The World Bank ’ s endeavours to reform the forest concessions ’ regime in Central Africa : lessons from 25 years of efforts

At the beginning of the 1990s, the World Bank introduced conditionalities to reform the forest concession regime in Central Africa and continuously intervened, inter alia, in Cameroon, Congo, Gabon and the DRC up to 2010. The reforms were designed to achieve two immediate objectives: (i) increase the price of the resource through competitive market procedures or, if inapplicable, through taxation, (ii) break down the patronage system that governed the forest permits allocation. These efforts have been supported by national reformers, and fought against by various vested interests such as the timber companies, but also officials in the forest departments. Because of the difficult political economy, reform ambitions have been lowered, and only a certain proportion of the policy measures initially contemplated has been implemented. This called for a completion of the reform agenda and fine tuning. But by the middle of the 2000s, the evolution of paradigms in tropical forestry gave the critics of the WB policy of concession reforms opportunities to challenge the orientations followed hitherto in Central Africa. The 2013 Independent Evaluation Group (IEG) report of the WB forestry policy, criticizing the WB’s management efforts to reform concession regimes in Africa, testified to this change of paradigm. The critics are of mixed mind, with community advocates favoring more pressure on the African governments to dismantle the concession regime and proponents of “performance-based aid” seeing REDD+ as an opportunity to withdraw from direct sectoral intervention. Paradoxically, in spite of REDD+ initiatives and rhetoric, forestry has tumbled in the national policy agendas and forest concessions or protected areas are encroached by agribusiness plantations and mines with the implicit support of large fractions within governments.

The WB took a prominent role in the forestry dialogue and the reform processes in these five countries 3 , with more or less success.This strong involvement and its modalities can be attributed to the convictions of the then WB staff in charge of the sectoral operations for Central Africa.But it was facilitated by the "structural adjustment" plans of that period that gave the WB levers to orient the policy dialogue in certain directions, notably for reforming the system of concessions.With the end of the structural adjustment period, large portions of these countries' foreign debts were cancelled through the HIPC (Highly Indebted Poor Countries) process toward the end the 2000s.Last, the emergence of REDD+ in the agenda, has completed the change in the manner the WB is conducting policy dialogues for the forest sector.

INTRODUCTION
Between the early 1990s and the mid-2000s, important changes occurred in the regulatory frameworks of several forested Central Africa's countries, in order to foster sustainable management of the forest resources and to increase transparency and co-benefits from the use of such resources.All the six forested countries (Cameroon, Congo, DRC, Gabon, CAR, and Equatorial Guinea) adopted new forest codes during the period, often supported and assisted by international organizations.In some of the countries, the involvement of the World Bank (WB) in both the preparation and the implementation of new legal principles through the publication of decrees and "arrêtés ("Orders"), has been especially noticeable, e.g., in Cameroon and the DRC, given the WB's financial levers 1 in these two countries through the attachment of "conditionalities" to critical loans and grants during negotiations with these countries.By comparison, the Republic of Congo and Gabon, where WB interventions for introducing reforms has been tangible but without comparable financial levers, have not experienced the same relationship with the international institution in the forest sector.The WB approach's footprints on the regulatory framework are less visible, even though they are not non-existent 2 .The Bank has has also been assessed critically by two specialized bodies of the WB group itself: the Inspection Panel (IP) for the DRC (2007) and the Independent Evaluation Group (IEG) (2013).Both the IP and the IEG reports thought ill of efforts to reform the concessions regime in Central Africa4 , arguing that it was a way of perpetuating if not supporting this system rather than replacing it by an alternative regime based on community forestry, conservation initiatives and other ways of capitalizing the forest resource (through carbon storage, for instance).Actually, neither of these two reports examined the theoretical background that supported WB interventions in the forestry sector.The policy recommendations, which were explicit in the IP report and implicit in the IEG report, betrayed the a priori thinking and beliefs of their authors, and also their poor knowledge of the political economy of those countries.
This article tries to shed light on the political economy context of the reforms conducted at the initiative of the WB in Cameroon and DRC between 1990 and 2010, and to explain the theoretical background of the WB intervention in the forestry sector.It takes stock of the reform process and its achievements, difficulties and limitations.This article also provides some background for explaining the changes in the modalities of the WB intervention in the forestry sector by the end of the 2000s decade and why it has been taken to task by its own Inspection Panel and its Internal Evaluation Group.Finally, it discusses the new context in which the intervention of donors (traditional ones such as the WB, or potential newcomers such as Norway) is likely to take place.It recalls the ambition embodied into REDD+, i.e., to leave the countries fully responsible for designing and conducting appropriate policies to curb deforestation, and to pay for the "performance" achieved.It argues that such an approach will turn out to be an illusion in countries with limited capacities and weak institutions.New policy approaches informed by the lessons of reform experiences in Central Africa, need to be promoted to shift from public national forest policies to forest-related policies.

THE SITUATION OF THE FOREST SECTOR IN THE EARLY 1990S IN CENTRAL AFRICA: A "RENTIER" SYSTEM BASED ON ECONOMIC EXTRAVERSION AND LAND RIGHTS ALLOCATION
The forest landscapes of Central Africa are characterized by low population densities, often below 2-3 inhabitants per square km.In an article called "La dialectique du vide", the geographer Roland Pourtier (1986) insisted on the impact of such phenomena on the perception of space and social organization.He dwelled especially on the blurred boundaries of the terroirs used by local communities.Land tenure is seen, therefore, as recognition of the effort devoted to clearing the land rather than ownership as is usually the case in Western conceptions.The land is not scarce, but the labour is, especially in local societies composed of relatively small villages that cannot easily mobilize large working forces for collective tasks.As Vermeulen (2000) also pointed out, conflicts within communities often lead to separation and the creation of new villages, since there were limited pressures on lands before.
On the other hand, forested countries of Central Africa5 enjoy abundant natural resources, such as oil (Gabon, Congo, Cameroon, and Equatorial Guinea), minerals (DRC and CAR, but also Congo and Cameroon) and timber.These assets, in countries with low education performances and weak institutions, tend to bring the economy of the region close to what economists call "the resource curse": extraversion of the export-based economy (raw material), under-development of agriculture, policy priority given to urban populations, and high social inequalities.

