Questions about land markets are central to development policy, as underlined recently in the 2008 World Development Report (World Bank Reference Peters and Kambewa2007: 9). In particular, markets in which land rights are transferred on a temporary or permanent basis are seen as playing a key role in the redistribution of land to more efficient users, thus increasing productivity and employment. In this view, any negative effects attributed to land markets, notably in terms of inequity and landlessness, result from failures in other markets, especially credit and insurance markets (Deininger and Feder Reference Deininger, Gardner and Rausser2001; de Janvry et al. Reference De Janvry, Gordillo, Platteau and Sadoulet2001; World Bank Reference Mathieu, Benjaminsen and Lund2003). In Africa, until late in the twentieth century it was these negative effects, and a perception of land as being relatively abundant due to low population densities in many parts of the continent, that influenced policy makers’ views. Consequently, land markets received relatively little attention in development policy, except, as exemplified in an earlier issue of this journal (Shipton and Goheen Reference Shipton and Goheen1992), as a means of securing credit flows for agricultural development. Over the past two decades, however, a wave of proposals for land tenure reform in many African countries (Toulmin and Quan Reference Quan, Toulmin and Quan2000; Berry Reference Berry2002; World Bank Reference Mathieu, Benjaminsen and Lund2003) has raised questions about land markets as a means of allocating land that have profound political and economic implications, rarely addressed by previous research. This collection of articles provides an opportunity to explore the nature of land markets in Africa.¹

EMERGENCE AND DYNAMICS OF LAND MARKETS IN AFRICA

In African contexts, the question of land markets is commonly framed by an analysis of how customary systems of land tenure have moved, or are supposed to be moving, towards private property. This transformation is typically attributed to the combined effects of demographic growth, development of cash crops and changes in cropping systems (development of tree crop plantations, shorter fallow periods, and the disappearance of ‘shifting cultivation’). As a consequence, it is anticipated that land values increase and land rights become increasingly identified with individuals, so that the ‘bundle’ of different rights relative to land, including the right to transfer, becomes concentrated in the hands of a single right holder. This then translates into increasingly monetized access to land through sales and rental (including sharecropping) (Boserup Reference Boserup1965; Johnson Reference Johnson1972; Ault and Rutman Reference Ault and Rutman1979; Feder and Noronha Reference Feder and Noronha1987; Migot-Adholla et al. 1991; Bassett and Crummey Reference Bassett and Crummey1993; Platteau Reference Platteau1996; Deininger and Feder Reference Deininger, Gardner and Rausser2001). This ‘evolutionary’ perspective on land markets is underpinned by the ‘theory of induced innovation’ (Hayami and Ruttan Reference Hayami and Ruttan1985), which draws upon economic models that understand institutional change as arising when existing institutions are incompatible with shifts in relative prices and profitability driven by changes in relative factor (labour, land, capital) scarcity, or technological innovation (Demsetz Reference Demsetz1967; Johnson Reference Johnson1972; Ault and Rutman Reference Ault and Rutman1979; Feder and Noronha Reference Feder and Noronha1987).

In recent decades, a growing number of authors have drawn on empirical studies to argue that market transfers of land have indeed become more common in Africa (Biebuyck Reference Biebuyck1963a, Reference Brebuyck and Biebuyck1963b; Feder and Noronha Reference Feder and Noronha1987; Barrows and Roth Reference Barrows and Roth1989; Bassett and Crummey Reference Bassett and Crummey1993; Bruce and Migot-Adholla Reference Bruce and Migot-Adholla1994; Ensminger Reference Ensminger, Drobak and Nye1997; Sjaastad Reference Sjaastad2003; Chauveau et al. Reference Chauveau, Kuba and Lentz2006; Chimhowu and Woodhouse Reference Chauveau, Kuba and Lentz2006; Colin and Ayouz Reference Colin and Ayouz2006; Holden et al. Reference Holden, Otsuka and Place2009). However – undermining somewhat the ‘evolutionary’ view of land markets outlined above – historical accounts make clear that land sales were occurring in the nineteenth century in Ghana (Hill Reference Hill1963), and in South Africa (Bundy Reference Bruce and Migot-Adholla1979; Lambert Reference Lambert1999). Similarly, Berry (Reference Berry, Downs and Reyna1993) refers to pre-colonial land sales in Kikuyuland, Kenya, and other references to pre-colonial land sales in both West and East Africa are found in Feder and Noronha (Reference Falloux, Davis and Schirmer1987), Bassett (Reference Bassett and Crummey1993), Ensminger (Reference Ensminger, Drobak and Nye1997), and Kuba and Lentz (Reference Lentz, Kuba and Lentz2006).

