Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Chapter Eleven:Economics and Ecosystems: the Case of Zimbabwean Peasant Households

Will Cavendish


In the past decade, bold claims have been made about actual or impending environmental crises in Africa, such as desertification, deforestation and soil erosion. Aside from the environmental cost of such crises, it is argued that these crises are serious because they are contribute to the pauperization of peasant households by undermining agriculture and increasing the scarcity of rural resources (for example fuelwood). Explanation of these grave developments has focused on the peasant as the agent of change, whether as a consequence of peasant characteristics (eg poverty, rapid population growth), historical processes (eg confinement on marginal lands) or poor incentives (eg distorted pricing, inadequate property rights). Given the strength of statements about these issues, it is surprising to find that there has been little in-depth work examining the interaction of peasant economy and environment - indeed, the study of African peasant households has been untouched by a serious integration of economic and environmental analysis. It would therefore seem important to conduct such analysis in order to scrutinise such claims.

This chapter takes as a case study the economically marginal households of the majority of Zimbabwe's Communal Areas, and examines the interrelations between peasant economy and the environment. From this examination, three main themes are developed. The first stresses the importance of an integrated analysis of economics and ecosystems: only through such work will meaningful analysis be possible both of Zimbabwean peasant production and of environmental change and/or degradation. On the one hand, traditional economic work, by ignoring the critical role of environmental resources both as generators of household welfare and determinants of household decision-making, has failed to understand adequately the economic systems it is analysing. On the other hand, comprehending the economic factors impinging on natural resources leads to a quite different perception both of the definition of environmental degradation and its causes from that commonly presented. Second, government policies, through faulty understanding of these economic and environmental interactions, have largely failed to ameliorate the problems they were intended to solve, and sometimes have even undermined the peasant economic production system and the survival of natural resources. Finally, it will be argued that 'sustainability' is in practice an extremely difficult concept to utilize: as a result, grandiose statements concerning environmental degradation and unsustainability should be regarded with some scepticism.

This paper is organized as follows. The next section outlines the production system (and in particular the agropastoral production function) of Communal Area farmers, emphasizing both their highly income- and asset-constrained status and their economic remoteness. As a consequence, these households rely heavily on the goods and services offered to them by natural habitats/environmental resources, a reliance explored (on pages 208-14). Given the pervasive interactions between economy and environment, it is then argued that the characteristics of local ecosystems have an important role to play in explaining household decision-making. As a consequence, it is unsurprising to find that peasant households have substantial environmental knowledge, and indeed are involved in managing their local ecosystem to retain flows of important natural habitat goods and services (see pages 218-22). In the light of this analysis, the final section discusses and concludes the main themes of the paper.


Zimbabwe's agricultural areas can be basically divided into two zones. The first zone comprises the Commercial Farming Areas - large, white owned farms located in favourable Natural Regions - while the second zone comprises the Communal Areas (CAs), the region of interest for this paper. These areas, and the agricultural system within them, are a colonial creation whereby the African population of Zimbabwe was relocated in agriculturally more marginal areas. As a result, peasant households in Zimbabwe's Communal Areas are characterized by features common to small peasant producers across southern Africa. There are relatively high population densities and the average farm household is large (circa 10 members), while the average farm is not large, consisting of three hectares of arable land, often located more than one km from the household. Communal Area farms are generally located in less favourable regions for agriculture: 74.2 per cent of CA farms are in Natural Regions IV and V, which - on account of their low rainfall levels (less than 650 mm pa) - are officially described as suitable for only extensive and semi-extensive farming.

The Agricultural Production Function

The characteristic of the agricultural production system usually emphasized is that it is agropastoral: in other words, the production system comprises a combination of crops, livestock and natural resources. Thus, while crops and vegetables are grown on the three areas that can, by and large, be considered to be under the private control of the household (namely arable fields, gardens and the homestead), another critical component of agricultural production is the livestocks' grazing lands, which are characterized by communal tenure. The characteristic emphasized here, however, is the heavy reliance of the agropastoral system on non-marketed transactions, and in particular the role played by natural habitats. Not only are these natural habitats used as the communal grazing lands and thus underpin the entire agropastoral production system, but also they contain the woodland, mountain and riverine areas from which most environmental utilizations are derived.