Ecological heterogeneity and high costs of resource mobilization means highly selective harvests
Forestry in Central Africa has been shaped by both the ecological characteristics of the forest and the economic features.Harvests are very selective ("high-grading") and concentrated on a handful of species: those whose commercial value is high enough to bear, notably, the high transport and handling costs incurred by the weak, poorly managed infrastructures.Given the spatial heterogeneity of the species, the volume harvested are, on average (with some exceptions when the commercial species are concentrated), around 8-12 m 3 /ha -and even as low as 3-4 m 3 in the hinterland of the DRC.As a result, the concessions tend to be large, with some in Congo and the DRC bigger than one million hectares.Around 50 million hectares are in the concessions (around 26% of the moist forest area) (De Wasseige et al. 2009).
The commercial harvest per surface unit is very limited compared to what is harvested in dense forests in South-East Asia.Industrial logging is rarely a direct factor of deforestation in Central Africa.There is no correlation between the importance of the forest industry and deforestation: Gabon used to have the largest timber industry, with up to 3.2 million m 3 extracted per year less than a decade ago, and it is also a country with one of the lowest rates of deforestation (less than 0.1% of gross deforestation).Conversely, population density is closely correlated with deforestation.As pointed out by Defourny et al. (2011) for the DRC, "the presence of a logging concession and a mining permit does not seem to play a role in deforestation/degradation, at least at the national and sub-national scales studied" and ". . . it is above all the importance of the on-site population which determine the quantity of forest affected by deforestation and degradation.These very clear results contradict several studies, more local-oriented, which often emphasized the distance to roads and the importance of the flux associated with roads as the first cause of deforestation".this pattern has changed over time: most of the timber is locally processed and low quality logs make up the bulk of the log exports (for China and India).The industry hardly produces for the domestic market -which is supplied almost exclusively by the informal sector.
The high cost of labour is partly linked to the fixed exchange rate between the CFA Franc and the Euro, and, for the DRC which has its own currency, to the cost of life.Coupled with the lack of trained workers and various logistical difficulties, this has been the rationale for the exportation of logs for decades.The governments' log export bans have modified this marketing pattern in Gabon and Cameroon, but has done less well in the Republic of Congo.On the other hand, the successive crises that struck the local populations put the industrial products beyond their financial reach, especially since the products were made more expensive because of more norms and taxes.Chainsaw loggers gradually replaced the formal industry as domestic suppliers.Now From an ecological perspective, the "creaming" of the timber species may have negative impacts: overexploitation of certain species, whilst other species, sometime abundant ones like the Fraké (Terminalia superba) in Cameroon, could be harvested more to supply the industry, without jeopardising their regeneration (Fredericksen and Putz 2003).Creaming would entail changes in the composition of the forest that could deplete the commercial value of future harvests (if the market demand targets the same given species over time).

Greater preference for raw material than for local processing
The other characteristic of Central African countries in the 1990s was the predominance of exports of logs6 .Typically, the better quality of timber was exported as raw material, while the lower quality was processed locally, partly for export, but also for local consumption.It is striking to see how the informal sector is producing more and more while the industrial sector is on the decline.

Low population density and abundant natural resources: the rationale for the concessions regime
For land-locked and low-populated countries such as Central African Republic (CAR), Northern Congo and most part of DRC, the rationale for large-scale concessions was clear.History, which shapes the collective preferences of local inhabitants, also matters.With a limited tradition of local self-organization for economic development, the collective preferences have favoured a bargaining-dependent relationship with the concession permit holder rather than unlikely community development externally promoted7 .On the other hand, the delivery of local public goods by foreign large-scale companies that are requested to replace an absent government cannot be seen as a sustainable and desirable model, since it can turn out to be an obstacle to the development of local and accountable representation schemes, especially public ones (elected local councils, etc.).
Land tenure is generally featured as "dual", with opposition between the State monopoly or ownership of land and resources, and the customary practices which allow the "land master" to allocate plots of land to newcomers who are guaranteed traditional family property rights over the cleared land, as long as the plot is cultivated.But another duality is even more important: the overlapping between community lands and forest concessions.Such overlaps are manageable since they are usually not mutually exclusive: land reserved for populations, timber for concessionaires (see Lescuyer et al. 2012).But there are also cases where trees are the subject of competition between concessionaires and populations: Moabi (Baillonella toxisperma) trees for seed oil and bark, Sapele (Entandrophragma cylindricum) trees for caterpillars.Yet, community claims for an economic autonomy are limited.Greater autonomy would be difficult to sustain, given the limited access to markets, credit and affordable transport.The relationships with concessionaires are ambiguous: on the one hand, logging activity creates trouble for hunting, cropfarming and so forth, but on the other hand, local communities see concessions as opportunities for employment, and for bargaining (various "compensations" are expected or claimed).

A discretionary allocation system at the core of the patronage system and politicians games: high private transfers and low public revenues
The core of the "rentier" system has always been the allocation system for concessions (whether for forestry, mining or agriculture).In some cases, the allocation was granted through a commission -or bidding -which made the decision processes opaque and plagued by corruption.After this point was raised in several WB documents in the 1990s concerning Central Africa (Grut et al. 1991), designing a transparent and competitive allocation system became a priority for the "conditionalities" set by the Bank for disbursing loans and cancelling debts.
The taxation system was also subject to scrutiny.Most of the tax collected was for the export of logs; the royalties proportional to the surface granted in concessions were inexistent or very low.This configuration was adapted to a discontinuous activity, characterized by "stop-and-go" phases, where taxes were due only when the timber was sold and exported.On the other hand, it was considered to encourage waste, "creaming" and the holding of huge areas by concessionaires at the expense of other potential uses or tenure (Grut et al. 1991;Gray 2002).In the DRC, at the beginning of the 2000s, more than 40 million hectares were held by concessionaires, but the country was only exporting 100,000 m 3 , meaning that most of these concessions were not exploited (at least for timber) (Debroux et al. 2007).
Public revenue from the timber trade was low, compared to what could have reasonably been expected for an activity spread over such large areas.In Cameroon, revenue amounted to $7.1 million in 1993, and for the DRC, the forest revenue was only around $100,000 in the early 2000s.This did not mean that the concessionaires had no "tax-like" costs: corruption costs were -and still are -significant and many "parafiscal" contributions requested by various administrations and bodies for their own sake, put a strain on companies' budgets.In sum, companies pay almost all the time to myriad of bodies, but only a tiny part of this money reaches the public treasury (see CIRAD &ID 2004, for DRC, andCerutti et al. 2012, for Cameroon).

