Push, pull and push-back to land certification: regional dynamics in pilot certification projects in Côte d'Ivoire

Abstract

Since 2000, many African countries have adopted land tenure reforms that aim at comprehensive land registration (or certification) and titling. Much work in political science and in the advocacy literature identifies recipients of land certificates or titles as ‘programme beneficiaries’, and political scientists have modelled titling programmes as a form of distributive politics. In practice, however, rural land registration programmes are often divisive and difficult to implement. This paper tackles the apparent puzzle of friction around rural land certification. We study Côte d'Ivoire's rocky history of land certification from 2004 to 2017 to identify political economy variables that may give rise to heterogeneous and even conflicting preferences around certification. Regional inequalities, social inequalities, and regional variation in pre-existing land tenure institutions are factors that help account for friction or even resistance around land titling, and thus the difficult politics that may arise around land tenure reform. Land certification is not a public good or a private good for everyone.

INTRODUCTION

Since 2000, there has been a renewed push within African countries for comprehensive land registration and, in many cases, land titling. Most of the new laws and registration programmes aim at assignment of rights to individuals and land commodification. Some examples are the effort announced in Kenya in 2013 to register and title all land within a decade, Côte d'Ivoire's law of 1998 (updated in 2011) which aims to register all land by 2023, Uganda's 1998 comprehensive land registration law, and the Burkina Faso 2012 and Malawi 2016 land policies, which also aim at comprehensive registration. In much of the donor community and among political scientists writing on land titling in Africa, registration and titling is defined as a universal good – i.e. as both a public good that will reduce conflict and promote investment in land, and as a pareto optimal change that creates only beneficiaries at the individual level.¹ Theories of ‘endogenous evolution of land rights’ suggest that privatisation will be a smooth and incremental process driven by incentives arising from within the rural economy and society, and thus require little ‘push’ or coercion from government.² Much research in the advocacy literature and the political science literature on titling casts recipients of land-rights formalisation as ‘programme beneficiaries’, and political scientists are modelling land titling programmes as a form of distributive politics by which politically favoured rural communities are the priority targets.³ In practice, however, land law reform rarely follows this smooth trajectory.

Land tenure reform proposals have been divisive, hard to develop, and difficult to implement in many African countries. As Ali et al. (2014a) wrote in a World Bank policy research paper, ‘During the 1990s and 2000s, nearly two dozen African countries proposed de jure land [law] reforms which extended access to formal, freehold land tenure to millions of poor households. Many of these reforms stalled.'⁴ Deininger et al. (2014: 79) write that implementation has been limited or non-existent almost everywhere. In Côte d'Ivoire, 20 years after the passage of the new land law, only about 4,000 land certificates, or less than 1% of estimated (and presumed) total demand, have been delivered. Land registration proponents, including the World Bank and the European Union, chafe at persistent obstacles to faster progress.⁵

This paper tackles the apparent puzzle of friction around land tenure reform. Existing explanations in the policy literature for the stalling or go-slow on land registration programmes in Africa tend to focus on ‘supply side’ factors related to project implementation. These include the high costs of registration and titling, challenges of communication and information dissemination, and low government capacity.⁶ Other explanations for slow progress focus on the ‘demand side’. The most common of these is that landholders believe that their land is secure and are therefore indifferent to titling. Yet much experience shows that many smallholders do face land insecurity, and are eager to obtain some form of formal state recognition of their land rights. A key demand-side explanation for slow uptake of land registration is the high costs of land certification that are imposed on the landholders themselves: farmers may want titles, but are unwilling or unable to cover the costs. It remains difficult to gain traction on the demand side of the registration and titling question, however, especially given the challenge of interpreting ‘non-responses’ to registration opportunities. As matters stand, pervasive friction in land rights formalisation programmes remains poorly understood, and only partially and unsystematically explored.

This paper advances knowledge in this field by probing the demand side of the registration equation. We analysed national pilot projects that registered land rights at virtually no cost to landholders in selected project zones across southern Côte d'Ivoire. This offered unique possibilities for gaining analytic traction on the issue of smallholders’ preferences for registration and titling. With the financial cost to smallholders near zero, the most common explanation for active smallholder reticence was largely neutralised.⁷ Observation within a single country allowed us to compare registration across subnational regions while holding the state's administrative capacity relatively constant across space.

Comparison across three spatially dispersed pilot zones revealed two demand-side phenomena that are unanticipated or unsystematically observed in the existing literature that advocates for titling in African countries. The first is geographic and social unevenness in small-scale landholder demand for land registration. The second is considerable political friction in the registration process, arising from local ambivalence and cross-purposes around the assignment of land rights.

We argue that explaining these patterns of friction and push-back around land registration requires a focus on spatially variant features of local political economies that are often overlooked or downplayed in the policy and advocacy literature. Political economy variables that may drive variation in local preferences for land registration are regional inequalities, regionally variant patterns of social inequality, and regional variation in existing land tenure institutions. Our empirical material on pilot land registration in Côte d'Ivoire traces the impact of these variables on local demand for registration, and does so in a way that reveals connections between regional agrarian issues and national politics.

The first section juxtaposes the mainstream policy literature's implicit assumption about homogeneity in positive rural demand for land registration and titling with hypotheses about why such demand may actually be uneven and socially fraught across space and social categories. The second section describes the Côte d'Ivoire land registration pilot programmes that were implemented in localities across the southern forest zone. Our study focused on pilot sous-prefectures in three regions. We describe our study sites, study design and research method. The third section focuses on each of the study areas, showing that hypotheses laid out in the first section offer leverage in revealing and accounting for tensions across localities on the ‘demand side’ of the land registration equation. In Côte d'Ivoire, these regionally specific differences have contributed to social tensions that have restrained land registration at the local level, hampered programme implementation nationwide, and played a part in the on-going salience and politicisation of land issues at the national level.

