Assigning Economic Value to Natural Resources (1994)

Claudia W. Sadoff

The World Bank

Washington, D.C.

In standard calculations of national income there is a dichotomy in the treatment of natural and man-made capital assets. The differences in accounting procedures effectively overestimate the income generated by natural-resource-based production relative to the income produced using man-made capital, creating market signals that recommend increases in resource exploitation without regard for their full costs. The United Nations has presented preliminary natural resource accounting (NRA) guidelines to modify the current System of National Accounts (SNA), from which national income is compiled, with regard to the treatment of natural resource-related activities. As yet, however, no consensus has been reached concerning a standard technique for valuing natural resources and environmental services in the revised SNA.

The importance of specifying valuation methodologies is highlighted in this paper by a comparison of the policy implications arising from different NRA valuation techniques. Specifically, the ''user cost" and "depreciation" approaches were used to modify Thailand's forestry sector income between 1970 and 1995. The two methodologies differed in their implicit evaluation of Thailand's 1989 forest logging ban. While user-cost-adjusted income projections showed the ban to yield net economic gains, the depreciation approach captured changes in wood volume which demonstrated that stricter control of commercial wood removals was required to economically balance forgone forestry income under a logging ban. Thus, while a user cost analysis recommended the ban as economically beneficial, the depreciation analysis suggested that it was not.

The divergence in policy recommendations, however, was not so much a result of differences in the theories themselves, but rather a result of the specific units of measure used to value the resource. The logging ban was found to slow forest area loss, the criterion upon which forest depletion penalties were calculated in the user cost analysis, by 86 percent. The rate at which the volume of commercial wood was removed from the forests, however, slowed only 26 percent; commercial wood volume losses were used as the basis for calculating depreciation adjustments. A continuous decline in the density of Thailand's forests, and dispro

portionate clearing of denser and more commercially valuable forest areas, gave rise to considerably different rates of forest area and commercial wood volume losses. The fact that the two valuation techniques were applied using these two differing rates of forest loss decline, in turn, led to differences in the methodologies' policy recommendations.

National income indicators, by which policymakers often judge their countries' economic performance, abstract the costs of natural resource depletion and environmental degradation. By abstracting these costs, national income indicators mislead policymakers, presenting the depletion of environmental assets as income generation. In the case of natural resource extractive activities such as forest logging, income will be recorded as total revenues less extraction costs. No costs are deducted either for the inherent value of the resource or for any damage associated with its removal. Failing to subtract these costs inflates the true value added generated by the activity and presents the sale of an asset as production.

Capital depreciation provides another example of the asymmetrical treatment of natural and man-made capital in the SNA. Depreciation is not an economic transaction but an imputation capturing the decline in income-generating potential of assets over time. Yet depreciation is imputed and deducted only for wear and tear on reproducible, man-made capital. When natural assets are depleted, no analogous depreciation is recorded. The exploitation of resources and degradation of the environment undoubtedly lessen an economy's productive capacity, particularly those developing economies which rely heavily on resource-extracting industries. This being the case, it is clearly inconsistent and misleading to deduct depreciation for productive man-made capital, while ignoring the analogous depreciation of productive natural capital.

Various modifications to the national accounts have been proposed. The different methodologies focus on specific shortcomings in the current income calculations and differ in their valuation of the services provided by the environment.

No standard valuation technique for natural resource accounting has yet been established. Opinion remains divided among market values, opportunity or replacement costs, and discounted future revenues as bases for valuation of the physical units. To remain consistent within the SNA, which was constructed to measure formal market activity, it is generally accepted that only the commercially salable value of forest timber would be included in the valuation of forest resources. Significant nonmarket values and externalities are not captured by these resource depletion adjustments. In this sense, most natural resource accounting adjustments are extremely conservative, reflecting only a small portion of the true social costs of deforestation.

There are two commonly discussed NRA approaches that address resource depletion and degradation in the national income accounts; they are the "user cost" (see El Serafy in Ahmad et al., 1989), and the "depreciation" methodologies (see Repetto et al., 1989). The two methodologies attempt to address a basic asymmetry in the way the SNA treats natural and man-made capital. They differ, however, in the valuation of natural resource depletion.