Empirical studies of land use also reveal widespread incidence of a variety of rental and sharecropping arrangements whereby mobile labour negotiated access to land, often in order to exploit opportunities for cash crop production. Examples are found in Robertson (1987), for Senegal, Ghana, Sudan and Lesotho, in Lawry (1993) for Lesotho, in Migot-Adholla, Benneh et al. (1994) and Migot-Adholla, Place et al. (1994) for Ghana and Kenya, in Amanor and Diderutuah (2001) for Ghana, in Lyne et al. (1994) for Uganda and South Africa, in Paré (2001) for Burkina Faso, in Edja (2001) for Benin, in Koné (2002) and Colin (2004) for Côte d'Ivoire, and in Holden et al. (2009) for Ethiopia, Kenya, Malawi and Uganda. More general accounts of African land rental markets are found in Lavigne Delville et al. (2001). The recent sharp rise in oil and agricultural prices has prompted interest from international investment funds in large-scale land acquisition in developing countries, including through sale or lease markets, in order to expand bio-fuel and food crop production (GRAIN 2008; Cotula et al. 2009). This has given rise to concern about possible impacts on land access for small-scale producers in these countries. However, there is not yet firm empirical evidence as to whether such investments will materialize and even less of their consequences for land markets in Africa.

Since all but 10 per cent of land in Africa is considered to be occupied under customary land tenure (World Bank 2003) – that is, without land titles formally registered in the names of individuals – the vast majority of sale and rental transactions are ‘informal’: without protection in the statutory legal system. Where governments have declared the sale of land held under customary tenure to be illegal, transactions in informal land markets may be disguised as ‘pledges’ or ‘loans’ (Feder and Noronha Reference Feder and Noronha1987; André and Platteau Reference André and Platteau1998; Quan Reference Quan, Toulmin and Quan2000). This characteristic of ‘informality’ in many land markets in Africa is associated with questions about the precise nature and content of transactions. Thus, while land sales are often reported in historical accounts as ‘outright sales’, typically from ‘autochthon’ land authorities to immigrants (for example, Hill 1963), they appear in more recent studies to be subject to contestation by other interpretations, with the ‘seller’ – or more often the seller's kin or heirs – disputing the extent to which their continuing interest in the land was extinguished by its ‘sale’ in the past. Such ‘sales’ may, for example, be interpreted by autochthons as establishing a long-term client–patron relationship between land ‘purchaser’ and ‘seller’ respectively, as exemplified by the tutorat in Côte d'Ivoire (Chauveau Reference Chauveau, Kuba and Lentz2006). It is perhaps significant that historical cases interpreted as ‘outright sales’ generally related to uncleared, and hence largely uninhabited, forest for which notions of ownership and control were substantially undefined. In such circumstances, a transaction could be seen as much in terms of establishing the rights of the seller to dispose of the land (hence also a preference for land disposals at potentially disputed boundaries with neighbouring authorities) as those of the purchaser to acquire it (Dozon Reference Dozon1985). Disputes over such sales of (undeveloped) forest land centre on the nature of the rights purchased, and the legitimacy of their transfer. As the case of Côte d'Ivoire exemplifies, questions may arise as to whether sales of forest relate only to the right to plant crops (cocoa or oil palm, for example), or to a complete set of rights to use and dispose of the land to others. Disagreements may also occur over whether acquisition of the land entails any enduring obligations on the purchaser, the extent to which the ‘sale’ of land is accepted within local custom, and, if so, who has the authority to decide on such sales. By contrast, sale and purchase of established plantations occasion few disputes, except in so far as they remain subject to continuing ‘autochthon’ contestation of authority to transfer the land to new users – as in the case of the tutorat (Colin and Ayouz Reference Colin and Ayouz2006).

