Zimbabwes Land Reform Program: Underinvestment in Post … · 2005-03-07 · Zimbabwes Land Reform Program: Underinvestment in Post-Conﬂict Transformation BILL H. KINSEY * Free

World Development Vol. 32, No. 10, pp. 1669–1696, 2004� 2004 Elsevier Ltd. All rights reserved

Printed in Great Britain
0305-750X/$ - see front matter
doi:10.1016/j.worlddev.2004.06.005www.elsevier.com/locate/worlddev

Zimbabwe�s Land Reform Program:

Underinvestment in Post-Conflict Transformation

BILL H. KINSEY *

Free University Amsterdam, Amsterdam, The NetherlandsUniversity of Zimbabwe, Harare, Zimbabwe

Summary. — In Zimbabwe�s current crisis, it is easy to overlook the fact that the country had aresettlement program for two decades before the large-scale, politically motivated land occupationsbegan in 2000. This paper does four things. First, it creates an historical bookmark for the earlierperiod of land reform in the hope that the lessons from that experience will not be lost. Second, itreviews some of the major outcomes of resettlement from almost a quarter century of research.Third, it partially sets the stage for some of the papers that follow in this section. Finally, as a cau-tion for those who will shape Zimbabwe�s future, it provides some reminders of the interlockingrelationships among property, poverty and conflict.
� 2004 Elsevier Ltd. All rights reserved.
Key words — Southern Africa, Zimbabwe, land reform, conflict, resettlement

* Final version accepted: 15 June 2004.

1. INTRODUCTION

The agrarian question and politically moti-vated violence are Zimbabwe�s most enduringcolonial legacies more than two decades afterindependence. The two are intimately linked.For the some 90 years the country was a settlercolony, the adoption of discriminatory agricul-tural policies and the alienation of most of thefertile, well-watered land to European settlersresulted in the oppression, marginalizationand impoverishment of indigenous rural peo-ple. In an effort to redress these inequities, thegovernment of Zimbabwe swiftly introduced aseries of agrarian reform measures after inde-pendence. The resettlement activities initiatedunder these measures peaked—well below tar-geted levels—just after the mid-1980s. There-after real progress slowed and commitmentweakened to sporadic administrative and legis-lative efforts to modify the operating environ-ment for resettlement.Following the rejection of the proposed new

constitution in the February 2000 referendum,a constitutional amendment and a modifiedLand Acquisition Act were promulgated inApril 2000 to effect land designation and com-pulsory acquisition without compensation.The same bill also declared Britain ‘‘liable’’ to

166

pay compensation. The British government—along with other donors—had indicated a will-ingness to fund land reform, but only if itbenefits the poor.Proponents of land reform have argued for

an expansion of the resettlement program tohelp redress the unequal distribution of landresources, to rectify acute land scarcity in com-munal areas, and to provide economic opportu-nities in a shrinking economy. Opponents offar-reaching land reform have asserted thesuperior efficiency of the commercial farmingsector and the adverse consequences that an ex-panded resettlement program would have onagricultural output, employment and the com-position and volume of agricultural exports.A government-appointed commission of ex-perts studied Zimbabwe�s systems of land ten-ure and detailed possible options in its 1994report. Few of its recommendations have beenacted upon.It is generally acknowledged that the nature

of international conflict has altered fundamen-tally, with intrastate civil conflicts replacingwars between states. More than 90% ofthe wars over the past decade have involved

9

1670 WORLD DEVELOPMENT

political violence between adversaries at thesubstate level fighting primarily within theboundaries of a single state (Jackson, 2001).In Africa, this has been the case for some 40years. Indeed, as Addison (2001, p. 1) remarks:‘‘Africa has become synonymous with conflict.’’Recent work suggests that many current con-

flicts differ from a breakdown of normallypeaceful political systems. 1 Rather, currentconflicts are often created civil disequilibria inwhich violence performs a variety of functionsin parallel, alternative systems of power, pun-ishment, profit and protection. In this view,the everyday politics of weak states providesthe soil in which many contemporary conflictsgrow. Biases in public spending, predatory tax-ation, and bad or shortsighted policy encourageconflict by reducing the real incomes—bothabsolute and relative—of groups in society thatsuffer this discrimination. In this way, weakstate politics inflames ethnic, economic andregional tensions—thus helping demagoguesrecruit and retain their followers. Addressingconflict within rather than between states re-quires new conflict management strategies.The rapidly growing literature in this areasuggests that new modalities of dealing withconflict need to be oriented toward the recon-struction and reformulation of weak state polit-ical practice and aimed not so much at themanagement of conflict as at its transforma-tion.The foundation that underlies the papers in

this collection is the role of land reform intransforming conflict in agrarian economiesin southern Africa, specifically in the economyof Zimbabwe. Originally, the focus was the roleof land reform in preventing conflict. When theanalyses here were being formulated in late1999, it seemed evident that the slow rate ofprogress on land reform in Zimbabwe couldnot continue. Numerical targets for the num-bers of households to be resettled had beenchiselled into political granite in the early1980s. Almost two decades later, however,nothing like the target numbers had been reset-tled. Yet lip service was regularly paid to landreform in the period just before every parlia-mentary and presidential election after inde-pendence. But few of the promises made inthese political campaigns were kept. After anauspicious beginning in the early 1980s, land re-form moved at a snail�s pace for some 15 years.With another parliamentary election due in2000, it seemed obvious that Zimbabwe�s rulingparty—ZANU-PF—would again herald its

commitment to providing land to the people.What was less apparent this time was whetherthe people would accept another round ofempty promises from a government that wasrapidly losing its credibility and acceptance.Multiple perspectives help explain why the

papers in this section have been written. Therewill be those who will read it solely as a re-sponse to Robert Mugabe�s political megalo-mania, to his cynical manipulation of the landissue as an election-winning tool. There is valid-ity in this perspective. The long history of bro-ken promises over land posed a clear threat tocivil order in a setting where economic condi-tions were deteriorating rapidly also. From thisperspective, it seemed sensible to try to make acase for Zimbabwe�s government to talk lessand do more about land reform. But equallyas valid is the view that the case for land reformneeded to be reinforced because of the failure ofZimbabwe�s friends and sponsors—the multi-lateral and bilateral agencies especially—to per-suade Robert Mugabe that land reform was tooimportant to be neglected.Zimbabwe did have an agrarian reform pro-

gram before 2000, but current debate is domi-nated by the nature of the multiple complexcrises that face the country. The papers in thissection therefore draw upon data spanningmore than 20 years in a search for understand-ing about what took place during Zimbabwe�sGolden Age of resettlement, before the govern-ment lost its conviction in the central role ofland redistribution in alleviating poverty. Thetwin themes are ‘‘what was’’ and ‘‘what couldbe.’’ We focus little on the history, underlyingresentments and the trigger mechanisms thatunderlie the current conflict. Similarly, weavoid most aspects of the unfolding crisis ex-cept certain ones that tell us how the door tothe past was closed. There is a growing bodyof literature that is highly variable in both qual-ity and the coverage of substantive issues, butthe definitive account of land in Zimbabwehas yet to be written. 2

This paper seeks specifically to do two things.First, in an attempt to test the intentions of theearly program—and implicitly to allow a con-trast with the present one, the paper identifiesthose who received land in earlier years. Sec-ond, the paper seeks to answer questions aboutwhat happened to those resettled in the early1980s. Before turning to these two empiricaltasks, however, some background is presentedto the broader issue of land reform and conflictin Zimbabwe.

ZIMBABWE’S LAND REFORM PROGRAM 1671

2. THE MAKING OF AN IMPASSE: LANDAND POLITICS AT THE MILLENNIUM

The 11 decades since the first colonial intru-sion have seen the creation in Zimbabwe ofmultiple tenure ghettos. The pattern of landdistribution at the turn of the millennium hadits roots in the 1890s, when an enforced racialdivision of land was first implemented. Thisdivision was formalized in 1930, when thecountry�s land was divided roughly equally be-tween the African majority and a relative hand-ful of European settlers. The fact that theEuropean population never constituted asmuch as 5% of the total population illustratesthe great disparity in the amount of land avail-able to each group.After independence, Zimbabwe, on the

strength of the historical record of high agricul-tural productivity, was assigned the food secu-rity portfolio in the then Southern AfricanDevelopment Coordination Conference. Thewhite commercial farmers believed the reassur-ances of the government that they were impor-tant to the achievement of national andregional economic objectives. With the decisionperhaps assisted by the inability to emigratewith their capital (see Davidow, 1983), thewhite farmers who were in Zimbabwe afterindependence generally stayed on. In the fol-lowing two decades, they transformed the oldsanctions- and war-constrained agriculturaleconomy dramatically. Nevertheless, the newgovernment honored its liberation war prom-ises by swiftly launching a land resettlementprogram based initially upon land abandonedduring the war. This program peaked by themid-1980s, and the number of families resettledeach year was to decline to what was a tricklecompared to the targets set. 3

The government of Zimbabwe increasinglylost all but rhetorical interest in the entire issueof land reform. The evidence comes in severalforms. The first is that, even before the match-ing grant funds from the British came to an endin the late 1980s, budgetary appropriations forresettlement—always inadequate—had begunto decline. Government�s commitment to landreform—whether as a way of addressing pov-erty or enhancing resettled farmers productivityand enabling them to participate in export mar-kets—dwindled very rapidly. When, at the endof the first phase of the program, the Britishfunds had not been exhausted, the British aidagency invited the Zimbabwe government tosubmit proposals for a new phase into which

the residual funds would be rolled. No proposalwas ever submitted, however, and the accountswere eventually closed with a positive balance.Instead, the government was turning its atten-tion increasingly to rewarding its political alliesby creating a body of black commercial farm-ers.Budgetary resources were never available to

do what was promised 4 and, before the gov-ernment resorted to hastily scrapping any lawthat got in its way, it failed repeatedly to appre-ciate the amount of time necessary to complywith the provisions of previous laws it hadintroduced. 5 This was particularly the case inlate 1997, when increasingly under pressure,the government designated 1,471 farms inOctober but failed to appreciate that therewas inadequate time to complete all the proce-dures, including appeals and responses, by thelegal deadline—especially given the well-knowntendency of the civil service to virtually shutdown in early December for a long Christmasholiday.Various commentators have defended the

government�s inaction on the grounds that itcould do nothing because its hands were tiedby the provisions of the Lancaster House Con-stitution. While this defence excuses an inabilityto act via certain modes of land acquisition, itfails to explain government�s inattention toalternative methods of acquiring land, such asthrough a land tax, reparations, reclaiming his-toric subsidies, 6 inviting nongovernmentalorganizations to take a more active role inresettlement, providing existing credit institu-tions with a window for ‘‘small’’ land transac-tions, or swaps for long-term governmentobligations matched to a corresponding pay-ment schedule for buyers. 7 In fact, given theincreasingly inadequate budgetary support forexisting resettlement areas, the governmentwould merely have been further embarrassedhad additional land been available for resettle-ment.For many years, the government of Zimba-

bwe appeared not to regard the land issue aspotentially conflict-inducing. Indeed, the Zim-babwean experience provides a sorry tale ofthe unintended consequences of flawed institu-tional solutions. Policies intended to regulateland use and ownership rights have often ledto a vicious circle: bungled strategies escalatedinto conflicts between ‘‘squatters’’ and land-owners; government agencies were then pres-sured to operate at levels for which theylacked the appropriate resources, which in turn


motivated large landowners to diversify intogame-ranching and wildlife management toprove they were putting their land to use toavoid expropriation.Equally, there are growing tensions between

the new landed elite and those whose pleasfor land have been ignored for so long. At thelevel of discourse that centers on amalgamatingdifferent tenurial systems, the current land re-form policy debate is caught between a rockand a hard place. On the one side lies the im-pulse to ‘‘modernize,’’ so that black agriculturemay contribute to exports and economicgrowth. The drive here is to utilize incentivesto individual producers to achieve greater pro-ductivity and maximize efficiency and profits.Security for this group of producers lies notin the legalities surrounding the tenurial systembut with perceived political loyalties. One unin-tended but inevitable outcome, however, isgreater social differentiation. On the other sidelies not so much a production system but rathera set of survival mechanisms that emphasizesaccumulation, community, security and equityof access. 8 The security of the communal resi-dent is founded on membership in a socialgroup and possession of an inalienable rightof access to share the group�s land. Almost aquarter of a century following independence,agriculture in Zimbabwe is scarcely any lessdualistic than was the system inherited in1980, even if race is no longer the differentiatingfactor.Fifteen years after it first became operational,

Zimbabwe�s land reform program ‘‘had neverbeen evaluated to determine the strengths andweaknesses as well as the need for its redesign’’(Zimbabwe, 1995, p. 2). Despite the absence ofan overarching performance review, majordevelopments took place throughout this peri-od that had significant effects on the policiesand procedures that shaped the program.Examples include the first and second five-yearnational development plans in 1986 and 1991(Zimbabwe, 1986b, 1991), the adoption of anew land policy in 1990 (Zimbabwe, 1990),enactment of a new land acquisition act in1992 and its revision in 1996, the introductionof structural adjustment in 1991, the hearingsof the Land Tenure Commission in the early1990s (Rukuni, 1994), and major revisions toland policy (Zimbabwe, 1996).The first substantial redesign of the program

(Zimbabwe, 1992b) also took place in the faceof inadequate evidence on what was happen-ing at the level of the resettled household.