DEMAND FOR LAND RIGHTS FORMALISATION: THE IMPORTANCE OF ANALYSING INEQUALITIES, INTERESTS, AND INSTITUTIONS

Smallholder land registration involves formal parcel delimination and assignment of ownership rights to a named owner (an individual, family, or corporate entitity such as a legally registered community). This is a critical step in land titling processes which aim inter alia to formalise and legalise land sales and land mortgaging, and to make such transactions legally enforceable. The pro-titling policy literature, epitomised by the World Bank advocacy of titling (World Bank 2009), does not explicitly recognise the existence of demand-side dynamics that complicate national land registration and titling programmes from the bottom-up. (See German & Braga 2019 for a review and critique.) This work assumes that land registration and titling is Pareto optimal in the long-run (that is, it leaves all those affected better off, and no one worse off), and does not anticipate smallholder ambivalence or resistance to titling. Much of the smallholder advocacy literature also suggests that all farmers will benefit from formalisation, since land titles will provide greater security of tenure and in some situations, allow farmers to access bank credit. There is recognition in some policy-advocacy literature that as a practical matter, the benefits of registration may be spread unevenly within families, and that under certain conditions (e.g. heavy reliance on a shared ‘commons’ for grazing) individual and nuclear-family registration may impose net costs on communities (Deininger et al. 2014: 78). However this is usually treated as a matter of negative externalities or outer limits on the scope of policy applicability, rather than in terms of uneven ‘demand’ for, or acceptance of, registration projects within or across rural localities, or in terms of controversy over registration and titling in national policy debates.

Recent political science studies do consider politics explicitly. These treat land-rights formalisation as a form of distributive politics (i.e. titles are a gift of the state that are likely to be targeted at regime supporters or would-be supporters). This work echoes assumptions about Pareto optimality, stressing that titling increases property owners’ autonomy vis-à-vis the state and the value of their land asset (Albertus 2015; Hassan & Klaus 2019).

These approaches share the tendency to adopt a space-blind view of ‘the rural majority'; to overlook heterogeneity and socio-economic inequality within rural societies; and to see the existing land tenure institutions as stable, and not politically contested or politicised. From this perspective, it is difficult to account for much of the controversy and political tension observed around land rights formalisation programmes.

We argue here that to explain observed friction and controversy around formalisation programmes from the demand side of the equation, a different starting point is required. It is necessary to set aside assumptions about the Pareto optimality of reform, and to focus instead on tensions arising from regional inequalities, conflicts of interest within rural society, and variation and bias in pre-existing land tenure institutions. Existing literatures engaging these themes in the more critical literature on land titling, as well as in the wider literature on the political economy of land tenure in Africa, offer hypotheses about where grassroots tensions may lie, and how these may connect to national-level political currents.

Territorial politics

Regional inequality and histories of the shifting and sometimes rivalrous and coalitional bases of national governments have created differentiated rural areas deeply marked by past and on-going state intervention. Legacies of regional competition and ‘geopolitical logics’ of rule have biased both land policy and its distributive impact in rural settings. Such considerations strongly inform existing work on land politics and policy in many African contexts (Berman & Lonsdale 1992; Klopp 2002; Boone 2003, 2014; Colin et al. 2009; Lavigne-Delville 2009; Gay 2014; Honig 2017; Klaus 2017; Lavers 2017; Bluwstein et al. 2018; Fatema 2019; Lavigne-Delville & Moalic 2019), including work on Côte d'Ivoire (Dozon 1985; Affo Yapi 1987; Chauveau 2000, 2003, 2009; Koné 2006a, 2006b; Ruf 2012, 2013; Colin 2013; Kouamé 2016; Tarrouth & Colin 2016; Montaz 2015, 2017). The current study takes this large body of literature as a starting point. A basic hypothesis that emerges from this work is that there will be variation both across and within subnational regions in the strength of demand for land rights formalisation, and in the type of land certification that actors may prefer or demand. For example, smallholders who are politically aligned with the incumbent national government, and who reside in regions that are also so aligned, may view formalisation initiatives more favourably than smallholders in an oppositional (distrustful) position vis-à-vis the state or regime.

Conflicts of interest within rural society

Agrarian studies literature shows that degrees and types of social and political hierarchy vary across localities and regions. These hierarchies tend to be based upon or reflected in differential distribution of land rights within localities and families. This suggests the hypothesis that land rights formalisation will solidify the positions of the wealthier and more politically and socially powerful actors within rural regions, communities and families. Case study research from a range of African countries provides much support for this hypothesis (Cotula et al. 2004; Manji 2012; Chitonge et al. 2017; Stein & Cunningham 2017), and scholars anticipated and indeed documented such trends in parts of Côte d'Ivoire in the 1990s and 2000s (Chaveau 2000, 2003: 353; Tarrouth & Colin 2016). Inequality within rural society is thus expected to be a source of variation in the actual or anticipated benefits of registration and titling, and thus a source of difference in ‘preferences for titling’ within rural society, with those posed to gain from titling likely to be the strongest source of demand for it.

Variation and bias in existing land tenure institutions

Closer analysis of existing land tenure institutions is also necessary to understand demand for land registration, and how this may vary across space and social categories. Policy advocates for registration and titling often suggest that this involves ‘simply conversion of de facto to de jure land rights’,⁸ or a ‘snapshot of existing land rights’. More historically attuned and fine-grained approaches often reveal the ways in which rights formalisation overwrites existing institutional matrices of land rights and entitlements (which are themselves sometimes disputed), some of which have been previously recognised by the state. Formalisation is thus shown to transform ways of legitimately claiming and holding land (Coldham 1978; Shipton 1992; Boone 2014; Chitonge et al. 2017). This suggests that departures from the status quo ante will create a politics around winners and losers of older rights and protections, and possible grounds for resistance or contestation, which will be strongly conditioned by the particular type of land registration and titling on offer.

Assumptions about uniformity, equality and institutional void or equilibrium in rural society underpin the implicit suggestion in much of the policy literature (and in the political science literature that defines all individuals targeted by reform as beneficiaries) that almost all rural landholders would prefer whatever form of land rights formalisation is on offer by the state. In contrast, approaches that foreground spatial and social inequalities and pre-existing institutional heterogeneity, as outlined in the literature referred to above, predict variation in both the level and the type of land-rights formalisation that different rural groups and actors would prefer. There are many possible dimensions along which rights-assignment and formalisation can be structured, including: individual versus collective registration (which individuals? which collectivities?); restricted or unrestricted transactability of title; local versus more centralised keeping of registries, rights enforcement, and conflict adjudication; short or long duration of contracting on the basis of different kinds of registration.