The literature provides accounts of increasing autonomy of decision by individuals (Bruce Reference Bruce, Bassett and Crummey1993) as well as continuing contestation over the nature of land transactions (Berry Reference Berry2002; Lentz Reference Chauveau, Colin, Jacob, Delville and Meur2006). In Burkina Faso, Mathieu et al. (2003) and Hagberg (Reference Hagberg, Kuba and Lentz2006) have reported efforts to disguise market transactions in land held under customary tenure, for example as ‘loans’ and through the observance of customary ritual, but also efforts to formalize transactions through recourse to written records or witnesses, or both. In Malawi, Peters and Kambewa (2007) found that increasing land sales reflect land supply and demand relationships among land-using households, but that it is not uncommon for sales to be subsequently disputed by the seller's kin on the grounds that they were not consulted on, or did not agree with, the original transaction. In sum, therefore, the evidence from different parts of Africa indicates that sales of land are becoming more common, but that disputes over the meaning of such transactions are also increasing.

For some, the tensions inherent in these dynamics are interpreted as evidence that customary land rights are a barrier to the emergence of private property in land and its transfer via a land market, and therefore to investment in commercial agriculture (Falloux Reference Falloux, Davis and Schirmer1987). Others emphasize the importance of assisting spontaneous moves towards the required transition through conversion of customary rights to a formal register of individually owned land titles (Barrows and Roth Reference Barrows and Roth1989). In each case the state is perceived as key to the reform of customary land rights, either by legislating and administering their substitution by land titles (as in Kenya) or by facilitating and formalizing their ‘endogenous’ evolution. In this perspective, government policy that restricts such an evolution towards private property in land constitutes a barrier to agricultural development in Africa (Feder and Noronha Reference Feder and Noronha1987). Such views, which remain influential in current policy discussions (World Bank Reference Peters and Kambewa2007: 9), were first formulated in the later (post-1930) colonial period when a drive to develop export-oriented commercial agriculture in Africa led colonial authorities to abandon their earlier prohibition of land markets among Africans (Chanock Reference Chanock, Mann and Roberts1991; Cowen and Shenton Reference Cowen and Shenton1991) and to promote the adjudication and registration of individual land titles in the belief that this would provide security for credit and investment and encourage land transfers to the most efficient producers (Coldham Reference Coldham1978; Bassett Reference Bassett, Bassett and Crummey1993). In the event, land markets in Africa have proliferated, as we observed above, largely (with the notable exception of Kenya) in the absence of land titling.

The impacts of land sales leave open to question the expected gains in efficiency. In particular, land purchases have often obeyed logics other than those of agricultural efficiency. Purchasers may draw on earnings from non-agricultural, often urban-based, activities and acquire land for speculative or prestige reasons (Haugerud Reference Haugerud1983; Platteau Reference Platteau1996). Poorly managed agriculture and absentee land owners may therefore be associated with areas of active land markets in Africa (Atwood Reference Atwood1990; Bruce Reference Bruce, Bassett and Crummey1993). Equally, land sales may reflect sellers’ short-term financial distress, rather than any intention of investing the proceeds in more productive activity (Platteau Reference Platteau1996).

In this regard, the operation of a market for the temporary transfer of rights (rental or sharecropping) may offer better prospects for achieving greater efficiency in land use (Lawry Reference Lawry, Bassett and Crummey1993; Blarel Reference Blarel, Bruce and Migot-Adholla1994; Lyne et al. 1994; Kevane Reference Kevane1997; Lavigne Delville et al. Reference Deininger, Gardner and Rausser2001; World Bank Reference Mathieu, Benjaminsen and Lund2003; Holden et al. Reference Cotula, Vermeulen, Leonard and Keeley2009). Assessing the impacts of such markets is not straightforward, however, because the wide diversity of forms of land rental and sharecropping transactions, and the consequent separation of ‘ownership’ and ‘use’ of land, may obscure the extent of land markets and their effects on concentration of control over land (Berry Reference Berry, Downs and Reyna1988, Reference Berry2002). Although historical examples exist of ‘reverse tenancy’ by agricultural entrepreneurs ‘renting in’ land for large-scale farming (Robertson Reference Robertson1987 for Lesotho), such patterns are relatively rare and land concentration remains primarily associated with ownership, rather than use (with implications for scale and investment in tenant agriculture). As a consequence, rental markets tend to be associated with lack of concentration of land use. This may change, however, should proposed foreign investments in land for large-scale agriculture in Africa, noted above, materialize, since many of these specify leasehold tenure for what would become very large landholdings.