Although survey programs were launched thatwere intended to be continuing exercises, thesewere never given sufficient resources to operateon much more than an ad hoc basis. Most so-called surveys were in fact merely reviews ofthe status of implementation at any level lowerthan headquarters. 9 It is not surprising there-fore that the redesign failed to address manyof the deep-rooted problems of the resettlementprogram.The resettlement program was implemented

very rapidly during the early 1980s but sloweddramatically thereafter. The reasons why thepace of the program changed are complex;among them, however, was a flood of criticismsand negative evaluations—from both withinand outside government—that the programhad failed to have a positive impact on agricul-tural productivity and rural incomes. 10 Suchconclusions, however, based on evaluations ofoutcomes over short-term horizons, are likelyto miss many of the beneficial impacts thatemerge only as communities organize them-selves internally and engage fully with the out-side world of support services and economicagents (Scudder & Colson, 1982). Thus, espe-cially in a context of great seasonal variability,conclusions based on single cross-sectional sur-veys are likely to misstate seriously the out-comes arising from land reform.Past agrarian policies in Zimbabwe have been

intended to encourage the modernization andgrowth of the agribusiness sector and/or to en-hance state power. These policies—with theexception of those during the approximate per-iod 1980–87—generally have not benefited theZimbabwean peasantry; rather, they havehelped to polarize national and local politicsand create extremes of wealth. Consequently,Zimbabwe is now in political and economiccrisis. Since the early 1980s, state-directed polit-ical violence has increased. Although this vio-lence has yet to reach the scale observedduring the early 1980s in Matabeleland, it ismore pervasive and extends seemingly to everycorner of the country.

(a) Other arguments in the land reform debate

Neoclassical analysts, characterized by Leh-man (1978), concentrate on the inefficient allo-cation of productive resources associated withfactor market imperfections. If only large-scalelandowners can gain access to credit, and ifsmallholdings operate more efficiently than lar-ger farms, then efficient producers may be de-


nied capital. Redistribution of land may there-fore be justified as a means of improving sub-optimal resource allocations. Other analystsfalling within this category include Griffin(1974), Griffin, Khan, and Ickowitz (2002),Lipton (1974), Schultz (1964) and Warriner(1969).Related rationales for redistributive land re-

form are the observed underutilization of landon large holdings and the inverse relationshipbetween farm size and unit yields. Historically,the underutilization issue has been of particu-lar importance in Latin America, but it is alsoimportant in southern Africa. The evidence forZimbabwe is reviewed by Weiner, Moyo,Munslow, and O�Keefe (1991). The empiricalevidence on the inverse farm size-yield rela-tionship is reviewed in Dorner and Kanel(1971) and Berry and Cline (1979). The mostcomprehensive recent treatment of thesethemes is found in two issues of the Journalof Agrarian Change. In the first, Griffin et al.(2002) make a case in favor of radical redis-tributive reform, arguing that mere tenurial re-form can never produce a superior outcome,and might even worsen matters, while ‘‘market-led’’ land reform is also rejected as a realisticsolution to the problems of inequality, povertyand growth. The second issue 11 contains a setof papers that challenges the proposition thatsmall farms are more efficient than large ones,and contains two papers (Sender & Johnston,2004; Bernstein, 2004) addressing some of thecomplexities and contradictions surroundingland-related issues in Zimbabwe.The inverse relationship has not been rigor-

ously tested in Zimbabwe, but it is unlikely thatit would be discovered because large-scalefarms have always benefited from vastly supe-rior access to inputs and technical services.A recurring theme in the land reform litera-

ture is the function of reform as an escape valvefor class tensions. Since the middle of the lastcentury, all major land reforms have occurredin the context of political crisis, imminent revo-lutionary threat or post-revolutionary ‘‘adjust-ment.’’ In virtually all cases, land issues havebeen central to the crisis.Attempts to subject the reform-conflict rela-

tionship to formal analysis have, however, beenless than satisfactory. Several approaches havebeen made to identify thresholds of landless-ness beyond which conflict is inevitable. 12 Rus-set (1964) utilizes a simple Gini coefficient in anattempt to identify a critical level of distribu-tional inequality. Prosterman and Riedinger

utilize an array of measures of deprivation,including measures of inequality in land hold-ings plus indices of physical quality of life, civiland political liberties, and demographic factors.Other works devoting particular attention tothe links among landlessness, reform and con-flict include King (1977), Paulini (1979), andTuma (1965).More recent work includes the two-volume

comparative study assembled by Nafzigeret al. (2000) that traces the economic, socialand political roots of conflict that leads tohumanitarian emergencies. They emphasizethe significance of protracted economic stagna-tion and decline, high and increasing inequal-ity, government exclusion of distinct socialgroups, state failure and predatory rule. 13

The central strategic choice lies between pov-erty reduction through faster economic growthand reduction through redistribution—although these may, of course, be complemen-tary. The relationship between growth andpoverty was explored in depth at the WorldInstitute for Development Economics Re-search in Helsinki in mid-2001. Althoughmany of the papers highlighted the need forserious research on the effects of redistributionon incentives, investment and growth, a pow-erful consensus emerged that (i) a redistribu-tive growth path is always likely to besuperior to a distribution-neutral path (the dis-credited notion of ‘‘trickle-down’’) for reduc-ing poverty; (ii) a redistributive growth pathis always superior if a country�s per capita in-come and inequality are relatively high; and(iii) a once-for-all redistribution from the richto the poor is superior to a protracted redistri-bution process in its effect on poverty inmost cases. 14 Further, empirical work isbeginning to make clear that the impact of pol-icies such as adjustment hinge critically on thenature of the asset distribution in a country atthe time they are introduced.Shifting alliances in post-revolutionary Zim-

babwe are outlined by Herbst (1989). Localeconomic and political elites, he notes, can callupon the repressive apparatus of the state forsupport when they feel challenged, as has hap-pened frequently since 2000. But local elitesmay also protect the poor and landless againstthe predations of the state. Local elites may aswell have their own grievances against the state.And because they are local, they are sur-rounded by the more numerous poor. In certaincircumstances, governments or national politi-cians may intervene to vindicate the rights of


the poor against abuses by local economicelites, although examples are rare.In this same vein, Esman (1989) asks how

defensive resistance to abuse gets turned intomore positive assets of group interest. By theirown actions, he questions, are the weak anddisadvantaged able to achieve a more ‘‘tolera-ble accommodation’’ to harsh realities thanwould otherwise be possible? He refers to theaxiom of contemporary political thought thatthe weak can be empowered only when theysucceed in capitalizing on their superior num-bers through organization—political, economicor cultural. This is precisely what the groupsidentified by Bernstein (2004)—particularlycommercial farm workers and women—havenot yet been able to do. But, Esman cautions,experience shows that ‘‘Effective organizationis likely sooner or later to produce confronta-tional tactics. Such tactics are inherently riskyand dangerous even when they are nonvio-lent, . . .’’ (p. 226). 15 Various forms of minorcivil disobedience may be useful where a weakstate is unable to compel compliance or isunwilling to confront groups directly that itcounts among its constituents—such as hap-pened with Zimbabwe�s war veterans. Nor-mally, an effective state can override suchtactics at minor cost in supporting determinedpolitical elites. 16

Jacobs (2000) notes how the various claimsmade around land reform employ the discourseof democracy as validation. She offers as exam-ples Mugabe�s claims that actions are justifiedby the historic wrongs done by the appropria-tion of African lands, and the war veterans� val-idation for their actions in terms of theirdeserving compensation for helping to van-quish settler society and undemocratic rule.At the other pole, white farmers� claims are jus-tified in terms of commercial agriculture�s cru-cial role in the economy—and implicitly asevidence of a structure in which democraticsociety can exist.Thus, institutional reforms and revolution-

ary changes may produce some significantgains, such as land redistribution or civilrights, for subordinate groups. Sooner or later,however, emergent elites will consolidate statepower in a new or reformed social order,where some will be politically and economi-cally strong and the majority relatively weak.New elites, among them large-scale blacklanded property owners, are jockeying moreintensely for power in Zimbabwe with eachpassing day.

(b) Temporal dimensions of land reform

Comparisons of land reforms in the past cen-tury reveal tremendous variation in the pace ofimplementation. On the one extreme are thevery rapid post-war reforms of Japan, Koreaand Taiwan, while at the other are the pro-tracted reforms in India and Pakistan that haveproceeded haltingly for more than five decades.Between these extremes lie numerous interme-diate cases, including some where land reformshave been reversed altogether.Considering the economic motivation for

land reform, it is difficult to justify implementa-tion trajectories that span long or indetermi-nate periods of time. If an alternativedistribution of land were considered desirableon efficiency grounds, then a prompt redistribu-tion would appear to be better than a laggedone. In reality, of course, there are many fac-tors that can retard reform based on redistribu-tion. In Zimbabwe, there can be little doubtthat limited administrative and bureaucratic re-sources have precluded more rapid implemen-tation. Even in the best of circumstances, theadministrative requirements of land reformcan be immense; and the technocrats responsi-ble for implementing Zimbabwe�s programhave had to contend with almost constant cutsin budgetary resources.The need to make compensation payments

for the land acquired for redistribution may,of course, also constrain the pace of reform.The provisions of the Lancaster House Consti-tution requiring compensation for land are oneof the most cited excuses for the previouslyslow pace of Zimbabwe�s reform efforts; but,as noted elsewhere, this weary refrain may belittle more than a convenient cover-up for inac-tion and lack of imagination.The source material indicates a wide range of

opinion as to the most desirable path of reform,but all underscore the fact that the pace of re-form is one of the policy instruments open togovernments. In Zimbabwe, the pace for twodecades was dictated by a lethargic govern-ment. Now other groups appear in control.

(c) Beyond the zero-sum game

The relationship between reform and classconflict has received a good deal of discussionbut little formal analysis. Reform has beenwidely discussed in terms of its function as anescape valve for class tensions and has beenshown to be an adjustment process by which


material endowments—specifically land—arebrought into balance with strategic capabilities(Horowitz, 1989).Papers in the volume edited by Wohlgemuth,

Gibson, Klasen, and Rothschild (1999) begin toprovide a conceptually sound framework foranalyzing the nature of modern conflict in Africa.They argue that analysis must start from ‘‘thecontext of decreasing state legitimacy’’ (p. 11)central to the weak-state framework. The fac-tors that spark violent conflict are still only par-tially understood, but they lie within theconstrained policy choices available to or uti-lized by the political elite of Africa�s weakstates: ‘‘There can be no doubt that that Afri-can governments and elites have behaved in amanner that has made certain conflicts inevita-ble or have given new life to certain dormantconflicts’’ (p. 16).With this analytical underpinning, it is easier

to understand why the ‘‘traditional’’ approachto conflict management is unlikely to work insettings such as contemporary Zimbabwe.The conflict over land in Zimbabwe shouldnot be regarded simplistically as an ‘‘old-style’’clash, in which two adversarial groups—rich,white, commercial farmers and poor, landlesspeasants—are locked in a battle in which thelosses of one group imply a corresponding gainfor the other. 17 A solution will not be foundin the ‘‘victory’’ of one side over the other—resulting in the transfer of land in a zero-sumoutcome. The reason this is not a solution isthat the transfers are induced by conflict, andthe costs of conflict are borne in a nonzero-sum manner. Importantly, significant costs ofconflict are borne by all participants in theconflict, as well as by varied ‘‘by-standers.’’ 18

Nor is a solution likely to be found throughsome new political accommodation that legiti-mizes one side or the other through negotia-tions or support, thus perpetuating theconflict.The conflict over land in Zimbabwe should

not be regarded as incomprehensible, merelyas some form of anarchy engineered by greedyopportunists. To follow this path is likely tolead to an emphasis on conflict managementrather than resolution and to the tactic of min-imizing social pain through provision ofhumanitarian assistance. While internationalinvolvement in situations involving conflict inAfrica typically wavers between indifferenceand some form of humanitarian intervention,such assistance simply helps to sustain the con-flict (Uvin, 1998). 19

From the perspective of the internatio-nal community, the real issue is the one in thebackground: the feasibility and desirability ofexpropriation and redistribution of land, or tra-ditional land reform. It is easy to make bothvalid economic and equity arguments for landreform in Zimbabwe. Indeed, the phrase ‘‘Eve-ryone agrees that land reform is needed’’ hasbecome a cliche. But rhetorical acknowledge-ment of the need for land reform does not im-ply acceptance of the scale and approaches toland reform proposed. As true land reforminevitably has an element of confiscation, how-ever, it can never be a purely economic policy.Economic policies anyway tend to generate dis-tributional side effects, so that judgements ofefficacy take place in the political arena. Inthe case of land reform, though, the desiredeconomic outcomes are the side effect, whilethe distributional aspects are fundamental. Gi-ven the inherent and dominant distributionalaspect, land reform is essentially political innature. 20

