Explaining the recent reduction of Indonesia's deforestation

Angelsen, Arild; Dermawan, Ahmad; Ladewig, Malte Research Report Explaining the recent reduction of Indonesia's deforestation Centre for Land Tenure Studies Report, No. 01/25 Provided in Cooperation with: Centre for Land Tenure Studies (CLTS), Norwegian University of Life Sciences (NMBU) Suggested Citation: Angelsen, Arild; Dermawan, Ahmad; Ladewig, Malte (2025) : Explaining the recent reduction of Indonesia's deforestation, Centre for Land Tenure Studies Report, No. 01/25, ISBN 978-82-7490-330-2, Norwegian University of Life Sciences (NMBU), Centre for Land Tenure Studies (CLTS), Ås, https://hdl.handle.net/11250/3176540 This Version is available at: https://hdl.handle.net/10419/311679 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ Centre for Land Tenure Studies Report 01/25 ISBN: 978-82-7490-330-2 Norwegian University of Life Sciences (NMBU) Explaining the recent reduction of Indonesia’s deforestation Arild Angelsen, Ahmad Dermawan and Malte Ladewig Explaining the recent reduction of Indonesia’s deforestation By Arild Angelsen, Ahmad Dermawan and Malte Ladewig School of Economics and Business Norwegian University of Life Sciences, Ås, Norway [email protected]; [email protected]; [email protected] 1 Foreword This report is prepared for by a team at the School of Economics and Business, Norwegian University of Life Sciences (NMBU). The report includes data publicly available by the end of May 2024, although news stories until the end of June 2024 have been included. We are grateful for comments and critical inputs from David Gaveau (Nusantara Atlas), Liz Goldman (GFW, World Resources Institute) and seminar participants at the Rainforest Foundation Norway. We also thank the 32 resource persons in Indonesia that spend their valuable times in the interviews. The assessments and views presented in the report are those of the authors. The report is part of the BEDROCK project: https://bedrock-project.earth/ Ås, Norway, December 2024 2 Execu�ve summary Deforestation in Indonesia has declined sharply since 2015-2016. We use a combination of spatial land use change data, interviews with stakeholders and statistical analysis to explain this reduction. The exact extent of the reduction in deforestation depends on the data set used, and the years or time periods compared, varying between 40% and 90%. A robust estimate is that deforestation has fallen by at least 50% since ca. 2016. The decline is uneven across islands and provinces. The two main “deforestation islands” Kalimantan and Sumatra experienced a major decline, with a relatively larger decline in the former. The provinces with the largest absolute decline were Riau, South Sumatra and Central Kalimantan, followed by West and East Kalimantan. In terms of commodities driving deforestation (“direct drivers”), palm oil is still the most important commodity (46% for 2018-2022, according to MapBiomas Indonesia data, compared to 55% for the period 2010-2017). Pulp is on the rise as a deforestation driver, particularly in Kalimantan. In the three high deforestation provinces on that island, the demand for pulp was behind almost 1/3 of the natural forest conversion in 2022. Mining for coal and valuable minerals such as nickel is also on the rise, with MBI data suggesting its share of national deforestation being 5.4% (2021-2022), compared to less than one (0.9%) for the 2010-2017 period. Mining is particularly important in Sulawesi, where it according to one estimate accounted for 30% of the deforestation in 2021-2022. Five hypotheses are put forward to explain the reduction, related to public policies, private (corporate) policies, civil society pressure, commodity prices, and forest scarcity. New public (government) policies have been a main reason for the decline. The moratorium of new permits of primary forest and peatlands in 2011 has had an impact, but also took time to produce an effect on the ground effect (in part as it was a ban on issuing new licences). Interviewees also stressed that better sectoral coordination related to the moratorium and other reforms has been key for the slowdown. A second set of reforms relate to fire management and peatland protection, sparked by the devastating 2015 forest fires. The interviewees highlight this as a key policy reform by, for example, increasing the accountability of subnational government officials for fire management. Results-based payments (RBP) or result-based contributions were central in the Letter of Intent between Indonesia and Norway in 2011 and in the new MoU of 2022, although the first payment was not made before in 2022. An RBP-based project of more than USD 100 million was approved by the Green Climate Fund in 2020. RBP has also been implemented at subnational level starting in East Kalimantan and Jambi with external donor support. Some 3 reports indicate positive results, although they are likely to be too location-specific and came too late to have had major impacts on the observed decline in national deforestation figures since 2016. Private regulation such as certification and corporate pledges show promising signs with an increasing share of oil palm plantations being certified. Yet, breached are reported, and the forest encroachment factor (i.e., the share of new land being converted from forest) has not dropped as much as to be expected. Moreover, certification of pulp production and mining is lagging behind. Civil society pressure has played a role in particularly two areas, although the exact role and contribution are hard to assess. First, CSOs are active actors on the policy arena, also to influence private sector initiatives and policies. Second, CSOs are important watchdogs for both implementation of public and private regulations and pledges. This has helped “bringing the forests to the court”, as one interviewee observed. Key commodity prices were relatively stable during the 2016-2020 period, and thus cannot explain much of the decline. Yet, no major increase in the prices of deforestation-risk commodities made the implementation of both public and private policies less costly (both for politicians and producers), and made violations of laws and regulations less profitable, increasing policy effectiveness. The prices of coal and minerals such as gold and nickel have, however, increased steadily over the period. The nickel price quadrupled between 2016 and 2022. The increasing role of mining in deforestation can largely be explained by the high and increasing profitability, and demand is likely to grow steadily in the coming years due to the global energy transition. The statistical analysis suggests that up to 1/3 of the reduction in deforestation can be explained by forest scarcity (forest transition). When a province hits ca. 40% forest cover, deforestation tend to decline. While we conclude that all five hypotheses put forward are relevant to explain the recent decline in deforestation, we tentative conclude that public polices – in combination with a forest transition (scarcity) effect – has been the most important factors. Private policies show more mixed results. Behind the public and private policy changes, civil society has played a key role in policy reforms and in promoting more effective implementation though its watchdog role. Non-increasing agricultural commodity prices have made the implementation of forest conservation policies less costly. Deforestation is still a profitable activity for land users and (sub)national governments, and forest conservation is a continuous battle, with future challenges emerging: the effectiveness of current policies may weakened over time, political priorities may change, prices of deforestation-risk commodities may rise, and the composition of direct drivers change – requiring a shift in the policy focus. 4 Recommendations 1. Be ahead of the curve: The forest transition suggests a natural development with decreasing deforestation in low forest cover regions (provinces) and increasing deforestation in high forest cover regions. The major future deforestation threats have moved eastwards in Indonesia. Mining is likely to be a key driver of deforestation in Sulawesi, while food estates, oil palm and pulpwood plantations in Papua are emerging as direct drivers in new deforestation hotspot. 2. Incentives for high-forest, low-deforestation areas: Much focus in the REDD+ discussion has been on reducing emissions from high-deforestation areas, and rightly so. Taking a more long-term view and preventing future increases in deforestation, mechanisms that incentivise the conservation of high-forest, low-deforestation areas are also needed, while noting the challenges of estimating additionality – particularly if carbon credits from these areas enter carbon trading. 3. New deforestation-risk commodities: The deforestation debate in Indonesia has traditionally focussed on timber logs and palm oil. Timber production is in decline, and the share of palm oil-driven deforestation is also declining, although oil palm cultivation is still the main immediate land use after forest clearing. Demand for biodiesel based on palm oil may make it maintain its role as the no. 1 deforestation-risk commodity. Pulp and mining (nickel) have increased its share of direct deforestation drivers, and with pricing of key minerals on the rise this pressure is likely to continue. Plans for large food estates is also likely to make it an increasingly important driver. 4. The energy transition and deforestation: The global energy transition has increased the demand for both renewable energy (palm oil and wood pellets) and minerals (such as nickel). Paradoxically, these also pose a threat to Indonesia’s natural forests, and balancing these trade-offs remains a major challenge both for policy makers and advocacy groups. 5. CSOs watchdog role important for effective implementation: The private sector initiatives are commendable, but their impact hinges on their effective implementation. Clearing new forest land remains a profitable option from a business perspective, and without clear sanctions of violations (both formal and informal such as reputational risk) illegal and semi-illegal forest conversion is likely to continue. 6. Better data on “other agriculture” commodities: The share of “other agriculture” as a direct driver is increasing, yet this remains a black box in many data sets and analyses. More data on what constitutes this driver and their relative importance (including cocoa, coconut, coffee, rice, rubber sugar) are needed to design and implement targeted policies, also related to the EUDR. 7. Transparency in data sets and their uses: Several data sets on deforestation or tree cover loss are available, with quite different primary data sources and forest/deforestation definitions. This creates a wide range of figures, opening for cherry picking and selective uses. Full transparency on definition and data transformations is needed to enable better comparison and detect underlying trends. 5 Contents Foreword .................................................................................................................................... 1 Executive summary .................................................................................................................... 2 Recommendations ..................................................................................................................... 4 1 Introduction ........................................................................................................................ 7 2 Approach, methods and data .............................................................................................. 7 2.1 Analytical approach and hypotheses .......................................................................... 7 2.2 Methods...................................................................................................................... 9 2.3 Data sources ............................................................................................................. 10 3 Disaggregation of deforestation ....................................................................................... 12 3.1 Deforestation over time............................................................................................ 12 3.2 Deforestation by location ......................................................................................... 15 3.2.1 Main islands ..................................................................................................... 15 3.2.2 Provinces .......................................................................................................... 17 3.3 Deforestation-risk commodities (direct drivers) ...................................................... 19 3.3.1 Land use transitions ......................................................................................... 19 3.3.2 Forest encroachment factor .............................................................................. 21 3.4 Synthesis of direct drivers ........................................................................................ 25 3.4.1 Oil palm driven deforestation slowed down, but still plays a major role ........ 25 3.4.2 The emergence of pulp wood as a deforestation driver ................................... 