Spatial-temporal aspects of cost-benefit analysis for park management: An example from Khao Yai National Park, Thailand

ARTICLE IN PRESS

Journal of Forest Economics 13 (2007) 129–150 www.elsevier.de/jfe

Spatial-temporal aspects of cost-benefit analysis for park management: An example from Khao Yai National Park, Thailand Heidi J. Albersa,, Elizabeth J.Z. Robinsonb a

Department of Forest Resources, Peavy 279, Oregon State University, Corvallis, OR 97330, USA b Centre for the Study of African Economies, University of Oxford, Oxford, UK Received 3 October 2006; accepted 15 February 2007

Abstract Using a model calibrated to Khao Yai National Park in Thailand, this paper highlights the importance of generating explicitly spatial and temporal data for developing management plans for tropical protected forests. Spatial and temporal cost-benefit analysis should account for the interactions between different land uses – such as the benefits of contiguous areas of preserved land and edge effects – and the realities of villagers living near forests who rely on extracted resources. By taking a temporal perspective, this paper provides a rare empirical assessment of the importance of quasi-option values when determining optimal management plans. r 2007 Elsevier GmbH. All rights reserved. JEL classification: Q01; Q23; Q24; Q20 Keywords: Cost-benefit analysis; Spatial-temporal modeling; Tropical forest; Parks; Thailand; Quasi-option value; Protected areas; Non-timber forest products

Corresponding author. Tel.: +541 737 1483.

E-mail address: [email protected] (H.J. Albers). 1104-6899/$ - see front matter r 2007 Elsevier GmbH. All rights reserved. doi:10.1016/j.jfe.2007.02.002

ARTICLE IN PRESS 130

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

Introduction Resource managers in tropical poor countries create and manage forest parks within a complex setting of ecological characteristics, local people’s needs, competing land uses, and divergence between who bears costs and who captures benefits. Spatial challenges arise because positive and negative externalities across adjacent areas of land in different uses may influence both the benefits themselves – such as creating minimum habitat size – and the incidence of costs and benefits – such as whether park neighbors bear significant opportunity costs (Albers, 1996; Ferraro, 2002). Social and institutional challenges arise because of the complex policy and property rights setting in such countries and because park benefits accrue to a wide range of people, from neighboring villagers collecting vegetables to the distant populations that enjoy tropical parks’ contributions to global climate control (Albers and Ferraro, 2006). Resource managers also face intertemporal challenges because the future benefits from these parks are often highly uncertain, and some changes in forest use are irreversible or reversible at a cost (Albers et al., 1996). Yet, although developing countries increasingly dedicate land to parks, few managers take these complex interactions into account in their park siting and management plans due, in part, to restrictive mandates, inflexible plans, and data/information limitations (Repetto, 1988; West and Brechin, 1991; Ghimire, 1994; Albers and Grinspoon, 1997). This paper combines the modeling structure of Albers (1996) – which presents a model for tropical forest management and explores the interactions of spatial and temporal characteristics through stylized parameter values – with information and data from Khao Yai National Park (KYNP), Thailand to explore issues about park management decisions, incentives facing local people, and the need for spatialtemporal data. As is true for all tropical parks, sufficiently detailed spatial-temporal data do not exist to undertake a complete case study. But that problem is part of this paper’s point; we make a case for the importance of generating spatial and temporal data for tropical forest management by grounding the analysis and discussion in an empirically-relevant range of values and parameters. The results from the calibrated simulation model demonstrate how forest managers and other stakeholders would manage the forested area differently depending on which spatial, temporal, and equity considerations they take into account. This paper demonstrates the usefulness of cost-benefit analyses to park management decisions, in addition to providing justification of a park as whole, when interactions between areas of different land use, the distributional impact of land use management zones, and temporal changes in values are accounted for. The paper also explores how cost-benefit analyses that identify the incidence of costs and benefits, rather than simply estimating a forest’s total value, better equip park managers to address spatial issues and improve conflict resolution. The following section provides background information about the spatial, temporal, institutional, and social aspects of tropical forest park management. The next section describes and solves a spatial-temporal optimization model for park management using KYNP as an example. Much of this section discusses dividing the

ARTICLE IN PRESS H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

131

park into zones, the land use options, the value of those land uses per zone, and the values generated by spatial patterns of land uses. The penultimate section solves the optimization model with different sets of parameter values to discuss differences in management decisions across regulatory settings and across various stakeholder groups. The final section discusses the implications of the findings for the role of cost-benefit analysis in establishing and zoning parks, determining the permitted activities within parks, and highlighting areas of potential conflict. In addition, the final section comments on the importance of rural people’s welfare in forest management, the role of the international community in tropical conservation, and the impact of quasi-option values on management decisions.

Spatial, temporal, institutional, and social aspects of tropical forest park management Even when a forest is managed as a homogenous unit, the ecological and humaninduced characteristics of land often vary across the forest. For example, ecological edge effects – such as changes toward light-tolerant species and drying of the ecosystem – and human-induced edge effects – such as degraded trees due to lopping and changes in species composition due to enrichment with economic species – occur at the boundaries of forests. Similarly, some areas of a forest may provide higher recreation benefits than others because of scenic features or wildlife viewing stations. Further, in general, tropical ecology suggests that biodiversity benefits increase with contiguous area. Large areas of contiguous forest land provide habitat for large animals such as elephants and guar and attract more visitors for trekking, wildlifewatching, and wilderness opportunities (MacArthur and Wilson, 1967; Soule´, 1987). These effects create spatial variation in the amount and types of benefits provided by subunits or zones of the forest. Managing a protected forest in distinct zones (heterogeneous management) provides opportunities to create additional value by, for example, locating tourist facilities adjacent to the park. Conversely, a park can negatively affect an abutting village if wildlife from the park damages crops (Studsrod and Wegge, 1995; West and Brechin, 1991; Kiss, 1990; Schultz, 1986). Alternatively, a developed area located next to a forest park can degrade the forest resource through pollution and edge effects. Homogeneous management approaches and valuation exercises that aggregate values to determine a value for the forest park as a whole cannot explore and manipulate these externalities. Efficient tropical forest management requires an explicitly spatial analysis that takes into account the interactions between different land uses when determining the flow of benefits (Albers, 1996). In addition to spatial considerations, a planner should consider how land use values may change over time. Most valuation studies consider some temporal aspects of benefit flows but uncertainty often remains about future values for many benefits. For example, if a species becomes endangered over time, the value of