3. WHO RECEIVED LAND IN THE 1980S?

Since early 2000, events relating to land inZimbabwe have followed one another at sucha pace that even long-term observers of thecountry�s politics have difficulty in ascertainingwhat is happening behind the curtain of chaos.Those less familiar with Zimbabwe are some-times prone to judge entire aspects of the coun-try�s post-independence history based onfragmentary reporting in contemporary media.Although reports from many different sourcesindicate that much/most of the land distributedsince early 2000 went, or has subsequentlygone, to a wealthy, politically elite group, weshall probably never know the true facts ofwho received land. It is still necessary, however,both to avoid terminology such as ‘‘the land re-forms’’—as if all land redistribution has pro-ceeded in the same manner—and to specifythe period in which the land in question wasallocated.For one thing, patterns of land allocation

seemed to differ markedly from one part ofthe country to another after 2000. For example,in the Mashonaland provinces, near Harare,where chefs grabbed prime farms for them-selves and often encouraged violence by warveterans, it was clear that one needed to be arecognized ZANU-PF supporter to get accessto land (Palmer, 2003). In other provinces,


especially Manicaland and Midlands, moremoderate local political regimes pursuedmore even-handed approaches. 21 This issue,however, is not the focus here.Some have asserted that land tended to go to

the undeserving from the very beginning of theresettlement program. For example, referringto those selected for resettlement over the firsttwo decades of independence, Sender and John-ston (2004, p. 156) comment: ‘‘There is no rea-son to believe that these beneficiaries wereselected from amongst the poorest rural house-holds living in the highly differentiated commu-nal areas. . ..’’ 22

While certainly true since 2000, to what ex-tent is this allegation accurate for the previous20 years? The material that follows attemptsto answer this question.Alexander (2003) argues that the current

debates over land stand in stark contrast to offi-cial practice over most of the post-independ-ence period. While the ‘‘land issue’’ wasresurrected in the mass media every five yearsfor the purposes of electioneering, 23 the rheto-ric of reclaiming lost land that had animatedthe liberation war was ousted in favor of tech-nical and economic interpretations of land re-form. Emphasis was increasingly shifted tothe ability of the resettlement program to pro-duce marketed surpluses and, to meet thisobjective, the selection rules for beneficiaries,as well as the practices employed, were moreand more handed over to bureaucratic manage-ment. The language that animated conflictingclaims to resources had no place it seems onthe agenda of a centralized, modernizing state.Official criteria for the selection of land re-

form beneficiaries differed slightly during the1980s depending upon the ‘‘model’’ used forresettlement. 24 For the dominant resettlementmodel A, the criteria originally emphasizedneed and gave priority to refugees and dis-placed persons and those with no, or inade-quate, land for subsistence. Settlers were alsosupposed to be married or widowed, aged 25–50 and not in formal employment. Governmentpolicy as well was to move individuals and theirimmediate families—not communities (seeDekker, 2004).Formal criteria determining eligibility for

resettlement were spelled out in several differentversions over the years following 1980. Formore than a decade, these criteria emphasizedprimarily the neediness of applicants, although,as noted, there was a growing stress placed onthe productive potential of those receiving land.

In the early years, in order to qualify for reset-tlement, a would-be settler had to be (Zimba-bwe, 1985, pp. 23–24):

—‘‘effectively landless, i.e., no or little landto support self and dependents;’’ and—‘‘not employed (nor spouse);’’ and—poor: ‘‘The intention is to reach the ruralpoor; not, as many development pro-grammes have, the rural rich;’’ and—‘‘married or widowed with dependents;’’and—‘‘aged 18–55 and able to ‘‘make produc-tive use of the land allocated;’’ and—prepared to give up all rights to land inthe communal areas; or—a returned Zimbabwean refugee, who isgiven special consideration; or—an experienced or master farmer who iswilling to give up land rights and wageemployment elsewhere.These criteria are essentially the same as

those first established in 1980 (Zimbabwe,1980) except for the stipulation regarding anability to use the land productively and the pro-vision made for master farmers. Applicants forland had to meet all the first five criteria and therelevant one of the last three.Beginning in the mid-1990s, preferential ac-

cess to resettlement land began to be given towar veterans as ‘‘a special category’’ (Zimba-bwe, 1996). From the first quarter of 1996, a20% quota was to be reserved on all newschemes for war veterans, but by then therewere very few new schemes coming along.The procedures actually used to select benefi-

ciaries for the resettlement program variedfrom place to place, with some involving a greatdeal of community participation and othersnone at all. In most cases, however, those inter-ested in being resettled were supposed to regis-ter their names on a waiting list, typically byfilling in a widely distributed form. It was fromthat point on that the selection practices dif-fered. In the case of Hoyuyu model A scheme,for example, interviews were then carried outby the settler selection committee at the districtoffice in order to ensure that the ‘‘selected peo-ple were not employed and that they were notin possession of land’’ (Choga, 1999, pp. 24–25). Ex-farm workers of foreign origin werealso considered, and settlers did not necessarilycome from the district or neighboring commu-nal lands, but could come from anywhere inthe country.In other cases, an initial screening of appli-

cants was done within local communities before

Table 1. Areas of origin for settler households inChinyika Resettlement Areaa

Area of origin or

category of household

Percentage of 1,231

households

Local communal areas 40.6

Local landless persons 25.7

Refugees and foreigners 16.2

Commercial farms and mines 9.7

Urban areas 7.8

Total 100.0

a Data represent the original households resettled inChinyika in October 1980 and were provided by theresettlement administration office in Rusape.


applications were submitted for bureaucraticprocessing. Vhutuza describes one such casein terms of the preliminary events as they oper-ated for one model A scheme in MasvingoProvince (1991, pp. 18–19):

Settlers on the dryland areas,. . ., were mostly land-less and war displaced people . . . Traditional chiefsand councillors were actively involved in the identifi-cation of those in need of land. Chiefs and some vil-lage development committee members fromsurrounding areas converged on district and provin-cial offices. . . to collect forms. These forms were inturn given to people classified. . . as needing land.Selection of the needy in the villages was donethrough a public vote. The chiefs called for publicmeetings. . . and it was at these meetings that peopleto be resettled were chosen. As a result in the earlyyears destitutes were resettled. . ..

In this instance, at least, local participationseems to have been remarkably effective inselecting precisely those defined officially asthe target group.There were also cases in some areas included

in the Zimbabwe Rural Household DynamicsStudy (ZRHDS) where the new communitiescreated by resettlement took it upon themselvesto become more inclusive than the official crite-ria allowed. For example, villagers would putpressure on local government staff to createadditional plots to accommodate aging formeremployees on commercial farms who had nowhere else to go. Those in this category wereoften destitute Zambians or Mozambicanswho had lived in Zimbabwe for many years. 25

Local participation did not always, however,have such outcomes. It is worth noting that notall resettlement in the early 1980s was com-pletely voluntary. Many of the households inone of the schemes studied by the author, forexample, did not themselves choose to be reset-tled but rather were ‘‘selected’’ by their formercommunal area neighbors to leave their homesand go away for resettlement. The exact mo-tives for such selection are difficult to ascertain,but they appear to be related primarily to var-ious forms of ‘‘anti-social’’ behavior (thievery,drunkenness and witchcraft) and perhaps sec-ondarily to membership in the wrong politicalparty at the time of independence.Another form of selection was ‘‘self-selec-

tion’’ in the form of unofficial occupation ofabandoned farmland, or ‘‘squatting’’ as it wascalled. Because the resettlement program wasso slow to become operational in certain partsof the country after independence, land occupa-

tions were common in some areas, perhapsmost notably in Manicaland. 26 In one Manica-land scheme studied by the author, those whospontaneously occupied land in the early1980s managed their areas autonomously forseveral years until the resettlement bureaucracyfinally took over and recognized them as legiti-mate settlers. Unfortunately little is known ofthe backgrounds and personal characteristicsof these early land occupiers so a comparisonwith contemporary occupiers is difficult. 27

Some notion of the heterogeneity amongthose selected for resettlement can be had fromthe data in Table 1, which categorizes thehouseholds initially allocated land in ChinyikaResettlement Area, one of the largest schemesin the country. Strictly speaking, the categoriesemployed are not mutually exclusive, but theydo indicate that two-thirds of beneficiariescame from nearby communal areas or werelandless people living locally.The question of the extent to which the

households allocated land in the original reset-tlement areas actually met the primary selectioncriteria is explored further here. In principle, ifthe criteria had been applied uniformly acrossthe country, those selected to receive land inthe early 1980s would have constituted a homo-geneous group of disadvantaged people. Towhat extent was this the case?This question can be answered through exam-

ination of the personal and household attributesof those who received land at the time. Usingdata from the ZRHDS panel, covering some400 resettled households over 20 years, it is pos-sible to test whether the households resettled inthe early 1980s met the defined criteria; that is, itis possible to explore the homogeneity of thehouseholds based on the defined attributes.The first step in the analysis is to examine the

criteria that were supposed to be employed in


selecting those to be resettled. For this purpose,a set of variables has been selected for its abilityto capture aspects of households likely to bestrongly related to poverty outcomes (see Table2). The reasoning guiding the selection is thatthe variables should represent, first, characteris-tics of the household, or head of household,that correspond fairly closely with the criteriaspelled out in the official policy. Second, thesecharacteristics were in principle verifiable bythe administrators running the program andthus could have been used to screen applicantsfor land.From the full data set, the variables selected

therefore represent the landlessness of thehousehold prior to resettlement; the age andmarital status of the household head; and theextent to which the household had been ad-versely affected by the war. Descriptive statis-tics for these criteria variables are given forall the households in the ZRHDS panel inTable 2. Descriptive statistics are shown fortwo categories: the five variables representingthe main screening criteria—identified as pri-mary variables—and a set of seven secondaryvariables.One comment is needed on the secondary

variable representing the number of bovines

Table 2. Descriptive statistics for selection criteria anin 1980 an

Variable

Primary

Number of acres of arable land prior to resettlementb

Age in years of the household head

Marital status of the household headc

Household was female-headed at the time of resettlemen

Condition in late-1979e

Secondary

Years of farming experience before resettlement

Years of formal education of the household head

Household head had a savings account before resettleme

Total number of persons resident in the household

Proportion of household members economically activef

Number of bovines owned at resettlement

Source: Zimbabwe Rural Household Dynamics Study (ZRa The total number of households in the analysis is 394. Fb The mean excludes the one household that had 165 acresc Zero if single; 1 if married monogamously; 2 if married pd Zero if male-headed; 1 if female-headed.e Proxy for the impact of the war and the household�s state bfollowing: wounded/assaulted/tortured; personal effects/monhouse/crops destroyed; livestock killed/stolen; or destitute.f Proportion of all persons resident in the household regar

owned. All model A resettlement was predi-cated on farming systems using oxen for draftpower. Although smallholders short of draftpower often yoke together a cow or a large hei-fer and an ox, and many have no draft power atall, the planners assumed that all those resettledunder model A would have at least two draftoxen at the outset. 28 The mean for the variableindicates that on average households ownedabout four head of cattle in the early 1980s.This mean is, however, very unstable, and infact over 40% owned no cattle at all (Kinsey,Burger, & Gunning, 1998).The pattern in resettlement areas in the early

1980s did not contrast sharply with that forother rural areas of Zimbabwe. In various stud-ies nationwide (Corbett, 1994; Coudere & Mar-ijsse, 1988; Moyo, 1995; Scoones, 1996; Zindi &Stack, 1992), it has been found that from 20%to 60% of households own no cattle. Such com-parisons suggest that the resettled householdson which the analysis here is based started froma point no better than most other rural house-holds. 29

It has been estimated by the government thatover 80% of those selected fell into the catego-ries of the most-needy: refugees and the war-af-fected, the landless, and those with insufficient

d related variables for ZRHDS households resettledd 1981a

Mean Minimum–maximum

4.13 (8.95) 0.0–165.0

43.71 (13.20) 18–83

1.14 (0.54) 0–2

td 0.05 (0.22) 0–1

0.93 (0.90) 0–4

14.08 (13.94) 0–60

4.43 (2.84) 0–11

nt 0.07 (0.26) 0–1

8.06 (3.82) 1–24

42.22 (19.47) 0–100

4.09 (4.89) 0–36

HDS).igures in parentheses are standard deviations.in Zambia before returning to be resettled.olygynously.

efore resettlement. Counts the number of states from theey stolen/destroyed; family members killed; unemployed;

ded as economically active.


land to maintain themselves (Zimbabwe, 1995).The proportions of households in the ZRHDSpanel, which is drawn entirely from land reformbeneficiaries, that should not have qualified forresettlement according to the different criteriaare presented in Table 3. The analysis shownin Table 3 suggests therefore that the figure of80% is an overestimate.The results in Table 3 are interesting in sev-

eral respects. First, although over 43% ofhouseholds fail to meet at least one of the de-fined criteria, no household would be disquali-fied on all four criteria considered. Moreover,less than a quarter are disqualified on two ormore criteria. These levels—although seeminglyhigh—probably do not indicate egregiousabuse by beneficiaries of the original land real-location program—implemented nationallywith great speed by personnel who learned asthey went along. Instead, it is suggested thatthe figures are consistent with: (i) flexible appli-cation of the criteria in the setting of a large-scale public welfare program, or (ii) theunworkability of the criteria and failure to ap-ply them at all. It can certainly be considereda failure in implementing public policy thatonly 27.4% of the beneficiaries fell into the in-tended target group.

Table 3. Proportion of resettled house

Criterion

Land Strictly landless (no land

Nearly landless 1 (land h

Nearly landless 2 (land–m

Neither strictly landless n

(holding larger than 4 acr

Neither strictly landless n

(land–man ratio above th

Age Head of household aged

War-affected Head of household repor

from the war for him/her

of his/her family or his/he

Marital status Unmarried male head of

Single or multiple Households disqualified o

Households disqualified o

Source: Derived from Kinsey (2002b).a Excludes eight single-person households, one ‘‘refugee’’ hhouseholds for which data on prior land holdings are notb No household head was aged below 18 at the time of renearly 12% were aged above 60.c Widows and divorcees with dependants were entitled to bdivorcees were very seldom resettled.d The given value is an understatement. The proportion heron all criteria is not available for all households.e The given value is an understatement. No household wasbasis of the data available.

