An overview of Fast Track Land Reform in Zimbabwe: editorial introduction

Events in the last decade around the land question in Zimbabwe and the broader political context in which they have played out have been dramatic and transformational. Sparked by land occupations (locally referred to as jambanja meaning ‘violence’ or ‘angry argument’), and involving contested land expropriation and violent episodes, the process has not surprisingly proved contentious among policy-makers and commentators, nationally and internationally, and among all those who have sought to explain, justify or criticise it. With few exceptions, those who have engaged in writing or political rhetoric have tended to take positions on one or other end of the spectrum in what has been a highly polarised debate, between welcoming a reversal of a racial distribution of land – some of them bewailing the manner of implementation and its distorting of the state – and those who condemn the end, in principle, as well as the means. The fervour surrounding these dramatic events and their explanation was vastly heightened, as well as being framed by, a massively debilitating economic crisis. This was marked by a world record hyper-inflation, for the moment resolved, and by a vast shrinkage in GDP. Debate continues as to what extent the overall economic meltdown was caused by or generated declines in post-land reform production or whether and how these processes interacted (see Davies 2005; Mamdani 2008; Scoones et al.2010; UNDP 2008 for different positions in this debate). The political context was no less dramatic and transformational. A nationalist party, Zimbabwe African National Union-Patriotic Front (ZANU-PF ), dominant for 20 years was seriously challenged for the first time by a new party, the Movement for Democratic Change (MDC), and reacted with a string of repressive laws and actions. These events also arguably shaped and were shaped by the land reform; one view explored below (Alexander 2006) is that repressive mechanisms were a requirement of enacting FTLRTP.

The very intensity of the debates and the widespread international interest in these events make a strong case for more careful and detailed analysis of the Fast Track Land Reform Programme (FTLRP) without preconceived conclusions. The passage of eleven years since the FTLRP began to be implemented is another trenchant reason to review these processes. The emergence of a range of studies into what has transpired over a lengthy period provides a ‘reality check’ and an opportunity to extend debates beyond policy prescriptions and their initial implementation to an assessment of what has actually been happening on the ground as a result of the land redistribution that occurred in the early 2000s. The present collection, while rehearsing past events and identifying the social and political forces that drove them, is principally concerned with summarizing the findings of investigations of the range of outcomes. Some material from these case studies has seen publication in one form or another, others are in the works, but nowhere has such a wide range of results been brought together between the same covers. It is hoped that this collection will aid in a much-needed process of rooting the vigorous debate on land reform in Zimbabwe in the available and emerging empirical evidence.

Crucial evidence is now becoming available on some of the economic and social outcomes of the FTLRP. The contributions to this collection will be focused more on these broader questions of what diverse pictures are emerging about local livelihoods rather than looking back to explore again the big political issues about the origins of the FTLRP, although this background will be reprised in this overview. Instead the following questions will be addressed: who have been the beneficiaries of the programme and what other groups have been affected (farm workers, former farm-owners, etc.) and how have their livelihoods fared? What new patterns of production have emerged and how successful are they in terms of levels of output, productivity and impacts on poverty? What new ‘communities’ and social structures have emerged?

Findings on these topics are valuable contributions to the debate, given that diverse views about the likely outcomes of a radical land redistribution emerged at the outset as a priori predictions – of agricultural disaster or widespread abundance – matching the polarised stances taken on basically political grounds. Those who condemned the illegality and brutality of the operations were prone to see them as also (and inevitably) ill-advised economically with disastrous effects predicted for agriculture and for confidence, investment, exports and the overall macro-economy. Equally predictably, defenders of radical redistribution and the justice of widely spreading rights to land as ‘progressive’ (e.g. Moyo and Yeros 2005) tended to expect a long-term expansion of land-based livelihoods and more intensive land use, similar to what happened after the ‘Old’ Resettlement of the 1980s (see Dekker and Kinsey here). Other logically possible combinations of views on process and outcomes – for instance, that however repressive the means, increased food security and a wider spread of livelihoods might result, or alternatively, that however justified the ending of a racial basis for land rights might be, it would not deliver at the level of production – hardly seem to figure in debates.1

This particular polarisation of views about outcomes was not simply a translation of assumptions into predictions, but embodied a deep difference in the way the problem was posed. As Mudege (2008, 12) remarks about debates in the late 1980s that led to the slowing of Old Resettlement, ‘ …the political problem of lack of access to land was turned … into the technical problem of poor access to relevant knowledge’ (by the beneficiaries). Hopefully this careful scrutiny of actual outcomes can move discussion away from predetermined views of what they would be.

There are broader, comparative reasons as to why the Zimbabwe experience deserves careful scrutiny by anyone concerned with land and agrarian reform in Africa, the future of small-scale producers (or ‘peasantries’) and alternative agricultural and rural futures in that continent. First, the sheer scale of the land redistribution in Zimbabwe has surpassed that of almost all other African countries, and thus like all ‘extreme’ cases it is potentially instructive. This of course challenges Bernstein's conclusion (2003 and 2005) about comparative experience, which sees Zimbabwe as an ‘exception’ – although he also saw South Africa as an exception (Bernstein 1996). Moreover, this scale has qualitative implications: there has not only been a significant shift in agrarian structure and thus in the scale and character of different rural classes but a major transformation of the agrarian sector as a whole. From a structure where the better half of the agricultural land was in the exclusive hands of large-scale commercial producers, there has been what van der Ploeg (2008) termed, controversially and in another context, a ‘repeasantisation’ where small-scale, household-based agricultural production is now the predominant form.

Second, however, the Zimbabwe case is not characteristic of the majority of African countries, where systems of land tenure and ownership rooted in pre-colonial societies persisted albeit within an evolutionary transition to capitalism. Its direct relevance is only to that minority of countries where ‘settler-colonialism’ (the commonly used term in Zimbabwe and elsewhere in southern Africa) was the dominant pattern of agrarian system and of state. This category (see Amin 1976, Mandaza 1986) included Kenya, Algeria, Namibia, Swaziland, Mozambique, Angola and, of course, South Africa. Thus Zimbabwe's experience can be assessed as an instance of that form of transition where there has been acquisition of some part of the large, white-owned holdings and some degree of redistribution, as in all these countries mentioned, usually with some sub-division of those large farm holdings, and the special challenges that such reform poses (Bernstein 2003, 2005, Bush and Cliffe 1984, Cousins 2003, Lahiff and Scoones 2000, Moyo 1995, 2005b, Moyo and Yeros 2005). Zimbabwe can be seen as one variant of this type of transformation; and its similarities and differences make for interesting and potentially explanatory comparison. Indeed, Zimbabwe may hold out lessons for these other former settler colonies in that it is the only one which after a similar, partial land redistribution in the 1980s has undertaken a second round of reform – an option that may provide an exemplar to be followed, or avoided.

Third, of course, there are insights that can be gleaned from Zimbabwe's experience for categories of African countries other than the former settler colonies, and even other developing countries, from comparing the differences in the starting point and the context, in relation to reform of ‘traditional’, ‘customary’ or ‘communal’ land tenure systems, which are central features of rural society in most African contexts, and sometimes found in some version or other on redistributed land. In the post-independence period Zimbabwe pursued changes in the colonial version of ‘traditional’ land tenure imposed in the areas ‘reserved’ for African farming and residence, the labour reserves as they have been conceptualised since Arrighi's (1970 and 1973) seminal work. In the 1980s land administration roles were transferred from ‘traditional’ authorities to local government committees, and then under a 1998 Act back to a version of the traditional authority regime of the post-1964 era of Rhodesian rule. What further changes might be made and how consistent they could be with the tenure rights of new beneficiaries of the second stage of land reform are issues still under debate, and some recommendations from findings from the studies of the FTLRP are discussed below. One opinion is that there should be a single form of land tenure extending across old and new redistributed land, remaining large commercial farms and ranches, and Communal Areas (CAs), but aside from comparing that view with other rationales not based on a singular system, the specific processes and problems of tenure in the communal areas will not be centrally addressed in the collection.

The political and economic context of land reform in Zimbabwe

A first stage of land reform got under way in Zimbabwe immediately after Independence in 1980 and the end of the liberation war that had preceded it (among the many accounts of the war and of how central an issue was land in it see Ranger 1985, Mandaza 1986, Kriger 1992, Bhebe and Ranger 1995) . That process and its legacy in the Old Resettlement Areas (ORAs), as they are labelled here, are placed alongside the recent Fast Track experience in the contribution by Dekker and Kinsey to this collection. Even this early phase was on a significant scale in that a higher proportion of the formerly white-owned farm land in Zimbabwe (23 percent) was transferred to African use by 1996 (Bowyer-Bell and Stoneman 2000; Kinsey 1999, 2003) than in Kenya (15 percent) (Leys 1975, Wasserman 1976, Kanyinga 2000), and considerably more than in the 17 years of land reform in South Africa. Latest available figures from South Africa's new Department of Rural Development and Land Reform (quoted in Umhlaba Wethu 2011) are 6.9 percent from the Land Redistribution programme, similar to Zimbabwe's Old Resettlement, and the Restitution programme together, against a target of 30 percent (for overviews of progress, see Hall 2009, especially p.2; Kleinbooi 2010, 43).