ARTICLE IN PRESS 132

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

protecting a population in a particular park could increase. The reversibility of land use allocation decisions also plays a potentially important role in the decision framework and hence optimal land use patterns (Arrow and Fisher, 1974; Pindyck, 1991). Some activities can destroy the benefits of tropical forests irreversibly – such as large-scale ‘‘permanent’’ agriculture or habitat restriction to the point of species extinction. Yet many forests can recover over time after less disruptive practices – such as small-scale shifting agriculture with long fallow periods (de Rouw, 1995). The benefits of tropical forest parks mentioned above can be categorized more generally as direct and indirect use, non-use (also known as existence value), option, and bequest values while the costs of parks fall into the categories of direct and opportunity costs (Le´le´ et al., 2001; Pearce, 2001). Studies that address the willingness to pay by tourists to visit a particular site, such as that undertaken by Naidoo and Adamowicz (2005), provide important information on how park management could optimize gate revenues. As economic valuation techniques have advanced, so has the ability to measure non-market benefits of tropical parks and forests, thereby increasing recognition of their importance (Emerton, 2003). Yet many cost-benefit analyses of parks concentrate on just one particular benefit from a protected area (Emerton, 2003). Even those studies that assess multiple values from a tropical forest typically seek to define a value for the forest as a whole, aggregating across time, area, and beneficiaries (Dixon and Sherman, 1990). Such valuation studies support park establishment but do not inform decisions about management zones, management activities, or park size. Choice modeling represents a fruitful direction for developing benefits assessments of more use to park management decisions (Hanley et al., 2001). Cost-benefit exercises that describe the incidence of these costs and benefits provide important information to managers. For example, particularly in a tropical setting, people at a distance from the park often benefit but bear few costs. In contrast, nearby people bear a potentially large opportunity cost from park establishment through the loss of land and access to non-timber forest products, NTFPs (Ferraro, 2002). Moreover, excluding nearby villagers from the park typically requires costly enforcement and, without compensation to villagers, tension between locals and managers places the protected area at risk (Pearce, 2001). Identifying the costs and benefits for stakeholder groups provides information about areas in which property rights enforcement is likely to be costly which in turn can contribute to cost-benefit analysis. Similarly, other stakeholders emphasize different costs and benefits from a park. For example, international conservation organizations may consider biodiversity existence values and carbon storage values that accrue to the international community but disregard local tourism benefits. Government environment departments may consider the total benefits and costs that accrue to that country’s citizenry, but not their incidence. Park managers often have a mandate to manage without regard to who bears the costs and who captures the benefits of forest conservation. Local communities and development organizations are unlikely to consider non-local benefits that they do not capture. In contrast, a social planner

ARTICLE IN PRESS H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

133

accounts for use and non-use values and the incidence of costs and benefits across space and time. In the following section, we apply a spatial-temporal model to the example of KYNP to consider how such a social planner would manage the park and to demonstrate the role of regulations and the incidence of costs and benefits in decisions.

Example: Khao Yai National Park KYNP in central Thailand provides an opportunity to demonstrate the empirical importance of spatial, intertemporal, and social constraints on tropical forest management. KYNP has been the subject of much research, including both ecological studies and economic valuation studies, which has generated more information than is available for most similar tropical forests. KYNP also presents clear management challenges. The park is not an island of pristine forest surrounded by cropland; rather, it has substantial tourist facilities both at its center and at its entrances, a de facto buffer zone from which villagers extract NTFPs, and one portion abutting a mountainous wilderness (Khao Yai Ecosystem Project Final Report, 1982; Albers and Grinspoon, 1997). The Royal Forestry Department (RFD) has managed KYNP since it became Thailand’s first national park in 1962. RFD aims to preserve the diversity of flora and fauna that led to the park’s selection as both an ASEAN Heritage Park and Reserve and part of a World Heritage Site, which continues to attract domestic and international tourists, especially birdwatchers. Both the Thai national regulations and the IUCN classification of National Park prohibit any other uses of the forest park, such as extractive collection of NTFPs or selective logging. Consequently, RFD manages KYNP’s 2200 km2 as a preservation area, with tourism and research as the only permitted land use activities. It views the park as one cohesive management unit and focuses on offering tourist information at park headquarters and attempting to enforce the preservation mandate at the park boundaries. Yet in reality, a number of relatively distinct zones can be identified (Fig. 1) from KYNP management documents and from interviews with park managers, development organizations, and villagers (Khao Yai Ecosystem Project Final Report, 1982; Albers and Grinspoon, 1997). These zones did not result from conscious management decisions. Rather, they are the consequence of the underlying ecology, settlement patterns, and incomplete protection of the park boundaries. The five zones comprise (1) an encroached and degraded outer ring at the periphery of the park; (2) an inner ring that is less degraded but still used by villagers for (illegal) NTFP extraction; (3) an inner core that offers a relatively undisturbed habitat with easily accessible trails for tourists, including birdwatchers and hikers; (4) a distinct, largely undisturbed, mountainous area that lies outside but adjacent to the park’s borders to the northwest; and (5) a zone that includes a highway bisecting the park, the park headquarters, tourist accommodations, restaurants, managed grassland, and park employees’ lodging.

ARTICLE IN PRESS 134

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

Fig. 1. KYNP stylized zones.

The model The spatial-temporal model of Albers (1996), adapted to reflect KYNP’s de facto zones, forms the basis of an economic management model. This framework, paired with the calibration below, permits an exploration of the empirical importance of different spatial, temporal, and social aspects of tropical forest management. Here, actual land uses identified in the region define four park management land uses that can be adopted in each zone: preservation (P), which permits recreation, research, and sustainable levels of NTFP extraction; development (D), which includes roads, restaurants, forest plantations, and resorts; temporary agriculture (A), which reflects local shifting cultivation and NTFP collection; and recovery (R), which includes only forest regeneration and no NTFP extraction on land that gradually recovers following temporary agriculture, A. These park management land uses incorporate implicit temporal characteristics of the region: conversion to D is irreversible and conversion to A is reversible only after a period in R. With these dynamics, the initial land uses in the zones of KYNP (Fig. 1) corresponds to zone 1 in temporary agriculture, A; zones 2, 3, and 4 in preservation, P; and zone 5 in development, D. A spatial framework permits an analysis of the management of the park as one single unit versus multiple units and accommodates externalities associated with different land uses. This framework captures two key spatial characteristics of the benefits from land in and around the park. First, to characterize the added benefits of creating large habitats, the model adds a value, named P-annex, to the total benefits for every instance of adjacency between two zones under preservation, P. Second, to include the increase in the value of resort development due to proximity to preserved land,