Fewer than 20% of those resettled were trulylandless when they were allocated land. 30 Anydefinition of near-landlessness is fraught withdifficulties, however, as it hinges not only onsize of holding, soil quality and topographybut also on family size relative to the areacultivated and farming system employed. Twoapproaches are taken here. One is to use asomewhat arbitrary area of four acres as thedividing line between landed and near-landless-ness. On this basis, just over 48% of those reset-tled were nearly landless. On the same basis,nearly one-third of households would not bequalified for resettlement, as they did not meetthe criterion.A second approach is to calculate the land–

man ratio for all households and to assume—again arbitrarily—that households below themedian value are ‘‘nearly landless.’’ 31 This ap-proach has the merit at least of correspondingfairly closely to the concept of what it meansto be land-poor in Zimbabwe, i.e., lacking ade-quate land to feed one�s family. This compositeindicator suggests that just over half the fami-lies resettled could have been considered land-less or nearly landless, compared to morethan two-thirds with four acres as the thresh-old.

holds not qualified for resettlement

Level Percent

prior to resettlement) 19.6

olding of 4 acres or less) 48.1

an ratio below the median value)a 51.0

or nearly landless 1

es)

32.3

or nearly landless 2

e median value)a49.6

below 18 or above 55b 18.6

ted no adverse effects

, members

r property

38.3

householdc 3.8

n at least a single criteriond 43.2

n 2 or 3 criteriae 23.7

ousehold that had 165 acres prior to resettlement and 18available.settlement. In contrast, 18.6% were aged above 55 and

e resettled, but in practice widowers, bachelors and male

e is less than the highest row above because information

disqualified on all four of the criteria considered on the


Of those households in possession of land be-fore they were resettled, 48% reported that theywere continuing to cultivate land in their homeareas in 1983–84—despite the requirement thatthey relinquish all rights to such land. This dec-laration should not be taken too literally, how-ever, because it is likely to mean in practice thatthese households had left family members be-hind on their old land. At the same time,though, this figure is both indicative of the highpressure on land in Zimbabwe and of the ulti-mate limitations that resettlement faces in alle-viating population pressures in the country�scrowded communal areas.Nearly a fifth of households selected to re-

ceive land had household heads above thecut-off age of 55. It is not surprising in the so-cial context of rural Africa that those agedabove 55 years should have received land.Any young and relatively inexperienced admin-istrator would have found it very difficult todisqualify one of his male elders on the groundsof age.Over 60% of households had reported ad-

verse impacts from the war that led to Zimba-bwe�s independence. Most of these werefamilies that lived in rural areas or had fled rur-al areas leaving behind property and livestock,which were subsequently destroyed or stolen.More devastating, many had experienceddeaths in their families or had been subjectedto torture. Conversely, just under 50% reportedno negative consequences from the war andthus would not have qualified for resettlementunder this criterion.Less than 4% of households would be dis-

qualified for having an unmarried male head,and it is possible that the occurrence of suchhouseholds is due to death, desertion or divorcesoon after resettlement took place.

Table 4. Characteristics of households

Characteristic in 1980–81

Applied for resettlem

and was selected (me

Arable land available (acres) 3.7a

Number of trained oxen owned 1.5

Age of head of household 41.3

Years of education of the

head of household

5.2

Source: ZRHDS surveys, various years.a The 3.7 mean acreage excludes one settler who had prZimbabwe. If his acreage were included, the mean figure w

The data considered above report only onthose ZRHDS households that applied forand received land in the early 1980s. Thereare two other groups in the panel that also needmention here: those who applied for land butdid not receive a plot and those who had nowish to be resettled and thus did not apply. 32

There are significant differences among thesegroups that shed further light on the character-istics of those who went through the resettle-ment procedures in the initial program.Table 4 emphasizes four key differences

among the three groups of households. Asmight be expected, those who did not applyto be resettled were the oldest, had the largestaverage holdings and the greatest number ofdraft oxen and, interestingly, the lowest levelof education. Those who applied but were re-jected, in contrast, were the best-educated andyoungest group and held intermediate-sizedholdings; but on average they lacked sufficientoxen even to manage the land they had. Thoseaccepted into the program had the smallestland holdings but reasonable numbers of draftoxen; they were also intermediate in age andeducational attainment.The discussion in this section is more lucid if

one understands something of what motivatedpeople to make the decisions they did regardingresettlement around the time of Zimbabwe�sindependence. More than two-thirds of thosewho were given land applied for precisely thisreason—to gain access to more land and/orland of better quality (Table 5). The secondmost common reason for applying—personal/social freedom—largely reflects the wishes ofmarried adult sons who wanted to move awayfrom their fathers� farmsteads and establishtheir own homes. This response is shaped bothby the general acute shortage of land in the

resettling and not resettling, 1980–81

Category of household

ent

an)

Applied for resettlement

but was rejected (mean)

Did not apply for

resettlement (mean)

4.3 5.9

0.6 1.8

36.5 44.9

5.6 3.9

eviously had 165 acres in Zambia before returning toould be 4.1 acres.

Table 5. Reasons for applying or not applying for resettlement when the opportunity was first offered

Why beneficiaries applied to be resettled

(proportion of 466 responses)

Why communal area families did not apply

(proportion of 103 responses)

To obtain more/better land than we had 66.9 Disqualified by age, marital or employment status 34.0

To achieve personal/social freedom 13.9 Had land and assets/satisfied here 28.2

To improve economic welfare 12.4 Inadequate farm equipment/draft oxen 9.8

To escape problems with wild animals 2.6 Not interested 8.7

To savor the fruits of independence 1.7 Inadequate labor/health problems 6.8

Other 2.8 Social reasons/roots/networks 4.9

Lacked information/didn�t know how 2.9

Other 4.9

Total 100.3 Total 100.2

Source: ZRHDS surveys: 1983–84 and 1999 survey rounds.


communal areas as well as by particular famil-ial circumstances.It is clear from the reasons given for not

applying that many households selected them-selves out of consideration because of whatthey knew about the qualifications for accept-ance (even though a small proportion of re-sponses indicated a lack of information aboutresettlement procedures). More than a thirdof all the responses received are linked to anawareness that the household would not qualifyin terms of a specific criterion, such as age, ormarital or employment status. Over 16% of re-sponses related to the fact that householdslacked resources—labor, equipment or oxen—to cope with a larger land area. Overall, how-ever, nearly 42% of responses reveal that thecommunal area families were content wherethey were and had no wish to leave their land,fixed assets and social networks.The examination of resettlement procedures

above views the selection criteria in the way aland reform administrator might view them.The bureaucratic mind may have attempteduniform application, but inevitably subjectivity,bias, past experience and pressures of time andwork colored judgment and decisions. Never-theless, it is striking how the existing evidencepoints to a selection process for land reformbeneficiaries that appears to have been remark-ably equitable, and also efficient in targetingthose who were government�s and donors� pri-orities at the time. How well the continuingimplementation of the resettlement programwas matched to those households selected inthe early 1980s will be addressed in the follow-ing section.It is not possible to be so sanguine about the

beneficiary issue following the chain of events

set in motion in early 2000. For one thing, de-spite its total authoritarian control, ZANU-PF still presides over a country in which thereare rapidly deepening chasms in the nationalsocial fabric. While there has undoubtedly beenan irreversible shift in the de facto distributionof land since the land invasions began in2000, and while many landless households haveaccessed land, there has also been what someobservers refer to as ‘‘a parallel, competingprocess of �accumulation from above� throughwhich a narrow class of politically-entrenchedaccumulators are positioning themselves forthe future’’ (Hammar & Raftopoulos, 2003, p.23).The ‘‘land issue’’ is now only one dimension

of a complex matrix of dilemmas and chal-lenges that need to be addressed simultaneouslyin Zimbabwe. Any new political accommoda-tion will of necessity have to address the bene-ficiary aspect of land redistribution. To fail todo so will sow the seeds of conflict for yearsto come. It is clear, for example, that thetake-up rate the government boasted aboutwas wishful thinking. Moreover, simple obser-vation in rural areas reveals that some farmshave been abandoned and others left unutilizedwith but a single caretaker. Thus, there island available that could be used to achieve amore equitable and transparent distribution,although matching further redistribution withadequate programs to support new farmers willnot be easy. A related issue is what can alreadybe perceived as almost certain conflict betweenthe early farm occupiers sent onto the land asshock troops and the politically powerful accu-mulators who are ejecting them now that theyhave served their political purpose. In addition,there remain unresolved sets of questions both


about, on the one hand, the entitlements ofhundreds of thousands of dispossessed andimpoverished farm workers and, on the otherhand, those who have taken over large-scalefarms in flagrant disregard for existing law.

4. WHAT HAPPENED TO THOSE WHORECEIVED LAND?

Whether Zimbabwe�s land reform should beregarded as a ‘‘success’’ is a matter both ofopinion and for further debate. Perspective alsomakes an important difference in judgment. Ifone compares what was achieved before thewave of land invasions starting in 2000 withthe original targets of the program—resettling162,000 households on about nine million hec-tares of land in the 1980s, then land reformshould be considered an outright failure. Butif one considers merely the number of house-holds resettled (as Mhishi, 1995, does forinstance), then performance has been satisfac-tory. By this criterion, almost one out of every10 communal area households benefited fromthe program.Other perspectives also need to be weighed.

Whereas the political returns to land redistribu-tion in Zimbabwe have tended to dominatemuch of the contemporary debate, the donorswho supported earlier phases of the pro-gram—as well as potential donors—were con-cerned with the economic returns to landreform. Two available analyses (Cusworth &Walker, 1988; Robilliard, Sukume, Yanoma,& Lofgren, 2001) examined the costs and bene-fits of the early land reform program in Zimba-bwe and found an internal rate of return to theprogram above 20%, exceptionally high for anagricultural program. 33

Deininger, Hoogeveen, and Kinsey (2004)update this earlier work using more recentand more detailed data that permit useful in-sights into what has happened to the originalresettled households since the 1980s. Using adifferent methodological approach, they con-firm earlier work (Gunning, Hoddinott, Kin-sey, & Owens, 2000) that showed householdincome from cropping is much higher amongresettled households. They also, however, pro-duce a more modest, but more nuanced, setof rate of return estimates based on the tempo-ral dynamics of resettled households.Together these analyses are indicative of the

kinds of scenarios that might have unfoldedin a different world—one in which the govern-

ment of Zimbabwe had retained an interest insmallholder resettlement, in which donors hadnot sidestepped the relationship between prop-erty and poverty, and in which the negative ef-fects of adjustment did not divert the attentionof Zimbabweans from the land issue for severalcritical years. These, however, are largely polit-ical and macro-level considerations. It is thetask of the remainder of this section to add towhat we can learn about the impact of the orig-inal resettlement program on the householdsthat acquired land.The approach followed is to query a wide

range of available literature for evidenceregarding what has happened to a set of out-come variables that it can be argued are criticalin evaluating the performance of any landredistribution program. These variables in-clude: production and yield levels, income,accumulation of assets and savings, consump-tion levels, poverty alleviation, and responsivecapacity.

(a) Production and yields

Enhanced levels of crop production becamean implicit if not explicit primary raison d’etreof the resettlement program very quickly onceit was underway. To what extent have theybeen achieved? The evidence is at best fragmen-tary—and the more macro-level data are not tobe trusted at all, but the available sources do al-low a picture to be composed.Recent data from a study by Chigwenya

(2001) are illustrative of the findings from re-search that compares outcomes in resettlementareas with those from the communal areas.Table 6, for example, compares productionand productivity levels for maize across twodifferent tenurial regimes—a model Ascheme—Gutu South—and an adjacent com-munal area—Ndawi, from which many of thesettlers came. Particularly striking are the lowfigures for total household production andyields for the communal area. While exact com-parison is impossible, there is a clear suggestionthat both production levels and productivityrise substantially following resettlement.Caution is needed in interpreting such results

however. The ratio between the observedchanges in the lowest production level post-and pre-resettlement is 5 to 1 while that for pro-ductivity is 2 to 1. Since such productivity gainscannot possibly generate the production gains,it is clear that much of the apparent improve-ment comes simply from cultivating more land

Table 6. Maize production and productivity: before and after resettling in a resettlement area (RA)and neighboring communal area (CA)

Tonnage category Total production (tonnes)

(percentage of growers)

Yield (tonnes/ha)

(percentage of growers)

Prior to resettlement After resettlement Ndawi CA Gutu South RA

Below 0.50 80.4 15.9 69.6 34.1

0.51–1.00 6.1 6.1 17.4 17.1

1.10–1.50 11.0 17.1 13.0 13.4

1.51–2.00 0.0 17.1

Above 2.00 2.0 61.0 0.0 18.3

Total 99.5 100.1 100.0 100.0

Source: Derived from Chigwenya (2001, pp. 31–32).


with the old standards of management. 34 Nev-ertheless, it is also clear that resettlement hasenabled substantial numbers of farmers toachieve greater productivity. Yields above1.5 tonnes/ha were not reported at all in thecommunal area, yet they are common in theresettlement area.Other data also suggest that, on average,

resettled farmers have exceeded planning expec-tations. Despite not following recommendedcropping programs, 73% of farmers in the Msa-sa-Ringa model A scheme realized a total grossvalue of production above target (Chikondo,1996). On average, Chikondo found that a ran-dom sample of farmers in Msasa-Ringa culti-vate 3.7ha—22% above the planning figures(Table 7). Yet, while some 54% of householdscultivate more than the average, a quarter ofhouseholds still cultivate less than the plannedthree hectares. Only 27% of households inChikondo�s study are classified as poor, but thisgroup constitutes most of the farmers cultivat-ing less than three hectares.Yields are also generally above expectation,

however 58% still realize maize yields below3tonnes/ha. This performance level is probably