The context of this Old Resettlement also has to be appreciated, which requires the briefest survey of the country's geography. The squashed diamond shape of Zimbabwe consists of a mainly well-watered and fertile core of ‘Highveld’ at altitudes of over 1000 metres, which is slightly off-centre to the north and east, and corresponds roughly to the three Mashonaland Provinces, with drier terrain descending in steps as one goes down to the country's peripheries in the five other provinces. The highveld was characterised by large-scale farms, mainly owner-managed, growing food (maize, wheat, oilseeds, fruit) and export crops (tobacco, cotton, flowers, vegetables), some under irrigation, plus even larger ranches and irrigated sugar estates and, in recent years increasingly, ‘conservancies’ for wild-life based tourism in the low-lying areas. Most of the ‘reserves’, wherein African populations had been dumped at periods from the late nineteenth century until as late as the 1950s (Palmer 1977), now termed ‘Communal Areas’ (CAs),2 are in the drier areas, though some are in Mashonaland and fertile parts of Manicaland.

That simplified geography is not just the basis of the agrarian structure which has been transformed but also gave shape to the liberation war. It was fought in the Communal Areas, some of which came under de facto control (at least at night) by guerrilla fighters backed by mobilised rural dwellers, from where encroachments and intimidation of farmers near the CAs picked away at the borders between the two types of agricultural areas (Ranger 1985, Sadomba 2009, especially Appendix 4). Much of the land that was bought up and redistributed by government in the 1980s was semi-occupied or made vacant by these forays during the war, which had established ‘facts on the ground’, which were then regularised by the official Resettlement programme (Moyo 1995, Alexander 2006). It should be remembered then that the programme was a consequence of a violent contestation in part over land. But the circumstances of that struggle delivered a great deal of land in the semi-arid areas, classified in Zimbabwe as Natural Regions (NR) IV and V, plus NR III designated as of marginal arable use, rather than much in the high potential NRs I and II (Vincent and Thomas 1961). But, as the Dekker and Kinsey paper shows, summarising studies such as Cusworth and Walker (1988) and Kinsey (1999), the outcomes of the Old Resettlement (OR) process were ‘successful’ in several respects until the 1990s, notably in making productive use of the land and increasing the number and range of livelihoods, largely as a result of the availability of farming inputs, advice and at times management, infrastructure and services. This account of earlier land redistribution experience has value not merely as a temporal comparison with FTLRP, but it can be used to rebut any crude argument that reductions in output under Fast Track were an inevitable consequence of the replacement of large by small units of production and of ‘skilled’, white farmers by black smallholders, for their predecessors had shown themselves capable of a different result when conditions were favourable. Of course that argument does not address more subtle concerns of critics about the undeniably real threat to smallholders taking over highly capitalised and technical production in fields like dairying and horticulture, and the contributions will explore those kinds of outcomes. A similar conclusion about the potential of Zimbabwe smallholder production is offered by the entry of communal area farmers in some areas, together with OR farmers, into significant smallholder production of marketable surpluses in crops like cotton, burley tobacco, and sunflowers (Cliffe 1988, Muir-Leresche 2006). The Dekker and Kinsey paper also documents how that ‘success’ in production and livelihoods levels did not continue into the later 1990s and 2000s, as a result of the macro-economic failure and its consequent denial of inputs like fertiliser and seed and of credit and extension to this sector. Again, this points to the weight that needs to be given to that factor in explanations of poor performance in post-Fast Track output, if even established smallholders were hit so hard.

This Land Reform and Resettlement Programme (LRRP) – referred to here as OR – had run out of steam by the late 1980s. Almost all the land acquired was already in public hands by 1984, although there was a trickle of new people resettled in the 1990s, and the UK budget agreed upon immediately after Independence to finance half of approved schemes under LRRP was never completely used up. But discussion about a further programme of renewed large-scale land redistribution moved from mere rhetoric to actual planning in the late 1990s.3 Government actually produced a detailed designed and costed LRRP-2 and launched a two-year Inception Phase for it, presenting it at a Donors Conference in June 1998, which endorsed the need for such a programme and agreed upon aims for it. The agreements entered into were followed up by a joint UK/EU mission to examine their possible support for the Inception programme in May 1999, and by a National Stakeholder Workshop in June, brokered by UNDP, which considered a National Land Policy Framework Paper (the Shivji 1998 report). The Inception Phase did receive promises of funding for technical assistance from the US, Norway, Netherlands and Sweden and a loan from the World Bank. Some of the elements that were later to form components of FTLRP were outlined in these documents: the conception of two main groups of beneficiaries, the A1 (smallholder) and the A2 (medium-scale); and a three-tier scheme for extending grazing areas of CAs through incorporation of ranches, comparable to the one Model D scheme implemented under OR (see Alexander 1991, 2006, 130–31, Robins 1994), to be discussed below.

However, funding for this LRRP-2 did not materialise, especially from the UK, the donor country with which the most detailed technical discussions had been taking place. The talks were at an advanced stage and had outlined a plan for transferring some 5 million ha of the remaining farms (about half of the total area eventually obtained in the FTLRP from 2000 to present), acquired by due legal process with compensation, and for which donor contributions would be forthcoming (GoZ 1999). It is still not clear how and why a jointly worked out blueprint was not consummated, and this issue will not be specifically addressed in the present collection. However, the intensity of the rhetoric about who should be blamed became a central dimension of debates over the next decade – Mugabe picturing this as a denial of the UK's historical responsibility and commitments at the Lancaster House Conference that arranged Zimbabwe's Independence; the UK Government seemingly reneging on support of a plan whose targets were not exclusively the ‘poor’.4 Whatever the actual roots, ZANU-PF's main justification of the land occupations from 2000 and of FTLRP when it was formalised in 2002 became and remains an alternative to a donor-supported LRRP-2, and a gesture of defiance to the old colonial power.

Land occupations and the genesis of Fast Track

The emerging pattern of FTLRP differed from the prescriptions of LRRP-2 in the mode of acquisition of land; it was not through the due legal processes as previously specified, involving designation of farms, with the possibility of appeal, and the payment of full value compensation. A replacement Land Acquisition Act 2000, plus regulations proclaimed after enabling changes in the Constitution, allowed for compulsory acquisition without compensation for the value of the land itself, only for the value of ‘improvements’. The priorities for the type of land that ought to be acquired also differed, as did the criteria for selecting beneficiaries, although the two main official ‘models’ on which schemes for allocating holdings did more or less correspond to the A1 and A2 formulae put forward in LRRP-2. The former was closely based on the similarly-labelled Model A used in OR, providing an individual smallholding for residence, homestead garden and field crop cultivation, together with a large area set aside for common grazing to which all members of the new community settled on the former farm would have access (regardless of whether devoting such high proportions of fertile land in Natural Regions I and II was appropriate). This became referred to as the ‘villagised’ sub-version of A1 as dwellings were clustered around a contiguous settlement area, as opposed to another variant of A1, to be known as ‘self-contained’, where the grazing was also sub-divided and combined with the individual arable plots to make up small farms. It may be that in the long sweep of history the mode of acquisition is not the main determining factor of outcomes, but the resulting short term circumstances of implementation – hurried, with no clear lines of responsibility, partly through occupation, and with planners trying to play catch-up with little finance for post-transfer support – meant that the mode of acquisition has been a critical factor in the first decade. It should be recognised that some of these factors were present in the OR programme (see Dekker and Kinsey in this collection), if not to the same extent.

Whatever the plans and their roots, it is generally accepted that this process of new land acquisition and redistribution was sparked by incidences of ‘occupations’, ‘incursions’ or ‘invasions’ (the terms used are themselves partisan) of farms inhabited by whites, and in some cases Africans who had obtained farms in the years after 1980. The first such instances occurred in 1998; these were often reined in by the ZANU-PF Government, but began again in earnest in 2000 (Moyo 2001, Moyo and Yeros 2005, Sachikonye 2003a). Varying from political protest to noisy disruption to violent intimidation, beatings and killings, they often dislodged owners and managers, and sometimes their farm workers, and were often named jambanja (Chaumba et al. 2003). Such incidents and their consequences on the ground are explored in some of the case studies presented here, but they illustrate that occupations were not the only means of acquiring land; in other cases it was a more administrative process. What is more a matter of dispute is the extent to which the farm occupations were a consequence of a coherent strategy articulated and orchestrated by the leadership of ZANU-PF and the Government of Zimbabwe (GoZ), or whether these two institutions had to be forced or manipulated into supporting them by a separate initiative taken by ‘veterans’, as argued by Moyo and Yeros (2005) and Sadomba (2008). Some who took the former position argued in a more nuanced way that the Fast Track formula inevitably involved a decisive restructuring of state institutions, from the top down to the local level, to implement it, and this marked a shift from a technicist approach to rural issues to a form of state that was partisan and violent (Alexander 2006, Tendi 2010).