ARTICLE IN PRESS H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

135

the model adds a site externality value, S, for every instance of a border between zones that are allocated to development, D, and preservation, P.1 Three different perspectives for the forest manager are used as a basis for analysis: ‘‘full-spatial’’ manager; ‘‘independent zones’’ manager; and ‘‘block’’ manager. A ‘‘full-spatial’’ manager recognizes different zones within the forest and the interactions among those zones. An ‘‘independent zones’’ manager views the park as a set of unlinked zones and does not incorporate externalities (S and P-annex) among zones in his decisions. A ‘‘block’’ manager internalizes the externality values but views the park as one unit and assigns only one of the four land uses above to all zones. The differences between the land use allocations across these stylized manager types demonstrate the role of spatial modeling in determining optimal land management patterns. Although the Thai RFD’s management of KYNP does not fall strictly into any one of these three management perspectives, its mandate is to preserve the entire area, similar to the block manager. Solving the economic management model developed here determines the optimal land use per zone from the perspective of each of the 3 managers using the starting point of de facto land use in each zone (Fig. 1). Calibration The dataset used to calibrate the model brings together information from several studies, primarily ecological information from the Khao Yai Ecosystem Project Final Report (1982), planning ideas from the KYNP Management Plan (1987), and economic valuation from Dobias et al. (1988) and Dixon and Sherman (1990). To fit the spatial aspects of the model, the data are divided into per-zone land use values (P, D, A, and R) and cross-zone interaction values (P-annex and S) to form the relevant dataset. Because the primary data were largely collected to determine a value for the park as a whole and not for management purposes, the dataset developed in this paper must, by necessity, contain assumptions about uncertainty, per-zone values, and spatial values. However, the key contribution of this example – demonstrating how optimal management patterns change depending on whether spatial aspects are accommodated – remains (see Albers, 1992, 2001 for more detailed sensitivity analysis of the assumptions for this example). The values from the preservation land use, P, sum benefits from three categories: hydrology and erosion control benefits; recreation values; and sustainable NTFP extraction (Albers, 1992). The park protects watersheds and provides water for irrigation to the surrounding agricultural lands; that ecosystem service is valued at 31.7 million Baht, or approximately US$1.27 million, per year (Dobias et al., 1988).2 Dobias et al. (1988) use a nutrient-loss cost-of-replacement method to value the 1

To represent the potential for crop loss on agricultural land bordering preserved land, the model could accommodate a crop-loss cost when a zone under temporary agriculture shares at least one border with a zone under preservation. However, because KYNP interviews with farmers revealed little concern for this problem, this spatial interaction is not included. 2 All values in this paper are reported in 1990 million Baht increments based on the timing of data collection. In 1990, 1 Baht ¼ US$0.04.

ARTICLE IN PRESS 136

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

park’s soil erosion control at 75.3 mB/year. Khao Yai Ecosystem Project Final Report (1982) discussion of the location of rivers and watersheds supports the assumption that the park-land generates these hydrology and erosion control benefits across the entire area. We divide the benefits across zones by weighting the total benefits by each zone’s fraction of the total area of the park. The park generates recreation benefits and consumer surplus of approximately 37.6 mB/year, which includes 2.6 mB/year from trekking and camping opportunities (Brockelman and Dearden, 1990; Dixon and Sherman, 1990; Dobias et al., 1988). Zones 2 and 3 contain all of KYNP’s hiking trails, campsites, and wildlife viewing stations. Because of its small size and awkward shape, zone 2 alone is assumed to generate 25% of its area-proportional value. The remaining 75% is included in the P-annex term – that is, the additional value of having both zones 2 and 3 preserved and therefore providing contiguous habitat and recreation area. Trekking originates in local villages and leads deep into the park. The contribution to preservation values comesboth from zone 4 alone and the P-annex value for preserving zones 2 and 4 contiguously. Finally, if a zone is in recovery, R, it generates a fraction of the preservation value, with that fraction equal to the average biomass of land in recovery divided by the average biomass in preservation. R also creates a P-annex value at this biomass-proportional rate (R-annex). During the 6-month non-growing season, KYNP generates extractive values of approximately 100 B/day per household from the vegetables, medicines, herbs, and meats extracted by the 3000 local households, which totals to 54 mB/year (Albers, 1992). The value of this extraction, included in the park’s preservation values despite its illegality, is divided into per-zone extractive values that reflect actual use in zone 2, estimated use without enforcement of the ban on collection for zone 3 (in interviews, villagers said they extracted little in zone 3 due to the large presence of guards there, but provided estimates of their extraction levels if there were no enforcement), and values proportional to area for the other zones (Albers, 1992; Dixon and Sherman, 1990). The benefits from the zones in the development land use (D) come from a variety of large, ecosystem-disrupting uses, such as permanent agriculture, eucalyptus plantations, and resort construction. The values used in the model contain assumptions about the amount of land converted to agriculture or eucalyptus in each zone and range from 25% to 40% and from zero to 20% respectively. The eucalyptus plantation values reflect the social profitability of the medium-scale eucalyptus planter (Tongpan et al., 1990). Development in or near KYNP receives a positive externality (S) from proximity to preserved areas if development includes resort ownership, thereby reflecting the benefits of access to the park for recreation and for local climate control. The profitability of a resort in the area is approximated by examining the profits generated by Thailand’s Association of Tourism at its golf course, restaurants, and lodging within zone 5 of KYNP, correcting for capital outlays and competition (Albers, 1992; Dobias et al., 1988). To accommodate the intertemporal analysis, the preservation values include an assumption about uncertainty over future values – preservation values are higher if elephants survive in KYNP, lower if they do not (Dobias, 1985). The values