Until now, increasing inequality has been primarily associated with land transfers through sales, as documented by André and Platteau (1998) in Rwanda and Murton (1999) in Kenya. In both instances, income from non-agricultural activity was a factor distinguishing those able to invest in increased agricultural productivity and to purchase land. Equally, in both cases distress sales by households unable to sustain land productivity were a factor in supplying a land market. Empirical evidence from other African contexts suggests more equivocal distributional effects, with land markets reported as having little effect on land concentration in Côte d'Ivoire (Colin and Ayouz Reference Colin and Ayouz2006), Kenya (Holden et al. Reference Cotula, Vermeulen, Leonard and Keeley2009) and Uganda (Baland et al. Reference Baland, Gaspart, Platteau and Place2007; Holden et al. Reference Holden, Otsuka and Place2009). More generally, while land markets may improve security of access to land for individuals with limited customary rights – notably women (André and Platteau Reference André and Platteau1998; Daley Reference Daley2005) – commoditization may favour those with purchasing power (Chimhowu and Woodhouse Reference Chimhowu and Woodhouse2006). However, such outcomes will depend on the interplay between the commoditization of land and the exercise of customary authority, since those with greatest influence over land under customary tenure will be best placed to gain, as sellers or renters-out of land, from a land market (Bernstein and Woodhouse Reference Bernstein and Woodhouse2001; Kasanga and Kotey Reference Kasanga and Kotey2001; Woodhouse Reference Woodhouse2003).

The tensions sketched in the above discussion explain the widespread contemporary concern with insecurity of land rights in Africa. This is not simply a question of ‘institutional transition’ but also one of the broader political context and its role in framing debates about property, identity, and individual and collective rights:

In African contexts, Peters (2004: 301) has described the present conjuncture as one of ‘social turmoil over deepening social differentiation, particularly with regard to land and landed resources but also, necessarily, to questions of legitimate authority within and between key social units of families, lineages, villages, “ethnic groups” and so forth’. An important discourse that surfaces from this turmoil relates to the perceived fairness or justice of contemporary land access in general, and land markets in particular (Platteau Reference Platteau1996: 55–6), in which questions of autochthony and the linkage of land rights to ethnicity are prominent, and in which the role of the state (as suggested by the quotation above from Platteau) is important.

SOCIALLY AND POLITICALLY EMBEDDED LAND MARKETS

Thus far, it seems clear that simple evolutionary models are inadequate to the task of analysing the significance of land markets in Africa. Yet empirical study of existing land markets, on which alternative models must rely, remains sparse. Moreover, most published work focusing specifically on these markets tends not to explore the content of land transactions (the rights and duties exchanged), but concentrates analysis on factors associated with market development (especially increasing land scarcity) and on its efficiency and equity implications. To the extent that such work addresses the social relationships embodied in markets, these tend to be restricted to narrowly defined (quantitative) parameters of equity (in terms of land concentration and exclusion), rather than analysing the social and political content and consequences that such parameters may reflect. This special issue of Africa, concerned with land transactions in rural areas of West and Southern Africa, contributes to addressing this gap. The articles consider transactions that transfer ‘ownership’ as well as markets for temporary access rights, such as rents or sharecropping contracts. Moreover, our focus is on the meaning and content of land transfers, rather than their effects on land use efficiency or on the social distribution of land holdings. In the remainder of this introduction we identify the principal themes that emerge from these articles.