The cases discussed in this collection do provide some local evidence of differing processes of occupation and the extent to which these were the key mechanism for land acquisition, without claiming to provide a definitive answer to the larger historical question. Irrespective of the origins of the impetus of the de facto occupations of farms, the ruling ZANU-PF party did shift to a policy of encouraging invasions when it was faced for the first time by an emerging, organised political opposition, the Movement for Democratic Change (MDC). The ruling party had to compete with a serious opposition in a 2000 referendum on a new constitution, which included a provision for land acquisition without compensation and which was voted down, in the subsequent parliamentary election in mid-2000 and in the presidential elections of 2002. This No vote in the 2000 referendum may have been an important early trigger for the decision to embark on the FTLRP. Gaidzanwa, as a participant in the 1999 constitutional reform process, recalls that the GoZ-appointed Constitution Commission's original contribution had accepted the principle of compensation for improvements to land. Indeed, the South African legislation on land was actually modelled on the 1990s legislation from Zimbabwe. The FTLR was turned into jambanja after the President changed the Clause 57 on land to disclaim any responsibility for the Zimbabwe state to compensate farmers whose land was taken for redistribution. The No vote hardened that resolve as spelled out in the revised Clause 57 in the rejected constitution. Giving the go-ahead to farm occupations and to the use of violence in acquiring land in 2000 was probably seen by ZANU-PF as a means of mobilising rural support and punishing its political foes, the white farmers, ‘their’ workers and other supporters of MDC (as argued by Alexander 2006). The ZANU-PF government then introduced a process of after-the-fact regularisation from 2002 with the formal introduction of FTLRP.

These differing mechanisms of de facto dispossession gave way to a new Land Acquisition Act in 2002 that legalised compulsory purchase without payment for the value of the land. (A legal commitment to pay for the value of improvements made to the land was written in, but in practice this has occurred only on a very limited scale.) And an official FTLR Programme got under way in July 2002, specifying what land was targeted for acquisition. Whatever had been the role of veterans or even spontaneous movements onto farms, once state control of past occupations and future designations was established, power over the process of redistribution of the land acquired was spread to include a melange of other actors: party and government leaders (local and national), traditional authorities as well as veterans (sometimes operating through district land committees, sometimes self-appointed). Who was then to get land was not simply a matter of satisfying popular need, and exercising patronage among rural people. Certain targets for percentages of beneficiaries from designated categories of people, such as veterans, women and former farm workers, had been mooted in earlier discussions like the 1998 Donors Conference on LRRP-2, but only the first category were specified in formal guidelines like GOZ 1998. But within any such designations actual allocations were seen as a means of satisfying the interests of higher, and also middle, levels of politicians, officials and local notables. How the differing local patterns of selecting beneficiaries played out, with what mixture of class interests satisfied, is documented in some of the cases here, but the overall outcomes were certainly varied.

The amount of farm land targeted for acquisition was over 10m. hectares (Utete 2003), about twice what had been specified in LLRRP-2 (GoZ 1998). Targets were also specified in the Programme for the proportion of A1 and A2 holdings. Most of this land had been acquired and sub-divided by the later part of 2003 and the first of a number of land audits was then made by a GoZ-appointed Committee (Utete 2003). This reported that at that date 4,324 farms had been redistributed covering 6.4m. ha., and 34 percent of this area had been allocated to 7,260 beneficiary households who received A2 (medium-scale) holdings and 66 percent to 127,192 holders of A1 smallholdings. Gaidzanwa, who was a member of the Committee, recalls the impression of considerable unevenness of the extent of the process and reliability of the reports across provinces. Those contributions to this collection that present more up to date national figures (Moyo and Scoones et al.) suggest slight increases in A1 beneficiaries by 2009 (145,000 is the lowest estimate) and perhaps a doubling of A2. Some case studies obtained numbers from provincial and district offices of AREX (Agricultural Extension Service, Ministry of Agriculture, formerly and now once again in 2010 AGRITEX), or surveyed samples, and their findings show regional variations around these national proportions – although these investigative efforts tend to confirm the observation that, ‘Data are hard to get hold of and often inconsistent, and the situation has been very fast-moving with official statistics not covering contested areas and new invasions’ (Scoones et al.2010, 33). Inconsistencies are compounded by the fact that some A1 occupants still have no offer letters.

Since 2003, the land reform is changing in shape and emphasis as the political struggles emerge and change. Some politicians use their muscle to dispossess recipients of land, especially A1; but some ‘land grabbers’ have in turn been dispossessed of land by war vets and others so the redistribution continues! Again, some of the small farmers and elite land grabbers are renting land back to the white farmers so the formal offer letters may not always reflect the actual use of land on the ground.

A further limitation of the local findings from micro-studies lies in the fact that often each pre-existing farm was targeted for sub-division on the basis of a single model, A1or A2, so reliable figures for the overall balance between types of holdings are only possible at the district or provincial level not from a study of one or a few ‘schemes’. Still, compared with the national ratios, where A2 beneficiaries made up less than 10 percent of the total, and received about 25 percent of the land, findings reported here range from the semi-arid Masvingo Province where only about 3 percent of beneficiaries were under the A2 model, but received 23.7 percent of the acquired land, reflecting that the larger areas needed to make a medium commercial enterprise from ranching (2009 figures from Scoones et al.2010, 33); to the high value commercial farming heartland with its developed infrastructure and proximity to the capital, where in Mazowe District, the A2 farms formed a higher percentage, and they received 211 of the former farms compared with only 113 allocated as A1 (Matondi forthcoming). Overall 70 percent of all A2 farms allocated in the country were in the most productive and accessible three Mashonaland Provinces, according to aggregates referred to by Moyo in this collection.

The broad proportions between these two main categories have to be interpreted with care as the distinctions between, for instance, the larger A1 self-contained and smaller A2 were in practice blurred (see Scoones et al.2010 for examples). The differences were further qualified and made more complicated by other additional categories of ‘beneficiaries’. There was, for instance, a ‘Three-Tier Scheme’ seen as suitable for a few areas in the drier south and west of the country, whereby the acquired land was to be used for livestock ranching, and was now made available as a common grazing area by rotation to groups and communities in neighbouring CAs, essentially a variant of the Model D scheme in one location under the OR Programme. According to evidence from case studies done by Moyo (this collection), AIAS (2007) and Matondi (forthcoming) Three-tier was only pursued in two districts in Matabeleland South Province: Mangwe where 13 farms were acquired for this purpose, and Matobo (14 farms). But, as with the original Model D (Robins 1994, Alexander 2006), when it came to implementation there was controversy about the precise aim, potential beneficiaries and patterns of land use. Official plans sought originally to restrict access to beneficiaries with more livestock which they would be expected to build up into a commercial beef herd – restrictions and aims which have been contested by communities, by encroachment as well as verbally, to the point where it became more common to modify the model to make grazing land available to whole communities. But then there seemed to be little communal management and something closer to open access.

In some arable areas various residual provisions (as spelled out by Chambati in this collection) were made for those former farm workers who remained on the land: through participation in occupations or via applications, or simply staying put, as well as de facto setting aside of very small plots, essentially ‘allotments’. Fox et al. (2007) report similar practices in Kadoma of farm workers being allocated plots half the 6 ha. of those of A1 beneficiaries or less. Zamchiya also cites instances of space being made for farm workers after initial allocations, and other cases of farm workers being subsequently displaced.

Another significant set of exceptions to all these formulae was those who received the whole of a pre-existing farm, or even more than one. This was sometimes within and at the time of the original allocation process, sometimes through subsequent, extra-legal displacement of beneficiaries or of unallocated acquisitions of other state land. This process of enforced land grabbing still goes on and is perpetrated by those who might accurately be described as ‘cronies’; indeed political muscle is a requirement for such acquisition. Some of the processes and the politically well-connected who are involved are documented in some of our case studies. Estimates of the scale of this pattern are made for Masvingo Province as a whole, where Scoones et al . (this collection) report 58 farms (111,330 ha.) as having been allocated as whole units under FTLRP, plus subsequent land grabs. Moyo (2011b, 261, 2011d) has recently estimated those who he designated ‘land grabbers’, which in his definition comprised those who have acquired farms of between 500-1,000 ha. (roughly a third of the average size of former ‘large-scale commercial’ holdings – although that does aggregate across the very wide range of what was ‘average’ in different Natural Regions). Accordingly, he estimates that 80 percent of the 3,000 holdings of this size or larger are now black owned. He suggests that his (rather arbitrary) category includes ‘urban and rural based professionals, public and private sector executives, other petty bourgeois elements and black capitalists’. Marongwe (this collection) offers detailed evidence that the reallocation of smaller holdings originally acquired, by occupation or allocation, as A2, and displacement of beneficiaries in favour of more politically connected people was ‘quite widespread’ after 2002.

For sure, ‘cronies’ did get some farms. But that is a far cry from assertions that all or most of the land went to them, and much depends on the definition of this political category. A significant number of farms have been acquired by prominent elite figures either by getting whole or large parts of pre-existing farms through FTLRP or by informal and often strong-arm methods, and this is an on-going process. Short of the promised Land Audit (see below) it is hard to estimate the scale. The references in the above paragraph give some indicators. But any statement that beneficiaries of FTLRP are confined solely to ‘cronies’ in the sense of those in the immediate power circles at the core of the ZANU-PF regime, who received farms on a scale of the previous owners, is indeed a ‘myth’ on both counts, as Scoones et al. assert in their study (this collection); and there is a consensus among contributors that smallholder farms now predominate in number and overall area, and thus a major agrarian restructuring has occurred, even if there is a difference in estimates between authors and regions.