ARTICLE IN PRESS H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

137

employed reflect the benefits associated with the preservation of a viable population of elephants in Thailand weighted by the fraction of those benefits that KYNP provides (Albers, 1992; Dixon and Sherman, 1990).3 That contribution to preservation value, 122.5 mB/year, added to the other preservation values, forms the high end of the range of preservation values used in the model, corresponding to the case of elephant survival. The probability assigned to the high and low states is 0.5 for lack of specific information about the probability of elephant survival and about trends in the value of elephant preservation over time. For each zone, a highstate and a low-state preservation value is given although there is no difference between the high and the low for zone 1, which does not generate elephant habitat benefits. Temporary agriculture, (A), includes both shifting cultivation and NTFP extraction. At any point in time, some plots within the zone are in agriculture, some are natural forest generating extractive NTFP values, and some are fallow. The value of such a zone therefore depends on the amount of land in each of those uses at a particular moment. In our model, within a 5-year period, the returns to agriculture fall each year and, because the level of NTFP extraction is not sustainable, these benefits also fall each year (TDRI, 1986; Khao Yai Ecosystem Project Final Report, 1982; Albers, 1992). The total value of temporary agriculture therefore changes as a function of the number of years a given zone has been in that use. Based on pre-park patterns of farming, the model assumes that 10% of the zone is in agriculture and the benefits include the value of farming rice and corn for two years before fallowing that plot (Khao Yai Ecosystem Project Final Report, 1982). By the third consecutive 5-year period of this temporary agriculture, agricultural values at the onset of the period fall off by 10% to reflect the lack of available high quality land. Similarly, unsustainable extraction forces extraction values to decline by 50% in the second 5-year period and 25% in the third 5-year period. A C++ computer program solves the dynamic optimization problem to find the land use for each zone that each of the three manager types chooses for the first 5-year period (of three considered in the optimization), subject to the constraints described above. Table 1 summarizes the estimated per-zone and per-period values of the different land uses, including the costs and benefits that one zone imposes on another, with a 10% discount rate. Discussion of results The solved model results in a set of first-period optimal land use management plans from the perspectives of each of the three managers (full-spatial, independent zones, and block), reflecting the initial conditions that exist in and around KYNP. These plans are presented schematically in Fig. 2 to allow for a visual comparison of management plans. 3

The future preservation values assumed for the model reflect only a small portion of the possible range of values, simply because information about the probability of different events is extremely limited.

138

Table 1 Approx area (km2)

Time period

P value (H ¼ high, L ¼ low)

A values

ia

R value

ii

(a) Values of zones in Khao Yai National Park under different land uses in 3 time periods (million Baht, mB) 1 (A) 92 1 16.7 29.0 2 10.4 18.0 14.5 3 6.4 11.2 9.0 975

1 2 3

3 (P)

1043

1 2 3

4 (P)

21

455.0 404.0 282.5 250.9 175.4 155.8

(H) (L) (H) (L) (H) (L)

305.8

424.4 288.4 263.6 179.0 163.7 111.2

(H) (L) (H) (L) (H) (L)

267.0

iii

1.2 0.8 5.7

423.9

189.9

153.5

117.9

95.3

181.6 66.0

145.6

102.9

90.4

64.8

8.1 5.0 3.1

11.2 6.9 4.3

4.1 2.5

1.4

n.a. n.a. n.a.

n.a. n.a. n.a.

1 2 3

n.a. n.a. n.a.

n.a. n.a. n.a.

Interactions between zones

Time period

P-annex

R-annex

S

197.7

129.6

1 2 3

Not avail.

224.5

353.5

165.8

5 (D)

24.9 23.1 2.7

310.5 156.2

3.6

9.4 3.6 3.6

n.a. n.a. n.a.

22.1 20.4 3.9

ARTICLE IN PRESS

2 (P)

D value H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

Zone (and current land use)

2 and 3

1 2

2 and 4

1 2 3

3 and 5

1 2 3

46.2

6.1 1.1

28.7

b

17.8

b

b

0.6 0.0

b

b

b

b

b

b

b

b

b

73.2 45.5 28.2

39.9 0.09 24.8 0.06 15.4 0.03

(H) (L) (H) (L) (H) (L)

n.a. ¼ not applicable because of starting point in (irreversible) development. *All values are in 1990 million Baht increments (1 Baht ¼ US$0.04). All values are the sum of 5 years (one period) of benefit flow. All sums are in present value terms, calculated with a discount rate of 10%. Where only one value is entered, the same value is used for the high and low events. a ‘‘i’’ is the value of the A option during the first period of its use, ‘‘ii’’ is that value in the second period of use, and ‘‘iii’’ is that value during the third period of use. b Value not applicable due to zone’s initial use or that ‘‘R’’ is not used in period 3.

ARTICLE IN PRESS

3

b

414.3 (H) 74.2 (L) 257.3 (H) 46.1 (L) 159.8 (H) 28.6 (L)

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

(b) Value of spatial interactions between land uses in adjacent zones (million Baht, mB) 1 and 2 1 26.4 (H) 2.3 111.1 0.0 (L) 0.0 2 16.4 1.0 69.0 0.0 (L) 0.0 b 3 10.2 (H) 42.8 0.0 (L)

139

ARTICLE IN PRESS 140

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

4 2

3

5

3

2 1

Block Manager

P – preservation

Full-spatial Manager

R – recovery

D – development

Indp. zones Manager

A – temporary agriculture

Fig. 2. First period optimal land use allocations for different managers.

The three managers adopt very different management plans for the park. The full spatial manager converts zone 1 to development and retains the rest of the park under preservation. In contrast, the block manager assigns the entire park to preservation land use P, which means that zone 1 is in recovery, R, in the first period. However, this management plan generates only 84% of the value of the full-spatial manager in the first period (Table 2). The independent zones manager converts zones 1 and 4 to temporary agriculture, A, preserving zones 2 and 3 and generating just 66% of the value of the full-spatial manager in the first period. The full-spatial manager preserves zone 4 because the added value of large contiguous blocks of preserved areas 2 and 4 (P-annex – see Table 1b) dominates the alternative of temporary agriculture, A. Second, the full-spatial manager develops the outer zone 1, whereas the independent zones manager maintains the zone in temporary agriculture, A. The large positive externality of preserved land, S, in zone 2 on resort development in zone 1 dominates the decision on zone 1 for the full-spatial manager. Relative to the independent zones manager, the actions of the full-spatial manager favor tourism at the expense of local communities but do take account of positive and negative externalities across zones.

How would different values and stakeholders change KYNP management? This section looks in more detail at how different stakeholders, regulations, and costs change the optimal land use allocations across KYNP zones. National park managers, for example, must comply with certain regulations; local people may not capture enough preservation benefit to merit preserving the area; and the international community may capture different preservation benefits. In what follows, we vary the values from those in Table 1 to reflect regulations and the incidence of costs and benefits.

