l51- E N V I R O N M E N T - «a - ~~~~D E P A R T M E N T PAPERS PAPER NO. 57 TOWARD ENVIRONMENTALLY AND SOCIALLY SUSTAINABLE DEVELOPMENT -^i r^ ENVIRONMENTAL ECONOMICS SERIES Estimating National Wealth: Methodology and Results Arundhati Kunte Kirk Hamilton John Dixon Michael Clemens January 1998 ESD Environmentally Sustainable Development F L CWorld Bank ESD - V Indicators and Environmental Valuation Estimating National Wealth: Methodology and Results Arundhati Kunte Kirk Hamilton John Dixon Michael Clemens January 1998 Papers in this series are not formal publications of the World Bank. They are circulated to encourage thought and discus- sion. The use and citation of this paper should take this into account. The views expressed are those of the authors and should not be attributed to the World Bank. Copies are available from the Environment Department, The World Bank, Room MC-5-119. I Contents Abstract v Foreword vii Chapter 1 Introduction 1 Chapter 2 Natural Capital 4 Discount Rates 5 Agricultural Cropland 5 Pasture Land 7 Timber Resources 7 Non-Timber Forest Benefits 7 Protected Areas 8 Sub-Soil Assets 8 The Unfinished Agenda for Natural Capital 10 Chapter 3 Produced Assets 11 Chapter 4 Human Resources 13 Chapter 5 Caveats 15 Chapter 6 Conclusions and Policy Implications 17 Appendix A Estimates of National Wealth: Total and Components 19 Environmental Economics Series iii Natural Habitats and Ecosystems Management in Drylands Appendix B Estimating Rental Values for Timber and Sub-Soil Assets 27 Rent from Oil 27 Rent from Natural Gas 28 Rent from Metals and Minerals 30 Rent from Hard Coal 32 Rent from Brown Coal (Lignite) 34 Timber Rent and Mean Annual Increments 36 References 39 Figures 1 Composition of wealth in low-income natural resource exporters, 1994 3 2 Components of natural capital, 1994 6 3 Human resources and years of education 14 B1 A comparison of several natural gas export prices from around the world 29 Tables 1 Wealth per capita by geographic region, 1994 2 Al Estimates of national wealth: Total and components 19 A2 Natural capital estimates, 1994 22 B1 "Surrogate" countries used to estimate oil production costs 28 B2 "Surrogate" countries used to estimate natural gas production costs 30 iv Enviromnent Department Papers Abstract As the nations of the world move from paper provides details on the methodology commitment to action in achieving and assumptions used in estimating these sustainable development the need for magnitudes. The results show that human measuring progress toward this end is resources play a predominant role, thus heightened. This paper builds on the concept lending support to investments in education of wealth (or the asset base) as the foundation and health. They also highlight the of generating well-being. National wealth importance of agricultural cropland and takes on a much broader definition and is pasture land, pointing to the need for embodied in natural capital; human resources sustainable land management practices. The that include education, raw labor, and social estimates support our intuitive capital; and produced assets (machinery, understanding of the importance of people equipment, buildings, and urban land). The and the environment in development. Enviromnental Economics Series v Foreword To estimate the total wealth of nearly 100 below are applicable in any context, the nations, as presented in summary form in methods applied are often constrained by the Expanding the Measure of Wealth: Indicators of data. Analysts working in a given country can Environmentally Sustainable Development generally employ more copious and pertinent (World Bank 1997b), a number of strong data, and so should be able to apply the conceptual assumptions are required. If these estimates approach more directly than we have below. While are to be credible, it is vitally important that we have espoused sound economics in what these assumptions be spelled out carefully. follows, the details of the methods applied This is the prime motivation for this working should not be construed as the last word in paper. expanded wealth accounting. In addition, this paper demonstrates that The final key point is inherent in the title of expanding the national accounts to include this working paper: the numerical results the environment and natural resources is a presented are estimates and so do not carry practical exercise, even when working within the authority of the World Bank to the extent the limits imposed by the availability of of the development data appearing in World international data sets. This raises a key point Development Indicators or the World however: while the economic concepts applied Development Report. Environmental Economics Series vii 1 Introducfion In the five years since the Rio Earth summit natural resources and also presents a much has been achieved not only in terms of complete picture of the costs incurred in the raising awareness of environmental concerns, process (Hamilton and Lutz, 1996). Building but also in instituting specific innovative on the same concept, "genuine" savings policies that capitalize on the potentially measure the true rate of savings in an positive link between economic development economy after taking into account depletion and the environment (World Bank 1997a, of natural resources and damage caused by Steer 1996). If governments are indeed pollution. For greater detail on green NNP moving from commitment to action it is and genuine savings, interested readers are . . . ~~~~~referred to Hamilton and Lutz (1996) and important to be able to measure and assess the results of such actions. Recent years have World Bank (1997b, chapter 2) respectively. seen much attention devoted to measuring The motivation of this paper is to elaborate progress toward sustainable development. on wealth accounting as a measure of Indicators to measure the pace and direction sustainable development. As defined by the of environmental change have ranged from report of the Brundtland Commission (World improved physical indicators of Commission on Environment and environmental resources, to measures linking Development, 1987), sustainable the macro-economy and the environment. development is development that meets the needs of the current generation without Prominent among indicators linking the compromising the ability of future macro-economy and the environment are generations to meet their own needs. The measures of "green" net national product simplest interpretation of this (and the one (green NNP), genuine saving, and wealth adopted here) is to leave future generations accounts. The field of integrated economic as many opportunities as we ourselves have and environmental accounting has made had, if not more. Here opportunity is significant advances in improving NNP as an measured by the capital stock which forms indicator of sustainable development by the basis for well-being (Serageldin 1996). To suggesting adjustments that bring the the extent that the available stock of assets- suggestirongajsmenta ts that bronomin acth y produced assets, natural capital, and human environmental effects of economic activityreocsanscilapt-mows int th .antem NPmaueh resources and social capital-,empowers into the mainstream. NNP measures the economic actors (in this and future annual flow of economic production based on generations) to create well-being, it is central market transactions, thereby leaving out the to comprehending and operationalizing the impact of economic activity on a very concept of sustainable development. important national asset-natural capital. By Maintaining, and/or enhancing the accounting for the degradation and depletion productive potential of an economic entity of natural capital, green NNP reflects the requires the creation, maintenance and sound productive services of labor, capital, and management of wealth, broadly defined. Environmental Economics Senes Estimating National Wealthi Methodology and Results Our approach is to seek the answers to two gross national product that can be questions. First, what are the components considered "returns to labor" irL and contributing factors to national wealth? agricultural and non-agricultural sectors, Since the sustainable performance of an taking the present value of this stream economy is influenced by the portfolio of over the mean productive years of the assets over time, this leads to the second population, and then subtracting the question: how best to manage and maintain stock of produced assets and urban land. this portfolio to promote sustainable Included in this component is the return economic development? The wealth estimates to social capital. are the sum of the following three major components: Not surprisingly, the human resources component which combines returns to raw * Natural capital. This is calculated as the labor, human capital, and social capital is the sum of the stock value of the following most important constituent of wealth. It renewable and non-renewable should be mentioned at the outset that natural resources-agricultural land, pasture capital values are primarily based onI land, timber, non-timber forest benefits, instrumental or use values of the environment protected areas, oil, coal, natural gas, and that important ecological and life support metals, and minerals. functions of natural systems have not been * Produced assets. This is the sum of the valued. This is in part due to the fact that value of a country's stock of machinery methods of valuation are more well developed and equipment, structures and urban for instrumental values. We have, however, land. included the value of protected areas and * Human resources. This is calculated as a some non-timber forest benefits such as minor residual by estimating the percentage of forest products and recreation. Table 1. Wealth per capita by geographic region, 1994 Total Human Produced Natural Human Produced Natural wealth resources assets capital resources assets capital dollars per capita percent share of tot4I wealth North America 326,000 249,000 62,000 16,000 76 19 5 Pacific OECD 302,000 205,000 90,000 8,000 68 30 2 Western Europe 237,000 177,000 55,000 6,000 74 23 2 Middle East 150,000 65,000 27,000 58,000 43 18 39 South America 95,000 70,000 16,000 9,000 74 17 9 North Africa 55,000 38,000 14,000 3,000 69 26 5 Central America 52,000 41,000 8,000 3,000 79 15 6 Caribbean 48,000 33,000 10,000 5,000 69 21 11 East Asia 47,000 36,000 7,000 4,000 77 15 8 East and Southern Africa 30,000 20,000 7,000 3,000 66 25 10 West Africa 22,000 13,000 4,000 5,000 60 18 21 South Asia 22,000 14,000 4,000 4,000 65 19 16 Notes: The aggregate figures for West Africa do not include Nigeria because of data quality issues. Similarly, figures for North Africa do not include Algeria. Pacific OECD includes Japan, Australia, and New Zealand. Source: Authors' estimates. 2 Environment Departnent Papers Introduction The different components of wealth were Figure 1 estimated for nearly one hundred countries Composition of wealth in low-income and aggregate results are presented below natural resource exporters, 1994 (see tables 1 and 2 in Appendix A for details on wealth estimates and the various components of natural capital for each Human country). The selection of countries was Pesources Natuma governed wholly by considerations of data 59% Capital availability and quality. Since we used 20% consolidated sources this was a particular concern. A greater investment of time and resources in collecting and validating datad from countries would certainly improve the Assets coverage of countries and the reliability of 21% estimates. The main results are presented in Source: Authors' estimates. table 1 which shows wealth and its components in dollars per capita and the the least variation across income groups and relative shares in different regions of the regions (15 to 30 percent). world. An exercise of this nature necessarily involves We observe that in all regions of the world several strong assumptions and the details human resources form the lion's share of are critical to understanding the results. In wealth. Across regions the relative share of what follows we discuss the methodologies wealth4 and assumptions used. The next chapter humaent. resources ransrom 40 to8 takes each component of natural capital and pvercent. The share of natural capital alsoprsnstevlainmholgydpe. varies considerably (2 to 40 percent). The pts 3 valuandon methodology adopted. small share of natural capital in industrialized Chapters 3 and 4 deal wrth produced assets economies is deceptive and does not imply fna hap resentrespectively and the that natural capital is insignificant. The . c preponderance of human resources and conclusions and guidance for public policy. produced assets in high income countries Most of these results reinforce our intuitive masks the percentage share of natural capital. understanding of the development process. For example, although Canada has only 11 By explicitly accounting for the different percent of total wealth per capita in natural components of wealth, these estimates have capital, in dollar per capita terms it ranks in brought to the fore the concept of portfolio the top five. In contrast Figure 1 shows that management where a nation's portfolio low income economies that depend primarily consists of natural capital, produced assets, on revenues from the export of natural and human resources. It is hoped that this resource commodities have a much larger endeavor will first and foremost strengthen percentage share in natural capital (this does the emerging view of the importance of not include countries dependent on people and the environment by supporting petroleum revenues). Interestingly the intuition with numbers, and secondly, spur relative share of produced assets, the main improvements in the measurement of these focus of national planners in the past, shows critical components. Environmental Economics Series 3 2 Natural Capital The environmental resource base is we have primarily considered the value of comprised of regenerative resources- natural capital as resource inputs into atmosphere, animal, bird, plant, fish production: land as an input into agricultural populations, land (arable and grazing land), crops and animal husbandry, forests as a underground basins of water-and non- source of timber, and sub-soil assets as a renewable resources-oil, coal, natural gas, source of metals, minerals, and fossil fuels. metals, minerals. Economic development in We have also considered certain extractive both industrialized and developing and non-extractive values embodied in the economies, but particularly so in agrarian component of non-timber forest benefits. Also economies, relies crucially on environmental included is an estimate of the values people resources, and yet they "make but place on keeping open the option of such perfunctory appearances in government extractive and non-extractive uses in the planning models" (Dasgupta 1993). Natural future, as revealed in the establishment of capital, like any other asset, contributes a protected areas. Clearly, there are other flow of services to the economy. These environmental resources and types of values services can be direct contributions to missing from the calculus. economic activity via inputs (raw materials, energy) or goods and services for final Having identified environmental resources consumption. Capturing the economic value and types of values, we have used the of the latter and bringing it into mainstream concept of economic rent to place a value on economic analysis has been the raison d'etre natural capital. Economic rent is the return on of environmental economics. Services a commodity in excess of the minim um provided by the environment range from required to bring forth its services. Rental current values such as extractive uses (fish, value is therefore the difference between the pharmaceuticals), non-extractive uses market price and cost of production / (recreation, aesthetic), maintenance of life- extraction.' To ease international support systems (watershed protection, comparability, we have used international nutrient cycling), to future values (options market prices for all the components of and existence values). All of these contribute natural capital where rental values are used - to well-being, and it is this total economic agricultural crop and pasture land, timber, value of environmental goods (over and and sub-soil assets. above its value as resource inputs) that ought to be reflected in a measure of natural capital. The capital value of an asset is the present Several methods have been identified and value of the stream of services it generates used to monetize these different values. over its life time. This is the method we have used to calculate the stock of natural capital To capture the total economic value of all based on the above-mentioned streams of environmental goods is a huge endeavor and services. 