Beyond opportunity costs: who bears the implementation costs of reducing emissions from deforestation and degradation?

We found wide variation in the number of organizations involved in the start-up phase of the different initiatives. The Brazilian and Indonesian initiatives that cover entire jurisdictions involve the most organizations, but there are also a large number involved in a donor-funded demonstration project in Indonesia (the Kalimantan Forests and Climate Partnership or KFCP).

As shown in Table 2, international organizations played a significant role in the implementation of all but five of the initiatives, and national and subnational organizations were involved in all but four (organizations that operate in an area larger than a single village but smaller than an entire country are considered subnational). In contrast, there were no local (village) level organizations significantly involved in the start-up phase of more than half (13 out of 22) of the initiatives.

Civil society organizations were involved in the start-up phase of all initiatives except one, while government agencies were involved in 14 initiatives, and private sector organizations were involved in 11 initiatives. Table 2 suggests three major categories of initiatives based on the sectors involved in their start-up phase, i.e., those implemented by

There is an apparent negative relationship (not statistically significant in our sample of 22 initiatives) between the importance of forest management in the strategy for a REDD+ initiative and the involvement of civil society institutions. Specifically, the larger the percentage of civil society institutions involved in an initiative, the less focus on sustainable forest management. Sixty-four percent of organizations that do not promote SFM (except to prevent illegal logging) are from civil society, compared to 35% of the organizations in initiatives that have certification as a core element of their strategy. On the other hand, the percentage of implementing organizations that are for-profit increases with the initiative’s degree of attention to SFM (from 11% in initiatives that do not actively promote SFM to 27% in initiatives that are pursuing certification).

We also note a possible relationship between an initiative’s approach to the carbon credit market and involvement of the private sector in its implementation. Across initiatives that have sold carbon credits, more than a third (37%) of the implementing organizations are for-profit, and no government institutions are involved. In contrast, initiatives opposed to credit sales had no private sector involvement. Across the four initiatives that had expired by the beginning of 2015, a third of implementing organizations were public sector, and only a fifth (21%) were from the private sector.

The 139 organizations involved in the start-up phase of the 22 initiatives in our sample provide a picture of who is involved in the implementation of REDD+ on the ground (counts given in Fig. 2). There are only 19 for-profit organizations participating across all 22 initiatives, including private investors, firms providing management or technical services such as carbon accounting, financing, and forestry expertise, and companies operating in the carbon offset market. Civil society organizations are the most frequently occurring (60/139) and government the second most frequently occurring (43/139). In addition to organizations from the public, private, and civil society sectors, we identify a small number (17) of organizations that are best characterized as donors or research institutions, such as universities.

Considered by level, the largest numbers of organizations involved in the start-up phase of initiatives operate at the subnational (45 out of 139) and international (42 out of 139) levels (Fig. 3). When we compare the level and sector of these organizations, we find that most subnational organizations (71%) are government agencies, whereas most of the international organizations are from civil society. Private sector organizations are either national or international. The few village level organizations (n = 13 across all initiatives) are mostly from civil society (77%). Subnational organizations are most prevalent (50%) among initiatives funded primarily by the government, and national organizations are most prevalent among initiatives funded primarily by NGOs. Three of the four initiatives that lacked national level organizations had expired as of 2015. It is also interesting to note that initiatives funded primarily by government sources involve more subnational and for-profit organizations. This suggests that government funding for initiatives may be associated with broader participation, including private sector and subnational institutions.

To assess the concern that the costs of REDD+ will fall on institutions in REDD+ countries, we assess which types of organizations are more likely to (i) have their costs fully covered (or possibly more than fully covered) by the initiative budget, and which are more likely to (ii) share the financial burden of implementing the initiative.² A high percentage of village and subnational level organizations are burden-sharing (62 and 40%, respectively), while 60 and 67% (respectively) of international and national institutions cover their costs (Fig. 4). Comparing across sectors, we find that burden sharing is most common among government (56%) (Fig. 2). Thus, we find some basis for concerns that REDD+ country institutions will end up shouldering part of the cost of REDD+, at least in the start-up phase.

