MAP OF ZIMBABWE................................................................................................................XIV
LIST OF TABLES
LAND APPORTIONMENT IN (SOUTHERN) RHODESIA AND ZIMBABWE 62
LAND APPORTIONMENT (1968) 64
NATURAL REGIONS 65
LOCATION OF MATERIAL INFRASTRUCTURE 66
LOCATION OF COTTON PRODUCING AREAS AND CMB FACILITIES 73
SALE OF HYBRID MAIZE SEED AND FERTILISER TO SMALLHOLDERS1 (in mega tonnes) 78
NOMINAL PRODUCER PRICES FOR MAIZE AND COTTON1 (in Zimbabwe Dollars/Cents) 80
ANNUAL GROWTH RATES OF OFFICIAL PRODUCER PRICES
TOTAL PRODUCTION AND SALES OF MAIZEFROM THE COMMUNAL AREAS 1975-1991 84
COTTON PRODUCTION FROM THE COMMUNAL AREAS 1975-1991 86
HOUSEHOLDS AND TOTAL POPULATION IN MOUNT DARWIN DISTRICT, 1969-1992 102
STANDARD RECOMMENDED ALLOCATIONS OF LAND AND STOCK UNDER THE NATIVE LAND HUSBANDRY ACT 136
TOTAL INDIGENOUS AFRICAN POPULATION IN SOUTHERN RHODESIA 171
POPULATION DEVELOPMENT IN MOUNT DARWIN DISTRICT, 1903-1992 171
REGISTERED MAIZE SALES FROM AFRICAN CULTIVATORS, 1937-1965 (in 202lb/91kg bags) 174
MAIZE SOLD BY AFRICAN CULTIVATORS TO GRAIN MARKETING BOARD (in 91kg bags) 179
REGISTERED MAIZE SALES FROM SMALLHOLDERS (in 91kg bags) 184
COTTON SALES TO COTTON MARKETING BOARD (in tonnes) 188
MAIZE SALES TO GRAIN MARKETING BOARD (in 91 kg bags) 215
AGRITEX' MAIZE CROP FORECASTS* FOR KANDEYA COMMUNAL AREA (in 91 kg bags) 217
MAIZE SALES TO GRAIN MARKETING BOARD: SHARE OF TOTAL DELIVERY FROM THE RELEVANT SUB-SECTOR (in 91 kg bags) 220
MAIZE SALES TO GRAIN MARKETING BOARD: AVERAGE SALES PER HOUSEHOLD (in 91 kg bags) 221
DISTRIBUTION OF TOTAL MAIZE PRODUCTION AND TOTAL INCOME FROM CROP SALES IN KANDARE VILLAGE 225
DISTRIBUTION OF TOTAL INCOME FROM CROP SALES IN KANDARE VILLAGE 228
HOUSEHOLD LABOUR AND MARKETED FARM PRODUCTION 236
STAGE IN LIFE COURSE AND MARKETED FARM PRODUCTION 237
ARABLE LANDHOLDING AND MARKETED FARM PRODUCTION 238
LANDHOLDING, LAND QUALITY AND MARKETED FARM PRODUCTION 239
POSSESSION OF DROUGHT POWER AND MARKETED FARM PRODUCTION 241
ACCESS TO FARM EQUIPMENT AND MARKETED FARM PRODUCTION 241
ACCESS TO COMMODITISED INPUTS AND MARKETED FARM PRODUCTION 242
CONTACT WITH AGRITEX AND MARKETED FARM PRODUCTION 243
QUALITY OF MAIN HUT BY ECONOMIC STRATA 261
STANDARD OF LIVING BY ECONOMIC STRATUM 263
ADULT CHILDREN’S AVERAGE EDUCATION LEVEL BY ECONOMIC STRATUM 265
LEADERSHIP POSITIONS BY ECONOMIC STRATUM 270
RELIGIOUS AFFILIATION BY ECONOMIC STRATUM 273
DISTRIBUTION OF CASH CROP INCOME AND CASH FARM INCOME AMONG HOUSEHOLDS IN KANDARE VILLAGE 278
DISTRIBUTION OF CASH INCOME FROM FARM SOURCES VERSUS INCOME FROM ALL SOURCES (FARM AND NON-FARM) IN KANDARE VILLAGE 279
DISTRIBUTION OF OFF-FARM AND FARM CASH INCOME IN KANDARE VILLAGE 1989-1991 280
AGE OF HOUSEHOLD HEAD AND SIZE OF LANDHOLDING 329
ECONOMIC STRATUM AND ARABLE LANDHOLDING 330
ECONOMIC STRATUM AND QUALITY OF LAND HOLDING 331
ECONOMIC STRATUM AND CULTIVATED AREA 332
ECONOMIC STRATUM AND ACCESS TO UNPAID HOUSEHOLD LABOUR 346
SOCIO-ECONOMIC GROUP AND ACCESS TO UNPAID HOUSEHOLD LABOUR 348
SOCIO-ECONOMIC GROUP AND EMPLOYMENT OF CASUAL LABOUR 356
SOCIO-ECONOMIC GROUP AND TYPE OF CASUAL LABOUR HIRED 356
SOCIO-ECONOMIC GROUP AND EMPLOYMENT OF PERMANENT LABOURERS 358
ECONOMIC STRATUM AND ENGAGEMENT IN CASUAL, AGRICULTURAL WAGE EMPLOYMENT LOCALLY 359
CATTLE OWNERSHIP IN KANDARE VILLAGE 365
ECONOMIC STRATUM AND ACCESS TO DRAUGHT POWER 367
ECONOMIC STRATUM AND OWNERSHIP OF FARM IMPLEMENTS 368
GENDER OF HOUSEHOLD HEAD AND CATTLE OWNERSHIP 373