4 Environment Department Papers Natural Capital Concerns about sustainable management of discount factor of 4 percent, because one of resources enter the calculation through the the objectives of this exercise is to enable choice of the time horizon over which present cross country comparisons. This discount rate value is computed. The extractive and non- is applied to all assets including human extractive uses of the environment and the resources. use of environmental resources as inputs into production can be enjoyed in a sustainable or We will now go into greater detail on the unsustainable manner. Since the calculation of each component of natural (un)sustainability of current use patterns capital and the data sources used. affects the stock of capital being left for the future, this ought to enter a measure of Agricultural Cropland natural capital. Thus, for renewable resources nandral non-renewablesresourcesethelcalculation The quality of land, not often thought of as a and non-renewable resources the calculation reeal reore ca emitandol mae adusmet fo (u)u, ial use. renewable resource, can be maintained only matters.The precise adjustments are by careful management. It is an important patterns,usse becise detare factor in generating well-being particularly in discussed below with the detailed countries where a large part of the population calculations for each component of natural is involved in agricultural and pastoral capital. activities. Indeed our results show that Discount Rates agricultural crop and pasture land account for a little over 80 percent of natural capital in In computing present values the choice of the low income economies (Figure 2). discount rate is critical to the calculation. There are several reasons suggested in the Country-level data on agricultural land prices literature for choosing a positive discount are not widely published, and so an rate such as social rate of return on estimation methodology for land asset values investment, pure time preference, and is required. Even if local price data were opportunity cost of capital (for a more available, it is arguable that land markets are detailed discussion see Pearce and Turner, often so distorted that meaningful 1990). The relevant discount rate for resource comparisons across countries would be allocation decisions over time is the social difficult. We have chosen therefore to rate of return on investment (SRRI). The SRRI estimate land values based on the present is defined as s = r + yc, where r is the pure discounted value of land rents, assuming the rate of time preference (people prefer their products of the land are sold at world prices. benefits now rather than later because they are impatient), y is the elasticity of marginal The return to land is computed as the utility of consumption, and c the rate of difference between the market value of the growth of real consumption per capita. output crops and crop-specific production Pearce and Ulph (1995) have estimated the costs. We have considered the production of SRRI for developed countries to be in the three major cereals-maize, wheat and rice- range of 2 to 4 percent. Clearly for fast and valued each at international prices. The growing developing economies this rate annual economic return to land is measured should be higher and for those experiencing as a percentage of crop value-34 percent for weak or declining per capita consumption maize, 31 percent for wheat, and 50 percent growth, the rate should be lower. for rice (World Bank 1992,1993a, 1995). These percentages are based on farm-level crop Although the SRRI varies from country to budgets from World Bank agricultural sector country, we have chosen to use a standard reports, and include the return to irrigation Environmental Economics Series 5 Estimating National Wealth- Methodology and Results Figure 2 Components of natural capital, 1994 High income economies Hgh income economies Nontimber Timber Nontimber benefits Pastures 3% benefits 4%Protected 4% 2% Timber areas Ptee 10% 11% Protected areas Pastures Su-ol2% 15% aSubsoil Sub-soil asset assets 19% 8% Agricultural Agricultural cropland cropland 80% 41% Source: Authors' estimates. but exclude return to other inputs such as been valued at 80 percent of the average per labor and fertilizers. Therefore, implicit in hectare return to the major cereals on the this calculation is the economic value of assumption that these crops will yield lower water in agricultural production. returns per hectare. While this is a reasonable assumption for coarse grains and tubers, this These crop-specific ratios of economic rent to is less accurate for higher value crops such as world prices are then multiplied by values of coffee, tea, rubber, and cocoa. Therefore, for production at world prices for the major countries that cultivate large amounts of cereal croplands in each country. This has the high-value non-grain crops this methodology effect of assigning higher land rents to more underestimates the value of land. productive soils, which makes perfect sense. However, applying average crop-specific In order to reflect the sustainability of current ratios in this manner probably understates cultivation ractices the annual return for the value of the most productive lands, and 1994aho practes, the annuased on overstates the value of the least productive. 1994 is projected to the year 2020 based on country-specific and crop-specific projected To calculate the return to land on which crops annual growth rates of cereal yields and area other than maize, wheat and rice are being under cultivation (Rosegrant, Agcaoili- grown we use a weighted average of the per Sombilla, and Perez, 1995). The growth rate hectare return to land under maize, wheat of production is therefore estimated and rice cultivation (weighted by the area as(l+Ga t)*(1+G t), where Gat is the growth sown to each crop in any individual country). rate of area under cultivation and Gyt is the Eighty percent of this weighted average is growth rate of yields. After 2020 the value of then used to value land on which other crops production is held constant to infinity. The are being grown (calculated by subtracting discounted present value of this flow was the area under maize, wheat, and rice from then calculated using a discount rate of 4 arable and permanent crop area). Land under percent to arrive at an estimate of the asset crops other than maize, wheat, and rice has value of agricultural land. 6 Enviromnent Department Papers Natural Capital Data sources. The percentage of crop value fluctuations in production. Data on attributable to land as an input is based on international prices (1994) are from the "Pink World Bank agriculture sector reports (World Sheet" published by the Commodity Prices Bank 1992, 1993a, 1995). Data on land areas and Analysis Unit of the World Bank. Milk is under cultivation of different crops is taken valued at $500 per metric ton. Area of pasture from the Food and Agricultural land (1994) is from BESD's FAO/Fertilizers Organization's (FAO) Production Yearbook. database. Data on crop production in metric tons were taken from BESD's' FAO-Production Timber Resources datab'ase. To smooth fluctuations in The predominant economic use of forests has production we use average production over been as a source of timber. The annual flow of the period 1990-94. Data on international roundwood production is valued using prices are from the "Pink Sheet" put out by timber rents (price minus average production the Commodity Prices and Analysis Unit of costs) and then capitalized using a 4 percent the World Bank. discount rate to arrive at a stock of timber resources. The concept of sustainable use of Pasture Land forest resources is introduced via the choice Returns to pasture land are assumed to be a of the time horizon over which the stream is fixed proportion of the value of output based capitalized. If roundwood production is on a study of sheep and bovine budgets greater than net annual increments, then the (World Bank, 1993a). On average, costs of time to exhaustion is calculated based on production are 55 percent of revenues, and estimates of forest volume divided by the therefore returns to pasture land are assumed difference between production and to be 45 percent of output value. Value of increment. Forest volume is calculated as 80 output is based on the production of milk, percent of forest area multiplied by an area- beef, mutton, lamb, and wool valued at to-volume factor for tropical and non-tropical international prices. As with croplands, this forests taken from Mather (1990). If, however, rental share of output values is applied to logging rates are below net annual country-specific outputs of pastureland increments, timber flows are capitalized over valued at world prices. The present value of an infinite time horizon. this flow is then calculated using a 4 percent Data sources. Roundwood production data discount rate over an infinite time horizon. (averaged over 1990-94) are taken from the In countries where there are significant feed- BESD/FAO-Forestry database. For further lot operations, we find that the apparent per details on the calculation of timber rents and hectare return to pasture land is greater than net annual increments see Appendix B. Forest the return to cropland. Lacking data on the area data for 1994 are taken from the World animal feed that 'subsidizes' the returns to Resources Institute database 1996-97. pasture land in these situations, we have Non-Timber Forest Benefits simply capped the maximum value of pasture land to be less or equal to that of cropland. Timber revenues are not the only contribution forests make. Non-timber forest benefits such Data sources. Production volumes (averaged as minor forest products, hunting and fishing, over 1990-94) for dairy and meat products are recreation, watershed regulation, options and taken from BESD's FAO/Production existence values are significant benefits not database. As with crop output, we use an explicitly accounted for. This fact leads to average over 1990-94 to smooth out undervaluing of forest resources and could Environmental Economnics Series 7 Estimating National Wealth Methodology and Results explain much of the difference in deforesta- capitalized over an infinite time horizon, tion rates in developed and developing using a 4 percent discount rate. countries. A comparison of non-timber forest benefits in developed and developing Data sources. Data on protected areas, IUCN countries reveals that returns per hectare per categories I-V for 1994 are from the World year from such benefits vary from $145 per Resources database 1996-97. hectare in developed countries to $112 per hectare in developing countries (Lampietti Sub-soil Assets and Dixon, 1995). Assuming that only a tenth Sub-soil assets-metals, minerals, oil, coal, of forest area is accessible, this per hectare and gas-form a large share of natural capital value is multiplied by one-tenth of the forest in oil-rich countries such as Saudi Arabia area in each country. Norway, and Mexico and mineral-rich countries such as South Africa. In the absence Data sources. Forest area data (1994) are taken of competitive markets for stocks of sub-soil from the World Resources Institute database assets, our approach to valuing these assets is 1996-97. to calculate present values of the economic Protected Areas profits on extraction (net operating surplus less a "normal" return on produced assets) Protected areas provide a number of benefits over the life of the resource deposit. that range from existence values to recreational values. They can be a significant Calculating the value of subsoil assels source of income from a thriving tourist requires some strong assumptions. Assuming industry. These values are revealed by a high a constant stream of resource rents, the value willingness to pay for such benefits. The of the stock of sub-soil assets, V(O), is establishment and good maintenance of calculated based on the following expression: protected areas preserves an asset for the future and therefore protected areas form an important part of the natural capital V(O)=Xoqo (1+_ (l+r)°, (I) estimates.r 1+r We have valued protected areas at the Here no is the total rent per unit of resource, q, opportunity cost of preservation-that is, the is the production of the exhaustible resource costs of demarcating these regions as smoothed over a 5 year period, r is the costs f demrcatig thes regins asdiscount rate, and T0 the time to exhaustion. protected areas are the foregone benefits from Epsson (1) is jT the presto valuetof convering tem topastue or gricuturalExpression1 (1) iS just the present valtue of a converThing them to pasture or agrcultural constant stream of total resource renl:s over a land. The