Elite ‘land grabbing’ has been a dominant trend in another former settler-colony, Kenya. If its two decades longer resettlement experience is any clue to the future of Zimbabwe, there is likely to be a continued and powerful thrust by such interests, but it is also likely to be contested by more democratic forces, especially at the local level. So the eventual outcome may not be the same as in Kenya, not only in the amount of elite grabbing, but also as regards the resulting agrarian structures (Kanyinga 2009, Rutten and Owuor 2009). One tendency in that country has been for many large landowners to settle for informal renting out of land to smallholder tenants rather than persist in being capitalist farmers. The studies presented here do provide a few anecdotal references to the borrowing (Mwenezi, reported in Scoones et al.2010), leasing or selling off (of A1 grazing land in Kadoma reported by Chigumira 2009) of redistributed land, even though such things are proscribed by the laws governing FTLRP. Such tendencies might be prompted by what seems to be a pattern that large farms and even medium A2 holdings are underperforming compared with A1 smallholders, examined here. It may take perhaps a decade or so before it is clear how far any such forms of landlordism become general.

Another form of political patronage beside that of regime cronies receiving large farms was detected in some areas. Marongwe's contribution (this collection) stresses the party political bias favouring those applicants who sought medium-sized A2 holdings in the effectively peri-urban setting of Goromonzi: civil servants, members of the security forces, ZANU-PF members and those who were the clients of party notables were privileged. However, villagised A1 beneficiaries formed the majority of the larger sample used in his study (Marongwe 2009), as they did by and large nationally, although party patronage even operated among them. Moyo (this collection) and Murisa (this collection) among other contributors to this collection note similar bias in the actual processes of allocation of A2 to government and party notables elsewhere, but the beneficiaries in those areas more distant from Harare were drawn more from middle and/or local level institutions. Zamchiya (this collection) takes his analysis of allocation processes in Chipinge further by looking specifically at bias in selection of A1 beneficiaries on self-contained and villagised schemes. He argues that this large segment were drawn principally from local ZANU-PF members and supporters. Though such land recipients may be ‘poor’ there is a bias toward those with ‘political capital’. Mkodzongi (2011) observed a similar process in Mhondoro in Midlands Province but interprets the implications differently, pointing out that anyone might claim party membership or to be a supporter for opportunistic reasons including jumping the queue for land, as Zimbabweans have done at least since the pre-Independence elections in 1979 and 1980 when voters were urged ‘to eat Muzorewa but vote Mugabe’ (personal observation, Cliffe 1980). Scoones et al. (2010, 54) and Mujere (this collection) similarly point to tendencies for people to assume several identities which they deploy to suit circumstances. One essential difference between such accounts lies in their interpretation of how deep party identity goes, though of course the depth of that identity might vary between areas, not all of which are marked by the same historical dominance of ZANU-PF as in a district such as Goromonzi. Of course even if patronage on party lines has not been as permanent and exclusive as ZANU-PF elites intended, such a process entrenches the partisan and patronage-based institutional character of the state and limits the free expression of political views.

The lack of clarity about who the beneficiaries are and what are the continuing dynamics of land acquisition and contestation was recognised in the political process from 2008 leading to a power-sharing government that included ZANU-PF and the two factions of MDC. The Global Political Agreement (GPA) made two pronouncements on land issues. It specifically called for a Land Audit ‘for the purpose of establishing accountability and eliminating multiple farm ownership’ (GOZ 2008 Article 5.9a), although this has yet to be initiated. But the GPA also set out the broader context within which future policy on land reform should be framed. It asserted that FTLR was irreversible, stating that:

“while differing on the methodology of acquisition and redistribution, the parties acknowledge that compulsory acquisition and redistribution has taken place” (Article 5.4) but in fact differences over acquisition are now basically matters for the past. They did agree that “the primary obligation of compensating former land owners for land acquisition rests on the former colonial power” (GOZ 2008 Article 5.6).

This formula, to which both partners have signed up, presents a particular challenge to prospective donors to land and agricultural programmes and to the international community generally in considering whether to resume normal economic relations with Zimbabwe. There is no doubt that the economic prospects have changed with the dollarization that halted hyper-inflation, and the recent expansion of mineral exports, making extra revenues available and reducing reliance on western largesse. The political circumstances have also changed with the inauguration of coalition government, but the partnership remains an unequal one and policy differences, not always along the party divide, prevent action on key issues. There has, for instance, been protracted discussion about a revision of land tenure, in all sectors but especially with regard to fast track areas, but still no draft legislation. It is one of the aims of this collection to present findings that might illuminate such debates.

The scale and type of land redistribution through FTLR, and the numbers and categories of people involved, begin to indicate the possible longer run outcomes of processes initiated eleven years ago. But what differences FTLRP has made to land production systems, the changed livelihoods pursued, the resulting overall agrarian and class structures will depend on what has taken place on the ground, in a range of different agro-ecological and political settings, in the past decade. The contributions to this collection provide as comprehensive a set of findings as have yet been assembled in one place on these outcomes, covering beneficiary origins, the fate of agricultural production and pursuit of other livelihoods, and the dynamics of the emergence of new communities and agrarian structures. An attempt will be made here to summarise, compare and contrast findings on these topics.

The case studies

Three major, longitudinal studies have been carried out in Zimbabwe covering the whole period of FTLRP, each involving surveys at different dates carried out in eleven of the country's 53 districts by three institutions, the African Institute for Agrarian Studies (AIAS), Harare; the Ruzivo Trust (RT), originally under the Centre for Rural Development, Harare; and a group from the University of Zimbabwe, AGRITEX and the Institute of Development Studies (IDS), University of Sussex, Brighton, UK, coordinated by the Institute for Poverty, Land and Agrarian Studies (PLAAS) at the University of Western Cape in South Africa. We have been fortunate to get three summary studies drawing on this work from Moyo and Chambati (AIAS) and Scoones et al. (IDS). In addition several other researchers have provided work based on individual studies all of which have involved fieldwork in one or more locations.5

Chronologically, mention can first be made of Dekker and Kinsey who have provided a summary of the Old Resettlement (OR) beginning in 1980 and they report on findings from the same sample of beneficiary households who have been surveyed over the whole of that period. Their contribution offers not only invaluable comparative material on that earlier process and evidence of broadly ‘successful’ production and livelihoods outcomes in the first decade, thereby refuting views that simply dismiss the potential of smallholder production and capabilities of Africans on a priori grounds, which were voiced then and are still repeated. They also give evidence of how this performance was not maintained in the second decade or so, pointing to possible long term issues such as new generations that might affect sustainability, but also emphasising the impact of economic stresses on inputs and markets from structural adjustment in the 1990s and the economic meltdown of the 2000s on these established farmers.

Moyo has drawn together material from six district studies of outcomes from all but two of the regions (outside the capital, Harare) conducted by the African Institute for Agrarian Studies (AIAS). Most of these data are presented in aggregate or average form, but many tables provide figures for each district as well. He combines that with a decade-long observation of a wide range of policy documents and processes nation-wide. He goes on to look at national developments in production, investment and marketing over the last couple of years and claims to discern, finally, signs of recovery in the contribution of the reformed sector.

Scoones et al. present a summary and up-date of their book length study (2010)6 reporting on developments in differing local environments in the southern province of Masvingo whose previously large holdings in drier conditions were given over more to ranching, wildlife and (in some areas) irrigated agriculture.

Another set of contributions, while examining local case studies and providing evidence of production, livelihood and social outcomes, take as a main focus the consequence to a particular social group: women in one study; farm workers in another. Other studies also provide snippets of information on these two categories. Mutopo sets her localised study in the historical and national context of how far women were beneficiaries of old and new land redistribution, referring to their limited land access rights in pre-colonial and colonial society which was not dramatically changed despite women being involved in the liberation struggle. That review of overall evidence in comparing the target of 20 percent suggested by the 1998 Donors Conference, for instance, with actual proportions of women who received land in their own name, is further substantiated by Moyo and Scoones et al. in their contributions, and in their other studies (Moyo 2009, Scoones et al.2010), and in official reports (Utete 2003). Other, gender-focused studies like WLZ (2007) and Goebel (2005) and local case studies, like Mutopo's, confirm the narrow openings for women to be granted land officially in their own right but also show how women farmers who manage to get use of land have often had to acquire it indirectly and by wheedling it out of spouses or wider family or political influentials; and having got land have difficulty in utilising it fully, experiencing the usual problems of scarcity of inputs, draught power, water, labour and secure tenure even more severely than men. Mutopo goes on to use an even more micro-focus to explore options faced by women on a former ranch in the south-eastern district of Mwenezi. She brings out the enterprising way some of them have been able to build on the security offered by their having land, in their own right or via husbands, and on the narrow opportunity of growing a surplus of beans, nuts and vegetables in a semi-arid area to deliver their produce to local markets in towns and over the South African border. She shows how realising these new economic prospects in turn empowered them in terms of controlling their lives more through involvement in a wider social universe and of changing gender relations at home.

Chambati utilises the AIAS survey sources and national data to undertake an exploration of the outcomes of the FTLRP for farm workers. He offers a broader analysis, not just of what has happened to them, but of the differing kinds of labour relations that have emerged to replace the paternalistic farmer/farmworker relationship that typically characterised previous patterns of ‘belonging’ on the former large farms (Rutherford 2003, 2008). He acknowledges that many early reports such as Sachikonye (2003) and FCT (2003) noted that disappointingly low proportions of beneficiaries were former farm workers, seldom above 10 percent, but points out that were no specified targets in the FTLRP policy document (GOZ 2001). He suggests that on the basis of a different ratio, the proportion of sampled farmworkers who took part in invasions and those who got land, as a group they fared better than might be thought. He also shows how some received tiny plots or stayed in their houses on allocated holdings, and how they can now be seen as making up a new kind of agricultural labour for new land holders, but often not tied to one farm, sometimes working in itinerant organised groups and on occasions able to bargain for reasonable wages with labour-starved A2 farmers. A complex new regime of labour relations has been emerging.