The literature reviewed above generates hypotheses about how and why preferences for land registration – and over the institutional design of such programmes – are likely to vary both within localities and across regions of smallholder agriculture. There may be multiple sources of bottom-up friction in, or even resistance to, smooth implementation of land rights formalisation programmes, as well as many potential sources of tension at the national level over programme scope and design. The analysis below presents evidence of these dynamics in Côte d'Ivoire that is based upon our study of the national land registration pilot projects that ran from 2011 to 2016. We show that divergent ‘preferences for registration’ generate friction in land rights formalisation programmes both within and across regions, scaling up to connect to political and policy debates at the national level.

CÔTE D'IVOIRE LAND RIGHTS FORMALISATION: FROM THE 1998 LAW TO PILOT PROGRAMMES

A large body of work on Côte d'Ivoire traces regional variation in land politics, and in the politicisation of both land law and land rights since the 1940s. This documented history of agrarian and regional tensions is the context in which the current land registration and titling effort is playing out. The rapid economic development of the country from the 1950s to the 1970s was driven in large part (but not exclusively) by state-orchestrated land pioneering that drove expansion of export crop production across the southern forest zone. Ivoirian land policy offered uneven recognition to the customary land rights of indigenous populations throughout the country. These were robustly upheld in the north and east, but compromised and encroached upon in the west, in the zones of large-scale in-migration. Grievances over land rights encroachment and a sense of regional injustice developed in the centre-west and far west. This found expression in militant and oppositional politics in the 1970s and the 1990s.

In 1998, in a context of high national political tension around issues of land rights and citizenship rights, the Ivoirian National Assembly passed a new land law, Law 98–750 on the Rural Domain. It was built around seemingly contradictory commitments to both autochthony-affirming principles (affirming the ability of village-level traditional authorities to adjudicate land rights, protecting autochthonous land rights, and denying non-Ivoirians the right to own land) and to pro-individualisation, pro-privatisation and market-building principles (requiring that all land certificates be transformed into individual and fully transactable land titles within three years of registration). The law declared that all rural land in Côte d'Ivoire would be either registered and individually titled, or immatriculated in the name of the state, within 10 years.

National elections in 2000 brought to power Laurent Gbagbo, a champion of the land rights of the indigenous people of the west. During his decade in office (2000–2011), Gbagbo undertook to deliver on the protective and autochthony-affirming promises of the new land law, especially in his core regions of political support in the west. With donor funding, the government introduced a village demarcation programme to lay the foundation for certification and titling (Boone 2018). In the regions and villages, pro-Gbagbo elites reinforced the insistence on autochthones’ claims to land that had been ceded (or lent) to in-migrants in the past.

Target villages for these early programmes were selected strategically, reflecting the Gbagbo regime's interest in delivering more secure land rights to supporters in its electoral strongholds, and project designers’ interest in building regional variation into the pilot project (for learning purposes and also to defuse the appearance of a partisan regional bias in programme implementation). These efforts, financed in part through a European Union debt swap programme (Devis-Programme, or DP) called DP3, led to the demarcating of 170 villages (1.1 million hectares of territory) across the four pilot zones in 2005–2006. Of the 198 certificats fonciers (CFs) delivered during Gbagbo's time in office, 121 were within the framework of the pilot projects. These programmes were interrupted by the political upheavals and crisis that ousted Gbagbo in 2011.⁹

In 2011, the new government of Alassane Ouattara revived and extended the village demarcation and land certification pilot projects. Under the DP4 debt swap programme financed by the EU, the Ministry of Agriculture targeted 17 sous-prefectures in five departments with the goal of certifying 28,000 hectares of land. The 2005–2010 and 2014–2017 pilot projects targeted sous-prefectures in which village demarcation was already complete.¹⁰ Processes in each locality were largely the same (although some divergences are noted below), following the regulations governing adjudication and certification.

Under the 1998 land law, land certification in Côte d'Ivoire is designed to offer state recognition to customary land rights and customary land rights ceded to third parties. The key actors in this process are thus the holders and adjudicators of customary land rights. To this end, the law called for the formation of a Comités Villageois de Gestion Foncière Rurale (CVGFR) within each village jurisdiction, a territory that may range from 1,000 to 10,000 hectares. These are headed by village chiefs (who serve under the sous-prefects as the lowest rung of the territorial administration) and composed of elders and notables in their roles as representatives of customary rights-holding families. Regulations specify that the CVGFR should be broadly representative and thus include women, youth and migrants, but the spirit of rule has not been observed in practice (Varlet 2014). Under the DP3 and DP4 projects, sous-prefects initiated the formation of CVGFR and the certification process by asking village chiefs to develop lists of landholders who were interested in obtaining a land certificate (applicants, or demandeurs). Based in part on the strength of expressed local demand, sous-prefects recommended villages as project sites. Once villages were selected, the Ministry of Agriculture deployed investigating commissioners (commissaires enquêteurs) to work with the CVGFR to carry out the adjudication process, and to prepare a land certificate dossier for each applicant.

Applicants were required to submit a written request, including the parcel description, parcel limits as determined by the surveyor, the name of the customary rights holder, and the names of family members and tenants holding interests in the parcel. Applications were posted publicly for three months to solicit possible objections. At the end of the publicity phase, the CVGFR and commissaires enquêteurs held a public meeting in which claims and objections were vetted. Most objections were resolved through negotiations between the disputing parties. Often, a payment in cash or kind persuaded the objector to desist, allowing the dossier to proceed to the departmental office of the Ministry of Agriculture for preparation of a land certificate which, after vetting by a sous-prefectural level committee, was signed by the prefect.

