rental shelter vary widely by location, culture and built form. In Latin American cities, many families rent rooms in consolidated self-help settlements (Gilbert, 1983) and most households make space for other families in order to supplement their income (Gilbert & Varley, 1991; Gilbert et al., 1993). In cities such as Bangkok (Yap & de Wandeler, 1990), Calcutta (Roy, 1983) and Cairo (Abu-Lughod, 1971), households often rent land on which to build their own shelters. Rakodi (1997) asserts that wherever access to land and home ownership is limited the majority of people become tenants. In South African cities, it was estimated that by 1996 more than a million households lived in backyard shacks or formal rooms in the backyards of other households (Crankshaw et al., 2000). Watson & McCarthy (1998) emphasise the pivotal role of the household sector, or small private landlord, in providing rental shelter for the urban poor.

As African cities continue to expand at rates often double national growth rates their coping mechanisms are strained to capacity. One area of extreme pressure is the provision of housing and basic services. Limitations such as prohibitions on squatting, high costs of new housing and acute shortages of formal rental shelter force many African urban dwellers to turn to lodging, or the informal rental of rooms and part-rooms, as the only available, affordable option. In some cities, such as Nairobi, lodging includes renting rooms in squatter settlements where illegality of tenure has not precluded commercial development (Amis, 1996).

With the exception of informal shelters erected for the sole purpose of renting, ‘lodging’ involves the creation of living quarters by the process of involution (the division of existing space). Often low-income households either add a room to their houses (usually self-built over a long period of time as resources become available) or increase the density of their existing space in order to rent to lodgers and ensure a steady, modest source of monthly income. Schlyter & Tipple (1998) have coined the term ‘multi-habitation’ for the situation where people who do not define themselves as one household share a living space that is not designed for multifamily purposes.

Space for lodgers may be created by dividing houses, rooms or part-rooms or renting out available servants' quarters, cottages or shanties. Space may be altered quickly to conform to immediate needs. For example, lodging space may be expandable to include a room, two rooms or a house, or contractible when an owner decides that more space is needed for their household or extended family members, and that this takes precedence over lodging income. With flexible terms and the absence of a written contract, lodging is the rental shelter type that is most sensitive to demand.

In some situations, however, this is a two-edged sword, since the absence of regulation may cause hardship to vulnerable low-income households. Owners can arbitrarily raise rents, levy extra charges for female household heads with children or demand that lodgers leave at short notice. In particular, lodgers with children have to worry about the possible ‘annoyance’ factor of their children's activities and noise, and many may feel they have to send their children to live with rural relatives in order to retain their urban lodging space. Divided families will also result from the limitations of living in one room, particularly when children become teenagers. In addition, where ordinary houses have been expanded into patchworks of ten or more lodging rooms, the basic water and sewage provision is woefully substandard and inadequate. These lodging houses increase immediate urban density and exert tremendous strain on urban services. They are also characterised by a complete lack of privacy and other stresses that result from severe overcrowding. Auret (1995) identifies lodgers as the most stressed urban group.

Despite these drawbacks, ‘lodging is the one adaptation that successfully circumvents, however precariously, a system characterised by bottlenecks, high and inflexible standards, bureaucracy and prices that ensure that the most modest homes remain but pipedreams for the majority of urban poor’ (Grant, 1996:254–5). Lodging is one of the most popular choices in rapidly expanding African cities with acute housing shortages. For recent migrants, it offers an affordable urban locale while they seek employment. Since many urban residents maintain strong rural ties, lodging provides the flexibility to accommodate this mobility. Although many lodgers have extended periods of urban residence, long-term plans do not usually include either home ownership or retirement in urban areas. Tacoli (1998) describes households where some members have migrated but maintain strong links with extended family in rural home areas and regularly exchange resources as ‘multi-spatial’.

In a context where new low-income houses are unaffordable for the vast majority of poor urban residents and where formal rental shelter is scarce, lodging is one of the few viable housing options other than squatting. Without the restrictions of formal leases, migrants seek out lodging that meets their particular situation, needs and aspirations. This article reveals the interface between lodging and the needs of urban migrants as they seek to negotiate an urban existence that will provide resources for both urban- and rural-based family members during a period of accelerated urbanisation and economic turbulence. It discusses the role of lodging as a critical component in urban migrants' priorities and long-term strategies.

