WPS6495 Policy Research Working Paper 6495 A Resource Allocation Model for Tiger Habitat Protection Susmita Dasgupta Dan Hammer Robin Kraft David Wheeler The World Bank Development Research Group Environment and Energy Team June 2013 Policy Research Working Paper 6495 Abstract Habitat conservation is critical to the survival of second stage, additional user-specified weights are used endangered tigers. This paper develops a resource- to combine the composite indices into priority scores and allocation model for the protection of tiger habitats, potential project budget shares for all 74 habitat areas. using information on threats to particular tiger Results suggest that changes in user-specified weights can subspecies, the quality of remaining habitat areas, the have very a significant impact on habitat priority scores. observed effectiveness of habitat protection by country, Illustrative scenarios indicate that the model can make and the potential costs of protection projects in74 a useful contribution by identifying priority orderings habitats in Asia. This model will be implemented in two that are consistent with different sets of preferences. It stages. The first stage involves using user-specified weights will also provide feedback to decision makers regarding to combine numerous sub-indices into composite indices, the implicit preferences associated with their resource covering threats to species, habitat quality, potential allocation decisions. project costs and the effectiveness of protections. At the This paper is a product of the Environment and Energy Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at SDasgupta@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team A Resource Allocation Model for Tiger Habitat Protection Susmita Dasgupta, Dan Hammer, Robin Kraft, David Wheeler* Keywords: Biodiversity Conservation; Tiger Habitat: Resource Allocation JEL Classification: Q23, Q57 World Bank Sector: ENV This research was funded by the Knowledge for Change Program. * Authors’ names in alphabetical order. The authors are respectively Lead Environmental Economist, Development Research Group, World Bank; Ph.D. Candidate, Department of Agricultural and Resource Economics, University of California, Berkeley; Research Associate, World Resources Institute; and Senior Fellow Emeritus, Center for Global Development. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Corresponding Author: Susmita Dasgupta. MSN: MC-3-308, The World Bank. 1800 H Street, Washington DC 20433, USA. Telephone: 1-202-473-2679. E-mail: sdasgupta@worldbank.org 1. Introduction The International Union for Conservation of Nature has classified tiger as endangered in its Red list of Threatened Species of (http://www.iucnredlist.org/details/15955/0 ). The wild tiger population of tropical Asia dropped from about 100,000 to 3,500 in the past century. The Bali, Javan and South China subspecies are believed to be extinct in the wild. An estimated 2,380 Bengal tigers survive, along with 340 Indochinese, 500 Malayan and 325 Sumatran tigers. The surviving wild tiger population of tropical Asia inhabits a scattered arc from southwest India to northwest Indonesia, much of it in upland areas. Bengal tigers survive in India, Nepal, Bhutan, Bangladesh and northern Myanmar, while the remaining Indochinese tigers are found in western and southeastern Myanmar 1, Lao PDR, Vietnam, Cambodia and Thailand. In contrasting geographic concentration, Sumatran tigers are confined to one Indonesian island and Malayan tigers exist only in Peninsular Malaysia and one small area in southern Thailand. Long term survival of the tiger is dependent on conservation of tiger habitats. 2 The global community has mobilized to conserve the tiger’s remaining habitat through the Global Tiger Initiative, which is supported by all countries with known tiger populations, the World Bank, and over 40 civil society organizations. 