willingness to pay to preserve fnt iehrzn finite time horizon. specific natural regions varies considerably, and there is no comprehensive data set on The unit total resource rent (i0) is the same as this. Limiting the value of protected areas to that used in the genuine savings estimates of the opportunity cost of preservation probably chapter 2 of Expanding the Measure of Wealth captures the minimum value, and not the (World Bank 1997b). It is calculated as the complete value, of protected areas. difference between the international price and the average cost of extraction for the year Area protected (IUCN categories I-V) is 1994 as detailed in Appendix B. The valued at the lower of per hectare returns to production of sub-soil assets is smoothed pasture land and cropland. This is then over a 5 year time period to eliminate sudden 8 Environment Department Papers Natural Capital fluctuations. In most cases it is smoothed q0 - over 1990-94, but for some countries data _qo =__ were not available for these years and an (1 +ry (3) average for the most recent five year period was used. Again, a discount rate of 4 percent Here S is the stock of proven recoverable is employed. reserves as reported in World Resources (World Resources Institute 1993). This is a Optimal extraction of a subsoil resource conservative measure of resource extent- requires the Hotelling rule to hold: unit more elaborate estimates of subsoil assets scarcity rents (price minus marginal cost of could include probable and possible reserves, extraction) must rise at a percentage rate but these would have to be adjusted to reflect equal to the discount rate. By adding one the probability of the resources being more assumption, an isoelastic extraction cost available in these quantities and the likely function with increasing marginal extraction extraction costs associated with them. costs, it is possible to define the path that the percentage change in unit total resource rents In some cases the margin between price and (it) must follow: average production costs is very thin, yielding a negative unit scarcity rent for the chosen value of the elasticity of the extraction * rp -clq cost function. For these cases we assume for C(q) = aq , e > 1. scarcity rents are equal to half of total rents -x e p-c / q and calculate the appropriate value for E from expression (2). In short hand: Denoting this growth rate of total rents as , p-£c/q = 0.5 * (p-c/q) it is easy to see that this rate varies over time => e = q/c * (p-x/2) as the ratio of scarcity rent (p - e c/q) to total and y = 3/4 * (r/e) rent (p - c/q) varies. However, under an optimal extraction program we know that In countries with large stocks of subsoil these two rent measures will be equal at the resources we find that the assumed constant point of exhaustion, and so a further rate of decline in the quantity extracted yields simplifying assumption is made: y is assumed no solution to expression (3). This is to be constant and equal to its average value equivalent to saying that the resource stock over the extraction program, so that, will not be exhausted in finite time. Therefore 1\'rr- in countries where T. tends to infinity we use l=I r r. p-ec! /q 1 the following equation to estimate the present 2 e £ p-cIqq ( value of the resource stock: In calculating expression (2) we assume that V(0) = itcq, (1+1/r). the elasticity of the extraction cost function e Where we have no data on reserves we equals 1.153. The assumption of constant assume the time to exhaustion This 20 years. resource rents over the extraction program In our estimates countries that are not implies that the quantity extracted, q, must extracting their reserves are assumed to have fall at the same percentage rate y as unit total a stock value of zero. rents are rising. Time to exhaustion T, is therefore calculated by solving the following A comparison of this valuation methodology equation: (based on the assumption of rising scarcity Environmental Economics Series 9 Estimating National Wealth: Methodology and Results rents and a constant revenue stream) with cropland (see above). It should be possible to others reveals that these estimates are greater value industrial and household uses of water than those from a simple present value by estimating economic values, based on approach (implicit in the method of El Serafy willingness to pay, for different uses. 1989) and lower than the values from the net However, finding consistent international price approach (as employed in Repetto and data sets with sufficient country coveragve is others 1989). Unlike these alternative problematic. Moreover, there are significant methods, our 'constant revenue' approach is conceptual issues associated with trying to consistent with optimal resource extraction assign a stock value to what is inherent]y a and the currently observed flat long term flow resource. trend in prices of sub-soil assets. Valuing fish resources is another clear gap in Data sourcesfor metals and minerals. our estimates of natural wealth. However, Production data on the following metals and both conceptual and practical issues stand in minerals - Bauxite, Copper, Iron Ore, Lead, the way. While the ownership of many inland Nickel, Phosphate Rock, Tin, and Zinc - are fisheries is fairly unambiguous, this is rnot the from BESD's METMIN database. Estimates of case for pelagic marine species. An overriding metals and minerals reserves (1990) are from issue is the fact that resource rents appear to the World Resources Database 1992-93. For have been driven to zero in many fisheries, a greater detail on the methodology used to clear indication of the mismanagement of estimate resource rents see Appendix B. these resources. Data sourcesfor crude oil, soft and hard coal, natural gas. Production data are from BESD's Endnotes UN Energy Statistics database. Estimates of proved recoverable oil, coal, and natural gas reserves are for 1993 and are taken from the 1. The difference between output measured at World Resources Database 1992-93. For world prices and production costs is, strictly greater detail on the methodology used to speaking, the total economic rent. The estimate rents see Appendix B. scarcity or 'Hotelling' rent is measure,1 as price minus marginal cost-see the The Unfinished Agenda discussion of subsoil assets below. for Natural Capital 2. BESD is the World Bank's Economic and Social Database that draws upon A prominent element for future consideration consolidated data sources of other is the explicit valuation of water resources. At international institutions such as the Food present the value of water as an input in Agriculture Organization, and International agricultural production (the predominant Energy Agency. utilization of water in most countries) enters 3. Vincent (1996) estimates this value for oil implicitly via the value of agricultural extraction in Malaysia. 10 Environment Department Papers 3 Produced Assets Produced assets (or physical capital) have rates, assumed asset lives and an initial been the focus of national economic planning assumption about the capital stock (based on for years. Physical capital was thought to be an assumed initial capital-output ratio). The the bottleneck to development hence estimates of physical capital stock for 1990 warranting high rates of capital accumulation are in constant 1987 local currency. These and a great emphasis on the optimum rate of have been converted to 1990 US dollars using accumulation. While no doubt an important 1987 exchange rates and the US GDP deflator. factor of production, physical capital is not The 1990 estimates are then extrapolated to necessarily the limiting factor and we 1994 by adding gross domestic investment in increasingly find that natural capital is taking the current year to the previous year's stock on this role. For instance, in the fisheries and subtracting depreciation. The sector it is the availability of fish and not depreciation series until 1990 have been fishing boats that is the problem. When we derived from the same physical capital stock look across different regions, the estimates database and then extrapolated to 1994 loo acos difrn regon , th esimte assuming that for a given country the show that the percentage share of physical proportion of depreciation in GNP is the capital in total wealth does not vary very pooto fdpeito nGPi h mch,ita raing froma wealh dofs 30t pveret isame as the average for the period 1980-90. mnuch, ranging from a high of 30 percent mivn Pacific OECD countries to a low of 15 percent To account for the fact that not all in East Asia and Central America. However, components of a country's phsical capital the variation in human resources is fary y stock are tradable, these are separated out as greater. follows. The non-tradable components, namely structures and urban land, are valued The effective use of physical capital itself is using puchasing power parit (PPP) depedentupo humn caita. Ifther isusing purchasing power parity (PPP) dependent upon human capital. If there is exchange rates while tradable components under-investment in human capital the rate at are valued at nominal exchange rates. which additional physical capital can be productively utilized will be limited. Structures = s * produced assets * (PPP/ nominal exchange rate) The estimates of produced assets include buildings and structures, machinery and Machinery and equipment = (1-s) * produced equipment, and urban land (as a natural asset assets whose value is closely related to the value of produced assets), and are based on physical Urban land = s * produced assets * (PPP/ capital stock estimates taken from Nehru and nominal exchange rate) * u Dhareshwar (1993). Their database of physical capital stock is created by a where, s (share of structures in total produced perpetual inventory method from investment assets) and u (ratio of the value of urban land Environ}mental Economics Series 11 Estimating National Wealth Methodology and Results to the value of structures) are assumed to be detailed national balance sheet information 72 percent and 33 percent respectively. for the Canadian economy (Statistics Canada, 1985). On average these balance sheet Data sources. Physical capital stock estimates accounts show structures accounting for for 1990 were provided by the Development roughly 72 percent of total produced assets, Data Division of the International Economics while urban land in turn is 33 percent of Department of the World Bank. The GDP structures. While this introduces a likely bias, Department ofteWrdBn.TeGPCanada being a land-rich country, the deflator is taken from BESD's International Canada balancersh counts are Canadian balance sheet accounts are among Financial Statistics database. Data on gross the most detailed in the world. PPP exchange domestic investment in current US dollars are rates are from the Development Data from BESD's World Bank National Accounts Division of the International Economics database. Values for s and u are based on Department of the World Bank. 12 Environment Department Papers 4 Human Resources Estimating the asset value of the return to population; and (iii) the value of produced human resources is the most difficult and assets is subtracted from this present value of contentious aspect of the wealth estimates. non-resource NNP - the residual is defined We use the term 'human resources' as distinct to be the value of human resources. from 'human capital' because the latter is generally considered to be the return to These calculations can be represented by the education. Aiming to be more inclusive, the following equation system: wealth estimates place a value on the returns to education, raw labor, and 'social capital' R = Present value (agricultural wages + non- (the value that is added by institutions and agricultural GNP - sub-soil rents - depreciation) other social structures-for a more comprehensive treatment of the concept of Kp = stock value of structures, and machinery social capital the reader is referred to chapter and equipment 6 of Expanding the Measure of Wealth (World Human Resources = (R - Kp - urban land) Bank 1997b)). (PPP/nominal exchange rate) While 'development' was at one time The share of wages in agricultural value synonymous with accumulation of produced added is calculated to be 45 percent, based on assets, recent thinking has emphasized the information on crop budgets from importance of human resources in the agricultural sector reports (World Bank, development process. Evidence suggests that 1992). Data on agricultural and non- expenditures on education, training, and agricultural shares of GDP and GNP are from health contribute to development outcomes. World Development Indicators (World Bank, Moreover, such expenditures yield a 1996). Because 'agricultural GDP' in these sustained return in the future. Improvement data includes forestry and fishing, the only in the quality of the human factor of resource rents remaining to be subtracted production is at least as important as from GNP are those for subsoil assets. investment in physical capital. For purposes of taking the present value, We have used a residual approach, similar to remaining years of productive life are pioneering efforts in Norway (Central Bureau calculated as: age 65 or life expectancy at age of Statistics of Norway, 1992), in estimating 1 (whichever is lower) minus mean (average) human resources. The calculation proceeds in age of the population. Mean age is calculated three stages: (i) returns to natural resources using the detailed age distribution of the are stripped out of NNP; (ii) the present value population. The lower of age 65 or the life of non-resource NNP is taken over the mean expectancy at age 1 is taken on the remaining years of productive life of the assumption that beyond 65 individuals are no Environmental Economics Series 13 Estimating National Wealth Methodology and Results longer working and producing. For example, The residual approach to valuing human in India life expectancy at age 1 is 61 years resources is biased to an extent, however, and the average age of the population is 27 because some of the net flows of wages and years, yielding remaining years of productive property income from abroad in national activity of 34 years. In Sweden life expectancy product should properly be attributed to at age 1 is 78 and the mean age of the produced assets. For countries where there is population is 40. Using an upper bound value a net outflow of wages and property income of 65 to reflect productive life rather than life to abroad, therefore, the value of produced expectancy, yields a time horizon of 25 years. assets is overstated and the value of hurnan Note that we take life expectancy at age 1 and resources is understated in the wealth not at birth because in many developing calculations. countries the highest risk of death is in the first year. Because the residual approach to valuing human resources stacks so many assumptions Recall that the non-tradable portions of on top of one another, it is important to produced assets-buildings and structures- corroborate the resulting valuations. Human were valued at purchasing power parity capital is undoubtedly the largest component (PPP) rather than nominal exchange rates. of the value of human resources. If the Human resources are similarly valued at PPP residual approach makes sense, therefore, the exchange rates, in an effort to reflect the resulting values for human resources should consumption possibilities presented by the correlate well with independent measures of returns to human resources. the quantities of human capital in different countries. To test this, Figure 3 plots mean The wealth calculations begin with net years of education per capita in a sample of national product (GNP minus depreciation of countries on the x-axis against the estirmated produced assets) because the question of value of human resources per capita on the y- ownership is important in wealth accounting. axis. Figure 3 Human resources and years of education (thousand dollars per capita) 325 300 275 250 ,, 225 2200 ° 175 a) c150 E 125 100 75*. 