26 3.4.3 Mining surging ................................................................................................. 26 4 Policy changes and the links to deforestation .................................................................. 27 4.1 Public policies .......................................................................................................... 27 4.1.1 Moratorium of new permits ............................................................................. 30 4.1.2 Forest and peatland management ..................................................................... 33 4.1.3 Mandatory verification and certification ......................................................... 36 4.1.4 Results-based payment..................................................................................... 37 4.1.5 Law enforcement and integrated fire management .......................................... 38 4.2 Private sector (industry-initiated) policies and measures ........................................ 38 4.3 Civil society organizations’ (CSOs) actions ............................................................ 41 4.4 Linking policy changes to deforestation changes on the ground ............................. 42 5 Linking changes to broader economic and ecological changes ....................................... 43 6 5.1 Commodity prices and economic fluctuations ......................................................... 43 5.2 Forest scarcity .......................................................................................................... 46 5.2.1 The forest transition ......................................................................................... 46 5.2.2 Regression model ............................................................................................. 47 6 Discussion and conclusions ............................................................................................. 52 6.1 What has happened? ................................................................................................ 52 6.2 The hypotheses......................................................................................................... 53 6.2.1 H1: Government policies ................................................................................. 53 6.2.2 H2: Private policies and industry action .......................................................... 54 6.2.3 H3: Civil society pressure ................................................................................ 54 6.2.4 H4: Markets and prices of deforestation-risk commodities ............................. 54 6.2.5 H5: Forest scarcity and the forest transition .................................................... 55 6.3 Concluding remarks ................................................................................................. 55 References ................................................................................................................................ 57 Annex 1: Interview guide ......................................................................................................... 67 13 Since its recent peak, deforestation has fallen by 69% according to GFW data (2016-2023) and by 79% according to TMF data (2015-2022). KHLK and MBI data show an even larger decline of 90% for both data sets (2015-2021 for KHLK, and 2011-2022 for MBI). Deforestation fluctuates by year, and cherry picking the top year exaggerates the decline. A more representative figure is to compare a period before and after the change. Taking the average for the years 2010-2017 and 2020-2022, deforestation has dropped by 64% (GFW), 58% (TMF), 52% (KLHK) and 59% (MBI).4 Taken together, we can conclude that deforestation has been reduced by at least 50% since ca. 2016. Data from the Indonesian Ministry of Environment and Forestry (KLHK) is useful for comparison, while it should be noted that it is referring to net deforestation and forest cover include tree plantations. In all data sets except MBI, deforestation has stabilized since 2020. At the time of finalizing this report, 2023 data are only available from GFW and Nusantara Atlas, both based on Landsat data. These suggest an increase from 2022, with an increase of 21% (GFW) and 34% (Nusantara Atlas) from 2022 to 2023. Reports suggest an increase in oil palm and pulp and paper driven deforestation in 2023, thus 2022 may represent a turning point.5 These data also suggests that deforestation from plantation is also moving from Sumatra to Kalimantan and Papua. While the main purpose of this report is to explain the decline since 2016-2017, we discuss in Chapter 6 whether another turn point happened in 2022-2023. The distinction between deforestation and forest degradation is important. TMF data allow to distinguish between the categories "undisturbed tropical moist forest”, “degraded”, “deforested” and “regrowth”. Figure 3 shows the TMF trends in deforestation and degradation: they tend to move together (positively correlated, with a correlation coefficient of 0.64). Furthermore, for both variables we observe a sharp decline after 2015. 4 Annual deforestation data represents both random fluctuations (due to fluctuations in factors such as weather end commodity prices) and medium-long term trends. The choice of periods is therefore a matter of discussion. We took a conservative approach and looked at the period up to 2017, to make the analysis less sensitive to random annual fluctuations. 5 Nusantara Atlas | 2023 Marks a Surge in Palm Oil Expansion in Indonesia (nusantara-atlas.org) See also: Palm oil deforestation makes comeback in Indonesia after decade-long slump (mongabay.com) 14 Figure 2: Deforestation (ha) in Indonesia, comparing different data sets. Figure 3: Deforestation and forest degradation (ha), 2000-2022. Source: TMF data. - 200,000 400,000 600,000 800,000 1,000,000 1,200,000 1,400,000 1,600,000 Deforestation data sets, Indonesia TMF-defor MBI KLHK Nusantara Atlas GFW-prim forest - 500,000 1,000,000 1,500,000 2,000,000 2,500,000 Deforestation vs. degradation (TMF data) Deforestation Degradation 15 3.2 Deforesta�on by loca�on The change in deforestation is uneven across islands and provinces, which makes it necessary to disaggregate numbers and identify which regions have driven the national level decline. We use mainly two data sets, the TMF and MBI. KLHK data are not publicly available at subnational level. We also supplement with some analysis using the Nusantara Atlas data. 3.2.1 Main islands The changes in deforestation over time in the four main forest islands of Indonesia are depicted in Figure 4. Given that TMF deforestation is detected from disturbances in forest cover and MBI transitions from land use following forest loss, MBI likely does not account for firerelated loss. To account for this difference, we added fire loss from Tyukavina et al. (2022) to MBI forest conversion and find that fire correction makes the trends overall more comparable, with the exception of Sulawesi. According to TMF data, the largest forest loss between 2000 and 2022 has occurred in Sumatra (11.4 Mha), followed by Kalimantan (7.7 Mha) and Sulawesi (1.9 Mha). To understand the noticeable gap between MBI and TMF data for the later island, we compare both figures with estimates from Supriatna et al. (2020) who find 2.1 Mha of forest loss between 2000 and 2017, confirming the TMF estimates. Besides mining and oil palm, they identify corn as a major commodity behind the loss in Sulawesi, frequently cultivated by smallholders who use the improved forest access created by logging activities. A hypothesis for the low deforestation estimates in MBI compared to TMF could then be that smallholder activities are not fully captured in the detection algorithms. The trend analysis for the islands shows that the decline started earlier in Sumatra and Kalimantan compared to Sulawesi and Papua6. Although forest fires played an important role in the peak years around 2015, especially on Kalimantan and Sumatra, they do not explain all the observed increase, indicating that other factors also were important in explaining the surge and following decrease of deforestation. The islands with the largest decline in 2021-2022, compared to the base period 2010-2017, were Sulawesi, with 81.2% reduction in 2021-2022, followed by Kalimantan with 69.7% reduction, whereas Sumatra deforestation levels remained comparatively high with “only” 55.8% reduction. 6 We use the term “Papua” for the Indonesian part of the island of New Guinea, also referred to as Indonesian New Guina or Indonesian Papua. Similarly, “Kalimantan” refers to the Indonesian part of the island of Borneo. 16 Figure 4: Deforestation on main islands for different data sets, 2000-2022. Sources: MBI, TMF, Tyukavina et al., (2022). The two plots shown in Figure 5 picture the composition of deforestation across the major islands, based on the MBI and TMF data. The dominance of Sumatra and Kalimantan is clear, while the changes in their shares are not consistent between the two figures. Keeping in mind the specifics of the data described in Section 2.3, an explanation for the differences possibly lies in the type of deforested forest, since TMF deforestation specifically concerns previously undisturbed (old growth) forest whereas MBI’s forest loss encompasses a much broader forest category. Whereas TMF data depicts that most of the remaining undisturbed forest is lost in Sumatra, MBI data indicates Kalimantan as the major contributor to forest loss according to its specific forest categorisation. Despite the differences, both data sets show an increasing share for Sumatra and declining share for Kalimantan in recent years, suggesting that the deforestation drop in Kalimantan was more pronounced than in Sumatra. 17 Figure 5: Deforestation 2001-2022 by island. Source: TMF and MBI data. 3.2.2 Provinces The decline in deforestation varies greatly across provinces. Table 1 shows the decline in provinces with more than 1 Mha. Among the large deforestation provinces during the period 2010-2017 (> 100 000 ha), Riau and Kalimantan Barat (West Kalimantan) had a decline of ca. 60% in 2021-2022, while Sumatra Selatan (South Sumatra) and Kalimantan Tengah (Central Kalimantan) had an even larger drop of 73% and 80%, respectively. Furthermore, we note that Riau’s two neighbouring provinces, Sumatra Barat (West Sumatra) and Jambi had much smaller reduction of 18-20%. The largest relative reductions are found in Sulawesi Utara (North Sulawesi), Sulawesi Selatan (South Sulawesi) and Maluku. The Indonesian part of the island of Papua, which was recently divided into six provinces but here is aggregated into one, has seen comparably little deforestation over time, except for the peak year in 2015. Roughly 20% of TMF forest loss on Papua during 2015 was caused by fires (see Figure 4). 18 Province Deforestation (ha) Change (%) compared to 2010-17 Change (ha) compared to 2010-17 2010-17 2018-20 2021-22 2018-20 2021-22 2018-20 2021-22 Aceh 39 031 20 053 17 446 -49 % -55 % -18 977 -21 584 Bengkulu 18 246 12 408 8 732 -32 % -52 % -5 839 -9 514 Jambi 70 972 57 111 57 132 -20 % -20 % -13 861 -13 840 Kalimantan Barat 163 724 77 957 69 154 -52 % -58 % -85 767 -94 570 Kalimantan Selatan 33 152 14 318 5 911 -57 % -82 % -18 834 -27 241 Kalimantan Tengah 134 797 50 688 26 810 -62 % -80 % -84 109 -107 987 Kalimantan Timur 94 667 47 846 27 483 -49 % -71 % -46 821 -67 183 Maluku 8 839 2 214 976 -75 % -89 % -6 624 -7 863 Maluku Utara 8 892 2 110 2 457 -76 % -72 % -6 783 -6 435 Papua 56 367 18 887 20 891 -66 % -63 % -37 480 -35 477 Riau 180 311 101 280 72 260 -44 % -60 % -79 031 -108 051 Sulawesi Barat 13 636 7 191 4 769 -47 % -65 % -6 446 -8 868 Sulawesi Selatan 30 408 10 089 3 720 -67 % -88 % -20 319 -26 688 Sulawesi Tengah 40 609 15 433 8 527 -62 % -79 % -25 176 -32 082 Sulawesi Tenggara 20 092 7 622 3 746 -62 % -81 % -12 470 -16 345 Sulawesi Utara 8 879 3 945 1 142 -56 % -87 % -4 934 -7 737 Sumatra Barat 26 692 20 103 21 863 -25 % -18 % -6 589 -4 829 Sumatra Selatan 144 629 64 738 39 453 -55 % -73 % -79 891 -105 175 Sumatra Utara 71 846 42 010 33 088 -42 % -54 % -29 836 -38 757 Table 1: TMF deforestation in hectares for provinces with the highest forest cover (>1Mha). Source: TMF data.7 7 For convenience, and due to changes in provincial border, Papua includes all the provinces on the Indonesian part of the island. 19 The absolute reduction shows a slightly different picture. Three provinces have reduced their deforestation by more than 100 000 ha: Riau, Kalimantan Tengah (Central Kalimantan) and Sumatra Selatan (South Sumatra). Large-scale reductions have also taken place in Kalimantan Barat (West Kalimantan) and Kalimantan Timur (East Kalimantan). The latter province is an interesting case, as it has been part of the World Bank’s Forest Carbon Partnership Facility (FCPF). Under the FCPF, the Government of Indonesia and the World Bank agree to reduce 22 million CO2 equivalent for the period of 2019-2024 (Forest Carbon Partnership Facility, 2022).8 By December 2020, East Kalimantan had achieved emission reduction by 30.8 million tons of CO2 equivalent. Subject to verification, the province successfully met the agreement within 1.5 years of the five-year plan and could opt to sell the additional 8.8 million tons of CO2 equivalent (Forest Carbon Partnership Facility, 2022; Green Climate Fund, 2022). Overall, we see highly varied patterns, with different trends among neighbouring provinces. We note a major decline in several of the Sulawesi provinces. The pattern in Sumatra is varied, with several provinces having small reductions. In Kalimantan the picture is more uniform, with the four major provinces having significant reductions (58-82%). 