In this section, though, we focus on the agropastoral system. First, consider the role of livestock: their importance lies chiefly in their provision of draft power and manure. Draft power is a critical input into crop production, as ploughing at the optimal time has a marked impact on crop yield (around 50 per cent) (Ashworth 1990, Cumming and Bond 1991). Households without livestock generally rely either on hand-hoeing or wait to hire/borrow livestock for ploughing after the optimal time. Manure is also an important resource: for Communal Area farms, mostly located on arid, nutrient-poor soils, regular fertilization of fields is necessary if adequate yields are to be obtained. Manure is the most common fertilizer used by Communal Area farmers (Dewees 1992). Indeed, this is the point of the agropastoral system, which involves the close interaction between cropping and livestock activities, and where the value of livestock primarily comes from its role in providing inputs to the agricultural system. Ownership of livestock - and in particular cattle - hence becomes crucial to these households' agricultural production: for example, arable yields for households with large cattle herds are triple those of stockless households.

However, the uses of cattle are not just restricted to draught power and manure. Other services they offer the household include the supply of milk, the provision of power for transportation, the production of saleable offspring, transformation into consumption items (meat, hides, hooves etc), and a vehicle for savings and investment - in other words, multiple use assets (Scoones 1992). In fact, the key role of cattle is to provide the household with non-market substitutes for inputs (draught, manure, work) and consumption goods (milk, meat) that the household chooses not to purchase in the market.

This point can be extended by considering inputs into livestock maintenance and the sources of fertilizer. With regard to the former, few households (especially those in NRs III to V) purchase livestock feed: instead, households rely primarily on forage and browse in the communal grazing areas in local natural habitats; and secondarily on the recycling of crop residues from their own fields to provide livestock sustenance (Campbell and Grundy 1991). Similarly, although (purchased) chemical fertilizer is available to households, in fact it is organic fertilizer that is more widely used, again especially in the less affluent farms of NRs III to V (Conroy 1990). The role of manure has already been mentioned. The second most important source of fertilizer is leaf litter which households collect from woodlands in surrounding natural habitats. Again, use of this resource is widespread: Campbell et al (1991a) report that, in a survey spanning three Natural Regions of Zimbabwe, 68 per cent of respondents claimed to use leaf litter, making this utilization the most significant direct transfer of nutrients from woodlands to fields. Finally, households also use termitaria, though this utilization is highly variable across Communal Areas (Campbell and Grundy 1991).

Thus, in the critical areas of agricultural production and livestock maintenance, households rely heavily on non-marketed goods - and in particular, goods that are derived from the natural habitats in the rural life space surrounding the household - despite the existence of marketed substitutes. What explains this pattern of consumption? Two overriding determinants are suggested:

  1. the extremely low cash income status of most CA households, and
  2. their economic remoteness implying the existence of pervasive market
    failures for poor rural households.

Communal Area Household Cash Incomes

The extremely low cash incomes of Communal Area households have been charted by a number of studies. The most insightful of these is a study by Jackson and Collier (1991), based on a survey during the 1984/85 season of 600 households spread over the five Natural Regions (Table 11.1). This establishes a household benchmark mean 'furl income' figure of only Z$701 per year (US$437): converting using Adult Equivalent Units, this gives the extremely low mean per capita 'furl income' figure of Z$227 pa (US$142). Even this, however, overstates average pure cash income, as both remittances and crop income include valuations of subsistence flows. Adjusting for these subsistence valuations, mean annual cash income becomes Z$421 per household (US$261), with mean per capita annual cash incomes a mere Z$130 (US$80). Finally, these average income figures disguise considerable income inequality: median incomes are less than two-thirds of these mean income figures. A corollary of these extremely low income levels is the very low asset levels of Communal Area households. While there is no comprehensive survey of assets levels in Communal Areas, Table 11.2 provides a summary of available data. There are two main points:

Table 11.1 The income structure of rural households

  Households with
income source (%)
Total household
income (%)
Mean income
from source
Crop income 95.8 50.4 368
Remittances 37.0 18.5 348
Local off-farm wages 10.7 14.7 960
Self-employment 23.1 7.9 225
Livestock income 66.8 5.4 56
Estate wage 6.0 2.5 295
Local farm wage 12.8 <1.0 32
Mean aggregate annual income data
(per houehold) (per adult equivalent)
Z$701 z$227
Median aggregate annual income data
(per household) (per adult equivalent)
Z$450 Z$138

Source Jackson and Collier 1991.