THE WORLD BANK INTERVENTION IN CAMEROON
Cameroon, by the end of the 1990s, was a country with high debts and serious cash flow problems.The country went into a quite brutal process of "structural adjustment", with severe cuts in public expenditures and civil servant wages.Many of the civil servants were dismissed (people coined an ironic term saying that the civil service was "déflatés") in the early 1990s.Financial and technical assistance from the WB and the IMF had been granted, on the condition that reforms would be introduced in various sectors8 .These "conditionalities" triggered for the reform of the forest sector in Cameroon, which will inspirit eventually similar processes in some of the neighbouring countries.

A diagnosis formulated in economic terms
During the 1980s and 1990s, there were several very active schools of thought on natural resources management, especially on timber.Economists such as Malcolm Gillis, David Repetto, Jeffrey Vincent, John Gray and William Hyde, amongst others, wrote several influential papers on the economics of forest concessions.Most of the analyses were derived from Malaysia and Indonesia where forest resources had been rapidly depleted whilst concessionaires enjoyed high profits.For West and Central Africa, a report by Grut et al. (1991) prepared for the WB, carried these analyses further and formulated policy recommendations which were to have a strong influence of the WB's work in Central Africa.
The "underpricing" of timber resource, thanks to an unbalanced fiscal structure (taxes on exported wood rather than on the volume cut or the surface granted), was considered as a source of waste and inefficiency affecting the entire value chain.It was also deemed as being a transfer of the economic rent to the concessionaires at the expense of the public treasury.Such underpricing was attributed to the structure of the fiscal system, but also to the discretionary allocation system that prevented transparent and fair competition.Competition could be a tool for showing willingness to pay the loggers to access the resource, as was later revealed in Cameroon (Topa et al. 2009).Focusing on bringing competition into the allocation system was considered as an opportunity to kill two birds with one stone: raise the stumpage value of commercial timber and introducing transparency and fairness into an opaque mechanism.Efficiency and governance could, then, go hand in hand.In addition, the expected higher fiscal revenue could be shared between the central government and the local governments: in Cameroon, the reform entitled local councils on whose territory the logging was taking place to 40% of the area fee (set through auctioning) and 10% was granted to the neighbouring communities.
The reform in Cameroon was prepared in the early 1990s by a mixed team composed of high profile officials from various ministries, WB staff, and consultants.An important document on forest policy was published in 1992 by the Cameroonian government.It summed up one year of discussions by the mixed team, and set out much of the content of the 1994 forestry law.This law, thus, was not "imposed" by the WB on the Cameroonian government, as is often believed.But the small number of Cameroonian civil servants involved in the process suggests that the "ownership" of the process was too limited.Ekoko and Brunner (2000) mentioned the absence of implication of the Parliament's Representatives which eventually led to the Parliament's eliminating a number of items previously agreed upon with the Bank and adopting unwelcome (from the WB and the Government's perspective) amendments to the Forestry Law (such as the introduction of a log export ban).In fact, at the onset of the reform process, members of Parliament (with many of them involved in forestry activities) were keener on defending their own privileges than advocating on behalf of the national economy, the environment and the interests of the rural population.Local NGOs, although mentioned as major stakeholders in the WB policy documents, did not participate in the discussions, and no real efforts were made to explain the orientations afterwards: the Bank staff was perhaps convinced they would welcome the Bank's effort to promote better governance, community forestry and a better fiscal system in a way that would benefit the local councils and the communities.
Cameroon has been a laboratory for introducing an auction system for forest permits that is unique in tropical forestry 9 .The principle was to auction an annual fee based on the area conceded10 .The negotiations with the Government led the WB to a concession that seemed minor at the time but turned out to be critical later on: the combination of "technical criteria" with the financial offer for assessing the bids.The WB managed to have the financial offer account for 70% of the total, and the technical proposal for only 30%.But the technical proposal did not only have the disadvantage of being assessed on a "subjective basis" (with the risk of corruption of the commission members) that distorted competition; a too low score could also disqualify the tender.This eliminatory threshold was overused in several cases to eliminate competition and allow a bidder to remain the only one on the list. . .

The second part of the agenda: community forestry, independent watching and SFM
The institutionalization of "community forests" (with a legal status) was also innovative in Central Africa.This agenda was pushed notably by IUCN, and received the support of the WB which shared the idea with the group of Cameroonian reformers in the mixed commission, in spite of the reluctance of the majority of the staff in the ministry in charge of forestry.Unlike it is sometimes considered, this was not a contingent objective of the WB in Central Africa, but it was conceived as the second step in a sequential dynamic of reforms.For WB staff involved in the process, no community forestry effort would have been possible or conceivable at the margin of an all-powerful, unregulated forest industry.
The community forests (CFs) were limited to a maximum of 5,000 ha, and the Ministry of Forestry obtained a modification in the draft law that excluded the community forests from the future Permanent Forest Estate (PFE), in spite of the legal requirement to draw up a (simplified) management plan that would guarantee long-term forestry use for the CFs11 .Since the highest quality forests were located in zones planned to be gazetted (to be included in the PFE12 ), community forests have often been located in less rich forests.
Many other policy orientations have also been promoted: compulsory management plans for concessions and protected areas, reorganization of the institutions involved in forest supervision and plantations, etc.The achievements in terms of disclosure of public information, transparency and accountability have been quite impressive (Topa et al. 2009).The most salient one is probably the institutionalization of independent watching for forest permits allocation.It was rounded out by the introduction on independent monitoring of forest infractions.The last innovation took place first in Cameroon and was implemented later in Congo and DRC, with the hiring of a specialized body, Resources Extraction Monitoring (REM) (which replaced Global Witness in Cameroon as "watchdog").But, to be a driver of changes, the information has to be used by the authorities.In Cameroon, the report of the independent observer of permits allocation mattered in the first years, but it gradually ceased to be used in the dialogue between the Government and the donors.As for the independent monitoring of forest infractions, with all the reports available on the Internet, they did not trigger the changes and the sanctions that could have been expectedprobably because of the diminishing interest in the forestry reform processes during 2000s.
A holistic view of the reform processes and their institutional implications helps to explain why the Bank -at least the forestry team -was seeing the forestry sector, and especially the reform of the concessions system, as an entry point for improving governance in Central Africa, beyond the forestry sector itself (see also Brown et al. 2002).