3.3 Deforesta�on-risk commodi�es (direct drivers) 3.3.1 Land use transi�ons Given that the MBI data set displays the transition between different land cover categories from one year to the next, we use that to understand the commodities that drove deforestation and how they changed over the past two decades. The land cover categories encompass forest formation, oil palm, pulpwood, rice, mining, water, other agricultural land and other nonvegetated land (such as infrastructure). Unfortunately, the disaggregation is at a high scale, and, for example, rubber does not have its own category but is classified as “other agriculture”. The classification of land cover in MBI data allows to reconstruct what area of each category is converted into something else each year and also into what it is converted. Figure 6 shows such flows for the four main islands, averaged over the time periods 2001-2009, 2010-2017 and 2018-2022. (NB: Note the different scales on the y-axis.) The corresponding figures are presented in Table 2. For all islands, with the exception of Sulawesi, we observe a significant decline in the total land use transitions from the 2010-17 to the 2018-2022 period. For Sumatra, this is part of a longer trend (i.e., a decline also from the 2001-2009 period). A second observation is that in Kalimantan and Papua, forest is the main “supplier” of new land in the land use transitions, while the category “Other agriculture” is a major supplier of land use transitions in Sulawesi and even more so Sumatra. 8 The Emission Reductions Payment Agreement is available at the following link: https://www.forestcarbonpartnership.org/system/files/documents/FCPF%20Carbon%20Fund%20ERPA%20Tran che%20A%20&%20B.pdf 20 A third observation, we observe a change in the composition of post-forest land uses, as we summarize in section 3.4. Figure 6: Land use transitions in the major islands. Source: MBI data. 21 Table 2: Post-forest land uses. Source: MBI data. 3.3.2 Forest encroachment factor Forest loss due to the expansion of the cultivation area used for agricultural commodities can be decomposed into two factors: forest loss to crop ≡ total increase in area of crop * forest encroachment factor for crop where the “forest encroachment factor” is defined as the share of new cropland that is converted from forest. While some factors, such as crop prices, determine the overall demand for new cropland, other factors may influence the encroachment factor more directly, such as product certification and moratoriums. On the land supply side, the scarcity of forest will matter both for decisions on whether to intensify (increase yield) or extensify (increase land area), as well as the type of land to expand on (if an extensification strategy is chosen). Table 3 presents the encroachment factors for oil palm and the category “Other agriculture”, for different island and time periods. For oil palm, the encroachment factor varies significantly across islands. It has since 2010 remained above 90% in Papua, while it is lower in Kalimantan and Sulawesi, and the lowest in Sumatra. The encroachment factor also dropped significantly for Sumatra. However, a major drop in the extent of new oil palm in 2018-2020 in Sumatra made the aggregate encroachment factor for Indonesia increase in this period (higher share, while declining again in 2021-2022. Island Category 2001-09 2010-17 2018-20 2021-22 2001-09 2018-20 2021-22 (ha) (ha) (ha) (ha) (pct. of 2010-17) (pct. of 2010-17) (pct. of 2010-17) Kalimantan Mining 969 3 258 4 339 4 844 29.7 133.2 148.7 Oil palm 74 144 156 770 68 077 39 867 47.3 43.4 25.4 Other agr. 11 367 24 662 22 256 12 574 46.1 90.3 51.0 Pulpwood 8 812 22 324 18 914 4 507 39.5 84.7 20.2 Papua Mining 5 1 1 4 973.1 109.6 732.7 Oil palm 1 475 14 506 17 732 4 988 10.2 122.2 34.4 Other agr. 874 2 238 2 701 1 252 39.1 120.7 55.9 Pulpwood 67 317 510 242 21.1 161.2 76.5 Sulawesi Mining 58 787 712 2 713 7.4 90.5 344.9 Oil palm 1 716 5 949 936 4 090 28.9 15.7 68.7 Other agr. 1 900 1 961 3 174 1 674 96.9 161.9 85.4 Sumatra Mining 180 304 292 656 59.1 95.9 215.7 Oil palm 64 270 78 815 6 677 30 204 81.6 8.5 38.3 Other agr. 105 990 86 237 50 574 31 672 122.9 58.7 36.7 Pulpwood 48 288 36 559 3 634 619 132.1 9.9 1.7 Total Mining 1 212 4 350 5 343 8 217 27.9 122.8 188.9 Oil palm 141 605 256 041 93 423 79 149 55.3 36.5 30.9 Other agr. 120 131 115 098 78 705 47 171 104.4 68.4 41.0 Pulpwood 57 166 59 200 23 058 5 368 96.6 39.0 9.1 22 Oil palm 2001-2009 2010-2017 2018-2020 2021-2022 Kalimantan 72% 76% 76% 63% Papua 68% 91% 93% 92% Sulawesi 36% 75% 65% 53% Sumatra 33% 40% 26% 18% Indonesia 46% 60% 69% 9 32% Other agriculture 2001-2009 2010-2017 2018-2020 2021-2022 Kalimantan 40% 60% 51% 55% Papua 56% 71% 72% 69% Sulawesi 8% 12% 29% 28% Sumatra 72% 70% 50% 53% Indonesia 50% 55% 45% 45% Table 3: Forest encroachment factors for oil palm and “other agriculture”. Source: MBI data. We also observe a decline in the encroachment factor for “Other agriculture” after 2017/2018, driven by a decline in both Kalimantan and Sumatra. The provincial differences are smaller than for oil palm. The Nusantara Atlas data offers one additional direct driver of deforestation, namely forest conversion to pulp, and report three post-forest land uses: oil palm (split between smallholder and industrial), industrial pulp, and other land uses. The encroachment factors for oil palm for the three provinces with the highest deforestation in 2010-2017 in Sumatra (Riau, Jambi and South Sumatra) and Kalimantan (West, East and Central Kalimantan), as well as the overall for Indonesia, are reported in Table 4. The absolute numbers are not comparable as the Nusantara Atlas data presented here only include the three largest “deforestation provinces” on each island, in addition to differences in definitions and raw data and processing. Yet the generally much lower encroachment factors for palm oil is notable for the Nusantara Atlas data.10 However, the main story emerging is similar for the two data sets. Compared to the previous period (2010-2017), total land expansion for oil palm declined substantially in 2018-2022, by at least 60 % depending on island and data set. Hence for Indonesia as a whole, the main cause of reduced expansion of oil palm into forest was the overall decline in the expansion of oil palm area. In Sumatra, is however, an exception: the reduced forest loss due to oil palm was due to a decline in the forest encroachment factor. 9 The increase is due to a shift in the composition of oil palm driven deforestation between the islands, with Sumatra (low encroachment factor) reducing its share. 10 We have not investigated further the reasons for the difference, beyond noting that different definitions, raw data and procedures are used in the two data sets. 29 • Moratorium map has been revised several times. coordination with other agencies with different interest, and there are some exceptions, such as National Strategic Programmes. the huge challenges to align maps of different ministries and national agencies. It also triggers deforestation outside moratorium area. It also has exemptions for existing permits and national strategic interests. Moratorium oil palm permits (2018) • Simulation shows potential impact to reduce deforestation. • No known study on the actual impact. • It provides a spatial restriction but only applicable for new permits. • It requires coordination between authorities of state forest areas and areas for other uses, and between national and subnational governments. • Resolution of the oil palm encroachment into state forest is captured by the Job Creation Law, focusing more with administrative procedures. • Low - moderate impact (low evidence) . While it aims to take a similar pathway as the primary forest and peatland moratorium, coordination cost is high as it involves subnational government and palm oil companies, which generally are not willing to share their spatial data. Subnational governments have shown their willingness to revoke permits. Forest management unit (KPH) • KPHs has the potential to work as an intermediary among global, national, and local interests in reducing deforestation. • KPHs face capacity challenges and reduced roles after the Job Creation Law. • KPHs face challenge to deal with local conflicts • Forest management unit provides on-theground support of policy implementation. • They face challenges in budgeting, staffing, and entrepreneurship. • Low impact (medium evidence) . While KPHs as organization have been formed, they lack capacity to implement their functions. Peatland protection and restoration • Serious attention from the Government of Indonesia esp. after 2015 fire event. • Creation of Peatland Restoration Agency and regulatory framework on peatland protection. • Canal blocking as a cost-effective peatland restoration • More preventive measures needed. • Protecting and restoring peatlands reduce emissions significantly. • Government of Indonesia issued Regulations on peatland protection, established Peatland Restoration Agency, and prioritize several provinces for peatland restoration activities. • Medium impact (medium-high evidence). Peatland restoration could reduce fires, which eventually reduce fire-driven deforestation. 30 Social forestry • Potential reduction in village forest and production schemes • But forest loss at the starting point was high • Show increased forest cover in Jambi and West Sumatra • Mainly through the sense of ownership and CSO facilitation • Low impact (medium evidence). Some scattered success stories, but lack of scale. Commodity certification • Perception that deforestation has reduced with the introduction of SVLK. • Difficult to assess the impact of ISPO. • SVLK requires actors to have good business process, including purchase and sales records. • The records include the origin of timber. • Low impact (low evidence). SVLK is perceived to have contributed to reduce deforestation. ISPO does not prevent deforestation. Rigorous studies are required to assess actual impacts of commodity certification on deforestation. Results-based payment • Under the FCPF, East Kalimantan has reduced emission earlier than projected. • National level RBP (Norway, GCF) hard to assess • It requires availability of data and presence of a functioning government-led multistakeholder process • RBP been “an icing on the cake”. • Medium impact (lowmedium evidence). FCPF program in East Kalimantan has proved to be an achievement. The challenge is to scale out to other provinces. National level RBP one among several motivations for policy reforms. Law enforcement • Initially focus on encroachment, illegal logging, illegal wildlife trade, and fires • Use multidoor approach • Low deterrent effects and no losses have been restored • Strengthened law enforcement, and the use of multidoor approach. • An increased number of civil courts for land and forest fire cases. • Fines do not materialize. • Low impact (medium evidence). While enforcement efforts have been acknowledged, they give low deterrent effects and do not return the loss. Integrated fire management • See peatland restoration • Commitment to look for permanent solutions to fires. • Increased coordination among national government agencies • Development of monitoring systems (i.e., Sipongi) • Medium impact (high evidence). Fire prevention and suppression might have contributed to the reduced firedriven deforestation. Table 6: Overview of major public policies aimed at forest conservation. 