34515 W O R L D B A N K W O R K I N G P A P E R N O . 7 1 Consensus, Confusion, and Controversy Selected Land Reform Issues in Sub-Saharan Africa Rogier van den Brink Glen Thomas Hans Binswanger John Bruce Frank Byamugisha THE WORLD BANK W O R L D B A N K W O R K I N G P A P E R N O . 7 1 Consensus, Confusion, and Controversy Selected Land Reform Issues in Sub-Saharan Africa Rogier van den Brink Glen Thomas Hans Binswanger John Bruce Frank Byamugisha THE WORLD BANK Washington, D.C. Copyright © 2006 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First Printing: November 2005 printed on recycled paper 1 2 3 4 5 07 06 05 World Bank Working Papers are published to communicate the results of the Bank's work to the development community with the least possible delay. The manuscript of this paper therefore has not been prepared in accordance with the procedures appropriate to formally- edited texts. Some sources cited in this paper may be informal documents that are not read- ily available. The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the International Bank for Reconstruc- tion and Development/The World Bank and its affiliated organizations, or those of the Exec- utive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank of the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission promptly to reproduce portions of the work. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Dan- vers, MA 01923, USA, Tel: 978-750-8400, Fax: 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, Fax: 202-522-2422, email: pubrights@worldbank.org. ISBN-10: 0-8213-6440-5 ISBN-13: 978-0-8213-6440-6 eISBN: 0-8213-6441-3 ISSN: 1726-5878 DOI: 10.1596/978-0-8213-6440-6 Rogier van den Brink is Senior Economist for the Africa Region of the World Bank. Glen Thomas is Director General of the Department of Land Affairs in the Ministry of Agriculture and Land Affairs of the Government of South Africa. Hans Binswanger was formerly Senior Advisor to the Vice President for the Africa Region of the World Bank. John Bruce is Senior Counsel for the Environmentally and Socially Sustainable Development Network and Inter- national Law for the World Bank. Frank Byamugisha is Operations Advisor for the Africa Region of the World Bank. Library of Congress Cataloging-in-Publication Data has been requested. Contents Abstract v Preface vii Introduction 1 1. Land Tenure and Land Policy 3 What is "Property"? 3 Definition 3 Security 4 Confusing Private, Common, and No Property 5 Full Tradability's Tradeoffs 7 The Evolution of Property Rights 8 Scarcity 8 Flexibility 10 Policy Recommendations 11 Formalizing Property Rights Regimes 12 Titling is Not Always Necessary 12 If Not Titling, Then What? 13 The Consensus 15 2. Land Redistribution 17 The Case for Land Redistribution 18 Efficiency 18 Equity, Growth and Poverty Reduction 21 Fairness 23 Land Redistribution Policies: Lessons Learned 24 Land Reform in South Africa 25 Land Markets Need Help 29 Land Acquisition and Resettlement Can Be Slow and Costly 32 Land Reform is Usually Underfunded 39 Large Farmers Will Lobby Against Land Reform 41 Policy Recommendations 43 Land Market Reforms 43 Land Acquisition Methods 44 Resettlement Models 45 National Implementation Strategy 46 Conclusion 47 References 49 iii Abstract and reform can broadly be divided up into land tenure reform--the establishment of L secure and formalized property rights in land--and land redistribution--the transfer of land from large to small farmers. The paper is therefore divided into two chapters. The first chapter gives a short narrative of some of the key land tenure and land pol- icy issues. While these issues remain politically sensitive, there is a solid consensus emerg- ing on how to deal with them, but only once the confusion surrounding private and common property, and formal and informal rights, is cleared up. In particular, secure property rights should not be confused with full private "ownership." Under certain eco- nomic conditions, property rights tend to become more individualized and formalized. However, the introduction of private title in situations where such economic conditions do not exist can be a waste of effort. The second chapter addresses the redistribution of property rights in land from large to small farmers--redistributive land reform. There is a heightened sense of urgency on the need to address land redistribution, especially in the former settler colonies in South- ern Africa, but controversy exists on the appropriate implementation mechanisms. This controversy, combined with the sensitive political nature of land reform, often results in costly inaction. We highlight the case of South Africa, because success there would have tremendous regional and international implications for land redistribution. A policy framework for redistributive land reform is outlined within which the competing paradigms compete there where it actually matters: on the ground. One can think of worse things than a situ- ation in which the various stakeholders in government, the private sector, and civil society agree on an overall framework for implementation and compete with each other on the ground to demonstrate the success of "their" model. Major land redistribution can be implemented peacefully: history need not repeat itself ad nauseam. This paper is targeted at land reform practitioners and stakeholders in government and civil society. The paper would also be of interest to students and researchers as a primer on some of the main issues. v Preface T his paper was written by Rogier van den Brink (Senior Economist, Africa Region, the World Bank), Hans Binswanger (Consultant, formerly Senior Advisor to the Vice President for the Africa Region of the World Bank), John Bruce (Senior Counsel, Environmentally and Socially Sustainable Development Network and Inter- national Law, the World Bank), Glen Thomas (Director General of the Department of Land Affairs, Ministry for Agriculture and Land Affairs, the Government of South Africa), and Frank Byamugisha (Operations Advisor, Africa Region, the World Bank). Natasha Mukherjee edited the text. Thepaperhasitsoriginsinabackgroundreportforaregionalworkshopon"LandIssues in Africa," organized by the World Bank and other donors, and held in Kampala, Uganda (April 29­May 2, 2002). This workshop, together with three other workshops held in other regions of the world, formed part of the consultations held for a World Bank Policy Research Paper entitled "Land Policies for Growth and Poverty Reduction" (Deininger 2003). During the next two years, versions of the background paper were used at various work- shops and seminars to give an overview on land issues in sub-Saharan Africa. These can be found on a number of public websites, including that of the International Federation of Sur- veyors,1 the Southern African Regional Poverty Network,2 the Friedrich Ebert Stiftung and the Economist,3 the International Land Coalition,4 the Institute for Public Policy Research,5 Oxfam,6 and the Foro Mundial sobre la Reforma Agraria.7 The second half of this paper was commissioned by the Harold Wolpe Foundation, and will be published as part of an edited book on land reform in South Africa by the Human Sciences Research Council Press of South Africa. It is based on a review of South Africa's Land Redistribution for Agricultural Development Program. Earlier versions of this paper benefited from comments by Dan Bromley, Ben Cousins, Alain De Janvry, Vincent Hungwe, Odenda Lumumba, Michael Nowak, Francisco Pichon, and Gine Zwart. The idea of a multi-option policy framework for redistributive land reform owes an intellectual debt to Vincent Hungwe and Sam Moyo. It came out of our (failed) attempt to reach a negotiated agreement on the Zimbabwe land reform program between 1998 and 2000. In that context, we are also grateful for discussions with Tom Allen, Caesar Chidawanyika, Mike Mispelaar, Joseph Msika, Jorge Munoz, Robbie Mupawose, Nick Swanepoel, Greg Brackenridge, and many anonymous others. 1. www.fig.net/commission7/pretoria_2002 2. www.sarpn.org.za/documents/d0000115/index.php 3. http://fesportal.fes.de/pls/portal30/docs/FOLDER/INTERNATIONAL/AFRIKA/Land_Reform_ SSA_Rogier_vd_Brink.PDF; and http://www.economist.com/displayStory.cfm?story_id=4157564 4. http://www.landcoalition.org/pdf/wbqrdbaf.pdf 5. http://www.ippr.org.na/Non-IPPR%20research/Land%20Reform%20Policy%20in% 20SSA%20Nov%201%202002.pdf 6. http://www.oxfam.org.uk/what_we_do/issues/livelihoods/landrights/africa_gen.htm 7. http://www.fmra.org Introduction L and reform can be divided broadly up into land tenure reform--the establishment of secure and formalized property rights in land--and land redistribution--the transfer of land from large to small farmers. The paper is therefore divided into two chapters.Thefirstchapterfocusesonpropertyrightsinland--itgivesashortnarrativeofsome of the key land tenure and land policy issues. While these issues remain politically sensitive, there is a solid consensus emerging on how to deal with them, but only once the confusion surrounding private and common property, and formal and informal rights, is cleared up. The second chapter addresses the redistribution of property rights in land from large to small farmers--redistributive land reform. There is a heightened sense of urgency on the need to address land redistribution, especially in the former settler colonies in South- ern Africa, but controversy exists on the appropriate implementation mechanisms. We highlight the case of South Africa, because success there would have tremendous regional and international implications for land redistribution. A policy framework for redistribu- tive land reform is outlined within which the competing paradigms compete where it actu- ally matters: on the ground. 1 CHAPTER 1 Land Tenure and Land Policy What is "Property"? What do we mean when we say I "own property," or I "own land" and assert "this is mine"? Property is a social relation. It is about rights and duties. It defines what an individual, a com- munity, or the state can and cannot do with a certain commodity, and what needs to be respected by others--think of property as a "bundle" of "my" rights and "your" obligations. If "landownership" is a social relation, it immediately follows that making policy rec- ommendations about landownership is not a technical matter. Land policy and land reform are about social relations, and therefore are invariably about "politics." To say that land reform is political is a tautology. Definition To define what property exactly means in a particular context, one needs to ask the "who, what, where, when, and how" questions about it. For instance, who defines property rights? And who enforces them? Defining and enforcing those rights and obligations is up to the community, or, when property rights become more formalized and legislated, the state. How is the property right acquired? Was it through sale or inheritance; was it by virtue of being a member of a certain group or community; through first or "good faith" occupa- tion; because certain investments were made, or through land reform? What benefits and income streams are included? How precisely are these defined? Where and when can these rights be exercised? What is the defined time period for the right to undertake a certain activity and reap its rewards, or incur the liability? In other words, what exactly is in the bundle of property rights we are talking about? 3 4 World Bank Working Paper The answers to these questions are derived from customs, norms, legal traditions and principles, laws, negotiations, and revolutions. The answers, and the social relations, or "politics," they represent, always differ from place to place and from time to time, both in the developing and in the developed world. For instance, there is no universal, technical definition of "ownership." Sometimes one can "own" a plot of land but one is not allowed to farm on it--only to build a house on it. In some places one cannot just build any house--it has to be a house of a certain type and a certain color. In other words, there is no universal definition--empirical or normative-- of the bundle of rights we loosely refer to as "ownership." There are many ways of defin- ing property rights, none of which have a claim to optimality, because the definition is, firstly, the outcome of a political or social process, and, secondly, the result of a wide range of economic and environmental factors. During the 1970s and 1980s, the consensus among development practitioners was that "optimal" property rights were best guaranteed under a formalized (that is, docu- mented) and private property regime, and that economic growth and environmental stewardship would be promoted by making the bundle of rights as large as possible, ter- ritorially exclusive, of infinite duration and fully tradable. In the next sections, a num- ber of reasons are given for the changing consensus on what constitutes "optimal" property regimes for development. The emerging consensus is far more agnostic about the "optimal" definition of property rights, and about the urgency of arriving at that state of affairs, than the earlier view. Security Property rights should be defined by the community (or the state), accepted and understood by all, and be able to be enforced. When a community, or the state, is able to enforce what it decides, property rights acquire a very desirable characteristic. They become certain--and tenure, the holding of the right, becomes secure. Security matters for investment. A farmer needs to know that if she sows maize she will own the harvest--this way she will do her best to farm well. Further, she will logically start thinking about future seasons and invest in maintaining the fertility of the soil. Anything that makes a farmer worry about whether or not she will be able to reap this harvest--this year and all the next years--will make her wonder about investing in her crop and in her field. This is why there is consensus that property rights need to be secure. Sjaastad and Bromley (2000) remind us that we should define "security" as assurance, and not extend its meaning to other dimensions of a particular property right, such as its duration or its marketability or its "breadth" (the number of different activities captured under the right, in effect the "size of the bundle"). In other words, a short-term lease, for its duration, is not less secure than a long-term lease. It may provide adequate tenure, depending on the nature of the land use. Maize production does not require a fifty-year lease, but hardwood plantations do. As long as the period of the lease allows the leaseholder to reap the full benefits of the investment to be made on the land, the lease will be perfectly incentive compatible--it is an appropriate property right from this particular investment perspective. Similarly, it is not necessary to ask for the largest possible bundle of property rights because, for example, to care for the fertility of your maize field, you do not need to own the mining rights, as well. Consensus, Confusion and Controversy 5 What is the situation in sub-Saharan Africa, where formally codified property rights regimes are still quite rare, and most land can be seen to fall under customary law? Bruce and Migot-Adholla (1993) summarized the research on the security of customary tenure in rural Africa and concluded the following: Overall, the security provided by customary tenure is quite strong and customary tenure has evolved to accommodate new needs, such as making land tradable. The existing property rights systems in many rural areas do not seem to be in need of a wholesale replacement with new property rights regimes. Titling programs may be needed for most urban, peri-urban and high-intensity agricultural areas, but alternative policies are needed to strengthen security of tenure in most rural areas. The authors suggest recognition by the state of customary tenure systems, and increased for- malization of those systems. This approach has been tested in World Bank-supported activities in a number of coun- tries, including Côte d'Ivoire, Ghana, Mozambique, Uganda, and Tanzania. To illustrate the findings, we will discuss two such projects, one just concluded, the other just begun. IntheRuralLandManagementandCommunityInfrastructureDevelopmentProjectin Côted'Ivoire,approvedin1997,theBankfundedthetitlingofcustomaryratherthan"West- ern"tenurerights.Thetitlingwasproceedinginseveralregionsofthecountryuntil2000,but was then seriously disrupted by the growing civil war in the country. Two insights emerged. While the intent of the project was to accurately reflect customary rights, in practice, it was found difficult to include all "secondary rights" of the community and other groups or indi- viduals within the community. These can be extensive and complex, and it was found that the process used to recording these rights failed to capture many of these, with the result that the project was merely accomplishing a simplification of rights. This tended to strengthen the position of the individual landholder at the expense of other rights holders. Second, it was found that basing the adjudication process on customary rights disadvantaged mem- bers of immigrant communities who, in some cases, had been in possession of the land for generations. The Ghana Land Administration Project (2004), in addition to titling in urban areas, is beginning to work with chiefs and their councils to support traditional land adminis- tration, including strengthening its processes, recordkeeping, transparency, and account- ability. The project started in 2004, and fifty customary areas (five per region) have been identified for support over the project period. Confusing Private, Common, and No Property The distinction between private and common property continues to be an area where there is neither consensus nor controversy, only considerable confusion. Common property is simply the property of a group, or, to put it differently "common property represents pri- vate property for the group" (Bromley 1992).8 A logical extension of common property is public property or state property. This is, in essence, common property but with the 8. "Common property" in the natural resource management literature is often used to specify a man- aged commons, used simultaneously or sequentially by members of the group that owns the land, in con- trast to the free-for-all on open access land. As the usage of this term suggests, "Common property" is managed by the group because it is its property and it has a right and every incentive to manage it. This is generally sound reasoning, but here the term "common property" is used in a narrower, legal sense: prop- erty owned in common by a group. It may or may not be used as a commons, and it may or may not be managed effectively. 6 World Bank Working Paper community now being a much larger community, say a city, or the nation as a whole, for- mally represented by the state. The confusion around private and common property stems from prejudices and per- ceptions. For instance, we associate the concept of private property with individual freedom. Under private property, we imagine, a person can do as she pleases. For instance, we think of private property as a tradable right which can be sold by the individual to anyone, without asking anybody else for permission to do so. We associate it with a sign that says "Keep Out. Private Property."--in other words, with a territorial boundary that excludes others. When we think of common property, on the other hand, we imagine non-tradability: either very restricted, permissible use of the asset, or the tragedy of a complete free-for-all. These stereotypes can be misleading. Everywhere in the world what one can and can- not do with private property is always regulated, for instance by zoning laws, building restrictions, the obligation by the owner to allow the public access for hiking or fishing, or to allow another individuals to establish a mining operation (in the event that mining rights are sold separately from other rights). In reality, the presumed unfettered sale of private property is commonly restricted. For instance, restrictions on the sale of agricultural land to foreigners exist in many countries in the world, including many states in the United States, Switzerland, and Denmark, and within Sub-Saharan Africa in Zambia, Nigeria, Kenya, Uganda, Ghana, and Tanzania. These restrictions can range from a total ban on sales to foreigners to the requirement of obtaining a special permission. Or they can mean that foreigners are not allowed to obtain freehold property, only leasehold. Common property regimes do the same: they may ban the sale of property to outsiders (non-membersofthecommunity),orrequireaspecialpermissionfromthecommunity.Yet, common property regimes can, and often do, allow the sale of the "shares" to others, just like private property regimes. The only difference is that--in the case of private property--one seeks permission from the state, while under common property one asks the community. To add to the confusion, within a common property rights regime, property rights can be individual or group. Simply put, a community can decide to give certain rights to some- thing to an individual or to the group. It may allocate a right to produce crops on a partic- ular plot to an individual, but allocate a grazing area to a group--for instance, all families living in the community. Sometimes these individual rights, within a common property regime, can be inheritable, exchangeable, rentable and even saleable. Under a common property regime, individual rights are regulated by the community. While, under a private property regime, individual rights are regulated by the state. Common property does not imply that the entire bundle of rights is given only to the group as a whole, or that the community engages in collective production.9 Within a com- mon property regime, rights can be assigned to individuals, like in the case of condo- minium apartments in the United States. The latter system is a good example of highly efficient common property systems in the developed world. 