To separate the cost of resource depletion from value added, the user cost approach explicitly calculates and subtracts a capital consumption component from the standard stream of recorded income. The premise of this approach is that revenues from resource-based activities include a component which represents the final sale of a natural asset, a component that is not value added, but rather disinvestment. El Serafy argues that if the owner of a natural asset is to consume only his true income, he must lend or invest a portion of his current revenues that would be capable of generating income to compensate the loss of revenues from his wasting asset in the future.¹ The sale of nonrenewable resources, or the unsustainable exploitation of renewable resources, is similar to the sale of an asset. It contains both a capital component or user cost, and a value added or income component.

In the case of renewable resources such as forests, El Serafy claims that an appropriate maintenance or reforestation cost, the cost required to sustain the productivity of natural capital, should be charged against the gross revenues of those activities which deplete or degrade the asset. If the asset is not, in fact, sufficiently maintained, the costs which would have been incurred in doing so should be charged against income.²

Calculations of user cost will necessarily differ by resource, and difficulties are certain to arise in defining maintenance. To maintain forest resources, for example, should the area of forest or the volume of wood be the appropriate criterion by which to measure depletion and, hence, replacement? Furthermore, if numerous technically acceptable means of maintenance exist, with widely varying costs and benefits, what criteria should be used to choose among them? In most cases, data availability will be the practical arbiter of such questions, but standardization and comparability will be comprised as a consequence.

The depreciation approach emphasizes actual natural resource depreciation, in contrast to the user cost approach which focuses on losses in future income resulting from the decline in natural resource productivity.

Standard calculations of national income impute and subtract the depreciation of man-made capital from gross domestic product (GDP) to arrive at the net domestic product (NDP). The rationale behind imputing depreciation is that the wear and tear of capital usage today will decrease the productivity of this capital in the future, and that this decreased future

productivity should be reflected in net national income. The depreciation approach argues that this same logic applies to the degradation of natural assets, yet in the case of natural resources, no decrease in productivity, or depreciation, is reflected in the accounts.

Stocks of natural assets therefore ought to be compiled and depreciated in the manner of man-made capital, to more accurately reflect the country's declining productive capacity. Specifically, the value of the net change in natural resource stocks over each accounting period should be calculated as depreciation. This natural resource depreciation would then be subtracted from GDP, in addition to the depreciation calculated for man-made capital, to arrive at a resource depletion-adjusted NDP.

Thailand's ban on forest logging provides an illustrative backdrop to examine the policy implications of natural resource accounting methodologies. Here, both the user cost and depreciation valuation techniques were applied to adjust forestry sector income in Thailand.

Despite a long history of forest management, deforestation has been significant in Thailand. In the beginning of the twentieth century, over 75 percent of the kingdom was covered with forest. Since 1960, the forest has shrunk at an average annual rate of roughly 2.5 percent, peaking at over 3.3 percent in the 1970s. Official estimates claim that less than 27 percent forest cover remains. Unofficial estimates suggest the actual figure may be closer to 18 percent.

In the fall of 1988 Thailand suffered floods and mud slides which killed more than 370 people. The intensity of the floods was blamed on the area's severely deforested surroundings and watershed. A nationwide logging ban was issued in 1989 in response to the floods and growing public concern for Thailand's forests. The ban itself is not totally restrictive. It allows, for example, the felling and sale of trees in privately operated forest plantations, harvests of designated species and trees which have been damaged by age or natural disasters, and the clearing of forests for infrastructure projects.

Official data reported by the Royal Forest Department (RFD) showed a significant decline in forest area losses following the ban. The rate of deforestation fell from 1.7 percent in the 1980s before the ban, to less than 0.3 percent after 1989—a decline of nearly 86 percent below projected trend levels.³ The success of the ban, however, is not as clear as these numbers make

it appear. Each year since the establishment of the ban, an average of over 300 square kilometers of forest has been cleared. In addition, the ban has been considerably less successful in slowing the volume of commercial wood removed from the forests.

The estimated annual volume of commercial wood losses has fallen just 26 percent from projected trend levels since the ban. The forests that have been illegally logged since the 1989 ban must therefore be significantly more dense and/or support a higher percentage of commercial species than those previously logged. In other words, forest loss has been successfully slowed only in those areas where it is least profitable.

Moreover, if the government-awarded concessions had been optimally managed to maximize wood extraction and minimize environmental damage, then this shift from previous patterns must imply a more ecologically harmful system of forest logging.