GENDER OF HOUSEHOLD HEAD AND OWNERSHIP OF FARM EQUIPMENT 373
ECONOMIC STRATUM AND ACCESS TO COMMODITISED INPUTS 375
SOCIO-ECONOMIC CLUSTER AND CONTACT WITH THE AGRICULTURAL EXTENSION WORKER 381
SOCIO-ECONOMIC CLUSTER AND EDUCATION LEVEL 383
MAIZE BAGS PRODUCED, ANNUAL AVERAGE 1980-81 431
MAIZE BAGS SOLD, ANNUAL AVERAGE 1980-81 433
MAIZE BAGS PRODUCED, ANNUAL AVERAGE 1989-91 434
NUMBER OF MAIZE BAGS SOLD, ANNUAL AVERAGE 1989-91 436
NUMBER OF COTTON BALES SOLD, ANNUAL AVERAGE 1980-81 438
AVERAGE NUMBER OF COTTON BALES SOLD PER YEAR 1989-91 439
Mean 2,3 Median 2 439
AVERAGE COTTON SALES AMONG THE COTTON PRODUCERS 1980-81 440
AVERAGE COTTON SALES AMONG THE COTTON PRODUCERS 1989-91 440
THE ASSOCIATION BETWEEN LEVELS OF MAIZE AND COTTON PRODUCTION (1989-91) 442
TOBACCO PRODUCTION 1991 443
TOBACCO PRODUCTION 1992 443
TOBACCO PRODUCTION AND AGE OF HOUSEHOLD HEAD (1991) 444
TOBACCO PRODUCTION AND TYPE OF HOUSEHOLD HEAD (1991) 444
TOBACCO PRODUCTION AND AGE OF HOUSEHOLD HEAD (1992) 445
Chi-Square: Pearson: DF 10 Signific.: ,12932 445
TOBACCO PRODUCTION AND TYPE OF HOUSEHOLD HEAD (1992) 446
GROSS INCOME FROM CROP SALES (ANNUAL AVERAGE 1989-91) 447
CONTACT WITH AGRITEX 449
TYPE OF HOUSEHOLD HEAD 449
CURRENT PRIMARY OCCUPATION OF HOUSEHOLD HEAD 450
THE ASSOCIATION BETWEEN LEVELS OF CROP SALES (ALL CROPS) AND LEVELS OF MAIZE PRODUCTION 450
HOUSEHOLD AND PER CAPITA LEVELS OF CROP SALES (ALL CROPS) 451
TOTAL CASH INCOME 452
ECONOMIC STRATUM BY AGE OF HOUSEHOLD HEAD 454
ECONOMIC STRATUM BY GENDER OF HOUSEHOLD HEAD 454
THE ASSOCIATION BETWEEN ECONOMIC STRATUM, ANNUAL SCHOOL EXPENDITURES AND SHARE OF TOTAL CASH INCOME SPENT ON EDUCATION PER YEAR 455
TOTAL CASH FRM INCOME 460
TOTAL OFF-FARM INCOME 462
Mean 1,503.6 Median 650 463
SIZE OF LANDHOLDING AND GENDER OF HOUSEHOLD HEAD 464
LAND USED AND GENDER OF HOUSEHOLD HEAD 464
AVAILABLE HOUSEHOLD LABOUR (GROUPED) 465
AVAILABLE HOUSEHOLD LABOUR
(ABOSLUTE VALUES OF «LABOUR POWER POINTS»*) 466
HOUSEHOLD MEMBERS’ EMPLOYMENT AS CASUAL, AGRICULTURAL WORKERS IN THE LOCALITY: WHAT WORK 467
ECONOMIC STRATA AND EMPLOYMENT AS CASUAL, NON-AGRICULTURAL WORKERS IN THE LOCALITY 467
HOUSEHOLD MEMBERS’ EMPLOYMENT AS CASUAL, NON-AGRICULTURAL WORKERS IN THE LOCALITY: WHAT WORK 468
CATTLE HOLDING BY ECONOMIC STRATUM 468
ACCESS TO DRAUGHT POWER BY AGE OF HOUSEHOLD HEAD 470
ACCESS TO FARM EQUIPMENT BY AGE OF HOUSEHOLD HEAD 470
CONTACT WITH AGRITEX BY GENDER OF HOUSEHOLD HEAD 472
MAP OF ZIMBABWE PREFACE
The plan to make this study was conceived in 1982/83, when I was about to complete my Magister Artium dissertation on peasants and social change in Kenya. Zimbabwe had recently gained independence, and the Mugabe Government was busy attacking the discriminating regulations that had been the backbone of the racially segregated Rhodesian society. The agricultural production of the peasants showed a phenomenal increase, and the former African reserves appeared to be undergoing a transformation from impoverished labour reserves to commodity producing localities with greater degree of socio-economic differentiation. The transformation seemed to have many similarities with changes that took place in Kenya in the late 1950s and early 1960s. The transformation from a settler colony to an independent African state had in Kenya, among other things, resulted in much sharper differentiation among the peasants. My research background from Kenya would therefore provide the basis for fruitful comparisons in a study of a similar transition in Zimbabwe.