4.1.1 Moratorium of new permits There are two major policies related to land allocation that are considered to have provided major contributions to the decline of deforestation. The first is the moratorium of permits on primary forests and peatland, which started in 2011. The then-President Susilo Bambang 31 Yudhoyono signed the Presidential Instruction 10 of 2011 on 23. May 2011. The President instructed to take the necessary steps to support the postponement of issuing new licences for primary natural forest and peatland in conservation forests, protected forest, production forest (restricted production forest, regular permanent production forest, convertible production forest) and areas for other uses (Area Penggunaan Lain, APL). The GoI created the indicative moratorium map (known by its Indonesian abbreviation PIPPIB), which is updated every six months. President Joko Widodo made the moratorium permanent by signing the Presidential Instruction 5 of 2019 on 7. August 2019, aiming to protect 66.2 million ha of primary forests and peatlands (MoEF, 2022). The area was about 3 million ha lower than the initial PIPPIB figure of 69.1 million. One interviewee stressed that the reduction does not reflect the loss of forest, but rather the adjustments of permits data across ministries and levels of governments that took place between 2011 and 2019 when the moratorium was made permanent. The adjustments were conducted to find and verify existing permits and exclude them from the moratorium map. According to some resource persons, the moratorium works in tandem with the One map policy that was initially developed during Mr. Yudhoyono’s administration. Consistent with the One map policy’s objectives, the moratorium’s initial activity was to conduct inventory of primary forests and peatlands both inside and outside state forests. Since these areas are not only under the management of the Ministry of (Environment and) Forestry, it took several years before the synchronization process could be completed (interview 32). Studies on the impact of the logging and peatland moratorium show that the policy may have reduced deforestation (Busch et al., 2015; Chen et al., 2019; Groom et al., 2022). Busch et al., (2015) simulate using 2000-2010 data, and concluded that should the moratorium have started in 2000, it would have been able to reduce the deforestation by 153–399 000 ha. Considering leakage (geographic displacement of deforestation to locations not covered by the moratorium), the reduction would have been 116–305 000 ha. Another study, using 2001-2017 data, indicated that there was a significant reduction of forest loss on the areas under moratorium, and concessions under moratorium were shown to have significantly lower forest loss than those outside the moratorium areas (Chen et al., 2019). A third study finds a small difference of dryland forest in forest loss between areas under and outside moratorium areas, although no difference in wetlands (Groom et al., 2022). Others claim that fire occurrences have been lower on the areas under moratorium (Lee, 2023, interview 28, 32). In contrast, others claim that deforestation tends to increase in moratorium areas, where the annual deforestation was 97 000 ha in 2005-2011 and increased to 137 000 ha in 2012-2018 (Greenpeace Indonesia, 2019). Others have pointed to the risk of deforestation outside the areas under the moratorium (Leijten et al., 2021; Suwarno et al., 2018). There are also general concerns that some aspects make the moratorium may be less effective. The PIPPIB have been adjusted several times to accommodate the exceptions made explicit in the presidential instructions, which lies lower at the hierarchy of Indonesia’s legislation (Greenpeace Indonesia, 2019; interview 15, 23). As such, there is no penalties if the instruction is neglected. Moreover, 32 since some of the primary forests and peatlands are located outside the state forest, the PIPPIB map must be developed and synchronized among the Ministry of Environment and Forestry, National Land Agency, and other national agencies. Finally, the moratorium contains some exceptions, such as existing permits and areas designated as national strategic programmes (Estherina & Sedayu, 2024). The second main policy in this category is the moratorium of new oil palm plantation permits. On 19. September 2018, President Widodo signed the Presidential Instruction 8 of 2018 on the moratorium and evaluation of licensing of oil palm and improvement of oil palm plantation productivity. The moratorium aimed to improve the governance of the sustainable palm oil production in Indonesia and to resolve the problem of oil palm areas inside state forests and associated conflicts. One major area was to clarify which agency should be designated as the data manager for the oil palm, and the Directorate General of Estate Crops of the Ministry of Agriculture was assigned that role. More coordination was required since the data on land allocation, land use permit, and plantation business permit are managed by different levels of government agencies (interview 32). The moratorium was effective from 2018 until 2021. The palm oil moratorium was part of a larger effort to solve challenges on the encroachment of oil palm plantations into the state forest areas, which the GoI responded by issuing the Presidential Regulation 88 of 2017 on the resolution of land control inside the state forest areas. Under the Job Creation law 6 of 202314, the 2017 Regulation was replaced by the Presidential Regulation 62 of 2023 on the acceleration of the implementation of agrarian reform. The 2023 Law contains two articles that provide the legal basis for palm oil plantations located in the state forest zone, at least temporarily (Muttaqien, 2023; Nababan, 2023). The article says that every person who carries out business activities that have been established and have permits in a forest area before the enactment of this Law and who has not fulfilled the requirements in accordance with the provisions of statutory regulations in the forestry sector, is obliged to complete the requirements no later than three years since this Law came into force (article 110a). NGOs are particularly concerned about the article because allowing palm oil companies in the state forest areas could backfire the progress made to resolve the oil palm encroachment into the state forest areas (Anonymous, 2023; Nurhadi Sucahyo, 2023; Siti Sadida Hafsyah, 2020; interview 3). Several NGOs raised concerns about the decision to not extend the moratorium because they have seen some early outcomes of the moratorium, although several areas need improvements, such as law enforcement (Madani Berkelanjutan, 2021). In addition, at least in East Kalimantan, the Presidential Instruction is used as a guidance for provincial government to sort out issues related to palm oil land and permit and to develop sustainable palm oil regulations at the provincial level (interview 13, 21). While NGOs were largely concerned about the palm 14 Job Creation Law was issued as Law 11 of 2020 on November 2, 2020. The Constitutional Court decided that the 2020 Law was conditionally unconstitutional. On November 3, 2021m, the Constitutional Court, through the Decision No.91/PUU-XVIII/2020 gave the Government of Indonesia and the Parliament two years to revise the Law. On December 30, 2022, the Government of Indonesia issued the Government Regulation in lieu of Law (Perpu) 2 of 2022, which was lifted as Law 6 of 2023. 33 oil permit moratorium ending in 2021, it was argued that it ended because the follow up or resolution of the oil palm permit issue was integrated into the Job Creation Law 6 of 2023 (interview 28). It was reported that more than 1 000 companies have plantations located in the state forest area (CNN Indonesia, 2023). While studies on the actual impacts of the Law on deforestation are not available (partly due to the Law being contested and re-issued in 2023), analyses point to several implications of the Law on forest governance and deforestation. For example, there are concerns that the Law has relaxed the need for ex-ante environmental impact assessment, centralized the approval and monitoring processes, and weakening strict liability stipulations (e.g., Sembiring et al., 2020). In addition, the Law has revoked the need to maintain a minimum of 30% of watersheds or island (Sunarto et al., 2021). Studies evaluating the impacts of the palm oil moratorium on reducing deforestation are not available. One study, however, used a simulation to forecast that the large-scale oil palm moratorium could reduce deforestation by 28% and cut greenhouse gas emissions by 16% compared to a no-policy scenario for the period of 2010-2030 (Mosnier et al., 2017). Such simulation studies are useful to demonstrate the potential but may also fail to incorporate several implementation hurdles. 4.1.2 Forest and peatland management Forest management units Some interviewees argue that the establishment of Forest Management Units (Kesatuan Pengelolaan Hutan, KPH) is a policy that contributed significantly to reduced deforestation. While the idea of KPHs dates back to the pre-2000 era, it gained momentum after the issuance of the Government Regulation 6 of 2007 (interviews 3, 17). According to Indonesian forestry regulations, KPH was seen as a prerequisite to achieve sustainable forest management (Sahide et al., 2016). KPH was even seen as a response to tackle deforestation and forest degradation (Ramadhan et al., 2023). KPHs were established in 2009 after the issuance of the Minister of Forestry Regulation 6 of 2009 on the establishment of Forest Management Units, and Minister of Forestry Regulation 51 of 2010 on the determination of KPH areas throughout Indonesia and the establishment of model KPHs. KPH has become one of the government programmes with large donor support, including the German government (Marianne Scholte, 2019; interview 17). KPH aims to achieve sustainable forest management, and in that way also reduce the influence of palm oil as a driver of deforestation (Sahide et al., 2016). KPH has the potential to play an important role as intermediaries to achieve REDD+ targets, due to its capacity to connect national and international actors and local realities (Bae et al., 2014; Kim et al., 2016). Despite the potential and strong donor support, KPHs would still need time to achieve its intended outcomes. First, KPHs are not equipped with sufficient human resources and budget for the operations, including the engagement with communities living in and around the management units (Bae et al., 2014; Kim et al., 2016; Scholte, 2019). Second, in the beginning KPH was designed as decentralized field-level forest management units, where they could operate as executor of forest management programs, including making contract with third parties on certain business aspects. However, under the 2023 Job Creation Law, the role of 34 KPH was reduced to focus only on facilitation without having executive decision power (Nugroho et al., 2023; Ramadhan et al., 2023). One implication is the reduced budget for KPHs, as they are more dependent on working with other government agencies such as the Watershed Management Agency (Balai Pengelolaan Daerah Aliran Sungai, BPDAS) and other agencies that have activities and programs located on areas under the facilitation of the KPHs. Third, being located at the site level, KPHs often have to allocate resources and time to manage and solve the issues related to land tenure and conflicts. Clarification on land boundary and tenure, including ownership, is frequently an issue in KPH areas (Budiningsih et al., 2022; Ichsan et al., 2021). Peatland rehabilitation and restoration Peatland has been well-recognized as a major pool of carbon. When converted to agriculture, it releases large amounts of carbon (Abrams et al., 2016; Ansari, 2011; Dohong et al., 2018). A major policy agenda of the GoI, after the huge fires in 2015, was to protect the country’s peatlands. The agenda consists of establishing an organization that works specifically on peatland restoration, creating a regulatory framework on peatland protection, and undertake national and international collaborations on peatland restoration (Nugroho et al., 2023). The GoI established the Peatland Restoration Agency (Badan Restorasi Gambut, BRG) in 2016. The mandate of the Agency was extended to include mangrove in 2021, hence the name becomes Peatland and Mangrove Restoration Agency (Badan Restorasi Gambut dan Mangrove, BRGM). The GoI adopts the 3R (rewetting, revegetation, and revitalization) approach to peatland restoration. Rewetting focuses on rewetting peatlands that have been drained by building canal blocking. Revegetation focuses on efforts to increase peatland vegetation cover. Revitalization refers to strengthen the capacity of communities or grass-root institutions, mainly at the village level (Badan Restorasi Gambut dan Mangrove, 2023). Some critics, however, mentioned that the 3R approach lack preventive measures such as Reduction of fire (hence 4R) (Harrison et al., 2020) and Reporting and monitoring (5R) (Terzano et al., 2022). In the same year, the GoI also issued Government Regulation 57 of 2016 on peatland ecosystems protection and management. Under this framework, the GoI use peatland hydrological unit (Kesatuan Hidrologis Gambut, KHG) that defines peatland management areas into cultivation and protection purposes. KHG is expected to hold the expansion of plantations, mainly oil palm (interview 23). The GoI has established 865 KHGs covering 24.6 million ha. BRGM focuses its activities in priority provinces, which consist of 522 KHGs covering 12.9 million ha (Badan Restorasi Gambut dan Mangrove, 2023). The Ministry of Environment and Forestry also has strengthened its peatland-related units, by, for example, building several information systems that monitor fires (called Sipongi) and water table of peatland areas (Simatag) (interview 28). In terms of restoration activities, BRGM collaborated with various stakeholders. They built 7,697 canal blockings in 746 villages of the seven priority provinces. BRGM also developed 2,147 hectares of revegetation plots. For the revitalization of livelihoods, BRGM has 35 established 1,246 packages of land-based, water-based, and ecosystem services-based activities. During 2020-2021, the number of hotspots in BRGM work area drops from 97 to 71, which is associated with the drop of burned area from 3,651 ha to 313 ha (Badan Restorasi Gambut dan Mangrove, 2023). In total, BRGM is mandated to restore 1.2 million hectares of degraded peatlands. By 2023, BRGM reported to have restored nearly 830,000 hectares through rewetting and revegetation (Badan Restorasi Gambut dan Mangrove, 2024). Similarly, reviews on peatland restoration in Indonesia shows the positive effect of canal blocking as a rewetting activity to reduce fires on peatland, and subsequently reduce peatland cover loss, making it a cost-effective measure to combat peatland fires (Kiely et al., 2021; Yuwati et al., 2021). Social forestry While social forestry has been one of the priority agendas of the Ministry of (Environment and) Forestry at least since Mr. Yudhoyono’s administration, efforts to raise the profile of social forestry has been strengthened during the Widodo presidency. The administration sets out a target to allocate 12.7 million hectares of state forests for social forestry, equivalent to about 10% of the total areas of state forests. By 2022, the Ministry of Environment and Forestry has issued or established social forestry agreement of 5.3 million ha (Kementerian Lingkungan Hidup dan Kehutanan, 2023). One highlight is the clarification of schemes that fall as social forestry. The ministry issued a regulation that defines the different schemes of social forestry, that is indigenous or customary forest (hutan adat), community forestry (hutan kemasyarakatan), smallholder timber plantation (hutan tanaman rakyat), village forest (hutan desa) and forestry and environmental partnership (kemitraan lingkungan). Social forestry has a large potential for achieving reduced deforestation, insofar it was argued that forest rehabilitation through social forestry would be able to meet Indonesia’s target for increasing carbon sequestration (interview 28). Village forest has the potential to reduce deforestation (Gunawan et al., 2022; Santika et al., 2017, 2019; Wahyu et al., 2020). Village forests in Sumatra and Kalimantan have avoided deforestation, although in the longer term the performance depends on the type of forest function. Village forest reduces deforestation on areas allocated for watershed protection and limited timber extraction (Meijaard, 2017; Santika et al., 2017). However, an analysis of three schemes - village forests, community forestry, and community timber plantation –that cover 2.4 million ha found that schemes that aim at conservation tend to increase forest loss, and these aimed at production tend to reduce deforestation, although the reductions came from an initially higher forest loss (Kraus et al., 2021). The performance of social forestry also depends on strong support from CSOs. In Jambi and West Sumatra, one analysis found that forest cover increased after the introduction of a social forestry scheme. While the forest cover under social forestry areas large recover through natural succession, it was argued that social forestry increases sense of ownership and responsibility of the agreement holders. They conduct patrol and adopt agroforestry systems in the areas allocated for cultivation (Yola Sastra, 2024; interview 7). While the cases in Jambi and West 36 Sumatra give some hope for the potential of the social forestry in reducing net deforestation, social forestry permits have been facing significant challenges in terms of flawed land administration processes, conflicting interests among local actors, and lack of institutional engagement beyond permitting process (Fisher et al., 2018). In addition, more small patches of forests under social forestry could lead to forest fragmentation (Gunawan et al., 2022). It is therefore too early to conclude on the nationwide impacts of these initiatives. 4.1.3 Mandatory verifica�on and cer�fica�on There are two main products that are covered by mandatory verification or certification: timber and palm oil. Timber is regulated by the Timber Legality and Sustainability Verification (SVLK). For timber concessions, SVLK is supposed to be a stepping stone to achieve Sustainable Production Forest Management (PHPL) (Kosar et al., 2019). The SVLK is compatible with the EU Forest Law Enforcement Governance and Trade (FLEGT), and it became the first scheme in the world to issue FLEGT licenses (2016). According to the Ministry of Environment and Forestry’s Information System on Legality and Sustainability (known as SILK), there are 3,983 SVLK-certified management units, both upstream and downstream, by April 2024. The second commodity is palm oil. Indonesia is the first country that created mandatory sustainability certification for palm oil. The Indonesian Sustainable Palm Oil (ISPO) was created in 2011 and has since been strengthened both in terms of substance and legal backing. In terms of substance, ISPO was revised in 2015 and 2020. ISPO has also been legally strengthened from a ministerial regulation to a presidential regulation. By 2023, ISPO covered 4.2 million ha, of which 3.9 million ha were large-scale plantations. There are 707 certified companies and 79 smallholders (Antara, 2024). These figures translate to approximately 25% of the total oil palm area. 31% of the total oil palm plantation companies being ISPO-certified, while the share of smallholders is miniscule (79 among literally millions of small palm oil producers). The deforestation reducing effect of both schemes remain understudied. One study shows that actors in Indonesia perceived SVLK to contribute to reduced deforestation, mainly through better implementation of forest management plan and the reduction of illegal logging (Goetghebuer et al., 2023; Tropical Forest Alliance & LPEM Universitas Indonesia, 2022). Another possible impact channel would be that having an SVLK certificate would mean the raw materials come from clear and recorded sources which follow the regulation of the Ministry of Environment and Forestry (Antara, 2020). In the case of ISPO, any impacts must also be seen in connection with private sustainability commitment or pledges. There are concerns, however, that certified companies do not necessarily carry out activities that conserve forests and comply to the national regulation. An example is a Greenpeace report which shows that over 200 ISPO-certified companies allegedly encroach state forest and therefore part of the problems being addressed by the moratorium of palm oil permits (Arumingtyas, 2021; Greenpeace Indonesia, 2021). 37 4.1.4 Results-based payment One of the mitigation strategies of the Government of Indonesia (GoI) is to establish multibilateralism, that is, several bilateral relations under multilateral settings (Leony Aurora, 2012). One mechanism of bilateral agreement is results-based payments (RBP) as part of multilateral frameworks (Dwisatrio, 2021). RBP can be made for both the national and subnational levels. A major example of a national level RBP is the Letter of Intent between Indonesia and Norway in 2010, whereby Norway committed to pay USD 1 billion to Indonesia for verifiable emission reductions15. After discontent and disagreements on several issues (including delayed payments), Indonesia terminated the agreement in September 2021 (Jong, 2021b). A year later, in September 2022, however, Indonesia and Norway agreed to a new and modified agreement, where Norway would pay Indonesia to keep the forest standing, while still honouring the outstanding payment from the previous agreement (Jong, 2022). The first RBP payment of USD 56 million was made in 2022, for emissions reduction achieved in 2016-2017. In total, payments of USD 156 million have been made in 2022-2024 for the three forest years 20162019. The Green Climate Fund (GCF) also approved in 2020 a project worth USD 103.8 million, based on emissions reductions achieved in the period 2014-2016. The money was disbursed in 2021, with UNDP as the accredited entity, and are to be used “to invest in activities that support the implementation of the country’s national REDD+ action strategy (STRANAS). ... This includes working with key agencies at the national, provincial, and local levels to strengthen the development, coordination, and implementation of Indonesia’s overall REDD+ architecture, as well as providing support to decentralised sustainable forest governance. This includes establishing forest management units and expanding implementation of the country’s social forestry programme.”16 RBP is also operational at the province level, where the central government select the province based on several criteria, including the political will of provincial government and a functioning multistakeholder process in place to support provincial government (interview 13, 23, 28). The Forest Carbon Partnership Facility (FCPF) is an example of a subnational RBP, in the form of an agreement between GoI and multiple donors coordinated by the World Bank. A FCPF grant agreement was initially signed in May 2011 and revised in November 2016. FCPF aims to contribute to the development of Indonesia’s capacity to design a sound national REDD+ strategy, develop national and subnational reference scenarios and establish a forest monitoring and carbon accounting system, consistent with local, regional and national conditions and circumstances (Ministry of Environment and Forestry, 2023). In East Kalimantan, one of the activities under the FCPF was the protection of forested areas outsize state forest zone. The size of the area is ca. 850 000 ha, with an aim to protect ca. 640 000 ha. Other activities included climate village program (Wiati et al., 2022). The achievement was noteworthy, as East Kalimantan managed to reduce emission beyond the target and earlier 15 https://www.regjeringen.no/globalassets/upload/smk/vedlegg/2010/indonesia_avtale.pdf 16 https://www.greenclimate.fund/project/fp130 38 than planned (Forest Carbon Partnership Facility, 2022; Green Climate Fund, 2022; interview 13, 26). However, the achievement is debated; deforestation remains high, and the impact of the carrots is questioned as the period for calculating results started not before 2019-2020 (interview 30). It is important to note that while emission reduction measurement was started in 2019, East Kalimantan issued and conducted several policies and programs that lead to its capacity to reduce emissions, started from the “Green East Kalimantan” (“Kaltim Green”) as early as 2010. East Kalimantan is also a progressive province in terms of issuing a provincial sustainability regulation for the estate crop sector. Assessing the impact of national level RBPs is inherently difficult. One approach is to “follow the money” and investigate the impact of the imitated projects. The GCF support was designated for specific activities, while the Norwegian contribution is not. However, a major aim of RBP is to induce policy reforms that reduce emissions from tropical forest. This boils down to tracing down policy decisions, and the impact of those decisions. As discussed in section 6.2.1, the promise of RBP from Norway and other actors is one among several factors that may explain Indonesian forest policy reforms since ca. 2010, but we are not aware of any detailed scientific studies on RBP’s role relative to other factors. 4.1.5 Law enforcement and integrated fire management One important change in recent years has been the strengthening of law enforcement. As one interviewee puts it, “forests has been brought into the courts” (interview 17). Together with the independent monitoring, law enforcement officers successfully reveal illegal timber and wildlife trade, including timber originating from forest-rich regions such as Papua. Furthermore, enforcement by the government has led to prosecution of actors using fire to clear the land. The trigger was the successful prosecution of a palm oil company, Kalista Alam, who was found to have burned forests in Aceh for oil palm plantations. After this, the GoI has been building court cases against companies allegedly burning forests and peatlands. While the big and publicised cases involve companies, there are also cases when individuals are brought to the legal system for using fire. In terms of fire management, one important policy feature was the inclusion of fire prevention and early suppression as the key performance indicators of high-ranking officers at the subnational level. The President threatens to remove officials from the position of provincial head of military and police if they failed to prevent and manage fires (interview 5, 10, 15, 24, 25). The firefighting was carried out in a coordinated manner, led by the fire brigade of the Ministry of Environment and Forestry (known as Manggala Agni) (interview 28). 4.2 Private sector (industry-ini�ated) policies and measures Private regulation, which refers to industry-wide standards initiated by the industry, have been in place since the early 2000s for timber and palm oil. There has been a steady increase of market uptake and production of certified palm oil (RSPO 2022). The uptake of certified timber, such as through the FSC and PEFC schemes, is also increasing. 