25 * *$* * 0 2 4 6 8 10 12 14 Mean years of educational attainnient Source: Authors' estimates. 14 Environment Department Papers 5 Caveats While the foregoing description has been where there is infrastructure to support it, careful to highlight many of the intricacies and not over the whole extent of the forest, and controversial aspects of the wealth our methods value only the economic margin estimation methodology, it is worth re- of the forest. Moreover, as documented emphasizing the key points and adding some above, only 10% of the forested area is new caveats. As noted in the Preface, the assumed to provide non-timber forest divergences between best practice and the benefits. As a result, the largest forested methods actually applied in estimation are countries (Brazil, Indonesia, Malaysia and dictated by inadequacies in the data. Canada are the obvious examples) have some portion of their forest resource valued at zero. The first point to note is that, while arbitrary To the extent that we are valuing commercial assumptions have been kept to a minimum, resources in this exercise, this is not a there are still a few 'expert judgements' that problem. But sequestering carbon, harboring were used. Non-cereal croplands are assumed biodiversity and regulating flow in to have 80% of the average value of land watersheds all have economic value as well, under cereals. Ten percent of forested land is and to that extent the wealth of these assumed to be accessible to provide non- countries is underestimated by our timber forest benefits. For subsoil resources, methodology. either the elasticity of the cost function was assumed to be 1.15, or scarcity rents were Cases of under-valuation of natural resources assumed to be half of total rents. Where the are most obvious for protected areas. While data show production of subsoil resources the opportunity cost approach applied is but no corresponding reserve estimates could probably quite adequate for the most remote be found, the reserve to production ratio was protected areas, the more accessible areas are assumed to be 20 years. The latter two certainly undervalued. Willingness to pay to assumptions in particular are fairly visit the most popular natural parks, even in innocuous-the assumed proportionality of developing countries, is substantial and is not scarcity to total rents only applies when captured by our methodology. rental margins are particularly thin, and missing reserves data only occurred for very The estimates of total wealth are pessimistic minor producers of resources. to the extent that existing under- and unemployment of production factors is The lacunae, notably water and fishery extrapolated into the future. Where capable resources, have already been highlighted. For human resources have been idled by bad countries with extremely large forest policies, therefore, their potential value is not resources there is another type of missing reflected in the wealth estimates. This is an element. Because timber harvest occurs only inherent byproduct of 'pulling apart' existing Enviromnental Economics Series 15 Estimating National Wealth: Methodology and Results GNP figures as the basis for calculating total the most overvalued exchange rates and wealth. inflated price levels end up appearing to be the richest nations on Earth. The use of purchasing power parities (PPP) to value both non-traded assets (buildings and As the foregoing indicates, there are reasons structures) and human resources is at one to believe that we have undervalued nature level simply following the logic of and natural resources in the methods applied. international economic comparisons, by Many of the limitations highlighted valuing relatively non-mobile production disappear when truly local data are available. factors at local prices. However, the resulting Working at the country level to overcomLe mixing of exchange-rate based and PPP- these limitations seems a better way forward based valuations is not particularly satisfying than by introducing more arbitrary from a theoretical point of view. That point assumptions and extrapolating from being granted, the alternative of using only extremely shaky foundations in order tc, exchange-rate based valuations gives results arrive at larger numbers for natural, and that violate our intuitions: the countries with national, wealth. 16 Environment Department Papers 6 Conclusions and Policy Implications The process by which nations combine and termed 'social capital'. Building people's employ their assets to generate well-being is capabilities through education and health can a complex one. Unquestionably, human enhance a nation's human resources and help resources and natural capital are major in realizing a sustainable path to contributors to this process. The wealth development. We need to learn more about estimates provide many noteworthy findings building social capital; at a minimum we and policy messages. must be careful not to destroy it through poor policies. First, agricultural crop and pasture land are a very important component of natural capital Third, for countries rich in sub-soil assets the in all income groups, with a relative share of importance of investing, rather than 50 percent or more. The implication for policy consuming, returns from extraction of oil, making is clearly the overriding importance minerals, coal, gas and other exhaustible of judicious management of land resources. resources needs to be stressed. Unsustainable cultivation practices could imperil the agricultural production system In conclusion, our analysis suggests the need itself and its supporting ecosystems, for a more holistic approach to development effectively curtailing opportunities for future planning, an approach that places due generations. This is particularly true of low emphasis on all the different components of income countries where the relative share of wealth. While investments in and agricultural cropland and pasture land is as maintenance of infrastructure are important, high as 84 percent of total wealth. this is equally the case for agricultural land or people. Economies have available to them an Second, human resources - the return to initial endowment of natural resources, raw education, raw labor, and to social capital - labor and social capital. This initial dominate the portfolio overall. This yields an endowment together with investments in optimistic message for policy-making. A produced assets and human capital form the nation's wealth lies predominantly in its foundation of the development process. The people and the amalgam of individual and sustainability of this process relies crucially institutional relationships that we have on sound management. Environmental Economics Series 17 Appendix A Estimates of National Wealth: Total and Components Table Al. Estimates of national wealth: Total and components Wealth Country Total wealth Human Natural Produced Human Natural Produced rank resources capital assets resources capital assets dollars per capita percent share of total wealth 24 Argentina 147,000 124,000 10,000 13,000 84 7 9 6 Australia 297,000 195,000 35,000 67,000 66 12 23 10 Austria 286,000 201,000 8,000 78,000 70 3 27 70 Bangladesh 22,000 17,000 3,000 2,000 76 14 10 9 Belgium 287,000 225,000 2,000 62,000 79 1 21 67 Benin 25,000 19,000 2,000 4,000 76 8 16 57 Bolivia 36,000 21,000 6,000 9,000 59 17 25 35 Botswana 89,000 68,000 6,000 15,000 76 6 17 35 Brazil 89,000 66,000 7,000 16,000 74 8 18 84 Burkina Faso 14,000 10,000 2,000 2,000 68 17 15 88 Burundi 10,000 6,000 2,000 2,000 58 20 22 60 Cameroon 32,000 17,000 7,000 8,000 53 21 26 3 Canada 331,000 227,000 37,000 67,000 69 11 20 71 CAR* 21,000 12,000 6,000 3,000 55 30 15 81 Chad 15,000 8,000 6,000 2,000 51 37 12 23 Chile 148,000 116,000 14,000 17,000 79 10 12 56 China 37,000 28,000 3,000 6,000 77 7 16 37 Colombia 85,000 67,000 6,000 12,000 79 7 14 62 Congo 31,000 17,000 4,000 9,000 55 14 30 34 Costa Rica 96,000 73,000 8,000 15,000 77 8 15 71 Cote d'lvoire 21,000 11,000 4,000 6,000 52 18 30 8 Denmark 295,000 213,000 11,000 71,000 72 4 24 42 Dom. Republic 68,000 51,000 8,000 8,000 76 12 12 43 Ecuador 67,000 41,000 11,000 14,000 61 17 22 49 Egypt 52,000 33,000 2,000 16,000 64 5 31 54 El Salvador 40,000 35,000 1,000 5,000 86 3 12 17 Finland 241,000 135,000 16,000 90,000 56 7 37 6 France 297,000 218,000 8,000 70,000 74 3 24 74 Gambia, The 18,000 13,000 2,000 3,000 73 12 15 11 Germany 281,000 211,000 4,000 66,000 75 1 23 65 Ghana 27,000 21,000 2,000 4,000 78 7 15 25 Greece 142,000 106,000 5,000 31,000 75 4 22 49 Guatemala 52,000 43,000 2,000 7,000 84 3 13 74 Guinea-Bissau 18,000 8,000 8,000 2,000 43 44 13 85 Haiti 13,000 9,000 1,000 3,000 70 7 23 58 Honduras 34,000 23,000 3,000 8,000 66 10 24 Environmental Economics Series 19 Estimating National Wealth: Methodology and Results Table Al. Estimates of national wealth: Total and components (continued) Wealth Country Total wealth Human Natural Produced Human Natural Produced rank resources capital assets resources capital assets dollars per capita percent share of total wealth 73 India 20,000 12,000 4,000 4,000 58 20 22 46 Indonesia 60,000 45,000 7,000 8,000 75 12 13 18 Ireland 219,000 162,000 18,000 39,000 74 8 18 16 Italy 257,000 187,000 3,000 67,000 73 1 26 52 Jamaica 45,000 22,000 3,000 20,000 49 7 44 4 Japan 304,000 208,000 2,000 94,000 68 1 31 44 Jordan 64,000 48,000 1,000 16,000 74 2 24 74 Kenya 18,000 9,000 2,000 7,000 51 9 39 22 Korea, Republic Of 168,000 138,000 3,000 27,000 82 2 16 64 Lesotho 28,000 22,000 1,000 5,000 80 3 17 79 Madagascar 16,000 8,000 7,000 1,000 49 42 9 91 Malawi 7,000 5,000 1,000 2,000 61 12 28 26 Malaysia 137,000 101,000 12,000 25,000 73 9 18 86 Mali 12,000 5,000 5,000 2,000 43 41 15 68 Mauritania 24,000 14,000 5,000 4,000 60 22 18 32 Mauritius 99,000 80,000 1,000 18,000 80 1 18 30 Mexico 113,000 87,000 7,000 19,000 77 6 17 48 Morocco 54,000 42,000 2,000 9,000 78 4 18 89 Mozambique 9,000 5,000 1,000 3,000 53 13 35 41 Narnibia 71,000 54,000 7,000 10,000 76 10 14 79 Nepal 16,000 11,000 3,000 2,000 68 18 15 13 Netherlands 272,000 196,000 4,000 71,000 72 2 26 12 New Zealand 277,000 162,000 51,000 63,000 59 18 23 65 Nicaragua 27,000 19,000 4,000 4,000 71 14 15 69 Niger 23,000 8,000 12,000 2,000 36 54 10 5 Norway 302,000 172,000 30,000 99,000 57 10 33 58 Pakistan 34,000 28,000 2,000 4,000 83 6 12 33 Panama 97,000 75,000 6,000 16,000 77 6 17 55 Papua New Guinea 39,000 25,000 7,000 6,000 64 19 17 45 Paraguay 61,000 43,000 7,000 10,000 72 12 17 47 Peru 59,000 40,000 5,000 15,000 67 8 25 53 Philippines 44,000 35,000 3,000 7,000 79 6 15 20 Portugal 175,000 137,000 4,000 34,000 78 2 19 92 Rwanda 5,000 2,000 1,000 2,000 39 22 39 21 Saudi Arabia 171,000 69,000 72,000 30,000 40 42 18 60 Senegal 32,000 22,000 5,000 4,000 70 17 13 87 Sierra Leone 11,000 6,000 3,000 2,000 58 28 14 38 South Africa 83,000 62,000 4,000 17,000 75 5 20 19 Spain 201,000 152,000 6,000 43,000 76 3 22 51 SriLanka 47,000 36,000 3,000 8,000 76 7 17 15 Sweden 260,000 176,000 15,000 70,000 68 6 27 2 Switzerland 352,000 237,000 3,000 111,000 68 1 32 90 Tanzania 8,000 2,000 2,000 4,000 21 27 52 29 Thailand 117,000 93,000 8,000 17,000 79 6 14 74 Togo 18,000 11,000 3,000 4,000 64 15 21 27 Trinidad &Tobago 128,000 77,000 12,000 39,000 60 9 30 20 Environment Department Papers Appendix A Estimates of National Wealth: Total and Components Table Al. Estimates of national wealth: Total and components (continued) Wealth Country Total wealth Human Natural Produced Human Natural Produced rank resources capital assets resources capital assets dollars per capita percent share of total wealth 39 Tunisia 81,000 58,000 6,000 17,000 71 8 21 40 Turkey 79,000 63,000 4,000 11,000 81 5 14 81 Uganda 15,000 8,000 2,000 6,000 49 15 37 14 United Kingdom 266,000 209,000 5,000 51,000 79 2 19 1 United States 401,000 308,000 17,000 76,000 77 4 19 28 Uruguay 127,000 99,000 15,000 13,000 78 12 11 31 Venezuela 110,000 57,000 21,000 32,000 52 19 29 74 Vietnam 18,000 12,000 4,000 2,000 68 22 10 81 Zambia 15,000 5,000 5,000 4,000 38 38 25 63 Zimbabwe 30,000 17,000 3,000 10,000 59 8 33 Note: Estimates for Eastern Europe and countries of the former Soviet Union are not included because of uncertainty about data quality. Similar problems exist for Nigeria and Algeria. * CAR is Central African Republic Environmental Economics Series 21 Estimating National Wealth: Methodology and Results Table A2. Natural capital estimates, 1994, total and components Timber Non-timber Natural capital Pasture land Crop land resources forest resources Protected areas Subsoil assets $ per capira (percent of total) Argentina 9,850 3,270 5,200 280 480 100 520 (33) (53) (3) (5) (1) (5) Australia 35,340 7,270 14,150 1,030 2,150 1,650 9,080 (21) (40) (3) (6) (5) (26) Austria 7,570 1,480 2,410 1,720 150 1,580 230 (20) (32) (23) (2) (21) (3) Bangladesh 3,110 60 3,000 0 0 10 20 (2) (97) (0) (0) (0) (1) Belgium 1,750 470 1,110 100 20 50 10 (27) (63) (6) (1) (3) (1) Benin 1,930 70 1,030 440 250 120 10 (4) (54) (23) (13) (6) (1) Bolivia 6,060 690 2,520 160 1,820 240 640 (11) (42) (3) (30) (4) (11) Botswana 5,620 1,180 260 420 2,700 490 570 (21) (5) (8) (48) (9) (10) Brazil 7,060 1,070 2,740 1,200 960 190 910 (15) (39) (17) (14) (3) (13) Burkina Faso 2,400 210 1,870 100 120 90 (9) (78) (4) (5) (4) Burundi 1,940 90 1,820 10 10 10 0 (5) (94) (0) (1) (0) (0) Cameroon 6,800 270 4,840 650 430 270 340 (4) (71) (10) (6) (4) (5) Canada 36,590 2,310 9,910 6,230 4,560 6,830 6,750 (6) (27) (17) (12) (19) (18) Central 6,470 440 2,010 520 2,600 900 African Republic (7) (31) (8) (40) (14) Chad 5,550 470 4,110 340 500 120 (9) (74) (6) (9) (2) Chile 14,440 1,100 4,910 