ARTICLE IN PRESS H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

141

Table 2. Patterns and expected value generated by different managers in the first period Manager type

Number of zones in P or R in period 1

Value of optimal land use allocations generated in first period

Percent of full-spatial manager’s value (using Benchmark values)

Full-Spatial Manager Block Manager Independent zones Manager All-Preservation Constrained Manager, NTFP extraction permitted Full-Spatial Manager, no NTFP extraction Block Manager, no NTFP extraction permitted Independent Zones Manager, no NTFP extraction permitted All-Preservation Constrained Manager, no NTFP extraction permitted

3 4 2

1289.9 1084.3 848.2

100 84.1 65.8

4

1084.3

84.1

3

919.2

71.2

0

833.8

64.6

0

839.7

65.0

4

786.63

61.0

All-preservation regulation, NTFP extraction permitted National park regulations constrain the Thai RFD to manage KYNP as a single unit – as for most park managers across the globe – despite increasing pressure for buffer-zone management (Ghimire, 1994; MacKinnon et al., 1986). This management plan is equivalent to the Block manager’s first period plan and results in an overall value for the park over the 15 years of the model that is 6% less than the value generated by the full-spatial manager and almost 16% less in the first period (Table 2). Such a homogeneous approach to managing the park excludes the possibility of explicitly developing the outer zone as a buffer area for local villagers or as a tourism development area. Park regulations prohibit NTFP extraction In the benchmark cases discussed throughout this paper, the preservation values include a sustainable level of NTFP extraction. Although park managers state that they do not strongly object to such extraction when it is sustainable and for home use, such extraction is clearly against park regulations and the definition of a ‘‘national park.’’ If these illegal extraction values are removed from the preservation

ARTICLE IN PRESS 142

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

values considered, in the first period the full-spatial manager only recognizes 71% of the park value but his choice of land use allocation remains the same – primarily P with resort development in the outer zone 1. In a more striking case, without the illegal NTFP extraction values the independent zones manager changes the land use in the core of the park (zones 2 and 3) to development with plantations (not resorts) and puts zones 1 and 4 into temporary agriculture, thereby preserving none of the park (Table 2). Without taking into account the illegal values, the block manager converts the entire forest to development, D, with plantations – although the value is quite close to that from preserving the area. If the managers are given a choice between a land use P that includes NTFP extraction and one that does not, they all choose the less restrictive P that permits such extraction. In this example, recognizing the spatial externalities generated by the preservation use dramatically alters the landscape if illegal – but sustainable – extraction values are not considered. All-preservation regulation and no NTFP extraction Combining these sets of rules – no zones and no NTFP extraction – with the inability to convert away from P in later periods (which the block manager can do) mimics management of KYNP according to national park regulations. This allPreservation manager with no NTFP extraction – no zones, no NTFP extraction, and no option to change to a different land use later – generates 24% less value than the full-spatial manager’s plan over 15 years and 39% less than the full-spatial manager in the first period (Table 2). Because valuation exercises often provide only a lower bound on preservation values, unmeasured preservation values might offset the differences between managing the forest as a homogeneous block (over space and time) and managing according to the full-spatial manager’s plan. Also, the less restrictive P values reflect a sustainable level of extraction or hunting but that level could negatively affect non-extractive portions of the P values. Information that assesses the impact of levels of extraction on the quality and magnitude of other preservation values would be useful in determining the appropriate level of restrictions on extraction. Similarly, demand analysis that links regulations and levels of extraction to tourism demand – such as whether international tourists will visit forests that are not designated national parks – would also be useful in these decisions. ‘‘Landscape’’ manager Park managers such as those at KYNP typically face a mandate to generate preservation values and therefore do not consider agricultural and regional development benefits that accrue to neighboring areas. The recent policy literature’s discussion of the impact of parks on rural people and of parks as part of a broader landscape raise questions about how park management might differ if managers considered non-preservation benefits (McNeely et al., 2005; Muller and Albers, 2004) In our model, such a ‘‘landscape manager’’ – whose objectives include

ARTICLE IN PRESS H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

143

rural welfare – allocates zone 1 to temporary agriculture A (its current de facto use). Because this land allocation generates benefits for local people, it potentially also creates goodwill between the park and local people and increases the social value of the land. In addition, the landscape manager chooses a less restrictive definition of preservation that permits limited extraction of NTFPs, though not agriculture, within zones allocated to preservation, P. Indeed, this perspective is consistent with Stræde and Treue (2006) who propose that, given the economic importance of resources extracted from Nepal’s Royal Chitwan National Park by nearby villagers, the reality of this extraction should be internalized into park management. Similarly, and in correspondence with the full-spatial manager, a KYNP manager who is unconstrained by national park regulations and who considers broader regional development rather than benefits to the nearby rural poor might encourage resort development at the park boundaries to boost the local economy and allow a wider group of people – though not local villagers excluded from parklands – to ‘‘capture’’ the benefits of proximity to the park. Although these simulations force a choice between resorts in the development use, D, and temporary agriculture, A, in reality some land uses that combine agriculture and resorts could capture both types of values. However, with current regulations against buffer-zone management, KYNP managers spend much of their budget on costly but relatively ineffective boundary enforcement patrols that have been unsuccessful in eliminating extraction and agricultural encroachment. Management with enforcement costs Although the results above consider costs of each land use, enforcing a property right for a land use such as ‘‘preservation’’ has proven quite costly in most protected areas in poor countries and in KYNP in particular. In addition, the choice of which areas to protect has often relied on assessments of benefits without regard to cost despite ample evidence that considering costs can lead to greater net benefits (Ando et al., 1998; Ferraro, 2003). To examine the impact of this oft-underestimated cost on management decisions, we examine the change in the land use patterns for the full, independent zone, and block managers when they face some enforcement costs if they choose the P land use. Enforcement spending in this calibration is based on the average spending for Thai parks, approximately 8.7 Baht per rai or US$2.20 per hectare, because enforcement spending was not available for the individual park (MIDAS Agronomic, 1991).4 If managers spent this amount per rai in KYNP, the total expenditure would be approximately 12 mB per year, or over 30% of tourist gate receipts. Even with this cost, the block manager receives a high enough net benefit from preservation to merit preserving (P and R) the entire area. Similarly, the spatial values from P-annex continue to dominate the full-spatial manager’s decisions and that manager’s plan remains the same as in the benchmark case above. 4

The Thai unit of measurement, the rai, is equal to approximately 1600 km2.

ARTICLE IN PRESS 144

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

The independent zones manager, however, does not preserve as much of the area when the enforcement costs reduce the net benefits of P. This manager continues to dedicate zones 1 and 4 to A and to maintain zone 2 in P but now converts zone 3 to the development use, D. For some managers, then, the reality of enforcement costs can reduce the benefits to preservation such that other land uses dominate. Although economic theory shows that costly enforcement implies that some illegal activity occurs at the optimum, the current enforcement in KYNP leads to enough extraction and encroachment to threaten some species (Milliman, 1986; Shavell, 1993).