45 measured in Rupiah. Thus, the pre-2020 decline in agricultural commodity prices (in USD) is smaller when measured in Rupiah. Figure 10: Palm oil prices and deforestation due to palm oil expansion: Source: indexmundi.com, MBI and Nusantara Atlas (NA). Note: first y-axis is price, while the second y-axis is ha (deforestation and total expansion of oil palm into forests and other land). No other country loses more forest due to industrial mining than Indonesia. During the first two decades of this century, 58.2% of the forest lost globally to industrial mining occurred in Indonesia, amounting to 190 098 ha (Giljum et al., 2022). Coal mining in Kalimantan was the main source behind this figure. In addition to coal, key minerals include nickel, copper and gold in the mining industry of Indonesia. Figure 11 shows the price trends for these commodities. The most striking development is the spike in coal prices, a result of the energy shortage after the Russian invasion of Ukraine in February 2022. By early 2024, coal prices are, however, back to a more normal level. - 100,000 200,000 300,000 400,000 500,000 600,000 - 200 400 600 800 1,000 1,200 1,400 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 Palm oil price and oil palm deforestation and expansion Price (global) Deforestation (MBI) Deforestation (NA) Oil palm expansion (NA) 46 Figure 11: Price index of key minerals, coal and timber (average price in 2014 = 100). Data source: indexmundi.com The price of the three minerals included have shown an increasing trend since 2016. Since its low in February 2016, the price of nickel quadrupled by March 2022, but has since dropped again by more than 50%. Gold prices have shown a steadier increase and are by early 2024 about 50% higher than the 2014 level. The same net increase is true for copper, although the price fluctuations have been higher. Timber (hard logs) prices are also pictured in the figure, and we observe no major change until 2020, after which the price has dropped by ca. 30% (January 2021 to end 2023). Overall, minerals and coal have witnessed a steady increase in its price for the period of study, unlike agricultural commodities which saw a decline until 2020 and an increase afterwards. This is consistent with the finding of section 3: mining has in recent years assumed a bigger role as a direct driver of deforestation. 5.2 Forest scarcity 5.2.1 The forest transi�on The Forest Transition (FT) theory predicts that forest cover undergoes a stylized trajectory of change, from an initial stable state through a period of accelerating rate of decline, to decelerating decline, and stabilization and partial recovery (Figure 12) (Wolfersberger et al., 2015). The theory is often used to explain and predict deforestation trends but is also criticized for its deterministic prescriptions and weak empirical foundations (Perz, 2007). Note that there are two important turning points in the FT trajectory: when the deforestation rate starts to - 25 50 75 100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 500 525 550 575 600 625 650 Jan-15 May-15 Sep-15 Jan-16 May-16 Sep-16 Jan-17 May-17 Sep-17 Jan-18 May-18 Sep-18 Jan-19 May-19 Sep-19 Jan-20 May-20 Sep-20 Jan-21 May-21 Sep-21 Jan-22 May-22 Sep-22 Jan-23 May-23 Sep-23 Jan-24 Price index of mined commodities and timber (2014 = 100) Hard Logs Nickel Copper Gold Coal 47 decline (shifts from accelerating to decelerating deforestation), and when deforestation ends and reforestation starts (deforestation shifts from positive to negative).22 Figure 12: Stylised scheme of the forest transition. The FT can be linked to different but interrelated factors. One of the main explanations put forward in the literature is the forest scarcity hypothesis: as forest cover declines, forest loss inevitably decreases simply by the fact that less forest is left to be cleared. Increasing scarcity of forest reduces the supply of forest products and hence leads to higher value of the forest that is left, making conversion less profitable and even partly reverses it (Angelsen & Rudel, 2013; Barbier et al., 2010; Rudel et al., 2005). The second main explanation in the FT literature is linked to stages of economic development. With increasing demand for food and other agricultural commodities, forest is first transformed into agricultural area, before urbanisation and decreasing agricultural employment halt agricultural expansion (Rudel et al., 2005). 5.2.2 Regression model If the FT theory plays a role in explaining the decline in forest loss in Indonesia, we would thus expect that it has mostly dropped in regions where: (i) forest cover has approached or surpassed the first turning point, or where (ii) economic development has led to lower agricultural dependency. We use a panel regression model of the following form: 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑖𝑖,𝑡𝑡= 𝛽𝛽1𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑖𝑖,𝑡𝑡+ 𝛽𝛽2𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓2 𝑖𝑖,𝑡𝑡+𝛽𝛽3𝑓𝑓𝑓𝑓𝑓𝑓𝑐𝑐𝑓𝑓𝑐𝑐𝑐𝑐𝑐𝑐𝑓𝑓𝑓𝑓𝑖𝑖,𝑡𝑡+ 𝛾𝛾𝑦𝑦𝑓𝑓𝑐𝑐𝑓𝑓𝑡𝑡 22 In mathematical terms, if we consider forest cover to be a function of time, the first turning point is when the second derivative changes sign, while the second turning point is when the first derivative changes sign. Forest cover Time Turning point 2 (defor  refor) Turning point 1 (increasing  decreasing rate of defor) Forest/plantations/ agric. mosaics Undisturbed forests Forest/agric. mosaics Forest frontiers 48 The variables 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑖𝑖,𝑡𝑡 and 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑖𝑖,𝑡𝑡 are forest loss and forest cover calculated as the share of land in province i and year t, and 𝛾𝛾 are year fixed effects to control for time trends that are common across provinces. We include squared terms to test for the hypothesised bell-shaped relationship between deforestation and forest cover. As an indicator of economic development, we used expenditures per capita (also including a squared term to allow for non-linearity), the agricultural employment share, and population density. All of these are provided by the Indonesian Database for Policy and Economic Research, accessed through the World Bank data portal.23 TMF MBI (1) (2) (3) (4) (5) (6) Forest cover 0.0346*** 0.0351*** 0.0307** 0.0099* 0.0104* 0.0079+ (0.008) (0.0061) (0.0099) (0.0040) (0.0039) (0.0046) Forest cover squared -0.043*** -0.0430*** -0.044*** -0.0106* -0.0111* -0.0117* (0.009) (0.0074) (0.010) (0.0046) (0.0043) (0.0049) Expenditure per capita 2.4e-08 4.0e-08* 1.8e-08 3.0e-08+ (1.6e-08) (1.8e-08) (1.3e-08) (1.6e-08) Expenditure per capita squared -9.7e-15 -1.6e-14+ -6.3e-15 -1.1e-14 (6.5e-15) (7.7e-15) (5.7e-15) (7.4e-15) Agric. employment share 0.0081 0.0061 (0.0048) (0.0040) Population density -0.00035 -0.00020 (0.00022) (0.00014) FT turning point (% forest cover) 0.405 0.407 0.349 0.466 0.465 0.335 Observations 660 602 500 616 544 488 R2 0.381 0.424 0.490 0.137 0.224 0.352 Std. Errors by: province by: province by: province by: province by: province by: province Fixed Effects: year X X X X X X + p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001 Table 8: Panel regression model results for deforestation as a share of land. Regression results with both TMF and MBI data are shown in Table 8. The results indicate the existence of a forest scarcity effect, where forest loss first increases with forest cover, but decreases once a turning point has been reached. Using only forest cover and squared forest cover as covariates, the model can explain 38% and 14% of the variation in deforestation for TMF and MBI data, respectively (columns (1) and (4)). With all variables included, our model explains 49% of the variation in forest loss for TMF data and 35% for MBI data. Depending on the model specification, the turning point was located between 35% and 43% forest cover according to TMF data and between 34% and 47% for MBI data. The forest scarcity effect remains significant also when socio-economic variables were added, such as expenditure per capita in columns (2) and (5), or agricultural employment share and population density in columns (3) and (6). It therefore appears that the forest scarcity effect dominated over the economic development effect in Indonesia. 23 https://databank.worldbank.org/source/indonesia-database-for-policy-and-economic-research/preview/on 49 Figure 13: Forest transition curve for model (1). Province trajectories are plotted for average deforestation and average forest cover (TMF data). We further investigated the trajectory of each province along the forest transition curve for the forest scarcity turning point of 41% as indicated in model (1) in Table 8, i.e., the first turning point of Figure 12. We find that many provinces’ forest cover indeed declined towards the turning point threshold and beyond (Figure 13, Table 8). Comparing average deforestation for the years 2001-2005 and the post-peak years 2017-2022, the only provinces with increasing rates of forest loss over the considered time interval were initially located on the right side of the turning point. Among the provinces with the highest forest cover decline are Jambi and Riau, as shown by the distance between the blue and the yellow points along the x-axis (cf. section 3.2.2). Whereas Jambi has moved beyond the forest transition turning point and shown an increase in forest loss between the two time periods, Riau was located close to the turning point in the 2000-2005 period with extremely high deforestation but has seen a substantial reduction in forest loss since then. A general observation is that provinces to the left of the 50 curve appear to have experienced a stronger reduction of deforestation rates and a lower loss in forest cover between the time periods considered. Figure 13 gives a better indication of the location of different provinces along the different forest scarcity turning points estimated in Table 8. Considering the implied turning point of model (1) of 41%, eight provinces (Jambi, Riau, Bengkulu, Sumatra Utara, Sulawesi Selatan, Bali, Bangka Belitung and Sumatra) have moved from the right to the left side of the turning point between 2001 and 2022. Another ten provinces were already below the forest scarcity threshold in 2002. Riau is an extreme province, in the sense that the deforestation rate in 2001-2005 was much higher than the estimated FT curve suggests. The 2017-2022 deforestation is still above that curve, but the huge decline observed between the periods can be seen as Riau becoming a more “normal” province. Table 9 shows the forest cover and forest cover loss across the archipelago, sorted by forest cover in 2002 (forest share of total land). The FT theory predicts increasing rates of deforestation for high forest provinces, while medium-low forest cover provinces should experience a decline in the rate of deforestation. More generally, the higher the forest cover, the higher the expected increase in the deforestation rate. In the first category of very high forest cover provinces, the deforestation rate increased by 1.6 percentage points (pp). In line with FT predictions, the rate of deforestation in the medium forest cover group, was reduced by 2.8 pp, but with large variations within the group. Overall, the analysis of the forest transition suggests that a forest scarcity effect has been a significant factor in the reduction of forest loss in Indonesia over the past decade. Assuming there is a forest cover turning point from which deforestation rates decline as forests get scarce, we showed that several provinces have surpassed this critical point in recent years and seen reduction in forest loss thereafter. However, as a very stylised theory, it is prudent to be cautious about the predictive power of the forest transition and to interpret the results in the context of the actual ecological, socioeconomic and political situation in each province. 51 Table 9: Forest cover (reduction) by province for the years 2002, 2012 and 2022. Kalimantan Utara and Kalimantan Timur are not included in the table, since Kalimantan Utara was only created in 2012 when it split off Kalimantan Timur to form a new province. Source: TMF data. Category/Province TMF cover 2002 TMF cover 2012 TMF cover 2022 Change 20022012 (pp) Change 20122022 (pp) Change 20022022 (pp) Change in def rates between 1. and 2. period Very high forest cover Papua Barat 93.0% 91.5% 89.0% -1.5% -2.6% -4.1% -1.1% Papua 84.3% 82.7% 79.7% -1.6% -3.0% -4.6% -1.4% Maluku Utara 82.9% 78.5% 71.8% -4.4% -6.7% -11.1% -2.3% Category average 86.7% 84.3% 80.2% -2.5% -4.1% -6.6% -1.6% High forest cover Sulawesi Tengah 78.9% 71.7% 62.7% -7.2% -9.0% -16.2% -1.8% Sulawesi Barat 74.3% 62.9% 53.5% -11.5% -9.4% -20.8% 2.1% Kalimantan Tengah 73.9% 61.7% 53.8% -12.3% -7.9% -20.2% 4.4% Maluku 71.6% 67.9% 62.8% -3.7% -5.1% -8.8% -1.3% Sumatra Barat 71.3% 61.4% 53.8% -10.0% -7.5% -17.5% 2.4% Kalimantan Barat 69.8% 56.1% 46.9% -13.8% -9.2% -22.9% 4.6% Aceh 69.7% 62.8% 56.7% -6.9% -6.1% -13.0% 0.8% Sulawesi Utara 68.5% 63.9% 52.4% -4.6% -11.5% -16.1% -6.9% Gorontalo 66.8% 63.1% 56.3% -3.7% -6.8% -10.4% -3.1% Kepulauan Riau 66.4% 58.2% 48.3% -8.3% -9.9% -18.1% -1.6% Sulawesi Tenggara 63.4% 56.5% 49.3% -6.9% -7.2% -14.1% -0.2% Jambi 62.6% 44.0% 29.0% -18.6% -15.0% -33.6% 3.7% Category average 69.8% 60.8% 52.1% -9.0% -8.7% -17.6% 0.3% Medium forest cover Riau 54.0% 30.1% 20.2% -23.9% -9.9% -33.8% 14.0% Bengkulu 53.6% 43.7% 36.5% -9.9% -7.2% -17.1% 2.8% Sumatra Utara 49.9% 37.8% 31.2% -12.2% -6.5% -18.7% 5.6% Sulawesi Selatan 47.2% 39.9% 33.2% -7.3% -6.7% -14.0% 0.6% Bangka Belitung 39.4% 26.6% 15.6% -12.8% -11.1% -23.8% 1.7% Sumatra Selatan 39.3% 24.7% 15.0% -14.7% -9.7% -24.3% 5.0% Bali 36.3% 27.8% 16.5% -8.5% -11.3% -19.8% -2.8% Kalimantan Selatan 35.1% 28.2% 22.2% -6.9% -6.0% -12.9% 0.8% Nusa Tenggara Barat 32.4% 29.3% 23.5% -3.1% -5.8% -8.9% -2.7% Category average 43.0% 32.0% 23.8% -11.0% -8.2% -19.3% 2.8% Low forest cover Banten 28.9% 23.4% 17.4% -5.5% -6.1% -11.6% -0.6% Jawa Barat 24.2% 17.2% 10.4% -7.0% -6.7% -13.7% 0.2% Jawa Tengah 17.5% 12.4% 7.1% -5.1% -5.2% -10.3% -0.1% Lampung 16.6% 13.5% 11.2% -3.1% -2.3% -5.4% 0.8% Jawa Timur 16.0% 10.8% 6.2% -5.2% -4.7% -9.8% 0.5% Nusa Tenggara Timur 9.7% 8.0% 6.4% -1.6% -1.7% -3.3% 0.0% Category average 18.8% 14.2% 9.8% -4.6% -4.4% -9.0% 0.1% 52 6 Discussion and conclusions The starting question was: “Why has deforestation in Indonesia declined over the past few years?”