3 Participating countries and institutions have endorsed the Global Tiger Recovery Program (GTRP), which aims to double the number of wild tigers by 2022 through habitat conservation programs and cooperation across national boundaries to stop poaching and illegal trade in tiger parts. Operating under tight budget constraints, the GTRP confronts several complicating factors, including the need to conserve specific habitats large enough to accommodate this keystone 1 The ranges of the Bengal and Indochinese tigers may overlap in Myanmar. 2 This paper focuses on forest habitat loss, but we recognize that poaching and illegal trade in tiger parts can devastate remaining tiger populations, even when forest habitat is intact. 3 For more complete information, see http://www.globaltigerinitiative.org/html/participants.php 2 predator; differential threats to important regional subspecies that must be preserved (Bengal, Indochinese, Malayan and Sumatran tigers); divided national jurisdictions; differences in countries’ institutional capabilities, conservation management costs, and willingness to pay for conservation; and, not least, widely-differing and rapidly-changing opportunities for commercial exploitation of remaining habitat areas. Cost-effective resource allocation under these dynamic conditions involves frequent reassessment of threats and opportunities in many habitat areas scattered across the tiger range countries. Timely analysis requires near-real-time forest monitoring information, which is now available from FORMA (Forest Monitoring for Action), a new database developed by a research consortium that includes the Center for Global Development (http://www.cgdev.org/forma), the World Resources Institute, the University of Maryland, and Resources for the Future, in consultation with World Bank staff, Conservation International, the Nature Conservancy and WWF. Drawing on data from NASA’s MODIS system, FORMA provides monthly updates on forest clearing at high spatial resolution for tropical Asian countries with tiger habitat. From a formal analytical perspective, saving the tiger involves optimal spatial resource allocation with a limited budget, a short time horizon (to prevent extinction), a complex, constantly-changing spatial distribution of potential conservation benefits and costs, and the prospect of rapid, irreversible losses in areas where conservation is weak. Combining the FORMA information with a spatially-disaggregated database of economic, demographic and geographic information, previous papers in this series have focused on identifying area-specific drivers of habitat destruction (Dasgupta, et al. 2012a), and testing the effects of interventions intended to protect habitat (Dasgupta, et al., 2012b). 3 In this paper, we draw on information and insights from the previous papers to develop a model that can inform the allocation of GTRP resources for tiger habitat conservation. Designed for frequent updates as new information becomes available, the model estimates priorities and potential program budget shares for 74 tiger habitat areas in 9 countries: Cambodia, India, Indonesia, Lao PDR, Myanmar, Malaysia, Nepal, Thailand and Vietnam. 4 It incorporates four key factors for resource allocation: Differential threats to tiger subspecies, habitat quality, potential project implementation costs, and recent evidence on the effectiveness of countries’ protection policies. The remainder of the paper is organized as follows. Section 2 develops the modeling framework, while Section 3 introduces the data that drive the model. Section 4 describes the operation of the model. Section 5 illustrates its operation in three different scenarios, and Section 6 summarizes and concludes the paper. 