Illegal logging is not a new phenomenon in Thailand. A ban on legal logging could only serve to increase the well-established demand for illegally procured logs. In 1980 the total illegal harvest was estimated to be twice the magnitude of the legal cut,⁴ and in 1991 it was believed to be more than six times as high.⁵ In some villages where government officials exercise their full authority, and where powerful political and business interests are absent or uninvolved, the ban has been successful in halting deforestation. In many areas, however, losses of legitimate logging employment and rising log prices have led to actual increases in cutting.⁶

Enforcement of Thailand's logging ban is likely to become even more difficult in the future. In order to insulate Thailand's wood-related industries from the adverse effects of a restricted domestic log supply, policies to facilitate regional timber imports were adopted in concert with the logging ban. Certification requirements were lifted, customs procedures were expedited, and duties fell to negligible levels. Thailand's log imports grew more than four-fold in the first two years following the ban and are causing a sizable spill-over effect of deforestation and environmental damage in the region. It has become increasingly apparent that while Thailand's forests may be spared by the logging ban, Thai import demand has contributed to widespread and relatively uncontrolled logging in the forests of its trading partners. The isolated national ban is simply shifting deforestation across borders. In response, however, the Southeast Asian nations have begun to tighten regulations and control the flood of logs into Thailand. The upward pressure this will exert on the market price of logs will in turn bolster the demand for, and profitability of, illegally procured logs.

Income growth in the wood-based industries has not been significantly affected by restrictions on the domestic log supply. Since the 1970s Thailand has encouraged the production of higher value added processed wood products. As a consequence, the total income earned by

wood-based industries exceeded that of forestry in 1977, and by 1990 it was nearly five times as high. Wood-based industries include paper and paper products,⁷ furniture, and processed wood products, all of which have continued to grow since the ban. The construction sector, which accounts for the use of roughly 90 percent of domestic logs and 65 percent of imported logs,⁸ has been one of the strongest performers in the economy, showing strong uninterrupted growth following the 1989 ban.

The adoption of legislation to preserve natural forests presupposes that those forests embody social value. Current calculations of national income, however, assign no value to standing forests. Standard economic indicators are thus inappropriate tools for analyzing the economic effects of such policies because they would tally only the costs, and assign no value to the benefits, of forest preservation. Natural resource accounting can be used to impute and incorporate a value for standing forests so that the economic success of forest protection can be more comprehensively assessed.

The user cost of forest exploitation in Thailand was calculated to adjust forestry sector income before and after the 1989 logging ban. User cost, following El Serafy, was defined as the expense incurred in replanting the forest area cleared each year, and minimally maintaining it to a harvestable age. The calculation of this charge requires specification of the timing, costs and extent of reforestation.

The timing of forest replacement will affect both the total area and cost per unit area reforested. Here, reforestation was calculated for the net area cleared of forests at the end of each accounting period. Yet, as will be discussed below, a considerable decline in the density of forested areas has occurred as well. Charging the costs of reforestation only when an area is fully cleared postpones the penalty to an accounting period in which most, though not all, of the forest loss occurs.

Before an area is cleared, however, it still retains great natural regenerative capacity. Because it is the loss of this productive capacity that is measured by user cost, it would be inappropriate to apply the penalty to an area in which natural rehabilitation was not significantly compromised. Even if a practicable definition could be found to identify the point at which forests no longer can effectively renew themselves, the lack of forest volume data in Thailand

makes it impossible to distinguish declining forest density until the forest's crown disappears from LANDSAT images altogether, and the area is recorded as non-forest.

Furthermore, past experience replanting forests as they are thinned, rather than after they are cleared, suggests this method of reforestation is prohibitively expensive and complex in Thailand. Unsuccessful replanting of selectively logged forests led to calls for forest clear-cutting in the mid-1980s; a practice which was then officially agreed to on an experimental basis, but which had always been the de facto system of logging. The only ''technically acceptable criterion" for reforestation in Thailand would therefore appear to be total replanting of cleared forest areas.