. There is no straightforward method to answer such a complex question, and the approach has been to build an evidence base and develop a coherent story by decomposing deforestation figures (year, location and direct driver), through interviews, and by statistical analysis. While such a triangulation is the most viable approach, there are several challenges. First, the data do not tell one uniform story. As seen in section 3, the numbers on the level of deforestation and its direct drivers differ considerably, due to different definitions, methods and primary data sources. Second, there is no single, best approach to trace impacts. This was exemplified by the interviewees on the impacts of the moratorium. Some stressed how the moratorium has – or has not – changed economic incentives though new “sticks and carrots” (including legal enforcement and punishments). Others stressed the process, and how the moratorium can change values and attitudes, and also has stimulated a process for better coordination across different sectors and level of government. 6.1 What has happened? Deforestation in Indonesia has dropped by at least 50% since 2016. Although forest loss to fires has been extraordinarily high in the years before the drop, this alone is not sufficient to explain the trend. The reduction has occurred for most provinces and direct drivers, suggesting that national policies and structural changes has been a major factor, more than commodity-specific ones (commodity prices, certification, etc.). Yet there are major differences across provinces, with some (previously) high-deforestation provinces having reduced deforestation by up to 80%. An interesting regional pattern is observed for the forest encroachment factor (EF) for different commodities - the share of land new agricultural land that is from forest conversion. The reduction in the EF for Sumatra is noteworthy, while it has remained relatively stable for Kalimantan. There has, however, been some marked changes over time in the direct drivers of deforestation, i.e., the commodities that are produced on the land after the trees have been removed. Palm oil still is the single most important commodity, and the MBI data suggest that 46% of the postforest uses during 2018-2022 was for oil palm, compared to 55% during 2010-2017. At the same time, the encroachment factor of palm oil has declined in recent years, i.e., a smaller share of the newly established oil palm is on previously forest land. This is true particularly in Sumatra. For Indonesia, however, the main factor is reduced oil palm expansion into any forms of land, while reduced encroachment factor comes second. Several reports suggest that pulpwood is on the rise as a direct driver, although MBI data does not support that claim. Nusantara Atlas data does, however, point to a major increase in Kalimantan (and a move in pulp-driven deforestation from Sumatra to Kalimantan). In the three high-deforestation provinces of Kalimantan, pulp made up 29% of the deforestation in 2022. 53 Mineral mining is another direct driver, although the overall share remains relatively low. According to MBI data, mining was responsible for 3.4% of the deforestation in Indonesia for the 2018-2022 period, up from 0.9% in the 2010-2017 period (with a much higher share in the last two years (2021-2022): 5.4%). Mining is particularly important in Sulawesi, where it accounted for 30% of the deforestation in 2021-2022, compared to 8% during 2010-2017. The expansion of other agricultural commodities, such as rubber, is not as easy to identify, as detailed data is missing and challenges exist in distinguishing, for example, jungle rubber from forests with remote sensing methods. The share of land categorised as “Other agricultural land” has, however, increased between the latest periods (MBI data). Finally, timber production seems to play a decreasing role, for reasons discussed below. 6.2 The hypotheses We put forward five possible explanations (hypotheses) of the decline, and we review each of them in turn. 6.2.1 H1: Government policies Public policy reforms take time to have an impact on the ground. The reduction of deforestation over the last few years is the result of policies and measures that were enacted much earlier. Several events led to major policy reforms during the decade of 2010-2020. In 2007, Indonesia hosted COP13 of UNFCCC, where REDD+ was officially adopted and became part of the Bali Road Map, and thus a main element of the international climate mitigation agenda. As the COP13 host, also Indonesia played a major role in the follow up work. In May 2010, Norway and Indonesia signed a Letter of Intent, in which Norway pledged up to USD 1 billion as payment for reduced emissions. A key result of the agreement was the moratorium of 2011. The impact of that (and the later additions) is still debated. Scholarly studies lend some support to the moratorium having a positive conservation effect, although the effect size is limited compared to the commitments. An important side effect of the moratorium has been more coordination across sectoral ministries and levels of government, also part of the One map initiative. In 2015, massive forest fires and resulting haze in Indonesia and its neighbours, which by some estimates caused about 100 000 premature deaths due to the smoke exposure.24 It caused major domestic concerns and put fire management and peatland protection at the top of the political agenda. Several important policy measures where implemented, and the interviewees highlights this among the most important policies in explaining the deforestation reduction. Other policy initiatives include schemes for results-based payment and social forestry. An overall assessment of these is lacking, but we conjecture that these have been important in the 24 https://seas.harvard.edu/news/2016/09/smoke-2015-indonesian-fires-may-have-caused-100000-prematuredeaths#:~:text=Schools%20and%20businesses%20closed%2C%20planes,Indonesia%2C%20Malaysia%2C%20 and%20Singapore. 54 locations where implemented. RBP has worked in East Kalimantan and potentially Jambi. Similarly, village forest shows its capacity to reduce deforestation. However, our assessment is that they have had more limited impact on national deforestation rates. 6.2.2 H2: Private policies and industry ac�on The story of private sector forest initiatives and their impacts is a mixed one. The area under certification, at least for oil palm, has increased, and a few studies suggest a positive forest impact of certification schemes and pledges. However, if fully implemented, we should have seen a sharp decline in the forest encroachment factors, something which only can be observed for Sumatra and is more likely to be due to a forest scarcity effect (see below). Much of the certification debate and initiatives has focused on palm oil. For timber, pulp, companies can apply FSC or PEFC certification, yet it is not widespread. In particular, the large conglomerate Sinar Mas (which includes the subsidiary Asia Pulp & Paper), despite having adopted PEFC certification, are still working to end its disassociation with FSC. Since the last decade, palm oil and pulp and paper companies have increased their commitment to eliminate deforestation from their supply chains by issuing NDPE policies. There are early signs that the pledges have produced some reduction in deforestation, but companies with NDPE policies still face challenges in ensuring that smallholders and indirect suppliers comply to their NDPE policies. 6.2.3 H3: Civil society pressure Civil society organizations (CSOs) play an indirect but important role by influencing public policy makers and corporate actors. The role and impact are harder to assess, as one has to track specific policy creation and implementation processes. CSO plays an important role in several areas. First, they are active actors on the policy arena. What is noteworthy in Indonesia is that this has not been limited to public policies, but there are several examples of their role in private sector initiatives. Second, CSO are important watchdogs for both implementation of public and private regulations and pledges. Several such breaches have been brought to public attention and in some cases ended up in the courts. Finally, CSO work locally in the implementation of specific projects, with potentially significant positive impacts locally, but most likely insufficient in numbers and coverage to affect national deforestation rates significantly. 6.2.4 H4: Markets and prices of deforesta�on-risk commodi�es Several commodities are involved in the deforestation process. Overall, prices of key deforestation-risk commodities were relatively stable during the 2016-2020 period, while some have increased during the pandemic (2020-2022) and also after 2022 as a result of the war in Ukraine. Changing commodity prices were not a driver behind the reduced deforestation. Nevertheless, we conjecture that stable price of key commodities made implementing both public and private policies less costly for politicians and producers, which also improved policy effectiveness. 61 Veen, F. J. F. (2020). Tropical forest and peatland conservation in Indonesia: Challenges and directions. People and Nature, 2(1), 4–28. https://doi.org/10.1002/pan3.10060 Heijlen, W., & Duhayon, C. (2024). An empirical estimate of the land footprint of nickel from laterite mining in Indonesia. The Extractive Industries and Society, 17, 101421. https://doi.org/10.1016/j.exis.2024.101421 Heilmayr, R., & Benedict, J. (2022, September 14). Indonesia makes progress towards zero palm oil deforestation. Trase.Earth. https://trase.earth/insights/indonesia-makes-progresstowards-zero-palm-oil-deforestation Heilmayr, R., Carlson, K. M., & Benedict, J. J. (2020). Deforestation spillovers from oil palm sustainability certification. Environmental Research Letters, 15(7), 075002. https://doi.org/10.1088/1748-9326/ab7f0c Ichsan, A. C., Prayoga, A. P., & Dalya, N. (2021). Study of forest management unit performance in Indonesia. IOP Conference Series: Earth and Environmental Science, 681(1), 012004. https://doi.org/10.1088/1755-1315/681/1/012004 Jong, H. N. (2021a, March 9). Deforestation in Indonesia hits record low, but experts fear a rebound. Mongabay Environmental News. https://news.mongabay.com/2021/03/2021deforestation-in-indonesia-hits-record-low-but-experts-fear-a-rebound/ Jong, H. N. (2021b, September 10). Indonesia terminates agreement with Norway on $1b REDD+ scheme. https://news.mongabay.com/2021/09/indonesia-terminates-agreement-withnorway-on-1b-redd-scheme/ Jong, H. N. (2022, September 12). Indonesia and Norway give REDD+ deal another go after earlier breakup. Mongabay. https://news.mongabay.com/2022/09/indonesia-and-norway-giveredd-deal-another-go-after-earlier-breakup/ Jong, H. N. (2024, February 15). Palm oil deforestation makes comeback in Indonesia after decade-long slump. Mongabay. https://news.mongabay.com/2024/02/palm-oil-deforestationmakes-comeback-in-indonesia-after-decade-long-slump/ Kaimowitz, D., & Angelsen, A. (1998). Economic Models of Tropical Deforestation: A Review. CIFOR. Kementerian Lingkungan Hidup dan Kehutanan. (2023). Statistik 2022 Kementerian Lingkungan Hidup dan Kehutanan. Kiely, L., Spracklen, D. V., Arnold, S. R., Papargyropoulou, E., Conibear, L., Wiedinmyer, C., Knote, C., & Adrianto, H. A. (2021). Assessing costs of Indonesian fires and the benefits of restoring peatland. Nature Communications, 12(1), 7044. https://doi.org/10.1038/s41467-02127353-x 62 Kim, Y.-S., Bae, J. S., Fisher, L. A., Latifah, S., Afifi, M., Lee, S. M., & Kim, I.