2. Allocating Resources for Tiger Habitat Protection To provide a consistent modeling framework, we adopt an approach to optimal allocation that draws on prior theoretical and empirical work by Behrman, Pollak and Taubman (1982), Bolt, et al. (2003), Buys, et al. (2004), Pandey, et al. (2005) and Wheeler (2011). In this model, the welfare impact of conservation project expenditures is a function of their levels and distributions across tiger habitat areas. Resource allocation decisions by the Global Tiger Recovery Program must incorporate and balance three factors: tiger subspecies preservation; representation for all participating countries; and overall welfare maximization. We cannot realistically characterize the program’s objective function as linear (infinite elasticity of substitution across habitat areas), because sole allocation to one area is infeasible, whatever the 4 Bangladesh and Bhutan are not included in this exercise because FORMA data are not yet available for their tiger habitat areas. They will be included in future updates as the data become available. 4 relative scale of its protection problem. Broader geographic coverage is implied by the program’s charter. At the same time, the objective function is not purely fixed-coefficient (zero elasticity of substitution across areas), because nothing forces it to maintain cross-country parity in allocation. This is good, since the distribution of habitat protection problems is likely to be far from even across areas. We adopt an intermediate assumption: that the objective function is characterized by unit- elastic substitution across areas. A unit-elastic (Cobb-Douglas) welfare function permits tailoring of programs to area-specific conditions, while encouraging portfolio diversification through the operation of diminishing returns. Expected welfare gains from expenditures are related to both the scale of habitat protection problems and the cost of successful protection under local conditions. Formally, we specify the resource allocation problem as follows: Ï?i Welfare Function: (1) W = Π Tijk ijk where Tijk = Expected number of surviving tigers of subspecies i in habitat j, country k Ï?i = Exogenous extinction risk for tigers of subspecies i. Extinction risk is a function of overall deterioration of the subspecies’ habitat in the region (D), the number of surviving tigers in the subspecies (T), and the number of countries (N) that still harbor them: (2) Ï? i = γ 0 Diγ 1 Ti γ 2 N iγ 3 The expected number of surviving tigers is a function of habitat quality and the effectiveness of habit protection in country k. α2 α Qα (3) Tijk = α 0G jk1 1 jk Pk where Gjk = Scale of GTRP activity in habitat j, country k Qjk = Quality of habitat j, country k 5 Pk = Effectiveness of habitat protection in country k Habitat quality is a function of habitat size (H) and fragmentation (F): (4) Q jk = δ 0 H δ δ2 jk F jk 1 Resource allocation is limited by an overall budget constraint: ï?· (5) B = ∑ C jk G jk jk where B = Available budget Cjk = Unit cost of protection for habitat jk Unit protection costs reflect a combination of economic incentives for local conservation and the direct cost of habitat protection, which is principally a function of local labor cost. 𝛽 𝛽2 (6) Cjk = 𝛽0 𝐿𝑗𝑘 1 𝑊𝑘 where L = Opportunity cost of land in habitat jk W = Labor cost in country k Substituting (3) into (1) yields the following welfare function: α2 α Qα (7) W = ωΠ G jk1 1 jk Pk ijk Maximization of W subject to the overall budget constraint yields the following ratio of optimal GTRP resource allocations for protection of arbitrarily-chosen habitat areas 1 and 2 with tiger subspecies m and n: α1 α 2 α 3 C1 G1* Ï? m Q P (8) = α1 1α 2 1α 3 C2G2 Ï? n Q2 P2 * Substituting and re-arranging, we obtain: α1 ˆ α 2 α 3 ˆ α 4 G1* Ï? ˆm Q1 P1 C1 (9) = α1 ˆ α 2 α 3 ˆ α 4 G2 Ï? * ˆ n Q2 P2 C2 The program priority score of habitat area 1 is the numerator of (9), where Ï? ˆ are ˆ and C ˆ ,Q calculated from (2), (4) and (6), respectively. Its share of the program budget is 6 S1 (10) s1 = N ∑S j =1 j 3. Quantifying Habitat Protection Factors To implement the model, we quantify the allocation factors using the database developed in Dasgupta, et al. (2012a). For each