9. In fact, whereas many areas in Africa fall under common property regimes, cultivation is tradi- tionally a family affair and the modern African experience with group farming or collective production has been an unmitigated failure (for example, Ujamaa in Tanzania, and numerous cooperative farming experiments elsewhere in Africa). What is common, though, is a system under which people set aside some time to collectively work a particular field for the traditional head of the group, to be used by him to pro- vide the food or drinks consumed by all during public ceremonies, or to needy households. Consensus, Confusion and Controversy 7 If there is no security at all, one ends up with a situation in which there is no property-- a situation of open access. Unfortunately, this situation is often wrongly referred to as the "tragedy of the commons" giving the concept of common property a bad name. The phrase was coined by Hardin (1968), and should be rephrased as "the tragedy of res nullius"--no one's property. The tragedy of the commons is not the tragedy of common property rights, it is the tragedy of open access, which occurs when communities are no longer able to define and enforce the property rules which apply to natural resources. It then becomes a free-for-all--open access--and everybody has a rational interest in depleting the resource as much as possible, because if they do not, somebody else will. Degradation of the resource is a foregone conclusion in this situation. When communities are able to define and enforce common property rights and rules, no "tragedy of the commons" need occur, as parts of the Alps in modern Switzerland demonstrate, having been under common property since the Middle Ages (Netting 1976). As the incidence of land degradation the world over demonstrates, neither private nor common property regimes can claim to be inherently more environmentally sustainable than the other. Poverty, technology, greed, short-sightedness, and a host of other factors come into play. Full Tradability's Tradeoffs The other idea under the old consensus was that investment incentives are strongest, and economic efficiency best served, if the property right is private, of infinite duration and fully tradable. In this way, it was argued, the productivity of the resource is maximally exploited, and if a particular owner is unable to extract the maximum profits from it, the property would be sold to someone who would. In a world of perfect markets, the argument holds. It breaks down when information and credit market imperfections are considered. For instance, one of the great advantages of many common property regimes in Africa is their risk insurance, or social safety net func- tion: community members can claim access to land for farming when necessary. In the absence of a formal social security scheme, this insurance function of common property regimes has reduced the poverty impact of the many external shocks and macroeconomic crises that have hit Africa. Similarly, in societies that do not have pension schemes, the pivotal role that the elders play in the allocation of individual property rights to land ensures that they have a strong bargaining position vis-à-vis the young adults. The elders use these powers to ensure that the young contribute to their "pension." Common property regimes often constitute very important social insurance mecha- nisms for the old and the poor--we need to be mindful of this when we add up the costs and benefits of moving to private property (Jodha 1992). Under conditions where holding land is one of the most important forms of social security, moving toward property regimes under which land can be freely sold can have very negative social, and developmental, con- sequences. In these cases, better functioning land rental markets can combine the social insurance benefits of holding land for the poor with the production benefits of allowing land to be cultivated by those with the greatest capacity to do so. Just as it is wrong to vilify common property regimes, it is equally misguided to roman- ticize them. "Community failure" occurs, just as does "market," or "government failure." 8 World Bank Working Paper Common property regimes can provide important insurance functions, but they also can be used to exclude people, especially those who are politically weak or not "real" members of the community, for instance, women, especially widows, and outsiders. The Evolution of Property Rights Economic institutions--or what we might call "the rules of the economic game"--governing propertyrightsemergetocapture,orinternalize,thefullincomestreamsfromscarceresources (land, capital, labor). In a world of imperfect information, the emergence of economic insti- tutions is constrained by transactions costs, that is, the costs of making an economic contract work properly in the face of possible opportunistic behavior by the parties involved. In short: propertyrightsregimesareeconomiccontractswhichareconstrainedbytransactionscosts-- the costs of information gathering, contracting and enforcing of contracts. In the literature on property rights, there is substantial emphasis on transactions costs to explain why certain property rights regimes exist (for example, Demsetz 1967). In par- ticular, the emergence of formalized private property is often directly linked to the value of the resource. The more valuable the resource becomes, the higher is the need, and the ability, of a property regime to sustain increasing transactions costs (in terms of the preci- sion and formalization of its definition), the monitoring of its use, and the legal protection necessary to defend it. Conversely, less valuable resources are unable to sustain property rights regimes with high "maintenance costs" and are therefore not optimal. Scarcity If property rights focus on scarce resources, the most obvious question to ask, in any given setting, is "what is the scarce resource?" Comparing factor scarcity across the continents, one sees that substantial areas of sub-Saharan Africa are basically land abundant and labor scarce, as compared to the land scarcity and high population densities of many parts of Europe and Asia. It stands to reason, then, that many social and economic institutions in Africa, unlike those in Europe or Asia, have evolved to enhance or secure the productivity of the individual, rather than of the land. Compare for instance the African practice of "bride price," under which the husband's family "pays" for the reproductive and labor services of the bride, with the Asian practice of the "dowry," under which the compensation flows exactly the other way. In Africa, labor is scarce, and an institution--the "bride price"--is induced to capture this value. In Asia, labor is abundant, and "dowry" emerges. Another example is the use of "labor tenancy" by white settlers in Southern Africa under which African peasants--the scarce labor factor-- were allowed to reside, farm a plot, and graze cattle on the settler's land in return for a cer- tain number of days labor.10 In Asia, in contrast, the more common form of tenancy is share cropping, with land as the scarce factor. 10. Labor tenancy occurs when labor, especially during seasonal peaks, is scarce, supervision capacity of the landowner is limited, and landowners have abundant land (Lastarria-Cornhiel and Melmed-Sanjak 1998). Consensus, Confusion and Controversy 9 This may explain why, in land-abundant areas of Africa, where forms of shifting cul- tivation may still be practiced, an individual will rarely say that he or she "owns the land."11 Instead, the customary law principle will usually be that every adult man and woman has a right to reap what he or she sowed.12 People have a right to the fruits of their labor: the property right to the productivity of the individual is secured. A farmer sows maize, weeds, and harvests, and she owns the harvest. That property right is very secure. If animals come and graze in a yet-to-be-harvested maize field, their owner will be fined for taking some- one's property. Owning the fruits of one's labor is very different from asserting that one owns the land. However, in areas of Africa where land is scarce, land is owned by lineages established by first settlers. Their descendents inherit their land, farm it for their lifetimes, and pass it to the next generation. These landholders may not say that they are the "owners" (owner- ship may be considered to reside with the lineage, or the ancestors) but they undisputedly have a strong sense that the land is their property. In such cases, strong individual rights exist within the lineage's common property. In some areas, the individual right is so strong that the common property is more of a legal fiction than a reality. Population growth and the increased profitability of crop cultivation affect the balance between land and labor. The generalized right to use land in a particular way, such as for cul- tivation, is derived from the person's membership in the community or lineage. This right is usually inheritable and secure. In situations where land is abundant, and soil fertility is main- tained by some form of shifting cultivation, this right will not be tied to a specific parcel of land. Yet, when population growth makes land scarcer, and economic and agronomic con- ditions permit crop and tree cultivation to intensify, property rights to particular parcels become more permanent and inheritable. Individualization is driven by the intensification of agriculture which, in turn, is caused by population growth and increased market access (Boserup 1981, Bruce and Migot-Adholla 1993). Typically, communities start by individualizing permanent resi- dential and garden plots, then allocate individual rights to nearby fertile farming plots, and progressively extend individualization to the remaining areas under community ownership, until only wasteland and land for common infrastructure and facilities is owned by the community (Binswanger and Rosenzweig 1986, Binswanger and McIntire 1987). Many common property regimes will allow individual usufruct rights to a spe- cific plot to become more permanent, often as a direct result of investments in the land. In other words, the security of individual property rights is created by the act of invest- ing, rather than the other way around. Yet, agronomic and economic conditions in many parts of Africa put severe con- straints on intensification. These include: shallow and infertile soils, high temperatures which decompose organic material quickly, or make fertilizer "burn" the crop, erratic but sometimes torrential rainfalls which leach away nutrients fast, and show extreme 11. The "ownership" is most often in a lineage, and may be seen as belonging to ancestors. In other cases, the original occupants may be a different ethnic group than the group which has political control over an area. 12. There may be restrictions on who is allowed to sow what, in line with the social and political organi- zation of the ethnic group. For instance, sometimes women are only allowed to sow annual, not perennial crops.Thelatterimplylonger-termusufructrights,followingtheprincipleofowningthefruitsofone'slabor. 10 World Bank Working Paper variability over time and space. The simple agronomic constraints to agricultural inten- sification can be formidable. The economic incentives that would encourage increased intensification are often weak, with output prices low and input prices at the farm-gate high. Many imported "miracle," yield-increasing agricultural practices fail to take root, simply because of their lack of profitability given the prices for outputs and inputs that farmers face. Under these conditions, subsistence farming or shifting cultivation need not be the result of a lack of knowledge about agricultural innovations. These are sim- ply economically rational responses to the prices and transactions costs farmers face (Sadoulet and de Janvry 1995). Flexibility Similarly, livestock production systems based on nomadism or transhumance need not be a demonstration of economic backwardness, but instead a perfectly rational response to economic conditions. In semi-arid and arid areas, rainfall variability, and hence the avail- ability of water and fodder, may be so high, that livestock production will be based on a system which allows the herd to move over great distances. The spatial mobility of pastoral systems--nomadic or transhumance--exploits the economic benefits associated with flexibility--a benefit which can be shown to increase with increased rainfall variability (van den Brink, Bromley, and Chavas 1995). Pastoralists do not want fences because they know that their potential grazing area, given highly-variable rainfall, would be very large, and probably, given the regularity of serious droughts, the fence could never be large enough. Compare these flexible livestock production systems with fisheries. Surely, no one would seriously argue that the optimal property rights for fisheries consist of erecting fences in the middle of the ocean. One would prefer to manage the resource through other types of property rights systems, such as by restricting the number of fishermen, the type of nets they are allowed to use, the max- imum allowable catch, or the time period for catching fish. Fish in the Atlantic are as "fugi- tive" as grass in the semi-arid areas of Africa. Weighing the costs and benefits, fishermen and pastoralists would rather have a non-exclusive access to a large area, than exclusive access to a relatively small area. On the grazing lands of Africa, property rights regimes have emerged where the income streams from mobility have been captured, not the income streams from irrele- vant or inferior production techniques, such as attempts to establish intensively-grazed ranches in areas with extreme rainfall, and hence, extreme fodder availability. Optimal property rights in land--rights which capture the full income stream of live- stock grazing--have taken the form of access rights to grazing land, which need to be nego- tiated with other users, such as sedentary farmers and other pastoralists. In order to prevent overgrazing and conflict, these pastoral access rights are not "open access," but specific rights restricted to a well-defined number of property right holders. The areas where such property rights apply are not "unused" or "vacant." Traditionally, pastoralists and farmers would sit down to discuss how to make sure that both parties could exercise their rights without getting in each other's way. This type of coordination is often most efficiently achieved under a common property regime. How- ever, in some parts of Africa supported by the earlier private property consensus, pas- toralists have been told to become sedentary farmers, or fences have been erected and new Consensus, Confusion and Controversy 11 property rights have been created which obstruct the movements of the pastoralists, essen- tially depriving them of access rights they traditionally held. Sometimes the resulting ten- sions have led to violence.13 What pastoralists want are property rights that match their activities: access rights and rules to prevent over-use of the resource. To start with, pastoralists would like their historic economic rights to be respected by the state and farming communities. The new consensus therefore recommends that governments create the possibility of resolving such potential conflicts and support dialogue so that communities can find ways of deciding together how the bundle of property rights should be allocated and enforced (McCarthy and others 2000). The idea of taking down existing fences to increase ani- mal production and reduce environmental degradation in semi-arid areas is a promis- ing one, but, so far, has been applied mainly in game farming. Fortunately, this situation is changing. In Tanzania, a Range Lands Bill which protects access rights of pastoralists to grazing land is being processed. In Uganda, so-called "cattle corridors" have been created that give pastoralists enough room to roam, and protect the rights of sedentary farmers. Policy Recommendations As markets become more accessible, economies develop, population pressure grows, and, if agronomic conditions permit, farmers intensify production and invest in the sustainable exploitation of the natural resources. The value of the resources goes up, and that means often that the potential for competition and conflict grows. Sometimes communities are perfectly capable of handling the situation, by continuing to enforce and adapt the traditional rules that govern property rights, even going as far as creating "informal" doc- umentation and adjudication mechanisms. However, whereas the legitimacy and enforce- ability of these property rights, backed as they are by the community, may be quite strong, this does not, by definition, extend beyond the boundaries of the community. Sometimes the capacity of the community to manage competition and conflict declines, or is under- mined by the intervention by external forces, or even by the state itself. As a result, the com- munity, or the state, may decide that it is time that rules become more formalized, documented, and more easily enforced in a court of law.14 North and Thomas (1977) and De Soto (2001), among others, argue forcefully that it is this process of more carefully defining and formalizing property rights which has been one of the driving forces of eco- nomic development in world history. What can be done to assist this process? And what are the pitfalls? 13. Van den Brink, Bromley, and Chavas (1995) coined this problem the "Cain and Abel" problem-- the first recorded clash between a pastoralist and a farmer in history. 14. Examples of development programs to assist communities to register property rights: (i) Côte d'Ivoire, where both individual and communal rights are registered under customary law; (ii) Colombia titles natural resources exclusively to Afro-American and indigenous communities rather than to indi- viduals; and (iii) Mexico gives communities the choice between individual and communal title, with well- defined mechanisms which allow the transition between either--more than 50 million hectares were registered in about six years. 12 World Bank Working Paper Formalizing Property Rights Regimes The African case, with its customary land tenure systems, invites a closer look at the term "informality." "Informality," as it is used by De Soto and others, carries with it a vision of squatters on state lands outside Lima. They are the politically powerless, using land to which they have no rights, and being kept in legal limbo. Such a situation can often easily be for- malized, especially when the landowner is the state, which can agree to give land to squatters with zero or low compensation. In Africa, such formalization of informality in urban areas is currently undertaken in Tanzania, where the Government has formalized squatter rights for 5,000 plots of unplanned settlements since July 2004. Plans exist to expand the program to cover the remaining 400,000 plots in Dar Es Salaam. This will be done by issuing occu- pants with two-year renewable residential licenses, without the need to do a cadastral survey. After three years, these licenses assure the holders of full compensation if evicted by the state. The licensed occupants are required to pay ground rent and can have their rights upgraded to certificates of occupancy, which are 33-year long term leases, which can be mortgaged. These certificateswillbebasedoncadastralsurveys,providedthoseoccupantshaveconsistently met the earlier licensing conditions. Unlike a squatter in Lima or Dar Es Salaam, the African landholder in most rural areas does not hold land informally, but according to an alternative formality, a community-based formality. His rights may be unrecorded, but they have a legitimacy that originates with the local community, and are part of a larger land tenure system of the community. Some of the rules concerned are culturally embedded, especially those concerning inheritance. There will be traditional institutions within the community with vested interests in the maintenance of that system. Even if such a system is in decline, it can make assumption by the state of the responsibility for providing security land tenure a difficult and chal- lenging task. Titling is Not Always Necessary During the 1970s and 1980s, there was a consensus on the need to formalize property rights by creating documentary evidence--title deeds. These were underpinned by detailed sur- veys of property boundaries, so-called cadastral surveys. The earlier consensus on this issue has since changed and become more nuanced. For instance, most policy analysts now no longer simply assume that formalization in a given context necessarily increases tenure security, and leads to collateralized lending. The original assumptions have now become questions for empirical research. One of these questions is: will security be higher if someone has a formal property right--a title deed or a lease, which is issued by the State? This need not be the case, as some title deeds are not worth the paper they are written on, and create more confusion than security (Bruce and Migot-Adholla 1993). Simply put, title deeds in themselves do not cre- ate secure property rights from an insecure situation. Creating secure and formalized rights can be a formidable task, involving major political, legal, and social reform. In many African countries title deeds create room for opportunistic behavior. Political and eco- nomic elites can use their influence with the land administration agencies to acquire title deeds in a non-transparent manner, confiscating existing informal property rights of local communities or unregistered state lands. Conversely, some property rights which are only informally agreed on and enforced, can be very secure. Consensus, Confusion and Controversy 13 When does land start to function as collateral for lending? The evidence shows that simply introducing title deeds may not lead to collateralized lending. For instance, as dis- cussed above, if land is not the scarce factor, it will have little value as collateral. Even under common property regimes which have allowed individual property to emerge, barriers against land