The work on the dissertation on Kenya had furthermore taught me how essential it is to have intimate, broad and multidimensional knowledge about the society one studies. I felt that literature studies combined with a localised field work of limited duration would never give me sufficient basis. Only by living in Zimbabwe over some time, could I acquire a necessarily broad and deep «sense» of the society. From 1986 to the end of 1989, I was living in Harare, working for the Norwegian Agency for International Development (NORAD).
The study itself was made possible by a three and a half years’ doctoral research grant from the Norwegian Research Council. I started on the research in 1990 and worked on it until the middle of 1995, with some long and short interruptions connected with other engagements (including giving birth to my son Øyvind). The research funding ran out half a year before that, and after draining the family resources for some months, I eventually had to find a paid job. Over the last year, I have again been working for NORAD, this time in Maputo. These necessities of life have delayed the completion of the dissertation, and certainly tested the patience of my family as well as my own stamina.
Two academic institutions have been of particular value during the research process. The Centre for Development and the Environment at the University of Oslo has provided a stimulating multidisciplinary environment and an amicable frame for the daily work. It has also offered excellent working facilities and technical assistance. I thank you, my colleagues at the Centre, collectively for the interest, encouragement and useful advice and comments you over these years have given. The other institution is the Institute for International Development Studies at Roskilde University, where I have been a distance Ph.D.student. I have benefited greatly from participating in a large number of researcher training courses organised by IDS. I have also benefited from researcher training offered by the Institute for Sociology at the University. of Oslo. In Zimbabwe, I have throughout been affiliated as research associate with the Centre for Applied Social Sciences at the University of Zimbabwe. I am grateful for the practical assistance and constructive comments I have been given by the permanent staff and research students there.
I am indebted to many people - in Zimbabwe and in Scandinavia. Above all, I am indebted to Mary Sandasi. Mary acted as my research assistant during three of the four field works, and continued working in Kandare Village during my long absences in Norway. Without her professional and patient asssistance, I would not have been able to obtain all the rich material we collected in Kandare Village. She is also a particularly nice and easygoing person, and staying together with her in Mount Darwin made the field work periods all the more joyful. Among the others I am indebted to, are the people of Kandare Village, who generously spent hours and hours with us and gave us the requested information, even though they sometimes questioned its value. Neriso and Erani housed us whenever we stayed in the village. Officials of the Department of Agricultural and Technical Extension Services (AGRITEX) and other government departments in Mount Darwin shared with me their knowledge and facilitated the field work. Juliana Kadzinga gave me the best possible introduction in the district and the village, and provided me with transport and accomodation.