tiger habitat area, we develop composite measures as follows. Variables are denoted by letters from the previous equations. (1) Subspecies extinction risk (Ï?). Our measure has three components: a. Species numbers (T): Table 1 provides recent estimates of surviving tiger populations, by subspecies and country. Our measure for each subspecies is its share of the total tiger population. b. Overall habitat loss (D): For each habitat area, we compute natural forest cover remaining by August, 2011. This combines measures from Hansen, et al. (2006) for 2000, Hansen, et al. (2008) for 2000 – 2005, and FORMA data for January, 2006 to August, 2011. Across all areas, we total original and remaining forest cover by subspecies 5, and compute the overall percentage of original forest that has been subject to clearing. c. Species distribution across countries (N): In standard portfolio terms, spreading the remaining tigers across polities should reduce aggregate risk. Our index of “portfolio riskâ€? is the intercountry entropy measure (E) for each subspecies, where K Ei = − ∑ sik ln sik k =1 and sk = country k’s share of surviving tigers of subspecies i. 5 See the Introduction, para. 1, for the geographic ranges of Bengal, Indochinese, Malayan and Sumatran tigers. 7 Table 1: Surviving Wild Tiger Populations* Bengala Indochinese Malayan Sumatran India 1,706 Thailand 200 Malaysia 500 Indonesia 325 Bangladesh 440 Myanmar 85 (Peninsular) (Sumatra) Nepal 155 Vietnam 20 Bhutan 75 Cambodia 20 Lao PDR 17 Total 2,376 342 500 325 * Midrange estimates a No current estimate for Myanmar Source: Dasgupta, et al. (2012a) (2) Habitat Quality (Q). This includes two components: Habitat area (H) and the degree of fragmentation (F). For the latter, we use the percentage of habitat area not subject to clearing by August, 2011. (3) Effectiveness of protection (P): For all tiger habitat areas within a country, we total original and remaining forest area and compute the shares of original forest that were subjected to clearing by 2000 and August, 2011, respectively. The increase in share from 2000 to 2011 is a proxy for the risk of protection failure during the period. (4) Protection project Cost (C). This includes two components: a. Habitat land opportunity cost (L): In Dasgupta, et al. (2012a), we specify a model that relates local forest clearing to the profitability of alternative uses for the land. By implication, the best single measure of the opportunity cost of forested land in a locality is the intensity of recent clearing. Accordingly, our proxy for opportunity cost in each habitat is the increase in the percent of its area cleared from 2000 to August, 2011. b. Labor cost (W): We incorporate relative protection project costs using differential wages, with income per capita as the proxy and an exponential weight 8 (0.6) that reflects the findings of Harrison (2002) on the labor share of income in low- and middle-income countries. 6 4. Model Implementation We transform all model variables to ranks for ease of use and interpretation, and to ensure robustness against outlier effects. To facilitate the nested procedure that we describe below, our rank-ordering for each variable ensures a positive relationship to the priority scores in (9). Accordingly, higher numerical ranks are assigned to lower values for subspecies numbers (T), subspecies distribution (E), land opportunity cost (L) and labor cost (W). Higher numerical ranks are assigned to higher values for subspecies’ overall habitat loss (D), habitat area (H), the percent of habitat remaining (F), and the effectiveness of country protection (P). Our model incorporates equations (1) – (10) in a two-stage exercise. In the first stage, the user specifies relative weights for the determinants of Ï?, Q and C in (2), (4) and (6) above. 