-A. (2016). Indonesia’s Forest Management Units: Effective intermediaries in REDD + implementation? Forest Policy and Economics, 62, 69–77. https://doi.org/10.1016/j.forpol.2015.09.004 Kosar, M., Pramudiana, D., Ferdyan, D. T., & Maulita, A. (2019). PHPL: From legality to sustainability. JPIK. Kraus, S., Liu, J., Koch, N., & Fuss, S. (2021). No aggregate deforestation reductions from rollout of community land titles in Indonesia yet. Proceedings of the National Academy of Sciences of the United States of America, 118(43), e2100741118. https://doi.org/10.1073/pnas.2100741118 Lee, G. (2023, July 4). Explainer: Why deforestation has hit a historic low in Indonesia. EcoBusiness. https://www.eco-business.com/news/explainer-why-deforestation-has-hit-a-historiclow-in-indonesia/ Leijten, F., Sim, S., King, H., & Verburg, P. H. (2021). Local deforestation spillovers induced by forest moratoria: Evidence from Indonesia. Land Use Policy, 109, 105690. https://doi.org/10.1016/j.landusepol.2021.105690 Leony Aurora. (2012, June 27). Indonesia seeks “multi-bilateralism” to help get REDD+ off the ground. Forest News. https://forestsnews.cifor.org/9932/indonesia-seeks-multibilateralism-to-help-get-redd-off-the-ground?fnl=en Madani Berkelanjutan. (2021). Palm oil moratorium, importantto be extended and strengthened. https://madaniberkelanjutan.id/palm-oil-moratorium-important-to-be-extendedand-strengthened/ Meijaard, E. (2017, September 17). Does social forestry always decrease deforestation and poverty? (Commentary). Mongabay. https://news.mongabay.com/2017/09/social-forestrysometimes-but-not-always-decreases-deforestation-and-poverty-commentary/ Meutia, A. A., Lumowa, R., & Sakakibara, M. (2022). Indonesian Artisanal and Small-Scale Gold Mining—A Narrative Literature Review. International Journal of Environmental Research and Public Health, 19(7), 3955. https://doi.org/10.3390/ijerph19073955 Ministry of Environment and Forestry. (2023). Indonesia: Emission reductions program. MoEF and Fofrest Carbon Partnership Facility. Miteva, D. A., Loucks, C. J., & Pattanayak, S. K. (2015). Social and Environmental Impacts of Forest Management Certification in Indonesia. PLOS ONE, 10(7), e0129675. https://doi.org/10.1371/journal.pone.0129675 MoEF. (2022). The state of Indonesia’s forests 2022 (p. 96). Ministry of Environment and Forestry. Mosnier, A., Boere, E., Reumann, A., Yowargana, P., Pirker, J., Havlik, P., & Pacheco, P. (2017). Palm oil and likely futures: Assessing the potential impacts of zero deforestation 63 commitments and a moratorium on large-scale oil palm plantations in Indonesia (177; CIFOR Infobrief). CIFOR. https://www.cifor-icraf.org/publications/pdf_files/infobrief/6468infobrief.pdf Muttaqien, A. (2023, July 18). Pemutihan Kebun Sawit di Kawasan Hutan Pupuskan Langkah Pemulihan Lingkungan. Mongabay. https://www.mongabay.co.id/2023/07/18/pemutihankebun-sawit-di-kawasan-hutan-pupuskan-langkah-pemulihan-lingkungan/ Nababan, W. M. C. (2023, June 24). Perbaikan Tata Kelola, Pemerintah Akan Putihkan 3,3 Juta Hektar Sawit di Kawasan Hutan. Kompas. https://www.kompas.id/baca/ekonomi/2023/06/23/perbaikan-tata-kelola-pemerintah-akanputihkan-33-juta-hektare-sawit-di-kawasan-hutan Nugroho, H. Y. S. H., Indrajaya, Y., Astana, S., Murniati, Suharti, S., Basuki, T. M., Yuwati, T. W., Putra, P. B., Narendra, B. H., Abdulah, L., Setyawati, T., Subarudi, Krisnawati, H., Purwanto, Saputra, M. H., Lisnawati, Y., Garsetiasih, R., Sawitri, R., Putri, I. A. S. L. P., … Rahmila, Y. I. (2023). A Chronicle of Indonesia’s Forest Management: A Long Step towards Environmental Sustainability and Community Welfare. Land, 12(6), Article 6. https://doi.org/10.3390/land12061238 Perz, S. G. (2007). Grand Theory and Context-Specificity in the Study of Forest Dynamics: Forest Transition Theory and Other Directions. The Professional Geographer, 59(1), 105–114. https://doi.org/10.1111/j.1467-9272.2007.00594.x Purwanto, E. (2015). Challenges of the Social Forestry Program in Indonesia and the Roles of CSOs. Tropenbos Indonesia. Rahayu, S., Laraswati, D., Permadi, D. B., Sahide, M. A. K., & Maryudi, A. (2023). Only a Noise? The Role of Non-governmental Organizations in the Policy Processes of a New Social Forestry Model in Indonesia. Small-Scale Forestry, 22(2), 253–270. https://doi.org/10.1007/s11842-022-09525-9 Ramadhan, R., Karim, S., Fisher, M. R., Mursyid, H., & Indrawan, M. (2023). Assessing the Governance Modes of Indonesia’s Forest Management Unit. In Environmental governance in Indonesia (pp. 151–166). Springer. RSPO. (2023). Impact update 2023. RSPO. Rudel, T. K., Coomes, O. T., Moran, E., Achard, F., Angelsen, A., Xu, J., & Lambin, E. (2005a). Forest transitions: Towards a global understanding of land use change. Global Environmental Change, 15(1), 23–31. https://doi.org/10.1016/j.gloenvcha.2004.11.001 Rudel, T. K., Coomes, O. T., Moran, E., Achard, F., Angelsen, A., Xu, J., & Lambin, E. (2005b). Forest transitions: Towards a global understanding of land use change. Global Environmental Change, 15(1), 23–31. https://doi.org/10.1016/j.gloenvcha.2004.11.001 64 Sahide, M. A. K., Maryudi, A., Supratman, S., & Giessen, L. (2016). Is Indonesia utilising its international partners? The driving forces behind Forest Management Units. Forest Policy and Economics, 69, 11–20. https://doi.org/10.1016/j.forpol.2016.04.002 Samnuzulsari, T., Kristanti, D., Susanto, H., Yudiatmaja, W. E., Yunanto, M. K., Meilinda, S. D., & Akbar, D. (2021). The strategic role of civil society organisations in handling climate change: A case of Riau in Indonesia. IOP Conference Series: Earth and Environmental Science, 824(1), 012104. https://doi.org/10.1088/1755-1315/824/1/012104 Santika, T., Meijaard, E., Budhiharta, S., Law, E. A., Kusworo, A., Joseph A. Hutabarat, Tito P. Indrawan, Matthew Struebig, Sugeng Raharjo, Imanul Huda, Sulhani, Andini D. Ekaputri, Soni Trison, Madelaine Stigner, & Kerrie A. Wilson. (2017). Community forest management in Indonesia: Avoided deforestation in the context of anthropogenic and climate complexities. Global Environmental Change, 46, 60–71. https://doi.org/10.1016/j.gloenvcha.2017.08.002 Santika, T., Wilson, K. A., Budiharta, S., Kusworo, A., Meijaard, E., Law, E. A., Friedman, R., Hutabarat, J. A., Indrawan, T. P., St. John, F. A. V., & Struebig, M. J. (2019). Heterogeneous impacts of community forestry on forest conservation and poverty alleviation: Evidence from Indonesia. People and Nature, 1(2), 204–219. https://doi.org/10.1002/pan3.25 Sastra, Y. (2024, January 24). Social Forestry in West Sumatra Reduces the Rate of Deforestation—Kompas.id. Kompas. https://www.kompas.id/baca/english/2024/01/24/enperhutanan-sosial-di-sumbar-tekan-laju-deforestasi Scholte, M. (2019). Indonesia curtails deforestation: German development cooperation lends support, a journalistic view (p. 35). GIZ Forclime. Sembiring, R., Fatimah, I., & Widyaningsih, G. A. (2020). Indonesia’s Omnibus Bill on Job Creation: A Setback for Environmental Law? https://doi.org/10.1163/24686042-12340051 Seymour, F. J., Aurora, L., & Arif, J. (2020). The Jurisdictional Approach in Indonesia: Incentives, Actions, and Facilitating Connections. Frontiers in Forests and Global Change, 3. https://doi.org/10.3389/ffgc.2020.503326 Sucahyo, N. (2023, June 24). Pakar: Pemutihan Kebun Sawit di Area Hutan Untungkan Korporasi. VoA Indonesia. https://www.voaindonesia.com/a/pakar-pemutihan-kebun-sawitdi-area-hutan-untungkan-korporasi-/7151177.html Sunarto, S., Shafira, M., & Anwar, M. (2021). Implications of the Omnibus Law on Job Creation towards the Indonesian Forestry Sector. Fiat Justisia: Jurnal Ilmu Hukum, 15(3), Article 3. https://doi.org/10.25041/fiatjustisia.v15no3.2302 Supriatna, J., Shekelle, M., Fuad, H. A. H., Winarni, N. L., Dwiyahreni, A. A., Farid, M., Mariati, S., Margules, C., Prakoso, B., & Zakaria, Z. (2020). Deforestation on the Indonesian island of Sulawesi and the loss of primate habitat. Global Ecology and Conservation, 24, e01205. https://doi.org/10.1016/j.gecco.2020.e01205 65 Suwarno, A., van Noordwijk, M., Weikard, H.-P., & Suyamto, D. (2018). Indonesia’s forest conversion moratorium assessed with an agent-based model of Land-Use Change and Ecosystem Services (LUCES). Mitigation and Adaptation Strategies for Global Change, 23(2), 211–229. https://doi.org/10.1007/s11027-016-9721-0 Tabacek, K. (2010, March 19). Nestlé sets timetable for palm oil decision. The Guardian. https://www.theguardian.com/sustainable-business/nestle-cargill-palmoil Terzano, D., Attorre, F., Parish, F., Moss, P., Bresciani, F., Cooke, R., & Dargusch, P. (2022). Community-led peatland restoration in Southeast Asia: 5Rs approach. Restoration Ecology, 30(8), e13642. https://doi.org/10.1111/rec.13642 TheTreeMap. (2023, January 5). Palm oil deforestation in Indonesia levels off in 2022. Nusantara Atlas. https://nusantara-atlas.org/palm-oil-deforestation-in-indonesia-levels-off-in2022/ Tropical Forest Alliance & LPEM Universitas Indonesia. (2022). Investor brief: Sustainable land use of soft commodities—Challenges and opportunities in Indonesia (01/2022; p. 18). Tropical Forest Alliance. https://www.lpem.org/wpcontent/uploads/2022/11/Investor_Brief_Soft_Commodities_Designed.pdf Turubanova, S., Potapov, P. V., Tyukavina, A., & Hansen, M. C. (2018). Ongoing primary forest loss in Brazil, Democratic Republic of the Congo, and Indonesia. Environmental Research Letters, 13(7). https://doi.org/10.1088/1748-9326/aacd1c Tyukavina, A., Potapov, P., Hansen, M. C., Pickens, A. H., Stehman, S. V., Turubanova, S., Parker, D., Zalles, V., Lima, A., Kommareddy, I., Song, X.-P., Wang, L., & Harris, N. (2022). Global Trends of Forest Loss Due to Fire From 2001 to 2019. Frontiers in Remote Sensing, 3. https://www.frontiersin.org/articles/10.3389/frsen.2022.825190 University of California Santa Barbara, Woods and Wayside International, TheTreeMap, Stockholm Environment Institute, & Global Canopy. (2023). Deforestation surge ends a decade of progress for Indonesia’s pulp sector. https://doi.org/10.48650/KWRA-E756 Vancutsem, C., Achard, F., Pekel, J. F., Vieilledent, G., Carboni, S., Simonetti, D., Gallego, J., Aragão, L. E. O. C., & Nasi, R. (2021). Long-term (1990-2019) monitoring of forest cover changes in the humid tropics. Science Advances, 7(10), 1–22. https://doi.org/10.1126/sciadv.abe1603 Wahyu, A., Suharjito, D., Darusman, D., & Syaufina, L. (2020). The Development of Community-Based Forest Management in Indonesia and Its Contribution to Community Welfare and Forest Condition. IOP Conference Series: Earth and Environmental Science, 528(1), 012037. https://doi.org/10.1088/1755-1315/528/1/012037 Wiati, C. B., Dharmawan, I. W. S., Sakuntaladewi, N., Ekawati, S., Wahyuni, T., Maharani, R., Hadiyan, Y., Naibaho, Y., Satria, W. I., Ngatiman, N., Abdurachman, A., Karmilasanti, K., Laksmita, A. N., Angi, E. M., & Khadka, C. (2022). Challenges to and Strategies for the 66 Climate Village Program Plus: A Lesson Learned from Indonesia. Sustainability, 14(9), Article 9. https://doi.org/10.3390/su14095530 Wolfersberger, J., Delacote, P., & Garcia, S. (2015). An empirical analysis of forest transition and land-use change in developing countries. Ecological Economics, 119, 241–251. https://doi.org/10.1016/j.ecolecon.2015.08.018 Yuwati, T. W., Rachmanadi, D., Pratiwi, Turjaman, M., Indrajaya, Y., Nugroho, H. Y. S. H., Qirom, M. A., Narendra, B. H., Winarno, B., Lestari, S., Santosa, P. B., Adi, R. N., Savitri, E., Putra, P. B., Wahyuningtyas, R. S., Prayudyaningsih, R., Halwany, W., Nasrul, B., Bastoni, & Mendham, D. (2021). Restoration of Degraded Tropical Peatland in Indonesia: A Review. Land, 10(11), Article 11. https://doi.org/10.3390/land10111170 67 Annex 1: Interview guide Resource persons were selected based on the information obtained in the literature review, as well as representatives for a diversity of actors in the field: government (national and subnational), private sector, donor community, and civil society organizations. We started asking the resource persons with open questions on the trends, policies, and measures. For each policy and measure, we asked more detailed information on the characteristics of the policies and measures that may contribute to the reduction of deforestation. In some interview sessions, we focus on certain policies that we found in the literature review. Most of the 32 resource persons interviewed requested to remain anonymous, and we have therefore not included the names and affiliations. The guiding questions were: 0. Introduction a. Introducing the interviewer, its organizational affiliation, and the project. b. Respondents may answer the questions partly without any consequences. c. Permission to record the audio only for research purposes. No recording if the resource person does not approve. 1. General a. Describe about you and your organization. b. Describe how you or your organization are related to the analysis of deforestation trends in Indonesia (policymaker, business, analyst, campaign, observer, etc). 2. Deforestation trends a. Explain what you understand about the trend of deforestation in the last 5 years, and in the longer term. b. Explain what you understand about the commodity, location, type of land use, and magnitude (size). c. Explain how you know that deforestation has decreased in the last few years. 3. Factors that contribute to the reduction of deforestation. a. Explain what public policies, programs, and measures that you think have contributed to the reduction of deforestation. b. For each public policy, how has the policy contributed to the reduction of deforestation? c. Explain what private actions that you think have contributed to the reduction of deforestation. d. For each private action, how has the action contributed to the reduction of deforestation? e. What are the roles of civil society organizations in reducing deforestation? To what extent they contribute to the reduction of deforestation? f. What are the roles of other stakeholders (e.g donor programs) in reducing deforestation? To what extent they contribute to the reduction of deforestation? g. To what extent the emergence of customary forest contributes to the reduction of deforestation? h. What other factors that you think may contribute to the reduction of deforestation? How did these factors play out? 4. Do you think the deforestation reduction will continue? What are the risk factors that may increase deforestation in the future?