AN UNFINISHED AND CONTROVERTED REFORM PROCESS
The reform process in Cameroon has never been such as the WB staff envisioned initially.The 1994 Forest Law was not enforced until 1998, primarily because of lack of political will.This was illustrated in 1996 by the discretionary awarding of 6 concessions; this act infringed the 1994 law that stated that all concessions had to be allocated through a public auctioning system.In spite of assurances to the Bank that these cases would remain the exception, the first concessions award in 1997 was fraught with irregularities.This showed that, despite apparent transparency and political willingness, practices had not changed (Ongolo 2015).Until 1998, neither the Bank nor Cameroon had defined adequate legal and regulatory instruments to carry out the forest sector reforms (Topa et al. 2009).

The allocation system criticized and circumvented
A second round of reforms took place towards the end of the 1990s, with conditionalities tied to the SAC III loan.A fiscal reform that also addressed the allocation system has been prepared and partially implemented.After long, tense discussions the Government finally accepted an independent observer (a Cameroonian citizen) of the allocation process.Allocation rounds in 2000 and 2001 were well organized, with financial offers surprisingly high, reflecting genuine competition to access the concessions 13 .But a "learning process" took place in the aftermath; companies in cahoots with corrupt civil servants took advantage of the loopholes in the allocation process (eliminatory threshold linked to "technical criteria", etc.) to eliminate genuine competition in most of the subsequent allocation rounds.In 2005, the allocation was clearly biased, with, in most cases, one competitor remaining listed for a concession and with unusually (suspiciously) low financial offers (CIRAD 2006).
But when the competition was effective, questions were raised about a potential "winner's curse" (Vincent et al. 2005).The long-term commitment (15 years) for the annual area fee (a fixed cost) had to be compared with the annual volatility of the timber market, and the risk a high financial offer would entail for the winning bidder.This triggered criticism from the industry associations, relayed by the French and German governments whose timber companies had vested interests in the sector14 .There are measures that could be adopted for mitigating this risk, such as indexing the area fees on the export price of timber, but such proposals are still part of the "unfinished agenda", along with many other measures for "fine tuning" the regulatory framework (Topa et al. 2009).

Community forestry: the great disappointment
Community forests in Cameroon, which covered over 1.5 million ha in 2011 (Cuny 2011), have been plagued by the "elite capture" phenomenon, a phenomenon facilitated by cumbersome, costly red-tape connected to the creation and functioning of community forests.Since 2005 the CFs, like the industrial concessions, have been subject to an Environmental Impact Assessment (EIA) study before being accepted.(Cerruti and Lescuyer 2011).Some chainsaw lumber operators are collecting their timber in village fallowlands, while others in this informal production circuit who are better equipped have a "small title", a category that is hardly controlled and is known for being a pillar of illegal logging (REM 2010).
The World Bank has always been queasy about "small titles", calling for more stringent regulations but at the same time wanting to keep them for the benefit of Cameroonians and to facilitate the development of small-scale enterprises, especially in the Community Forests.Globally, the WB has overlooked the "informalization" of the forest sector, and has been uneasy in its efforts to formulate appropriate policy responses (beyond the usual calls to "formalize the informal" and remove the red tape that strangles small and medium enterprises).

A Permanent Forest Estate still virtual
Cameroon is one of the rare countries of the sub region with a zoning plan covering most of its dense forest areas.This could have been the basis for the decision to limit the size of the PFE and would have given a clear sign of the will to keep forest-related uses restricted to around 14 million hectares.The gazetting process, which provides the legal basis for granting legal status (PFE) to a mere proposal (the zoning plan), has required two decades.But, except for the protected, previously gazetted, areas, none of the Prime Ministers ever signed decrees for completing the gazetting of around 7 million hectares of production forests (with concessions already granted).This part of the government wants to keep the future land use options open.The WB did not consider this issue extremely seriously until around 2010 when authorisations The EIA study costs around US$13,000 to which the community is to add around $15,000 as an upfront investment before any forest exploitation (Julve et al. 2007).This type of unrealistic requirement paved the way to greater growth in the informal sector and prevented the CFs from becoming competitive on the national market.Cerutti and Lescuyer (2010) showed that chainsaw logging was more profitable than small-scale enterprises operating within community forests.CFs have been used by corrupt officials, in cahoots with illegal loggers for "laundering" illegally exploited timber.The international partners who at one point actively backed the community forestry experience in Cameroon (such as the DFID and the GTZ/GIZ) gradually withdrew their support.

The expansion of the informal sector
The economic crisis that hit the Cameroonian population, especially as of 1992 (cut in public and social expenses) and in 1994 (CFA devaluation), has led to growing informality in economic relationships.The "structural adjustments plans" piloted by the IMF and the WB also contributed to this phenomenon: impoverished citizens turned entirely to the chainsaw lumber operators to buy the timber they needed for construction work.The table below shows the evolution between 1980 and 2004.In the 1980s, exports (expressed as roundwood equivalents) were significantly less than registered production, meaning a large share of the domestic markets was supplied by the formal production sector.This share diminished slightly in the 1990s and at the end of the decade, exports even surpassed registered production! 15he current volume produced by the informal sector is comparable to the volume produced by the industrial sector Source: author