I also stand in debt to my supervisors in Denmark, Mogens Buch Hansen, Henrik Secher-Marcussen and Peter Gibbon. Over the last two years, the latter two have given me invaluable criticism and advice. Peter Gibbon, in particular, has, through his great knowledge of the area and his exceptionally thorough reading of draft chapters, done far more than can be expected from a supervisor. Jan Hesselberg, Tore Linné Eriksen, Norbert Tengende, Elias Madzudzo, Tawana Kupe, Mungai Lenneyie and Roger Leys have all read and given useful - at times very critical - comments to earlier drafts of selected chapters. Roger has also «polished» the English in most of the chapters. My gratitude to them all. Above all, am indebted to my husband, Helge Rønning. He has patiently read, commented and discussed several drafts of this dissertation, from the first scetches to the final version. Over the last couple of years, he has also had to put up with having a wife whose mind was constantly engaged elsewhere. His great intellectual skills, love and patience have all been of invaluable assistance.
Having acknowledged my indebtedness to many people, I wish to emphasise that none of them should be held responsible for the possible mistakes, views or judgements expressed in this work.
Oslo/Maputo, June 1996.
1.1. SETTING THE PROBLEM
This study deals with two related topics: 1) The expansion and commoditisation of Zimbabwean peasants' agricultural production after 1980; and 2) Related processes of differentiation among the peasantry.
When Zimbabwe in 1980 obtained independence, it had an extremely uneven economic and social structure, where underdeveloped, overpopulated and impoverished African reserves existed side by side with European farming areas, towns and cities that had the most advanced production and infrastructure in Africa north of the Republic of South Africa. The African reserves functioned primarily as labour reserves. A large proportion of the households in the reserves did not even produce enough on their plots to satisfy their subsistence food requirements, and more than half of all men aged between 15 and 55 years were at any time absent from the reserves, engaged in - or looking for - migrant wage employment. Although about 65% of the total African population resided on peasant plots in the reserves, they accounted for only a negligible proportion of the country's registered agricultural commodity production.
By the mid-1980s, this picture was substantially altered: The peasants in the former African reserves - now called communal areas - had multiplied their agricultural output, and produced over 40% of Zimbabwe's cotton crop and between 50% and 60% of the marketed maize crop. Over the subsequent decade, this pattern has largely been maintained. In this study, I will try to identify which forces have stimulated and which have constrained the cash crop boom, and also assess the sustainability of the expanded peasant production.
It has become evident that this impressive commodity expansion is a very uneven one. It is largely confined to a few communal areas located in favourable agro-ecological zones in the northern part of the country, and even within those areas commodity production is concentrated to a small stratum of the peasantry. It has thus been assessed that at most 20% of Zimbabwe's peasants have benefited from the post-independence expansion in cash crop production (Moyo 1986:188). A central objective of this study is to answer the question why the agricultural expansion has been concentrated to such a small minority of Zimbabwe's peasants. This concentration of production has two dimensions, one spatial and one social. The spatial concentration of peasant cash crop production to a few communal areas located in favourable agro-ecological regions, can easily be explained by the fact that in the remaining communal areas, rainfall is too low to permit substantial production increase with the production techniques currently applied. I shall, therefore, not spend time exploring the spatial dimension. In stead, I shall concentrate on the social dimension, and investigate why commodity production is so heavily concentrated to a small stratum of the peasantry even within the cash crop producing localities.
This skewed expansion of agricultural commodity production appears to have created sharper socio-economic differentiation among peasant households than there used to be prior to independence. But the increase in cash crop production is not the only factor which has influenced processes of differentiation among Zimbabwean peasants after 1980. Given their traditionally heavy dependence upon income from migrant wage labour, two other developments are of particular importance:
Firstly, the removal of racial barriers and deliberate "indigenisation" (meaning Africanisation) of both the public and the private wage employment sectors. During the colonial period, Africans were only able to obtain jobs as manual workers or low-level office workers. The highest possible career was to become a teacher or nurse. The removal of those racial barriers has opened up unprecedented career opportunities for people who possess a certain level of education and the right connections. One effect of this has been reduced inequality between white Zimbabweans and a small stratum of the black population. However, it also brought about a sharp increase in the social differentiation among black Zimbabweans. The inequality between, on the one hand, this groups of well educated Africans in remunerative (wage or self-) employment, and, on the other hand, the peasants (as one social class), is very much greater that the socio-economic differentiation that today exists within the peasantry. But because many of the privileged employees maintain close ties to their home village, either by sustaining a plot and agricultural production there or more indirectly through supporting parents and other close relatives, their upward mobility influences the social differentiation within the peasantry as well.