Initiatives funded primarily by governments in their start-up phase involve a much higher percentage of burden-sharing institutions (56%) compared to initiatives funded primarily by NGOs (16%). Thirty-four percent of institutions involved in donor funded initiatives and 46% of institutions involved in private sector funded projects are burden-sharing. At least in our sample, donor and NGO funded initiatives are most likely to balance compensation with actual costs incurred from the beginning of the initiative, whereas the government funded initiatives are more likely to involve organizations that do not fully recover their costs, at least during the start-up phase. However, this could be due to the fact that all government funded initiatives in this sample occur in Brazil and three out of four of these initiatives are opposed to credit sales (Table 1). Initiatives with lead proponents opposed to sale of carbon credits have the highest percentage (50%) of organizations that are burden-sharing. Thus, opposition to credit sales is accompanied by willingness to share the costs of REDD+, as might be expected. Another pattern is that initiatives with a stronger focus on SFM have a higher percentage of organizations that have their costs fully covered, e.g., 73% of organizations involved in initiatives that focus on forest certification cover their costs (Fig. 5).

There are 139 organizations are involved in the implementation of the 22 subnational REDD+ initiatives in our sample. Based on a review of carbon forestry projects, Milne (1999) suggests that involving a large number of parties in project implementation increases implementation costs. This raises the question of why there are so many organizations involved in REDD+ initiatives. In our data, there is a weak, but positive, correlation between the number of implementing organizations and the proportion of those organizations that are burden-sharing (i.e., cost-sharing the implementation of REDD+) in any given initiative. That is, the more organizations involved, the more likely there is to be cost-sharing. Alston and Andersson (2011) suggest that the costs of REDD+ can be reduced by encouraging a variety of organizations to take on related tasks. Adding organizations can also help reduce total costs if they have complementary expertise (e.g., staff with different backgrounds and training). Furthermore, they may be able to capture additional resources available only to certain types or levels of organizations. Agrawal et al. (2011) suggest that efforts to promote complementarity of interests and capacities among different actors “may help achieve multiple objectives that REDD+ efforts have come to symbolize for different stakeholders.”

Start-up of the REDD+ initiatives that we studied has been supported by numerous government institutions. In particular, many government entities bear non-monetary transactions costs, such as staff time, which have also been identified as important hidden costs in community forestry (Schreckenberg and Luttrell 2009). Government subsidization is more apparent in Brazilian initiatives and less common among Indonesian initiatives, both because there are more government institutions involved in Brazil and because many of those institutions are cost-sharing. The willingness of both national and subnational governments to bear some of the costs of establishing initiatives suggests that their goals go beyond immediate cost-recovery. For example, governments may seek to use the early initiatives to build readiness and support for REDD+, or to generate non-carbon co-benefits. An important area for further research is whether these public investments in REDD+ are “crowding out” or “crowding in” investments by the private sector and civil society (cf. Andreoni and Payne 2003; Albers et al. 2008).

A diverse range of stakeholder institutions, across all governance levels, were involved in the start-up phases of most of the initiatives studied. Subnational and international organizations are particularly numerous. The burden on subnational institutions was predicted by Nepstad et al. (2007), who highlighted the costs that the federal and state governments in Brazil were likely to incur to reduce emissions. Brazilian initiatives are supported by a particularly large number of subnational organizations: all but one of the Brazilian initiatives that we studied have more subnational than national or international organizations significantly involved in implementation. This is related to on-going decentralization of environmental governance in Brazil, including federal efforts to hold local governments responsible for slowing deforestation, e.g., through a “blacklist” or “embargo” of municipalities that deforest the most (Assuncao and Rocha 2014; Cisneros et al. 2015). Our data also reveal a large number of subnational actors involved in Indonesian initiatives, in contrast to arguments that this level has been excluded from REDD+ policy discussions in Indonesia (Gallemore et al. 2015). Phelps et al. (2010) also warned that REDD+ could encourage recentralization of forest management across the tropics, due the requirements of performance-based financing. However, our data suggest that even if policy discussions are dominated by national level organizations, organizations at the subnational level are highly involved in implementation. Indeed, four of the initiatives in our sample (in Tanzania and Brazil) had no significant involvement of national organizations.