The second issue to be clarified in defining user cost, is whether forests should be maintained in terms of land area or wood volume. If a wood volume approach were adopted, larger tracts of natural forest cover could be replaced with smaller, higher-density, monoculture plantations. Wood, however, is only one of many forest products. Many nonmarket forest functions would in fact be hampered by high volume monoculture plantations. Plantations of this type crowd out undergrowth, limiting the diversity of flora and hence fauna supportable in the secondary forests. Widespread public protests have occurred regarding the harmful effects of commercial plantations, particularly eucalyptus which has been found to significantly draw-down the water table of surrounding agricultural lands.

Moreover, there is a clearly articulated social preference in Thailand to maintain significant, specified areas of forest cover. Each of Thailand's seven National Economic and Social Development Plans (1961-1996) have called for a target of at least 37 percent forest cover in the country, though in nearly all periods actual forest area has fallen well below the target level. These repeated calls for ecological balance suggest that conservation of forest area in itself is a priority for Thai society.

Deforestation was therefore defined as the net forest area cleared each year. Forest losses by region over the period were calculated by interpolation of periodic aerial photography and LANDSAT survey data published by the RFD.

The cost of reforestation was calculated as the sum of the present discounted costs of establishing a forest plantation, and maintaining it until a harvestable age. This would be the amount necessary to set aside in the year deforestation occurred in order to fund complete forest renewal.

Reforestation is by nature a labor-intensive operation. The cost of reforestation is thus driven by the cost of labor. Calculations of reforestation costs over time were therefore derived primarily from labor requirements and regional wages. A conservative estimate of the cost of labor was calculated at the prevailing legal minimum wage.⁹ Labor requirements for

reforestation were drawn from a study of the employment effects of forest plantations by Tingsabadh, and based on a ten-year profile for the establishment and maintenance of a mixed forest plantation.¹⁰ A mixed plantation standard was used on the assumption that, though monoculture plantations have somewhat lower maintenance costs, mixed forests present closer ecological approximations of natural forests. Similarly, while fast-growing tree species such as eucalyptus have shorter maturation periods, these foreign species are less acceptable substitutes for Thailand's natural forests, and their use has met with stiff local resistance. Costs over the ten-year cycle were discounted back to the year deforestation occurred in order to arrive at the present value of reforestation per rai by region in each year.

The series of present value regional labor costs was then weighted by the proportion of terrestrial deforestation, and hence required reforestation, in each region during the relevant period. To this, capital costs were added. Capital requirements, which account for roughly ten percent of total costs, are relatively small and unchanging. Capital inputs were therefore assumed to be a constant ten percent of labor cost in each period. The time series of reforestation costs calculated under these assumptions fell well within the range of published cost estimates for plantations of various types. Finally, total costs were deflated to create a time series of total, real reforestation costs per rai. The resulting series was multiplied by net deforestation in each period to arrive at the real user cost penalty for that period. This forest depletion penalty was then subtracted from forestry sector income to determine the user cost-adjusted forestry income in each period.

A similar series of resource depletion-adjusted forestry income was calculated following the depreciation approach. Depreciation was defined as the net change in value of the forest asset. It was calculated as the volume of commercially valuable wood removed from the forest annually, priced according to its stumpage value. This represents the opportunity cost of wood still standing in the forest.

This methodology captures only the commercial timber loss resulting from deforestation. It is an understatement of the true loss to society because it excludes the value of environmental services, nonmarket forest production, and non-commercial timber species. Restricting adjustments to commercial timber, however, will maintain the consistency of the SNA, a system which is designed to reflect only market transactions and their clear proxies.

The physical volume of the forest stock will depend on both forest area and density. Forest area estimates were derived from published and secondary RFD data sources, and broken down into four types in four regions using intermittent RFD surveys. Proportions of the forest types for each region were interpolated between the two survey years, and elsewhere assumed

constant. The constructed forest area time series reflects the net changes in each period, by forest type and region. These area estimates are the same as those used to calculate the user cost penalties.

No current or time series forest density data exist for Thailand. The First and Second National Forest Density Inventories performed in Thailand between 1969 and 1973, and 1975 to 1979, are still considered the best available density measures. A considerable decrease in density, approximately six percent annually, was seen between the two surveys, suggesting that forests were being thinned as well as cleared over time. A straight application of the late 1970s densities would thus almost certainly overstate actual forest volume by failing to account for such thinning. On the other hand, a projection assuming a continuation of the rate of decline in density seen between the two inventory periods might well understate actual densities, because the period in which the measurements were made was a time at which forest clearing was at its peak.