Three studies build on further local cases covering a range of agro-ecological and political circumstances to look at beneficiary selection. All cover localities within districts covered by one of the three main studies, so cross checks and comparisons of findings are available to the discerning reader, although this Introduction will only draw attention to some particularly interesting and contested issues. These cases use follow-up surveys to provide more depth to the findings from official land audits, like Utete (2003), and from the three institutions' studies on the selection of beneficiaries. They spell out what actually happened in allocating land rather than just the laid-down procedures, and reassess who the beneficiaries were. As mentioned above, Marongwe shows that in one A2 scheme in Goromonzi, located conveniently close to Harare, actual distribution did not accord with the formal process of application nor the procedures for selection according to technical competence and by a land committee. Rather they were made on an ad hoc basis, depending on local political circumstances, and were politically partisan with the result that the great majority of actual recipients did not meet the criteria laid down and were not on the original lists. He also argues that this privileging of political and governmental elites resulted in new farmers not having technical capacity, which added to their problems of financing inputs and labour required by the scale and type of production envisaged on A2 farms, and the underutilisation of land at least up to 2003. Those consequences are projected not so much on the basis of production outcomes as on notions that underpin continuing debate, on the assumed potential of ‘real farmers’, a category which is not clearly defined (Cousins and Scoones 2009). Marongwe's focus is primarily on A2 farmers, and smaller A1 farms are discussed only in cases where their land is being grabbed by the A2 elites. Zamchiya also portrays a pattern of politicisation of the selection and allocation processes in a very different district in the eastern border areas, Chipinge. As indicated above this picture differs in that he detects a party political bias even among beneficiaries in self-contained and villagised versions of A1. Mkodzongi (2011) enters this particular debate in arguing that there is little that is fixed about the categories of ordinary party ‘members’ or ‘supporters’ at the local level. But Zamchiya partly anticipates such counter evidence by suggesting that demonstrations of loyalty to ZANU-PF are persistent conditions for maintaining rights to land in his district.

Mujere offers a different perspective on the selection of beneficiaries and its consequences. Moving away from considerations of patronage and political bias in the process of allocation, he focuses on the subjective distinctions between ‘strangers’ and ‘locals’ among new farmers in A1 village schemes in the lower potential, but heavily populated Masvingo district of Gutu (also covered in Scoones et al.). In the process of exploring the interaction between these different categories and those in other studies, he brings out the issue that, although FTLRP did not involve an explicit commitment or a sub-component, like the Restitution Programme in South Africa, many would-be beneficiaries were motivated by claims of restoring access to specific land from which their kin or their ancestors had been dispossessed. To that extent relations between different land recipients were, and remain, matters of their ‘belonging’. Moreover the identities that came into play involved complex sets of claims about who had rights to land. As well as the division between those with some historical claim as opposed to ‘strangers’, mostly from more distant localities, there were those between restitution claimants. These could be based on ethnicity, such as the predominant group in the district, the Kalanga, or those based on autochthony. And the latter could base claims to ‘their’ land on the basis of their community having been displaced within living memory, often involving the loss of chiefdoms – the right to return – or the spiritual claim that their ancestors were buried there – the right conferred by ‘bones’. Conflicting claims at the start of FTLRP as to who should receive an allocation have evolved into disputes as to who should control the schemes now in place. These latter involve traditional leaders and not just the new farmers themselves, but also interact with the committees of seven put in place to run schemes, and the District Land Committees to whom they are supposedly answerable under government directives but not under existing legislation, and other government and political bodies. Sketching in this dimension of complexity about the new realities of land relations deepens understanding of the outcomes. But the study does not identify what mechanisms might resolve these persistent and potentially debilitating disputes. This perspective does call attention to the need to clarify not just legal land tenure rights but the provision of land administration systems, including dispute resolution procedures.

Murisa's particular focus in two Mashonaland districts goes beyond individual households and how they are making out to consider what patterns of social organisation are emerging, especially those that provide cooperating networks for farmers. The very act of resettlement under FTLRP, as with the earlier OR, brought together in one place a set of people from diverse localities and backgrounds (see also Barr 2004). With much local variation across his case studies, Murisa finds that people from nearby and distant communal areas, from towns and former farms sharing little if any past contact or other ties, now find themselves neighbours and facing the challenge of developing local institutions for specific functions (to run social facilities and farm support among other things). How far they have started to build some networks in place of former ‘farm communities’ is explored in passing in several case studies, but local farmer groups are potentially a vital contribution to building new ‘social capital’, given the unfamiliar challenges to small and medium size farmers of coping with larger scale technology, like irrigation, tobacco curing and other crop processing, in a context where mechanised but also simple inputs are scarce, and where credit to pay for them and access to new markets is difficult. Murisa documents the presence in his areas of farmers' groups pooling efforts to access inputs, reach markets, obtain tilling services, and also engage in cooperative production activities in irrigation and larger-scale operations such as wheat harvesting. They also join hands in resisting encroachments from A2 and politically connected farmers and in seeking greater tenure security together. His findings reveal that such initiatives do not always survive or prove beneficial. They are sometimes subject to corruption and nepotism (as with cooperatives in many rural areas), with an inverse correlation to size of group.

Overall, the studies in this collection provide first-hand information and observation from 14 different districts in seven of the eight provinces of the country outside Harare. The exception not covered is Matabeleland North. A neighbouring district, Mangwe, in Matabeleland South, is only covered in the form of returns to a sample survey there reported along with those from other sites in Moyo's contribution. The underplaying of these areas is to be regretted, especially for political reasons as they are generally seen along with Manicaland as the rural core for the opposition MDC, but will hopefully be covered by future research.

The eastern province of Manicaland, which contains the whole gamut of agro-ecological conditions found in Zimbabwe, none of which can be regarded as ‘typical’, is the subject of only one study (Zamchiya) with only limited data on post-settlement economic and social outcomes. The high potential heartland of former settler agriculture in Mashonaland's three provinces is the target of seven investigations in four districts. And four districts in each of the other (more marginal for arable) provinces, Masvingo and Midlands, are also included.

Of course no one case study, whether of a former farm, scheme, sample surveys from a locality or wider district – in any country – can give a ‘representative’ assessment of the impact of a widespread upheaval such as Zimbabwe's FTLRP. The significance of the findings in the book length account of Scoones et al. (2010) has already been dismissed by some reviewers (e.g. Andoh 2010) on the grounds that Masvingo is so distinctly different, even though four differing ‘clusters’ are surveyed, an argument which does not challenge any of its findings; it just serves as a caution before drawing any wider generalisation of conclusions or lessons, as the book is careful to do, and as Andoh's review explicitly calls for. One aim of this collection is to enhance the value of every one of the case studies in placing them alongside each other in order to have a better, though not foolproof, basis for such generalisations – as will be attempted in comparing findings and considering future implications in the rest of this Introduction.

Findings

New agrarian structure

One basic conclusion with which all broadly agree, but with differing assessments of details and actual numbers, is that the ‘dualist’ pattern of land use inherited from the settler colonial past, with a large, capitalist white sector and African smallholder reserves, modified to a degree in the 1980s and 1990s, has now undergone fundamental structural change. The overwhelming form of agriculture is small-scale, relies mainly but not exclusively on household labour, and produces for domestic consumption as well as the market. In fact that racially based dualism of land use organisation was always over-simplified since the establishment from the 1930s of a middle stratum of ‘small-scale commercial or African Purchase Area farms’, owned by individual African farmers outside the communal area system, with a hired labour force (Cheater 1984). The A2 farms created under FTLRP were estimated (Moyo in this collection, quoting GoZ figures) in 2010 to constitute 22,373 farms. However, Matondi (forthcoming) quotes a lower figure again derived from official figures of only 16,386 – a significant difference pointing up the need to complete the Land Audit agreed in the GPA. Whatever the exact number, these can be considered to be the other part of the middle band of what is now in Moyo's phrase a ‘tri-modal’ structure. The top stratum in 1980 had consisted of some 6,000 white-owned large commercial farms, and even larger estates and ranches, occupying some 15m. ha., the better half of the country's agricultural land. This was reduced to 12.5m. ha. by the 1990s, and through FTLRP is now down to perhaps 3m. ha. or less. There has been a corresponding increase in the area devoted to smallholder production which included only the CAs in 1980 and which ran to 16 m. ha., but now represents 62 percent of the farming area (CAs plus OR plus A1).

The processes of FTLRP and the new structure emerging have also to be understood as involving class reformation. Many of the very largest of the former commercial farms, especially the ranches and agro-estates, most under corporate ownership, remain in private hands (Moyo 2011d). Together with the few remaining farms owned still by white individuals, many of them now operating on reduced areas – estimated by Moyo (this collection) at 200–300 nationally; 4 percent of the land area in Masvingo Province (Scoones et al.2010) – these represent a continuing but much reduced class of large capitalist farmers directly employing a labour force. Formerly there were about 6000 such farmers and the evidence in keeping track of them comes from studies like Selby (2006) and the still existent Commercial Farmers Union (CFU) who conducted a survey of former members. An interview with two CFU officials (Cliffe, personal communication, 2010) revealed that over 60 percent of their survey had no desire to resume farming in Zimbabwe, even if it were feasible; most cited age as the decisive factor. It is known that some have moved to farm in countries as diverse as Malawi, Zambia, Mozambique, South Africa and as far afield as Nigeria (Hammar 2010).