Demandeurs could seek an individual land certificate (CFI) or a collective land certificate (CFC). Collective certificates are intended primarily for use by nuclear families; they list the names of parents and children, or the names of siblings (in the case of deceased parents) requesting certification of a shared panel. The interpretation of the 1998 law that informed policy from 1998 to 2017 was that both individual and collective certificates are valid for a period of three years, during which time collective certificates were to have been parcelled among the named individuals. Once parcels are immatriculated in the national land registry, a title is issued. The period of validity of all land certificates was extended to 10 years in 2017. Certificates issued to non-Ivoirians (as customary rights ceded to third parties, in this case, to foreigners) are immatriculated in the name of the state and the certified landholder enters into a long-term lease with the state. The 1998 law declared that all land not immatriculated by 2023 would be declared ownerless (sans maître) and revert to state ownership. This clause was under debate within the Ivoirian government during the period of the certification pilot projects that we studied.¹¹

The pilot land certification projects – known locally as the DP3 and DP4 programmes – covered virtually all the costs of the process described above, sparing the applicants’ land registration costs that were estimated in 2019 to run to between FCFA 500,000 and 1 million (US$1,000–2,000) for a 3 hectare parcel in southern Côte d'Ivoire.

Table I presents an unofficial balance sheet for the registration pilot projects in three regions of Côte d'Ivoire – Daloa, Agboville and Abengourou – as of 2017. These are the study zones we analyse in this paper. Not included in our study are the pilot zones of Agnibilikro and Soubré, which ran into blockages and resulted in the delivery of no CFs.¹² Target hectarage in our study zones accounted for 75% of the target hectarage in the DP4 programme. As reported in Table I, about 70% of the initial demandes in these departments resulted in the delivery of a CF.¹³ Collective land certificates were 30% of the total, with wide cross-regional variation. Individual certificates issued to women ranged from 3–15% of the regional totals.¹⁴ The administration did not record the proportion of CFs issued to non-autochthone Ivoirians – i.e. Ivoirian citizens claiming land in localities in which they do not have customary land rights ('ethnic outsiders'). Table I offers an estimate of this figure for each project zone, based on our interviews with state agents and local residents, project reports, and the grey literature. Estimates run from 25–70%. These cross-regional variations are discussed below in analyses of the three study zones.

Table I DP4 Programme de Certification Foncière (2017 estimates), CFs delivered.

^a Estimate based on authors’ surname analysis only.

^b Does not include Baoulé names.

^c Abnibilikro: 124 parcels demarcated but CFs not issued.

^d As of May 2019, 401 CFs had been delivered to women countrywide, out of a countrywide total of 5,550.

^e Plus about 190 in Abengourou from the DP3 and APROMAC projects.

Source: Various unverified Ministry of Agriculture DFR and DRA documents obtained by authors, 2015–2019.

By the end of 2018, a cumulative total of 800 villages had been delimited and the land certification pilot projects had delivered 2,245 (56%) of the 4,500 CFs that had been issued so far. Less than 1% of the land that the government aimed to certify by 2023 had actually been registered.¹⁵

Our field research targeted project and non-project villages in the main DP4 ‘intervention zones': six villages in Daloa, six in Agboville, and six in Abengourou.¹⁶ We interviewed administrative agents and village residents on village selection, land adjudication, perceptions of local demand for certificates, and assessments of why and to whom certificates were delivered. Local agents of the Ministry of Agriculture provided project documentation and other information. Chefs de village (i.e. central government appointees, usually respected senior men who are selected from among autochthonous community members, who serve as village-level state agents) convoked members of the CVGFR, as well as representatives of the Associations of Women, Youth, and Migrants. Our main meetings took place in village meeting venues with approximately 15–40 village residents in attendance. We conducted one-on-one and smaller group discussions with the association heads and other knowledgeable persons, for a total of 3–4 hours per village by our team of four researchers, with repeat visits to some sites. To protect interviewees we have not used individuals’ or village names.

Given our strategy of access, we were able to gather information mostly from autochthonous community members (not ethnic outsiders). Side interviews with autochthonous youth, women and long-established ‘migrants’ allowed us to expand the spectrum of views, and gather additional information. Although social cadets, women and migrants were under-represented in our information-gathering process, the thrust of the information we obtained on age, gender and indigène/stranger biases in the land certification process was consistent with findings in the secondary literature on these topics in western Côte d'Ivoire (Chauveau 2000; Koné 2006a, 2006b; Galoui Bi 2012; Montaz 2015, 2017; Balac 2016; RCI 2017a, 2017b), the Agboville region and the larger south-east (Ruf 2012, 2013; Mel Meledje & Siméon 2016; Tarrouth & Colin 2016), and Abengourou (Gastellu 1989; Koffi Kouassi 1999; Maclean 2010), and with reports by the Ministry of Agriculture, government consultants, and Ivoirian NGOs.¹⁷

General findings were consistent with the hypotheses and expectations laid out in the first section. We found regional variation in the intensity and nature of customary rights holders’ ‘preferences for titling’, and this variation tracked the regional geography of land grievances and advantage. Social inequality and hierarchy were strong predictors of social cleavages and division around titling, both within and across communities. Where land tenure institutions previously recognised by the state were overwritten (superseded) by land certification, customary rights-losers harboured grievances. And as expected, local leaders had strong preferences around the institutional design of land registration (and indeed, for securitisation strategies more generally) that often differed from what the government had on offer.

LAND CERTIFICATION IN THREE STUDY ZONES

This section presents the results of our research in the three study zones (see Figure 1). For each zone, we describe the course of the pilot project and relate our findings to the hypotheses and expectations laid out in the first section.

Figure 1. Study zones in Côte d'Ivoire. Map by Mina Moshkeri.

Daloa: ‘He who has a CF does not need an autochthone'

The Daloa region is in the heart of the western forest zone and, along with the Gagnoa region, has been an epicentre of migratory inflows from other parts of Côte d'Ivoire and neighbouring Sahelian countries since the 1960s. Large in-migrant populations, drawn by expansion of the coffee and cocoa economy, now exceed 60% of the population in most localities¹⁸ and occupy a very large share, almost certainly most, of the land. Over the course of the 1960–80s, in-migrants gained access to land through clientage or tutorat relations of land patronage with autochthones, or through informal purchases of land or farming rights. This process was aided by the agents of the central government who pressured autochthones to offer land to the cash-crop pioneering migrants. On the side of the autochthones, control over land was and is atomised, with lineage heads and household heads (rather than a superordinate authority such as a village chief) in charge of mediating land-access relations with in-migrants. Household heads have been responsible for the more or less informal land cessions and sales to migrants. These land deals are resented inter alia by younger members of households who – especially since the late 1970s, when the supply of good job opportunities in the formal sector began its precipitous decline relative to the number of qualified job-seekers – feel deprived of their land inheritance.