Table 7. Mean performance indicators comp

Crop Area (ha) Yielda

Planned Actual Planned A

Maize 1.00 2.65 3.00

Groundnuts 0.50 0.62 0.85

Sunflowers 0.90 0.28 1.50

Other 0.60 0.11 –

Total 3.00 3.65 –

Source: Adapted from Chikondo�s study of Msasa-Ringa Ra Yield is in tonnes/ha and retention and sales are tonnes.

attributable mainly to capital constraints; only8% of farmers were granted credit, whereasthe planners expected that all farmers wouldbenefit from the resettlement credit scheme.On average, farmers used less than recom-mended input levels—but half used the recom-mended levels or even more. Most whomanaged to realize higher yields had large live-stock numbers—above the permissible limits.Shortages of extension staff and transport forinputs are said to have militated against morewidespread use of purchased inputs. Three-quarters of the total value of production camefrom crop production and the balance fromlivestock. On average, retentions were abovethe levels assumed. Households retained 23%of maize output—a proportion similar to theplanned level—but in real terms they retained1.9 tonnes—almost three times the planned700kg per household (Chikondo, 1996).Table 7 below suggests powerful discrepan-

cies between planners� preconceptions andfarmers� objectives. While areas planted, as wellas yields, are reasonably close to expectations,retention levels and sales differ significantly—probably due to the fact that household sizes

ared to planning targets, 1993–94 season

Retention Sales

ctual Planned Actual Planned Actual

3.62 0.70 1.85 2.30 7.46

1.42 0.25 0.39 0.15 0.44

0.70 0.00 0.02 1.85 0.18

– – – – –

– – – – –

esettlement Area (1996).


are far larger than assumed by planners. 35

Failure to follow proposed cropping programsreflects adaptive behavior as farmers allocateresources among different uses in order tosatisfy individual and household needs mostefficiently. The discrepancies between theplanned and actual outcomes highlight plan-ners� misconceptions about household behav-ior. Differences in endowments, motivationsand capabilities were not considered in themechanistic planning models.More systematic work (Harts-Broekhuis &

Huisman, 2001) has compared productivity dif-ferences between large samples of resettled andcommunal area farmers in a resource-poor partof the country—Insiza District in Natural Re-gion IV. On average, they find that resettledhouseholds have higher levels of productionthan other households, but the relative differ-ence is attributable entirely to the fact thatthese households have access, first, to more ara-ble land and also to more labor, farm equip-ment and livestock. They conclude thatresettlement has not resulted in significantimprovements in agricultural productivity perhectare.Chiremba (2002) studied the productivity of

resettled farmers over four years in the earlyand mid-1990s. 36 She finds that, on average,farmers do not show consistent growth inproductivity over the time period studied. Alimitation on this conclusion, however, is thatperformance was compared across years ofnormal rainfall and below normal rainfall.Although there was some evident decrease inthe proportions of individual farmers operatingat the lowest levels of efficiency, the proportionof farmers operating at or near the efficiencyfrontier was low. Hence, one conclusion to bedrawn from the analysis is that there was enor-mous scope for many farmers to catch-up tothe best farmers in the sample. Nevertheless, amajor finding was that resettled farmers werefound to be more productive, on average, thancommunal farmers.Efficiency gaps among resettled farmers were

explained mainly by possession of farmingequipment and frequency of extension visits.Farmers with the basic ox-drawn plough werecloser to the frontier, hence more productive,than those without. Chriremba postulates thatequipment could be a proxy for farmer ability.The best farmers performed well despite hav-ing fewer visits from extension workers. Itappeared moreover that extension workers con-

centrated their visits among the least produc-tive farmers.The efficiency measure used in this study was

limiting as it incorporated only land, labor andfertilizer in the input vector. Several importantunmeasured input variables—such as manure,seed and chemical use—were not included.The only farmer-specific characteristics that ex-plained variations in household performancewere the number of years of farming experienceand possession of farming equipment (Chir-emba, 2002).

(b) Family and per capita income levels

Chikondo (1996) studied the outcomes ofresettlement in terms of the total gross valueof production (TGVP) of crops and livestockachieved by resettled farmers compared to theurban minimum wage—Z$4,800—at the timeof her study. She found that 31% of resettledhouseholds overall realized less than theminimum wage. 37 Because aspirations andobjectives influenced agricultural output sostrongly, Chikondo could find no direct rela-tionship between endowments and householdproduction. Both households that were highlyendowed and less well endowed but hard work-ing realized higher values of sales than others.Chikondo also generally corroborates the nega-tive relationship between farm income and ageobserved in other studies of resettled farmers.Chikondo differentiates farmers into four in-

come strata according to their TGVP (Table8). 38 Twenty-seven percent of the householdssampled fall into the lowest group, and Chik-ondo concludes that the benefits of resettlementare marginal for this group. Even with a smallsample, however, intergroup heterogeneity isso great that it is difficult for Chikondo to domore than identify broad patterns. Across allgroups, except for those aged below 45, she ob-serves that those with many cattle are also the‘‘richest.’’Since female-headed households are poorer

than those headed by males, there is evidenceof the feminization of poverty. Thirty-threepercent of female- and 25% of male-headedhouseholds are ‘‘poor.’’ Overall, 91% of fe-male-headed and 69% of male-headed house-holds fall in the lower half of the incomedistribution. While female-headed householdsare poorer in cash terms, Chikondo observesthat they are more food-secure than male-headed households.

Table 8. Distribution of total gross value of production by category of household

Stratum (TGVP (Z$)) Variable used for categorization of households

Sex of head Age of head Cattle ownership Years of

education

Total %

(N = 96)

Female

(n = 24)

Male

(n = 72)

<45

(n = 22)

>45

(n = 74)

<11

(n = 32)

>11

(n = 64)

<3

(n = 46)

>3

(n = 50)

Rich (>18,917) 8.3 13.9 18.2 10.8 6.3 15.6 13.0 12.0 12.5

Upper-middle

(12,612–18,917)

0.0 16.7 0.0 16.2 0.0 18.8 17.4 8.0 12.5

Lower middle

(6,306–12,611)

58.3 44.4 63.6 43.2 43.8 50.0 43.5 52.0 47.9

Poor (<6,306) 33.3 25.0 18.2 29.7 50.0 15.6 26.1 28.0 27.1

Total % 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Source: Derived from Chikondo�s study of Msasa-Ringa Resettlement Area (1996).


Using a wide range of secondary sources,Tawonezvi (1995) reviewed the performanceof the livestock subsector in resettlement areas.His main conclusion is that the schemes exam-ined were failing by very wide margins toachieve planned target incomes from cattlesales. Tawonezvi attributes this poor outcometo the unreality of initial assumptions regardingownership, herd growth and composition,slaughter grades and offtake rates. Cattle salesmay, however, be too restrictive a concept bywhich to assess livestock income. In someareas, at least, income derived from livestockcontributes much more to the cash flow ofhouseholds than does the sale of agriculturalproduce (Harts-Broekhuis & Huisman, 2001).The most comprehensive studies of income

changes in resettlement areas, however, cometo conclusions that differ substantially fromthose cited above. Controlling for differencesin initial conditions and differential access tocredit and agricultural services, Deininger,Hoogeveen, and Kinsey (2000) find that com-munal farmers produce 40–50% less than reset-tled farmers, confirming other findings (Kinsey,1999; Harts-Broekhuis & Huisman, 2001). AndGunning et al. (2000) show that, since 1983,resettled households in three differing environ-mental zones have greatly increased their cropincomes, productivity and household income.It is ironic therefore that resettled households

do well in terms of agricultural productivity butthat these differences do not translate well toimproved poverty outcomes (see below). Asnoted, controversy over the impact of the reset-tlement program stems in part from the per-spectives taken by the analysts. Foragricultural productivity indicators, careful

studies from the perspective of the householdlevel suggest substantial gains in some casesbut not in others. For welfare indicators, theperspective at the level of the average individualsuggests no significant improvement—a findingthat appears to be universal.In large measure, this result arises simply be-

cause resettled households are substantially lar-ger (they comprise 10.0 resident members onaverage in contrast to 6.4 members for commu-nal households), so that household differencesdisappear if they are expressed in per capitaterms. Deininger et al. (2000) explore this issuefurther, as do Deininger, Hoogeveen and Kin-sey in this section. They find that resettledhouseholds attract new members as a result oftheir success in agriculture 39 and show thatthe additional members do not contribute tothe household�s agricultural output. The impli-cation then is that relatively well-off householdsprovide a social safety net function. Other stud-ies (see Harts-Broekhuis & Huisman, 2001) con-firm that resettled households attract additionalfamily members and are larger, but the relativedifference in size between resettlement and com-munal areas is less.Summarizing the income patterns of the

resettled households in the low-potential areathey studied, Harts-Broekhuis and Huisman(2001) note that, while both communal andresettled households have similar portfolios ofincome sources, the outcomes vary substan-tially:

—Settler families have much higher levels oftotal income than communal area families,even on a per capita basis, and the differenceis more pronounced in good seasons than inbad;


—The contribution of agriculture to theaverage income is higher in resettled areasthan elsewhere, although cash income fromcropping is minimal in resettlement areas—income from livestock and nonfarm activi-ties, including wage labor, contributes muchmore to households� cash flow;—Valuing maize retained for consumptionat its opportunity price increases the contri-bution of agriculture to total income tomore than half;—Eighty percent of resettled households areinvolved in self-employed nonfarm activi-ties—a level lower than in the communalareas; and—Resettlement has reduced the incidence oflabor migration among household heads, sothat income from remittances is less impor-tant among resettled households.This last outcome is not surprising, however,

as resettled household heads were prohibitedfrom migrating for employment for more thana decade; and it is difficult to interpret sincethe opportunities for off-farm employmentbegan to decline rapidly at the time the prohibi-tion was lifted.

(c) Asset accumulation and savings

Asset accumulation in model A schemes hasoccurred in two main forms: the acquisitionof cattle (and other livestock) and throughinvestment in farm equipment and other dura-ble assets.While cattle are essential to the farming mod-

els on which resettlement agriculture is based,the provision of grazing land within schemescan cope with only limited numbers of live-stock. In his review of livestock, Tawonezvi(1995) shows that, although resettlement areasas a whole were still generally understocked adecade after resettlement began, some 8–20%of settlers had already exceeded the recom-mended livestock targets. By 1990, some modelA schemes—Copper Queen and Chikomba 2,for example—were already overstocked as awhole by between 11% and 19%. 40

In their examination of changes in cattlenumbers over time in three model A schemes,Kinsey et al. (1998) found that the number ofcattle owned by the average household in-creased from less than three in 1983 to 10 in1995. The proportion of households withoutcattle also decreased from 42% in the early1980s to less than 10% in the mid-1990s. Themean real value of family herds increased on

average some 16% a year, making cattle aninvestment that would have outperformedevery financial instrument available in the Zim-babwean economy.Few if any attempts have ever been made to

enforce regulations governing the size of indi-vidual herds in resettlement areas. The reluc-tance to ensure that farmers� herds arerestricted to recommended stocking levels hasbeen partly because schemes have generallybeen understocked, and there has been littleenthusiasm to implement what would beunpopular regulations among those growingincreasingly wealthy. Moreover, innovativeproposals to make grazing rights a marketableasset (as in the Chinyika scheme) so as to ben-efit those with few or no cattle were neverimplemented. In the absence of mechanismsto assist settlers without cattle and to ensurethat cattle owners do not exceed recommendedherd sizes and stocking rates, Tawonezvi con-cludes that the resettlement areas can be ex-pected to deteriorate to the situation incommunal areas.Very little work appears to have been done

on savings patterns by resettled households.The study by Kinsey et al. (1998) finds thatthe use of savings accounts grew rapidly amongthe households they analyzed. Whereas the pro-portion of households with at least one savingsaccount was only 6% in 1980, the figure was52% in 1995. The growth in the number of ac-counts was steady over this period with thenotable exception of the drought years. Majordroughts were reflected in dramatic falls in boththe number of savings accounts and mean bal-ances, however a single good season could seethe number of accounts restored. Cash holdingsand savings accounts are however unattractiveassets in Zimbabwe. In the mid-1990s, beforethe current hyperinflation, inflation fluctuatedbetween 22% and 42% and interest rates onmonetary savings accounts were negative inreal terms.Recent research (Kinsey, 2003) gathered

information on the accumulation of severalcategories of moveable assets (farm equipmentand implements, household durables, and oth-ers—including vehicles, bicycles, grinding mills,solar installations and so on). The analysis cal-culated the mean and median change in thenumber of items over the period since resettle-ment. 41 For all farm equipment, the medianchange in farming assets was 8% (mean240%); for household durables, the medianchange was 41% (mean 321%); while for all


other types of assets enumerated, the medianchange was just 1% (mean 90%). This patternof relatively small medians compared to themeans is indicative of quite modest asset accu-mulation overall but dramatic gains on the partof some households. Given evidence that animportant source of productivity gains in reset-tled agriculture is increasing returns to farm as-sets (see Owens, Hoddinott, & Kinsey, 2003),the slow rate of accumulation of such assetsin resettlement areas should be cause for con-cern. 42

A final observation needs to be added here inrelation to investments in housing and otherfarm structures. Although no analysis to datehas focused on this aspect of asset accumula-tion, 43 it is obvious from both the author�sfield experience and the data collected thatresettled households have made very substan-tial investments in their stock of on-farm build-ings. Some of the improvements to housingwere assisted by a special credit facility thatwas utilized in the 1980s, but many householdshave gone far beyond what this scheme pro-vided. This type of investment has been encour-aged by farmers� knowledge that, whateverhappens to their entitlement to the land itself,they can sell the assets they place upon theland.