FIGURE 2 Comparison between registered (official) production and registered exports (converted into roundwood equivalents) from 1980 to 2004
Nevertheless, coupled with other regulatory measures, 25.5 million ha were returned to the public domain in 2002, and the attribution of new concessions was suspended (Debroux et al. 2007) through a moratorium that was to remain in force until the completion of a new regulatory framework. . .that still does not exist.The moratorium was to be circumvented repeatedly by the new minister of forests (appointed as of 2003), and millions of hectares were allocated illegally.A "conversion process" of all the concessions was launched in 2006 and applied for a few years, with the aim to legalize concessions allocated according to the law.This participatory process was satisfying, but the results have been a foreseeable compromise between the rule of law and the vested economic interests.And last minute conversions even occurred after the completion of the process, a situation sharply denounced by international NGOs (RainForest Foundation Norway & UK, Greenpeace, Global Witness 2009).Around 12 million ha of concession are now legally attributed in DRC, but the exact situation in the field remains unclear.
A new forest code was promulgated in 2002.It includes provisions for allocating the concessions through a competitive process (but with exceptions!) and paves the way for major development of community forestry.The code also aims to have forest rent shared more equitably with local governments and local communities.In many respects, this new code can be compared with the 1994 Cameroonian law, and has provided opportunities for in-depth reform of the practices in the sector.But most of its provisions have still not been implemented, especially those related to community forestry: the decree has been in the governmental pipeline for 5 years and issued only in 2014.
In DRC, thanks to the experience gained in Cameroon, the WB worked more effectively with civil society, Parliamentarians, the private sector, civil society advocates and environmental NGOs.The difficulties came with some of these latter.The DRC has been highly invested -more than in Cameroon -by a few international NGOs, such as RainForest Foundation, that are opposed to logging16 .Such organizations established strong links with the emerging civil society, providing them with access to Medias, international fora and financial support.Their agenda is to avoid the development of industrial logging in the DRC and to promote community forestry instead.
The zoning plan was, like in Cameroon, supposed to be one of the pillars of the reform in DRC.But the World Bank faced criticisms from Congolese NGOs who denounced the risk of a top-down exercise resulting in the marginalization of the local communities and indigenous people.Hence the Bank decided to defer the zoning plan experiment after it had been planned, in order to concentrate on other issues seen as urgent priorities, such as breaches of the moratorium.This decision eventually landed the WB team in hot water.
for large-scale oil palm plantations and mining permits were granted on the virtual PFE.The WB, nonetheless, insisted for the gazetting decrees being signed, and used notably the lever of the HIPC (Highly Indebted Poor Countries) initiative negotiation in 2010-2011, with only partial success (Ongolo and Karsenty 2015).

DRC: THE UNLIKELY DUPLICATION
Beyond Cameroon, the other country where the WB had a strong lever for fostering reforms was the DRC.Years of civil war exhausted this country.After 2000, the new government solicited external assistance for rebuilding the country.It was, for the WB, a unique opportunity to reshape the forest sector of one of the largest tropical forests in the world (around 140-150 million ha including dry forests and other woodlands).The same WB team that was involved in the Cameroonian reform started working in 2001 to design a new regulatory framework and governance system for the forestry sector in the DRC.But the difficulties were to be very significant, proportional to the size and the complexity of the country.While there were similarities, the WB work done in DRC and Cameroon were different.The similarities stem from the application of key principles (transparency, good governance fiscal accountability, sustainability, benefit sharing, and fight against corruption).The differences with Cameroon related to the fact that, in the DRC, WB forestry assistance was part of a broader re-engagement post-conflict intervention which had to address head-on the legacy of many years of war, destruction, lawlessness and institutional vacuum.This included primarily addressing a situation where most of the productive forests had been arbitrarily assigned to title holders in exchange for political support and other favours.

Overestimation of government capacity and sustained will to reform
One of the first measures pushed by the WB, when starting its forest cooperation with the DRC government, has been to help the government to return into the public domain millions of hectares of concessions that had been granted in the past.On 43.5 million hectares allocated in 2002, numerous concessions totalling 35 million hectares were inactive and kept by their owners only for speculation.(Debroux et al. 2007).One of the instruments for returning these concessions back to the public domain, was the tentative introduction of an annual area fee of $0.5 per ha in 2002.This endeavour immediately spurred opposition from the forest industry (a dozen active companies), whose members owned concessions, some with surfaces over one million ha.A more gradual change in the fiscal system was adopted as of 2004.implementation has been extremely limited.Nonetheless, this statement reveals the beliefs of the IP members: for them, there was a tangible alternative to the concessions system, namely "sustainable use of forests" that was not explored by the WB team and could be achieved through community forestry.
This IP agenda is expressed clearly in the conclusion of the report: "the Panel notes the potential importance of developing a more balanced approach by emphasizing appropriate models of community forests as well as other actions to support community participation, land tenure and use rights in the forests and by linking to the recently proposed Bank administered fund [the FCPF] to pilot instruments for reducing carbon emissions from deforestation and forest degradation [REDD]".This IP report blaming the WB management has had a considerable impact on the modalities of WB interventions in Central Africa.Actually, it marked a turning point in the involvement of the WB in the reforms of the forests sector.The WB became increasingly risk-adverse, and the Central Africa forest policy work was seen as overly complex and prone to reputational exposure.