Given the implication of the central state in organising migratory inflows to the Ivoirian west, autochthone-migrant relations have been highly politicised since the 1960s. Autochthonous populations maintained an oppositional stance vis-à-vis the regime in the first three decades of Ivoirian independence and this was met by political marginalisation and repression. These tensions fuelled land-related violence in the 1990s and the political rise of Laurent Gbagbo, who championed the land grievances of the autochthones of the west and who presided over the first phase of village delimitation and land certification in this region. Gbagbo's replacement by his rival Alassane Ouattara in 2011, strongly backed by migrant communities throughout southern Côte d'Ivoire, was a reversal of political fortune for the autochthones of the west. These facts shaped the implementation of the land registration pilot programmes in the Daloa sous-prefecture.

In this sous-prefecture of 12 villages, village demarcation began in the 1980s and was completed under the Gbagbo government in 2006. Under the DP3 programme, the Ministry of Agriculture collected a total of 95 applications for land certificates in 2009. The process stalled when disputes arose between the CVGFR and state agents over whether CFs should be issued to in-migrants, and with the national-level political crisis of 2011, the entire DP3 project came to a halt.

Under the Ouattara regime, land certification resumed in this sous-prefecture under the DP4 programme. DP3 villages were bypassed due to the past difficulties and three new target villages were selected.¹⁹ The selection prioritised large villages where many applications for land certificates had been gathered by local agents of the Ministry of Agriculture – the authorities wanted to undertake certification where this was most likely to be successful. One chief explained that his village was selected for the pilot project because ‘relations with migrants are good in our village – we have no problems here’. Of approximately 800 original demandes across the three villages, 669 CFs were delivered under the DP4 project. This represented the vast majority of all CFs that had been delivered in Daloa Department by the end of 2018.²⁰

Autochthonous individuals and families received about 25% of all land certificates delivered under the pilot land registration programmes. From Ministry of Agriculture records we calculated an average parcel size of 7 hectares, with a strong skew in favour of non-autochthones (i.e. individuals with non-Bété sounding surnames), who registered larger parcels (often 10–12 ha, but some three or four times larger than this), and against autochthones (i.e. individuals with Bété sounding surnames), most of whom registered very small parcels of 1–2 ha (some running up to 3–4 ha). Collective land certificates (certificats fonciers collectifs, or CFCs) (13%) and land certificates to women (3%) were few and far between. Most land certificates went to non-indigenous Ivoirians (45%), and non-Ivoirians hailing from neighbouring CEDEAO²¹ countries (28%). The uneven distribution of CFs across the autochthonous and in-migrant communities reflected the power balances on the ground, both in terms of land occupation and in terms of economic advantage more generally. The in-migrants had the upper hand in both respects. After 2011, the national political dispensation had also shifted in their favour.²²

Interviewees reported that CFs were issued through a process that usually unfolded according to the formal regulations. The customary landholder had to formally register ‘no objection’ to issuance of a land certificate to an in-migrant. Most objections arose around in-migrants’ claims to more land than had been ceded originally, and often resulted in in-migrants’ agreement to (re)purchase all or part of the parcel. Deadlock usually simply halted the certification process. In a few instances, individuals claiming ceded rights hired lawyers to press their claims. According to local interviewees, many non-autochthonous CF holders proceeded quickly to immatriculation and the receipt of a land title (the costs of which are borne in full by the CF holder).

As hypothesised in the first section, a regionally specific land and political history has exerted a strong pull on preferences around registration and titling in target villages in this sous-prefecture, and in the denouement of the pilot projects. The perceived structural disadvantage of autochthones in the state-sponsored processes of in-migration to this region, and the west's history of ‘reversal of fortune’ in the 2000s and 2010s (with the rise and fall of the Gbagbo regime), shaped autochthonous community leaders’ interpretation of bias in the land certification programme.

CVGFR members and other heads of household expressed a strong sense that land certification, rather than securitising their customary claims on land, had become the vehicle of their dispossession. Certification had, in their perception, undermined the autochthony-affirming tutorat land tenure institutions that the state had endorsed in the past. Interlocutors in each village explained clearly that through certification, non-autochthones had established autonomous claim over the land: ‘He who has a CF has no need for an autochthone'; ‘Social relations have been turned upside-down'; ‘Migrants now strut around our village without showing respect for us'; and ‘Next, we will be laborers on their plantations!’. Indeed, many expressed the general anxiety that many autochthonous families were now indisputably landless. Moreover, many were convinced that autochthones in the west were worse off than autochthones elsewhere in Côte d'Ivoire because the policy environment was singularly pro-migrant in west.

In terms of social inequality or hierarchy, the main cleavage underlying local tension around certification was the autochthone–outsider cleavage. Social hierarchy also activated a line of cleavage within autochthonous communities and families, especially between elders and male youth (i.e. social cadets, generally those under about 40 years old). Social cadets were generally excluded from the CVGFR because these positions were reserved for household heads. When a land certificate (and then a title) when to a non-autochthone, the young man knew for sure that he would never inherit his ancestors’ land. And by raising land prices, certification also made it more difficult for him to rent or buy land in the sous-prefecture.²³

Closing of the land frontier and the land shortage felt by autochthonous families clearly aggravates all tensions around land certification, and colours local community leaders’ preferences for state-led land securitisation efforts. All six chefs de village asserted their strong preference for a rule that would require non-autochthones to obtain a chief's permission before proceeding to land immatriculation and title. They believed that in general, land titles should be assigned to autochthonous families only. Families could then lease or rent land to non-autochthones and others. Their unanimity on this issue was in part the result of political coordination. All 27 chefs de village of this sous-prefecture had met as a group with their member of parliament (who hails from northern Côte d'Ivoire) in 2018 to protest the way land certification and titling was progressing, and to call for a halt in the process until the rules can be revised by a National Assembly renewed by the 2021 elections.

Agboville: ‘The CF is for fonctionnaires!'