1,560 180 1,110 5,580 (8) (34) (11) (1) (8) (39) China 2,670 100 2,010 90 30 10 420 (4) (75) (3) (1) (1) (16) Colombia 6,100 1,160 2,490 390 410 270 1,380 (19) (41) (6) (7) (4) (23) Congo 4,420 20 200 1,040 2,200 0 960 (1) (4) (24) (50) (0) (22) Costa Rica 7,860 1,480 5,690 180 100 410 (19) (72) (2) (1) (5) 22 Environment Department Papers Appendix A Estimates of National Wealdtt Total and Components Table A2. Natural capital estimates, 1994, total and components (continued) Timber Non-timber Natural capital Pasture land Crop land resources forest resources Protected areas Subsoil assets $per capita (percent of total) Cote dlvoire 3,790 80 2,870 570 210 10 30 (2) (76) (15) (6) (0) (1) Denmark 11.070 270 7,210 380 30 1,930 1,260 (2) (65) (3) (0) (17) (11) Dominican 8,380 560- 7,310 90 30 280 100 Republic (7) (87) (1) (0) (3) (1) Ecuador 11,330 1,160 4,880 440 270 2,610 1,970 (10) (43) (4) (2) (23) (17) Egypt 2,360 420 1,540 0 0 70 330 (18) (65) (0) (0) (3) (14) El Salvador 1,150 250 890 10 10 0 (22) (77) (1) (0) (0) Finland 15,930 90 4,670 6,970 1,660 2,420 110 (l) (29) (44) (10) (15) (1) France 8,120 1,350 5,210 700 90 700 60 (17) (64) (9) (1) (9) (1) Gambia, The 2,120 190 1,850 10 20 50 (9) (87) (1) (1) (2) Germany 4,150 430 2,100 490 30 750 350 (10) (51) (12) (1) (18) (8) Ghana 1,920 60 1,510 190 150 10 10 (3) (78) (10) (8) (1) (1) Greece 5,210 1,490 3,080 170 90 60 320 (29) (59) (3) (2) (1) (6) Guatemala 1,720 300 930 170 110 150 60 (18) (54) (10) (6) (9) (4) Guinea-Bissau 7,970 200 7,440 330 (2) (93) (4) Haiti 840 110 720 0 0 0 0 (13) (86) (0) (0) (0) (0) Honduras 3,380 410 1,610 820 210 230 100 (12) (47) (24) (6) (7) (3) India 3,910 90 3,440 50 20 110 210 (2) (88) (1) (0) (3) (5) Indonesia 7,480 60 5,780 720 150 100 670 (1) (77) (10) (2) (1) (9) Ireland 17,780 11,770 4,810 510 40 120 530 (66) (27) (3) (0) (1) (3) Italy 3,400 430 2,430 110 40 230 160 (13) (71) (3) (1) (7) (5) Environrmental Economics Series 23 Estimating National Wealth: Methodology and Results Table A2. Natural capital estimates, 1994, total and components (continued) Timber Non-timber Natural capital Pasture land Crop land resources forest resources Protected areas Subsoil assets $ per capita (percent of total) Jamaica 3,080 110 280 50 10 0 2,630 (4) (9) (2) (0) (0) (B5) Japan 2,300 120 1,360 220 70 490 40 (5) (59) (10) (3) (21) (2) Jordan 1,020 260 360 0 0 100 300 (26) (35) (0) (0) (9) (29) Kenya 1,730 740 840 10 10 120 0 (43) (49) (1) (1) (7) (,0) Korea, 2,940 50 2,290 120 40 390 50 Republic Of (2) (78) (4) (1) (13) (2) Lesotho 940 340 600 0 0 0 (36) (64) (0) (0) (0) Madagascar 6,510 500 5,350 310 330 20 (8) (82) (5) (5) (0) Malawi 880 60 600 90 80 40 (7) (68) (11) (10) (4) Malaysia 11,820 20 6,190 1,310 230 840 3,230 (0) (52) (11) (2) (7) (27) Mali 4,840 530 3,620 270 340 70 (I1l) (75) (6) (7) (1) . Mauritania 5,100 1,060 2,270 0 70 50 1,640 (21) (45) (0) (1) (1) (32) Mauritius 1,240 20 1,180 10 10 10 (2) (95) (1) (1) (1) Mexico 6,630 810 1,520 200 140 110 3,860 (12) (23) (3) (2) (2) ( 58) Morocco 2,210 480 1,480 60 100 10 80 (22) (67) (3) (5) (0) (4) Mozambique 1,130 90 360 400 280 0 0 (8) (32) (35) (25) (0) (0) Namibia 7,180 1,400 1,230 2,310 380 1,B60 (20) (17) (32) (5) (26) Nepal 2,900 380 2,150 90 60 210 10 (13) (74) (3) (2) (7) (0) Netherlands 4,140 560 1,020 80 10 230 2,250 (14) (25) (2) (0) (6) (54) New Zealand 51,090 22,130 12,600 4,340 770 9,950 1,300 (43) (25) (9) (1) (19) (3) 24 Environment Department Papers Appendix A Estimates of National Wealth: Total and Components Table A2. Natural capital estimates, 1994, total and components (continued) Timber Non-timber Natural capital Pasture land Crop land resources forest resources Protected areas Subsoil assets $ per capita (percent of total) Nicaragua 3,690 540 2,110 580 360 90 0 (15) (57) (16) (10) (2) (0) Niger 12,340 310 11,600 50 80 300 0 (3) (94) (0) (1) (2) (0) Norway 30,220 110 1,680 2,520 700 5,110 20,090 (0) (6) (8) (2) (17) (66) Pakistan 1,880 140 1,480 0 0 100 150 (7) (79) (0) (0) (6) (8) Panama 6,300 930 3,960 270 310 830 (15) (63) (4) (5) (13) Papua New 7,490 10 560 1,550 2,370 20 2,980 Guinea (0) (7) (21) (32) (0) (40) Paraguay 6,990 1,490 3,590 1,150 650 100 (21) (51) (16) (9) (1) Peru 4,630 350 2,770 220 800 50 430 (8) (60) (5) (17) (1) (9) Philippines 2,730 50 2,400 140 30 30 80 (2) (88) (5) (1) (1) (3) Portugal 4,040 280 2,140 1,140 110 190 190 (7) (53) (28) (3) (5) (5) Rwanda 1,110 100 930 0 10 70 0 (9) (84) (0) (1) (6) (0) Saudi Arabia 71,880 330 3,600 20 20 67,910 (0) (5) (0) (0) (94) Senegal 5,300 290 4,180 310 250 210 60 (6) (79) (6) (5) (4) (1) Sierra Leone 3,040 60 2,570 180 110 0 120 (2) (84) (6) (4) (0) (4) South Africa 4,200 880 1,790 90 30 80 1,340 (21) (43) (2) (1) (2) (32) Spain 5,740 940 3,690 430 140 390 140 (16) (64) (8) (3) (7) (3) Sri Lanka 3,480 140 2,970 90 30 250 0 (4) (85) (3) (1) (7) (0) Sweden 14,590 440 4,390 5,890 1,160 2,300 410 (3) (30) (40) (8) (16) (3) Environmental Economics Series 25 Estimating National Wealth: Methodology and Results Table A2. Natural capital estimates, 1994, total and components (continued) Timber Non-timber Natural capital Pasture land Crop land resources forest resources Protected areas Subsoil assets $ per capita (percent of total) Switzerland 3,050 950 820 600 50 620 0 (31) (27) (20) (2) (20) (0) Tanzania 2,200 310 920 530 310 120 0 (14) (42) (24) (14) (6) (0) Thailand 7,600 110 6,270 110 50 980 80 (1) (83) (1) (1) (13) (1) Togo 2,670 50 2,250 0 90 170 120 (2) (84) (0) (3) (6) (4) Tnnidad And 12,110 70 2,540 40 40 100 9,310 Tobago (1) (21) (0) (0) (1) (77) Tunisia 6,370 550 5,070 10 20 10 710 (9) (80) (0) (0) (0) (11) Turkey 3,940 490 2,950 170 90 40 200 (12) (75) (4) (2) (1) (5) Uganda 2,230 120 1,680 210 90 130 0 (5) (75) (9) (4) (6) (0) United 4,940 1,540 1,820 110 20 710 730 Kingdom (31) (37) (2) (0) (14) (15) United States 16,500 2,570 7,210 1,730 410 1,400 3,180 (16) (44) (10) (2) (8) (19) Uruguay 14,810 6,040 8,530 160 60 10 (41) (58) (1) (0) (0) Venezuela 20,820 860 3,130 40 570 1,270 14,960 (4) (15) (0) (3) (6) (72) Viet Nam 3,990 70 3,490 70 30 260 70 (2) (87) (2) (1) (7) (2) Zambia 5,490 160 3,330 660 940 30 360 (3) (61) (12) (17) (1) (7) Zimbabwe 2,520 450 990 400 220 280 170 (18) (39) (16) (9) (11) (7) Notes: 0 means less than 10 dollars per capita. (0) means less.than I%. .. means no data. Estimates for Eastern Europe and countries of the former Soviet Union are not included because of uncertainty about data quality. 26 Environment Department Papers Appendix B Estimating Rental Values for Timber and Sub-Soil Assets What follows is a detailed account of Tobago, Argentina, Bolivia, Brazil, Chile and methodology for calculating the rental value Colombia (IADB, 1981); Iran, Iraq, Saudi of sub-soil assets (oil, natural gas, metals and Arabia, Kuwait, United Arab Emirates, Oman minerals, and coal) and timber. It documents (IEA, 1995b; Jenkins, 1989); Indonesia (IEA, all assumptions made, operations performed, 1995b; Repetto et al., 1989); Canada (IEA, and bibliographical sources utilized. These 1995b; Smith, 1992); and Europe (EIA, 1995). rental values were used in the wealth estimates and in the genuine savings Because production cost data were frequently estimates (Chapter 2, World Bank, 1997). available for a single year only, one of two methods was used to obtain year-by-year Rent from Oil estimates of production cost: 1) If data were As in the case of all other nonrenewable available for a single year only, it was resources, rent was estimated as: assumed that production costs remained constant in real terms. Production costs for Rent = (Production Volume) ( International each year in current dollars were obtained Market Price - Average Unit Production Cost) from the single data point and a times series GDP deflator (obtained from BESD, "WB-IEC In other words, rent equals volume produced Data" Database, "NY.GDP.MKTP.XU.E" by a particular country in a particular year indicator). 2) if data for two different years times the unit rent for that country in that were available with an interval of no data, year. Production volume data for 1970-1992 estimates for the intervening years were were taken from the Bank Economic and calculated as a linear interpolation between Social Database, or BESD, an internal on-line the two points. Those countries for which no database of the World Bank Group ("UN- production cost data were available were Energy Statistics" database, "CR Crude assigned a surrogate production cost from Petroleum" indicator, "production volume" another country. The choice of surrogate transaction). Production volume data for country was made on the basis of 1) 1993-1994 were obtained from West (1996). geographic proximity and 2) similarity International price data came from UNCTAD between the ratios of offshore active drilling (1989; 1993; 1996). Average production cost rigs to total active drilling rigs between the data were taken from from, by country: two countries. The numbers of active Ukraine (IEA, 1996a); Russia (IEA, 1994a; IEA offshore and total drilling rigs were obtained 1995a; IEA 1995b; Sagers, 1995); Venezuela from Meyer et al. (1994), and selected and Mexico (IEA 1995b; IADB 1981); Libya, statistics on onshore vs. offshore production Malaysia, Nigeria, USA, Gabon, Egypt, North came from Whitehead (1983). Table B1 shows Sea/Great Britain (IEA, 1995b); Norway the resulting assignments of surrogate (Adelman, 1987); Ecuador, Peru, Trinidad & countries. Environmental Econornics Series 27 Estimating National Wealtih Methodology and Results Additional, general references on accounting 1996); and export from Algeria to Italy, for the depletion of oil reserves include France, Belgium, and Spain (McKeough, Dienes et al. (1994) and Stauffer (1984; 1986). 1989). Due to the similar historical pattern in Useful conversion factors came from the first all of these prices, a world price for natural page of Blackwell Energy Research (1996). gas was estimated as the average of all available prices in this set in any givren year. Rent from Natural Gas Since this data set only extended back in time The most difficult aspect of the natural gas to 1973, some method of estimating the price calculation was that natural gas, unlike crude on the period 1970-72 was required. Data oil, has no single de facto world price. Since were available for the US internal price for the object of correcting saving rates is to natural gas from 1970-1995 (Streifel, 1996), measure opportunities foregone by current and when the international price on the extraction, some estimate of this opportunity period 1973-1995 was compared to the US cost was required. Data on natural gas internal price on the same period, the two export prices from various countries were were found to follow each other so closely collected, and found to collectively follow a that the international price could be very similar historical trend. Figure B1 estimated as a linear function of the US presents annual average export price data internal price with R2 = 0.90218. This same from the United States and Europe (Streifel, linear function was then used to estimate the 1996; Meyer, 1994); the Netherlands, Norway international price for 1970-1972 based on the (IEA, 1996b); Canada (Tiratsoo, 1979, US internal price for the same periocl. This McKeough, 1989, Government of Canada, function was: Table B1. "Surrogate" countries used to estimate oil production costs Country ISurrogate Country I Surrogate Country I Surrogate Country I Surrogoate Count r I Surrogate AGO GAB COG GAB HUN EU NOR NOR TKM RUS ALB EU COL VEN IDN IDN NZL USA TTO TTO ARE ARE CSK EU IND IDN OMN OMN TUN LBY ARG ARG CUB TTO IRN IRN PAK IRN TUR IRN AUS USA CZE EU IRQ IRQ PER PER UKR UKR AUT EU DDR EU ISR SAU PHL IDN USA USA AZE RUS DEU EU ITA EU POL UKR UZB RUS BEN GAB DFA EU JOR SAU PNG IDN VEN VEN BGD MYS DNYK NOR JPN EU QAT OMN VNM: IDN BGR UKR DZA LBY KAZ RUS ROM UKR YAR OMN BHR SAU ECU ECU KGZ RUS RUS RUS YEM OMN BLR UKR EGY EGY KWT KWT SAU SAU YMD OMN BOL BOL ESP EU LBY LBY SUN RUS YSR EU BRA VEN FRA EU LTU RUS SUR VEN YUG EU BRB TTO GAB GAB MAR LBY SVK EU ZAR GAB BRN GAB GBR GBR MEX MEX SVN EU CAN CAN GEO RUS MMR IDN SWE NOR CHL ARG GHA NGA MNG RUS SWK MYS CHN RUS GRC EU MYS MYS SYR SAU CIV NGA GTM MEX NGA NGA THA IDN CMR NGA HRV EU NLD EU TJK RUS 28 Environment Department Papers Appedix B Estimating Rental Values for Timber and Sub-Soil Assets Figure Bi A comparison of several natural gas export prices from around the world Natural Gas Export Prices, Comparison 4.50 4.300X K 3 3.50- 3.00 j * mO 2.50 - E 2.00 XU ~1.50 X1 1.00 0.50 0.00 I l 1970 1975 1980 1985 1990 1995 Year International price = (1.2436) (US internal price) country whose production cost would be + 0.53245, in $/MMBTU. used instead. These assignments, as before, were made on the basis of 1) geographic Production volumes for 1970-1992 were proximity and 2) similarity in the ratios of extracted from BESD ("UN Energy Statistics" offshore gas production to total gas database, "NG Natural Gas" indicator, production between the two countries. These "Production Volume" transaction). ratios were obtained from British Petroleum Production volumes for 1993-94 were taken (1995). Table B2 shows the resulting from British Petroleum (1995). Average assignments of surrogate countries. Note production cost estimates were obtained that "WD" denotes a world average from, by country: Turkmenistan, Iran, Iraq, production cost. Qatar, Saudi Arabia, United Arab Emirates, Oman, Nigeria, Algeria, Libya, Venezuela, General references on the expanding Trinidad & Tobago, Norway (IEA, 1995c); international market for natural gas and the Tunisia, Cameroon, Morocco, Tanzania Julius future of natural gas in developing countries and Mashayekhi, 1990; Mashayekhi 1983); include IEA (1996c), Conant (1986), and Egypt, Thailand, Bangladesh (Julius and Homer (1993). Useful factors for converting Mashayekhi, 1990; Khan, 1986; Mashayekhi between British Thermal Units, joules, cubic 1983); India, Pakistan (Mashayekhi, 1983); meters, cubic feet, tons of oil equivalent, and USA (Meyer, 1994; Adelman, 1991); South barrels of oil can be found in Valais (1977), Asia region (Khan 1986); Europe region Mashayekhi (1983), Varzi (1983) and (Cornot-Gandolphe, 1994); Russia (IEA, Blackwell Energy Research (1996). 