sales to outsiders might exist, making foreclosure either difficult or impossi- ble, as has been the case in many rural areas of Kenya. Finally, it is unlikely that a banking system would emerge merely because of the introduction of title deeds. Successfully funding small-scale farming also does not necessarily need a title deeds register and collateralized lending. The emergence of cooperative banking in 19th century Western Europe can serve as an example. When farmers found that banks would not lend to them on an individual basis, farmers' groups--some of which were already organized into trading cooperatives--proposed to be jointly liable for any and all loans of its mem- bership. Furthermore, financial institutions can operate efficiently without the use of col- lateral, by carefully evaluating the intrinsic merits of a project proposal or an applicant's repayment record. Successful examples of this principle can be found in the management of U.S. credit cards, or within the micro-finance sector. In many rural areas of Sub-Saharan Africa the potential for mortgage-based lending is limited. Such potential, however, does exist in the urban areas and in areas characterized by intensive agriculture, in particular irrigated areas. In those areas, common property regimes may no longer be functional, land will have considerable value, and a banking sys- tem is often already in place. Under such conditions, the chances that a formalization of tradable property rights will lead to increased access to credit are significant. If Not Titling, Then What? How does land policy accommodate these differing needs within a single polity? Titling programs obviously have their place on high-value land, in particular in urban, peri-urban and high-intensity farm areas. However, there are many rural areas where wall-to-wall titling would make no sense, because most landholders would not have the opportunity to practice the new behavior that a title would permit, such as mortgaging the land. Though, a few may, and traditionally their interests have been accommodated through "sporadic" titling, in which an individual applies for and covers most of the costs of survey and regis- tration of his holding. This will work for the investor who wants to irrigate a holding or build a mill or a hotel or even a fine house. What should be done for the bulk of the rural people? Their customary interests in land may be protected reasonably well under many, but not all, customary tenure systems. In some societies in southern Africa, notably Swazi culture, successful investment on land excites jealousy, both on the part of traditional rulers and among neighbors, and can result in accusations of witchcraft. The "offender" may be expelled from the community; his land is not taken from him, he is taken from it. In countries that have gone through long conflicts resulting in displacement of mil- lions of people from their customary lands for many years (for example, Mozambique, Rwanda, Angola, southern Sudan, and northern Uganda), customary interests in land may not be adequately protected due to a breakdown in the traditional rules and leadership structures. In such situations, alternative arrangements would need to be explored to pro- vide reasonable protection of customary interests during the post-conflict era. 14 World Bank Working Paper Similarly, communities which have seen their social fabric disrupted may not longer be able to protect the customary interests in land for all or some members of the commu- nities. For example, the HIV/AIDS pandemic has in some African countries destroyed tra- ditional systems that used to protect the customary land rights of widows and orphans. Widows who lost their husbands through HIV/AIDS and whose access rights to land used to be protected through wife inheritance by their in-laws, get deprived of their land as their in-laws can no longer inherit them for fear of contracting HIV/AIDS. Often, after the death of their husbands, the widowed women are chased away by their in-laws with the aim of taking their family land. Customary tenure systems are not adequate to protect the land rights of these widows. The existing formal systems do not provide protection either. It is for this reason that Uganda, in 2004, amended its 1998 Land Act to require consent by spouses and children before family land could be transferred. Even where custom can provide adequate protection from neighbors and local author- ities, it cannot provide it against incursions from outsiders, who may be claiming land awarded by government officials under national law. Land titling, misused, can be a tool for land grabbing, by which official and commercial elites from national level appropriate land from rural communities. The root of the insecurity of rural landholders lies in the fact that much of the land they hold is considered state-owned land, and national government does not recognize rights under customary tenure. Rural people cannot be left in legal limbo, exposed to such incursions, and national governments need to find more ways other than titling, most cost-effective ways, to secure their tenure. Several countries in the region have in the last decade sought to deal with this issue. In Uganda, government has addressed this by legislating the broad recognition of customary land rights, providing them with the same status as private property rights under freehold ownership and the same protection under national law. In Mozambique and Tanzania, national land laws have recognized existing landholding under customary systems, not by reference to custom but to the good-faith occupation it has enjoyed. In these two cases, land continues to be owned by the state, but customary holdings are given full legal pro- tection as usufruct rights. In a number of countries, such as Ghana and Nigeria, custom- ary rights enjoy strong recognition under common law judicial decisions, and in Ghana, customary rights can be registered under statute. Even in countries where customary land tenure is recognized by the state, there are increasing cases of landholders seeking additional protection of their land rights, strangely enough, against the state. As many governments in Africa are notorious for not paying full and fair compensation when they compulsorily acquire land in the public interest, the cus- tomary landholders seek documentary evidence, such as a title deed, to increase their nego- tiating power against the state in the event they are evicted by the state. A second and supplementary approach is to provide a certification of land rights (less elaborate and less expensive than land registration foregoing for instance the cadastral sur- vey required for title registration). This is already being tried out in Tanzania and Uganda. In Madagascar, a new national land policy calls for the creation of land certificates pro- vided by the commune, combining simpler certification of land rights with a decentralized administration of the new system. Decentralization will be a key element in such cost- effective approaches, and yet it should be approached with caution. In the Uganda 1998 Land Act, an ambitious district-level staff was specified that could not be funded in the end, requiring the amendment of the law and resulting in some years delay in implementation. Consensus, Confusion and Controversy 15 A third and supplementary approach is the provision of a title to the community or village, leaving individual rights within the community or village to be taken care of by the community or village itself. This approach is cost-effective as it requires surveying the boundaries of community land only, and not the land boundaries of individual landhold- ings. The Uganda 1998 Land Act and the Tanzania 1999 Village Land Act provide for the issuing of village titles along these lines. What this implies is a two-track approach to land tenure reform. In some areas, where the demand for change has outpaced the evolution of customary tenure systems, there is a need to essentially replace those systems. Systematic titling is an effective way to approach this problem. Systematic titling adjudicates, surveys and registers land in mass, area by area, and thus enjoys, economies of scale. Moreover, transparency can be achieved through com- munity participation. Yet, equal attention must be given to other areas, through recogni- tion and protection under national law of customary land rights and landholdings, and the provision to rural people of more user-friendly and affordable options for recording and certification of those land rights. That attention will need to include not only support, but reform initiatives, especially if customary rules infringe on constitutional or other funda- mental rights, but more generally as well, to increase transparency and accountability in decisionmaking under these systems and to assist their evolution in accommodating the increasingly market-driven economies in which they exist. This implies that it is imperative that a number of approaches to securing land rights be maintained, perhaps for a generation or more. At independence, "unification of land tenure" was often a stated aspiration of new governments. It is still a worthwhile long-term objective, but in the short and intermediate run, the emphasis must to be on responding adequately to the different needs of different landholders for security of tenure. The Consensus The definition of property rights is, as illustrated above, a definition of social relations. At the end of the day, this definition is the business of the community (or the state) and there- fore a political process. For instance, a community may have very definite notions about fairness, and this will--and should--influence profoundly the decision about property rights regimes and the distribution of rights within that system. Summarizing the policy consensus on property rights in land: I Property rights are rules that govern relations between individuals with respect to land, and they should therefore be defined by the polity--the community or the state--to which such individuals belong. I Propertyrightsneedtobeclearlydefined,well-understood,andacceptedbythose who have to abide by them--and strictly enforced. I Propertyrightsneednotalwaysconferfull"ownership"andbeindividual--theycan, and should, be individual, common or public, depending on the circumstances. I Most important for sustainable development is that property rights are deemed secure. As economic conditions permit, property rights tend to become more individualized and formalized. The causality does not automatically run the other way: introducing formal 16 World Bank Working Paper and individualized property rights in situations where economic conditions do not make such rights efficient can be a waste of resources. The consensus is that economic conditions and institutions are defined recursively, with causality going both ways. Finally, the emergence of property rights regimes is the result of political or social processes, and not just a matter of relative prices and transactions costs. This implies that political processes within communities and states are not, by definition, inclusive, demo- cratic, and fair. As the next chapter of this paper demonstrates, the emergence, change, and destruction of particular property rights regimes is often profoundly influenced by these politics--the politics of power relations. CHAPTER 2 Land Redistribution "I was born at the foot of anthills that cast suspicion on the silent threat of mountains. I grew up in a mining town in Namaqualand. My parents grew up in so called colored reserves. Their parents lived on missionary land. These were places with rules that men with see through skin and black holy books defined punitively--blacks will not own land. Here, generation after gen- eration, we lived conditional lives, understanding how profoundly the development of mines and reserves and missionary stations sanctioned our castration from the land and ourselves. So we misplaced ourselves. These places made grown men faceless, weak, and angry. They made grown women dependently worn out, and left questioning children harshly chastised to silence the accusing presence in the hundred and ten half human trees that covered the hills. I was told that these trees tell the story of the Nama people. Who before the arrival of waist high poles and wire moved between shadows of seasons and open spaces, to honour the needs of the goats, the gods and the land. Namaqualand. The land of the Namaqua's. I say it often. Sometimes being is as simple as knowing the womb is not a place to overstay one's welcome. Sometimes it is as com- plex as knowing where your umbilical cord was buried, and what that means to be a part of . . ." T hese are the complex feelings of Elspeth, a contemporary South African woman, working for an NGO in South Africa15 which focuses on the land issue. Elspeth is not alone. James Gibson (2001) surveyed 3,700 South Africans and asked them about their perspective on the land issue. He first tried to assess how important the issue was com- pared to other issues. Not surprisingly, 89 percent of respondents rated unemployment as a very important issue, closely followed by poverty (86 percent). Some 57 percent judged the land issue to be very important, comparable to racism and discrimination (59 percent), racial reconciliation (56 percent), and pollution (55 percent). Gibson then asked his sam- ple to agree or disagree with the following statement: "Most land in South Africa was taken unfairly by white settlers, and they therefore have no right to the land today." An astonish- 15. Surplus Peoples' Project (SPP) in the Northern and Western Cape Provinces. 17 18 World Bank Working Paper ing 85 percent of the black respondents agreed with that statement. Two-thirds of blacks agreed that "land must be returned to blacks in South Africa, no matter what the consequences are for the current owners and for political stability in the country." We start with the above quote and survey results to emphasize that land redistribution is foremost a matter of fairness and equity. It will almost always quickly take us back to his- tory when people feel that the way that property rights were established earlier was not fair. A history in which farmers were dispossessed of their land often corresponds in people's minds to a grave injustice--a wrong to be righted, no matter what. Unfortunately, it is exactly this link to feelings of injustice which makes land redistrib- ution in many countries such an urgent development issue, on the one hand, and too polit- ical, sensitive, and controversial to be dealt with as a part of the economic development and poverty reduction strategies by governments and development partners alike, on the other. It does not help that even among those who are essentially in favor of land redistribution, there does not exist consensus on the "how to do it" part. This confuses policymakers and the development community at large, providing another excuse for inaction, and avoid- ing the heart of the matter--the actual redistribution of property rights in land. This chapter is organized in several sections. First, we restate the case for land redistri- bution, based on equity and efficiency grounds. Second, we describe the lessons learned with respect to land redistribution policies, with specific reference to South Africa, and in the process define a policy framework for the third section, in which we suggest a set of specific policy recommendations. We conclude by stressing that there is a heightened sense of urgency on the need to address land redistribution in South Africa. This sense of urgency should lead the stakeholders to now agree on an overall policy framework for redistributive land reform within which the competing paradigms can actually compete there where it matters: on the ground. The Case for Land Redistribution In countries with a highly unequal distribution of land, the case for land redistribution-- redistributing property rights from the rich to the poor, from large to small farmers--is a strong one, both from a theoretical and empirical perspective. It rests on strong arguments about conflict prevention, equity, economic growth, jobs, and poverty reduction. In spite of this, considerable controversy exists about land redistribution. The roots of this contro- versy are to be found in ideology, politics, history, economic theory, and various efficiency and implementation arguments--a daunting list. A list of the most frequently cited issues that describe the controversy with respect to land redistribution follows. Efficiency Controversy exists about what economists call the "large versus small farms" or the "farm size-productivity" debate. Nearly a century of research by agricultural economists, in par- ticular, all over the world16 has produced a counter-intuitive stylized fact: generally, family 16. The debate started with Chayanov in 1918, when he opposed Stalin's "factories-in-the-field" strat- egy. His critique was based on his PhD research (1910) and that of other "neo-populist social agrono- mists," which empirically documented the efficiency of peasant family farming (Chayanov 1966). Stalin quite literally eliminated Chayanov, his work, and millions of peasant farmers. Consensus, Confusion and Controversy 19 farmers use resources--land, labor, and capital--more efficiently than large, commercial farmerswhodependprimarilyonhiredlabor.Thisstylizedfactisknownasthe"inversefarm size-productivity relationship." It implies that agriculture is generally characterized by dis- economies of scale, which means that redistributing land from large to small farmers can bringefficiencygainstothe(local)economy.Thisoftencomesasashocktothosewhoequate efficiency with the visible signs of modernized, highly mechanized farms which achieve very high crop yields.17 However, the economists' notion of higher efficiency of family farmers and disec- onomies of scale does not equate with higher yields. It does not mean that small farmers have higher yields. Yields are quantities, not values. For example, yields can be raised enor- mously by applying lots of fertilizers and pesticides, but that does not necessarily mean that a profit will be made. In other words, achieving high yields can be inefficient. If high yields are achieved through state subsidies, the economist will still call this inefficient. In prac- tice, large farmers often achieve higher yields than small farmers on the land they actually cultivate. However, at the same time, large commercial farmers often use only a small frac- tion of their land for crops, leaving much arable land to pastures and forests, which pro- vide lower values of output per hectare than the crops of family farmers. The underuse of the land is the most visible sign of large farmers' inefficiency. Not visible, but consistently showing up in the research results, are the higher profits (in kind or in cash) for every unit of investment (either in kind or in cash)18 on small farms. Note that this does not mean that small farmers are richer than large farmers. Often it sim- ply means that they make relatively more out of the little they have. What causes these results and how do we define a "small" farmer? Farm size per se is not the defining feature of family farmers, but instead it is their primary reliance on fam- ily labor, rather than hired labor. What constitutes a "small" farm will vary considerably because of differences in soil fertility, rain fall distribution, market development, technol- ogy, and the opportunity cost of capital and labor in the economy. For instance, from a profit-perspective, five hundred hectares of semi-arid shrub can be "small" when com- pared to half a hectare of irrigated roses. Physical size is not what matters. The productive capacity controlled by the farmer is. "Small" farmers operate their farm using mainly family labor, and employ capital and machinery that they can afford or hire in rental markets. This is the main cause of the supe- rior efficiency of family farms: the owner of the farm lives on the farm, manages the farm herself, and is aided by other family members. These do not need a lot of supervision to work their farm well, because they care about their own property. Hence, instead of "small" farms, it would be more appropriate to speak of "family-sized" farms. 17. The World Bank, since the publication of its Land Reform Policy Paper in 1975, favors sup- porting small-scale, family farming rather than large-scale or plantation-type farming (World Bank 1975). Today, in many countries in the world, direct World Bank support often only goes to the small- scale farm sector. 18. These findings are well-documented in the research literature. For instance, Rosenzweig and Binswanger (1993) show that small farmers get much higher rates of return to capital than large farmers. They also show, that despite being more efficient than large farmers, small farmers do not have profit max- imizing portfolios, because they face significant risk, and have to use the sale of assets, such as draft ani- mals, to deal with the consequences of risk. If small farmers could have profit maximizing portfolios, their profits would increase by another 25 percent. 