In the first place, the articles by Amanor, Lentz, Chimhowu and Woodhouse, and Chauveau and Colin underline the historic role of market transactions in pre-colonial and colonial customary land tenure, a point worth stressing given the frequency with which African land rights are referred to as ‘inalienable’ in policy debates. The roots of the non-market ideal of customary land tenure in Africa have been traced back to the colonial administrations of the early twentieth century by, among others, Colson (1971), Chanock (1991) and Berry (1992), but its persistence in contemporary political discourse suggests a need to recall Mair's (1971 [1948]) observation that the undoubted existence of sacred ritual concerning land should not obscure a primary reason for the ‘inalienability’ of land, which is the absence of sufficient economic incentive, rather than cultural conservatism. The political persistence of a non-market discourse of ‘customary’ land rights is evident in all the articles in this volume, in the ‘illegal’ status of market transactions in customary land. In the cases of dualistic tenure regimes, in Zimbabwe and Zambia, exclusion from a land market is a principal measure by which ‘customary’ land is defined as distinct from ‘commercial’ land allocated through formal markets. In the markets for customary land, designated variously as ‘informal’, ‘clandestine’ or ‘vernacular’ markets by authors of these articles, there is nonetheless often recourse to ‘informal formalization’ to make the transaction more solid.

In all the cases considered in this special issue, the commoditization of land access emerges in newly settled ‘pioneer’ areas at the land frontier, or is closely related to arrival of ‘outsiders’, migrants from within or outside national frontiers. The development of market relations over land is thus primarily among pioneers or their heirs, or between ‘autochthon’ landholders and migrants seeking land – in other words concerning ‘rights to settle’ – but only rarely between autochthons. We are aware that this raises questions about the stage at which ‘pioneer’ settler populations become ‘autochthon’ with respect to later-arriving settlers. More significantly, however, there is no support from these cases for the notion of a progression from market exchange of land among individuals within communities towards exchange with outsiders. Similarly, there is little clear evidence to support an ‘evolutionary’ model of land market development as a sole response to demographic pressure, since in both Ghana (Amanor) and Côte d'Ivoire (Chauveau and Colin) the purchase of land for cash was taking place in sparsely populated areas (for the express purpose of commercial agricultural development). Moreover, and associated with this sale of ‘undeveloped’ land, the commoditization of such transactions – not necessarily expressed in monetary terms (Bohannan and Dalton Reference Bohannan and Dalton1962; Parry and Bloch Reference Parry and Bloch1989) – appears to proceed independently of the establishment of ‘full’ ownership rights by the seller.

In all the cases presented here, except those of Amanor, Kouamé and Sitko, the state, both colonial and post-independence, is visible as a brake on the dynamics of market formation, mobilizing political discourses in which ‘inalienable’ land rights are linked to notions of citizenship and ethnicity. This political dimension is played out at a more local level, particularly in the cases presented by Lentz and by Chauveau and Colin, as competing claims to land seek justification through the mobilization of different versions of history and tradition: ‘I bought the land’ versus ‘the land cannot be sold’. The malleability of the ‘customary’ is a central theme of the Zambian case presented by Sitko, one linked to shifting social relationships that underlie differential control over land, now (re)interpreted in terms of customary rights to buy and sell land. In Chauveau and Colin's article divergent interpretations of the nature of land transactions reflect not just a difference about the record of an event, but a more fundamental question about the separation of the transfer of land rights from the social and political dimensions of the transaction. Thus, the commoditized character of access to land is conditioned by enduring social and political relationships. In a number of the West African cases (Amanor, Chauveau and Colin, Kouamé) the contestation of market transactions centres on the rights of individuals to ‘family land’ controlled by the broader corporate lineage. Intergenerational struggles, in particular, focus not only on rights of the ‘youth’ to use land, but also on the control of income generated from family land rented or sharecropped by non-family ‘outsiders’. In this sense the rental market is both ‘inclusive’ in broadening access to land to those without customary rights, while at the same time ‘exclusive’ in its effects on individuals (women, youth) with weaker claims on land held by their lineage. Amanor, in particular, sees here an erosion of customary rights derived from kinship in favour of a more generalized recourse to the (rental) market in order to gain access to land. In Sitko's study in southern Zambia, shifting interpretations of customary rights, and their relationship to land commoditization, have a more marked gender dimension. Here, the subordination of women's customary land rights to those of their husbands and sons becomes magnified when translated onto the terrain of the informal market for customary land. Women's access to land may thus be affected by land sales in this market, but they themselves are unable to participate in it, even as sellers of the land they farm.