1995c; Liefert, 1988)-roubles were converted Rent from Metals and Minerals using data from The Economist Intelligence Unit (1989; 1991; 1992). Metal production was measured as the metal content of ore production. With the exception Single-year data for production costs were of gold and silver, production volumes for converted to a 1970-1994 time series exactly 1970-1990 were extracted from BESD ("WB- as described above for the calculation of oil Metals & Minerals" database; rents. Those countries for which no "BAUXITE_GW,- "COPPER_OREMC," production cost data were available were, as "IRONORE_MC," "LEAD_OREMC," in the case of oil, assigned a "surrogate" "NICKEL_OREMC," "PHOSPHAT_ROCK," Environmental Economics Senes 29 Estimating National Wealth: Methodology and Results "TIN-OREMC," "ZINC_OREMC" indicators: York price and a London price) were "PROD_VOL" statistic). Production volumes averaged to arrive at an approximate for 1991-1994 came from World Bureau of international price. Bauxite prices for 1970- Metal Statistics (1995), with the exception of 1991 were taken from World Bank (1993b). iron ore (United Nations, 1995) and Since the approximate price ranges presented phosphate rock (FAO, 1995a). In the in Serjeantson (1996) show that bauxite price- occasional case that no data for 1994 were at-mine did not vary significantly during the available, a linear extrapolation of period 1991-1995, it was assumed that the approximately the past four years was World Bank (1993b) price for bauxite in 1991 performed. For gold and silver, production held constant on 1992-1994. Since iron ore volumes for 1970-1975 came from United production volumes are reported as mass of Nations (1977a), for 1975-1984 came from metal content and iron ore prices are given by United Nations (1986), for 1987-1991 came UNCTAD as mass of ore, an adjustment to the from United Nations (1995), and for 1992- iron price was necessary. Assuming that iron 1994 came from World Bureau of Metal metal is the only valuable component of iron Statistics (1995). ore, the iron ore prices were divided by their fractional metal content (reported by With the exception of Bauxite, international UNCTAD along with the price) to arrive at an price data for all metals were obtained from approximate price per mass of metal content UNCTAD (1989, 1993; 1996). In each case, the (e.g. Brazilian iron ore in 1980 wEts being two or three prices presented (often a New exported at $28.12 per metric ton of ore, with Table B2. "Surrogate" countries used to estimate natural gas production costs Country Surrogate Country I Surrogate Country I Surrogate Country I Surrogate AFG BGD COL VEN IRL EU QAT QAT AGO WD CSK EU IRN IRN ROM EU ALB EU CUB TTO IRQ IRQ RUS RUS ARE ARE CZE EU ISR IRQ RWA NGA ARG WD DDR EU ITA EU SAU SAU AUS THA DEU EU JPN EU SUN RUS AUT EU DFA EU KAZ RUS SVK EU AZE WD DNK EU KGZ RUS SWK IND BEL EU DZA DZA KWT IRQ SYR IRQ BGD BGD ECU WD LBY LBY THA THA BGR EU EGY EGY MAR MAR TJK RUS BHR SAU ESP EU MEX VEN TKM RUS BLR RUS FRA EU MMR BGD TTO TTO BOL WD GAB WD MYS IND TUN DZA BRA VEN GBR NOR NGA NGA TUR EU BRB TTO GEO RUS NLD EU UKR RUS BRN THA GRC EU NOR NOR USA USA CAN USA GTM VEN NZL WD UZB RUS CHE EU HRV EU OMN OMN VEN VEN CHL WD HUN EU PAK PAK VNM BGD CHN WD IDN PAK PER WD YSR EU COG NGA IND IND POL EU YUG EU 30 Enviromnent Department Papers Appendix B Estimating Rental Values for Timber and Sub-Soil Assets an Fe content of 64.5%. The price used for the price minus average unit production cost) calculation was thus $28.12 / 0.645 = $43.60 was negative. Since the assumption of per metric ton Fe). constant real production costs is rather strong, it is most likely that such numbers are Production costs for metals and minerals are errors indicative of the rough estimation proprietary information and very difficult to employed. This problem was evident in cases obtain for research purposes. In addition to of large price drops, such as the price of tin, bibliographical sources, the project received which fell by approximately two thirds from volunteer expert assistance from the US 1980 to the early 1990's. Such a drop would Geological Survey in Denver, Colorado probably force greater efficiency of (Bleiwas and Wagner, 1996). Sources for production and lower unit production costs, metal production costs, together with the but in the absence of complete production dates of data presented therein, were: Bauxite cost time-series data, a constant real 1984-1992 (World Bank, 1994), 1989 (Bleiwas, production cost was assumed. The remedy 1996) and 1985 (Bureau of Mines 1987); selected was, for each country with Copper 1975-1992 (World Bank, 1994), 1989 "negative" unit rents, to calculate an average (World Bank, 1989), 1988 (Bleiwas, 1996), and rental rate (the difference between market 1985 (Bureau of Mines, 1987); Gold 1992 price and production cost, as a fraction of (World Bank, 1994), 1991-1992 (Bleiwas, market price) for those years in which unit 1996), and 1985 (Bureau of Mines, 1987); Iron rent, as initially calculated, was positive. It Ore 1985 (Bureau of Mines, 1987); Lead 1988- was then assumed that the same rate would 1991 (World Bank, 1994), 1990 (Bleiwas, 1996), hold approximately constant in those years and 1985 (Bureau of Mines, 1987); Nickel where initially calculated unit rents were 1990-1992 (World Bank, 1994) and 1981 negative, and for those years the unit rent (Bleiwas, 1996); Phosphate Rock 1985 was set equal to the average positive rental (Bureau of Mines, 1987); Silver 1985 (Bureau rate multiplied by the market price. of Mines, 1987); Tin 1989 (World Bank, 1991) Significant numbers of such corrections were and 1985 (Bureau of Mines, 1987); and Zinc necessary for nickel in 1992-1994, phosphate 1988-1991 (World Bank, 1994), 1990 (Bleiwas, rock 1983-1994, tin 1970-71 and 1986-1994, 1996), and 1985 (Bureau of Mines, 1987). In and zinc 1970-71, 1985-87, and 1991-1994. In all cases, the most recent cost estimate for the absence of more precise production cost each country was used. A 1970-1994 time data, such approximations appear to be the series of costs was constructed from each most attractive alternative. single-year figure, as for oil, by assuming constant real costs and adjusting with a GDP Similarly to the technique used for oil and gas deflator. Cost data in most cases were the rents, countries for which no production cost sum of mining costs, milling costs, smelting/ data were available were assigned a refining/transportation costs, capital recovery (depreciation), and 15% lDiscounted "surrogate" country's production cost. These rcashvFlo (precateiof un), minuDiscbyprodute assignments were made strictly on the basis Cash Flow Rate of Return, minus byproduct o egahcpoiiy ol vrg credit. In rare cases where data were of gegapi prxmt A wol,vrg less detailr(e.g. presented simply production cost was assigned to countries presented in whose eograph was dissimilar to all those as a single, aggregate figure for "production gor ge ograta wy availar yl ,, . ~~~~~~~~~~for which cost data were available. cost"), that number, for lack of any other, was used. Rent from Hard Coal In some cases, for certain country-years, the Excepting countries of the Former Soviet calculated result for unit rent (unit market Union, production volumes for 1970-1992 Environmental Economics Series 31 Estimating National Wealth: Methodology and Results were extracted from BESD ("UN Energy average price and production cos t for all hard Statistics" database, "CL Hard Coal" coal, based on parallel figures for steam and indicator, "Production Volume" transaction). coking coal. There would clearly be many Production volumes for 1993-1994 were taken ways to carry out this estimation; the from National Mining Association (1996), methods below were chosen for their except Former Yugoslavia which came from expediency and plausible conclusions. Blackwell Energy Research (1996). Data for the Former Soviet Union came from IEA The price estimation was carried out as (1994b) for the period 1985-1991, the same follows. Data on export prices between 1985 BESD database cited above for 1992, and and 1991 for steam coal from three Australian Blackwell Energy Research (1996) for 1993- sites and one Canadian site were compared to 1994. export prices of coking coal from three Australian sites and two Canadian sites (IEA, In the case of coal, the determination of price 1995d). All were made commensurable by and production cost is complicated by 1) the converting to US dollars (IMF, 1996). The fact that coal exists in different grades of heat values of steam coal from ea,h site were different value and 2) the lack of a single, obtained (IEA, 1995d) as were the heat values uniform world price even for the same grade of average coking coal from each country of coal. The hard coal production volumes (IEA, 1995e). This allowed the calculation of reported in BESD, which made up the vast a price in US dollars per unit heal: value majority of the volume figures used, were (US$/1000 kcal) for each site in each year. standardized at 29.3076 Terajoules per For each of the two countries in each year, thousand metric tons (i.e., 0.6995 tons of oil this value was averaged across ali sites to equivalent per ton). Since hard coal includes obtain a country-average price per unit heat both steam and coking coal, and a range of value for steam coal and a corresponding heat values within each type, the BESD price for coking coal-in each year. For each production volumes are clearly adjusted to country in each year, a ratio of the average represent an average heat value and quality price per unit heat value for stearn coal to the of the coal to which is referred. BESD average price per unit heat value for coking indicates the heat value that was the norm of coal was determined. When averaged across standardization, but not the quality. Steam all years (1985-1991) this ratio was 0.7677 for coal, however, makes up more than three- Australia, and 0.7756 for Canada--essentially quarters of global hard coal production (IEA, identical. This means that after adjusting for 1995d), and it was thus assumed that the the effects of heat value, the value of steam BESD figures represented an aggregate of all coal is approximately 77% of the value of steam and coking coal standardized to an coking coal, presumably because of its lower average "coal equivalent" of steam coal with a quality (ash content, sulfur content, utility in heat value of 0.6995 toe/ton. It was assumed industry). This ratio was assumed to hold for that all collected hard coal production all hard coal. volumes uniformly represented this type of coal. Next, data for free-on-board prices of coking coal were collected for exports from the Further difficulties arise because, although United States, Former Soviet Union, the hard coal production volumes are Australia, Canada, China, South Africa, presented as an aggregate of steam coal and Indonesia, Poland, and New Zealand (IEA, coking coal, prices and production costs are 1995d) on the period 1970-1994 (for any given reported separately for steam coal and coking year, prices for between two and nine of the coal. It was necessary to establish a weighted above countries were available). Again, all 32 Enviromnent Department Papers Appenix B Estimating Rental Values for Timber and Sub-Soil Assets data were converted to US dollars (IMF, standardized heat value in which hard coal 1996). For each year, a world average export production volumes were reported (0.699465 price was calculated as the average of toe/ton), it was scaled up slightly by a factor available prices for that year, weighted by of 0.699465/0.686583. national average coking coal heat value (IEA, 1995e; IEA 19950. This allowed the Since production costs were also separately calculation of a complete time series of world reported for steam coal and coking coal, a average coking coal export price from 1970- similar method of estimating a single, 1994. An identical process, utilizing the same aggregated hard coal production cost was bibliographical sources, was followed to required. Production costs for steam coal obtain a time series for world average steam were obtained from, by country: Australia, coal export price, the differences being that 1) USA, Canada, Colombia, South Africa, Colombia was included in the country set and Indonesia (IEA, 1995d); Poland (IEA, 1995g); New Zealand was not, and 2) the final time Czech Republic (IEA, 1994c); China (Doyle, series stretched only from 1980-1994 due to 1987) [converted to US$ using IMF (1996)]; unavailability of steam coal export prices Russia (Tretyakova and Heinemeier, 1986) before 1980. To complete the steam coal [converted to US$ using The Economist series, prices on the period 1970-1979 were Intelligence Unit (1991; 1992)1; Mexico (World approximated from the coking coal world Bank, 1989); and India (Bhattacharya, 1995). prices for the same period by 1) scaling the Production costs for coking coal came from, coking coal price down to reflect the by country: Australia, USA, Canada, South difference between the average heat value of Africa (IEA 1995d); Poland (IEA, 1995c and coking and steam coal in the countries 1995d); and India (Bhattacharya, 1995). Since included in the sample and 2) scaling down coking coal data were scarce, only the steam by an additional factor of 0.77 to reflect the coal data were utilized and an aggregate hard difference in quality. Finally, after the steam coal production cost was constructed by coal prices for all years were scaled down to assuming production cost varies with the be commensurable in terms of dollars per heat content and quality of coal just as the kilocalorie with the coking coal prices, the market price does. That is, it was assumed world average hard coal export price in each year that after adjusting for heat value, the ratio of from 1970-1994 was approximated as the steam coal production cost to coking coal average of the export prices for steam and production cost is 0.77. Since the production coking coal, with steam coal given a weight cost figures came with no information on heat of three quarters and coking coal a weight of content, it was assumed that the heat content one quarter to reflect the aforementioned of steam coal for this calculation was 0.5943 relative production of the two types at the toe/ton, the average heat content of the global level. Note that in averaging steam steam coal exports analyzed in the price coal and coking coal prices, the steam coal calculation. To obtain an approximate prices were adjusted to be commensurable in aggregate steam-and-coking coal production terms of heat value with the coking coal cost for each country, the steam coal prices, meaning that the final hard coal price production cost was first scaled up to be series is in terms of a mixed steam/coking commensurable in heat content with the coal with a standard heat content equal to the production volumes (i.e. multiplied by average heat content of coking coal in the 0.6995/0.5943) and a weighted average was nine countries included in the coking coal found between the heat-adjusted steam coal average price (i.e. 0.686583 tons of oil figure and an estimated coking coal equivalent per ton). In order to make this production cost, the latter being obtained by final price commensurable with the dividing the former by 0.77. The weights in Environmental Economics Series 33 Estimating National Wealth: Methodology and Results this weighted average, as above, were three- 1993 (National Mining Association, 1995); all