1. Cattle comprise as much as 80 per cent of total asset values: other than cattle, the average household owns only a few, mostly low value, agricultural implements, and some smaller livestock.

2. As we will see later, these assets are vulnerable, as livestock stocking strategies involve periodic die-offs. However it should be noted that, like income, the averages overstate the median asset level which is, in fact, significantly below that of Table 11.2.

These low figures are born out by a host of other studies, both country-wide and in specific locales (CSO 1988, Campbell et al 1989, MLARR 1990, Coudere and Marijse 1991, Mehretu and Mudimu 1991). Though carried out in different years with different methodologies, they all provide figures similar to those of Table 11.1. 5 This situation is replicated across Africa generally, despite the importance of assets in underpinning welfare and reducing household insecurity. However, see Deaton (1991) and Dercon (1992) for a start on this work. 6 Jackson 1989 provides data on the considerable inequalities in cattle ownership

Table 11.2 Household assets

  Households Average no. Average value Average value
with item (%) per household per item (Z$) per household
Scotch cart 34.9 0.4 314 126
Water cart 3.9 0.1 132 13
Bicycle 22.4 0.3 101 30
Ox-cultivator 31.6 0.3 68 20
Harrow 17.8 0.2 59 12
Wheelbarrow 39.7 0.5 43 22
Ox-plough 83.8 1.1 38 42
Yoke 82.2 2.1 8 17
Hoe 92.1 5.4 3 16
Sub-total       298
Cattle n.a. 7.00 287 2,001
Goats n.a. 6.67 30 202
Total       2,501

Source: Tables 1.5, 1.8, 5.1 and 5.2 in MLARR 1990 and authors own calculations

Why are income and asset levels so low? The work of Jackson and Collier (1991) demonstrated that rising incomes were directly connected to Communal Area households' ability to accumulate and diversify their number of income-generating activities. However, households are held back from accumulation and diversification by their highly-constrained nature. The economic effect, and severity, of these constraints is clearly indicated by high marginal returns to livestock accumulation: Scoones (1992) calculates the internal rate of return on cattle to be uniformally greater than 100 per cent. These constraints are obviously related in part to the low level from which these households are starting, a level which has been delineated above. However, as is discussed below, they are also a consequence of the economic remoteness of Communal Area households.

Communal Areas' Economic Remoteness

By economic remoteness, what is meant is economic distance (measured through proportional transactions costs) from markets in which significant income-raising trading can take place (ie. excluding markets within the rural life space, which have a much lower potential for raising incomes). Generally speaking, this economic remoteness is the cause of the widespread market failures that afflict peasant production and hence inflicts a substantial welfare loss on rural producers. While the literature on the existence of markets for Communal Area households is patchy, there are some characteristics affecting the existence of markets that are relevant to a majority of Communal Area households. The first is that Communal Area households are frequently physically distant from formal markets, shops, mills, clinics, roads and so on (MLARR 1990). Covering the distances from farm to services to carry out economic activities comprises considerable unit transactions costs as these trips must be undertaken mostly by foot: indeed, a study of one Communal Area found that the total energy costs of making key trips was over 25 per cent of the available household energy budget. Thus, the substantial distances separating Communal Areas from various amenities suggests market failure is a generic phenomenon (Mehretu and Mutambirwa 1992).

As far as specific markets are concerned, connections between Communal Area households and formal markets are weak (even though the same market in the rural life space may be active). For example, capital market failures are pervasive. Insurance markets for crop income failures do not exist, while less than 10 per cent of Communal Area farmers receive formal credit (World Bank 1991). Unsurprisingly, this minority is found in the better agricultural areas, meaning it is the resource-poor farmers for whom formal credit markets are failing (MLARR 1990). In order to compensate for this, informal credit mechanisms are common, but they exist firmly within the rural life space. Similarly, while on the surface the state marketing boards offer access to external markets for Communal Area farmers' agricultural outputs, in practice transportation costs to depots for poorer farmers are high, and it is households in NRs II and III that have mostly benefited from participation in these markets. If a poorer household does sell crops, this is usually to other households within the rural life space. The same pattern holds for non-agricultural goods: rural to urban sales are rare, but within-rural sales are common (McPherson 1991).