2. ZIMBABWE AND GWERU

The 1990s were characterised by considerable growth in Zimbabwe. The 1992 census reported a population of 10.4 million, with 31 per cent in urban areas, while the 1998 estimate was 11.9 million with an urbanisation rate of 40 per cent by the year 2000. The inter-censual (1982–1992) household formation rate of 3.04 per cent indicated that households were being formed nationwide at a level close to that of the population growth. This indicator, combined with the accelerated urbanisation rate, had serious implications for shelter provision in Zimbabwean cities (Grant, 1996:248).

Zimbabwe's economic woes began even before Independence in 1980 when President Mugabe inherited slightly more than US$697 million in private, bilateral and multilateral debt. This debt, and economic advice from Washington, were among the main causes of the current economic plunge that started in late 1997 (Bond & Manyanya, 2003). The IMF/World Bank sponsored Economic Structural Adjustment Program (ESAP) in 1991/92 laid the groundwork for extreme economic hardships. These included massive formal sector retrenchments, an official unemployment rate of 40 per cent and sudden and frequent increases in the cost of basic food and services. Average real wages fell by 33 per cent between 1990 and 1997. By 1997, inflation had reached 45 per cent, some 46 per cent of all households had incomes below the food poverty line and 42 per cent of urban households were poor. Given widespread retrenchments and food and basic commodity shortages, increasing numbers of households could not meet minimal basic needs (Raftopoulos et al., 1998:11). The situation has deteriorated drastically since then, with the result that Zimbabwe has been in a spiral of ‘nearly uninterrupted economic chaos’ since 1998 (Bond & Manyanya, 2003).

The other key event was the rapid spread of the HIV/AIDS pandemic. By 1997, between 20 and 50 per cent of pregnant women were HIV-positive at 23 surveillance sites (UNAIDS, 1998). World Health Organisation data indicated an HIV prevalence of 25.84 per cent of the population in the same year (WHO, 1998). This pandemic has had a devastating social, economic and demographic impact. Breadwinners have died, households have used scarce resources for medicine and funerals, widows have had to reconfigure households, and grandparents and other extended family members have had to take responsibility for AIDS orphans (Grant & Palmiere, 2003).

This article investigates how the flexible and relative inexpensive nature of lodging affects the economic coping strategies of the urban poor. It also explores the fluidity of household composition and the extent to which workers and dependants shift rural–urban locales as the cost of urban services and staples escalates.

It is based on a longitudinal study of 100 lodging households between 1993 and 1995, when economic and social circumstances worsened considerably. The medium-sized city (1992 population 128 000) of Gweru was the site of the study. Located halfway between Harare, the capital, and the secondary city of Bulawayo, Gweru is the capital of Midlands Province, an industrial city, a key centre of a large agricultural hinterland, and an important destination for rural–urban migrants and intra-urban migrants (Grant, 1995). Gweru is a typical medium-sized Zimbabwean city in that strict pre-Independence legislation was enforced, resulting in influx controls, chronic shortage of housing and prohibition of squatting. After 1980, the government legislated that most rental housing be converted to home ownership and promoted building campaigns for low-income urban homes. Despite these efforts, municipalities could not build or convert quickly enough to meet the insatiable demand fuelled by high growth levels and in-migration, and by 1990 Gweru's official waiting list was 14 468 (Grant, 1996).

This study entailed purposive, stratified sampling to capture lodgers in all density areas: peri-urban, low-, medium- and high-density. While the sample is far from representative, the study was intended to reveal the vulnerability, household fluidity and coping strategies of a diverse group of lodgers over a three-year period characterised by rapid economic and social change. Types of lodging ranged from primitive rooms in peri-urban cottages (formerly used to house farm labourers), to rooms or part-rooms in private homes (where owners were supplementing their incomes), to lodging houses in low-/medium- and high-density areas. In high-density areas there was a booming business in overcrowded, poorly serviced conglomerate houses that filled most of the lot and housed nine to 12 families. In low-/medium-density areas, former white-owned houses had been converted to lodging and every room on the premises was rented out, including former servants' quarters. In some areas a few blocks from the CBD, almost all these spacious houses had been converted to lodging. The properties were usually characterised by poor upkeep and general deterioration (Grant, 1996).