quarters for steam coal and one-quarter for Former Soviet Union countries except Russia coking coal. In shorthand form, the (IEA, 1994b). In certain cases where no estimated aggregate hard coal production production volume data for 1994 were cost for each country was derived from the available, 1994 volume was assumed equal to steam coal production cost by multiplying by 1993 volume. the following factor: Estimating unit rents for lignite was a (0.699465/0.594342)(1+((1/4)((1/0.77)-i))). daunting task, since no export prices are available (it is only traded internationally in As in the case of oil, this gave production minute quantities), domestic prices are often costs for a single year only, in almost all cases. distorted by subsidies, and production cost A time series of production costs from 1970- data were only available for a single country 1994 was generated by assuming constant real known to be a particularly heavy subsidizer production costs and adjusting the single- at the time of the study (Bhattacharya, 1995). year figure by a GDP deflator. The An estimation technique was required if assumption of constant real production costs lignite was to be included in the study. is a strong but necessary one, and as in the Again, as above, the following method was case of metals & minerals described earlier, it chosen for its expediency and plausible resulted in a few falsely negative unit rents conclusions, and has the potential to be for certain country-years-mostly on the improved. periods 1970-71 and 1993-1994. The same remedy that was applied to metals and An international price for lignite was minerals was applied again. Each negative estimated as follows. It was assumed that the unit rent was replaced by an estimated unit value of lignite is some affine function of the rent calculated by multiplying the average value of steam coal, and that the determining rental rate for all years in which calculated factors in this relationship are difference in unit rent was positive by the market price in heat value and some coefficient for "quality," the year of the unit rent being replaced. As in which together determine the slope of the the case of metals and minerals, countries for function. First, data on current free-on-board which no production cost data were available prices for exportable steam coal from 11 were assigned a "surrogate" country's countries were compiled (Coal Week production cost based on geographic International, 1996). The heat values of each proximity; when no such country was type were scatter-plotted against the f.ob. available, a world average production cost value, giving a least-squares affine was used. relationship of Rent from Brown Coal (Lignite) price = (0.0075)(heat value) - 11.73,' In most cases, production volumes for 1970- 1992 came from BESD ("UN Energy where price is in US$/ton and heat value in Statistics" database, "LB Lignite/Brown kcal/kg. This linear specification gave a Coal" indicator, "Production Volume" higher correlation coefficient than an indicator) and for 1993-1994 came from exponential, quadratic, or higher-order fit. Blackwell Energy Research (1996). Two different values were then compared: 1) Exceptions included: China 1984-1986 the price obtained from inserting 2693 kcal/ (Blackwell Energy Research, 1996); Albania, kg (the standardized heat content of lignite Austria, France, Italy, Japan and Myanmar used for the production volumes in BESD) 34 Environrment Department Papers Appendix B Estimating Rental Values for Timber and Sub-Soil Assets into the above affine function, i.e. $8.46/ton, internal coal prices (IEA, 1994c) gave a value and 2) for each country, the price that would for this lignite "quality deflator" of 0.85, a result if the price of both brown and steam figure which was discarded due to clear coal were assumed to be a simple linear distortions in the prices presented. The function of heat content alone, with no effects Czech Republic's recent history of central from coal quality (average across 38 coal planning makes all internal nominal price types in 11 countries of $15.16/ton). The data suspect, and in fact it is clear from the comparison was thus between a price source that steam coal was being sold below extrapolated from an observed downward cost at the time of the study. trend in price as both heat content and quality decrease, and an imaginary price ignoring the Lastly, a production cost for lignite was effects of quality. The ratio of the former to estimated. As for hard coal, production cost the latter, averaged across the sample, was was assumed to vary with coal heat value 0.562305, or approximately 0.56. That is, it and quality in the same manner as price. An was estimated that the value of brown coal analysis of lignite production costs identical would be 56% of the value which its lower to the above analysis of prices was heat content alone would imply, performed. Lignite and steam coal demonstrating the effects of higher sulfur production costs in Canada (World Bank, content, higher ash content, and decreased 1979) were compared with heat values (IEA, utility to industry. 1995e) to arrive at a "quality deflator" of 0.653; that is, lignite production costs were To verify this result, a separate analysis was approximately 65% of the costs that would be performed. Statistics Canada provided the expected from scaling down steam coal project with production volumes and values production costs solely to reflect the (internal prices) for Saskatchewan lignite and difference in heat value between the two steam coal from Alberta on the period 1979- types. In a similar analysis of Australia, hard 1994 (Born, 1996). Since the coal industry of coking, steam and brown coal production western Canada faces a relatively free costs (Abelson, 1983), hard coking coal heat market, these prices were assumed to be value (Abelson, 1983), and steam and brown competitive. Heat values for both types of coal heat value (IEA, 1995e) were compared coal came from IEA (1995e), and the prices to arrive at a "quality deflator" for lignite were converted to dollars with IMF (1996). production costs of 0.645, essentially The price of steam coal in each year was identical to the Canadian value. Note that in adjusted, assuming that price varies strictly this analysis, coking and steam coal linearly with heat value, to obtain an production costs were in 1983 Australian imaginary price for brown coal in the absence dollars and lignite costs were in 1980 US$, of quality effects. The ratio of the true price necessitating the use of a currency conversion of brown coal to this imaginary price, factor (IMF, 1996) and a GDP deflator (from averaged over 1979-1994, was, strikingly, BESD, "WB-IEC Data" Database, 0.562108, or approximately 0.56. The fact that "NY.GDP.MKTP.XU.E" indicator). An this ratio was identical to the ratio derived analysis of Czech coal production costs (IEA, above in a completely different fashion does 1994c) (note that production costs in the IEA not prove the analysis to be correct; however, study are estimates of real costs and therefore 0.56 was therefore judged to be a reasonable do not suffer from the aforementioned estimate of a "quality deflator" for lignite. distortion of the prices in the same report) This index was assumed to hold true for all showed the "quality deflator" for lignite lignite. An identical analysis of Czech production cost to be approximately 0.60. Envirorunental Economics Series 35 Estimating National Wealti: Medtodology and Results The only other study available that estimated countries came from FAO/UNECE (1992), brown coal production costs was for India and for a small number of tropical countries (Bhattacharya, 1995), where the costs from Kanowski et al. (1992), Lamprecht reported were indicated by the study's author (1989), and Duvigneaud (1971). Forest areas to be distorted by overinvestment in the for 1970, 1980, and 1990 were obtained, for sector during the oil crisis and ineffective most tropical countries, from the United monitoring of inefficient, nationalized lignite Nations Food & Agriculture Organization producers. Based on the above, a figure of (Singh, 1994), for most temperate countries 0.65 was assumed to be a valid "quality from FAO (1994), and for Former Soviet deflator" for global lignite production costs. Union countries from the International Institute for Applied Systems Analysis in The above assumptions and estimations Vienna (Nilsson, 1994). Some addlitional data allowed the calculation of a world price and were collected from WRI (1995). Nearly all estimated production costs for lignite, area figures were multiplied by a factor of 0.8 Lignite price was taken to be the world to allow for the fact that roughly a fifth of any Lignite price was taken to be the world ie onr' oetae sntacsil average export price for steam coal given country's forest area is not accessible for extraction due to steep slopes, rivers, etc. (calculated above) scaled down by a factor of (Cassells, 1996). In most cases, forest areas in 0.2693/0.5943 (the ratio of the BESD lignite 1971-1979 and 1981-1989 were eslimated by standard heat value to the average heat value linear interpolation between the above of steam coals used in the steam coal world figures and areas in 1991-1994 by price estimation) and scaled down again by a extrapolation of the 1980-1990 trend. Where factor of 0.56. Lignite production cost was data for 1970 were unavailable, data on the taken to be steam coal production cost (in period 1970-1979 were estimated by back- each country-year for which those data were casting based on the 1980-1990 trend. Thus, available) scaled down by a factor of 0.2693/ the product of the mean annual increment 0.5943 and again by a factor of 0.65. A time per hectare in commercial quality wood, the series for production costs in each country factor of 0.8, and the forest area was was made as it was for steam coal, and determined for each country in each year assignments of "surrogate" country from 1970-1994. This number will be called production costs were made as for steam the increment for that country-year. coal. In the minor case of two country-years, falsely "negative" unit rents were seen (both In certain countries, particularly in East Africa, in 1970; see the metals & minerals section), a large portion of roundwood production and were corrected as described in metals & comes from land which does not have minerals. sufficient tree density to be classified as a forest by FAO (Cassells, 1996). Table 6-5 in Timber Rent and Mean Annual Millington et al. (1994) indicates that Increments approximately 67% of roundwoocl production in East Africa comes from "forest" land, as Estimates of mean annual increment per opposed to 94% in Central Africa (Table 6-4 in hectare in commercial-quality wood mass Millington et al., 1994). The increment figures (m3/hectare/year), were calculated first by for certain countries were thus multiplied by a creating a table of "potential productivities" factor of 1/0.67 to reflect non-forestland based on a map of the same created from soil, increment. These countries were Rwanda, temperature and rainfall data (Mather, 1990). Burundi, Uganda, Kenya, Tanzania, Malawi, The resulting estimates were reviewed by a Haiti, Egypt, and Bangladesh. senior World Bank expert for substantial correction (Cassells, 1996). Estimates of mean Figures on total roundwood procluction on annual increment for most temperate the period 1970-1992 were extracted from 36 Environment Department Papers Appendix B Estimating Rental Values for Timber and SulbSoil Assets BESD ("FAO Forestry" database, price of coniferous logs, 2) an estimated "RNDWOOD_TOT" forestry code), and world average price for fuelwood, and 3) a figures for 1993-1994 were estimated as a third price, which varied according to the linear extrapolation of the trend on 1988-1992. region. In temperate countries, this third Figures for coniferous roundwood price was a world average export price for production were similarly obtained non-coniferous softwood logs. In tropical ("RNDWOOD_C" forestry code). These countries, this third price was an average figures were used to calculate the fraction of tropical hardwood export price of which total roundwood production that was there were three versions for Asia, Africa, and coniferous wood in each country-year. A Latin America. The weights in this average separate set of total roundwood production price were by relative proportion of total volumes and fuelwood production volumes roundwood that was coniferous, fuelwood, for each country-year on the period 1970-1994 and non-coniferous non-fuelwood. In precise was obtained from WRI (1996) and used to terms, the price was calculated as: P5 = (Qf)(P1) + (1-Qf)W(Qc)(P,)+(1-Q,)(P,)I where P. = Shadow price of roundwood P = Price of fuelwood = Price of coniferous roundwood PO = Price of "other" wood, depending on region: if temperate country, then non-coniferous softwood price if Latin America, then estimated Latin America tropical hardwood price if Africa, then Africa tropical hardwood price if Asia, then Asian tropical hardwood price = Fuelwood quotient, i.e. percentage of total roundwood production that is fuelwood = Coniferous quotient, i.e. percentage of total roundwood production that is coniferous calculate the percentage of total roundwood The technique of obtaining Qf and Qc has production that was fuelwood in each already been described. P, was the country-year (note that the WRI figures for "coniferous logs average world export unit total roundwood production were used for value" from FAO (1983; 1995b). Pf was the calculation of this percentage only; the estimated as an average value of reported total roundwood figures used in the main fuelwood prices in 21 developing countries calculation were those from BESD). The (Barnes, 1992), Kenya (Openshaw and original BESD figures were modified for Feinstein, 1989), Costa Rica and Nicaragua Malaysian roundwood production based on (van Buren, 1990), which came to an average data in ITTO (1996). The average ratio and relatively constant price of $25/cubic between the ITTO figures and the BESD meter around the end of the late seventies. A figures during the 1990's was assumed true time series of fuelwood prices was created by for the rest of the period as well, and this assuming constant real prices and adjusting ratio was used to scale down the BESD with a GDP deflator, as was done previously figures from 1970-1994. for production costs. The assumption that fuelwood prices do not follow the same Next, a shadow price for roundwood was trends as timber prices and can be assumed required. This price, in each country, was relatively constant is supported by Barnes estimated as a weighted average of three (1992). As was noted previously, P0 varied different prices: 1) the world average export depending on the region to which the country Environmental Economics Series 37 Estimating National Wealth: Methodology and Results in question belonged; there were four study a set of rental rate estimates was different versions, and thus four different collected from a group of experts and some prices P. were applied to total roundwood previous research. Rental rate is de'ined as production depending on the country. For temperate countries, P0 was "non coniferous ((Market Price - Production Cost)/(Market logs average world export unit value;" for Price)). Africa, P. was "tropical logs average export unit value Africa;" and for tropical Asia, P A rental rate of roughly 50% for Indonesia was "tropical logs average export unit value A estimate froughld for data Asia;" all from FAO (1983; 1995b). Since no was estimated from World Bank data time series data on Latin American tropical (Carbonnier, 1996; Douglas, 1996) Repetto et logs were available, an estimate was made: al. (1989) used a figure of 55% for Indonesia. the ratio of average tropical log export unit A study on the Philippines by delos Angeles value in Latin America to the same value for et al. (1988) found figures of 42% and 58% on the Asia-Pacific region in 1993 (Table 3-3 in two different sites (average 50%). Further ITTO, 1996) was assumed to hold constant on World Bank data were employed to estimate the entire period 1970-1994. This allowed the 48% for Thailand (Sadoff, 1996). Biased on Latin American price to be calculated from these figures, an approximate rental rate of the Asian price. 