Local villagers as park manager The local villagers’ illegal extraction of products and agricultural encroachment into the park raise the question: How would villagers manage KYNP land? Answering this question helps identify areas of conflict between these resourcedependent people and the park managers. Solving the optimization model with a selective set of values – only those that accrue to villagers – demonstrates that an optimal village management plan places zones 1 and 4 in temporary agriculture, A, rather than preserved/recovery, P/R. Maintaining zone 1 in temporary agriculture rather than developing the land reflects the inability of local villagers to capture the S externality benefits that preservation creates for the resorts, D. Similarly, the conversion of zone 4 to temporary agriculture reflects villagers’ inability to capture the value of large habitats, P-annex. Most strikingly, however, the villagers receive large enough benefits from the preservation of zones 2 and 3 – sustainable extraction, humidity control, and hydrologic and erosion control benefits – that they preserve these zones. On the surface, this analysis suggests that preservation of zones 2 and 3 is possible without government enforcement if local people are given access to a sustainable level of extractive goods; enforcement and incentives for preservation are needed only in zone 4. However, to avoid an inefficient open access situation some mechanism would be required to ensure that all villagers internalize the park externalities. For example, because the hydrology and erosion benefits are externalities created by the forest, without some enforcement or incentive mechanism an individual would choose to convert land to agriculture, disregarding the impact of the lost benefits on other local people. Local people may also exceed the sustainable level of extraction under free access to the area without fear of government reprisal. The villagers’ social optimum therefore requires social resource-access rationing or community resource management. Nevertheless, this analysis demonstrates that the social goals of local people and park managers are not dramatically different – a finding that should help prioritize enforcement expenditures and encourage communication between park managers and park neighbors. In addition, policies that enable local people to capture benefits from resorts, such as jobs or trekking businesses, can further pull local people’s incentives into line with those of the fullspatial manager, if the social/community welfare is not overwhelmed by individual, self-serving actions.

ARTICLE IN PRESS H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

145

International community as park manager The global benefits from preserved land suggest another question: How would the international community manage this land? Because the baseline data used in the model do not include the carbon storage and biodiversity values of importance to an ‘‘international manager,’’ some approximations must be made. If, for example, the international manager valued preservation in the high state by 50% more than the regional benchmark but faced more uncertainty that reduced the probability of a high state to 25%, his management patterns would be identical to those of the fullspatial manager. Indeed, for both the spatial manager and the international manager, only P values of 10 times the values used here induce recovery R and later P in the outer ring, zone 1. Hence, in this example, the international community need not financially support preservation in KYNP other than to continue to send international tourists to the park. This case is just one example of possible benefit flows captured by the international community but it makes a general point: the international manager’s perspective and compensation will prove important only where the local value of non-preservation land uses outweighs the local government’s value of preservation. As long as the government preserves the area, the international community can ‘‘free ride’’ on that preservation, capturing benefits without incurring costs. Only when the domestic government does not capture enough benefits to provide preservation will international support or compensation alter the amount of preserved tropical forest. In addition, questions remain about whether international conservation funds to support KYNP crowds-out local/domestic spending on preservation or will enable higher level of preservation from increased spending. Longer planning horizon and quasi-option value As described above, a key characteristic of the development land use is the irreversibility of that conversion – the land cannot function as part of an ecosystem ever again. As in many models of irreversible decisions under uncertainty, a KYNP manager has an incentive to remain flexible by postponing irreversible choices – here, the development of preserved or temporary agriculture land for alternative uses – until the value of the choices in the future is clarified (Arrow and Fisher, 1974). To accommodate this flexibility, a closed loop rather than open-loop solution to the model is determined, which allows a ‘‘flexible manager’’ to update his decisions each period based on information available at that time (see Albers, 2001 for details of the model). The difference between the closed loop solution’s expected value, and the open-loop solution’s expected values is a non-negative value called ‘‘quasi-option value’’ or QOV (Albers et al., 1996; Fisher and Hanemann, 1986a, b; Arrow and Fisher, 1974). This quasi-option value reflects the value of future information if the manager is in a flexible enough position to take advantage of the new information. For the 15 year timeframe of the model, the quasi-option value does not influence the optimal management patterns: the quasi-option value is simply not large enough to induce more P or R. However, solving the model for a 30-year time frame (that is,

ARTICLE IN PRESS 146

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

each period represents 10 years rather than 5) generates a quasi-option value that does influence the optimal land use patterns in the first period. A full-spatial manager who includes the QOV (by solving the close-loop model) preserves less land in the first period of a 30-year plan than in the 15-year case described above. He converts zone 4 to the reversible temporary agriculture land use A to take advantage of the up-front benefits while maintaining the ability to revert to P after a period of R, should high preservation values in period 3 arise. In contrast, the 30-year independent zones manager who does not consider the QOV assigns the heart of KYNP to development, naı¨ vely converting more than half of the park to plantations because he does not value the interactions across zones nor flexibility across time.

Conclusion The application of a multi-zone, intertemporal economic decision model to KYNP, Thailand demonstrates the empirical importance of taking into account intertemporal, spatial, social, and ecological factors in tropical forest park siting and management. The spatial relationships in particular – habitat size, recreational benefits as a function of size, and a positive externality of proximity to preserved land on resorts – appear especially important in determining optimal land management patterns.5 Yet, sufficiently detailed spatial data are not, in practice, available for any tropical forest. Although most cost-benefit analysis can support the establishment of parks, the collection of more detailed spatial and temporal data would inform three critical areas of protected area establishment and management: the types of uses permitted in the protected area and thus the IUCN designation the area receives; the zoning of protected areas to reflect the varying realities and benefits across the area; and the identification of areas of potential conflict through the incidence of costs and benefits of the protected area. During park establishment, the information about values that can be derived from different types of access – such as trekking, day-trips, and NTFP extraction – could be used to determine what access restrictions, and therefore what IUCN classification, the park should achieve. For example, if tourism values are estimated to generate a significant fraction of the forest’s potential value, an IUCN classification of ‘‘nature reserve’’, which allows no tourism, might not be appropriate or justified. Similarly, if NTFP extraction helps to justify the existence of a protected area, the IUCN classification of ‘‘national park’’, which allows no NTFP extraction, might be overly restrictive, particularly given that the costs associated with preventing such extraction are typically underestimated in cost-benefit analysis. For example, KYNP was designated a national park without zones and without permission for sustainable extraction by local villagers. Yet, as shown above, without the value from extraction included as a preservation benefit in cost-benefit analysis, preservation of the area is more difficult to justify. 5 Sensitivity analysis, not shown in this paper, suggests that negative externalities due to crop damage from proximity to wildlife on preserved land could play a part in determining the optimal spatial land allocation.