The loss of the usual levers: from debt reimbursement to investment
The change in the WB intervention methods is also due to the end of the levering effects of debt restructuration.By 2010, the HIPC process had been completed for most of the Central African countries.The "conditionalities" that fostered the reforms in the forest sector are now obsolete.Since the Central African countries are no longer indebted, and their annual growth rates are high thanks to the rise of oil and mineral prices, and new investors from emerging countries (notably China) are knocking at their door so they care far less about the WB and the traditional donors.The Bank now seeks to invest in large sectoral or multisectoral programs such as the forest sector's "Forest-Environment programs".Making disbursements, however, is proving to be extremely difficult due to the inability of the governments to assure the conditions for the correct use and investment of aid money.
In Cameroon, during the 2008-2009 economic crisis, which had a severe impact on timber exports, the government negotiated emergency measures, such as suspending area The Inspection Panel report: the end of WB active involvement in the reform processes Between 1990 and2005, the global agenda on tropical forestry has evolved significantly.New topics moved to the top of the agenda, pushing aside the former concerns about how to regulate concessions in order they deliver sustainable forest management and public revenue.Carbon emissions, illegal logging, and indigenous people: the three "big issues" of today were hardly addressed in the 1990s.The issues discussed in publications and at international conferences and the financial flows have changed in a manner that will impact the WB itself, especially at headquarters.
At the end of 2005, the WB Inspection Panel (IP)17 received a request from the "Organisations Autochtones Pygmées et Accompagnant les Autochtones Pygmées en République Démocratique du Congo".The requesters claimed "that they have been harmed and will be harmed by the forestry sector reform activities supported by the [WB] Projects, which includes a plan for zoning the forests and, they fear, the implementation of a new commercial forest concession system".The request was written in English18 and it is obvious that the whole process was orchestrated by at least one international NGOs.The final IP report, which was published in 2007, blamed the WB management for having neglected several safeguards related to indigenous people and environmental assessments (Inspection Panel, 2007).The report exploited a strategic error of the WB related to the pilot zoning plan mentioned above: the WB had dropped the plan.Since the "conversion process", set up as a response to the breach of the moratorium by the ministry of forests, was meant to legalize the concessions granted on a regular basis, the Panel noted that "in this context, the forest concession conversion process serves as de facto zoning" 19 .
Another criticism focused on the potential for levying forest taxes for the benefit of the State and the local governments.The IP pointed out that the WB "provided estimates of export revenue from logging concessions that turned out to be much too high.This had a significant effect, for it encouraged a focus on reform of the forest concession system at the expense of pursuing sustainable use of forests, the potential for community forests, and conservation".The estimates provided by the WB came from the 2003 "Economic and Fiscal Review" (CIRAD and I&D, 2004).The revenue projections were attached to what turned out to be in retrospect a quite optimistic scenario in which the changes embodied in the new forest law would have been reasonably implemented.But fees, directly with the industry.The WB was left out of the negotiations.The Bank is no longer a driver of change in the forest sector.In 2009 when the President of Gabon imposed a log export ban -which was fully enforced in 2010 -the Bank, admittedly seldom influential in Gabon, was not consulted.

The decline of the traditional forest industry and the rise of Asian investors in land business
Globally, industrial timber production in Central Africa has shrunk.The conjunction of higher taxes, escalating costs and log export bans (fully enforced or not) has squeezed the level of profits in the sector, at least for those who comply globally with the legal rules.The growing scarcity of high quality logs within some traditional species has created many more difficulties for the European operators than for Asian concessionaires who are looking for lower quality logs and can cope more easily with lesser known species.
The forest industry in Central Africa countries has, in the meantime, significantly evolved.Many European companies, once prominent, have been taken over by Asian ones (China or Malaysia).The industry split into four groups: (i) those who export mainly to Western Europe and have gotten or sought to get a FSC certification (7 companies on 5.5 million ha), (ii) those who are operating on various markets and are not looking for certification but will try to comply with FLEGT requirements, (iii) those who are operating in a "grey area", pay all taxes but do not comply with the requirements of management plans and wood processing (many Asian companies), (iv) those who are working on "small titles" and are barely complying with the law.
The entrance of the Singapore-based company Olam (a conglomerate controlled by Indian businessmen working in agri-business rather than forestry) in the forest sector in Gabon and the Republic of Congo (buying back the CIB, one of the largest FSC certified companies in the world) is significant of this shift in power from the European companies to the newcomers from emerging countries (Karsenty 2011, Cerutti et al. 2011, Putzel et al. 2011).The strategy of these profitable conglomerates is to enter sectors where profits used to be significant in the past but limited at present (the former Danish owner of the CIB recorded financial losses during the 4 years he owned the company).Questions about the midterm intentions of these firms also need attention.Olam, which is developing oilpalm and rubberwood plantations in Gabon, probably does not plan to earn a lot of money in forestry in Central Africa.Controlling a few million hectares of land (today forested) might be an advantage for developing various types of plantations in the future.

The doublespeak regarding REDD+
REDD+ has been a prominent issue since the mid-2000s for the Congo Basin countries.The UN-REDD focused especially in the DRC, where the Ministry of Forestry seemed to be particularly interested at becoming a regional REDD leader.A lot of papers and studies have been produced so far -mostly by international consultants -but nothing has changed in the field (Brockhaus and Di Grigorio, 2015).
Gabon, Cameron and the Republic of Congo are willing to become "emerging" countries at the horizon 2030-2035.They are counting on their oil, gas, mineral resources and large-scale agribusiness to reach this goal, not on the forests, where the potential to generate public revenue through taxes is comparatively low.When Gabon, historically one of the world largest log exporter, decided to implement a log export ban in 2010 (essentially for creating jobs in the wood processing industry), the government accepted the sacrifice of almost 80% of the forest fiscal revenue ("less than one month of oil receipts" said an official at a workshop in 2010).In Cameroon, oil palm permits have been granted on forest concessions (even on FSC certified ones), the government is reluctant to complete the gazetting process of Forest Management Units and barely conceals its preference for other land use options (Ongolo and Karsenty, 2015), and mines are sometimes dug in protected areas (Hoyle and Levang 2012).
Paradoxically, the deforestation rates in Central Africa have slowed by a third between 1990 and 2010, even though a recent acceleration is reported by Global Forest Watch20 for 2013 and 2014 essentially in the DRC.Less than 200,000 ha of dense rainforest were lost annually between 2000 and 2010 compared to almost 300,000 ha in the 1990s (Mayaux et al. 2013), in spite of the increase of the population.The limited if not the absence of public investments in infrastructure and agriculture during the decade is probably the main cause of this evolution, certainly not the forest-related public policies that the governments would have implemented in the name of REDD+.
At first, Central African countries hoped they could be remunerated under the REDD+ scheme for their "past efforts" which they claim to have contributed to conserve the forest, thus rendering an "ecosystem service" to the planet.They gradually understood that the REDD scheme did not allow payments for a "carbon stock"21 .Instead, they would have to negotiate convenient baseline scenarios in the hope of being rewarded for "deforestation below the prediction".COMIFAC proposed an "adjustment factor" that would allow Central African countries more deforestation without sacrificing remuneration.The World Bank and the GEF launched a sub-regional program for REDD+ preparation, which is dedicated to institutional reinforcement and studies.But nothing has emerged so far regarding forest-related policies (land tenure and land-use planning, energy and transport schemes, agriculture policies, population planning, governance, etc.) or measures needed to curb or contain deforestation and degradation in the near future.