20 World Bank Working Paper If the opportunity cost of labor in the rest of the economy is very low (the so-called "reservation wage"), calling the efficiency of family labor "superior" is little consolation. Put simply, if poverty is everywhere, and all there is left to do is to try to eek a living out of a small plot of land, the higher profits per hectare of the family farmer that show up in the researcher's books do nothing to reduce her misery. Again, efficiency and income levels are different concepts, and efficiency can be associated with hopeless exploitation by one's self19 or ruthless exploitation by others. So "small" is not necessarily "beautiful." The efficiency of the family farm also does not imply that the average farm size does not increase as farmers' incomes go up. As the history of the developed countries has shown over the last 50 years, farm size grows as a consequence of rural-urban migration caused by rising urban wages. Migration allows the remaining farmers to then earn higher incomes.Tomanagethelargerareas,thefarmerschangetocropsrequiringlowerlaborinten- sities, or they mechanize their operations. Yet, this increase in average farm size is not, it should be emphasized, due to economies of scale. Current U.S. agriculture, for instance, is characterized by diseconomies of scale. At the same time, though, average farm size con- tinues to rise due to the overall increases in incomes in the U.S. economy (Peterson 1997, Kislev and Peterson 1982). Under economies of scale, each additional unit of land brought into production can be cultivated at less cost than the previous one. Under economies of scale, marginal costs, and thus average costs, go down, and thus profits per hectare and the rates of return go up. Such economies of scale in agriculture are limited. They are limited to those crops where economies of scale exist either in marketing or where large farmers have better access to finance and information about technology and markets. Such advantages then spill over into production, but they are not caused in the production process itself. If farms increase in size (or in capital stock) beyond a size that a family can comfortably manage itself, more hired labor is needed. It is the increased cost of supervision of the hired labor--the transactions costs--which is the source of the economic inefficiency.20 Not sur- prisingly the most successful agricultural systems in the world, such as China, Thailand, and Costa Rica, are largely run by family farmers. By acknowledging that small farmers often use their resources--however meager they may be--better than their larger counterparts does not mean that there are no disadvan- tages to being small. The main disadvantages of small farmers lie in more difficult access to credit, markets, and information--especially information about new markets and tech- nologies. Public and private agricultural research has often been biased toward developing technologies suitable to large farms, given their lobbying power and financial wherewithal. Larger farmers usually have easier access to cheaper credit. This enables them to quickly respond to the market, especially when the market demands agricultural products with high investment costs, such as for example, horticultural products. Small farmers are also at a disadvantage when the market demands large quantities of standard quality be produced at exactly the right moment. Coordinating such production may be easier to organize on a large farm, even if it means managing a large labor force. This applies to many 19. Chayanov therefore called this the "self-exploitation of labor." Others have called it "under- consumption" (Kautsky) and "the plunder of labor" (Lenin). 20. In the literature, the difficulty is defined as caused by heterogeneity, seasonality, and the resulting problem of asymmetric information. Consensus, Confusion and Controversy 21 of the "plantation crops," such as bananas, sugar, and tea. Hence, there do exist situations were medium and large farms are more productive than small farms. Because large estates usually use less labor per hectare (or per unit of output) than fam- ily farms, they generate less employment per hectare (or per unit of output) for the econ- omy as a whole. This is a consequence of their low land-use intensity, where only a small fraction of the landholding is under crop cultivation. While they may utilize modern tech- niques and inputs, and achieve superior yields on the land they crop, their overall land use and employment intensities are usually low. Large farmers are often well-organized and well-connected, and are able to lobby gov- ernments for special tax breaks, subsidies, and other special distortions. The consequence of these distortions is invariably that they face lower effective capital costs relative to labor costs, and therefore overinvest in more machines that replace labor than they would have had they notbeenabletoobtainthetaxbreaks,subsidiesandcheapcredit.Thiscanmeanthatanagri- cultural economy based on large farms becomes socially inefficient. Even as production in the large farm sector rises, its contribution to employment may actually go down, while high unemployment in the overall economy persists or becomes even more severe. Equity, Growth and Poverty Reduction One of the most compelling reasons to support an agrarian structure based on smaller fam- ily farmers rather than large commercial farms comes from the body of international expe- rience on what it takes to achieve high growth and reduce poverty. Does a consensus emerge from the international experience? Recent research comparing many countries with each other suggests that equity, in gen- eral, is good for growth (Aghion, Caroli, and Garcia-Peñalosa 1999). In particular, equity in land distribution is also associated with overall higher economic growth (Deininger and Squire 1998, Deininger and Olinto 2000). Country case studies confirm this hypothesis. For instance, the initial phase of China's high and sustained growth and poverty reduction spurt was clearly linked to its change from collective large-scale farms to small family farms in 1979 (Ravallion and Chen 2004). Once collective production was abandoned and key agricultural markets were liberalized, China's peasant sector provided the initial engine of the rapid eco- nomic growth which dramatically reduced poverty. Recent research21 confirms this thesis in many other contexts and settings--namely, that equity is good for growth. More equitable land distribution is also beneficial for non-agricultural and non-rural growth.Accesstolandprovidesagoodsocialsafetynet,whichinducesmorefarmerstomove into non-farm businesses, given the higher risks associated with entrepreneurship. In China, the broad-based access to land allowed peasants to take increased risk, move into non-farm activities, and then produce the boom in small-scale entrepreneurs. This also explains why China spends significantly fewer fiscal resources on social security-type transfers than, say, India, where the poor have much more restricted access to land (Deininger 2003). OthersuccessstoriesarefoundinCostaRica,Indonesia,Malaysia,Taiwan,andThailand. These countries' agricultural sectors are all predominantly based on owner-operated, 21. For instance, Deininger and Squire (1998) find that only 2 of the 15 developing countries with a Gini coefficient for land higher than 0.7 managed to grow at more than 2.5 percent per year during 1960­92. 22 World Bank Working Paper small-scale family farms.22 When these countries also made substantial investments in rural infrastructure to help small-scale farmers, and had no or light taxation of agricultural pro- duction (this included avoiding over-valued exchange rates), they created the type of high and sustained agricultural growth which substantially reduced rural poverty. This is because family farmers spend more of their income on locally-produced goods and ser- vices than large farms, creating a positive relation between family farms and non-farm incomes in the local economy. An illustration of the point for the United States is found in a seminal study of two small towns, first published by the U.S. Senate in 1946. Walter Goldschmidt (1947) compared two California towns, Arvin and Dinuba, which were selected because they were virtually alike in all basic economic factors except farm size. While the total volume of agricultural pro- duction was about the same, the "small family farm" community supported about 20 per- cent more people at significantly higher living standards, had twice as many businesses doing 61 percent more retail business, and boasted a substantially higher level of public infrastructure than the "large corporate farm" town. When Goldschmidt proceeded to repli- cate his findings nationwide, his research was stopped due to political pressure from the large farm companies. However, his landmark study would later spawn a large literature, confirming his hypothesis that family farms were better for rural poverty reduction than large-scale corporate farming. The history of many land redistribution programs demonstrates that once poor peo- ple are given good farm land they can lift themselves out of poverty permanently, even without significant government support. In Africa, this has been shown to be true in the case of both Kenya and Zimbabwe. Even without substantial support services, Zimbabwe's land reform program of the early 1980s was deemed a success. Longitudinal household data sets on land reform beneficiaries show that after about 10 years "land reform beneficiaries cultivate nearly 50 per cent more land than non-beneficiaries, obtain four times as much in crop revenues, own substantially more livestock, and have expenditures that are higher by 50 per cent" (Hoogeveen and Kinsey 2001). Without a doubt, providing more support services would have sped up the process of establishing successful small farms. At the other end of the spectrum, one finds the countries that have been least success- ful in terms of rural poverty reduction. These include Brazil, Colombia, Guatemala, and South Africa. Not surprisingly, these countries are characterized by highly unequal landownership, with substantial public investments in large-scale farming. While these large-scale farms have usually become technically sophisticated, they make economically inefficient use of land and labor, and lead to rapid out-migration of labor from the agricul- tural sector into urban or rural slums. In short, by focusing too much on their large-scale farms, these countries created more rural (and urban) poverty. Deininger (2003) and the World Bank's World Development Report (2003) provide further compelling evidence of the long-term implications of inequality in landholdings and development. By tracing individual countries' long-term development paths within sets of comparable countries (Columbia, Costa Rica, Guatemala, El Salvador; Indonesia, the Philippines, and Thailand; states within India; and North and South America) they fur- 22. In the cases of Taiwan and China, the smallholder-based agrarian structures were created by land reform--transforming tenants into owners. In the case of Thailand, nineteenth-century legislation had set a four-hectare limit on freely acquirable agricultural land, constraining the emergence of large estates and feudal tenancy relations. Consensus, Confusion and Controversy 23 ther illustrate how initial inequality in landholdings leads to dramatically different develop- ment outcomes in the long run. Deininger (2003), and Acemoglu, Johnson, and Robinson (2001 and 2002) use cross-country time series to show the same "path dependent" develop- mentpattern:countrieswithamoreegalitariandistributionoflandtendtohavebetter,more inclusive institutions which in turn lead to higher levels of economic growth. Fairness Finally, maybe the most important reason to worry about equity is linked to the inherent political and social nature of property rights. History and culture, and many other factors, can mould what a community or a nation thinks is a fair use and ownership of land. As his- tory shows, communities may even change their views on what is appropriate and fair. Societies usually have strong feelings about how and by whom land should be used, becausetheoverallareaoflandinacountryisfixed,andagricultureis(orcouldbe)animpor- tant source of income for many people. Notions of fairness, or equity, are often very pro- nounced when it comes to land: there often is a general feeling that land should be equitably distributed to as many people as possible. A countryside populated by small family farmers tilling the land corresponds in many peoples' minds to a system that is fair and equitable. Unresolved land issues lead to violence, civil unrest, or even civil war, and demonstrate most effectively how strong these notions of fairness are. Land-related conflicts have plagued countries such as Algeria, Brazil, Colombia, El Salvador, Honduras, the Philippines, and many others. In Africa, the establishment of the settler colonies in Kenya, Namibia, South Africa, and Zimbabwe were all accompanied by fierce resistance from the displaced indige- nouspeople.Countriesthatarebesetbyunresolvedlandproblemsaredoomedtoprotracted periods of economic instability. Extremely unequal distribution of property rights in land, and water has been the legacy of white settler colonies. These settlers appropriated for themselves the best pieces of land, either for livestock or crop production. They then turned the indigenous black peasants into tenantsorwagelaborers,orsimplyexpelledthem(Wolpe1972,Bundy1979).23Whenmech- anization (subsidized by the state) made it feasible to depopulate the land, black tenants and wage laborers were removed at an even faster rate, and driven away from their homes into marginal areas, designated "homelands" or "communal areas." Apart from the fact that--by design--the homelands' resource base of these home- lands was poor, the manner in which property rights were dealt with further contributed to the break in their development.24 In the Southern Africa countries, "communal areas" or "homelands" are, in theory, governed by "traditional" or "communal" property rights regimes, but in fact, land is state-owned and subject to a particular colonial interpretation of "traditional" tenure. These interpretations have also become quite static, and sometimes lead to the insecurity of property rights, thereby undermining the development of land sales and rental markets. If these regimes had been allowed to change according to the 23. Similarly systematic expropriations, and sometimes outright exterminations, of indigenous peo- ple took place in many other parts of the world. North and South America, Australia and Tasmania all suffered the tragic consequences of settler actions, often justified under variants of Herbert Spencer's phi- losophy of "Social Darwinism." Many of these countries are wrestling with the aftermath of these human tragedies until today. 24. Many other brakes were put on development in the homelands. For instance, farmers were not allowed to produce certain crops and had to market through monopolistic marketing boards. 24 World Bank Working Paper needs of the communities, they would have probably slowly evolved toward land use based on private property where intensification was encouraged and toward forms of common property where private property did not make economic sense, as in very dry areas only suitable for livestock production.25 The removal of black peasants from their land was very systematic.26 Today, the most fertile lands in Southern Africa are occupied by very large, sprawling farms which are, on average, underused. The highest population densities--black population densities--are found in the most infertile rural areas and often close to natural parks. This is what is called the "rural geography of apartheid," brought about by economic policies that have favored the settlers and the forced removal of black people from fertile lands over a period of over a century.27 This inefficient geography continues to impose tremendous costs on the poor and the economy as a whole, but it is also highly inequitable and unfair. Because the legacy of the removal of black people from their land is often still fresh in peoples' minds, land reform is a highly emotive issue. Communities and nations will have to deal with this legacy. They will invariably form opinions about what is fair. They may simply look at the land issue as one of justice and one of redressing old wrongs. This is as it should be. People should reflect on the existing property rights and democratically make decisions about their distribution, because, as his- tory shows, ignoring a looming land conflict, is a very risky economic strategy indeed. In Southern Africa, like elsewhere, restoring a more equitable distribution of land will greatly contribute to more social cohesion, which will foster more inclusive institutions and poli- cies, and hence better long-term development. Land Redistribution Policies: Lessons Learned Preventing a looming land conflict means undertaking land redistribution on a substantial scale and at a rapid pace. Unfortunately, international experience suggests that substantial 25. Individualization is driven by the intensification of agriculture caused by population growth and increased market access (Boserup 1981, Bruce and Migot-Adholla 1993). Typically, communities start by individualizing land into permanent residential and garden plots, then allocate individual rights to nearby fertile farming plots, and progressively extend individualization to the remaining areas under community ownership, until only wasteland and land for common infrastructure and facilities is owned by the community (Binswanger and Rosenzweig 1986, Binswanger and McIntire 1987). Many common property regimes will allow individual usufruct rights to a specific plot to become more permanent, often as a direct result of investments in the land. In other words, the security of individual property rights is created by the act of investing, rather than the other way around. 26. One of the transitory phases in this removal was the phenomenon of "labor tenancy" under which African peasants--the scarce labor factor--were allowed to reside, farm a plot, and graze cattle on the set- tler's land in return for a certain number of days work. Labor tenancy occurs when labor, especially dur- ing seasonal peaks, is scarce, supervision capacity of the landowner is limited, and landowners have abundant land (Lastarria-Cornhiel and Melmed-Sanjak 1998). In contrast, in Asia--where land is the scarce factor--share-cropping in the more common form of tenancy. 27. The basic economic idea behind the "settler economy" is the following: capital and good land are owned by the settlers, while the cost of unskilled, indigenous laborers is reduced by restricting their eco- nomic alternatives and creating a migrant labor system. Under the migrant labor system, only male adults are allowed to work and reside in the settler areas. The wage paid to the migrant laborer can now be below the amount he would need if his family were living with him, while his "reservation wage" is reduced because of the poor agricultural profitability in the homelands, where his family resides. Consensus, Confusion and Controversy 25 land redistributions are most often done in periods of upheaval and political violence. Some observers even go so far as to draw the conclusion that there is no such thing as a peaceful and "orderly" land reform. The Zimbabwean "fast track" land reform of the early 2000s, under which the Government of Zimbabwe explicitly and publicly decided to set the con- stitution and the law aside, seems to confirm this stylized fact. To make such a historical precedent into a policy recommendation is not only false logic, it is also dangerous. If a society has the choice between doing land reform with or without violence and economic destruction, surely society would chose the "without" option. This then is the challenge: can land reform be implemented at scale, peacefully and successfully? The starting point is to analyse past failures and derive lessons. The following sections will argue that there are three main reasons for the lack of success of "peaceful" land reform: (i) land markets need help; (ii) land acquisition and resettlement can be slow and costly; and (iii) large farmers will lobby against land reform. However, we will start with a brief background on the opportunities and challenges of land reform in South Africa. Land Reform in South Africa Land and agricultural policy reform in South Africa holds the promise of increasing effi- ciency, equity and generating jobs. South Africa's farms also confirm the international evidence that size matters. Within the commercial, formerly "white" farm areas, smaller farms have consistently higher prof- its and employ far more labor per hectare than large farms (Christodoulou and Vink 1990; van Zyl, Binswanger, and Thirtle 1995). It would be unfair, and because of the general lack of data on black farming virtually impossible, to compare the formerly white farming areas with the formerly black areas (the so-called homelands), because of the centuries of sup- pression of black farming. However, there do exist case studies in the tea and sugar-cane industry which compare small black farmers benefiting from support services under con- tract farming to their large-scale counterparts. These case studies confirm the higher effi- ciency of the small farms. Moreover, in dry-land cotton, small black farmers were more efficient than white farmers, even under apartheid (Wheeler and Ortman 1990). Livestock production is characterized by the same diseconomies of scale, associated with farm size and the need to hire additional labor. It is, of course, true that arid lands are not very suitable for crop production, and that lumpy investments into wells and pumps will tend to favor larger units. Yet, this should not lead us to thinking that, therefore, land redistribution in dry areas (most parts of Southern Africa) is impossible. "Dryness" does not reverse the inverse farm size-productivity relationship, which is based on labor super- vision costs, not on rainfall. Furthermore, land reform in dry areas does not imply a whole- sale switch from livestock to crop farming. Arid and semi-arid lands are excellent livestock production areas (for example, Texas), because of the lesser risk of disease compared to more humid areas. After land reform, this comparative advantage remains, but production costs will go down due to the lower labor supervision costs on the new family farms. In addition, breaking up large-scale, fenced-in ranches would improve efficiency in several other ways. First, in semi-arid and arid areas, rainfall variability (and hence the availability of water and fodder) is very high, putting a premium on flexibility. The benefits of flexibility increase 26 World Bank Working Paper with rainfall variability (van den Brink, Bromley and Chavas 1995). Given highly variable rainfall, fenced-in areas will never be large enough, forcing owners to either move or sell their herd during severe droughts at