(d) Consumption levels

Work on estimating the benefits from reset-tlement in terms of changes in patterns and lev-els of consumption is limited and still at a fairlyearly stage. In analysis aimed at quantifyingthe gains in per capita expenditures betweenhouseholds that applied but were rejected forresettlement and households that benefited byreceiving land, Deininger et al. (2000, 2004)suggest that being a land reform beneficiary in-creases per capita expenditure by a very modestZ$165 per annum. At 1995 values and prices(with an exchange rate of Z$9.50 to the USdollar, the per capita benefit to land reformbeneficiaries is only approximately US$17 perannum.In terms of expenditure levels on a per-house-

hold basis, resettled farmers do substantiallybetter than those who were not selected forresettlement. Much of this advantage disap-pears, however, when expenditures are ex-pressed in per capita terms, an outcome thatshould be attributed to the fact that settlerhouseholds are now substantially larger thannonsettler households. The disparity in house-

hold size is such that, in per capita terms, thedifference in expenditure levels becomes negligi-ble.Data collected in the first half of 2002 using

the same methodology but from a differentsample of resettlement schemes generated amedian per capita expenditure of Z$1,334(mean Z$2,037) (Kinsey, 2003). This medianfigure is equal to US$24 at the highly overval-ued exchange rate in 2002. 44 The same studyfinds that the value of food consumed peradult-equivalent tends to be greater amongthe smaller households. For households witha mean size below eight persons, for example,median food consumption is valued at $3,484,whereas it is only $2,353 among larger house-holds. The opposite is true for nonfood con-sumption, where households with a mean sizebelow eight persons have a median consump-tion per adult-equivalent valued at $235,whereas the nonfood consumption of largerhouseholds is valued at $579.The evidence suggests, therefore, that while

resettled households do relatively well in farm-ing, this success does not translate through intopronounced improvements in per capita con-sumption. Resettled households have attractedadditional household members, who appear toserve chiefly to dilute benefits per person,although their presence does allow the benefitsof the program to be spread more widely. Nev-ertheless, consumption on a per capita basis isthereby reduced to levels that differ little fromthose in communal areas. These weak overallconsumption benefits are confirmed by otheranalyses showing that land reform beneficiariesdo not outperform nonbeneficiaries either interms of indicators of nutritional status or percapita expenditures (Kinsey, 1999; Hoogeveen& Kinsey, 2001).

(e) Poverty alleviation

At the national level, the representative In-come, Consumption and Expenditure Surveys(ICES) indicate that resettled households arealmost as poor as communal households (Zim-babwe, 1998; Alwang & Ersado, 1999). Alwangand Ersado find a degree-of-poverty incidenceof 65.3% among land reform beneficiaries and66.8% in communal areas. Second, the ICESinformation also suggests that during droughtyears, and despite their greater ownership ofassets, resettled households are just as vulner-able as communal ones. Caution needs to beexercised in interpreting the ICES results on


resettlement, however, since the surveys fail todistinguish between resettlement areas thathave just recently been established and thosethat have had time to reach projected economicmaturity.Other results also show that the success of

land reform in allowing households to obtainhigher productivity has not translated intosuccess in terms of poverty reduction amongmodel A households (Hoogeveen, Kinsey, &Bouwmeester, 2003). Those living in model Ascheme areas are no better off than peopleresiding in nearby communal areas. The sameanalysis also finds that rural poverty is sensitiveto the occurrence of drought and concludesthat the 1995–96 national poverty estimatesmay be overestimated by as much as 10 per-centage points. Simulation results indicate thatrural households benefited from the crop priceincreases associated with economic adjustmentbut that they were relatively isolated from theformal sector contraction that accompaniedadjustment. As shown elsewhere, resettledhouseholds were not, however, isolated fromthe dramatic changes in input prices that camealong with adjustment.Deininger et al. (2000) observed that resettled

households have grown in size as a result oftheir economic success. The movement of indi-viduals toward economically successful house-holds might be reversed if Zimbabwe�seconomy were to come out of its severe slump.At present, the flow of resources between ruraland urban areas is the reverse of the familiarpattern across much of southern Africa. In-stead of urban-based migrants sending remit-tances from employment to support familiesin rural areas, rural households currently sendresources to sustain urban migrants in the slimhope they will be able to secure wage employ-ment. The simulations reported by Hoogeveenet al. (2003) show that if economic recoverywere to happen, it would result in a very sub-stantial reduction in rural poverty. This reduc-tion would be much larger than the oneassociated with the increase in crop prices thatfollowed economic adjustment.Not only do differences in household size ex-

plain why economic success of the householddoes not translate into improved welfare indi-cators, but household size is also an importantfactor in explaining the difference between non-poor and poor households among resettled andcommunal households. The poor live in largerhouseholds. In resettlement areas, poor individ-uals live in households that comprise three

additional members compared to the numberof individuals in nonpoor households. In com-munal areas this difference is two.In a more recent study, Kinsey (2003) finds

extreme levels of distributional inequality interms of both farm revenue and nonfarm in-come across a sample of resettlement schemes.Gini coefficients of 0.80 or higher indicate pro-nounced concentrations of income and suggestthat the extent and depth of poverty in resettle-ment areas are much worse than in Zimbabweas a whole. The Gini coefficients for asset own-ership are also high, but those for consumptionare somewhat lower—indicating that a fewhouseholds at least have had some success insmoothing consumption across income strata.

(f) Responsive capacity

A number of analyses have been undertakento assess productive efficiency and the capacityof households to utilize the new resources madeavailable through resettlement. The first ofthese (Kinsey, 1987) was an analysis of the so-cial and economic correlates of productivity inmaize cultivation. This study found that, con-trary to expectations, it was the relativelyyounger farmers who achieved higher produc-tivity in the early 1980s. A more recent study(Chiremba, 2002) of the same households usingmore recent data reached similar conclusions,as did a study in a different area by Chikondo(1996). Chiremba found that the number ofyears of farming experience of the head of thehousehold had a negative impact on productiveefficiency more broadly measured. 45 In sepa-rate regressions for 1993 and 1996, long-termfarming experience actually impedes catch-upto the efficiency frontier, and hence retards pro-ductivity growth. Since it would be expectedthat greater farming experience would facilitateproductivity catch-up, it seems that little learn-ing of new technologies that optimize efficiencyor productivity occurs despite the number ofyears of farming experience. An alternativeinterpretation is that younger farmers are moreinnovative and therefore achieve higher efficien-cies than older farmers, who are perhaps resist-ant to adopting new technologies or who havedifferent utility preferences.The evidence noted above is contradicted,

however, by more recent cross-sectional data.An analysis of the factors underlying observedvariations in use of inputs and value of outputsamong farmers across several different resettle-ment models (Kinsey, 2003) estimated the re-


turns to management, where skill levels areproxied by years of farming experience. A sim-ple ordinary least squares regression model em-ployed years of farming experience as theindependent variable to explain differences inthe gross revenue from cropping. This formula-tion explained 25% of all variation in revenues,and the increase in revenue with each addi-tional year of farming experience was $13,794(at 2002 prices—with a standard error of$2,618). At the sample mean of 16.8 years offarming experience, this experience was worthsome $232,000 in additional revenue per sea-son.Other evidence points toward explanations

for these differences. Resettlement areas havealways had preferential access to agriculturalextension staff. It has been demonstrated thataccess to farm-level extension visits increasesproductivity significantly even after controllingfor innate productivity characteristics andfarmer ability (Owens et al., 2003). 46 It is alsoapparent that the entire resettlement experienceinvolves important learning-by-doing elements,as a number of papers in this section attest. In-deed, it would be very surprising if this were notthe case for what was after all a completelynovel experience for all participants. Suchopportunities are reflected both in the percep-tions of those who resettled and in empiricalwork that points to increasing returns to ‘‘reset-tlement experience’’ (Gunning et al., 2000).A further outcome variable that might also

usefully be considered is the impact of the reset-tlement program on the natural resource baseupon which agriculture depends. This dimen-sion, although both critical and controversial,is beyond the scope of this review. 47

5. CONCLUSION

The evidence examined here reveals a mixedpicture for Zimbabwe�s first generation of reset-tlement. On the plus side, generally positiveconclusions about the selection of beneficiariescan be drawn from what has been presented.Although extreme socioeconomic differentialsdo exist in the communal areas, there is no evi-dence whatsoever that the relatively wealthyand powerful were able to commandeer theprogram in any part of the country. To the con-trary, the evidence suggests that those who werealready well off had no interest in working up asweat tilling the five hectares to be given to eachfamily. Instead, through redistributing med-

ium-sized plots, the program did successfullycreate opportunities for those who were disad-vantaged.These early years were a Golden Age for the

resettlement program. Beneficiaries receivedexceptional levels of supporting services. Exten-sion coverage and access to credit were bothuniversal in the initial stages, and marketingfacilities, schools, clean water supplies andother infrastructure were provided. Althoughthe program was beset with periodic drought,beneficiaries became increasingly able to copewith drought and other adverse conditions. Inpart, the development of coping mechanismswas an individual family response, but at thesame time it was also an evolutionary adapta-tion within the new communities establishedunder the program. The research collaboratorson the ZRHDS have been made very aware justhow much beneficiaries value in noneconomicterms the opportunities made available tothem. The perceived benefits weight heavily fac-tors such as the chance to make a new start,the opportunity to take control of their ownlives, and the freedom to make their own deci-sions.The evidence is also consistent with an inter-

pretation that beneficiaries for the most partbecame better farmers. Both productivity andproduction levels have tended to rise, and reset-tled farmers continue to diversify their activitymixes. There are clearly problems, however, ingetting these agricultural gains translatedthrough into improvements in consumption,welfare and nutritional state. Analysis is stillbeing done to explore this phenomenon, butthe strongest explanatory hypotheses relate firstto the very considerable growth in mean house-hold size in resettlement areas and, second, tothe decline in support to the resettlement sec-tor. Benefits are being diluted by the largerhousehold sizes without concomitant gainsfrom the additional labor supply. The chief cul-prit in fueling these changes is thought to be thegeneral economic malaise that began withadjustment in the early 1990s. Whatever theexplanation, Zimbabwe appears to be becom-ing a more rural society as economic opportu-nities diminish in other areas.These factors point clearly to the conclusion

that the major gains from small-scale resettle-ment are exhausted within a single generation.Five hectares make an economically viablefarm for a nuclear, not an extended family.One adult son—or daughter—can succeed thepatriarch, but the other siblings will have to


move on. Where, in an economy with 70+%unemployment, do they go? The failure in polit-ical vision that allowed small-scale resettlementto disappear from the national policy agenda inthe late 1980s has been a costly one. It dailygrows more costly with the ill-conceived fast-track approach, the consequences of whichare seriously undermining the performance ofthe rest of the economy. As Dale (2001) andmany others have shown, poor agrarian poli-

cies have direct implications in terms of futureconflicts. The failure in political vision is, how-ever, a dual one. The donor community, andother friends of Zimbabwe, should have beenproactive in lobbying for active, continuingland reform a decade and a half ago. With en-ough creative energy and good will, it may stillnot be too late to rectify the obvious conflict-inducing aspects of the government�s currentapproach.

NOTES

1. See Kaldor (1999), Reno (1998), Berdal and Malone

(2000) and the papers in the 2-volume study edited by

Nafziger, Stewart, and Vayrynen (2000).

2. Readers wishing more background to the current

crisis are referred particularly to Addison and Laakso

(2003), Hammar and Raftopoulos (2003), Roth and

Gonese (2003), and Bernstein (2004).

3. In fact, the number of farms acquired for resettle-

ment peaked in 1982—at just over 600. Two years later,

the figure had dropped some 93%, and for the second

half of the 1980s it averaged only some 48 farms a year.

4. Nor, as I pointed out near the beginning of the

program (Kinsey, 1983), were the resources ever likely to

be provided given the competing demands from such

immensely popular sectors as education and health.

5. A powerful tendency, shrewd or naıve, within the

government of Zimbabwe is to overestimate its capacity

to implement complex programs. Complexities, instead

of being carefully evaluated as the basis for reformula-

tion of programs and practices, become instead merely

formulaic. Thus it was far easier to reduce the imple-

mentation of the resettlement program to a set of

mechanical procedures that operated merely by turning

a crank.

6. Because subsidies to white commercial farming were

high before independence (in part due to the motivation

to maintain a white buffer zone in rural areas during the

liberation war), this approach should have had both

political and practical appeal. Moreover, it would

usefully have distinguished between white farmers who

acquired their farms under the ‘‘old regime’’ and those

who committed themselves to Zimbabwe by purchasing

farms after independence.

7. In contrast, after an initially slow start, South Africa

moved decisively to allow a multiplicity of approaches to

redistributing land. Recent experience of the Land

Reform Credit Facility, one component facilitating

redistribution, is reviewed in Lyne, Zille, and Graham

(2000).

8. The issue of whether Zimbabwe�s overriding ‘‘na-

tional project’’ is one of modernization on the one

hand or accumulation by certain political and eco-

nomic strata on the other is delineated in Raftopoulos

(2003).