AMBIGUOUS OUTCOMES
Assessing the outcomes of around 15 years of WB endeavours for reforming the forest concession's regime in Central Africa cannot be exempt of controversy.The last example to date was provided by the 2013 report of the IEG, which is the evaluation team of the World Bank (Evaluation of World Bank Group Experience 2013).The evaluation contains criticisms against the WB work on "concession reforms in tropical moist forest countries with weak governance"22 .The Guardian summed up the IEG conclusions as follows: "The IEG panel (. ..) criticised the bank strongly for: (i) Continuing to support industrial logging, (ii) Not involving communities in decision-making, (iii) Assuming that benefits would accrue to the poor rather than the rich and powerful, (iv) Paying little attention to rural poverty" (J.Vidal, 29/01/2013).All the attention of the observers has been caught by two paragraphs related to the West and Central Africa WB policies towards concession reform: "In tropical moist forests in Central and West Africa, World Bank support for industrial timber concession reforms has helped to advance the rule of law, increase transparency and accountability (as compared to the "without the Bank" scenario) and put environmental standards in place.Evidence is lacking, however, that these reforms in tropical moist countries with weak governance have led to sustainable and inclusive economic development.
Attention to rural poverty has been lacking in World Bank supported concession reform projects.World Bank policy advice and projects that have supported the reform of industrial timber concession regimes have usually neglected or underestimated the nontimber values and uses of the forests with respect to the livelihoods of forestdependent people, their traditional claims, sociocultural values, and overall sense of security.Evidence is also lacking that concessioned natural forests are being managed sustainably".
Actually, the 2013's IEG report is walking in the footsteps of the 2007's IP report on DRC.It shares the same beliefs about the possibility of an alternative model of forest management based on community-based management -even if it is much more prudent than the 2007's IP report about the capacity of REDD+ to be instrumental in building such an alternative.It is striking that the IEG report did not mention the numerous studies that refer to trouble with community forestry in Cameroon (Ndanyou and Majerowicz 2004, Vermeulen et al. 2006, Lescuyer 2007, Julve et al. 2007, Ezzine de Blas et al. 2009, Cuny 2011, Dehu 2012) 23 and suggest that community forests in Central Africa are embedded in the governance problems of the African States.The functioning of the community forests is plagued with the difficulties of collective action (free riding and accountability) in societies where the "exit option", allowed by the limited density of population, has prevented the design of strong collective rules for sustainable use of the resources.The IEG report does not mention the meta-analysis of the literature by Robinson et al. (2011) which shows an "association between negative forest outcomes and communal land in Africa" (p. 22).
The IEG report also points out that "By neglecting the informal sector, the World Bank has missed an opportunity to reach more forest-dependent rural poor, while at the same time helping to achieve more environmentally sustainable forest management".This difficulty to deal with the "informal sector" has clearly been one of the weaknesses of the WB intervention in Central Africa.It is probably not specific to the forest sector: the WB is more prepared to talk with governments and formal companies than with the informal actors of the economy.But the "informalisation" of the forest sector is not an exception: the share of informal employment in non-agricultural employment, excluding South Africa, rose to 78% in sub-Saharan Africa, and is still tending to increase (Chen Alter 2008).The avoidance of "formalization" has something to do with the "predatory behaviour" of the administrations that prevent the development of legal small-scale enterprises (importance of the "parafiscalité").But it is also a way of coping with social and family pressures for jobs and financial aid.Such pressures are pushing the small and medium entrepreneurs into informality.Avoiding the creation of formal businesses allows them to "pretend to be poor" vis-à-vis their entourage (Baland et al. 2011).
The implicit hypothesis of the WB was that the development of community forestry should absorb the chainsaw logging activity, resulting in the creation of small-scale community enterprises to supply the domestic markets.But this did not happen.Chainsaw loggers, often professional operators, are more interested in higher quality logs they can select wherever they want.When community forests are timber-oriented, their production costs are higher than the timber costs in informal sector (Lescuyer et al. 2011).Community forests' enterprises prefer exporting rather than selling on the domestic market, which pays less.Finally, the administrative red tape encumbering the functioning of the community forests put hurdles on their legal production.Consequently, the needs of the domestic markets are still not met by legal timber, but by the informal operators.

The predominance of economic approaches over tenure and the informal economy issues
Taking a broader perspective, such debates reveal the importance of the underlying paradigms, especially when evaluations cannot rely on "counterfactuals" ("without the Bank") scenarios, as is often the case in policy evaluations.The WB intervention in Central Africa has been shaped by the "economics of concessions" debates of the 1990s, using a specific "theory of change".In this theory, what matters for achieving SFM was the incentives given by the pricing system (hence, the importance given to fiscal studies) and the allocation of harvesting rights.The agents who mattered were the concessionaires (including community forestry enterprises) and the governments that established the incentives systems.Land and forest tenure issues were not disregarded but considered as matters to be addressed after the concession regime being regulated.Informality was difficult to handle within this analytical framework, since the informal actors bypassed the regulations whose design is the core of the reform processes.
Combined with changes in costs of operations and in the international timber markets, one of the effects of the changes in the forest sector based on the reforms (notably compulsory management plans, norms, higher taxes, an auction system for permits allocations), has been the double movement of concentration (of the export-oriented sector) and fragmentation (of the domestic-oriented sector) with the expansion of the informal sector (Karsenty 2007).This dynamic has been overlooked in the "theory of change" of the WB24 .
Poor articulation of the agendas of zoning plans, community forestry and small-scale enterprises has been a major weakness in WB interventions.Forest sector reforms should be conceived and conducted in close connection with evolutions in the legal framework regulating land and tree tenure.Providing there is a political will to address also this sensitive issue.