significant cost. These costs are sometimes transferred to the surrounding areas: common property is turned into open access, or to the state: spe- cial subsidies are then accorded to prevent livestock prices from plummeting. Getting rid of fences, and organizing more flexible grazing systems in other ways, would increase the effi- ciency of production. Second, the suppression of bushfires to protect the costly ranching infrastructure leads to so-called "bush encroachment," thereby reducing the area under pasture.28 As Figure 2.1 shows, following this pattern, Namibia's commercial cattle herd has shrunk by 70 percent over 40 years. This is clearly not an efficient production system. If the lost grazing land were now to be recovered, the bush-encroached areas would need to be manually or mechani- cally de-stumped. Land redistribution from large to small farmers would make available the extra labor required to de-stump the areas affected by bush encroachment. Figure 2.1. Namibia: Cattle Numbers in Commercial Ranch Areas (1958­2000) 3,000,000 2,500,000 2,000,000 numbers 1,500,000 1,000,000 Cattle 500,000 0 1958 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 Year Third, at higher population densities, the depressions, where good soils and water accumulate, will be cultivated with crops, resulting in the very beneficial interaction between crops and livestock (through fodder and manure). In summary, even though the empirical evidence for the higher efficiency of family farms in South Africa remains scarce, the existing data confirm the international evidence. There is a case for land reform on efficiency grounds, including in dry areas. If this is so, then what has the Government of South Africa done about land reform so far? The Government's target is to redistribute at least 30 percent of the land in 15 year and completetherestitutionprocessbytheendof2005.In1993,ajointSouthAfricanandWorld BankteamestimatedthatreachingthelandredistributiontargetwouldcostbetweenR22and R26 billion in total, or about R1.5­1.7 billion per year, and create more than one million rural livelihoods, or the net equivalent of 600,000 full-time farm jobs at about R35,000 per job. We use the term "livelihoods," because family farm communities the world over consist of households which obtain only part of their income from farming. For instance, the con- 28. The suppression of fire has had similar negative results for the ecology of the grasslands of the prairie of North America (Licht 1997). Consensus, Confusion and Controversy 27 tribution from farming to the average farm household in the United States is only 11 per- cent, although this contribution rises to 60 percent on large family farms.29 Themaineconomicimpactofawell-executedlandreformprogramwouldthereforenot only come from a more intensive use of agricultural land, but perhaps, more importantly, from the multiple livelihoods created by a more dynamic local peri-urban and rural econ- omy, based on a substantial increase in the number of small family farms. In the short term, creating sustainable "pluri-activity" households with only a small portion (say up to 25 per- cent) of income coming from farming is especially feasible and attractive in the peri-urban areas, where there is a dearth of small-scale agricultural production for the informal urban markets nearby.30 In the medium term, stimulating pluri-activity households at higher incomelevelsandwithahighercontributionfromfarmingcouldalsobeachievedintherural areas. However, much more will be needed here in terms of support services and rural infra- structure investments to stimulate farm and non-farm incomes of rural households. These investments will need to be undertaken as part of the integrated local development plans and fiscal transfer systems underpinning decentralized development in South Africa. However, current trends are alarming. In 1996, nearly half (46 percent) of South Africa's population of 40.6 million people lived in the rural areas (Census, 1996), where 70 percent of the poor live. In spite of the dramatic political, economic and social reforms that have taken place, rural areas seem to have benefited less than the urban areas from the pol- icy changes introduced after 1994. When changes in household expenditure, poverty and inequality between 1995 and 2000 were examined,31 the following trends emerged: slow con- sumption growth (less than 1 percent per capita annually); no change in the overall poverty headcount; while the poverty gap, the severity of poverty, and overall inequality increased. The data show a deterioration of real expenditures at the bottom end of the distribution and an improvement at the top. Because most of the poor live in rural areas, this implies that the rural black population is impoverishing, both in absolute and in relative terms. What are the trends in agriculture? Parts of commercial agriculture responded well to the devaluation of the currency and trade liberalization, exploiting the new opportunities for South African products abroad. Agriculture now contributes 4.5 percent to exports-- a share which has been rapidly growing since the liberalization process started in the late 1980s. Yet, the increased export orientation was not matched by increased labor- intensity of production in the sector as a whole. Employment in commercial agricul- ture declined, from about 1.1 million in 1995 to 0.9 million in 2003, or about 10 percent of total employment.32 It is now estimated that about 1 million farm dwellers, or about 29. These data come from the website of the United States Department of Agriculture: http://www. ers.usda.gov/Briefing/FarmIncome/forenew.htm. In 2003, a large family farm was defined as a farm with farm sales between US$250,000 and US$499,999. 30. The absence of such high intensity small-scale farming "rings" around all of South Africa's cities is the direct result of apartheid, in the past, and the continued restrictions on subdivision and the absence of a land tax (leading to unused peri-urban land for speculative reasons), in the present. These distortions are discussed below. 31. Using comparable consumption aggregates from the Income and Expenditure Surveys (IES). 32. Oneexplanationofthistrendisasfollows:inthecommercialfarmareas(86percentofthetotalarea) the legacy of apartheid often strains labor relations. Expansion of agriculture means expansion of the labor force, and this is accompanied by increased supervision problems--not a preferred option for many white farmers. The commercial farmers' expectations that the post-1994 Government would provide increased protection against the eviction of labor tenants and farm workers often resulted in their pre-emptive expul- sion. These expectations proved to be true, and while exact numbers are not available, anecdotal evidence suggests that the eviction of labor tenants and farm workers has been quite dramatic. 28 World Bank Working Paper 200,000 households, were evicted between 1994 and 2004, continuing the trend established under apartheid.33 This trend needs to be reversed, given the imperative to reduce overall unemploy- ment in South Africa, currently measured at about 30 percent. Laid off and evicted farm workers would be an important target group for South Africa's land reform program. Even some of the unemployed youth in urban areas without any farm experience will find it worth their while to join the beneficiaries of land reform and work on these new farms. Yes, a job in town is much more desirable to them, but if there are no jobs, work- ing on a farm may be better than permanent unemployment in South Africa's sprawling squatter camps. In the former homelands, where about 12.7 million people--or 31.4 percent of South Africa's population--live,34 subsistence farming bucked the general jobs trend, and added a respectable 0.4 million livelihoods between 1995 and 2003.35 This is all the more remark- able because of the limited potential for agriculture, low consumption growth, the lack of investment (and maintenance of existing investments) in irrigation infrastructure, and poor agricultural support services. Much more can be done to promote farming in these areas, while land reform beneficiaries could also be drawn from here. What has the Government achieved so far on land reform? Even though the first law to be enacted by the new Government in 1994 was land-related (the Restitution of Lands Right Act),theGovernment'slandreformprogramswereofftoaslowstart.Between1994and2003, only about 3 percent of total agricultural land in South Africa, while the stated target is 30 per- cent of the land or about 25 million ha in 15 years. The land reform strategy has three main programs: (i) restitution of land to the victims of forced removals; (ii) land tenure reform to promotesecurityoftenureforall;and(iii)redistributionoflandtohistoricallydisadvantaged, landless people. Fortunately, after some key improvements were made to the restitution and redistribution programs, the delivery rate is currently accelerating. Hence, the challenges facing the Ministry of Agriculture and Land Affairs are daunt- ing. To attain the overall goals of ensuring peace and stability in rural areas, reducing rural poverty and increasing black empowerment in agriculture, the following objectives need to be achieved. First, agricultural and other policy reforms need to continue to increase the labor- intensity of the sector. Second, the pace of delivery of the Ministry's main programs needs to pick up substantially. This includes setting as a priority the restitution program which will need to come to closure by December 2007, by which time all outstanding claims need to be settled. It is of key importance from the perspective of justice, but, if unresolved, it threatens to undermine investment incentives in the farms affected. It also includes increasing the pace of the land redistribution program, because most of the target of 30 percent of land redis- tribution will have to come from the Government's flagship land redistribution program Land Redistribution for Agricultural Development (LRAD) and a better functioning land market, which continues to be constrained by a number of key distortions. Tenure reform 33. Briefing to Parliamentary Portfolio Committee for Agriculture and Land Affairs by the Nkuzi Development Association and Social Surveys, August 30, 2005. The estimate is a national extrapolation based on a random sample of 300 communities and 7,759 households. 34. Stats SA, Rural Survey 1997. 35. October Household Survey/Labor Force Survey. Consensus, Confusion and Controversy 29 needs to be fully implemented, especially to provide broad-based, democratic and secure access to land in the communal areas. To ensure improved livelihoods and productive and sustainable land use on all transferred land, beneficiaries need to be fully empowered with adequate resources and support services through the main agricultural support programs. Finally, the Ministry's monitoring and evaluation systems need to dramatically improve to provide timely feedback on the achievement of outputs and outcomes. In the following, we will elaborate on some of these challenges, with an emphasis on increasing the pace and improving the impact of the redistribution program. We follow the main lessons from international experience: land markets need help, land acquisition and resettlement can be slow and costly, land reform is usually under-funded, and large farmers will lobby against land reform. Land Markets Need Help What about letting the market solve the issue of land redistribution? Why would one need state intervention? If small farmers are so efficient, why does the market then not auto- matically transfer the land from inefficient to efficient users? Why do small farmers not go onto the land market and outbid large farmers for land--especially in light of the fact that large farmers usually do not even use all of their land? There are number of reasons why the land market, as it is defined in many countries that are characterized by very unequal landholdings, fails to redistribute land from large to small farmers. First of all, the poor do not have money to buy land, and either no, difficult or costly access to credit. Second, agricultural land prices may be high due to many factors: investors may value land not just to farm, but also for its value as insurance, as a hedge against inflation, as a tax shelter, or as a means by which to gain access to subsidized credit or public infrastruc- ture (for example, irrigation works). If subsidies in input and output markets are also biased toward large farmers, they will drive up the land price and increase the wedge between what small and large farmers can afford to pay.36 Hence, even if small farmers had access to credit, they would not be able to repay the credit. Third, in the context of imperfect markets for credit and insurance--a context typical of rural areas--land sales markets may instead lead to distress sales of production assets, such as draft animals or land, by poor small farmers during major droughts and other adverse shocks, creating even more poverty (Binswanger and Rosenzweig 1986). Land rental markets for land could overcome some of these problems, but they often do not result in a redistribution of access from large to small farmers in countries where there is a highly unequal distribution of land to start with (Deininger, Castagnini and González 2004). Non-economic factors may contribute to this lack of transfers (both sales and rentals) between large and small farmers. In Southern Africa, this includes the so-called "Not In My Backyard" (NIMBY) phenomenon--a legacy of apartheid. Better credit markets would overcome some of the imperfections in the land markets, but it is difficult to increase the credit supply to small farmers and the landless. Credit markets are constrained by many factors, including informational asymmetries which are 36. For an analysis of this land price wedge in South Africa, see van Schalkwyk and van Zyl, 1996. 30 World Bank Working Paper difficult to overcome. This is why credit markets usually require collateral, which brings the issue of land back into the equation. Given the political power of large landowners, and the poor track record of agricul- tural credit systems for small farmers, a policy to increase access to credit for small farm- ers and the landless is likely to fail, and might even make matters worse if large landowners end up with more credit, and subsequently even more land. Using legislation to improve the weak economic bargaining position of the landless in land markets is ineffective or counter-productive. For example, India has a long history of legally strengthening the rights of tenants, but in the end, the legislation inadvertently led to the illegal evictions of tenants. Such unintended, adverse effects take place when landowners trytopre-empttenantsandoccupiersfromacquiringstrongerrights.Landownersevicttheir tenants before the new laws are passed, or exploit lax enforcement of these laws once they are in place. In South Africa, the 1997 Extension of Security of Tenure Act (ESTA), the 1996 Labor Tenants Act (LTA), and the 1998 Prevention of Illegal Eviction and Occupation of Land Act attempted to prevent illegal evictions and confer certain property rights in land to farm workers. Their impact on illegal evictions and the security of tenure of farm workers is still unknown. What is clear, however, is that only a few cases have actually been settled under these laws, and that they have done little to stem the secular decline of farm employment on South Africa's commercial farms. In fact, it appears that these laws have contributed to pre-emptive evictions by landowners. The bottom line is that land markets need help, because they cannot be counted on, on their own, to redistribute land from large to small farmers. Current land markets in Southern Africa are in need of more than "help." They need serious reform, because, by historical design, they place severe restrictions on land sales to the poor. Restrictions on Subdividing Land. Poor small-scale farmers can usually only expand their landholdings slowly by re-investing their own profits--incrementally--because of the dearth of finance available to them. They can only self-finance the purchase of land, if land markets make small pieces of land available, on a continuous basis. Unfortunately, land markets in the commercial farm areas of Southern Africa do not function like that. South Africa, for instance, still has explicit legal and policy restrictions against the subdivision of farms into smaller units. The existence of such restrictions should profoundly worry those who believe that large farms are more efficient than small farms. If large farms were more efficient than small farms, why would it be necessary to legally restrict the subdivision of land? This begs the question: where does this restriction on subdivision come from? South Africa's subdivision policy--the Subdivision of Agricultural Land Act, 1970 (Act No. 70 of 1970)--was inspired by the danger of "die verswarting van het platteland"--literally, the "blackening of the country side." The official reason given at the time was that farms should not be allowed to decrease in size below the so-called "viable" size. This begs the next ques- tion: what is a "viable" size? The first thing to realize is that "viability" is not a notion related to production economies of scale. Instead, it is linked to a minimum income target. In former settler colonies, the "viable" size was calculated by setting a minimum income target for white farmers. On the basis of this income target, a simple calculation followed which determined Consensus, Confusion and Controversy 31 the size of the farm.37 Efficiency considerations, such as economies of scale, or employment generation, did not enter the calculation. The viability policy was a social policy which ensured that white farmers earned an income acceptable to white society. If "viability" norms were defended by a settler government on a "white income stan- dards" basis, this social policy objective should have quickly become obsolete at Indepen- dence in Zimbabwe or the end of Apartheid in South Africa. To date, unfortunately, neither Zimbabwe nor South Africa has removed such subdivision restrictions. The result is that the restriction on subdivision functions as a powerful barrier to racial integration in the com- mercial farm areas38 and in the peri-urban areas, where the unavailability of small parcels for sale leads to widespread squatting, or so-called "informal settlements." It makes it dif- ficult for a black person--in Southern Africa, on average, poor--to legally buy a small plot in a formerly white area--simply because no small subdivisions are on offer. The pent-up demand for land spills over into the informal land market where per hectare land prices quickly reach and sometimes overshoot the levels of the formal market. In other words, a policy that had been designed with the sole purpose of ensuring white living standards and segregating the races is still in place in the democratic, non-racial South Africa of today. This policy lacks any economic, let alone social, rationale. It restricts the land market and makes it difficult for small farmers to buy small farms. In order to cre- ate a level playing field between small and large farmers, and poor and rich urban dwellers, these land markets must be reformed. The removal of these subdivision rules is urgently needed. Fortunately, in the con- text of the Government's land reform programs, the Subdivision Act does not apply-- an exemption has been granted. The next step should be to extend the exemption granted to the beneficiaries of government programmes to the land market as a whole. Below, we will suggest a possible explanation for the remarkable failure by the democ- ratic governments in South Africa and Zimbabwe to remove this barrier to broad-based landownership. Distortions in Land and Other Markets Can Drive Up the Price of Land. These dis- tortions need to be removed and/or a level playing field has to be created with respect to them. In many countries, the absence of a land tax (or of a property tax which includes the land wealth of large farms) raises the attractiveness for the rich of holding land as an asset, but does not necessarily entice them to farm it and make full use of it. In South Africa today, the land wealth of large farms is either not taxed at all, or, based on a 1939 law, taxed at a rate 100 times less per hectare than that which applies to small farms (see Table 2.1). This extremely regressive tax produces artificial economies of scale, because land consolidation 37. A similar "viability" logic was followed in the pricing of agricultural products in apartheid South Africa. An income target was set, a preferred (i.e. capital-intensive) technology chosen, and then the state was asked to guarantee the resulting "cost plus" output price. That price would then be protected by not allowing cheaper imports in years of low production, and not allowing the domestic market to clear in years of high production, when arbitrarily declared "surpluses" would be exported at a state-subsidized loss. 38. Currently, a farm, or subdivision of a farm, needs to have the potential to produce a net income (gross margin) of R24,000. This is translated into farming on 60 large stock units, 20 hectares of irriga- tion land or 100 hectares of dry land. These minimum sizes are too high and inconsistent with the Gov- ernment's land reform strategy. 