Agboville is part of the south-eastern Côte d'Ivoire region that has been marked by a long history of land commercialisation. In the 1920–1940s, French and other European settlers created commercial plantations along the Abidjan–Bobo railway line. With the railway came in-migrants from other parts of the colony and Burkina Faso. They settled on or around the commercial estates or on the periphery of autochthonous Abé villages (see Segula 1978; Chaléard 1981–2). The colonial state zoned the Agboville region into freehold lands, classified forest, other state land, and ‘the customary domain’. In the 1970s and 1980s, the country's first president declassified tens of thousands of hectares of forest in the Agboville region to give favoured members of the political elite large private landholdings ('their slices of the national cake'), or to make land grants to autochthonous villages (Affo-Yapi 1987; Faussey-Domalain & Vimard 1991). A new trend of private land acquisitions by members of the Abidjan-based salaried and rentier classes began in the 1990s. It gained momentum dramatically in the 2000s and 2010s, fuelled by a boom in rubber (Hevea) prices, an acquisition rush by political elites in the terminal Gbagbo years, and the favourable investment climate prevailing since 2011 under the Ouattara regime (Ruf 2012, 2013; Tarrouth & Colin 2016). Much of the land-selling over time has happened incrementally and informally as Abé families have sold parts of their customary land. Today, the Agboville region is a chequer-board of different types of land tenure and land holding: land titles on corporate holdings and private holdings, the privatised lands of former state-owned companies, classified forests, tens of thousands of hectares of declassified forest that were given in land grants to individuals and villages, and the customary domain (domaine coutumier). As in Daloa, there is no open land frontier.

Private land certification started here in approximately 2010 and the first CVGFR were created at that time. By April 2014, 188 CFs had been issued in Agboville (average size, 28 ha), by far the most in any single department.²⁴ At the end of 2017, 640 CF had been delivered in the department of Agboville. Of these, 271 were issued from the DP4 land certification project in two project villages.²⁵ The majority of Agboville CFs have thus been obtained privately (outside the project framework) and by ‘new boom land acquirers’, and are dispersed across many village territories. There is a strong contrast with the Daloa and Abengourou sous-prefectures, where almost all CFs have been issued within the government's land certification pilot project and are concentrated in a few project villages.

Agboville sous-prefectures were included in the DP4 effort because project designers were eager to balance the wave of certification by new acquirers in this region with more certificates for local village residents. The two villages (combined population of 6,600 in 2018) were selected because village demarcation there was (almost) complete, and because local officials believed that the project could be successful in these localities. In one of the villages, the project apparently followed the official procedure: CVGFR were constituted at the initiative of the sous-prefect, lists of applications for CF were drawn up, commissaire-enquêteurs did their work, dossiers were constituted and publicised, and CFs issued at no cost to the demandeurs (except the charge for the fiscal stamp of about $10). In the second project village, there were short-cuts to formal procedure, but CFs were issued nonetheless.

Of the 271 CF delivered in the two project villages (of 328 applicants), about half went to autochthones, and about half to non-autochthones (of which, 40% to non-Ivoirians apparently from CDEAO countries to the north of Côte d'Ivoire).²⁶ Interviewees in both villages said that most of the project CFs that went to non-autochthones simply formalised land sales that had taken place in the past, often 30 years ago or more. Most of the certificates delivered to autochthones were collective certificates (CFCs); the Ministry of Agriculture agents implementing the DP4 programme had encouraged this option. This also represented a stark departure from the Daloa experience, where administrative agents had discouraged families from opting for CFCs and few CFCs were issued. Forty per cent of names that appeared on Agboville's 120 project CFCs (209/497) were women's names (Lestang 2018: 5). The average parcel size under the DP4 project was 16 ha (compared with 28 ha in the privately obtained CFs distributed before 2015). Even within the DP4 sample, some large parcels skewed the average upward (15 of 271 parcels were over 50 hectares). We can conclude that hectares per capita secured by the project-sponsored land certificates (collective and individual) obtained by autochthones in Agboville department was low, perhaps about 4 ha.

As hypothesised in the first section, regional specificities – land tenure and migration histories, as well as Agboville region's physical and economic place in the national economy – play a very strong role in shaping village residents’ perceptions and preferences around land certification and titling. A long history of both formal and informal land sales to outsiders defined the context in which commercial land transactions and outright land sales are seen as a normal fact of life. The short distance to Abidjan (50–75 km by paved road) has been a key factor in the local boom in land sales and certification: the location is very convenient for salaried Abidjanais absentee-landowners. In most villages, interviewees believed that land certification since 2011 had intensified outsiders’ demand for land and accelerated the pace of sales.

Although our interlocutors did not express a sense of betrayal of a land-tenure status quo ante, in two of the non-project villages the new wave of selling and (in some cases) certification and titling was perceived as an existential threat to the community because of its scope and rapid pace, cumulatively adding up to a ‘tragedy of the commons’ in which the collective land patrimony could dissolve away. Chefs de village said they try to discourage sales in favour of land rentals or leases, ‘but land-selling is a decision of household heads – it is difficult to prevent’.

As anticipated, strong perceptions of social hierarchy and economic inequality informed autochthones’ descriptions of land certification. Yet in contrast to Daloa, the line of social cleavage emphasised in the project and non-project villages in Agboville Department was the division between long-time village residents and wealthy outsiders, referred to as ‘the haves’ (les nantis) or new investors from outside (opérateurs économiques), who arrived to buy land and increasingly, to certify and title it. In the non-project villages, land certification was seen as an option made for salaried civil servants, or fonctionnaires. ‘The CF is for fonctionnaires who have purchased customary land. As soon as they receive the CF, they rush to get the land title.’ Perceptions of corruption, non-respect for the legal procedures in the land certification process, and/or the sense of being outbid by the wealthy, generated grievance and frustration. Where high-visibility politicians or regime personalities were said to have acquired large tracts through non-regular channels or even the use of force, there was a sense of outrage.

Chefs de village and CVGFR members described an acute sense of land shortage in the region. They were unanimous in explaining that the majority of autochthonous families had little or no land to pass on to their children, that unemployment and delinquency among youth were high ('there is not even a youth centre here'), and that wage jobs in agriculture were low-paying and exploitative. According to them, the land-buying opérateurs économiques rarely create jobs for locals: either they hold onto their land for speculative purposes without developing it, or they bring in workers from outside.