Secondly, mass unemployment has since the mid-1980s emerged in Zimbabwe. Unemployment is not a new - or post-independence - problem, but it is now taking much greater proportions than before. As reliable statistics are missing, there are no generally accepted unemployment figures. It is clear, however, that in the 1980s, job creation in the formal sector absorbed not more than 10% of school leavers. And the problem has been aggravated in the 1990s, for the liberalisation of the economy under the structural adjustment programme has resulted in large retrenchments and closing down of a number of industrial enterprises. As a result, peasant households who wish to combine (female) peasant farming with (male) labour migration, find it increasingly difficult to do so. The situation is further exacerbated by severe land pressure in the communal areas, which causes widespread landlessness. In order to eke out a living, therefore, people increasingly engage in various kinds of temporary and ad hoc activities in the so-called informal sector, which in most cases yield very meagre income. The mass unemployment and growing "informalisation" of the economy are impoverishing those affected, and thus have great impact upon differentiation processes in the communal areas. Both landlessness and unemployment hit young people of peasant background particularly hard, so they make up a large proportion of the rapidly growing stratum of impoverished Zimbabweans who are excluded from the peasantry, but not fortunate enough to become proletarianised wage workers.
Little is known about the current processes of differentiation within the Zimbabwean peasantry. Among the many unanswered questions are: How sharp socio-economic differentiation is there among peasant households today? What is the inequality based on? What counteracting forces exist, and how powerful are they? Which are the most significant changes that have occurred since independence? How does the skewed expansion of agricultural commodity production relate to the broader differentiation processes in the communal areas? What are the future prospects?
The broad ambition of this study is to answer - or at least elucidate - these questions, through an analysis of the contemporary processes of differentiation among the inhabitants of a selected commodity-producing locality in one of the communal areas. In this analysis, the post-independence changes in the wage labour market will not be examined in any detail - only the impact they have upon differentiation processes among the peasantry. The focus of the analysis will be on changes related to the expansion of commodity production in the communal areas.
1.2. DATA AND METHODOLOGICAL CONSIDERATIONS
1.2.1. Required and available data
The general objective of this study is to analyse social change. More specifically, it is to analyse certain key features of social change which have occurred in the communal areas of Zimbabwe over the last 15 years. To study social change implies by necessity so study society as process. The focus of analysis is on processes of transformation and reconstruction, on the forces that stimulate and constrain those processes, and on their - principally temporary - outcomes. Such analysis presupposes diachronic data.
An underlying perspective guiding this study is the conviction that «Neither the life of an individual nor the history of a society can be understood without understanding both» (Wright Mills 1959:3). A complete analysis does not only investigate macro structures and macro level changes, or individual strategies, values and actions, but also how macro level changes manifest themselves at lower levels of analysis, how actors respond to them, and what effects their responses have for themselves, for broader social groups, and for society as a whole. Such analysis presupposes data at several levels of analysis; macro (national and regional society), meso (village and community) and micro (individual and household).
I aim, furthermore, through this study to reach certain general conclusions about the «commodity boom», the processes of rural differentiation, and the relation between the two. This presupposes data that allows for generalisations beyond one specific, single locality. And finally, to answer the concrete research questions posed, I need quantitative data on a wide range of distributions, as well as qualitative data on, among other things, historical changes, inter-relations, individual perceptions and priorities. This study, in other words, necessitates triangulation of several methodological approaches.
The actual collection of data, however, has been restricted by three factors. The first is the general shortage of historical and socio-economic data from African societies. Although there are more - and apparently more reliable - data available from contemporary Zimbabwe than there are from most other African countries, existing information is still highly incomplete and often of dubious quality. And for the colonial period, information on conditions in the African reserves is extremely scattered indeed. This situation severely limits the scope both for comparing my own findings with other data sets, and for situating them within a national context. On the other hand, these conditions also imply that it is more gratifying to do research on Africa than it is on over-documented societies in the first world. For even if a study should not be among the most brilliant and innovative ones, there is little doubt that it will provide useful new information and insights.
The second constraint is my limited command over the local language - chiShona. In virtually all communication with local informants, I have depended upon the assistance of an interpreter. Such language constraints restricts what issues one can meaningfully investigate, for our ability to grasp the concepts and lifeworld of a Zimbabwean villager without even understanding her language, is very limited indeed. In this concrete study, however, I am convinced that the language constraint does not represent any serious limitation. Firstly, because the issues investigated are not very philosophical or conceptual. Secondly, because I throughout the long period of data collection was helped by a very good Zimbabwean research assistant/interpreter who fully understood what kind of information I was searching for.
The third constraint, is the limited time and resources available in an individual research project such as this one. For that reason, the collection of primary data has been confined to one locality only. The primary data obtained through observations and interviews with peasant households, have all been collected from one village, located in the fertile Kandeya Communal Area in Mount Darwin District, about 200 km.s north of Harare. The written primary data from the field study area is generally on a more aggregated level of analysis, namely communal area or district level.