Labour markets are, in contrast, more active. Data on the components of household income demonstrated the importance of remittances from non-resident labour who, by definition, are involved in external markets. However, the potential of external labour markets to raise rural incomes is heavily circumscribed by Zimbabwe's critical unemployment problem. Within the rural area labour markets are widespread. For example, one study found 39 per cent of households hiring at least one permanent worker, while as many as 77 per cent of households had hired casual workers in the last year (Adams 1991). However, these rural labour markets do not raise incomes appreciably as wage rates are low, at Z$30 to Z$40 per month (Scoones 1992). Similarly, markets for the key agropastoral input, livestock services, are common - indeed, Muchena (1989) catalogues five payment mechanisms for ploughing services - but they are often costly and are overwhelmingly within the rural life space. In general, then, participation in income-raising markets is very limited.


Earlier, the surprisingly widespread reliance of Communal Area households on non-marketed goods and natural habitats for important inputs into the production system was noted. This has been explained by the extremely low-income, highly-constrained nature of Communal Area households, which renders them unable to purchase these inputs in the market. However these same constraints face households in all areas of economic activity: in the following section, then, the analysis is expanded to look at the wide range of natural habitat utilizations by Communal Area households. To motivate this discussion, consider the low cash incomes of Communal Area households. It is clear from the data above that if all household consumption had to paid for, income levels of these sorts would place Communal Area households near the bottom of a global poverty profile, with scarcely adequate incomes to purchase anything other than the most minimal consumption bundle. Hence, there is something missing from the conventional economist's description of these poor, economically remote, rural households: namely the pervasive dependence of Communal Area households on natural habitats.


Evidence of the importance of wild resource use can be taken from Table 11.3, which lists the wide range of different natural habitat utilizations mentioned in various literatures. As will be shown, a fundamental problem in evaluating these utilizations is the lack of proper economic data on quantities used, prices, the pattern of household use and the like. However, while not all households use all the wild resources listed, casual empirical evidence suggests most utilize many of these resources, hence the high valuation given to resources (see below) and efforts made to conserve them. Indeed, the fact that many of these resources offer multiple uses to households suggest they also provide considerable flexibility of use, in contrast to the constrained nature of agricultural production noted previously: this has not, however, been adequately researched.

The Consumption of Wild Foods

There is no definitive analysis of the consumption of wild foods. A comprehensive listing is presented in Gomez (1988), who lists species of wild cereals, legumes, vegetables, mushrooms, fruit, insects, fish and game known to be consumed by Communal Area households. However, there is no quantification of resource use and hence no indication of the significance of these items in rural diets. As the evidence reviewed below includes a mere subset of the species listed in Gomez, it must be regarded as a preliminary analysis only.

First, there is evidence that wild meats can play a quite substantial role in rural diets despite the illegality of hunting. For example, de Vos (1978) suggests that hunting of wild animals yielded 2.5 million kgs of meat per annum in the mid-1970s, more than the local beef industry. More recently, Murindagomo's (1988) study of Angwa in the mid-Zambezi finds that adults consumed an annual average of 88 kg of bushmeat per person. Further, illegal poaching brought a return of Z$8.20 per hectare, far exceeding the estimated damage to crops incurred by the wildlife on whose presence such high consumption levels are predicated. Finally, an important role is played in rural diets by the consumption of smaller mammals such as rodents, birds etc. and of fish caught in local rivers and dams (Wilson 1989b).