Lack or ownership or insufficient capacities to generate coalitions?
WB policies resting on "conditionalities" have been criticized for not inducing ownership.This argument is only partly true, since it is a common practice for governments to be fully part of the reform process (as we have seen in the case of Cameroon) but at the same time to argue that they "were forced" to adopt measures that hamper their vested interests.Nonetheless, Brunner and Ekoko (2000) were right in noticing the overly exclusive focus on dialoguing with the executive branch of the government.Representatives of the parliament have often been neglected -leading to some surprises during the legislative process.
The Bank also tended to believe in spontaneous support for its actions by local NGOs flagged as "pro-poor", in favour of conservation and local communities.In reality, the economic paradigm underlying the WB's "theory of change" is generally not well understood by the local NGOs.In Cameroon, many NGOs opposed the auction system for concessions on the grounds that it would favour wealthy foreign operators over national ones25 .The Bank probably relied too much on the relationships established with key members of governments and it neglected creating coalitions for introducing reforms that included large parts of the civil society and the parliamentarians.Changes in governmental coalitions due to changes in the ministries' senior staff have often been synonymous to loss of leverage for reform processes.

The way forward: rewarding or investing?
REDD+ as a "performance-based" payment scheme is particularly appealing for donors since it relies on financial incentives but does not need any "conditionalities": the governments are completely sovereign in deciding on the means and policies they will implement for curbing deforestation.In principle, donors are in a more comfortable position: they just have to invest upfront for institutional building, and "the carbon market" will reward ex post for the "performance" reached (by comparing against a baseline scenario).The complex task of designing policies and verifying the implementation of the reforms are no longer the business of the donors, who, thus, can position themselves as mere lenders for the countries seeking to take advantage of the carbon market.
Beyond the question of how to fabricate the "performance" from scratch (Karsenty 2008), the difficulty with a Central African country is that the governments seem more interested in large-scale investments they can attract for agriculture, mining and infrastructure than in undertaking the (socially) costly reforms that would be required to conserve their forests.And, assuming that these countries had the will to really take action to protect their forests, most countriesstarting with the DRC -simply do not have the institutions needed to design and implement ambitious transformations, such as a land tenure reform or new agriculture policies.Donors will have to invest heavily and over a long haul not only in institution-building, they will also have to participate in designing the reforms they will be requested to sponsor, at least partly.
Whatever the likelihood of a future REDD+ scheme delivering offsets for the international carbon markets, it seems more and more evident that the public funding associated with REDD+ will have to be sustained and probably increased through time.This will be to give the poorest countries the financial means needed to address the domestic drivers of deforestation and to pay for the reforms of forest-related policies that are critical for tackling the issue.Phase II of REDD+ is designed along those lines and the recent Central African Forest Initiative (CAFI) is to be seen in this perspective.It is unlikely that most of the countries can move quickly to Phase III, which is performance-based."Investing" will be the key word for long to come, the "reward" dimension remaining a possible feature of the REDD+ mechanism once the numerous difficulties (additionality of the reductions, risks of leakage, permanence issue, side-effects on biodiversity, opposition of social NGOs, etc.) have been overcome.And as long as public aid for investment is concerned, donors will want to monitor the effective and equitable use of their money.In the end, with the exception of some emerging countries such as Brazil that might have the capacity to fund their reforms themselves, the REDD+ public sponsors will be facing the same issues as the WB faced in Central Africa during its "forest policy reform" years.Rather than unconditional reward once the "performance" has been monitored, i.e., achieved, as is often expected from REDD+, conditionalities that orient the sponsors' investments will be the cornerstone in the decision to invest public funds for implementing forest-related policies.

CONCLUSION
To qualify the situation prevailing in the forest sector in Central Africa, it is not uncommon to hear that the regulatory framework is globally suitable, but the implementation is failing due to a lack of political will and limited effectiveness of the public institutions.Even though the framework still needs many improvements, this statement accurately summarizes the achievements and limitations of the reform processes in Central Africa during the two last decades.
The issue of ownership, and more precisely in-depth and sustained ownership of the reform processes, remains central.But this issue goes well beyond the commitments of the President, the Minister of Forests and a small group of high-level officials.Increasing transparency and accountability jeopardizes vested interests entrenched in the public administration.These interests have the capacity to resist silently and to undermine the reform processes from the inside.On the other hand, convincing the public opinion of the benefits of the reforms for more equity and effectiveness is a key achievement when hoping to circumvent the passive resistance in the inner circles of the public administration.The WB probably did not pay enough attention to the in-depth work that needed to be done with journalists, leaders of associations, parliamentarians and academics, to explain the ins-and-outs of the reform processes26 .As for the NGOs, it is not enough to "bring them on board" through consulting contracts as it tried to do.Providing training for their members -especially in economics and forest ecology/management and institutional change -would be extremely useful to developing a shared understanding of the stakes associated with the reform processes.
Another dimension that is critical for the reform processes is enhancing the rule of law through accountability, downward accountability from the public administration to the citizens and from the citizens who hold the public administration accountable.This would help in sustaining the reform processes, regardless of changes at the top echelons of the ministries.Checks-and-balances in institutional functioning can be encouraged through support to parliamentarians and citizens' associations tasked to investigate and report publicly on the implementation of the rule of law.In 2005, the publication of the Lutundula report27 on the irregularities in the mining contracts granted to companies in DRC, was instrumental in raising the awareness of Congolese citizens to the importance of what was at stake in the concession allocation process in the mining sector.
The ultimate challenge for donors and committed governments will be to shift from the traditional approach in terms of national public forest policies to a new paradigm in terms of the coherence of forest-related policies (Singer 2008); this includes all the public and private policies that affect forests, directly and indirectly (land tenure, agriculture, transports, land-use planning, energy policies, etc. see Lele et al. 2013).The new paradigm does not only concern forest-related policies in developing countries, but also policies and consumption patterns in industrial countries, as was recently pointed out regarding the significant impact of EU policies and consumption patterns on deforestation (European Commission 2013) 28 .The limited capacity to take action beyond traditional sectoral boundaries is a traditional weakness of both the organizations and the ministries in developing countries.But, if tropical deforestation is to be reduced or contained, coherent cross-sectoral policies coupled with responsible trade and consumption policies in developed countries, will be essential.

FIGURE 1
FIGURE 1 Density of population in the Congo Basin