32 World Bank Working Paper Table 2.1. South Africa--Taxes Payable for a 100 Hectare Farm Valued at R400,000 in Four Municipalities Mogale City Madibeng Merafong Mbombela Tariff 7.62% 17.33% 13.00% 24.18% Agricultural Bracket (ha) Rebate Factor Tax Payable (R/year) 1 100% 304.86 693.20 520.00 967.20 4 25% 228.64 519.90 390.00 725.40 20 10% 487.77 1,109.12 832.00 1,547.52 >20 1% 243.88 554.56 416.00 773.76 total 1,265.15 2,876.78 2,158.00 4,013.88 leads to a sizeable reduction in the tax bill. It also makes the cost of holding onto unused or underused land very low and raises the attractiveness of agricultural land as an asset.39 For all these reasons, land prices of large farms often exceed what economists call the present value of farm profits. If the land price exceeds the present agricultural value of the land, small farmers will be unable to outbid the large farmer or repay the loan given to them. Thus, land markets will not redistribute land from large to small farmers. In this sit- uation, a strong economic justification exists for subsidizing land purchases by the poor. Land Acquisition and Resettlement Can Be Slow and Costly There exist different ways by which land can be acquired for redistribution to benefi- ciary farmers: either through compulsory acquisition, direct purchases by beneficiaries ("market-assisted" or "community-driven"), or negotiated transfers. Compulsory Acquisition. Land redistribution has often been slow and costly because many governments choose to redistribute land through legal processes of expropriation or compulsory acquisition. This legal process is rooted in the legal principle of eminent domain: the state's power to take private property for public use, following the payment of just compensation to the owner of that property. In the past forty years, countries such as Colombia and the Philippines accomplished little through this process. It took Mexico around sixty years to redistribute half of its agricultural area using this method. Only Brazil, and only since 1995, has made a considerable dent in its land problem by expropriating about 20 million hectares in seven years. 39. A new rationale to prevent subdivision is environmental. The underuse of arable land, or its con- version into private forests, game farms and nature conservancies, is sometimes seen as promoting the environmentally sustainable use of natural resources. This perspective leads some environmentalists to oppose subdivision and land taxation, as this would provide the owner an incentive to make more prof- itable, which may mean more intensive, use of the land. There are solid arguments in favor of conserva- tion and the sustainable use of natural resources. But in accepting the current unequal distribution of assets, and income, as a given, the new rationale against subdivision and land taxation seems merely to promote the opulence of large landowners as the best strategy to conserve the environment (Daniel W. Bromley, personal communication). Consensus, Confusion and Controversy 33 The legal process of expropriation or compulsory acquisition can be slow and costly. Amending the law to speed up the legal process of compulsory acquisition, can make a dis- cernibledifference,astheexperienceofseveralcountrieshasshown.Between1995and2002, Brazil redistributed some 20 million hectares to about 580,000 families, using expropriation as the main method.40 In seven years, this is twice what was achieved in the previous 30 years. However, there is a limit to speeding up the process. The legal process is based on due process--an important principle of justice--and implies that every farm owner can opt to have his or her day in court. For instance, half of the landowners in Brazil go to court to con- testthevaluationofthefarm,andthisslowstheprocessdownandmakesitmorecostly."Just compensation," moreover, is invariably interpreted by courts as at least as high as the prevailing market prices. By its very nature, then, the legal process is lengthy and costly, adding to the costs of compensation. On the other hand, a country which sets aside principles of legality and due process, as Zimbabwe has recently done, sends shockwaves to anybody wanting to invest in such a country, including its own citizens. It leads to disinvestment, devaluation of the currency, and economic contraction. Further, it creates new wrongs and legal complica- tions that need to be resolved later, prolonging the uncertainty around the land issue. This is essentially the situation that Zimbabwe finds itself in today. The case for compulsory acquisition is usually based on a planning argument.41 For instance, whereas the homelands or communal areas in Southern Africa generally have poor soils, they may have reasonable social infrastructure, such as schools and health facil- ities. There may exist large farms next to these areas which have better soils, but no social infrastructure. Compulsory acquisition of a group of such large farms could then be a smart way of giving poor farmers better land, while their families can continue to benefit from the infrastructure present in the communal areas. Even though the legal process of compulsory land acquisition by the state is inherently slow, the existing legislation can usually be improved upon, while safeguarding the con- stitutional rights of citizens. For instance, in some countries, landowners do not just have the right to challenge the level of compensation--which seems eminently fair--but also the possibility of challenging what the State defines as the public interest--which seems excessively conservative. In other cases, land acquisition procedures are so complicated and open to interpretation that it is virtually impossible for the state to acquire a significant number of farms within a reasonable period of time. Much can be done in many countries to create better legislation which fairly balances the public and the landowners' interest. 40. The Brazilian approach to expropriation is as follows: each district has a specific reference farm size calculated as "15 modules," with one module defined as the "viable" family farm. This allows for a wide variety of reference farm sizes depending on the agro-climatic zone the district is in. Subdivision below half a module is not permitted, and farms up to 15 modules cannot be expropriated. If a farm is above that reference farm size and declared "unproductive," the Government can potentially expropriate it. A farm is declared unproductive when, after deducting the area that cannot be cultivated (for example, rocky land), 80 percent of the remaining area is cultivated at a yield which is below the average yield for that district. 41. Note that the case for compulsory acquisition cannot be based on cost-savings. The international principle governing the compensation issue is that of "fair" compensation, which will invariably also be reflected in national law. In practice, the courts will interpret this to mean a value which reflects the mar- ket value, or something close to it. Moreover official value estimates, even when they value land improve- ments only and not the soil itself, are usually quite generous. 34 World Bank Working Paper In South Africa, the legal framework to support land reform is one of the most pro- gressive in the world. First of all, the South African Constitution and the property clause in it are quite clear on what constitutes "the public interest." It explicitly includes land reform and redresses the historical inequities. Second, the South African Constitution even allows the state to expropriate below market value, deducting, for example, the value of past subsidies. In addition, legislation to effect compulsory acquisition exists--the 1975 Expropriation Act. It is not unduly complex and restricts court appeals to contest only the value of compensation offered, even though it does not explicitly mention land reform as being in the public interest.42 In fact, it is related legislation (Provision of Land and Assis- tance Act 126, 1993) that provides financial assistance for land acquisition, which intro- duces an additional step in the process by calling for a hearing with affected landowners. The downside of compulsory acquisition is that it is often perceived as too confronta- tional. And in practice, compulsory acquisition is usually more costly and slower than other methods of acquisition. Expropriation transfers the title deed from the former owner to the state, which usually triggers procurement and disbursement regulations for the man- agement and disposal of state assets. These may delay the transfer of property rights to the beneficiaries. The delay sometimes results in a dangerous situation of de facto open access and asset stripping, which then needs to be prevented by employing costly security forces or hiring the former owner back as a caretaker or manager--with all the incentive prob- lems that this process entails. This has been the experience so far with the few expropria- tion cases that the South African Department of Land Affairs has undertaken, and it is this negative experience which largely explains why the Department has not pursued compul- sory acquisition at a significant scale. Market-assisted or Community-driven Land Acquisition. Given the difficulties associ- ated with compulsory acquisition, the question arises why land can not be acquired by the future beneficiaries themselves. The future settlers can independently decide what farm to buy and the land passes directly from the previous owner to the new owner without ever becoming state property. The process becomes much simpler and also more in tune with what the beneficiaries really want: Some may want a farm close to where they currently live, others may want a farm much closer to an urban centre; some may want a large farm suit- able for livestock production, others may want a small plot close to town for irrigated veg- etable production. This approach is often referred to as "community-driven" or "market-assisted" land redistribution. In such programs in Brazil, Central America, Malawi, and India, the com- munities, families, or groups of families are given a grant or subsidized loan by the state to buy their own farm. Of course, they often need help in negotiating these purchases, and there are various ways in which to manage this approach. There is now enough evi- dence to suggest that it is an approach which effectively transfers land at reasonable cost and speed. South Africa is implementing two variants of this approach. The first variant is known as the Settlement and Land Acquisition Grant (SLAG), and it was the Department of Land Affairs' main redistribution program from 1994 and 1999. The approach is demand-driven and does not involve the prior acquisition of land by the State for subsequent resettlement. 42. However, when defining compensation the 1975 Act makes reference to market value plus an addi- tional compensation, referred to as solatium. The latter partly compensates the owner for "suffering." Consensus, Confusion and Controversy 35 It makes a fixed grant available to self-selected beneficiaries, whose eligibility is confirmed or rejected on the basis of a means test--with the maximum total household income set at R1,500 per month. Grants can be used for land purchase for settlement and/or agricultural production. SLAG operations were suspended between 1999 and 2001, pending an internal review, after which SLAG was joined by the Land Redistribution for Agricultural Development (LRAD) program. The internal review had concluded that: (i) the limited involvement of the Department of Agriculture before, during and after project approval severely compromised the success of those redistribution projects which aimed to support agricultural production; (ii) projects undertaken by large groups (over 25 households) had a high failure rate in terms of income generation; (iii) the lack of an own contribution made it difficult to screen appli- cants; and (iv) the means test restricted the target group too much--emerging commercial farmers could not qualify. Asaresultofthisreview,anew sub-programwasaddedtotheredistributionprogram-- the Land Redistribution for Agricultural Development (LRAD) program, developed jointly by the Departments of Agriculture and Land Affairs. The main changes vis-à-vis SLAG included a broadening of the target group to include emerging black, commercial farmers, and the decentralization of project approvals to the provincial level. One mechanism used to achieve this was the "sliding scale" grant, which made higher grants available to benefi- ciaries conditional on the size of their own contribution. This own contribution consists of own labor ("sweat equity"), cash or assets to be used for the project, or a combination of at least two of these. The lowest grant, set at R20,000 per individual, is made available against an own contribution of R5,000 in labor, which is automatically factored in even in cases where cash or assets are contributed. The highest grant of R100,000 is made available to beneficiaries who have contributed R400,000. Part of the budget is outsourced to the Land Bank--the Government's agricultural bank--under an agency agreement, empowering the Land Bank to mix the LRAD grant with a loan to creditworthy applicants. Both SLAG and LRAD have demonstrated that, even as their implementation has been constrained by excessive centralization, the market-assisted approach is able to transfer land at reasonable cost and speed. Both programs experienced a slow start, as systems were being developed and officials gained familiarity with the implementation procedures. Yet, both programs also demonstrated their ability to rapidly accelerate in subsequent years. Under LRAD, project approval was delegated from the Minister at national level to the Provinces, which was the main factor explaining its faster delivery. LRAD started in 2001, but by late 2002 several provinces had exhausted their budgets in the middle of the fiscal year. The program is currently severely budget-constrained. Finally, initial fears that LRAD had abandoned the poor seem unfounded. The self-selection using the sliding scale of grants seems to have been effective in reaching the poor, as well as emerging farmers. Women and youth are also effectively participating in the program. The distribution of the number of grants (Figure 2.2) and the total value of transfers approved by the provincial offices follows a distribution in favor of the poor.43 43. During 2001/2 and 2002/3, the average grant size per beneficiary was R27,696, only about R7,500 above the minimum grant, while the distribution of grants followed a pro-poor pattern for the grants administered by the Department of Land Affairs. The pattern for the Land Bank-administered grants showed that the prospective farmers targeted by the Land Bank benefited from a higher average grant, consistent with its targeting objective. 36 World Bank Working Paper Figure 2.2. South Africa: Distribution of LRAD Grants (2001/2­2002/3) Provincial Land Reform Offices 4,000 3,500 3,000 2,500 beneficiaries 2,000 3,524 3,443 of 1,500 1,000 500 1,071 308 116 25 9 5 90 Number 0 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 100,000 Grant size categories Sales of Parcels by Developers. There are likely to be a number of beneficiaries who would prefer to acquire land as individuals rather than as members of a community. For such beneficiaries, developers could be encouraged to acquire farms in the market (or from the state as a result of compulsory acquisition) for subsequent subdivision and development. While often discussed as an option for land redistribution, we are not aware of actual experiences with this approach, and therefore suggest the developer model be included on a pilot basis first. Negotiated Transfers. Even if better expropriation or restitution legislation is in place, it will always be the case that an "out of court" settlement is far easier and cheaper for all parties involved. The mechanisms for such out of court settlements can range from media- tion, via non-binding arbitration, to binding arbitration. South Africa's restitution process started in 1994 when the first law that the new democratic Government passed was the Resti- tution of Land Rights Act. The Act operationalized the clause in the Constitution which allowed for the restitution of property (physically or financially) to persons who had been dispossessed based on racially discriminatory laws after June 19, 1913. The 1913 date is when the Land Act was passed, restricting the ownership and rental of 87 percent of South Africa's land to whites only. Initially, each restitution case had to be dealt with by a specialized court--the Land Claims Court. This caused the process to slow down to a trickle, with only 41 land claims (out of total of 68,000) to be settled between 1995 and 1999. When the Act was amended to allow for an administrative, out-of-court settlement, the pace picked up dramatically. As of August 2004, 48,000 claims have been settled. At the level of the broader restitution objective, not based on individual claims, but as part of a general redress of inequalities in landownership and land use, governments such as the South African one can anticipate the potential for negotiated settlements of various sorts and at various levels. A legal and policy framework can be put in place which maxi- Consensus, Confusion and Controversy 37 mizes the opportunities for reaching such settlements. Further, governments can create various forums, at the national and at the local level to promote negotiated settlements. Resettlement. Many land redistribution programs in the world are hampered by very bureaucratic and slow approaches to resettling farmers once land has been acquired. For instance, attempts to centrally plan and execute individual land reform projects, whether by single line ministry or dedicated parastatal, invariably end up slowing down the process to a snail's pace. Centralizing all aspects of land reform into specialized land agencies has usu- ally not been able to speed up the process either, as the examples from Mexico, Colombia, the Philippines, and Honduras show. Instead, the "one-stop-land-reform-shops" have spawned costly and paternalistic bureaucracies. Alternatively, several ministries have to work closely together, which is also very difficult to coordinate. Whatever the land acquisition method, much can be done to redistribute and resettle land in a faster and less bureaucratic way. Bureaucratic processes can be streamlined, but most importantly they can be decentralized and driven by the beneficiaries themselves. Decentralization should imply that the beneficiaries of redistribution should have much more say into the way in which resettlement--their resettlement--is carried out. For instance, it is eminently sensible to give beneficiaries much more say in how the farm will be planned, what services will be needed, and who should provide these services. Why not give beneficiaries the choice as to who should help them plan the farm, provide access roads, ensure water supply, and so forth, for example by giving them the financial resources to procure them themselves? Sure, these services can be provided by govern- ment ministries, but there may be private sector providers or NGOs who can deliver these services better and cheaper. Why not allow much more flexibility in how this is done? Why not allow much more community participation and decentralization? Why not allow for much more private sector and NGO involvement? Why not define national standards on how this should be done, but decentralize implementation and supervision to the local level? Central ministries often resist decentralization on transparency and accountability grounds. It is felt that requiring ministerial approval ("vertical accountability") reduces opportunities for collusion and corruption with regards to the selection of beneficiaries and/or the farm price. In practice, ministerial approval for individual land reform projects always results in long delays, while it is unclear how the minister would obtain the neces- sary information to detect opportunistic behavior at the local level. Decentralization, on the other hand, speeds up decisionmaking, but could indeed lead to more corruption if the "vertical" accountability is not replaced by more "horizontal" and "downward" account- ability. Such horizontal accountability could be achieved by stakeholder participation (government and non-government) in the decisionmaking process. In South Africa, the first land redistribution program--supported by SLAG--adopted an implementation strategy under which each project needed the approval of the Minister of Land Affairs. This approval would in turn largely be based on a perusal of farm and busi- ness plans drawn up by consultants hired by the Ministry. As a result, the program took considerable time to take off and reach significant numbers, and became consultant- driven. Many of the business plans may have looked quite compelling on paper, but were not "owned," or even understood, by the beneficiaries themselves. 