ARTICLE IN PRESS H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

147

Despite a growing literature that supports zoning protected areas, viewing parks as a part of a broader landscape, and recognizing heterogeneity across various subunits of a forest, most protected areas, including KYNP, are still established without zones or subunits. Cost-benefit analysis that reflects the benefits and costs from subunits of the area, and the interactions among different subunits, could inform decisions about the size and location of various zones. Yet thus far, the costbenefit literature provides little guidance to the establishment of zones, such as buffer zones, typically saying little more than that the ‘‘size of the zone will depend on the requirements of nearby residents’’ (p. 198, para. 4; Dixon and Sherman, 1990). Because the entire area of KYNP received the IUCN classification of National Park, official management cannot move toward zones and less restrictive access rules without a difficult process of expanding the area or de-accessioning parts of the park. Despite this constraint, unofficial policies to overlook NTFP extraction for home use and to work with rural development organizations reduce some of the burdens of KYNP on local people and indeed such policies move the de facto management of the park closer to the social optimum, as determined by the stylized full-spatial manager here. A small but growing literature identifies who bears the costs and who captures the benefits of protected areas. Le´le´ et al. (2001) considers how the establishment of a wildlife sanctuary in southern India has affected different stakeholder groups, including forest dwelling communities, farming communities, local tourists, and the global community. They suggest that a conservation strategy that combines strict protection from external pressures with sustainable use by local communities can be effective and relatively socially equitable. Van Beukering et al. (2003) identifies the benefits and costs faced by stakeholder groups of several land use scenarios, finding that conservation strategies improve equity while ‘‘deforestation’’ benefits the logging industry and increases income disparity. Ferraro (2002) estimates the opportunity costs of Ranomafana National Park in Madagascar borne by local residents to be $3.37 million. Such cost-benefit studies can contribute to a park manager’s ability to recognize potential sources of funding – such as international conservation groups – and potential sources of conflict – such as onerous burdens on the rural poor. Le´le´ et al.’s (2001) analysis explicitly identifies and distinguishes a number of different local users in terms of the extraction benefits that they get from a wildlife sanctuary and the damage they inflict. Similarly in KYNP, local users’ demands on the park are relatively low, such as extraction of fuelwood for subsistence, whereas non-local people’s activities are more damaging, such as the recent increases in incense wood extraction and wildlife poaching that have triggered an increased emphasis on law enforcement in KYNP (Mather, 2006). A disaggregated cost-benefit analysis that distinguishes between activities and between the groups that perform those activities could be used to support differential enforcement for more and less damaging activities. Also in the KYNP example presented here, we show that the international community could simply free-ride on Thailand’s provision of this preserved area. However, the international community recently granted KYNP, as part of the Dong

ARTICLE IN PRESS 148

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

Phayayen-Khao Yai Forest Complex, the UNESCO Natural World Heritage Status, which brings with it international support and prestige. This example of a disaggregated cost-benefit analysis that recognizes who gains and who loses from KYNP’s establishment and management suggests that KYNP might gain further from mechanisms to capture some of the benefits that accrue internationally – perhaps through higher entrance fees for international visitors – and from policies that meet local people’s needs in a sustainable manner and discourage open access extraction from the park. Finally, recent calls for more estimates of quasi-option value are thwarted by a lack of appropriate data. In recent years, quasi-option values have been proffered as an argument for the conservation of biodiversity, despite the lack of estimates of those values (Heal, 2004). As a rare estimate of quasi-option values in the tropics and that value’s role in land use decisions, this analysis (and the related work in Albers et al., 1996; Albers, 2001) makes four points to that end. First, even large quasi-option values may not alter the land use decision. Second, valuation studies rarely provide the information that is necessary to identify quasi-option values, which requires a description of both the probability of high values from preservation and the magnitude of those potential high values (Albers, 1996; Water Science and Technology Board, 2004). Third, the length of the timeframe considered matters for the size of the quasi-option value and its contribution to current period decisions. Fourth, the availability of land uses that do not irreversibly disrupt ecosystems – such as short term and small-scale agriculture – provides flexibility for managers, enabling them to take advantage of high-valued preservation in later periods. References Albers, H.J., 1992. Economic Management of tropical forests: uncertainty, irreversibility, and spatial relationships. Ph.D. dissertation. Department of Economics, University of California, Berkeley. Albers, H.J., 1996. Modeling ecological constraints on tropical forest management: spatial interdependence, irreversibility, and uncertainty. Journal of Environmental Economics and Management 30, 73–94. Albers, H.J., 2001. A Spatial-intertemporal model for tropical forest management applied to Khao Yai National Park, Thailand. Resources for the Future Discussion Paper 01–35. Albers, H.J., Ferraro, P., 2006. The economics of terrestrial biodiversity conservation in developing nations. In: Ramon, Lopez, Michael, A.Toman. (Eds.), Economic Development and Environmental Sustainability: New Policy Options. Oxford University Press, Oxford, UK, p. 486pp. Albers, H.J., Grinspoon, E., 1997. A comparison of the enforcement of access restrictions between xishuangbanna nature reserve (China) and Khao Yai National Park (Thailand). Environmental Conservation 24 (4), 351–362. Albers, H.J., Fisher, A.C., Hanemann, W.M., 1996. Valuation and management of tropical forests: implications of uncertainty and irreversibility. Environmental and Resource Economics 8, 39–61. Ando, A., Camm, J., Polasky, S., Solow, A., 1998. Species distributions, land values, and efficient conservation. Science 27 (5359), 2126–2128. Arrow, K.J., Fisher, A.C., 1974. Environmental preservation, uncertainty, and irreversibility. Quarterly Journal of Economics 88, 312–319. Brockelman, W.Y., Dearden, P., 1990. The role of nature trekking in conservation: a case-study in Thailand. Environmental Conservation 17 (2), 141–148. de Rouw, A., 1995. The fallow period as a weed-break in shifting cultivation (tropical wet forests). Agriculture, ecosystems, and environment 54 (1–2), 31–43.