The findings of these studies agree that there has also been a reduction of those who were farm workers, but offer a more complex and differential pattern to that of the wholesale dispossession of this class pictured in some accounts of the early 2000s, and even a different overview to those few studies that were done at that time (Sachikonye 2003b, FCT 2001). A significant minority remain on the estates and others live on remaining or newly acquired large farms; another proportion has retained or found a place as permanent workers on new farms, especially A2. Moyo and Scoones et al. report that a majority of the latter hire permanent labour, although Chambati suggests that a new form of wage labour where workers are no longer tied to one farm but may be itinerant, working in groups sometimes, are more characteristic of new labour relations. A large part of the hired labour is in fact casual. Former farm workers were absorbed in FTLRP schemes: in Masvingo they constituted 11 percent of A1 beneficiaries; in Mazowe a smaller proportion; in one scheme surveyed by Chigumira (2009) the proportion was 30 percent, while in her other two areas of study they were negligible. In other schemes some farm workers received a holding that was but a small proportion of the standard A1 plot (Chambati). Elsewhere some have hung on to their residences, in the midst of new villages or as ‘squatters’ on new holdings.

Who benefitted?

Reference has been made above to varied findings about the selection of beneficiaries and allocation of holdings, and to the different conclusions about how far the outcome skewed the list of those who got land politically, either in the narrow sense of highly placed cronies, or a more widespread preference for ruling party supporters. One crucial dimension is that of gender. At the donors' conference of September 1998 (discussed earlier), a quota of 20 percent for women was adopted. However, this quota did not become formal policy and was not included in the Inception Phase Framework Plan 1999–2000 and has not been put into statute (Jeanette Manjengwa, personal communication). There is a consistency in the reports about what proportion of actual beneficiaries (i.e. named on the offer letter in their own right) were women: Utete 2003 offered national figures of 18 percent of A1 and 12 percent of A2. In Mazowe these were 13 percent and 11 percent, respectively (Matondi forthcoming); very similar figures from a 2007 Provincial Land Audit are quoted for Masvingo by Scoones et al. (this collection). These findings compare with what seem to have been even smaller proportions in ORA allocations in the 1980s (Dekker and Kinsey, this collection). But those like Mutopo (this collection) and Scoones et al. who explore the gender dimension in more detail indicate that women received a slightly higher proportion in subsequent reallocations, and also use a variety of stratagems to obtain the use of land even if they were not the recorded beneficiary. These studies suggest that the disproportionality compared even with the ratio of women-headed households in rural communities generally, often as high as 30 percent, is a result of the patriarchal practices at work in civil society (Gaidzanwa 1988, Pankhurst 1991), although there seems to have been little pro-active attempt to reverse such patterns in the formal allocation process (Jacobs 2009).

Beyond these findings of who among certain social groups got land, the evidence that studies in Masvingo, Mazowe and in Moyo's contribution refer to, of national or provincial or district figures, use categories that merge the place of origin and the occupation of beneficiaries, adding up those who are from urban or communal areas, say, alongside civil servants or war veterans. Some of the difficulties of interpreting these figures and even of compiling them are underlined by Matondi (forthcoming) who in attempting to make a classification of ‘backgrounds’ from full lists from Mazowe District made available by local officials found it was impossible from the records to place 68 percent of households unambiguously in an appropriate box. Of the remainder two thirds were dubbed ‘ordinary’, clearly used as a residual category not in formal occupation, thus including farmers but others too. Figures covering Masvingo Province reported by Scoones et al. classify 50 percent as ‘ordinary’ (defined as those without a formal occupation pre-land reform) from rural areas and 18 percent from urban areas. Chigumira (2009) in a comparable set of case studies found over 30 percent on one A1/villagised scheme were from CAs, but only tiny numbers on an A1/self-contained and an A2 scheme, while some 30 percent of A1 were from urban areas, compared with 70 percent of the A2. In some of these cases, and also in Mazowe (Matondi forthcoming), there was a smaller proportion of veterans than the 15 percent that was laid down as a target. All the cases citing such evidence of ‘backgrounds’ categorised significant though varied minorities of beneficiaries as civil servants and from the security services, with higher percentages among A2. All of these reports use such figures of ‘ordinary’ (or some similar category) as evidence that the great majority of land recipients were not members of privileged strata of society and many were ‘poor’ (as of course were a very high proportion of Zimbabweans by the end of a decade of economic collapse). Zamchiya's figures for Chipinge, the only case study from Manicaland, are an exception: 50 percent listed as civil servants (including security branches), 22 percent traditional authorities (including headmen, village heads, messengers and family members); veterans were 17 percent and only 11 percent were ‘other ordinary’. Different studies define categories of beneficiaries in inconsistent ways, making direct comparison difficult. Different figures thus do not necessarily reflect regional variations, although these are of course inevitable.

Outcomes for production and livelihoods

To make any assessment of the impact of FTLRP on production or livelihoods invites the making of some kind of comparison, typically of a before-and-after kind. Various comparative indicators are used in the studies. One used by Moyo and by Scoones et al. traces the change in total output of various crops since 2000 using national figures. These show a general downward trend but with exceptions such as cotton, small grains, round nuts and beans. Moyo updates this picture by pointing to what might at last signify a reversal of some of the declines in the seasons after 2008, coinciding with the end of the worst period of macro-economic crisis. What might be a significant component of this shift is tobacco which not only recorded growth of production after a sharp decline in 2000–04 but a surge in producers from some 3,000 registered in 2000 to some 50,000 mainly small growers by 2010. These examples might also point to the role of contract farming when normal input supplies are limited, and in particular the Chinese corporate enterprises engaged in it, especially in cotton and tobacco (see Moyo in this collection).

Other contributors offer some account of output at local levels (Mazowe, Goromonzi and Zvimba) or even on what were individual farms and now consist of many smaller ones (Chigumira 2009 and Mutopo). Inevitably outcomes are varied between sites and within samples, although evidence of the central priority given to maize comes out and is consistent with the findings from the larger surveys of Moyo and Scoones et al. as are the significant proportions achieving the measure normally used as the threshold to ensure grain self-sufficiency, of 1+ tonne per household. Collecting data on before and after situations has proved very difficult given the chaos and confusion of FTLR, the intense politicisation of land, etc. However, the literature contains two individual small cases where new settlement farmers have been compared with nearby CA farmers and they offer contrasting findings: Zikhali (2008b) found that a sample of FTLRP used ‘significantly more fertiliser, and draught power per hectare’ than a sample in Chiweshe CA and realised three times higher yields (although their land was of higher potential and they got better access to inputs and extension). On the other hand, Chamunorwa (2010) found that among samples in three Mashonaland East districts fertiliser use on A1 farms was slightly below those of CA farmers, and that the former had poorer yields for commercial crops, but better for food.

But there are limitations inherent in any such comparisons: the figures indicate correlations between trends and the process of fast track; they do not prove causal relations. They do not compare impact of other simultaneous events or processes, especially the massive macro-economic decline and a range of accompanying factors such as declining public services, including those to agriculture, and worsening access to markets, credit, and inputs, which will have affected all agriculture including what had been the growing sectors of old resettlement areas and communal areas. In addition such comparisons do not compare like with like. Production figures may relate to the same physical areas but now measure output under different agrarian systems. For instance, former ranching areas are now producing grain and vegetable crops (Mutopo); and the livestock numbers and herd composition and the objectives of husbandry change from the simple commercial ranching logic of maximising profits by optimising off-take and/or quality to a more varied pattern when livestock fulfil a more complex range of functions including self-subsistence, draught power, bride-wealth, sources of finance for key household expenditure and a means of saving, especially when money is worthless as it was during the hyperinflation up to 2009 (Mavedzenge et al.2009).

A further difficulty in using output data as a basis for identifying ‘outcomes’ of fast track land reform is that such a formulation can easily descend into implying a causal relationship. But the very finding that the output record varies greatly between crops, regions and strata of beneficiaries in the same area suggests the need for exploring what intervening variables might have explanatory power. In particular there is a strong consensus among contributions that most beneficiaries report facing major constraints to realising the potential of their land and other resources because of lack of timely access to seed, fertilisers and other inputs into production, to credit, reliable markets with guaranteed fair prices, to labour, and to extension services, as well as the limited opportunities for non-farming livelihood sources given the massive downturn in the economy. The fact that such factors have affected other sectors of agriculture like the OR areas, reported by Dekker and Kinsey as we have seen, reinforces the possibility that such contextual factors may be responsible for poor performance rather than land redistribution per se. Scrutinising evidence that can illuminate that issue is of course important in clarifying future options for policy and practice.

A different approach is to look at livelihoods and not just farm output. Before and after numbers of livelihoods can be instructive. Several of the studies of areas and individual farms and schemes show that in more marginal areas more self-provisioning households of new settlers and casual labour can be provided for than the number of owner and worker families under the old farm system. But Scoones et al. and some of the individual case studies provide valuable evidence of the varied types of livelihoods and in turn the class dynamics that have emerged; the former summing up the outcome in terms of 10 percent ‘dropping out’, a third ‘hanging in’, 20 percent ‘stepping out’, while just over a third are ‘stepping up’. Similar categories which combine a range of measures and outcomes are used in other studies, e.g. by Matondi (forthcoming), to show complex patterns of ‘success’ and ‘failure’ in securing livelihoods which flesh out the simple counting of those households producing more than a ton of maize.