The relationship between the rate of change in forest area and the rate of change in forest density was found to be statistically significant at the 95 percent confidence level for each of the forest categories over the period between the two national inventory surveys. A simple model was therefore constructed using deforestation rates as a proxy for pressure on the forests to estimate the change in forest densities by type and region.

This derived density function was used to project changes in density from a base-year measurement of the second national forest inventory in 1977. The density model permitted both increasing and decreasing forest densities, capturing natural regeneration in previously thinned forest areas where decreased pressure on the forest allowed for net growth. Where this relationship projected an increase in forest density that exceeded the forest type's natural growth rate, the natural growth rate was used to project density changes. The estimated density series,¹¹ as expected, fell markedly over time, but did so at a declining rate after the late 1970s.

The total volume of the forest stock over time was calculated by applying forest densities by type and region, to the corresponding forest areas. These calculations, like those performed in the user cost analysis, capture only the wood loss resulting from total forest clearing; declining densities in remaining forest areas are not reflected by these measurements.

Changes in the physical volume of the forest stock must be quantified in monetary terms if they are to be incorporated into a national income accounting framework. The market price of wood products, however, is not an appropriate value to attach to the wood inherent in a standing forest. A stumpage value, the value of wood still on the stump, must therefore be calculated.

Stumpage values are calculated from the market prices of wood products by subtracting the costs and profits associated with their production. These costs generally include extraction,

transport and processing. Here, world average export log prices were used as a starting point for the calculation, as they reflect the economy's opportunity cost of wood.

Extraction costs were estimated from benchmark year reports on the cost of forest log felling and removal. Transport costs were calculated in two parts. Estimates were first made for the cost of transporting logs from their roadside felling sites to sawmills. The rates charged for these transfers were higher than the standard transport rates because they travelled more remote and less well kept road networks. Average distances between felling sites and sawmills were used for each region.¹² The second component of transportation cost was the transfer of logs from sawmills to Bangkok. Stumpage values calculated from freight-on-board prices must include all of the costs required to deliver logs to the location at which export prices would apply, in this case Bangkok. Regional distances were determined by weighting the distance from provincial capitals to Bangkok, by each province's sawmilling capacity. Cost rates for standard highway trucking were then applied to find the cost of transfer from mills to Bangkok.

Regional stumpage values were applied to that portion of wood volume loss that could reasonably have been expected to arrive at market. Wood that either lacks commercial value or is effectively irretrievable has no opportunity cost. Volume was therefore adjusted to include only commercial species in each region. Percentages of commercial species by region were taken from a study by Thammincha¹³ and were assumed to remain constant. Adjustments were also made to allow for the volume of timber normally damaged in the logging process, an amount which would contribute to net forest loss, but which could not be expected to reach market. A ratio of 1.71¹⁴ was used to reflect the total volume of wood felled or damaged for every cubic meter marketed.

This stumpage value, applied to the adjusted volume of wood loss over each accounting period, represents the depreciation of Thailand's forest assets.

Following both the user cost and depreciation¹⁵ methodologies, depletion-adjusted forestry sector income between 1971 and 1982 were actually negative. A reversal in the sign of adjusted forestry income results from the fact that climbing rates of deforestation in the 1970s were not matched by corresponding increases in the sale of timber. Standard measures of

forestry GDP over the period remained quite steady while depletion penalties rose, leading to negative depletion-adjusted incomes. An implication of this is that deforestation in the period was not effectively driven by commercial logging operations but was more the result of agricultural expansion and forest encroachment.

The magnitude of the forest depletion adjustments calculated using the depreciation approach were generally larger than the comparable figures produced following the user cost approach, particularly before 1979. Differences arise from the fact that the user cost approach, in this study, was calculated on a forest area basis, while the depreciation approach was based on the volume of commercial wood lost.¹⁶ A continuous decline in forest density made these differences more pronounced in the earlier years, when changes in forest area led to proportionately larger changes in wood volume.

Resource depletion adjustments indicate that the failure to account for the cost of deforestation in standard measures of GDP, has led to consistently overstated levels of national income in Thailand. Following the user cost approach, GDP adjustments for forest loss between 1970 and 1990 yielded an average annual income 1.45 percent lower than the figures derived by standard calculations. The depreciation approach called for an average downward adjustment of 2.17 percent.