In the two DP4 project villages, interviewees bemoaned the high cost of land certification on the private market, and expressed strong interest in more state-subsidised land certification. This was expressed as an individual-level, defensive response – a way to ensure that households and individuals would hold onto their land. Some pointed out that once their land was certified, they could rent it out. Among those holding certificates obtained through the DP4 project, there was worry and uncertainty about the spectre of break-up of collectively certified properties. Would there be enough land to divide up? How would they pay for the costs of survey and certification of the individual parcels? Could family disputes be managed?²⁷ In non-project villages, chefs de village and CVGFR members had somewhat different preoccupations and preferences: three of the four had asked the government for additional forest declassification, to give the village new land to allow for the creation of new farms.

Abengourou: ‘The [autochthonous] Agni regain land ceded in the past'

The economic and political history of Abengourou is defined largely by autochthonous Agni lineages’ successful pioneering of the local smallholder cocoa economy (économie de plantation), and by the fact that these same autochthones, with the backing of the Ivoirian state, have retained the upper-hand in land and labour relations with in-migrants. This has defined power relations around both village demarcation and land registration and ensured that locally dominant social strata were best positioned to take advantage of land registration and titling under the government-run pilot programmes.

The department of Abengourou is Côte d'Ivoire's oldest zone of cocoa and coffee production. Smallholder plantations developed strongly over the course of the last century within the hierarchical social structures of indigenous Agni kingdoms and chieftaincies that trace their origins to the 18th century. Since the 1920s there has been substantial in-migration from other parts of Côte d'Ivoire and from other parts of the West Africa subregion. Ivoirian and non-Ivoirian migrants gained access to land through grants from Agni chiefs and lineage heads (mostly before 1960), loans, tutorat-like sharecropping contracts and sales. Early waves of in-migrants founded their own settlements within the village territories of the Agni chieftaincies, cleared the land, and created coffee and cocoa plantations. Since the 1970s and 1980s, land saturation has slowed the land-pioneering dynamic and planters have tended to recruit in-migrants as workers or on limited-term farming contracts. Non-indigenes made up about 30% of the population in the 1960s and 1970s (Diaby 1996). Today, non-autochthones represent the majority of the population in many villages of Abengourou department. Even so, and in contrast to the two other regions, autochthonous inhabitants of the region have the upper-hand politically at the village and sous-prefecture levels.

The department of Abengourou has been included as a pilot zone in all the village demarcation and land certification projects since the 1990s, partly because the central government and foreign donors have seen local conditions as propitious for success. Where chiefs and Agni notables bought into village demarcation and land certification, the hierarchical structure of relations between villages within sous-prefectures (former cantons), and within village territories and villages, has indeed facilitated these operations. However, in some localities, village demarcation and land certification under pilot projects encountered push-back. The spectre of land taxation, non-respect of inter-village hierarchy or encroachment on established forms of land authority, the possibility of erosion of customary land rights or reduction of territory under autochthones’ control, aversion to direct state control over landholdings at autochthones’ expense, and resistance to dividing-up lineage or family land holdings made some local authorities reticent and diverted the projects to other localities (see also Koffi Kouassi 1999).²⁸

Land certification in 2008–2010 under DP3 began in a village near Niablé, producing 60 land certificates, approximately 90% of which went to Agni landowners (average parcel size, 50 ha). Resumption under the DP4 programme in 2015 targeted four villages, and in two years delivered about 970 land certificates (out of a total of 1034 for the Abengourou department) covering 12,000 hectares of land. Most certificates in the DP4 programme went to Agni families (about 74%). Most of the rest were issued to long-established in-migrants whose parents or grandparents had received land grants from Agni chiefs or lineage heads during the colonial period, or in the first two decades of independence. (Fourteen per cent of the programme-sponsored certificates went to non-Ivoirians and another 8–10% went to Ivoirians with non-Akan Ivoirian surnames.)²⁹ Recent in-migrants, however, did not fare so well. Those who purchased access to land in recent decades or were in long-term land-sharing arrangements with autochthonous families (such as land- or crop-sharing abusan or diochamchê), experienced an erosion of their bargaining power and their ability to claim permanent and transferable land rights, either because their land was certified in the name of the original (autochthonous) landholder or because of the general shift in the local balance of power against them.³⁰

Certification dynamics reflected a configuration of regional power and social hierarchy that is specific to this region, as anticipated in the arguments sketched in the first section. Chiefs invited the pilot programmes to work in their villages, and chiefs, male lineage heads and long-established in-migrants who had been assimilated into the local notability controlled the CVGFR. Formal certification procedures were followed only roughly in many cases; state agents implementing the project preferred to work within the existing ‘well structured’ social hierarchies. Where autochthonous landholders did not want to surrender land to in-migrants seeking to register parcels, they resisted certification and sometimes the certification process as a whole.³¹ Some Agni individuals and families in DP4 villages did not participate in the certification project for this reason, and some villages were not selected for DP4 because cantonal or village authorities were cool on certification on the basis of the same rationale. One village dropped out of the pilot project because the CVGFR turned out to be divided on this issue.

Social hierarchy within a still largely agrarian society defined winners and losers of certification and titling, and thus went far in shaping preferences in this region. As Koffi Kouassi realised in his 1999 study of an earlier land-rights assignment and securitisation project, the autochthones retained the initiative. There is no open land frontier in this region and Agni families who certified properties often reasserted rights to family or village lands that had been ceded in recent decades to in-migrants via informal purchases, loans or land-sharing contracts. Large family holdings are considered characteristic of this region (with holdings of 100, 200 or 300 ha not uncommon), but the modal size of parcels registered under the DP4 programme appears to have been about 10 ha. The certification projects thus appear to have contributed to more general processes of fragmentation of lineage land into nuclear-family (or individual) holdings. Thirty per cent of all DP4 land certificates in this region were collective certificates. On the basis of a review of 156 dossiers, we estimated that the collective certificates included an average of about five persons (with about one-third of collective certificates listing only one name).³² Certification thus probably strengthened leading individuals’ land prerogatives vis-à-vis those of lineages, families, female family members and youth. Many interviewees in the latter categories expressed frustration over their inability to benefit from the pilot certification project. By contrast, autochthonous youth and non-resident cadres belonging to locally influential families could benefit from certification when their parents were able to register land for the nuclear family and to recover land ceded in the past. Some Agni women with their own cocoa or Hevea plantations (typically inherited from their fathers) did certify their parcels under their own names through the DP4 programme (15% of all titles delivered).