While rental markets may generate intra-family tensions concerning their legitimacy – with those in Lower Côte d'Ivoire (Kouamé) eventually finding an outlet in inter-ethnic conflict – the ‘content’ of rental and sharecropping transactions appears largely uncontested. By contrast, the meaning of land ‘sales’ in these ‘vernacular’ land markets is understood in a variety of ways in these articles. For Amanor, for Chimhowu and Woodhouse and for Sitko, land sales are ‘outright’ sales in which the seller accepts the cessation of all further interest in the land in exchange for payment, whereas Lentz focuses on contemporary disagreement about the nature of past land sales. Chauveau and Colin's study encompasses cases of both ‘outright’ (uncontested) sales and also ‘incomplete’ sales. The latter are characterized not only by a partial transfer of rights, but also by conditions of patronage (tutorat) through which the ‘foreign’ land purchaser is incorporated into the local community. The often covert nature of these transactions (not least to hide them from family members with potential interests in the land) contributes to a lack of transparency as to the legitimacy of the seller and of the rights being transferred through the ‘sale’. As a consequence, contrary to what might have been the purchaser's expectations at the time of the initial transaction, it may transpire that payment does not conclude the relationship with the vendor, but instigates a continuing obligation to recognize (generally through further payments) the role of the vendor as tuteur. From the perspective of the tuteur, characterization of the transaction as a ‘sale’ refers rather to the monetarization of a token acknowledgement – albeit one subject to redefinition according to changing conditions – of social and political obligations of the land purchaser towards the ‘seller’ and/or the host community. This raises the question of whether such transactions constitute a land market. At a basic level, these transactions appear to conform to criteria of markets, in which rights over specific goods that have both use value and exchange value are transferred on the basis of a price (though one that is not necessarily monetized) that reflects a calculation of costs and benefits. However, these ‘vernacular’ land markets clearly do not conform to classical economic conceptions of market prices as unaffected by social relationships, and market exchange as occurring between mutually independent actors, whose independence remains unchanged at the close of the transaction. Rather, these markets exhibit a strong influence of social relationships in shaping their logic and outcome (Granovetter Reference Granovetter1985).

In this sense, African land markets are strongly embedded in contemporary African politics and society. This manifests, in all the articles in this special issue, not a simple evolution of customary rights towards greater efficiency in response to forces of supply and demand, but relations of exchange between ‘natives/autochthons’ and ‘settlers/migrants/foreigners’, or between earlier and later settlers in the Zambian case and in lower Côte d'Ivoire. The repercussions of these relations, for both sets of ‘communities’, are mediated by intra-family tensions over land access and by state and/or local politics of citizenship and identity. It is evident that the conflicts inherent in these political dynamics are exacerbated by the dynamics of supply and demand in land markets, which in many parts of Africa are influenced by both the closure of the land frontier, restricting the availability of land for new settlement, and also the stagnation or decline in the non-farm economy (most dramatically in the case of Zimbabwe) that has prompted the urban unemployed to seek land in rural areas.

The broader significance of this set of articles may be identified in their contribution to a more critical analysis of the forms which institutional change may take, and, particularly, of the institutional processes through which commoditization is embedded in society (including in its intra-kinship dimensions) and politics. Beyond the undeniable influence of increasing land scarcity, the concrete forms and consequences of land transactions can only be understood from a perspective that incorporates the macro- and micro-political evolution of African societies. We believe these studies mark a step forward in an African land policy debate that remains largely polarized between those seeking poverty reduction through promoting land markets and their opponents who seek to protect the poor from losing their land to the market (Chimhowu and Woodhouse 2006). In particular, they suggest that security of land rights is neither a simple question of formalization via issuing land titles, nor of preserving ‘communal’ tenure. Further, the delivery of efficient and equitable land use via land rental markets as frequently advocated today may confront unresolved conflicts over the rights to lease out land held under customary tenure. Above all, these studies suggest that an understanding of actually existing land markets is a key element in the analysis of the roots of conflict and insecurity in Africa today.