7 The model allows the user to specify the parameters for each equation in arbitrary units. After the parameters are specified, the model standardizes to Cobb-Douglas (CD) parameters by dividing each parameter by the sum of parameters for that equation. Then it forms the product of the relevant rank-transformed variables, each weighted by its exponential CD parameter. At the completion of the first stage, the model has created indices for Ï?, Q and C. P (protection effectiveness) has only one component, so that is the second-stage index for this variable. At the beginning of the second stage, the user specifies relative weights for species risk (Ï?), habitat quality (Q), protection effectiveness (P) and project cost (C). As before, the model allows the user to specify the weighting parameters in arbitrary units. Then it standardizes to CD 6 Formally, this index assumes a Cobb-Douglas (unit-elastic) cost function, internationally-traded capital and non- traded labor. The cost elasticity of the average wage (proxied by income per capita) in this function is the labor share of national income. 7 There is no need to specify the constants (γ0 δ0 β0), since they do not affect the relative values that determine priority scores. 9 parameters and forms the product of the four indices, each weighted by its exponential CD parameter. The most critical question for our modeling exercise relates to the potential variation in outcomes for different user settings of model parameters. To address this question, Table 2 provides rank correlations for the 8 variables incorporated in the model. The results are almost evenly divided between positive and negative correlations; some are strong but most are relatively weak. Overall, these results indicate that the habitat priorities and implicit budget shares calculated by the model are highly dependent on user-specified weights. Table 2: Rank Correlations for Model Variables Habitat Subspecies Habitat Land Habitat Subspecies Subspecies Habitat Habitat Percent Opp. Labor Obs = 74 Numbers Distribution Threat Size Remaining Cost Cost Subspecies Numbers 1.00 Subspecies Distribution 0.06 1.00 Subspecies Habitat Threat -0.75 -0.43 1.00 Habitat Size 0.22 -0.18 -0.21 1.00 Habitat Percent Remaining 0.41 0.20 -0.46 0.16 1.00 Habitat Land Opp. Cost -0.57 -0.13 0.62 -0.43 -0.30 1.00 Habitat Labor Cost -0.28 -0.35 0.51 0.13 -0.03 0.16 1.00 Country Protection -0.78 -0.44 0.87 -0.23 -0.43 0.66 0.41 Effectiveness 5. Three Illustrations To illustrate the possible range of variation, we implement the model for three sets of parameter weights. The first two give extra weight to subspecies threats and project cost elements, while the third assigns equal weight to all variables. The subspecies threat scenario assigns unit weights to first-stage variables except for subspecies numbers (T) and distributions across countries (N), which are assigned weights of 5. In the second stage, we assign a weight of 3 to subspecies threat and unit weights to the other 10 three indices. The cost scenario assigns a weight of 3 to the cost index in the second stage, leaving all other weights at unit values. The equal-weights scenario assigns unit weights to all variables in both stages. Table 3 shows that these weighting changes have very significant consequences for the priority rankings of the 74 tiger habitat areas in the model: The rank correlations of Species Threat with Cost and Equal Weights are .41 and .64 respectively, while the correlation between Equal Weights and Cost is .82. Table 3: Habitat Rank Correlations Species Threat Cost Species Threat 1.00 Cost 0.41 1.00 Equal Weights 0.64 0.82 Table 4 presents results for all 74 habitat areas, sorted by rank in the Species Threat Scenario. Inspection of the top entries indicates that the major beneficiaries of extra weighting for species threat are habitat areas in Indonesian Sumatra, the sole locale of the Sumatran tiger, which rank much higher than in the Cost and Equal Weighting scenarios. This applies particularly to Bukit Balai Rejang-Selatan and Gunug Leuser, which move to first and second in the priority