3. Migration and Demographic Profile of Lodgers

The lodgers in the study were almost evenly divided between male (45 per cent) and female (55 per cent) respondents, although at this time male household heads were still predominant (83 per cent) compared with female household heads (17 per cent). Almost three-quarters (74 per cent) of respondents had been born in a rural area and the majority of these lodgers originated from a rural area other than rural Gweru. Only one-quarter of lodgers were born in an urban area and most of these were born in smaller towns in Midlands Province, followed by Gweru and Harare. The small proportion of urban-born lodgers would largely reflect pre-Independence residential restrictions on African families.

The study revealed that these lodgers had made an average of four lifetime moves, with an average of five moves for males and three for females. Almost half the lodgers had moved to Gweru seeking employment and altogether 63 per cent had moved either in search of work or because of a transfer. After employment and transfers, the next most important reasons for migrating were family (often wives coming to Gweru to join their husbands) and education. Slightly more than half the lodgers had moved to Gweru directly from a rural area, underlining the town's importance as a target destination for rural dwellers. Just over a third had migrated from different cities and towns and this group was almost evenly divided between migrants from the two dominant cities of Harare and Bulawayo and those who had generally originated in smaller towns. Exclusive of transfers, Gweru is an important choice for migrants who are moving up or down the urban hierarchy in search of employment. As a medium-sized city, it is still perceived as an accessible locale for rural migrants and yet is large enough to hold the possibility of employment for both rural and urban migrants.

The predominant household type was nuclear (40 per cent), 22 per cent of households included extended family members, and only 10 per cent comprised couples only. Almost one-fifth (17 per cent) were single male or female households, a reflection of the growing trend of lone females migrating to the city to look for work. Altogether, 57 per cent of lodging households included children, household size ranged from one to nine people, and the average was 3.4 people.

Most lodgers (88 per cent) rented just one room, with the rest able to afford two rooms. In many cases this constituted a small house. Table 1 shows that most lodgers had been in Gweru for more than a year, with more than half having been there for more than five years. The table also shows, however, that lodgers had moved frequently, with the vast majority having been in their present lodging for less than five years and a substantial 42% for less than one year. This illustrates the constant search by lodgers for better conditions, price, location, space or overall circumstances. Since they generally find lodging by walking around neighbourhoods or by word of mouth, search behaviour is itself quite informal and precarious.

Table 1: Length of time in Gweru and in present lodging

Time	Time in Gweru (%)	Time at present lodging (%)
Less than 1 year	5	42
1–5 years	42	46
6–10 years	14	8
More than 10 years	39	4

4. Economic Characteristics of Lodgers

In 1993, the majority of lodging households were heavily tied to formal sector employment, with 56 per cent participating in the formal sector only, 22 per cent only in the informal and 17 per cent in both formal and informal. This reflects not only the industrial nature of Gweru but also the legacy of the colonial labour migration system, with the vast majority of formal work targeted at males.

Almost three-quarters of lodging households relied on just one mode of livelihood and for the majority (75 per cent) this was formal sector employment. This would render these households particularly vulnerable as the ESAP took hold and factories began massive retrenchments. One-quarter of households relied on two modes of livelihood and 2 per cent on three modes in 1993. Although the average wages were 31 per cent higher in formal sector employment than in informal, there was a considerable range in both sectors, with some formal employment paying as little as Z$90 per month while vegetable sellers were earning Z$300–500 or more per month. In some cases, perks available from formal employment, such as housing, were far more valuable than the wage itself. The data reveals a major gender gap. Comparing average incomes for men and women, women's average incomes were significantly lower, at 55 per cent and 82 per cent of men's average incomes in the formal and informal sectors, respectively.

Table 2 shows that rents increased by approximately 21 per cent over the three-year period. During this time, rents of less than Z$100 per month became more scarce, and by 1995 almost three-quarters of lodgers paid between Z$100 and Z$199 per month. A key determinant of increased rents was the cost of utilities, which were usually included in the rent. For example, the Central Statistical Office Consumer Price Index for 1993–1998 shows that, using 1990 as a base of 100, rents, rates and utilities had increased nationwide by 219 per cent by 1995. Lodgers were committing a higher proportion of their monthly incomes to rent, averaging 27 per cent by 1995. However, the percentage of lodging households that had to pay 30 per cent or more of their earnings for rent declined to 20 per cent in 1995. Some lodgers thus took advantage of the flexibility within the system to move to cheaper accommodation and reduce their expenditures.