Some observers have argued that in response to the substantial costs of REDD+, a polycentric (or multileveled) system is emerging (Ostrom 2010, 2012). They argue that multilayered, collective action problems need global institutions to channel finance, but local institutions to monitor management and distribute benefits (Corbera and Brown 2008). Ideal polycentric systems benefit from the complementary strengths of multiple institutions (Andersson and Ostrom 2008; Nagendra and Ostrom 2008; Agrawal et al. 2011), keeping transaction costs low (Alston and Andersson 2011; Leifeld and Schneider 2012). The variety of levels and types of implementing organizations in our sample is consistent with a polycentric system (Ostrom 2010, 2012). Reynolds (2012) examines variation in local, subnational, and national level institutions and finds that working across levels can make carbon forestry projects more likely to succeed. This is echoed in our finding that the initiatives that have expired did not have implementing organizations at all levels and from all sectors, with gaps especially in national level and private sector representation.

While the initiatives in our sample involved organizations from a variety of sectors and levels, there is one notable gap: in many (13 out of 22) initiatives, particularly in Brazil, Peru, and Indonesia, there were no village level institutions involved in the start-up phase, or in other words, the core of the polycentric governance system was hollow. The exception was Tanzania, where, in most initiatives, village level organizations were most common. This may reflect the way in which Norwegian funding for REDD+ in Tanzania specifically targeted local capacity building, in order to move forward while national institutions and rules for REDD+ were still being developed (Rantala et al. 2015). In the short-run, the hollow core does not appear to have been a major stumbling block for initiatives: there were no village organizations and relatively few subnational organizations involved in implementation of initiatives that were certified and that sold credits (one definition of “success”).

Cost compensation is a prominent part of the discourse on REDD+ benefit-sharing (Luttrell et al. 2013). A commonly held view is that international funding for REDD+ should cover both the opportunity costs of landowners and the implementation costs of the organizations involved, thus matching benefits to the costs incurred. Failure to cover costs could lead to a motivation deficit, and thus a sustainability problem. However, we find that not all organizations behave according to this “cost compensation” logic in the start-up phase of initiatives. For example, many (40%) subnational organizations supporting the implementation of initiatives are bearing the costs themselves (Fig. 4). This presents a stark contrast to predictions that funding for REDD+ would be diverted to support conservation efforts and other national and local priorities (Pagiola and Bosquet 2009; Harvey et al. 2010; Phelps et al. 2011). Rather than siphoning off REDD+ funds to support related activities, the organizations involved from both the public sector and civil society are sharing the costs of implementation.

This willingness to share the burden perhaps reflects a strategy of getting REDD+ up and running, with the hope that it will eventually generate a surplus. Specifically, governments may be willing to use public funds to reduce emissions, with the hope that their investment will leverage other investment or REDD+ funds that can be shared as benefits (e.g., for forest stewards). A second possible explanation is that governments (and local organizations) have the greatest interest in securing non-carbon co-benefits from REDD+ and are thus willing to invest in its establishment. Two of the four initiatives funded primarily by private sources indicated that non-carbon benefits were their primary objective. The proponent of one of these (Ulu Masen) stated that its primary objective was poverty reduction, while the proponent of Rimba Raya stated that its primary objectives were conservation, biodiversity, and SFM (Table 1). Both of these initiatives aimed to sell carbon offsets and thus may have been interested in generating co-benefits partly to secure access to the voluntary carbon market. Dixon and Challies (2015) also find a range of motivations among private sector investors in REDD+ in Indonesia. They suggest that those who seemed to be willing to derive little or no immediate financial return were the most active in REDD+, while those who prioritized financial profit had typically suspended or slowed their activities or spread their risk across other investments.

A third possibility is that host governments view investment in REDD+ as part of voluntary national contributions to climate change mitigation (i.e., nationally appropriate mitigation actions (NAMAs) and nationally determined contributions (NDCs)) (Boos et al. 2015; Fridahl et al. 2015). This could turn REDD+ into an unfunded mandate for subnational governments. Alternatively, it could be that government institutions are willing to share the costs in order to obtain more influence and control over implementation, thereby avoiding the paternalistic relationships associated with international financing (Sanginga et al. 2007).