The advantages of concentrating the collection of primary data to one locality are obvious. It enables one to obtain rich in-depth information on a wide variety of issues, and to contextualise the information. Because the informants are neighbours who know a lot about each other, it also allows one to cross-check information given by individual informants. The disadvantage is equally obvious. It severely limits one’s scope for generalisation of the findings. Ideally, my study should have comprised of equally rich in-depth studies from at least 3-4 different localities. That being impossible, I have put emphasis on making as systematic as possible comparisons between my own findings and data in available secondary sources. Reports from other contemporary research, national statistics and historical studies together broaden the scope, enrich the field and assist in identifying what is unique for this locality, what is general for the communal areas as a whole, and what is representative only for the fertile communal areas in the northern part of the country.
The village was carefully selected through a process which aimed to end up with a field area that is «typical» for localities in the surplus-producing communal areas. Kandeya Communal Area was chosen for two reasons. Firstly, because it was one of the communal areas which in the 1980s had exhibited very strong increase in marketed peasant production. Secondly, because I already was quite familiar with the area and had very good local contact persons there, who could introduce me and obtain acceptance from the villagers, local leaders and District officials. The actual locality - Kandare Village - was selected upon the recommendation of the district agricultural extension services (AGRITEX). In my own assessment, Kandare was a suitable locality, which has proved to be a quite «typical» example of the surplus-producing villages in Zimbabwe.
1.2.2. Sources of primary data
Four main sources of primary data have been used in this study:
i) A comprehensive village survey;
ii) Life histories and other types of qualitative information obtained from 30 case households in the village;
iii) Qualitative information obtained through observation, interviews and conversations with other key informants;
iv) Written sources found in the National Archives and in the local archives such as those of the District Administration, District Council and Community Court.
To complement these primary sources, I have made extensive use of national statistics, available research and consultancy reports, official documents, maps and areas photos, and fiction literature.
The primary data were collected over a very long period, stretching from August 1990 to April 1994. From August 1990 to December 1991, I myself and/or my trusted research assistant were continuously collecting interview- and observational data from the village. Repeated field visits in October-November 1992 and April 1994 gave me the opportunity to fill important lacunae discovered in the material, and also pursue new issues which the process of analysis had alerted me to.
The village survey
The village survey has been the principal data source for the synchronic analyses of the skewed distribution of peasant production, of the extent of socio-economic differentiation, and the relationship between inequality and access to material, social and cultural resources. It includes background data on the households (size, ages, composition, education levels, employment careers), assets (house quality, land holding, domestic animals, farm equipment and other valuables), data on production and marketing of main crops, inputs used, other major and regular expenses, access to loans, remittances and positions of power and authority.
As much as 50% of the 160 resident households in the village were selected as units. This sample is stratified and weighed, and thus reflects a representative picture of the residents of the village. The units were selected through the following process: After first having explained the objective of the study to the villagers in a large public meeting, we sat down for 3-4 days with the village leadership and other selected key informants, who gave us key information about each household. The key information included sub-village, age of household head, household size, arable land holding, cattle holding, levels of production of the principal crops, and engagement in wage labour. On this basis, I temporarily, as a methodological aide only, classified the units into three broad socio-economic groups («rich», «middle» and «poor»). The groups were further subdivided on the basis of sub-village, age (of household head) and engagement in wage labour. From each sub-category, 50% of the households were randomly selected.
The main weakness of this sampling method, and of any sampling method that uses a village as the universe, is that it excludes a large number of de jure village members who de facto live more or less permanently outside it1. In Kandare Village, this group represented as much as 60 households, and the vast majority of them were landless and rather young people. These marginalised, young households reflect a very important aspect of the communal areas today - namely the great and ever more serious land hunger. Special efforts were, therefore, made to ensure that, although being excluded from the village survey, substantial qualitative information was obtained from some of the young absentee households.
Data were collected from the 80 sample households annually in 1990, 1991 and 1992. These repeated efforts enabled me to check the consistency of their responses from one year to another and thus improve the reliability. It also enabled me to minimise the impact of annual fluctuations (caused by variations in rainfall and other natural conditions), because I could use the average for the three years.
The survey data were meticulously collected. The responses were to the extent possible checked against observation and information from other informants. All land holdings were actually measured by myself or an assistant2. In 1990, I did most of the interviews myself, and the remaining interviews were done under my close supervision. In 1991 and 1992, however, all the interviews were made by a research assistant, and the quality of those data is lower. To improve both validity and reliability of the most «difficult» variables, such as magnitude of remittances, main expenses, inputs used, labour applied and yields harvested, survey data were between October 1990 and November 1991 collected on a monthly basis from 26 of the sample households.