ordering. Subspecies scores for the critically-threatened Indochinese Tiger are also high, which produces high ranks for several Vietnamese, Laotian and Thai habitat areas (e.g., Northern and Southern Annamites, Phu Miang – Phu Thong, Taman Negara-Belum) . The Cost scenario shifts habitat scores toward areas that have low labor costs, low land opportunity cost indices, or both. As a result, the highest priorities are assigned to some habitat areas in Nepal (Corbett-Sonanadi, Royal Bardia, Royal Chitwan, Royal Suklaphanta), Myanmar (Northern Forest Complex - Namdapha - Royal Manas), and India (Dandeli – Anshi, Royal 11 Chitwan, Kanha û Phen, Western Ghats: Bandipur - Khudrenukh û Bhadra, Simlipal, Kaziranga – Garampani). At the same time, cost advantages in Vietnam and Lao PDR maintain high rankings for several areas with high priorities in the Species Threat scenario (particularly the Northern and Southern Annamites). The Equal Weights scenario favors many of the same protected areas in Nepal, Vietnam, Lao PDR and India. The three illustrative cases suggest that the priority rankings of some habitat areas are highly sensitive to variable weighting, while others are not. However, these are only three of many possible scenarios, and the correlations in Table 2 suggest that most habitat areas would change priority ordering substantially for some values of the 8 weighting parameters. We conclude that the principal value of our modeling system is educational rather than prescriptive. Undertaking numerous weighting experiments can provide a useful sense of the relationship between decision-makers’ preferences and habitat assistance priorities. In the same vein, the model can be used to reveal the preferences of decision-makers who have assigned priorities to different habitats in resource allocation. 6. Summary and Conclusions In this paper, we have developed and implemented a model that translates detailed information about 74 tiger habitat areas into consistently-derived priority scores and potential project budget shares for those areas. Drawing on the database constructed by Dasgupta, et al. (2012a), the model incorporates information about threats to particular tiger subspecies, the quality of remaining habitat areas, the observed effectiveness of habitat protection by country, and the potential costs of protection projects for different habitats. Implementation of the model moves through two stages. In the first, user-specified weights are employed to combine subindices into composite indices of species threat, habitat quality, cost and protection 12 effectiveness. In the second stage, user-specified weights are employed to combine the composite indices into priority scores and potential project budget shares for all 74 habitat areas. Our investigation of inter-variable correlations suggests that changes in user-specified weights can have very significant consequences for habitat priority scores. In three illustrative scenarios, we investigate the implications of equal weights for all model variables, higher weights for species threats, and higher weights for potential project costs. We find very substantial differences in high-priority habitats across the three scenarios, although habitats in some countries retain high positions in all three. In summary, we find that great habitat diversity is revealed by the introduction of eight critical variables for priority-setting. No single priority ordering can be prescribed in such a diverse setting, and actual priorities will depend on the preferences of decision-makers, as revealed in the weights assigned to species threats, habitat quality, cost elements, and effective protection. At the same time, we believe that our model can make a useful contribution by identifying priority orderings that are consistent with different sets of preferences. And it can inform policy discussions by allowing for extended exploration of alternative strategies, along with feedback to decision makers about the implicit preferences associated with their resource allocation decisions. 