What problem of land tenure security?

An issue that is hotly debated in the process of formulating a new Land Tenure Act in Zimbabwe is how big a priority is securing land rights in the new (and Old) resettled areas. On this question there is a major polarisation between the findings of some contributors. Several of them explore the actual processes of selecting beneficiaries and allocating land, as we have seen. Matondi (forthcoming) goes beyond documenting what happened and who benefitted to probe what those procedures mean for the presence or absence of tenure security. Surveys from Mazowe, and other districts in Mashonaland investigated by his team, ask respondents their views about the degree of insecurity and its causes. He tends to see insecurity of tenure as inherent in the redistributed holdings: A1 beneficiaries who, at best, have offer letters or some even less formal occupancy, and A2 who were promised leases that have not always been issued. The rights conferred by these forms of title are rendered less secure by the realities where ‘regulations’ passed by local power-holders interfere, and by several people sometimes being allocated the same plot. These research findings are used to support the conclusion that insecurity is a root problem and directly affects investment in land and productivity, although the survey data presented on people's attitudes seem to focus mainly on what they state they are doing to fend off or prepare for insecurity, partly begging the question, and does not explore, for instance, the link between tenure rights and credit. At the other end of the spectrum, Moyo, reporting on AIAS (2009) research, seems to downplay the problem of insecurity, offering as evidence responses to a different kind of attitudinal survey that asks people whether it is a problem that hinders them in utilising their land or in investing in it. But it also presents data about whether people are indeed investing in their land, and finds that they are not seriously deterred by only having a ‘permit’ or no document at all. These answers suggest that the problem is overstated, which in turn lends support to a policy stance that resists individualisation of tenure, and the need to seek alternative forms of securing credit not based on property. Another source of potential insecurity, not based on the actual documents related to the piece of land, is raised by Mujere's contribution. He records conflicts over land arising from the competing entities claiming power to allocate new holdings and mediate disputes over rights; on the one hand traditional authorities claiming jurisdiction on the basis of their customary role and the 1998 Traditional Leaders Act, which gave them some duties in Old Resettlement, and the new Land Committees set up under FTLRP regulations but so far without legislated authority. He argues that formal legislation to clarify which bodies have final juridical authority and a single channel for resolving disputes is needed if continuing disputes over access to land are to be avoided.

Existing legislation allows for a wide range of tenure forms, ranging from individual, freehold title to regulated leases to permits to ‘communal tenure’ (Moyo 2008). Several contributions point to the formal lack of clarity that exists over land rights on the FTLR schemes: A2 beneficiaries were supposed to be allocated land after successful applications, and then receive tenancies on a 25 or 99 year lease; many received land other than through that process and few have yet got leases. A1 allocates were only entitled to ‘permits’ to use land, comparable to OR land, which were not transferable and could be withdrawn.

Conclusions

This attempt to bring together some of the concrete investigations that have been conducted in Zimbabwe in the last decade, and to bring out in the last section above some of the key findings, will be rounded off by posing two issues about the further implications of this work. The first is to ask what lacunae are revealed by laying out this cross section of work. What further investigation is needed to fill crucial gaps or to resolve apparent differences in findings? The second matter is to review how far these materials illuminate the long term prospects for Zimbabwe after FTLR, matters that relate to current policies but go far beyond them.

Future research priorities

There is a clear need for widening of empirical knowledge, both geographically and thematically. The most obvious spatial gap in knowledge is with regard to Matabeleland: there are no case studies in the Matabeleland North Province, and only some limited surveys by AIAS and Ruzivo Trust in the Mangwe District of the Matabeleland South Province. The logic for more case studies goes beyond simply having a more representative area sampling. These provinces are distinct environmentally; most of them lie in the drier Natural Regions IV and V. Moreover, the large scale commercial enterprises that have been absorbed within FTLRP were very largely devoted to livestock ranching, which in turn implies a different kind of transformation in agrarian structure. How far has that outcome been the spawning of small or middle size commercial ranches, solely devoted to livestock and based on a logic of optimising off-take of beef for internal or export markets? Or is a more mixed farming pattern emerging based on a herding strategy of meeting local milk and meat demand, and supplying draught oxen and other four-legged inputs and complementarities to appropriate cropping? But two dimensions of the political context are potentially determinant of outcomes, likely to be different from those in other former ranching areas like those in Masvingo. First, the OR Programme had been a contentious issue. The fighting between ex-fighters of ZAPU and the new Zimbabwe army and the eventual brutal repression of civilians generally in this region in the 1980s interrupted implementation until 1987. Much land purchased by GoZ remained undistributed for several years; in particular the Model D scheme which was to make ranch land available as an extension of grazing to existing CAs was not implemented for several years and was resisted by local communities, as was Model A resettlement (Alexander 2006). The contemporary variant of Model D always seemed likely to be a target for resistance. Three-tier schemes have been limited to Matabeleland, and as far as our studies report, to two districts in Matabeleland South, Mangwe and Matobo. But it seems they did not benefit small commercial ranchers, as intended. Local pressures forced a shift to what amounted to open access to those with livestock in neighbouring CAs. But more investigations are needed into what mechanisms have been set up to manage these new grazing areas, and whether they work. Assessment of the Three-tier variant is in turn just one of the questions that need to be answered as part of the second difference in context: what are the consequences of introducing FTLRP in a region that has a long history of opposition to the former ruling party?

A more general imperative for widening the territorial reach of case studies would be that of carrying out new ones outside the central core of Mashonaland. Researchers have tended to do field work in areas situated conveniently close to Harare – for the same reasons as the national politically influential see the desirability of being within daily commuting range of the city. This proximity is itself likely to mean such samples are biased toward certain kinds of beneficiaries, as well as being typical only of areas with good infrastructure, reliable climate, more advanced agriculture (with all that means in terms of challenges and potential). A comparison even of the studies presented here does suggest that more peripheral areas – like Masvingo Province (by Scoones et al.), the Mangwe surveys in Matabeleland South by AIAS and Ruzivo (only reported on in comparative tables in this collection by Moyo, but see AIAS 2009 and Matondi forthcoming), Chipinge in Manicaland – have a smaller proportion of A2 farms and beneficiaries with lower education and that are more likely to be from CAs. There is also a need to study production and livelihood outcomes in these marginal areas but also in middling areas of NR III which are also underrepresented. Of course national surveys, such as the proposed Land Audit, but also ones that cover land use and economic outcomes, will also be invaluable in giving a representative picture in years to come.

The set of studies presented here do cover a wide range of themes: mechanisms of land allocation, origins of beneficiaries, their political and occupational status, the fate of special categories – women (see also WLZ 2007), farm workers – in FTLRP, the patterns of land use, of livelihood generation, the economic constraints and potential, the evolution of new social relations and conflicts. A range of possible broadenings of those agendas for research can be mentioned. Although many studies provide evidence of the extent to which women have been among official and unofficial recipients of land and a few, like Mutopo, have explored how some women have pursued livelihood options, other issues of gender can be added. For instance, in reports on access to land for women, that category is seldom differentiated. More exploration of sub-categories such as women heads of households would be valuable, as well as the question of who is responsible for managing new farms. Matondi (forthcoming) quotes the bare statistics that women do so in 46 percent of A1 plots, but more such data, including probing decisions about land use, investment, etc., are required. More broadly there is need to investigate whether the division of labour and gender relations are in any way different from those in CAs, and to link that analysis with changing patterns of migration and availability of other livelihoods. Such matters are also vital in exploring policy options for (re-)developing the fast tracked land and the rural and national economies.

Another dimension of ‘divided families’ that demands investigation are relations between the FTLR areas and the CAs: how widespread is the maintenance of two homesteads and is this changing? What economic and social linkages are there? For instance, the AIAS Baseline Survey (2007) reports that in their Mangwe sample 31 percent of new farmers keep livestock belonging to others. Is this common elsewhere and on what relational or contractual basis (mutual aid, kin obligation, rental)? What patterns of management of common land in A1 villagised schemes, perhaps in conjunction with patterns in nearby CAs, have emerged; in particular how do the special cases of Three-tier schemes manage their affairs? More generally there is need for studies within the CAs, of the impact of FTLRP to assess such issues as to whether FTLR has realised any ‘decongestion’.

A central question with policy implications is to chart changing patterns of production and marketing. Only the three longitudinal studies of the institutions (Moyo and AIAS, Scoones et al. and Matondi and Ruzivo Trust) have offered detailed quantitative evidence on such matters. A replication of such studies would aid future planning and would also thoroughly test Moyo's hypothesis that ‘an anticipated corner has been turned’ and production levels and number of benefitting livelihoods are increasing.

Future scenarios

Recovery of economy and livelihoods?

A number of recommendations for action are made in the contributions, some at odds with each other. It is to be hoped that this collection can prompt some potential applicable ideas for policy but also further debate around options – especially about ways forward beyond what most contributors see as a policy ‘impasse’. There are several suggestions for improving performance, livelihoods and daily life in fast track areas: infrastructure on and off farms, timely supply of inputs and credit, extension advice, technical improvements appropriate to new types of smallholder and middle-scale farms, farmers' organisations and social institutions. Most contributors, implicitly at least, seem to believe that past experience in Zimbabwe and favourable land resources indicate that another ‘agricultural revolution’ is a possibility (c.f. the challenging title of Rukuni et al.2006). There is also agreement that the realisation of any such potential depends on a political process as well as technical prescriptions. Here a major divide then opens up: some contributors, like other writers, feel that a major shift in the holders of political power and in reform of existing institutions is a prerequisite to any prospect of implementing alternative approaches (see Habib 2011). Similarly, donors seem to believe it would be premature to commit resources to the present government, even after coalition, and to the FTLR sector.