Table 2: Increases in rent 1993–1995

	1993	1994	1995
Average monthly rent (Z$)	103	119	125
% lodgers paying < Z$100 for rent	48	33	16
% lodgers paying Z$100–199 for rent	48	55	72
Average % monthly income spent on rent	22	21	27

Food prices also escalated but expenditures rose only slightly from Z$159 in 1993 to Z$185 in 1995. Also during this time, the total percentage of monthly income spent on food declined from 37 per cent to 23 per cent. Although most lodgers admitted that they cut back on food if they were short of funds at the end of the month, this is still a significant decrease in the proportion of income spent on food. Indeed, the CSO's Consumer Price Index (1993–1998) showed that food prices increased by 389 per cent between 1990 and 1995.

Total household expenses for lodgers increased by 35 per cent over the three-year period, from Z$324 to Z$437. In 1995, the CSO calculated the Total Consumption Poverty Line (TPL) (which includes food as well as non-food minimum need requirements such as housing and transport) for the urban Midlands as Z$235.95 per person per month. For lodgers, total household expenses absorbed 78 per cent of incomes in 1993, but this dropped to 67 per cent in 1995. However, by 1995 one-fifth of households had expenditures that exceeded income. Some households were awaiting a pension payout after the death of the wage earner, while others were hoping to receive help from extended family members. Whether they were scraping by on savings, short-term high interest loans or the goodwill of neighbours and relatives, these particular households were extremely vulnerable. Unless they could secure funds immediately, they would most probably be forced to leave their lodgings and possibly return to the rural areas.

The CSO's Consumer Price Index indicates that by 1995 the entire consumer package had increased by almost 370 per cent (based on a 1990 level of 100). Structural adjustment programme conditions included the removal of state subsidies on food, medical care and education. Food and medical expenses increased by 389 per cent and 411 per cent, respectively. There were also sharp increases in medical care costs coinciding with the escalation of the HIV/AIDS pandemic.

The next section examines how lodging households attempted to deal with retrenchments and sharp and sudden increases in basic commodities. It explores strategies of joining informal sector activities, encouraging more household members to work, increasing the number of modes of livelihood and reinforcing the safety network role of rural homes and extended family members.

5. Lodging Household Incomes, Earners and Modes

Faced with galloping inflation and frequent increases in everyday expenses, the economic strategies of these lodger households became critical to their survival and many coped by increasing their household incomes. Average household income increased from Z$464 (US$69) in 1993 to Z$954 (US$110) in 1995, as shown in Figure 1.

Figure 1: Average monthly household income

Figures 2 and 3 show two ways in which households increased their monthly incomes. One way was to increase the number of earners. In 1993, almost four-fifths of households had just one earner, but this decreased to just over half by 1995. The proportion of households with two earners doubled, from 20 per cent to 40 per cent over the three-year period, and the proportion with three earners increased from one to eight per cent. Another way was to increase the number of modes of livelihood. In 1993, 73 per cent of households relied on just one mode of livelihood and 25 per cent relied on two modes (Figure 3). This distribution changed significantly over the next two years. By 1995, only 40 per cent of households relied on one mode and 40 per cent on two modes. In addition, the proportion of households engaged in three modes grew from 2 per cent to 20 per cent. A number of households (37 per cent) added self-employment activities. Some retrenched husbands joined their wives in tailoring or selling vegetables at small stands. Almost one-fifth (17 per cent) of households added more wage earners, which in many cases included wives, youth or extended family members. A small number of households sublet their limited space and rented out to other lodgers for added monthly income. These changes are clear indicators that households quickly realised that they had to expand their economic base to provide some insulation against the shocks of retrenchments and high inflation.

Figure 2: Lodging households by number of earners

Figure 3: Lodging households by number of modes of livelihood

Opportunities for self-employment were more limited for lodgers than for home-owners. Lodgers usually did not have access to yards for gardening or for establishing workshops, nor did they have the option of subletting their lodging space. In addition, as Schlyter (2003) notes in her study of lodgers in Chitungwiza, lodgers are constrained by lack of security and lack of neighbourhood networks for small businesses, since they tend to be highly mobile.