Table 11.3 A taxonomy of natural habitat utilizations by communal area households

Utilization or service Source of item
in household
Type of activity
Wild meat State forest areas
Wildlife areas
Income generation
Wild fruits and nuts Local woodlands
Trees in grazing areas
Trees in fields
Trees in homesteads
generation (wine
Interseasonal stock
Income generation (sale)
Input into income
Fish Local rivers and dams Consumption
Income generation
Other wild foods:
Wild vegetables
Local woodlands
Local fields
Income generation
Natural medicines Local woodlands Consumption
Income generation (for
Other wild goods:
Wild shampoo and soap
Wild salt
Tooth-cleaning materials
'African' snuff
Local woodlands
Local grazing areas
Fuelwood for:
Cooking and heating
Beer brewing
Brick burning
Local woodlands
Trees in grazing areas
Leftover wood from
other uses (especially
Interseasonal stock
Input into income
generation (brewing,
Construction wood:
Hut walls
Roofing poles and
Door frames
Storage huts
Livestock pens
Storage racks
Drying racks
Fencing (live and dead)
Tree protectors
Specific trees in local
Local woodlots
Input into housing
Wood for making:
Agricultural implements
Household utensils
Musical instruments
Hunting and fishing
Local woodlands
Trees in grazing areas
Input into production
Input into income
Input into consumption
Input into consumer
Reeds, Grasses, Cane,
Bark etc:
Thatching grass
Sleeping mats
Other mats
Storage baskets
Winnowing baskets
Transport baskets
Herding whips
Gourds and calabashes
Contour ridges in fields
Grazing areas
River- and streambanks
Trees in local woodlands
Trees in resettlement
Input into consumption
Input into income
Input into production
Input into housing
Clays for pottery:
Storage pots
Cooking pots
Rural life space Input into consumption
Input into income generation
Leaf litter Local woodlands
Trees in fields
Input into agricultural
Termite mounds Local grazing areas
Crop fields
Input into agricultural
Livestock fodder and
Local woodlands
Grazing areas
Input into consumption
item (milk)
Maintenance of capital
Indirect input into
agricultural production
(manure, draught power)
Shade Trees in grazing areas
Trees in fields
Input into agricultural
Ecological services eg
Higher soil organic matter
better, soil moisture,
reduced soil erosion
Trees in fields Input into agricultural

There is more evidence concerning the prevalence of wild fruit consumption. For example, Campbell's (1987) study covering three natural regions and three tenure systems found wild fruit to be an important supplement to diets in all three regions and tenure systems. Ninety-five per cent of respondents stated that a member of their family regularly consumed wild fruit, with most consumed by young children. In a later survey, Campbell et al (1991a) found that an even higher percentage of rural households - 98 per cent - collected wild fruit. Further, wild fruits are used not just for immediate consumption, but are also stored for consumption in the dry season and can be sold in local markets, with 10 per cent of households raising income this way. This extensive use of wild fruits has been confirmed by the studies of Wilson (1990) and McGregor (1991), with evidence in the latter study that wild fruit is more heavily used by poorer households. Finally, the importance of wild fruit is confirmed by the considerable efforts undertaken to preserve wild fruit trees.

Quantitative evidence on the consumption of other wild foods is insubstantial. There is some work (Parent and Thoen 1977, McGregor 1991) indicating that wild mushrooms are used quite heavily, although evidence is scanty. Other important wild foods are honey, wild vegetables and edible insects such as termites, caterpillars and especially 'mopane worms' (Wilson 1990, Dewees 1992). These are shown in Wilson (1989b) to form the basis of most relishes (ie the accompaniment to the stiff cereal porridge made from maize). Many of these resources are dried and stored for consumption during the late dry season, when food shortages are common, and for a variety of them, urban markets exist.

Wild species are also used commonly as medicines. Up to eighty per cent of both rural and urban populations consult traditional healers for treatment of illness and other maladies (Mclvor 1989). This heavy utilization implies substantial wild resource demands, as traditional healers use these resources exclusively: indeed, over 500 species have been catalogued as having medicinal use by traditional healers in Zimbabwe (Gelfand et al 1985). However, many households also use wild resources for their own treatment of common complaints such as headaches, colds, indigestion and diarrhoea (Cavendish, forthcoming). Finally, there are a number of other known uses about which little has been written, for example, use of certain twigs for cleaning teeth; use of wild species for soap and shampoo; collecting wild salt; and extracting gum and glue.


Contents - Previous - Next