38 World Bank Working Paper Incidentally, South Africa's restrictions on subdivision caused many problems for the land redistribution program. SLAG was based on a fixed grant per beneficiary family, set at R15,000 initially, and later raised to R16,000. Because the land market did not have any R15,000 plots on offer, beneficiaries were forced to pool their grants together to be able to purchase the large farms available on the market. This led the more entrepreneurial bene- ficiaries to put large groups together with the sole purpose of reaching the required farm price. This became commonly known as the "rent-a-crowd" phenomenon. Failure to subdivide also created problems for the choice of property rights by the ben- eficiaries. When the title deed to a large farm had to be legally transferred to the beneficiary, it was administratively easy to transfer it to a group of beneficiaries as one title deed. Instead ofbeneficiarieschoosingthetypeoflegalentityunderwhichtheywouldholdtheland,which should have led many to opt for individual title or shares, the Department of Land Affairs steered the beneficiaries toward the newly legislated Communal Property Associations (CPAs). There is of course nothing "wrong" with common property (Bromley 1992; Netting 1976; van den Brink, Bromley and Cochrane 1994). However, the practice of creating CPAs became so widespread that practice turned into de facto policy. This restrained beneficiaries' freedom of choice and forced them to participate in often complex group processes. In addition, given the dearth of small family farms on the market due to the subdivision restrictions, many groups had to acquire large, often capital-intensive farms. South African consultants would then proceed to draw up business plans which attempted to continue the operations of the large commercial farm as a collective farming enterprise, ignoring the volu- minousandconsistentlynegativeinternationalexperiencewithcollectivefarming(Deininger 1995). Again, there were no explicit policy instructions to this effect, but under no project was a farm ever subdivided (by making use of the policy exemption granted to do so under the land reform legislation) into individual family farms of a size that the beneficiary fami- lies would be able to manage comfortably by themselves. As a result, these land reform proj- ects subjected the beneficiaries to unfortunate attempts at collective farming, in some cases managed by the previous farm owner at a fee. Repeating international experience with col- lective farming, these attempts invariably failed, or were prevented from failure only through intensiveandunsustainablecarebyNGOs.Mostofthemembersofthegroupsseefew,ifany, benefits. Many beneficiary groups are now saddled with complex problems of reorganizing and reconstituting the membership, as many members of the collective have lost interest. A land redistribution program which attempts to change an agrarian structure such as the one in South Africa, which is completely dominated by large farms, will need to be flex- ible enough to be able to fill in a large spectrum of farm sizes. It will need to accommodate peri-urban gardens, medium-scale commercial farms, irrigated vegetable plots, as well as small livestock ranches. It will need to cater for poor, vulnerable and marginalized groups, as well as emerging commercial farmers. The design of LRAD attempts to incorporate this flexibility, both by allowing a sliding scale of grants, and by allowing projects to allocate more or less to land acquisition, or more or less to agricultural development of that land. While purely residential projects are not supported under LRAD, but by SLAG, beneficiaries seeking to establish household gardens at their new residences can be supported. In addition, beneficiaries can use the LRAD grant to participate in so-called equity schemes and become shareholders in existing agricultural enterprises. Farm workers can use LRAD to participate in employee-ownership schemes. Other beneficiaries enter LRAD to engage in commercial agricultural activities. They access Consensus, Confusion and Controversy 39 the grant and combine it with normal bank loans, approved under standard banking pro- cedures, and their own assets and cash to purchase a farm. Finally, while many people liv- ing in communal areas already have secure access to agricultural land, they may not have the means to make productive use of that land. Such people would be eligible to apply for assistance to formalize their tenure and make productive investments in their land. The lessons learned so far during the implementation of LRAD and SLAG-supported projects suggest that the flexibility in design has not always resulted in flexibility in prac- tice. For instance, officials and consultants enter bilateral agreements during project prepa- ration, sometimes completely sidelining the beneficiaries. The government officials, not the beneficiaries, present the project proposal to the Provincial Grant Committees for approval. Even though beneficiaries are free to chose their legal entity, including individ- ual title on subdivisions, the majority of beneficiaries applying as a group has been steered toward CPAs and trusts, even under LRAD. Again, there is nothing inherently problem- atic with these forms of ownership, but one would have expected a wider variety of legal entities if beneficiaries were truly empowered to make these decisions. Finally, community procurement of goods and services has not yet been allowed under the program. This means that all procurement has been undertaken by Government, including the selection of design agents and technical advisers, and the purchase of agri- cultural inputs and assets. The lack of community procurement has made a scaling-up of the program very difficult, if not impossible. It has also constrained the flexibility of ben- eficiaries to make their own choices and thereby disempowered them. Land Reform is Usually Underfunded The most important lesson from experience is that much more than just the land needs to be funded for land reform to be successful within a reasonable time frame, say 5 to 10 years. International experience shows that in a typical land reform project, the land costs are only part (30 to 40 percent) of the total costs of land reform (Figure 2.3). The other costs, which are essential for the success of the undertaking, include the costs of housing, resettlement, start-up grants, inputs, tools, equipment, farm development, training and advisory ser- vices, and the overhead costs of the management of the land reform program. The relative importance of each of these "slices" varies depending on the particular livelihood created. A homestead for a rural artisan or worker, with a small vegetable plot, but close to town, will need to pay more for the "land slice" of the project than for, say, equipment and farm development. A beneficiary who wants to acquire a farm which has already substantially been developed, will have a much bigger land cost slice than a piece of undeveloped land. What are the options for its financing? One way of approaching this question is to look at the various stakeholders who could pay for land reform: (i) the beneficiaries themselves; (ii) current landowners; (iii) former colonial powers; (iv) donors; and (v) government. Beneficiaries. Almost all land reform programs in the world have asked the beneficia- ries to contribute to the costs of land reform. This happened either explicitly (through cash, sweat equity and loans), or implicitly (by simply dumping beneficiaries on the land with- out any support). To ask beneficiaries to contribute toward the costs of land reform makes sense. Requiring an own contribution will help to self-select people who are actually will- ing and able to run a farm. But this should be done within limits, so as not to exclude the 40 World Bank Working Paper Figure 2.3. South Africa: Land Only Part of the Costs of Typical Land Reform Project Land Other investments Inputs Resettlement Advice Overhead poor or saddle the new farmers with too much debt. Their loans should not exceed a cer- tain proportion of their assets. While this limit depends on the profitability and riskiness of the particular farming system adopted, an internationally reasonably robust "rule of thumb" is that the loans need to stay below about 30 percent of the value of assets (the so- called 30 percent debt-equity ratio). Landowners. In Latin America (for example, Mexico, Colombia, Central America, and Bolivia), expropriation below market value often has been attempted by governments to force land owners to contribute to the costs of land reform. Compensation for land was often paid for in land reform bonds, which could not be traded and had low interest rates. Not sur- prisingly,therefore,ownersfiercelyresistedlandreform.Inaddition,mostoftheLatinAmer- icanlandreformprogramswerechronicallyunderfundedandthereforeproceededatasnail's pace (such as in Mexico). Because these programs provided almost no finance beyond land and some housing, agricultural success was therefore very slow in coming. Expropriation is a useful approach as part of overall land reform strategy, but expro- priation does nothing to reduce the land cost. Rather, it tends to increase it, because of the likelihood of litigation. Only outright confiscation reduces the land costs, but it has many other undesirable consequences, such as a reduction of investor confidence. This, in turn, can easily lead to a devaluation of the currency, imposing the costs of land reform on the entire nation. Unless strong legislation is in place which limits the power of for- mer owners to go to court to block the expropriation, it will also slow down implemen- tation. Alternatively, expropriation can be managed in a chaotic way, as in Zimbabwe. Having landowners contribute to the costs of land reform by imposing a land tax on the value of the unimproved land is politically and economically perhaps more attractive, as argued above. Consensus, Confusion and Controversy 41 Donors. Former land owners in Kenya and Algeria were compensated for their farms by Great Britain and France. In Kenya, a World Bank loan financed complementary invest- ments, inputs and overhead costs, and the land reform was highly successful. In Algeria, little complementary finance was available, leading to much reduced success rates. After the end of the Cold War, many donors professed willingness to finance land reform pro- grams, but few have done so. The World Bank is an exception to this, currently funding programs in Brazil, Guatemala, India, Malawi, and the Philippines. Until recently most donors, including the World Bank, would not pay for the land, but only for the other costs. The World Bank has now changed that policy, and India and Malawi are among the first countries in which land is actually financed. Government. International experience shows that a sound financing plan must first rest on own fiscal resources. In the South African case, the prospects for this look good. Based on estimates of current budget trends, South Africa's fiscal support for land reform is increasing significantly. In the current three-year national budget, the 2007/8 land reform budget rises to R5.7 billion. If we assume that this level of financing is not reduced until 2014 (the year by which the 30 percent target needs to be reached), a cumulative budget of R56 billion will be available for land reform. Estimates of the total costs based on the current land reform costs per hectare put the total around R35 bil- lion. However, as explained above, in the current program the non-land costs are under-funded. Another way of demonstrating that the projected fiscal resources for land reform seem "in the right ball park" is to start with the value of commercial farm assets and to take 30 percent of that. This would amount to about R30 billion.44 This value constitutes more than just the land, because it also includes houses, buildings, and fixed improvements. The adverse consequences of inadequate finance are severe, including slowing down implementation, creating strong political resistance of the land owning classes, and under- mining the chances of success of the settlers--consequences we have already seen in South Africa. Instead, adequate financing by government can be used to leverage additional financ- ing by beneficiaries (by allowing them to safely borrow and increase their productivity), donors, and landowners. Fortunately, South Africa's recent land reform budget trends put the national targets within reach. International experience shows that on the basis of this commitment, an effective partnership with stakeholders can be built. Large Farmers Will Lobby Against Land Reform We have argued that economies of scale in agriculture cannot explain the emergence of agrarian structures based on large scale farms. In fact, as Binswanger, Feder, and Deininger (1995) demonstrate, these agrarian structures have emerged historically as the result of coercion and distortions by the politically powerful. Even today, large landowners repre- sent a powerful political force in many countries, able to profoundly influence govern- ment policy and extract subsidies. Locally, their political influence is often very direct by 44. The value of total commercial farm assets in 2002 was R98.4 billion, and 30 percent of R98.4 billion isR30billion.Seehttp://www.statssa.gov.za/publications/Report-11-02-01/CorrectedReport-11-02-01.pdf. To put this number in perspective: current spending on social welfare per year is R73 billion. 42 World Bank Working Paper being able to influence the votes of their farm workers or tenants.45 Large farmer lobbies in Southern Africa are, in general, opposed to a substantial restructuring and downsizing of the agrarian structure of commercial farm areas. As said earlier, removing the land market distortions and the subsidies conferred to large farmers through other markets (for example, for outputs and credit) improves the role of land markets in the redistribution of land from large to small farmers. After 1994, South Africa liberalized agricultural marketing and reduced most subsidies to commercial farmers to very low levels in one of the most complete agricultural liberalizations in the world. The political transition of 1994 allowed a substantial liberalization of the agricul- tural sector. However, other restrictions remain, demonstrating the considerable political power of the large farm lobby in South Africa. One can hardly find a better demonstration of the strength of the large farm lobby in Southern Africa than the fact that neither Zimbabwe nor South Africa relaxed their subdivision rules or imposed a land tax after independence. If there were uncertainties about compensation, the position of the farm lobby would be understandable and rational. Compensation uncertainties present personal financial risk, and influence expectations, which immediately reduces land prices. However, there is an other reason for the large farmer opposition to land reform: a reluctance to integrate poorer black neighbors into a more racially integrated farm community. Instead of view- ing integrated rural communities as providing increased long-term security, parts of the white farming community in Southern Africa view an influx of black families as causing more insecurity, given their experience with large farms that border the former homelands and are often prone to theft and vandalism. Political theory and history suggests that these anti-land reform lobbies may only switch strategy when they perceive that a large-scale land reform program is the price they have to pay for peace. Unfortunately, by then the situation may already have deteriorated to such an extent that an "orderly" land reform program will have become impossible, as the example of Zimbabwe so amply shows. Fortunately, the recent events in Zimbabwe have convinced a growing number of stakeholders that the status quo in South Africa is unsustainable. More private initiatives are being launched with the objective of promoting land reform and emerging black farm- ers. These initiatives come from a number of producer associations, NGOs, and commer- cial banks. At the same time, some stakeholders at the local level are also starting to think about working together to achieve integrated land reform implementation plans in their municipalities or districts. So far, however, the Government has not systematically tried to incorporate these initiatives into an institutional framework, which consistently promotes dialogue, consensus-building and partnership in planning and implementation. Finally, it should be mentioned that there exists another lobby that is not so much interested in cutting up and subdividing large farms, as in simply changing the owner- ship of the large farms. This is the so-called "same car, different driver" lobby. While they 45. But this influence may also turn against them. The radical nature of Zimbabwe's Fast Track land reform as illustrated by the Government's refusal to negotiate a compromise with white farmers, or to take any interest in the fate of the affected farm workers, was often rationalized by senior ZANU-PF lead- ers as the appropriate punishment meted out for supporting the opposition in defeating a popular refer- endum on a new constitution in 2000. Consensus, Confusion and Controversy 43 may not be opposed to land reform for small farmers per se, they would also argue that there be ample room created for large-scale farms owned and operated by black commer- cial farmers. Clearly, in Southern Africa this group represents a strong "nationalist" senti- ment that believes the commercial farm sector itself should be de-racialized. It would make a lot of political sense to accommodate this group, rather than exclude it. In this way, the land reform process becomes more inclusive and may benefit from a much broader polit- ical base. What would become problematic, though, if is this particular group would be able to torpedo the wider land reform agenda, or successfully lobby for their installation as "telephone farmers" (absentee landlords) and for continued farm and credit subsidies and other discriminatory policy distortions, and lobby against the introduction of a land tax and the relaxation of subdivision rules. If the Government's social objective is to achieve equity and efficiency in farming, such a lobby would be counter-productive. Policy Recommendations An internally consistent land reform strategy should have the following pillars: (i) it should boost land market forces that could redistribute land from the rich to the poor; (ii) it should improve on the processes of land acquisition and resettlement; and (iii) it should create consensus around the implementation strategy, together with stakeholders. In the remainder of this paper, some more detailed, concrete policy suggestions are identified, with a particular emphasis on South Africa. Land Market Reforms We have argued that in many countries, land markets--as they are currently designed-- cannot be counted on to redistribute land from the rich to the poor. That, however, does not mean that land markets should not play a far greater role in land reform than they often do now. In fact, irrespective of which land redistribution approach is taken, better per- forming land markets will make the land reform process work better, faster and cheaper. Recall that the price of land in the market reflects the value of the income stream from agriculture, plus its value as an asset, such as, for example, a hedge against inflation or its speculative value--conversion into residential property, a new road, and so forth. Poor farmers will only be able to afford to pay the agricultural value, and will therefore be out- bid by the rich in the land markets. To counteract this disadvantage, several policy reforms must be undertaken. First, as in South Africa, it is suggested that all distortions and subsidies be removed that favor large farmers only (because they will find their way into the land price--for example, a special mortgage interest rate subsidy, or a subsidized output price only acces- sible to large farmers given certain marketing arrangements). Second, it is recommended that targeted grants or subsidies be provided to the poor and other aspiring family farm- ers to purchase land. Third, subdivision restrictions should be removed and a progressive land tax should be put in place--in essence a user charge for land. In other words, it is imperative to level the playing field in agricultural and related markets between large and small scale farmers, boost the purchasing power of the poor, and eliminate the incentives for the wealthy to hold land for non-agricultural purposes. 