Short-term strategies were varied and included the following:

• Sending wives to rural areas to farm. This would provide households with food and extra income. Young children often accompanied their mothers, and school fees were less expensive in rural areas.

• Taking in lodgers (i.e. subletting) for extra income.

• Borrowing from friends (50 per cent) and family members (26 per cent).

• Making handicrafts to sell in Botswana and buying goods there for resale back in Zimbabwe.

• Engaging more household members in wage and self-employment activities.

• After a male wage earner was retrenched, expanding the wife's self-employment micro-enterprise.

Long-term strategies also covered a wide range but the two most prevalent included securing a better job or wage employment (34 per cent) and building their own house (34 per cent). Obviously lodgers who were hoping to secure better paying jobs did not think retrenchments would continue, or believed that they would still obtain decent paying work. Building one's own home is an important goal, but most lodgers were not aware of the cost of even the cheapest new houses. Other long-term strategies included doing more in the informal sector (11 per cent), moving to rural areas to farm (9 per cent), forming a cooperative (3 per cent), taking courses (3 per cent) and relying on savings (2 per cent).

Another key immediate economic strategy was to reduce expenses. Almost two-thirds of households admitted to cutting back on food. Many would substitute tea for a meal, and would cut back on the number of cooked meals to conserve both food and fuel. For the majority, meat became a luxury item. As transport costs increased (by 1995 they were more than 200 per cent higher than in 1990), workers started to walk into the city to go to work rather than taking buses or combis (informal mini-bus taxis). This was no small sacrifice, since it entailed rising very early in the morning and walking for an hour to an hour-and-a-half to get to work. Not only did this add two to three hours to the working day, it further exhausted workers and left them with much less time for their family. It also wore out shoes, another expense many could ill afford. They made do by stuffing their shoes with rags or newspapers where holes had been worn in them.

Some lodgers moved closer to work and others moved to rooms that did not have the extra expense of electricity. In the winter months, wives were sent out to the bush around the city to bring back firewood for household use and for sale. Household composition was often altered (see below). Later in the 1990s, children were kept out of school to save on school fees. This did not appear to be the practice for this group of lodgers at the time of this study, although some households were using the option of cheaper rural schools.

6. Household Composition

Despite the fact that the majority of lodging households were limited to one room, household sizes increased over the three-year period, as shown in Figure 4. Smaller one- to two-person households decreased from 34 per cent to 24 per cent and three- to five-person households decreased slightly from 51 per cent to 44 per cent. Disturbingly, households of six to ten persons more than doubled, from 15 per cent to 32 per cent.

Figure 4: Changing household size

In 1993, just over half of the households were split between their rural homes and Gweru. With wives and children away for months at a time, extended family members could come to Gweru and look for work. The lodging household can expand and contract more easily than others to meet priorities. For example, it can easily move to larger or smaller living space as required. The presence of extended family members who could contribute to the household was also important for vulnerable female household heads who had only one mode of livelihood and many dependents to support. Some female household heads lost access to rural homes through divorce or abandonment by their husbands. This diminished their support networks and access to food and a different locale, especially for their children. By 1995, one-fifth of households had at least some family members in the rural areas, including children in rural schools.

Just over three-quarters of lodging households had children or youth present. Youth who had finished school were encouraged to try to find work or to help out with micro-enterprises. Extended family children, including orphans, added to the dependency ratio of households and increased the burden on income earners to meet the basic needs of all household members. The presence of orphans indicates the heightened impact of the AIDS pandemic. Lodging households were almost evenly divided between those whose dependency ratio worsened (34 per cent) and those for whom it improved (32 per cent). One household increased in size from one to ten members, only one of whom was from the extended family. This indicates that the entire family moved to Gweru to join the male wage earner once he had become established, although the wife continued to commute back and forth in order to farm. In some cases, original lodgers had shifted back to their rural homes and had let adult sons or daughters and their families take over their lodgings. The rural areas were important default locales for those lodging households that needed a place to send children for school or whose retrenched members decided to farm and to use that food and income to supplement the family's urban existence.