At the same time, the adjustments indicate that the rate of Thailand's GDP growth is actually understated. Standard GDP calculations found 7.29 percent real average annual growth between 1970 and 1990. Over the same period the user cost-adjusted incomes grew at a rate of 7.38 percent, while depreciation adjustments resulted in a real growth rate of 7.61 percent. The upward adjustment in income growth was a result of the declining rate of deforestation, and hence the declining magnitude of resource depletion adjustments since the 1970s.

To examine policy implications of the two NRA accounting analyses, two different logging scenarios were projected to the year 1995—assuming continuation of the current ban and assuming that the ban had never been imposed.

The two approaches differed in their implicit evaluation of the logging ban. The ban severely restricts recorded forestry income, regardless of its effectiveness in terms of forest protection. Resource depletion penalties, however, are calculated in proportion to actual forest savings. The magnitude of foregone forestry income relative to savings in terms of forest resource depletion penalties will dictate whether the policy yields net economic benefits. User cost calculations in this case showed forest area savings to be of greater value than foregone forestry income, hence, the positive user cost-adjusted incomes. Forest savings in terms of commercial wood volume, the measure by which depreciation penalties are calculated, were considerably lower and did not outweigh foregone income. Depreciation-adjusted incomes were therefore negative.

Results of the user cost adjustments to forestry income recommend continuation of the current logging ban. The ban led to considerable decreases in forest area loss, and therefore decreases in user cost depletion penalties. The value of these forest area savings outweighed losses in recorded forestry GDP, making a continued logging ban the highest income scenario. An enforcement level of 80 percent, in forest area terms, was required to achieve net economic gains following the user cost approach.¹⁷ The current ban has decreased forest area losses by 86 percent, hence its recommendation by the user cost analysis.

The depreciation-adjusted income projections, however, recommend a different policy ranking. While the ban has produced environmental gains, enforcement must be tightened to yield economic benefits as well. An enforcement level of 50 percent, measured by commercial wood volume losses, was required to balance forgone forestry income and produce positive depreciation-adjusted incomes. The current ban, however, achieved only a 26 percent decline in commercial wood volume removals, thus creating net economic losses when depreciation adjustments were made.

The explicit accounting of commercial wood volume changes in the depreciation methodology called attention to important facts concerning the sufficiency of the logging ban's enforcement; facts which failed to be reflected in the user cost analysis. It is clear that effective forest protection in Thailand has been extended only to those areas in which logging is least profitable. Continued illegal logging in Thailand's most dense, commercially valuable forests has led to unintended and unnecessary economic and ecological damages. It should be noted, however, that in this instance the difference in illustrative power of the two methodologies is largely a result of the variables chosen in their application, rather than a consequence of the two techniques' theoretical underpinnings.

Recognizing the costs and consequences of its forest use policies, the Thai government banned commercial forest logging in 1989 to protect the country's forests and restructure its management system. The logging ban was originally expected to be an economic sacrifice for environmental gains; ironically, it has resulted in little of either. The economy has been largely insulated from the ban's anticipated negative effects, while deforestation continues in Thailand's most pristine and commercially productive forests and grows rapidly in neighboring countries.

On a national level, the economic and forest area losses that have been incurred under the ban are not inevitable consequences of the legislation, but rather the result of failed enforcement. When calculated in terms of forest area, the rate of deforestation declined by 86

percent following the logging ban. The loss of commercial wood volume, however, declined by only 26 percent. If the ban were more strictly enforced, and commercial wood volume losses were curtailed by 50 percent, the logging ban could achieve both net economic gains¹⁸ and forest preservation in Thailand.

Both the user cost and depreciation methodologies suggested that standard GDP calculations overstated Thailand's national income as a consequence of failing to account for forest depletion. The user cost methodology suggested that real GDP was overstated by an average of 1.45 percent annually, while the depreciation adjustments found GDP inflated an average of 2.17 percent. Gross capital formation was found to be 6.4 and 9.5 percent inflated, following the user cost and depreciation methodologies respectively.

A natural resource accounting analysis of Thailand's forest logging ban pointed to the differences between the two approaches. The user cost analysis found the ban to be both ecologically and economically beneficial, while depreciation adjustments suggested that forest protection was not sufficient to reap net economic gains. This divergence, however, is not so much a result of differences in the theories themselves, but rather of data choices in their application.