44 World Bank Working Paper When beneficiaries are given a grant to purchase land, it will be important that the land market supply farms in sizes that correspond to the grants. Otherwise, substantial trans- actions costs are imposed on the beneficiaries, because they need to organize themselves and pool their grants to purchase large farms. The land market will need to work in such a way that a supply of "grant-sized," small farms is available. It should be relatively costless to subdivide the farm (or that part of the farm that the group has decided should be indi- vidualized) after purchase by a group of beneficiaries. Part of the full costs of subdivision could be borne by the government. One needs to be mindful of the price-raising effect that land purchase subsidies have, if they are given on a substantial scale. In that case, it becomes even more important to ensure that the market can deliver farms of various sizes. This implies that subdivision rules need to be relaxed, that large farm subsidies are eliminated, because they raise the price of land, and that there is a financial incentive for large farmers to sell unused land (a land tax). It is important to understand that a land tax is different from a property rates tax. The ideal land tax would tax the potential agricultural profits from a particular piece of unim- proved or unused land. Unlike a property rates tax, a land tax would not tax the value of investments on that land, or the value of the farmhouse erected on that land. Taxing invest- ment in agriculture is probably the last thing a government should want to do in the context of a land reform program. A land tax supportive of land reform could be flat or progressive, and would exempt small farmers from making significant tax payments, if any. Alandtaxmustbesimpletoadminister,leavenoloopholes,andhavelittleroomfordis- cretionary valuations of the farm. One way of achieving this is to use existing agro-climatic zones as a proxy for potential agricultural profits, and to set different land taxes for different zones. Taxation within a zone would then be imposed on a simple per hectare basis, without any exceptions and exemptions. In theory, it is possible that tax brackets are defined in such a way that almost all truly productive farms are exempt from taxation. Such a tax would reduce the speculative land price premium and release unused land into the market. Finally, the land tax revenues could be both a source for local government revenues and the financ- ing of land redistribution. Land Acquisition Methods Develop a Menu of Land Acquisition Options. The overall policy objective would be to have a ready set of complementary land acquisition methods that have been tested and made operational. For instance, even if a government decided to pursue expropriation as its main strategy, it would be prudent to have the alternative of community-driven land redistribution at hand, to give government and landowners an alternative option to avoid litigation. This is sometimes referred to as the "sandwich" or "stick-and-carrot" approach. An improved pol- icy framework would thus consist of a package of at least three options for land acquisition: compulsory acquisition, market-assisted or community-driven land acquisition, or negoti- atedlandtransfers.Governmentsshouldaddvariants,adaptedtolocalcircumstances,ofthese optionstotheir"toolkit"andstarta"learning-by-doing"process,flexibleenoughtobescaled up when good results are obtained. Inimplementingcompulsoryacquisitionpilots,theGovernmentcouldtest,andimprove on, the Expropriation Act of 1975, ensuring that it is consistent with the Constitution. It is also advisable to find a legal mechanism which transfers the ownership directly, or almost Consensus, Confusion and Controversy 45 directly, from the former owner to the beneficiaries and avoid a lengthy transfer of own- ership during which the State has to ensure the security of the asset "in transit." Make Acquiring Subdivisions Easy. Compulsory, market-assisted, and negotiated land acquisition methods all need to be able to acquire subdivisions, rather than whole farms. In many cases, the state will not be interested in acquiring the whole farm for redistribu- tion, but rather a part of it, leaving the farm owner with the part that originally was used for residential and intensive farming purposes. This method of acquisition has several advantages. First, it has the advantage of causing the highest possible increase in agricul- tural production. Very little existing production is disturbed, while unused land is brought into production by the new, small-scale settlers. Second, it avoids costly experiments by beneficiaries to attempt keeping the commercial parts running under collective farming arrangements. Third, it saves on acquisition costs, by not acquiring what is probably the most expensive part of the farm, and also the part of the farm initially least likely to be effec- tively used by small-scale farmers. Fourth, acquiring subdivisions would create new farm "neighborhoods" in which the new neighbors may be able to work together and help each other. Such new neighborhoods will have substantial political benefits, in particular if there is a history of antagonism between classes or races. Resettlement Models Decentralize Decisionmaking. South Africa's experience with land redistribution since 1994 confirms the international lessons that underline the need to decentralize and make programs more community-driven. South Africa's flagship redistribution program, LRAD, made the important step of decentralizing decisionmaking down to the provincial level, and reaped immediate benefits in terms of speed and the quantity of projects. The logical next step is to decentralize even further to the district level, followed by further decentralization down to the municipal level. As the approval of land reform projects gets further decentralized to the district and municipal levels, and as beneficiaries, officials, and stakeholders can better exploit the flexibility of the LRAD policy, redistribution will become faster, cheaper, and more in line with local conditions and the capacities and needs of the beneficiaries. Strengthen Accountability. As vertical accountability is relaxed, horizontal and down- ward accountability, and integration between programs should be strengthened. All land reform programs should be channelled through the same screening and approval processes. These processes should be managed by multi-sectoral committees at the local government level that allow for stakeholder participation. The land reform programs can then become an integral part of the local development plans, which in South Africa are the basis for local development budgeting and implementation. District Land Reform Committees could be constituted as sub-committees of the District Councils. Allow Community Procurement of Goods and Services. In South Africa, land reform, and many other development programs which attempt to deliver services in a decentral- ized fashion, suffer from the prohibition on community procurement which the Public Financial Management Act of 1998 seems to have placed on procurement procedures. The 46 World Bank Working Paper restrictions are the result of an unduly conservative interpretation of this Act, preventing the direct transfer of public resources to communities. In fact, there do exist government programs which have transferred funds to communities without running foul of the Act. In order to accelerate and improve the LRAD program, however, community procure- ment needs to be introduced. Communities would be allowed to manage resources directly, following simple and transparent rules, such as gathering three quotes before pur- chases are made and documenting democratic decisionmaking with respect to procure- ment decisions. Pilot a "Developer" Model. Some beneficiaries would prefer to go it alone, and not as part of a group. The policy recommendation is to pilot a redistribution approach that fol- lows proper plan procedures and led by a developer. The pilot's objectives could be to fill in underused, but prime, areas close to urban centers, but also to settle contiguous, under- used or derelict farms bordering on communal areas. A developer would then be brought in to assist the communities to restructure the farms, create subdivisions and common areas, put in basic infrastructure, and further develop the organizational capacity and agri- cultural skills base of the communities. In South Africa, the developer model fits well with the Government's strategy to promote public-private partnerships (PPP). During the pilot phase, developers could benefit from a limited transactions cost subsidy. However, devel- opers should be fully privately financed if the pilots turned out successful. National Implementation Strategy As mentioned earlier, it would be in the interest of all stakeholders, especially the com- mercial white farmers, to participate voluntarily in the land reform process. This partici- pation should be seen as the price to pay for long-term social and political stability. Already, we have pointed out that it would make sense to create ample opportunities to come to a negotiated or market-assisted transfer of land. Rather than slowing a gov- ernment down in its goal to achieve land reform, voluntary transfers of land can accelerate implementation. Clearly, large farmers worry about compensation, while governments have ideas about whether they should compensate for the farm as a whole, or only for the improve- ments. Again, these are issues to be debated at a national level in a democratic way. There are many ways of looking at the issue, and many ways of arriving at compensation that is fair to both sides. More damaging than anything else is uncertainty about the level of com- pensation or the timing of it. Simply put, people have to be able to get on with their lives and be able to plan. If they cannot, their lives and overall confidence in the economy will be affected. If that happens, investment falls, the currency depreciates, and everybody pays the price for this uncertainty. At this point in time in South Africa's land redistribution efforts, the need for a broad- based consensus around a national strategy for implementing land reform is obvious. There is a broad-based consensus emerging among the various stakeholders that South Africa needs to solve its land question as a matter of urgency. What government needs to do now is build on this emerging consensus and involve stakeholders in a dialogue around policy implementation. Stakeholders, including local government structures, farmers asso- ciations, NGOs, and churches, can assist in a number of ways. They can identify urgent Consensus, Confusion and Controversy 47 landneeds,supportbeneficiariesinaccessingthevariouslandreformprograms,andprovide technical assistance as demanded by the beneficiaries. NGOs and research institutions can provide valuable monitoring and evaluation services, and assist in policy improvement. Conclusion The first chapter of this paper focused on property rights in land--what most would describe as "land tenure" or "land policy" issues. A new consensus on the analysis and solutions to these issues is emerging. There is a much stronger recognition today that prop- erty rights are social constructs and not merely technical attributes of "things." Property rights should be clearly defined, well understood, and accepted by those who must abide by them. However, secure property rights should not be confused with full private "owner- ship." Under certain economic conditions, property rights tend to become more individu- alized and formalized. However, the introduction of private title in situations where such economic conditions do not exist can be a waste of effort. Moreover, the emergence of cer- tain property rights regimes is not just a matter of shifting relative prices and changing trans- actions costs, but can be profoundly influenced by the political dynamics of power relations. The second chapter of this paper looked at the controversy that surrounds the redistri- bution of property rights in land. There is consensus on the need to address the issue of highly unequal land distribution with a renewed sense of urgency. However, the debate on large versus small farms and the optimal land redistribution implementation mechanisms continues. These controversies spill over into the choice of the beneficiaries (subsistence ver- sus commercial) and on the preferred approach to land acquisition (compulsory versus market-assisted). The controversy should not become an excuse for inaction. Therefore, the paper sug- geststhatallstakeholdersagreetodisagreeexanteontheoptimalapproach.Instead,theopti- mal way forward is to agree on a policy framework which allows a menu of options to be pursued, the results of which can then be evaluated as the program proceeds, and corrections made when ex post evaluation shows some negative aspects. Rather than endlessly debating the pro's and con's of each particular approach, we propose to create a policy arena in which the particular models can show their relative performance in competition with each other. Ofcourse,oneneedstoagreeontherulesofthegame,sothattheperformanceofeachmodel can be compared. In the short-term, performance would be defined by the "inputs" and "outputs" of a land reform program: the number of beneficiaries, the amount and speed of the land transfer, fiscal costs per beneficiary and per hectare, and the speed with which agri- culturalproductionisestablishedandincreased.Inthemedium-term,theperformancemea- sureswouldincludethe"outcomes":thereductionoftensionsandviolenceintheruralareas, the increase in rural and agricultural growth, the reduction of poverty, and the improved environmental condition of the land. The mid-term review of South Africa's redistribution program suggests that: (i) major land market reforms still need to be implemented (with respect to subdivision, land tax and zoning); (ii) the program is able to deliver land at a substantial rate and pace, but it is constrained by budget constraints and insufficient decentralization and inadequate empowerment of beneficiaries themselves; (iii) a "developer" model should be put in place, under which a developer would restructure a large farm and develop a range of small 48 World Bank Working Paper farms to be acquired by beneficiaries; and (iv) a consensus between the stakeholders around an overall implementation strategy needs to be reached. This consensus would clearly indicate and formalize the different roles of the stakeholders in the various land reform programs, allow for experimentation within the rules of the game, and also facili- tate negotiated transfers of land. The priority is to speed up and improve South Africa's land redistribution program. The debates around the need for radical versus incremental change, or the relative merits of the various approaches should not stifle and overshadow the discussions on the practi- cal lessons that have emerged from the post-1994 experience. This makes sense not just from a technical, but also from a political perspective. The politics of land reform can result in costly inaction. One can think of worse things than a situation in which the various stakeholders in Government, the private sector, and civil society agree on an overall framework for implementation and compete with each other on the ground to demonstrate the success of "their" model. Major land redistribution can be implemented peacefully: history need not repeat itself ad nauseam. References Acemoglu, Daron, Simon Johnson, and James Robinson. 2001. "The Colonial Origins of Comparative Development: an Empirical Investigation." American Economic Review XCI:1369­1401. ------. 2002. "Reversal of Fortune: Geography and Institutions in the Making of the Mod- ern World Income Distribution." Quarterly Journal of Economics 117(4):659­694. Aghion, Philippe, Eve Caroli, and Cecilia Garcia-Peñalosa. 1999. "Inequality and Economic Growth: The Perspective of the New Growth Theories." Journal of Economic Literature 37(4):1615­60. Binswanger, Hans, Klaus Deininger, and Gershon Feder. 1995. "Power, Distortions, Revolt and Reform in Agricultural Land Relations." In Jere Behrman and T. N. Srinivasan, eds., Handbook of Development Economics. Vol. 3B. Amsterdam: Elsevier. Binswanger, Hans, and John McIntire. 1987. "Behavioral and Material Determinants of Production Relations in Land-abundant Tropical Agriculture." Economic Development and Cultural Change 36(1):73­99. Binswanger, Hans, and Mark Rosenzweig. 1986. "Behavioral and Material Determinants of Production Relations in Agriculture." Journal of Development Studies 22(3):503­39. ------. 1995. "Wealth, Weather Risk and the Composition and Profitability of Agricul- tural Investments." Economic Journal 103(January):56­76. Boserup, Ester. 1981. Population and Technological Change: A Study of Long Term Trends. Chicago: University of Chicago Press. Bromley, Daniel W. 1992. "The Commons, Common Property, and Environmental Policy." Environmental and Resource Economics 2:1­17. Bruce, John W., and Shem Migot-Adholla, editors. 1993. Searching for Land Tenure Secu- rity in Africa. Dubuque, Iowa: Kendall/Hunt Publishing Company. 49 50 World Bank Working Paper Bundy, Colin. 1979. The Rise and Fall of the South African Peasantry. Berkeley and Los Angeles: University of California Press. Chayanov, A. V. 1966. The Theory of Peasant Economy. Edited by Smith, Thorner & Kerblay. Homewood, Illinois: Richard Irwin. Christodoulou, N. T., and N. Vink. 1990. "The Potential for Black Smallholder Farmers' Participation in the South African Agricultural Economy." Paper presented at the Conference on Land Reform and Agricultural Development in South Africa by the Newick Park Initiative, October, United Kingdom. Deininger, Klaus. 1995. "Collective Agricultural Production: A Solution for Transition Economies?" World Development 23(8):1317­1334. ------. 2003. Land Policies for Growth and Poverty Reduction. World Bank Policy Research Report. Washington D.C. and Oxford: The World Bank and Oxford University Press. Deininger, Klaus, Raffaella Castagnini and María González. 2004. "Comparing Land Reform and Land Markets in Colombia: Impacts on Equity and Efficiency." World Bank Policy Research Working Paper 3258, April 2004. Washington D.C.: The World Bank. Deininger, Klaus, and Lyn Squire. 1998. "New Ways of Looking at Old Issues." Journal of Development Economics 57:259­87. Deininger, Klaus, and Pedro Olinto. 2000. "Asset Distribution, Inequality, and Growth." Policy Research Working Paper Number 2375. Washington, D.C.: The World Bank. Demsetz, Harold. 1967. "Toward a Theory of Property Rights." American Economic Review 57(May):347­59. De Soto, Hernando. 2000. The Mystery of Capitalism: Why Capitalism Triumphs in the West and Fails Everywhere Else. London: Bantam Press. Gibson, James L. 2001. "The Land Question in South Africa: Clouds on the Horizon." The Institute for Justice and Reconciliation: www.ijr.org.za/art_pgs/art23.html Goldschmidt, Walter. 1947. As You Sow: Three Studies in the Social Consequences of Agribusiness. Montclair, N.J.: Allanheld, Osmun and Co., Publishers Inc. Hardin, G. 1968. "The Tragedy of the Commons." Science 162:1243­8. Hoogeveen, J. G. M., and B. H. Kinsey. 2001. "Land Reform, Growth and Equity: Emerging Evidence from Zimbabwe's Resettlement Programme. A Sequel." Journal of Southern African Studies 27(1):127­36. Jodha, N. S. 1992. Common Property Resources: A Missing Dimension of Development Strate- gies. Washington, D.C.: The World Bank. Kislev, Yoav, and Willis Peterson. 1982. "Prices, Technology and Farm Size." The Journal of Political Economy 90(3):578­95. Lastarria-Cornhie, Susana, and Jolyne Melmed-Sanjak. 1998. Land Tenancy in Asia, Africa and Latin America: A Look at the Past and a View to the Future. Draft Report for the Food and Agriculture Organization of the United Nations. Land Tenure Center Uni- versity of Wisconsin, Madison, Wisconson. Licht, Daniel S. 1997. Ecology and Economics of the Great Plains. Lincoln: University of Nebraska Press. McCarthy, Nancy, Brent Swallow, Michael Kirk, and Peter Hazell, editors. 2000. Property Rights, Risk, and Livestock Development in Africa. Washington, D.C., and Nairobi: Inter- national Food Policy Research Institute and International Livestock Research Institute. Netting, Robert. 1976. "What Alpine Peasants Have in Common: Observations on Com- munal Tenure in a Swiss Village." Human Ecology 4:135­46. Consensus, Confusion and Controversy 51 North, Douglass C., and Robert Paul Thomas. 1977. "The First Economic Revolution," Economic History Review 30:229­41. Ostrom, Elinor. 1994. Neither market nor state: Governance of common-pool resources in the twenty-first century. IFPRI Lecture Series 2. Washington, D.C.: International Food Pol- icy Research Institute. Peterson, Willis L. 1997. "Are Large Farms More Efficient?" Department of Applied Eco- nomics, University of Minnesota, Staff Paper Series, P97-2. Ravallion, Martin, and Shaohua Chen. 2004. "China's (Uneven) Progress against Poverty." World Bank Policy Research Working Paper Number 3408. Washington, D.C.: The World Bank. Sadoulet, Elisabeth, and Alain de Janvry. 1995. Quantitative Development Policy Analysis. Baltimore: Johns Hopkins University Press. Sjaastad, Espen, and Daniel W. Bromley. 2000. "The Prejudices of Property Rights: On Individualism, Specificity, and Security in Property Regimes." Development Policy Review 18:365­389. Van den Brink, Rogier, Daniel W. Bromley, and Jean-Paul Chavas. 1995. "The Economics of Cain and Abel: Agro-pastoral Property Rights in the Sahel." The Journal of Devel- opment Studies 31(3):373­399. Van den Brink, Rogier, Daniel W. Bromley, and Jeff Cochrane. 1994. "Property Rights and Productivity in Africa: Is There A Connection?" Development Southern Africa 11(2): 177­82. Van Schalkwyk, Herman, and Johan van Zyl. 1996. "The Land Market." In Johan van Zyl, Johann Kirsten, and Hans P. Binswanger, Agricultural Land Reform in South Africa: Policies, Markets and Mechanisms. Cape Town: Oxford University Press. Van Zyl, Johan, Hans Binswanger, and Colin Thirtle. 1995. "The Relationship Between Farm Size and Efficiency in South African Agriculture." World Policy Research Work- ing Paper Number 1548. Washington, D.C.: The World Bank. Wheeler, M. W., and G. F. Ortmann. 1990. "Socio-economic Factors Determining the Suc- cess Achieved Among Cotton-Adopting Households in Two Magisterial Districts of Kwa-Zulu." Development Southern Africa 7(3):323­333. Wolpe, Harold. 1972. "Capitalism and cheap-labor power in South Africa: From segrega- tion to apartheid." Economy and Society 1(4):425­56. World Bank, 1975. Land Reform: Sector Policy Paper. World Bank Development Series. Washington, D.C.: The World Bank. ------. 2003. Sustainable Development in a Dynamic Word: Transforming Institutions, Growth, and Quality of Life. Washington, D.C., and New York: The World Bank and Oxford University Press. Eco-Audit Environmental Benefits Statement The World Bank is committed to preserving Endangered Forests and natural resources. We print World Bank Working Papers and Country Studies on 100 percent postconsumer recy- cled paper, processed chlorine free. The World Bank has formally agreed to follow the rec- ommended standards for paper usage set by Green Press Initiative--a nonprofit program supporting publishers in using fiber that is not sourced from Endangered Forests. For more information, visit www.greenpressinitiative.org. In 2004, the printing of these books on recycled paper saved the following: Trees* Solid Waste Water Net Greenhouse Gases Electricity 307 14,387 130,496 28,262 52,480 * Pounds Gallons Pounds KWH 40" in height and 6-8" in diameter Consensus, Confusion, and Controversy: Selected Land Reform Issues in Sub-Saharan Africa is part of the World Bank Working Paper series. These papers are published to communicate the results of the Bank's ongoing research and to stimulate public discussion. Land reform can be divided broadly into land tenure reform (the establishment of secure and formalized property rights in land) and land redistribution (the transfer of land from large to small farmers). This paper therefore is in two parts. The first part focuses on proper- ty rights, giving a short narrative of some of the key land tenure and land policy issues. Though these issues remain politically sensitive, a solid consensus is emerging on how to deal with them--but only once the confusion is cleared up surrounding private common prop- erty and formal and informal rights. The second part addresses redistributive land reform--the redistrib- ution of property rights in land from large to small farmers. A height- ened sense of urgency surrounds the need to address land redistrib- ution, especially in the former settler colonies in southern Africa, but controversy exists regarding the appropriate implementation mechanisms. The study highlights the case of South Africa, because success there would have tremendous regional and international implications for land redistribution. A policy framework for redistrib- utive land reform is outlined, within which the competing paradigms compete where it actually matters--on the ground. World Bank Working Papers are available individually or by subscrip- tion, both in print and online. ISBN 0-8213-6440-5 THE WORLD BANK 1818 H Street, NW Washington, DC 20433 USA Telephone: 202 473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org