7. Conclusion

This article has analysed the range of strategies of a small group of migrant lodgers during a critical period in Zimbabwe. The first major effects of the ESAP were felt in the everyday lives of the urban poor through retrenchments and unemployment; increased health and school user fees; increased housing, service, transport and basic commodities costs; high inflation; and the overall stress of everyday struggles. In addition, the social and economic costs of the rising HIV/AIDS pandemic were felt in the loss of wage earners and savings and the reconfiguration of households, often in the form of taking in orphans. By 1995, many in the group of lodging households (one-fifth) whose expenses exceeded income had endured the shock of the death of wage earners and found themselves in an extremely vulnerable position. The urban poor are particularly vulnerable to any calamity – whether retrenchment and unemployment, abandonment through death or divorce, unforeseen major expenses related to illness and death or incomes that cannot match accelerated costs and inflation.

In response, some households moved elsewhere in Gweru or to other cities, and others moved back to rural areas to escape the cash-intensive existence of city life. However, those who remained (which most probably included only the more successful adaptors) provide some indication of strategies under difficult and precarious economic conditions.

Gweru remains an important destination for migrants from rural areas and smaller towns as well as from the larger cities of Bulawayo and Harare. Many migrants were attracted by Gweru's accessibility and its reputation as an industrial centre with jobs. This group of lodgers was almost evenly divided between relative newcomers (who had been there five years or less) and long-term residents, and almost all lodgers moved frequently between lodgings.

Among this group of lodgers, there was originally a heavy reliance on one wage earner and one mode of livelihood. When retrenchments and unemployment took hold, households were forced to act quickly and strategically to ensure their continued urban existence. Strategies included engaging extended family members as extra agents to earn cash or to help in the household while others went out to earn. Many households had divided families, with the return of wives and children to rural areas to farm and attend cheaper schools. This was the result of crowded lodgings, lack of access to yards and water for gardening and the high costs of urban food and school fees. The separation of families would have serious long-term consequences, particularly as prices of transport rose and families were less able to move easily back and forth between urban and rural locales. One could argue that colonial migrant labour system as a cause of separation of families has been replaced by the economic stringency driven by the ESAP. A longitudinal study of migrants based in Harare found that by 1994 only half as many migrants as in 1988 felt that their future lay in Harare, indicating that in-migrants felt less secure about urban life, employment and earning potential (Potts, 2000).

Lodging households in Gweru also encouraged more household members to earn cash, whether in wage labour or self-employment. The strong evidence for an increase in both the number of earners and the number of modes of livelihood indicates successful strategies. More subtle strategies were reducing food consumption, taking children out of school or sending them to cheaper – and lower quality – schools, and cutting back on transport and other consumer goods. This is a slippery slope that will have serious effects on health, quality of life and future prospects (especially for children) over the long term.

Finally, household composition was transformed, most peobably as a result both of calamities and longer-term strategies such as separation of families, and inclusion of extended family members' earning efforts. Immediate and longer-term strategies were working for many households and lodging played a key role in their ability to adapt quickly.

The present situation in Zimbabwe – slightly more than a decade later – is dire in the extreme. Three-quarters of the population is unemployed, the economy shrank by almost 50 per cent between 1999 and 2005, most people rely on remittances from the three to four million Zimbabweans living abroad, and at 1200 per cent the inflation rates are the highest in the world (Economist, 23 September 2006). In 2002, the WHO reported HIV infections at 33.7 per cent and also recently reported that Zimbabwean women now have the lowest global life expectancy at 34 years, compared with males at 37 years (Telegraph, April 2006).

The current situation was worsened considerably by the government-instituted operation called Operation Restore Order (Operation Murambatsvina), a major national urban demolition and eviction programme begun in May 2005 to ‘clean up’ its cities. Tibaijuka (2005) estimates that 700 000 people lost either their homes or their livelihoods or both and indirectly another 2.4 million people were affected. Homeowners were forced to demolish rooms or entire houses that were declared illegal, forcing massive displacement, family separation and increased vulnerability. Most of those affected lost most or all of their assets, plunging lodgers and homeowners into further destitution. At the time of her report, Tibaijuka (2005) witnessed thousands of people sleeping in the open either in the rubble of their destroyed homes, in rural areas or in official transit camps. Within this context and with the ensuing state of fear of soldiers, police and officials, only lodgers who are lucky enough to have ‘legal’ shelter will have a chance of remaining in urban areas in Zimbabwe.

The author wishes to acknowledge the generous support of the University of Calgary URGC (University Research Grants Committee) for this research.

Additional information

Notes on contributors

Miriam Grant

Associate Professor, Department of Geography, University of Calgary, Canada.