If the costs and benefits of resource use policies can be prescribed by the choice of valuation technique, then natural resource accounting without a standardized valuation methodology cannot provide a consistent analytical framework for the economic evaluation of resource management.

Ahmad, Yusuf J., Salah E1 Serafy, and Ernst Lutz, eds. Environmental Accounting for Sustainable Development. A UNEP-World Bank Symposium. Washington, DC: The World Bank, 1989.

ESCAP. State of the Environment in Asia and the Pacific 1990. Bangkok: United Nations Economic and Social Commission for Asia and the Pacific, 1990.

Forest Industries Organization. FIO Annual Report. Various years. Bangkok: FIO. (in Thai)

Forest Inventory Team Working Document. Bangkok: Royal Forest Department, 1991.

Hicks, J. R. Value and Capital: An Inquiry into Some Fundamental Principles of Economic Theory. (first edition 1939). Oxford: The Clarendon Press, 1946.

Hoamuangkaew, Wuthipol, and Prakong Intrachandra. The Structure of Sawmilling Industry and Sawnwood Markets in Thailand. Paper presented at the Thai Forestry Sector Master Plan Seminar "Demand for Industrial Forest Products and Roundwood in Thailand up to Year 2015." Bangkok, 25 November 1991.

Inter-Secretariat Working Group on National Accounts. "System of National Accounts (SNA) Review Issues." Paper presented at the 1990 Regional Commission Meeting on SNA. United Nations Statistical Office, New York, February 1990.

Markets for Industrial Forest Products and Roundwood in Thailand. Paper presented at Thai Forestry Sector Master Plan Seminar: Demand for Industrial Forest Products and Roundwood in Thailand up to Year 2015. Bangkok, 25 November 1991.

NESDB. Summary: The Seventh National Economic and Social Development Plan (1991-1996). Bangkok: National Economic and Social Development Board, 1991.

National Income of Thailand. Bangkok: National Economic and Social Development Board, various years.

Peskin, Henry. Accounting for Natural Resources Depletion and Degradation in Developing Countries. Environment Department Working Paper No. 13. Washington, D.C.: The World Bank, 1989.

Repetto, Robert, Magrath, W., Wells, M., Beer, C., and Rossini, F. Wasting Assets: Natural Resources in the National Income Accounts. Washington, D.C.: World Resources Institute, 1989.

Richards, John and Richard Tucker, eds. World Deforestation in the Twentieth Century. Raleigh: Duke University Press, 1988.

Royal Forest Department. Forestry Statistics of Thailand. Various years. Bangkok: Forest Statistics Sub-Division, Planning Division, Royal Forest Department.

Sadoff, Claudia W. Natural Resource Accounting: A Case Study of Thailand's Forest Management. Doctoral dissertation submitted to the University of California at Berkeley, December, 1992.

Sterk, Arjen and Pieter van Ginneken. Benefit-Cost Analysis for Forest Plantations in Phu Wiang. FO:DP/THA/84/002 Field Document 4. Bangkok: FAO, 1987.

Thai Forestry Sector Master Plan. "Forest Plantation Development." Working Draft 3. Bangkok: Royal Forest Department, 1992.

Thailand Development Research Institute. Thailand Natural Resources Profile. Bangkok: Thailand Development Research Institute, 1987.

Thammincha, Songkram. Thailand's Forest Resources Data. Bangkok: FAO, 1982.

The Efficacy and Economic Impacts of Thailand's Forest Logging Ban. The World Bank, forthcoming.

Tingsabadh, Charit. Employment Effects of Reforestation Programs. Bangkok: Thailand Development Research Institute, 1989.

Tropical Science Center and the World Resources Institute. Accounts Overdue: Natural Resource Depreciation in Costa Rica. Washington, DC: World Resources Institute, 1991.

United Nations Secretariat. Revised System of National Accounts: Draft Chapters and Annexes. Provisional ST/ESA/STAT/SER.F/2/Rev.4. 1991.

User Cost and Depreciation: A Practical Comparison of Natural Resource Accounting Methodologies. The World Bank, forthcoming.

Ward, Michael. Accounting for the Depletion of Natural Resources in the National Accounts of Developing Economics. Pads: Organisation for Economic Cooperation and Development, 1982.