The case households and their life histories
The most important data sources for the diachronic analyses have been the life histories of 30 selected case households, and written archival material. Additional, useful data were drawn from other oral sources and from secondary material. In the absence of diaries and other written sources, the life histories, like other oral accounts of past features or events, had to be reconstructed through the informants’ own recollection.
Recollection has the weakness, that it is selective and filtering. When we recall the past, we reconstruct it in accordance with our present ideas about what is important and not. It has, therefore, been important to confront the oral accounts with written sources from the relevant historical periods. Such sources are, first and foremost, reports from the colonial district commissioners for Mount Darwin, and other documents in the local archives of the District. Unfortunately, their information lies at more aggregated levels of analysis. The written sources refer to the district, or at best the African reserve, as the unit. The life histories, on the other hand, provides information at the level of household and locality. But in the absence of other sources, this archival material is still the best to check the life histories against. The fact that all the life histories were collected from the same locality, also reduces the problem of selective and filtered recollection: It allows me to cross-check much of the information, and goes a long way towards eliminating individual misrepresentations of the local reality. It does not provide the same scope, however, for correcting collective misrepresentations or misrepresentations in accounts of individual and household life trajectories.
Through the use of these oral and written sources, I have attempted to reconstruct the social and economic history of the field study area through the 20th Century. That account has some value in itself, for this history has not been written before. But more importantly, it provides context for analysing the life histories, and historical background for analysing the post-independence developments.
My interest in life histories as a methodological tool, has its basis in research from other African societies (Kitching 1980, First 1983). Reading this, inspired the hypothesis that the key to understanding contemporary rural differentiation, is to be found in the study of past wage work careers and economic dispositions of the peasant households. In the absence of detailed, longitudinal household budget studies and large statistical data sets on education and work careers, oral recollection of life histories is the best available source of information (Hagestad 1990, Bertaux 1994).
Life histories were collected from 30 households, which had been selected as cases. To be treated as cases, meant that the households were studied intensively and situated within (historically changing) contexts. In addition to their life histories, we obtained information from the case households through detailed, monthly survey interviews, observation and informal conversations. In 8 of the cases, I also engaged a household member (a secondary school student) to write a daily record of main events in the household, with particular emphasis on labour processes, division of labour, incomes and expenses. Most of these diaries cover the whole period between October 1990 and November 1991.
The case households were selected on their virtue of promising rich and interesting information. Extreme and exceptional cases, e.g. the poorest and richest households, were overrepresented. I thus make no claim that the case material is statistically representative. I will argue, however, that it is substantially representative. Given the fact that the life trajectories of peasant households in Kandare Village exhibit very limited variety, 30 cases should be safely above the «saturation» point, i.e. the point where additional cases yield very little new information (Bertaux and Bertaux-Wiame 1981:186-188).
1.2.3. Households as units of analysis
In the greater part of this study, the unit of analysis is the household. Most importantly, the analysis of socio-economic differentiation is structured around the household as the principal conceptual unit. I am fully aware that this choice has certain negative implications. As pointed out over the last two decades, not least by feminist anthropological literature, households are neither natural, nor stable units (Guyer 1981, Roberts 1991). Household does not necessarily coincide with family. African households, in particular, may frequently have non-family members who are co-opted into themselves for periods of varying lengths, on the basis of different contractual relationships. This practice, together with the prominence of labour migration, also imply that size and composition of the households tend to fluctuate a great deal over time. Moreover, the dominant concept in international peasant studies, of the household as a solidary, harmonious unit, is a gender-blind, false concept which neglects the inequality, conflicts and negotiations that take place within the households. The fallacy of that concept is particularly obvious in Africa, where «there are social units centering on an adult male with authority over land and over his wife/wives and children who often have their own separable stocks of property and authority» (Guyer and Peters 1987:207).
I have, still, found household to be a feasible conceptual unit for my analysis of peasant differentiation in Zimbabwe, for two reasons. Firstly, in contemporary rural Zimbabwe, the household (which consists of/is centred around a nuclear family) is the primary point of reference in the peasants’ social and economic strategies, as well as the principal locus of their production, reproduction and consumption. Broader socio-cultural institutions, such as the clan or extended family, still play an important role. It is, however, in most respects secondary to that of the nuclear family. The economic autonomy of married women in rural Shona society of Zimbabwe is also far more limited than studies have shown it to be in a number of West-African societies (Hill 1972, Berry 1985, Batezat and Mwalo 1989).