13 Table 4: Results for Three Weighting Scenarios Habitat Ranks 8 Habitat Scores Variable Ranks 9 Habitat Name (Tiger Species Equal Species Equal Subspecies Subspecies Subspecies Habitat Habitat Land Labor Effective Country Landscape) Threat Cost Weights Threat Cost Weights Numbers Distribution Habitat Size Percent Cost Cost Protection Bukit Balai Indonesia Rejang - 1 39 45 100.00 63.31 70.08 4 3 2 35 71 42 3 2 Selatan Indonesia Gunug Leuser 2 54 48 98.44 56.77 68.44 4 3 2 59 61 24 3 2 Northern Vietnam 3 14 2 97.78 82.07 97.43 3 1 3 64 67 32 5 7 Annamites Southern Vietnam 4 15 6 97.15 81.22 95.18 3 1 3 69 50 33 5 7 Annamites Southern Laos 5 8 4 96.27 85.73 96.49 3 1 3 69 58 36 6 6 Annamites Phu Miang - Laos 6 4 7 95.99 93.71 94.77 3 1 3 54 35 66 6 6 Phu Thong Bukit Barisan Indonesia 7 46 52 94.73 59.74 64.62 4 3 2 19 70 41 3 2 Selatan South Indonesia Kerinci Seblat 8 61 53 94.50 50.40 64.37 4 3 2 63 56 15 3 2 Taman Negara - Thailand 9 51 31 94.20 58.28 75.79 2 3 1 67 63 28 2 5 Belum Rimbo Panti- Indonesia Batang Gadis 10 43 54 94.17 60.75 64.04 4 3 2 15 72 47 3 2 West Vietnam Kon Ka Kinh 11 17 14 92.88 79.13 90.20 3 1 3 46 46 35 5 7 Nepal Royal Bardia 12 2 1 92.53 96.63 100.00 1 2 4 48 30 48 8 9 Nam Et Phou Laos 13 25 33 92.24 75.54 75.12 3 1 3 56 60 23 6 6 Loey Indonesia Sibologa 14 55 56 91.46 56.69 61.29 4 3 2 14 69 37 3 2 Thailand Khlong Saeng 15 52 39 91.24 57.41 72.25 2 3 1 40 65 31 2 5 Corbett - Nepal 16 1 3 90.81 100.00 97.22 1 2 4 44 19 66 8 9 Sonanadi Laos Nam Ha 17 19 19 90.47 77.00 86.71 3 1 3 33 64 29 6 6 India Royal Chitwan 18 9 12 90.00 85.12 90.79 1 2 4 36 34 66 4 8 India Western Ghats: 19 11 8 89.03 83.40 94.39 1 2 4 58 38 50 4 8 8 All three ranks in descending order from 1 to 74 9 See Section 4 for an explanation of rank calculations for these variables. 14 Bandipur - Khudrenukh û Bhadra Kaziranga - India 20 13 9 88.49 82.62 93.53 1 2 4 51 41 49 4 8 Garampani India Dandeli - Anshi 21 7 10 88.40 86.06 93.38 1 2 4 22 73 63 4 8 Northern Laos 22 35 23 88.16 65.65 83.41 3 1 3 64 54 13 6 6 Annamites Northern Forest Complex - India 23 37 11 87.93 64.09 92.64 1 2 4 73 20 20 4 8 Namdapha - Royal Manas Bukit Rimbang Indonesia 24 67 58 87.88 44.51 57.74 4 3 2 38 53 11 3 2 Baling Bukit Barisan Indonesia 25 66 60 87.68 46.23 57.54 4 3 2 28 55 14 3 2 South Nepal Royal Chitwan 26 3 5 86.76 94.94 95.93 1 2 4 36 27 51 8 9 Periyar - India 27 16 13 86.45 79.57 90.31 1 2 4 43 40 45 4 8 Megamala India Kanha û Phen 28 10 15 86.07 84.45 89.72 1 2 4 53 21 66 4 8 India Simlipal 29 12 17 85.25 82.77 88.44 1 2 4 24 44 62 4 8 India Indravati 30 22 18 84.54 75.98 87.33 1 2 4 66 23 39 4 8 India Royal Bardia 31 24 20 83.86 75.69 86.29 1 2 4 48 28 40 4 8 Laos Hin Nam Ho 32 36 27 83.85 64.64 77.37 3 1 3 27 57 16 6 6 Bukit Indonesia Tigapuluh 33 69 63 82.78 37.89 52.78 4 3 2 50 36 6 3 2 Landscape Vietnam Bi Dup-Nui Ba 34 38 32 82.53 63.51 75.55 3 1 3 16 59 19 5 7 Royal India 35 59 21 82.49 53.89 84.18 1 2 4 11 1 51 4 8 Suklaphanta Nam Et Phou Vietnam 36 41 16 82.21 62.70 89.26 3 1 3 56 17 18 5 7 Loey Anamalai- India 37 29 22 82.13 70.65 83.63 1 2 4 31 45 30 4 8 Parambikulam Indonesia Berbak 38 68 66 81.20 38.14 51.28 4 3 2 25 49 7 3 2 Corbett - India 39 20 24 81.09 76.21 82.04 1 2 4 44 16 51 4 8 Sonanadi Thailand Phu Khieo 40 50 37 80.19 59.45 72.36 3 1 3 42 39 38 2 5 Northern Forest Myanmar 41 6 25 79.71 88.42 79.95 1 2 4 73 68 34 9 4 Complex - 15 Namdapha - Royal Manas India Dandeli North 42 18 26 78.87 77.18 78.70 1 2 4 6 66 65 4 8 Vietnam Cat Tien 43 47 42 78.87 59.57 70.58 3 1 3 34 18 17 5 7 Cambodia Cardamom's 44 56 51 78.67 54.97 67.43 3 1 3 60 52 9 7 3 India Rajaji Minor 45 21 28 77.58 76.11 76.77 1 2 4 10 32 66 4 8 Pachmarhi - India 46 26 29 77.52 72.85 76.68 1 2 4 41 10 51 4 8 Satpura - Bori India Radhanagari 47 27 30 77.45 71.55 76.58 1 2 4 30 15 46 4 8 Phu Miang - Thailand 48 60 49 76.76 51.96 67.77 3 1 3 54 31 22 