9. The first such survey appeared four years after

implementation began (Zimbabwe, 1984). The first

‘‘annual’’ survey of resettled ouseholds appeared in

1986 (Zimbabwe, 1986a) while the second only became

available six years later (Zimbabwe, 1992c). The first

broadbrush evaluation of the less numerous Model B

schemes appeared only 12 years after they were first

launched (Zimbabwe, 1992a).

10. The official government critique of the program is

best spelled out in Zimbabwe (1993), while Moyo (1995)

and Masilela and Weiner (1996) provide useful overall

commentaries on the various perspectives that have

labeled the program a failure.

11. Journal of Agrarian Studies (January–April 2004) 4

(1–2).

12. See Russet (1964), Prosterman (1976), and Proster-

man and Reidinger (1987) for examples.

13. Less satisfactory are works such as the World Bank

volume Post-Conflict Reconstruction (1998), which de-

votes a scant two paragraphs to the role of land, and

these refer only to the issue of land for demobilized

troops.

14. See particularly papers by Dagdeviren, van der

Hoven, and Weeks (2001), Bigsten and Levin (2001) and

Quah (2001).


15. See Ackerman and DuVall (2000) for a set of global

case studies of nonviolent conflict.

16. Esman (1989) concludes ‘‘The informal, unorgan-

ized, nonconfrontational weapons of the weak can be

effective as long as their services are required and cannot

be dispensed with’’ (p. 226). He cites the case of

Malaysia, where peasants were partially compensated

because the ruling political party continued to need their

votes.

17. In an address to the Royal Institute of Interna-

tional Affairs in June 2000, I pointed out that the

international media—particularly the British press—

were misrepresenting events in Zimbabwe by their

portrayal of the conflict in black-versus-white terms or,

even more narrowly, as Robert Mugabe versus the

British. To do so was to frame the debate entirely in

Mugabe�s terms. Journalistic coverage generally im-

proved following the 2000 elections.

18. Horowitz (1989) demonstrates mathematically that

the resource expenditure that accompanies conflict is

best thought of as a fixed cost incurred by the parties

whether they win, lose or draw. This result holds even in

situations where external parties are eager to defray part

of the fixed costs of conflict.

19. Uvin (1998) focuses on the role played, or not

played, by the development aid community prior to and

during the Rwanda crisis. He details how international

organizations, in addition to failing to catch warning

signs of impending genocide and respond appropriately,

also failed to weigh carefully the effects and potential to

harm inherent in their development programs. The

development community indirectly intensified inequali-

ties and the social exclusion of peasants. The develop-

ment community failed while believing its efforts were

successful, pointing to traditional development indica-

tors that made the country�s outlook seem so promising.

20. It is for this reason that international involvement

in supporting land reform occurs only in exceptional

circumstances.

21. In Manicaland, a woman governor actually en-

sured that women received land in significant numbers.

But, as Palmer (2003) accurately notes, everywhere the

process was driven by ZANU-PF, the party, as distinct

from ZANU-PF as the head of government.

22. Although they are undoubtedly correct in their

comments on the gender bias is the resettlement

program, Sender and Johnston�s sources for this state-

ment include not a single micro-level study of resettle-

ment.

23. See Herbst (1990) for an illuminating treatment of

the frequency of appearance of land-related articles in

the press prior to one election.

24. Four models—designated A, B, C and D—were

used in the program, and each was based on a different

approach. By far the most common variant was model

A, the family-based smalholding, which accounted for

over 95% of resettlement.

25. There were, however, press reports in the early

1980s that indicated the opposite also happened—that

newly resettled communities refused to allow ‘‘outsid-

ers’’ to become part of their village.

26. See Alexander (2003) for a detailed treatment of

land occupations in eastern Zimbabwe.

27. Marongwe (2003) provides very useful insights into

the farm occupations that began in the late 1990s and

the motives of those who undertook them.

28. This was just one of numerous contradictions in the

planning models. Others are discussed in Kinsey (1998,

2000, 2002a).

29. For further details, see Kinsey et al. (1998).

Vhutuza (1991) reports that more than 30% of settlers

in Mushandike Model A owned no cattle a decade after

resettlement.

30. Chigwenya (2001) reports the proportion of truly

landless in Gutu South resettlement area as above 35%

and the proportion with less than four acres as greater

than 62%.

31. Clearly, those below the median also include the

strictly landless.

32. In our paper in this section, Klaus Deininger, Hans

Hoogeveen and I use these categories as the basis for

refining the opportunity costs of resettlement.

33. According to Cusworth, shifting political currents

made this outcome unpalatable with the British govern-

ment because Zimbabwe�s white commercial farming

community had successfully organized political support

in the UK (personal communication).

34. Farmers in the ZRHDS panel decreased their

average fertilizer application per acre planted by some

27% between the mid-1980s and the late 1990s.

35. See Deininger et al. (2004) for further implications

of this point.


36. See Chiremba and Masters (2003) for a more

succinct treatment of farm productivity.

37. Parity with the urban minimum wage was the

original target for incomes in resettlement areas. A

slightly greater proportion of female- than male-headed

households achieved TGVPs below the minimum wage

(33% compared to 31%). Chikondo attributes this

difference to the fact that female-headed households

produce less because they are less well endowed and

have older household heads.

38. The criteria for the cut-off points are not made

explicit in the study.

39. This interpretation assumes that resettlement areas

pull in extra family members, whereas there is consid-

erable evidence that the larger household sizes in

resettlement areas are in considerable measure a con-

sequence of the poor state of the nonagricultural

economy.

40. An earlier but partial analysis (Madzivanyika,

1991) found no evidence of overstocking for schemes

as a whole.

41. The analysis included other resettlement models in

addition to Model A.

42. Harts-Broekhuis and Huisman (2001) also find that

settlers are better endowed in terms of farm equipment

than farmers in the communal areas but do not indicate

whether the implements were brought in upon settlement

or acquired subsequently.

43. Mutimukulu (1991), among others, provides an

account of infrastructural development in resettlement

areas, but all such treatments tend to focus on the

provision of physical infrastructure by the program. See,

for example, Madhuku (2002).

44. The corresponding value is US$0.99 at the parallel

market rate—an illustration of the difficulties of doing

meaningful analysis in an economy as distorted as

Zimbabwe�s.

45. This conclusion is reversed in a more recent study

(Kinsey, 2003), but these later conclusions are still

preliminary in nature.

46. Such effects are not, however, uniform across

different types of seasons.

47. Some indications of trends can be obtained from

work recently completed or underway. See Chikondo

(1996), Choga (1999), Elliott (1995, 2000), Elliott &

Campbell (2002), Elliott, Kinsey, & Kwesha (2002).

REFERENCES

Ackerman, P., & DuVall, J. (2000). A force morepowerful: a century of nonviolent conflict. New York,NY: Palgrave.

Addison, T. (Ed.) (2001). Reconstruction from war inAfrica: communities, entrepreneurs, and states.Discussion Paper No. 2001/18. Helsinki:World Institute for Development Economics Re-search.

Addison, T., & Laakso, L. (2003). The political economyof Zimbabwe�s descent into conflict. Journal ofInternational Development, 15(4), 457–470.

Alexander, J. (2003). ‘‘Squatters’’, veterans and the statein Zimbabwe. In A. Hammar & B. Raftopoulos, &S. Jensen (Eds.), Zimbabwe�s unfinished business:rethinking land, state and nation in the context ofcrisis (pp. 83–117). Harare: Weaver Press.

Alwang, J., & Ersado, L. (1999). Changes in poverty inZimbabwe: 1990–96 (mimeo).

Berdal, M., & Malone, D. (2000). Greed and grievance:economic agendas in civil wars. Boulder, CO: LynneReinner.

Bernstein, H. (2004). ‘‘Changing before our very eyes’’:agrarian questions and the politics of land incapitalism today. Journal of Agrarian Change, 4(1–2), 190–225.

Berry, R. A., & Cline, W. R. (1979). Agrarian structureand production in developing countries. Baltimore:Johns Hopkins University Press.

Bigsten, A., & Levin, J. (2001). Growth, income distri-bution and poverty: a review. Paper presented at theDevelopment conference growth and poverty, May 25–26, World Institute for Development EconomicsResearch, Helsinki.

Chigwenya, A. (2001). An assessment of environmentalimpacts of resettlement programmes in Zimbabwewith specific focus on Gutu South resettlementscheme. Unpublished B.Sc. Honours dissertation,Department of Rural and Urban Planning. Harare:University of Zimbabwe.

Chiremba, S. (2002). Productive efficiency of resettledfarmers in Zimbabwe. Unpublished M.Sc. thesis,Purdue University.

Chiremba, S., & Masters, W. (2003). The experience ofresettled farmers in Zimbabwe. African StudiesQuarterly, 7(2–3), 1–2http://web.africa.ufl.edu/asq/v7/v7i2a5.htm.

Chikondo, K. (1996). Production and management ofnatural resources in resettlement areas in Zimbabwe:the case of Masasa-Ringa. Unpublished M.Sc.dissertation. Agricultural University of Norway.

http://web.africa.ufl.edu/asq/v7/v7i2a5.htm

http://web.africa.ufl.edu/asq/v7/v7i2a5.htm


Choga, E. (1999). Management of wood biomassresources in resettlement areas: a case study ofHoyuyu Resettlement Scheme in Mashonaland EastProvince. Unpublished M.Sc. dissertation, Depart-ment of Rural and Urban Planning. Harare: Uni-versity of Zimbabwe.

Corbett, J. (1994). Livelihoods, food security and nutritionin a drought prone part of Zimbabwe. Oxford: Centrefor the Study of African Economies, University ofOxford.

Coudere, H., & Marijsse, S. (1988). ‘‘Rich’’ and ‘‘poor’’in Mutoko Communal Area. Zimbabwe Journal ofEconomics, 2(1), 1–25.

Cusworth, J., & Walker, J. (1988). Land resettlement inZimbabwe: a preliminary evaluation. EvaluationReport no. EV 434. London: Overseas DevelopmentAdministration.

Dagdeviren, H., van der Hoven, R., & Weeks, J. (2001).Redistribution does matter: growth and redistribu-tion for poverty reduction. Paper presented at theDevelopment conference growth and poverty, May 25–26, World Institute for Development EconomicsResearch, Helsinki, Finland.

Dale, S. (2001). Investigating the impact of agrarianpolicies on conflict and peacebuilding. Availablefrom http://www.le_num=960. [Reference: IDRCProject Number 0030569]. Accessed: June 19, 2001.

Davidow, J. (1983). Dealing with international crisis:lessons from Zimbabwe. Occasional paper no. 34.Muscatine, IA: Stanley Foundation.

Deininger, K., Hoogeveen, J., & Kinsey, B. (2000).Productivity and equity impacts of land reform: thecase of Zimbabwe. Paper presented at the Minisymposium on improving land access and asset own-ership by the poor through land reform: empiricalevidence and policy implications, at the XXIV inter-national conference of agricultural economists, August13–18, 2000, Berlin, Germany.

Deininger, K., Hoogeveen, H., & Kinsey, B. H. (2004).Economic benefits and costs of land redistribution inZimbabwe in the early 1980s. World Development,32(10), 1697–1709.

Dorner, P., & Kanel, D. (1971). The economic case forland reform. In P. Dorner (Ed.), Land reform in LatinAmerica. Madison, WI: Land Economics for theLand Tenure Center at the University of Wisconsin-Madison.

Dekker, M. (2004). Risk, Resettlement and Relations:Social Security in Rural Zimbabwe. Tinbergen Insti-tute Research Series No. 331. Amsterdam: TheTinbergen Institute and the Free University Amster-dam.

Elliott, J. A. (1995). Processes of interaction acrossresettlement/communal area boundaries in Zimba-bwe. Geographical Journal of Zimbabwe, 26, 1–16.

Elliott, J. A. (2000). Resource implications of landresettlement in Zimbabwe: insights from woodlandchanges. In T. Bowyer-Bower & C. Stoneman(Eds.), Land reform in Zimbabwe: constraints andprospects (pp. 145–158). Oxfordshire: Ashgate,Abingdon.

Elliott, J. A., & Campbell, M. (2002). The environmentalimprints and complexes of social dynamics in ruralAfrica: cases from Zimbabwe and Ghana. Geoforum,33, 221–237.

Elliott, J. A., Kinsey, B. H., & Kwesha, D. (2002). Long-term changes in land-use and resource entitlementsaccompanying changes in land tenure in Zimbabwe.Paper presented at the 9th Biennial conference of theinternational association for the study of commonproperty, June 17–21, Victoria Falls.

Esman, M. J. (1989). Commentary. In F. D. Colburn(Ed.), Everyday forms of peasant resistance(pp. 221–228). Armonk, New York: M.E. SharpeInc.

Griffin, K. B. (1974). The political economy of agrarianchange: an essay on the Green Revolution. London:Macmillan.

Griffin, K. B., Khan, A. R., & Ickowitz, A. (2002).Poverty and distribution of land. Journal of AgrarianChange, 2(3), 279–330.

Gunning, J. W., Hoddinott, J., Kinsey, B., & Owens, T.(2000). Revisiting forever gained: income dynamicsin the resettlement areas of Zimbabwe, 1983–97.Journal of Development Studies, 36(6), 131–154.

Hammar, A., & Raftopoulos, B. (2003). Zimbabwe�sunfinished business: rethinking land, state andnation. In A. Hammar, B. Raftopoulos, & S. Jensen(Eds.), Zimbabwe�s unfinished business: rethinkingland, state and nation in the context of crisis(pp. 1–47). Harare: Weaver Press.