The second reason for choosing household as the principal unit of analysis, lies in the fact that it is in line with the whole tradition of peasant studies. It thus allows me to link up with important debates in that literature, and more specifically make comparisons with relevant research from Zimbabwe and other countries in the region.
In this study, the main weakness in using household as unit of analysis is the inherent inability to capture intra-household inequalities and conflicts. I have attempted to overcome that by specifically addressing issues of gender-based inequality as regards power and access to resources, subordination and exploitation within the households. I have, furthermore, attempted to minimise the problem of fluctuating size of the households, by using the respective households’ averages for the three years survey data were collected. During that period, we actually did not find very great fluctuations in the sizes of the households included in our sample.3
In this study, households are defined as economic rather than residential units. Absentee husbands in migrant employment are counted as members of the rural households, even though they live elsewhere for the larger part of the year. In several other contemporary studies from Zimbabwe, migrant husbands are not considered to be full household members, and only the proportion of their income that actually reaches the rural home is taken into account (Govaerts 1987, Pankhurst 1989, Stack 1992). That contribution is regarded as "remittances". For my purpose, however, it has been more relevant to take the migrant husbands' total wage income into account. It is the total household income which expresses the households' room for manoeuvre, and thus that total which is most relevant for analysing their economic strategies4. There are also additional reasons for defining households as economic rather than residential units. One is the fact that it tends to be the migrant husbands who, in their capacity as household heads, make the far-reaching decisions regarding farm production as well as uses of incomes. Another reason is the fact that this definition corresponds to the Kandare villagers’ concept of a household, and thus reduces the «translation problems» in our communication.
1.2.4. Research ethics and cross disciplinarity
The cross-disciplinary character of this study poses certain challenges of research ethical character. The sociological analysis of socio-economic differentiation and the peasants’ strategies to secure and enhance their access to material, social and cultural resources, rests on information which the villagers have given me in confidence. This information had to be anonymized in order to protect the informants. Qualitative life history data poses particular challenges, because it is quite easy to identify the individuals behind the stories. I have, therefore, avoided to refer to particular cases, and in stead presented imagined, but realistic, life trajectories that are based upon the «real» life histories. I have also found it necessary to anonymize the village where the field study was undertaken.
However, this study is not only a general sociological analysis of differentiation processes among Zimbabwean peasants. It is also an account of the economic and social history of a particular area of Zimbabwe. It would be meaningless to present that historical analysis without situating it in its real context. The sources used for that analysis are public records or oral information which cannot easily be associated with the informants, so the concrete references pose no ethical problem.
To balance between the need to protect the informants and the need to localise the historical analysis, I have chosen the following compromise: When referring to the more aggregated units of analysis, I have used the real names. Mount Darwin District, Kandeya Communal Area and Dotito Ward are the actual names. Only from the level of the village and downwards, have I attempted to anonymize the units. Kandare Village is thus an acronym, and so are all the names of kraals, households and individuals.
1.3. STRUCTURE OF THE THESIS
Chapter 1: Sets the research problem and discusses data and methodological issues.
Chapter 2: Outlines the theoretical framework for the study.
Chapter 3 and 4: Are two descriptive chapters which provide factual background for the study. Chapter 3 outlines the changes since independence in key policy areas and in production trends in the communal areas. Chapter 4 presents the village, the communal area and the district where the field study was undertaken.
Chapter 5: Provides historical background for the post-independence changes I am investigating. The social and agrarian changes in Kandeya African Reserve during the colonial period are situated in a wider national context, so the outline helps to identify in what respects Kandeya differs from most other communal areas and in what respects it is more generally representative, and also identifies trends and features which help to explain the current patterns and processes of differentiation.
Chapter 6: First analyses the magnitude of the post-independence commodity expansion in Kandeya Communal Area and subsequently identifies which factors have stimulated and which have constrained this expansion. More general conclusions is drawn on the basis of comparing the experience in Kandeya with data from other localities and from statistics at the national level. Thereafter, social changes in Kandeya related to the post-independence cash crop expansion are identified.
Chapter 7: Investigates the distributional effects of the agricultural commodity boom in Kandeya Communal Area.
Chapter 8: Explores the current social differentiation among the peasants in Zimbabwe in general and in Kandeya particular, considering not only their farm production, but also their engagement in other economic activities. After identifying the main economic strata among the peasantry, it discusses the significance of the stratification.
Chapter 9:Explores the relationship between the present socio-economic position of the respective households and their (current or past) wage labour careers.
Chapter 10: Investigates the respective strata’s (unequal) access to the essential means of production in peasant farming, and explores the strategies households use in order to secure and improve their access to them.
Chapter 11: Summarises and highlights the most important conclusions from the study.