ARTICLE IN PRESS H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

149

Dixon, J.A., Sherman, P.B., 1990. Economics of Protected Areas: A New Look at Benefits and Costs. East-West Center, Island Press, Washington, DC. Dobias, R.J., project coordinator. 1985. Final Report of WWF/IUCN Project 3001: Elephant Conservation and Protected Area Management. Dobias, R.J., Thorani Tech, Vute Wangwacharakul, and Nirand Sangswang. 1988. Beneficial Use Quantifications of Khao Yai National Park: Executive Summary and Main Report. Report on WWF Project 3757. Emerton, L., 2003. Tropical forest valuation: has it all been a futile exercise? Paper submitted to the 12th World Forestry Congress, Quebec City, Canada. /http://www.fao.org/DOCREP/ARTICLE/WFC/ XII/MS3-E.HTMS. Ferraro, P.J., 2002. The local costs of establishing protected areas in low-income nations: Ranomafana National Park, Madagascar. Ecological Economics 43, 261–275. Ferraro, P.J., 2003. Assigning priority to environmental policy interventions in a heterogeneous world. Journal of Policy Analysis and Management 22 (1), 27–43. Fisher, A.C., Hanemann, W.M., 1986a. Environmental damages and option values. Natural Resource Modeling. Fisher, A.C., Hanemann, W.M., 1986b. Option value and the extinction of species. Advances in Applied Micro-Economics 4, 169–190. Ghimire, K.B., 1994. Parks and people: livelihood issues in national parks management in Thailand and Madagascar. Development and Change 25, 195–229. Hanley, Nick, Susana Mourato, Robert, E.W., 2001. Choice modelling approaches: a superior alternative for environmental valuation? Journal of Economic Surveys 15 (3), 435. Heal, G., 2004. ‘‘Keynote Address’’. In: Proceedings of the USEPA’s NCEE Workshop ‘‘Valuation of Ecological Benefits: Improving the Science Behind Policy Decisions.’’ October. /http://yosemite. epa.gov/ee/epa/eerm.nsf/vwSER/96291273F5DF6C2085256F9B00733175?OpenDocumentS. Khao Yai Ecosystem Project Final Report, 1982. Volume III: Soil and Vegetation, by Wisut Suwannapinunt and Somkid Siripatanadilok. Volume IV: Wildlife, by Utis Kutintara and Somneuk Pongumphai. Volume V: Socioeconomics and Agricultural Development, by Sathit Wacharakitti, Prakong Intrachandra, Songkram Thammincha. Faculty of Forestry, Kasetsart University. December. Khao Yai National Park Management Plan 1987–1991, 1987. Report from the National Park Division of the Royal Forest Department and the Ministry of Agriculture and Cooperatives. Kiss, A., 1990. Introduction. In: Kiss, A. (Ed.), Living with Wildlife Resource Management with Local Participation in Africa. World Bank, Washington, DC. Le´le´, S., Srinivasan, V., Bawa, K.S., 2001. Returns to investment in conservation: disaggregated benefitcost analysis of the creation of a Wildlife Sanctuary. In: Ganeshaiah, K.N., Shaanker, R.U., Bawa, K.S. (Eds.), Proceedings of International Conference on Tropical Ecosystems: Structure, Diversity and Human Welfare. Oxford-IBH Publishing Co., New Delhi, pp. 31–33. MacArthur, R.H., Wilson, E.O., 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. MacKinnon, J., MacKinnon, K., Child, G., Thorsell, J., 1986. Managing Protected Areas in the Tropics. IUCN/Gland, Switzerland. Mather, R., 2006. Police getting tough in protection of Khao Yai. Worldwide Fund For Nature. Articles. /www.wwfthai.orgS. McNeely, J.A., Faith, D.P., Albers, H.J., 2005. Biodiversity’’ in Ecosystems and Human Well-Being, Volume 3: Policy Responses. Millennium Ecosystem Assessment. Island Press, Washington, DC, 621pp. MIDAS Agronomics Company Ltd, 1991. Study of Conservation Forest Area, Demarcation, Protection and Occupancy in Thailand. Unpublished report for the World Bank. Bangkok, Thailand. Milliman, S.R., 1986. Optimal fishery management in the presence of illegal activity. Journal of Environmental Economics and Management 13, 363–381. Muller, J., Albers, H.J., 2004. Enforcement, payments, and development projects near protected areas: how the market setting determines what works where. Resource and Energy Economics 26, 185–204. Naidoo, R., Adamowicz, W.L., 2005. Economic benefits of biodiversity exceed costs of conservation at an African rainforest reserve. Proceedings of the National Academy of Sciences 102 (Nov. 15), 16712–16716. Pearce, D., 2001. How valuable are the tropical forests? Demonstrating and capturing economic value as a means of addressing the causes of deforestation, Seminar paper for Conseil d’Analyse E´conomique, Seminaire Economie de L’Environnement Et du Developpement Durable, Paris, December.

ARTICLE IN PRESS 150

H.J. Albers, E.J.Z. Robinson / Journal of Forest Economics 13 (2007) 129–150

Pindyck, R.S., 1991. Irreversibility, uncertainty, and investment. Journal of Economic Literature 26 (3), 1110–1148. Repetto, R., 1988. The Forest for the Trees? Government Policies and the Misuse of Forest Resources. World Resources Institute, Washington, DC. Schultz, B., 1986. The management of crop damage by wild animals. The Indian Forester 112 (10), 891–899. Shavell, S., 1993. The optimal structure of law enforcement. Journal of Law and Economics 36, 255–287. Soule´, M.E., 1987. Viable Populations for Conservation. Cambridge University Press, Cambridge, UK. Stræde, S., Treue, T., 2006. Beyond buffer zone protection: A comparative study of park and buffer zone products’ importance to villagers living inside Royal Chitwan National Park and to villagers living in its buffer zone. Journal of Environmental Management 78 (3), 209–316. Studsrod, J.E., Wegge, P., 1995. Park-people relationships: the case of damage caused by park animals around the Royal Bardia National Park, Nepal. Environmental Conservation 22 (2), 133–142. Thailand Development Research Institute (TDRI), 1986. Phantumvanit, D. (Ed.), Thailand Natural Resources Profile. Tongpan, S., Panayotou, T., Jetanavanich, S., Faichampa, K., Mehl, C., 1990. Deforestation and Poverty: Can Commercial and Social Forestry Break the Vicious Circle? Research report no. 2 for the 1990 TDRI Year-End Conference, Bangkok, Thailand. van Beukering, P.J.H., Cesar, H.S.J., Janssen, M.A., 2003. Economic valuation of the Leuser National Park on Sumatra, Indonesia. Ecological Economics 44, 43–62. Water Science and Technology Board, 2004. ‘‘Valuing Ecosystem Services: Toward Better Environmental Decision-Making.’’ Committee on Assessing and Valuing the Services of Aquatic and Related Terrestrial Ecosystems. Division on Earth and Life Studies. National Research Council of the National Academies. The National Academies Press, Washington, DC /www.nap.eduS. West, P.C., Brechin, S.R. (Eds.), 1991. Resident Peoples and National Parks. University of Arizona Press, Tucson, AZ.