2 5 Phu Thong Myanmar Tenasserims 49 45 34 76.57 59.98 74.33 1 2 4 71 51 8 9 4 Thailand Cardamom's 50 57 44 76.50 54.40 70.32 3 1 3 60 33 25 2 5 Thailand Tenasserims 51 65 50 75.92 46.85 67.52 3 1 3 71 42 12 2 5 India Pench 52 28 35 75.91 71.34 74.31 1 2 4 29 11 51 4 8 India Satkosia-Gorge 53 44 36 74.81 60.64 72.70 1 2 4 26 25 21 4 8 India Palamau 54 31 38 74.53 70.05 72.30 1 2 4 32 8 51 4 8 Kualar Kampar- Indonesia 55 70 70 74.43 28.37 45.00 4 3 2 52 29 2 3 2 Kerumutan Royal Bardia Nepal 56 23 40 74.00 75.87 71.52 1 2 4 5 22 43 8 9 South India Shendurney 57 42 41 73.63 61.85 70.99 1 2 4 7 62 26 4 8 India Chandoli 58 33 43 73.18 67.10 70.33 1 2 4 17 14 44 4 8 Sunabeda- India 59 32 46 72.67 68.30 69.60 1 2 4 21 9 51 4 8 Udanti Western Ghats - India Sharavathi 60 30 47 72.39 70.47 69.20 1 2 4 2 73 63 4 8 Valley Thailand Khao Yai 61 53 59 68.90 56.83 57.63 3 1 3 20 7 72 2 5 Thailand Salak-Phra 62 62 62 67.58 47.33 55.98 3 1 3 8 37 27 2 5 Bandhavgarh - India 63 40 55 67.06 63.02 61.69 1 2 4 18 4 51 4 8 Panpatha India Purna 64 34 57 66.36 66.06 60.74 1 2 4 9 5 72 4 8 Tesso Nilo Indonesia 65 72 71 65.04 22.09 36.76 4 3 2 23 26 1 3 2 Landscape Cambodian Cambodia 66 63 64 64.64 47.28 52.37 3 1 3 62 6 10 7 3 Northern Plains 16 India Rajaji Major 67 49 61 63.29 59.48 56.57 1 2 4 3 12 51 4 8 Thap Lan - Thailand 68 64 69 60.19 46.87 47.05 3 1 3 39 1 51 2 5 Pang Sida India Biligiri Range 69 48 65 59.84 59.56 52.00 1 2 4 1 13 72 4 8 Mahabaleshwar India Landscape - 70 58 67 57.76 54.28 49.31 1 2 4 4 3 51 4 8 South Royal Nepal 71 5 68 57.34 90.84 48.78 1 2 4 11 24 66 8 9 Suklaphanta Taman Negara - Malaysia 72 71 72 56.42 22.48 35.13 2 3 1 67 47 4 1 1 Belum Malaysia Endau Rompin 73 74 73 53.09 20.16 32.06 2 3 1 47 43 3 1 1 Malaysia Krau 74 73 74 50.23 20.77 29.51 2 3 1 13 48 5 1 1 17 References Behrman, J., R. Pollak, and P. Taubman. 1982. Parental Preferences and Provision for Progeny. Journal of Political Economy, 90, 1, pp. 52 - 73. Bolt, Katherine, Piet Buys, Susmita Dasgupta, Kirk Hamilton, Craig Meisner, Kiran Pandey, Limin Wang and David Wheeler. 2003. Measuring Up: New Directions for Environmental Programs at the World Bank. World Bank Policy Research Working Paper 3097, July. Buys, Piet, Kenneth Chomitz, Susmita Dasgupta, Uwe Deichmann, Bjorn Larsen, Craig Meisner, Jostein Nygard, Kiran Pandey, Nat Pinnoi and David Wheeler. 2004. The Economics of Regional Poverty-Environment Programs: An Application for Lao People’s Democratic Republic. World Bank Policy Research Working Paper 3267, April. Dasgupta, Susmita, Dan Hammer, Robin Kraft and David Wheeler. 2012a. VyÄ?ghranomics in Space and Time: Estimating Habitat Threats for Bengal, Indochinese, Malayan and Sumatran Tigers. World Bank Policy Research Working Paper 6212. Washington, DC. Dasgupta, Susmita, Dan Hammer, Robin Kraft and David Wheeler. 2012b. Forest Habitat Protection in Tropical Tiger Range Countries: Measuring Results Since 2007. Washington, DC: World Bank. Mimeo. Hansen, M., R. DeFries, J.R. Townshend, M. Carroll, C. Dimiceli, and R. Sohlberg. 2006. Vegetation Continuous Fields MOD44B, 2001 Percent Tree Cover, Collection 4, University of Maryland, College Park, Maryland, 2001. Hansen, M., S. Stehman, P. Potapov, T. Loveland, J. Townshend, R. DeFries, W. Pittman, F. Stolle, M. Steininger, M. Carroll and C. Dimiceli. 2008. Humid tropical forest clearing from 2000 to 2005 quantified using multi-temporal and multi-resolution remotely sensed data. PNAS, 105(27), 9439-9444. Harrison, Ann. 2002. Has Globalization Eroded Labor’s Share? Some Cross-Country Evidence. UC Berkeley and NBER. Pandey, Kiran, Ramesh Ramankutty, Piet Buys, Ken Chomitz and David Wheeler. 2005. Biodiversity Conservation Indicators: New Tools for Priority-Setting at the Global Environment Facility. GEF Working Paper. November Wheeler, David. 2011. Quantifying Vulnerability to Climate Change: Implications for Adaptation Assistance. Center for Global Development Working Paper 240. January. 18