Recent in-migrants (arriving during or after the 1980s) got the short-end of the certification stick in the pilot project. Many farmers in this region hold a wide variety of informal documents ('petits papiers') attesting to land sales, some signed by sous-prefects and other state agents. These documents are often underspecified, typically not saying whether the buyer purchased the land, the trees, or the crops growing upon it, or time-limited use rights only. (This phenomenon is also widely documented in the Daloa region.³³) In such cases in Abengourou, when autochthones wanted to certify the land for themselves in the context of the DP4 project, past transactions were often interpreted in the autochthones’ favour to mean that the land itself had not been sold. (The opposite arbitrage was the norm in Daloa.) Similarly, in Abengourou, when the Agni customary landowner who originally ceded the parcel was no longer living, the transaction with the in-migrant was often considered obsolete (caduc). Migrants who obtained land through sales or loans in recent decades have thus experienced an erosion of their right to sell this land (to anyone but the customary landowner), or to pass it on to their heirs without permission of the customary landowners. In the course of the DP4 project, autochthonous landowners and state agents appear to have been united in their interpretation of customary land institutions and land transactions of the last few decades to the autochthones’ advantage.

Although there is little indication that new land certificate holders in the Abengourou region have moved to immatriculation and land titling, some did express desire for more private options for land certification at accessible cost. They would like to not depend so heavily on government agents and contractors, and erratic certification project cycles. Comparing across regions, there is a stark difference between Abengourou landowners’ hope for more private certification to consolidate their advantage, the fear of Agboville's customary rights holders that they will lose out in a race to land certification in which the land will go to those who can afford to pay for it, and perceptions of land certification in Daloa, where autochthones see themselves as the losers in a process that is stacked against them.

The character, functioning and operational existence of the CVGFRs varied by region. Almost all those surveyed by PARFACI (Varlet 2014: 73) in 2006 were 90–100% made up of autochthones. By 2011–2016, women's representatives and a representative of ‘the non-autochthonous community’ had been added to the committees in the sous-prefectures we studied. The CVGFR we encountered in Daloa seemed to operate mostly at the behest of agents of the local administration. In Agboville, several localities did not have an operational CVGFR, two committees had been disbanded or were non-functional due to disputes over non-compliance with formal guidelines, and several did represent local village leaders but felt they had little real power to structure the land certification pilot project. In Abengourou, the CVGFR were clearly agents of and staffed by locally influential autochthonous families. An implication of the analysis advanced in this paper is that these variations were symptoms of more general power-balances in relations between the government, in-migrant communities and autochthonous families in these different regions, rather than independent causes of the variation in outcomes of the land certification pilot programme.³⁴

CONCLUSION

Of the 4,500 CF delivered through the end of 2018, our focus has been on the land certification pilot project zones where certification was undertaken with no direct charge to the landholder, accounting for about half of all certificates delivered in Côte d'Ivoire as of mid-2019. We were thus able to observe ‘demand for certification’ in village settings in which the cost constraint was lifted from the landholders. There was considerable heterogeneity in village residents’ response to the titling project, both across regions and across social categories. Our observations suggest that uneven or ambivalent demand for registration can arise from perceptions of political, social and economic risk; fears of exclusion; differing preferences around the institutional design of tenure reforms; and tensions arising from the uneven distribution of costs and benefits associated with certification and titling. These are, in part, regional effects that are shaped by uneven development of the national economy, histories of differential incorporation into the national project, and the spatially uneven histories of land politics in each region. Concerns about the risk of exclusion, and conflicts over legitimate land ownership and the locus of customary rights, can be serious fetters on certification. Scaling-up to the regional and national level, these contribute to divergent regional preferences of autochthones (as well as of in-migrants) and regionally specific patterns of politicisation around land registration and titling. Such concerns may have strongly political overtones, and be politicised, at the local and regional level.

This study has not discussed the other 2,000 CFs delivered in Côte d'Ivoire since 2011 – i.e. those obtained outside of the pilot certification programmes, at private cost to the applicants. What is known about these? Tarrouth & Colin (2016) have shown that the biggest take-up is in the Agboville-Sikensi region (where investors are planting rubber trees) and in Assinie/Assinie-Mafia, a touristic region along the coast east of Abidjan. These are regions with weak traditional rights and high levels of land commercialisation. Informed estimates corroborated by our analysis of land certificates immatriculated in the Journal Officiel in 2017 and 2018 indicate that about 85% of the non-project CFs are going to individuals who reside in Abidjan (cadres, fonctionnaires, opérateurs économiques). By the end of 2018, privately secured land certificates (44% of the total) covered almost five times the land area, and were issued for parcels that were, on average, approximately five times as large (66 ha compared with an average of 12 ha in the pilot certification projects discussed above).³⁵ This reveals a class-consolidation aspect of the land certification and titling processes that can be understood as layering onto the regional and local dynamics that have been the focus of our analysis.

Land rights formalisation has been characterised by land securitisation advocates and in some political science research as providing a private good for beneficiaries – that is, as a type of pro-poor distributive politics, akin to the delivery of electricity or education. But as Bhattacharya et al. (2019: 53) argue in a review of land tenure reforms worldwide, certification initiatives may not improve tenure security for all land-rights holders, and are not always pro-poor. In African countries, land rights formalisation is taking place in the absence of pro-poor mobilisation, and in national contexts of high interpersonal and spatial inequality. Top-down drivers of land certification are strong, and these are likely to relate strongly to the quest for consolidation of territorial control, and the quest for asset-creation through the opening of land markets. Often an explicit goal is the eventual transfer of land to investors. The Côte d'Ivoire case underscores the fact that such programmes may have (re)distributive implications that cleave rural societies along the lines of pre-existing social and regional inequalities, and that may overturn previous land tenure institutions in ways that create losers and fuel grievances. This helps explain pushes, pulls and push-backs around land certification that we observe across the regions of southern Côte d'Ivoire and in other contexts in Africa.