Harts-Broekhuis, A., & Huisman, H. (2001). Resettle-ment revisited: land reform results in resource-poorregions in Zimbabwe. Geoforum, 32, 285–298.

Herbst, J. (1989). In F. D. Colburn (Ed.), How theweak succeed: tactics, political goods, and institu-tions in the struggle over land in Zimbabwe(pp. 198–220). Armonk, NY: M.E. Sharpe Inc.

Herbst, J. (1990). State politics in Zimbabwe. Harare:University of Zimbabwe Publications.

Hoogeveen, J., & Kinsey, B. (2001). Land reform,growth and equity: emerging evidence fromZimbabwe�s resettlement programme. A sequel.Journal of Southern African Studies, 27(1),127–136.

Hoogeveen, J., Kinsey, B., & Bouwmeester, K. (2003).Poverty in Zimbabwe: consequences of drought andeconomic reform for communal and land reformhouseholds (draft).

Horowitz, A. W. (1989). Time paths of land reform: atheoretical model and applications. Ann Arbor, MI:University Microfilms International, University ofWisconsin-Madison.

Jackson, R. (2001). Conflict resolution in Africa: inter-vention, indifference and indigenous solutions. Afri-can Affairs, 100, 321–328.

Jacobs, S. (2000). Zimbabwe: why land reform is agender issue. Sociological Research Online, 5(2).

Kaldor, M. (1999). New and old wars: organised violencein a global era. Cambridge: Polity Press.

King, R. (1977). Land reform: a world survey. Boulder,CO: Westview Press.

http://www.le_num=960


Kinsey, B. H. (1983). Emerging policy issues in Zimba-bwe�s land resettlement programmes. DevelopmentPolicy Review, 1(2).

Kinsey, B. H. (1987). Agricultural extension in intensiveresettlement schemes. Paper presented at the Work-shop on the National Extension Framework Study.May 12, 1987. Harare: Department of Agricultural,Technical and Extension Services.

Kinsey, B. H. (1998). Allowing land reform to work insouthern Africa: a long-term perspective on ruralrestructuring in Zimbabwe. In Proceedings of theinternational conference on land tenure in the devel-oping world. January 27–29, 1998, University of CapeTown.

Kinsey, B. H., Burger, K., & Gunning, J. W. (1998).Coping with drought in Zimbabwe: survey evidenceon responses of rural households to risk. WorldDevelopment, 26(1), 89–110.

Kinsey, B. H. (1999). Land reform, growth and equity:emerging evidence from Zimbabwe�s resettlementprogramme. Journal of Southern African Studies,25(2), 173–196.

Kinsey, B. H. (Ed.) (1999). Determinants of ruralhousehold incomes and their impact on poverty andfood security in Zimbabwe. Rome: Food and Agri-culture Organization.

Kinsey, B. H. (2000). The implications of land reformfor rural welfare. In T. Bowyer-Bower & C. Ston-eman (Eds.), Land reform in Zimbabwe: constraintsand prospects (pp. 103–118). Abingdon, Oxfordshire:Ashgate.

Kinsey, B. H. (2002a). After land, what? The long-termimpact of land redistribution on poverty in Zimba-bwe. Paper presented at the workshop: Causes ofpoverty and holistic strategies for poverty reductionin Southern Africa. University of Swaziland, Febru-ary 25–28, 2002.

Kinsey, B. H. (2002b). Who benefits from asset redis-tribution in Zimbabwe? Targeting, beneficiary selec-tion, and economic performance. Paper presented atthe conference: understanding poverty and growth inSub-Saharan Africa. Centre for the Study of AfricanEconomies, Oxford University, March 18–19, 2002.

Kinsey, B. H. (2003). Comparative economic perform-ance of Zimbabwe�s resettlement models. In M.Roth & F. Gonese (Eds.), Delivering land andsecuring rural livelihoods: post-independence landreform and resettlement in Zimbabwe(pp. 199–221). Harare and Madison, Wisconsin:Centre for Applied Social Sciences, University ofZimbabwe, and Land Tenure Center, University ofWisconsin-Madison.

Lehman, D. (1978). The death of land reform: a polemic.World Development, 6, 339–345.

Lipton, M. (1974). Towards a theory of land reform. InD. Lehmann (Ed.), Agrarian reform and agrarianreformism. London: Faber and Faber.

Lyne, M., Zille, P., & Graham, D. (2000). Financing themarket-based redistribution of land to disadvantagedfarmers and farm workers in South Africa: LandReform Credit Facility. Sociological Research Online,5(2).

Madhuku, N. S. (2002). An assessment of the pre-2000(1980–2000) period land reform programme withregard to infrastructure provision—A case study ofDeure Resettlement scheme in Bikita. UnpublishedB.Sc. Honours dissertation, Department of Ruraland Urban Planning. Harare: University of Zimba-bwe.

Madzivanyika, H. (1991). The resettlement pro-gramme in post-independence Zimbabwe: anassessment of Model A schemes with special focuson livestock production and use. UnpublishedB.Sc. Honours dissertation, Department of Ruraland Urban Planning. Harare: University of Zim-babwe.

Masilela, C., & Weiner, D. (1996). Resettlement plan-ning in Zimbabwe and South Africa�s rural andurban discourse. Third World Planning Review, 18(1),23–43.

Marongwe, N. (2003). Farm occupations and occupiersin the new politics of land in Zimbabwe. In A.Hammar, B. Raftopoulos, & S. Jensen (Eds.),Zimbabwe�s unfinished business: rethinking land, stateand nation in the context of crisis (pp. 155–190).Harare: Weaver Press.

Mhishi, S. G. (1995). A critical analysis of the landreform programme in Zimbabwe (mimeo).

Moyo, S. (1995). The land question in Zimbabwe. Harare:SAPES Books.

Mutimukulu, R. (1991). Infrastructural developmentin the resettlement schemes: a focus on Model Anormal intensive schemes. UnpublishedB.Sc. Honours dissertation, Department of Ruraland Urban Planning. Harare: University of Zim-babwe.

Nafziger, E. W., Stewart, F., & Vayrynen, R. (2000). Theorigins of humanitarian emergencies: war and dis-placement in developing countries (vol. 1) and Weakstates and vulnerable economies: humanitarian emer-gencies in developing countries (vol. 2). Oxford:Oxford University Press.

Owens, T., Hoddinott, J., & Kinsey, B. H. (2003).The impact of agricultural extension on farmproduction in resettlement areas of Zimbabwe.Economic Development and Cultural Change, 51(2),337–358.

Palmer, R. (2003). Future land and rural developmentpriorities in the new resettlement areas in Zimbabwe.Oxford, UK: Oxfam.

Paulini, T. (1979). Agrarian movements and reforms inIndia: the case of Kerala. Socio-economic studies onrural development no. 33. Saarbrucken: Verlag Bre-itenbrach.

Prosterman, R. L., & Reidinger, J. M. (1987). Landreform and democratic development. Baltimore: TheJohns Hopkins University Press.

Prosterman, R. L. (1976). IRI—A simplified predictiveindex of rural instability. Comparative Politics,8(343).

Quah, D. (2001). Some simple arithmetic on howincome inequality and economic growth matter.Paper presented at the Development conferencegrowth and poverty. May 25–26, World


Institute for Development Economics Research.,Helsinki.

Raftopoulos, B. (2003). The state in crisis: authori-tarian nationalism, selective citizenship and distor-tions of democracy in Zimbabwe. In A. Hammar,B. Raftopoulos, & S. Jensen (Eds.), Zimbabwe�sunfinished business: rethinking land, state and nationin the context of crisis (pp. 217–241). Harare:Weaver Press.

Reno, W. (1998). Warload Politics and African States.Boulder, CO: Lynne Reinner.

Robilliard, A.-S., Sukume, C., Yanoma, Y. & Lofgren,H. (2001). Land reform and poverty alleviation inZimbabwe: farm-level effects and cost-benefit analy-sis (mimeo).

Roth, M., & Gonese, F. (Eds.) (2003)Delivering land andsecuring rural livelihoods: post-independence landreform and resettlement in Zimbabwe. Harare andMadison, WI: Centre for Applied Social Sciences,University of Zimbabwe and Land Tenure Center,University of Wisconsin.

Rukuni, M. (1994). Report of the Commission of Inquiryinto Appropriate Agricultural Land Tenure Systems[The Rukuni Commission] vol. 1: Main Report &vol. 2: Technical Reports. Harare: GovernmentPrinter.

Russet, B. M. (1964). World Handbook of Political andSocial Indicators. New Haven: Yale University Press.

Scoones, I. (1996). Hazards and opportunities. Farminglivelihoods in dryland Africa: lessons from Zimbabwe.London: Zed Books Ltd.

Scudder, T., & Colson, E. (1982). From welfare todevelopment: A conceptual framework for theanalysis of dislocated people. In A. Hansen & A.Oliver-Smith (Eds.), Involuntary Migration andResettlement (pp. 267–287). Boulder, CO: WestviewPress.

Schultz, T. W. (1964). Transforming Traditional Agri-culture. New Haven: Yale University Press.

Sender, J., & Johnston, D. (2004). Searching for aweapon of mass production in rural Africa: uncon-vincing arguments for land reform. Journal ofAgrarian Change, 4(1 & 2), 142–164.

Tawonezvi, H. P. R. (1995). Status of livestock inresettlement schemes in Zimbabwe. Harare: Fried-rich-Ebert Stiftung.

Tuma, E. H. (1965). Twenty-six centuries of agrarianreform: a comparative analysis. Berkeley: Universityof California Press.

Uvin, P. (1998). Aiding violence: the development enter-prise in Rwanda. West Hartford, CT: KumarianPress.

Vhutuza, W. A. (1991). Review of the productionperformance of the Model A resettlement pro-gramme in low rainfall areas: focus on MushandikeResettlement Scheme in Masvingo. UnpublishedB.Sc. dissertation, Department of Rural and UrbanPlanning, University of Zimbabwe.

Warriner, D. (1969). Land reform in principle andpractice. Oxford: Clarendon Press.

Weiner, D., Moyo, S., Munslow, B., & O�Keefe, P.(1991). Land use and agricultural productivity inZimbabwe. In N. D. Mutizwa-Mangiza & A. H. J.

Helmsing (Eds.), Rural development and planning inZimbabwe (pp. 143–183). Aldershot, UK: Avebury/Gower Publishing Company Ltd.

Wohlgemuth, L., Gibson, S., Klasen, S., & Rothschild,E. (1999). Common security and civil society in Africa.Uppsala: Nordiska Africainstitutet.

World Bank (1998). Post-conflict Reconstruction. Wash-ington, DC: The World Bank.

Zimbabwe (1980). Resettlement policies and procedures.Harare: Ministry of Lands, Resettlement and RuralDevelopment.

Zimbabwe (1984). Report on results of the first nationalsurvey of the normal intensive resettlement pro-grammes. Harare: Department of Rural Develop-ment Ministry of Lands, Agriculture and RuralDevelopment.

Zimbabwe (1985). Intensive resettlement programme:policies and procedures. Harare: Department of RuralDevelopment.

Zimbabwe (1986a). First annual survey of settler house-houlds in normal intensive Model A resettlementschemes. Harare: Ministry of Lands, Agricultureand Rural Resettlement.

Zimbabwe (1986b). First five-year development plan,1986–90. Harare: Government Printer.

Zimbabwe (1990). The national land policy. Harare:Ministry of Lands, Agriculture and Rural Develop-ment.

Zimbabwe (1991). Second five-year development plan,1991–95. Harare: Government Printer.

Zimbabwe (1992a). Evaluation of the Model B resettle-ment schemes. Harare: Ministry of Community andCo-operative Development.

Zimbabwe (1992b). Policies and procedures: resettlementand the reorganisation and development of communallands. Harare: Ministry of Local Government, Ruraland Urban Development, Department of RuralDevelopment.

Zimbabwe (1992c). Second report of settler households innormal intensive Model A resettlement schemes(socio-economic survey 1988/89 and 1989/90): Mainreport. Harare: Ministry of Lands, Agriculture andRural Resettlement.

Zimbabwe (1993). Value for money project (specialreport) of the Comptroller and Auditor-General onthe land acquisition and resettlement programme.Harare: Office of the Comptroller and Auditor-General.

Zimbabwe (1995). Review of the land resettlementprogramme, 1980–95. Harare: Ministry of LocalGovernment, Rural and Urban Development, Dis-trict Development Fund.

Zimbabwe (1996). Memorandum to Cabinet onResettlement by the Minister of Local Government,Rural and Urban Development. Harare: Ministry ofLocal Government, Rural and Urban Develop-ment.

Zimbabwe (1998). 1995 Poverty Assessment Study Sur-vey Main Report. Ministry of Public Service, Labourand Social Welfare. Harare: Government PrintingOffice.

Zindi, C., & Stack, J. (1992). Income versus insur-ance: Preliminary findings on the contribution of


livestock to communal area farm household foodsecurity. In J. B. Wyckoff & M. Rukuni (Eds.), Foodsecurity research in Southern Africa: policy implica-

tions (pp. 293–316). Harare: University of Zimba-bwe, Department of Agricultural Economics andExtension.

Documents

Zimbabwes Land Reform Program: Underinvestment in Post … · 2005-03-07 · Zimbabwes Land Reform Program: Underinvestment in Post-Conﬂict Transformation BILL H. KINSEY * Free