Even bilateral and international agencies are providing significant amounts of farming and livelihood support in Zimbabwe but apart from UN agencies they still do not engage with GoZ to provide input support as opposed to emergency aid. Moreover, their support programmes to small farmers through NGOs do not touch those smallholders in the FTLR areas as these lands are still legally contested. The continued application of such conditionality is not a legal or logical formula, and flies in the teeth of the principles accepted by both sides of the GPA. Farm support through NGOs cannot provide some of the essentials for the sustained period needed for recovery – like research, extension, and widespread credit (Tupy 2007).

On the other hand Moyo and Scoones et al. among the contributors propose that political reform with emphasis on rights and redress as well as redistribution is essential – but as an accompanying dimension of technical and economic recovery and improvement not as a preliminary stage. Moyo's recommendations include the view that the nature of politics is taking new forms going beyond a one-dimensional inter-party struggle to a more complex expression of class struggles involving chiefs, lineage leaders, farmers old and new, bureaucracies and old and new international economic actors. Both he and Scoones et al. call for strong cooperating and representative organisations of producers, as does Murisa.

Future land tenures?

As indicated above the tenure forms on both A1 and A2 schemes have a temporary nature and will need eventual regularisation: on the former allocatees simply have permits or even only letters of allocation; on the latter leases are promised and have yet to be issued. But alternative measures for FTLR land are being considered alongside a process of enacting reform of tenure governing other sectors, remaining large commercial farms and estates, communal land and OR. In current debates, some analysts advocate forms of freehold title not only for fast track land holders but for all landholders, mainly through conversion of leaseholds and permits to Deeds of Grant, with an option of maintaining the 99 year leases for A2 farmers (Rukuni et al.2009). For others, leasehold and permit systems for the moment offer working security for A2 and A1 farmers respectively (Scoones et al.2010, Moyo 2008). Reform of communal tenure, which was last enacted as recently as the 1998 Act which involved traditional leaders in the allocative role of local government, is also being debated. A hybrid approach that combines communal tenure with legally binding arrangements as recommended by the Rukuni Commission of 1994 (Rukuni 1994) is still favoured by some, though no longer by Rukuni (2009) himself. The present ambiguity of whether traditional authorities also have some powers in Old Resettlement Areas, as stipulated in the Traditional Leaders Act 1998, or even by extension to FTLRP farm areas, needs clarification and is further complicated by proclamations of new regulations under the 1998 Act. Moyo (2008) and Scoones et al. (2010) also advance proposals under which communal areas and those newly made available through state action like FTLRP should be subject to some form of community control and not completely individualised and alienated for all time – but stipulate that such control should be through some democratic form rather than through ‘traditional’ leaders. (For a survey of earlier patterns and proposals about land in communal areas see Ranger 1993; for options being considered in contemporary debates see MLLR 2009). The Brooks Poverty Institute report of 2009 called for the creation of a unified land tenure system, but held back from advocating any particular tenure form (Chimhowu 2009). Of course all land tenure systems exist in a wider political-social-economic context, and the Zimbabwean experience clearly demonstrates that tenure security depends crucially on the wider political setting, land administration capacity and the ability to assure the rule of law. Without these, no tenure regime can assure security.

Similar debates over having a single land tenure system have also been central features of debates in Namibia and South Africa, although both have resisted arguments for complete individualisation across the board. How some form of community voice to temper individual rights should be enshrined in some parts of the system of law and administration over land resources remains unresolved, however, and in South Africa there is at present a vacuum as their Communal Land Rights Act to follow a traditional leaders formula for this has been declared unconstitutional.

There is no clear-cut evidence available in these or other studies that could definitively settle these debates as to the most appropriate policy, even for the land reform areas, nor is it likely that survey data of the here and now could ever be a foolproof guide to future policy (Andersson 2007). However, the material does help illuminate the realities on the ground as opposed to the political or legal formulae. Comparative studies that have sought to compare outcomes from different tenure systems can be potentially instructive although the Africa material is equally inconclusive (Bruce and Migot-Adholla 1994, World Bank 2003) – although they do tend to cast doubt on the idea that there is one ideal solution. An informed debate in Zimbabwe could well look particularly at one country's experience. Kenya was also a settler colony but is the only country in sub-Saharan Africa that has implemented individual freehold titling, both in the communal areas and resettlement land holdings, and thus where there is historical evidence over 50 years of what does happen when that happens (Shipton 2009, Mckenzie 1993). Wanjala (2000, 98) characterises the situation that developed in Kenya when the lending, which individualised titling of land was supposed to promote, led to massive defaults:

As the banks have moved in to realize their securities by exercising their statutory right of sale, the social consequences have forced the government (the very advocate for the system) to come up with ambiguous directives through the office of the Chief Justice directing that no sales of agricultural land should be conducted unless the Provincial Administration has given its consent.

Thus individual tenure has not facilitated agricultural credit nor provided security of tenure through the market, and does not even operate in practice (see also Berry 1993, Platteau 2000, Mackenzie 1993). A cautionary tale for Zimbabweans.

Future dynamics of reformed agrarian structures

These contributions concur that there has been major transformation of the agrarian structure in Zimbabwe since 2000, and that the central characteristic of that outcome is that smallholder production now predominates. But these structures are not yet set in stone. Other possible long term scenarios might emerge from new forms of class struggle, as Moyo anticipates, by efforts of the beneficiaries themselves and by the unresolved political struggles for reform.

Scoones et al. point to one emerging dynamic in on-going agrarian restructuring by pointing to the range of economic units and associated classes which have emerged from FTLRP. There are specialist entrepreneurs providing services such as mechanised operations and technical advice to down-sized commercial farms and A2 farms, as well as intact corporate estates and ranches; a wide range of types and strata of smallholders, and of permanently and casually employed types of workers reproducing themselves in a variety of ways and located on existing large farms, in the interstices of new communities. Anecdotal evidence indicates that there have been many unpredicted social and economic interactions between these elements: former commercial farms do contract ploughing for people who occupied their land; different types of farmers periodically swap use of land rather than set aside ‘fallow’ for tobacco; former farmers and managers act as ‘mentors’; new marketing mechanisms have evolved like the side-stepping of the classic tobacco auctions by state capitalist dealers from East Asia. Further blossoming of such interchange will no doubt emerge spontaneously but the multiplier effects could be increased with informed planning and creative institutional flexibility.

Several contributions report another dynamic: continued struggle to confirm de facto or assert new rights to land, especially efforts by large A2 owners, often politically connected, to oust smaller A2 and A1 farmers. Moyo stresses how this is likely to persist and thereby poses a question of whether encroachment by elites will reach a scale that represents a further structural transformation. One future scenario would be that such accumulation by a new class could signal a return to a basic dualism: black capitalist farmers replacing large scale white capitalist farmers. But some of our studies suggest a more complex future: evidence of A2 farms leasing out their unutilised land to get an immediate return. Such a trend might be reinforced as A1 farmers seek to defend their terrain, and their off-spring and land seekers from the CAs use squatting as a tactic for acquisition. It is possible that actual long term outcomes might be a complex combination of forms of landlordism together with partially successful resistance by existing and would-be smallholders. In general terms this has been the path that has emerged in Kenya but what actual mosaic may arise in Zimbabwe will depend on different overt political struggles.

Notes

¹Honourable exceptions who sought a more nuanced middle ground did exist, notably in the attempt by Hammar, Raftopoulos and Jensen (2003, 17) ‘to expose and interrupt such rigid polarities and the essentialisms that underpin them’, and in Campbell (2003) and Alexander (2006).

²For introduction to the considerable literature on reserves, see Palmer (1984).

³Tendi (2010) offers a new explanation for GoZ delay in readdressing land reform in the early 1990s, in addition to the lack of political will or of finances suggested in other analyses: that it was a forswearing by GoZ as a result of a covert understanding, brokered by Commonwealth Secretary-General Anyaoku, to rein in land reform to ease the climate of opinion during South Africa's transition from apartheid. His interviews with Anyaoku and ZANU-PF leaders seem to confirm there was some such agreement, but that does not rule out other explanatory factors, nor is it easily compatible with accounts that suggest GoZ was reluctantly pushed into the FTLRP initiative by pre-emptive action by the liberation war veterans, as argued by analysts such as Sadomba (2009).

⁴The notorious November 1997 letter from Clare Short, UK Secretary for International Development in the new Labour government, discussed in Tendi (2010, 87–93), stated, ‘we do not accept that Britain has a special responsibility to meet the costs of land purchase in Zimbabwe’. After further attempts to resolve the differences, the then Minister for Africa, Peter Hain, told the House of Commons Foreign Affairs Committee in May 2000 that release of funding was suspended because the Government of Zimbabwe was running ‘a corrupt, inefficient land reform programme’.

⁵Two other sets of case studies are available under the Livelihoods after Land Reform programme: http://www.lalr.org.za/zimbabwe/zimbabwe-working-papers-1

⁶Access to this and related material: www.zimbabweland.net