50% was used for East Asia, Southeast Asia, Since all FAO prices only extended to 1991, and South Asia. prices for 1992-1994 had to be estimated. A World Bank data suggest an approximate time series for a single type of wood, the rl rate of suggest a frican price of Okoume hardwood from Cameroon, rental rate of 30% for the West African was available for the entire period 1970-1994 rainforest (Rietbergen, 1996). A study by (UNCTAD 1989; 1993; 1996). Each of the four Gillis (1988) found the rate in Ghana to be types of primary roundwood price data was 26%. Based on these figures, a lower rate of estimated as an affine function of the 30% was used for all of Africa. Okoume price based on the period 1970-1991. All five prices were found to follow similar A study by Cottle et al. (1990) dernonstrated a trends on this period; the estimate of rental rate of approximately 55% for a coniferous roundwood price as a function of Brazilian forestry operation. Sol6rzano et al. Okoume price during 1970-1991 had an R2 of (1991) provide data on Costa Rican internal 0.948, for the non-coniferous softwood price prices which show an average rental rate in R2 was 0.898, for Africa tropical hardwood recent years of 68% when sawmill costs are price R2 was 0.957, and for Asian tropical excluded (probably too high, but effects on hardwood price it was 0.808. Since this set of the price of processing are difficult to four functions was a reasonable predictor of separate since data on unprocessed logs are the prices of the four types of wood as a K 1 v function of the Okoume price on 1970-1991, not presented). Kellenberg (1995) provides the same functions were used to estimate the data that show an approximate rate of 52% in prices of the four on the period 1992-1994, Ecuador. Based on these results, a rental rate again based on the Okoume price for those of 55% was assumed for all of Latin America. years. Finally, Carbonnier (1996) further estimated a Rather than collect information on rental rate of 40% for temperate country production costs, for the purposes of this forestry operations. 38 Enviromnent Department Papers References Abelson, Peter. 1983. Coal in Australia: Prospects Carbonnier, Louis. 1996. Agriculture and to 1990. Special Report No. 149. London: Natural Resources Department, Forestry The Economist Intelligence Unit. Systems Division, World Bank. Personal Adelman, M. A. 1987. "Offshore Norwegian communication. Development Cost Calculated From Project Cassells, David. 1996. Environment Data," Energy Exploration & Exploitation, Vol. Department, Land, Water and Natural 7, No. 1, pp. 51-62. Habitats Division, World Bank. Personal -. 1991. "US oil/gas production cost: Communication. recent changes," Energy Economics, Vol. 13, Central Bureau of Statistics of Norway. 1992. October issue, pp. 235-237. Natural Resources and Environment 1992. Barnes, Douglas F. 1992. Understanding Report 93/1A. Oslo Kongsvinger. Fuelwood Prices in Developing Nations. Coal Week International. 1996. "Current Steam Washington, DC: World Bank. Coal Price Table." Vol. 17, No. 28, July 9. Bhattacharya, Subhes C. 1995. "Estimation of McGraw-Hill. subsidies on coal in India." Natural Conant, Melvin A., ed. 1986. The World Gas Resources Forum. Vol. 19. No. 2. pp. 135-142. Trade: A Resourcefor the Future. Boulder, Blackwell Energy Research. 1996. Oil and Colorado: Westview Press. Energy Trends: Annual Statistical Review. [No Comot-Gandolphe, Sylvie. 1994. Adequate and city] Secure Gas Supplyfor Europe: How to Meet the Bleiwas, Donald. 1996. Denver Office of the US growing demandfor natural gas? Paris: Geological Survey. Personal CEDIGAZ, IEA International Center for Gas Communication. Technology Information. Bleiwas, Donald and Lorie Wagner, 1996. Cottle, Laura E., Gerard F. Schreuder and Denver Office of the US Geological Survey. Antonio A. A. de Barros. 1990. Brazil: A Personal Communication Country Profile of the Forests and Forest Born, Alice. 1996. Oil, Gas and Mineral Industries. Working Paper 27. Seattle: Accounts; National Accounts and Center for Intemational Trade in Forest Environment Division, Statistics Canada Products, University of Washington. (Government of Canada). Personal Dasgupta, Partha. 1993. An Inquiry into Well- communication. being and Destitution. Oxford University British Petroleum. 1995. BP Review of World Gas. Press: Oxford. London. delos Angeles, Marian S., et al. 1988. "The Bureau of Mines. 1987. An Appraisal of Minerals Private Costs of Commercial Forestry, Availabilityfor 34 Commodities. United States Reforestation and Social Forestry." Journal of Dept. of the Interior. Washington, DC: US Philippine Development. Vol. 15. No. 1. pp. Goverinment Printing Office. 113-135. Environmental Economics Senes 39 Estirating National Wealth: Methodology and Results Dienes, Leslie, Istvan Dobozi, and Marian Homer, John. 1993. Natural Gas in Developing Radetzki. 1994. Energy and Economic Reform Countries: Evaluating the Benefits to the in the Former Soviet Union. New York: St. Environment. World Bank Discussion Paper Martin's Press. No. 190. Washington DC: World Bank. Douglas, James. 1996. Agriculture and Natural IADB (Inter-American Development Bank). Resources Department, Forestry Systems 1981. Investment and Financing Requirements Division, World Bank. Personal for Energy and Minerals in Latin America, communication. Washington: IADB. Doyle, Guy. 1987. China'spotentialin IEA (International Energy Agency). 1994a. international coal trade. London: International Russian Energy Prices, Taxes, and Costs 1993. Energy Agency Coal Research. Paris: OECD. Duvigneaud, P., ed. 1971. Productivity of Forest . 1994b. Coal Information 1993. Paris: Ecosystems. Paris: UNESCO. OECD. EIA (Energy Information Administration). 1995. . 1994c. Energy Policies of the Czech Performance Profiles of Major Energy Producers Republic. Paris: OECD. 1993. Washington, DC: US Dept. of Energy. . 1995a. Energy Policies of the Russian FAO (United Nations Food and Agriculture Federation, 1995 Survey. Paris: C)ECD. Organization). 1995a. Fertilizer Yearbook . 1995b. Middle East Oil and Gas. Paris: 1994. Vol. 44. Rome. OECD. _____. 1983. Forest Products Prices 1963-1982. . 1995c. Oil, Gas & Coal: Supply Outlook. Forestry Paper 46. Rome. Paris: OECD. ______. 1994. Forest Products Yearbook 1993. . 1995d. Coal Information 1994. Paris: Rome. OECD. _______ 1995b. Forest Products Prices 1973-1992. . 1995e. Energy Balances of OECD Forestry Paper 125. Rome. Countries 1992-1993. Paris: OECD. FAO/UNECE. 1992. The Forest Resources of the . 1995f. Energy Statistics and Balances of Temperate Zones. Vol. 1: General Forest Non-OECD Countries 1992-1993. Paris: Information. New York, United Nations. OECD. Gillis, Malcolm. 1988. "West Africa: resource . 1995g. Energy Policies of Poland. Paris: management policies and the tropical rain OECD. forest." In Repetto, R. and M. Gillis, eds. . 1996a. Energy Policies of llkraine, 1996 Public Policies and the misuse offorest resources. Survey. Paris: OECD. London: Cambridge Univ. Press. . 1996b. Energy Prices and Taxes, 4th Government of Canada. 1996. International Quarter 1995. Paris: OECD. Natural Gas Division, Energy Sector, Dept . 1996c. Asia Gas Study. Paris: OECD. of Natural Resources Canada, http:/ / esl.es.emr.ca/ingd/tofc/pgll.html, July IMF (International Monetary Fund). 1996. 15th. International Financial Statistics Yearbook 1995. Hamilton, Kirk, and Ernst Lutz. 1996. Green Washington, DC. National Accounts: Policy Uses and Empirical ITTO (International Tropical Timber Experience. Environment Department Paper Organization). 1996. Annual Review and Number 39. World Bank, Environment Assessment of the World Tropical Timber Department, Vice Presidency for Situation 1995. Yokohama, Japan. Environmentally Sustainable Development. Jenkins, Gilbert. 1989. Oil Economists' Handbook, Washington DC. 5th ed. Vol. 2. New York: Elsevier. 40 Environrnent Department Papers References Kanowski, P. J. et al. 1992. "Plantation National Mining Association. 1996. International Forestry." In Sharma, Narendra P., ed. Coal. 1995 edition. Washington, DC. Managing the World's Forests: Lookingfor Nehru, Vikram, and Ashok Dhareshwar. 1993. Balance between Conservation and "A New Database on Physical Capital Stock: Development. Dubuque, Iowa: Kendall/ Sources, Methodology and Results". Revista Hunt Publishing Co. de Andlisis Econ6mico. Vol. 8, No. 1, pp. 37-59. Kellenberg, John. 1995. Accountingfor Natural Nilsson, Sten. 1994. International Institute for Resources: Ecuador 1971-1990. Ph.D. Applied Systems Analysis, Vienna. Personal Dissertation, Department of Geography and Communication. Environmental Engineering. Baltimore: Openshaw, Keith and Charles Feinstein. 1989. Johns Hopkins University. Fuelwood Stumpage: Financing Renewable Lampietti, Julian A., John A. Dixon. 1995. To See Energyfor the World's Other Half. Energy the Forestfor the Trees: A Guide to Non-Timber Strategy, Management and Assessment Forest Benefits. Environment Department Working Paper No. 270. Washington, DC: Paper Number 13. World Bank, World Bank. Environment Department, Vice Presidency Pearce, David W., and R. Kerry Turner. 1990. for Environmentally Sustainable Economics of Natural Resources and the Development. Washington DC. Environment. Harvester Wheatsheaf. Lamprecht, Hans. 1989. Silviculture in the , and David Ulph. 1995. A Social Discount Tropics. Eschborn, Germany: Deutsche Ratefor the United Kingdom. Centre for Social GJesellschaft fur Technische and Economic Research on the Global Zeseamschafbet fu ehicEnvironment. Working Paper 95-01. ,-usammenarbeit. University College Longdon.9501 Liefert, William Mark. 1988. "The Full Cost of Reioert etlae. 1989. Soie Oi.n auaGsPouto. Repetto, Robert et al. 1989. Wasting Assets: SovieratOil acond miNa Sturalias, Product Natural Resources in the National Income Comparative Economic Studies, Vol. 30, Accounts. Washington, DC: World Resources Summer issue, pp. 1-20 Institute. Mashayekhi, Afsaneh. 1983. Marginal Cost of Rietbergen, Simon. 1996. Africa Technical Natural Gas in Developing Countries: Concepts Department, World Bank. Personal and Applications. Energy Dept. Washington, communication. DC: World Bank. Rosegrant, Mark W., Mercedita Agcaoili- Mather, A. 1990. Global Forest Resources. Sombilla, Nicostrato D. Perez. 1995. Global London: Belhaven Press. Food Projections to 2020: Implicationsfor McKeough, Kay. 1989. Current International Gas Investment. Food, Agriculture, and the Trades and Prices, World Bank Policy, Environment Discussion Paper 5. Planning & Research Working Paper No. International Food Policy Research Institute. 159, Washington, DC: World Bank. Washington DC. Meyer, Sandra et al., eds. 1994. International Sadoff, Claudia. 1996. AFlEI division, World Energy Statistics Sourcebook. Fourth Edition. Bank. Personal communication. Tulsa, Oklahoma: PennWell Publishing Co. Sagers, Matthew et al. 1995. "Resource Rent from the Oil and Gas Sector and the Russian _______ 1994. Natural Gas Statistics Sourcebook. Economy," Post-Soviet Geography, Vol. 36, No. Oil & Gas Journal Energy Database 7, pp. 389-425. Publcations. Tulsa, Oklahoma: PennWell Serageldin, Ismail. Sustainability and the Wealth of Publishing Co. Nations: First Steps in an Ongoing Journey. Millington, Andrew C. et al. 1994. Estimating Environmentally Sustainable Development Woody Biomass in Sub-Saharan Africa. Studies and Monographs Series Number 5. Washington, DC: World Bank. Washington DC. Environmental Economics Series 41 Estimating National WealthL Methodology and Results Serjeantson, Richard, ed. 1996. Metal Bulletin's UNCTAD (United Nations Conference on Trade Prices & Data. Surrey, UK: Metal Bulletin and Development). 1989. Monthly Books Ltd. Commodity Price Bulletin, 1970-1989 Singh, K. D. United Nations Food and Supplement. New York: United Nations. Agriculture Organization, Rome. Personal . 1993. Monthly Commodity Price Communication. Bulletin. Vol. 13, No. 1. New York: United Smith, Roger. 1992. "Income Growth, Nations. Government Spending, and Wasting . 1996. Monthly Commodity Price Assets-Alberta's Oil and Gas." Canadian Bulletin. Vol. 16, No. 1. New York: United Public Policy. Vol. 18, No. 4, pp. 387-412. Nations. Sol6rzano, Raul et al. 1991. Accounts Overdue: United Nations. 1975. Statistical Yearbook 1974. Natural Resource Depreciation in Costa Rica. New York Washington, DC: World Resources Institute. N York. Statistics Canada. 1985. National Balance Sheet York. Accounts 1961-84. Statistics Canada, Financial Flows Section. . 1977a, Yearbook of Industrial Statistics Stauffer, Thomas R. 1984. "National Incomes of 1975, Volume II: Commodity Production Data Oil Producers Are Overstated." OPEC 1966-1975. Dept. of Economic and Social Bulletin. February 1984 issue. pp. 17-19+. Affairs, Statistical Office of the United . 1986. "Accounting for 'Wasting Nations. New York. Assets': Measurements of Income and . 1977b. Statistical Yearbook 1976. New Dependency in Oil-Rentier States." The York. Journal of Energy and Development. Vol. 11, . 1978. Statistical Yearbook:1977. New No. 1, pp. 69-93. York. Steer, Andrew. 1996. "Ten Principles of the New . 1979. Statistical Yearbook :1978. New Environmentalism". Finance and York. Development. December 1996. International Monetary Fund and the World Bank. . 1982. Statistical Yearbook 1981. New Washington DC. York. Streifel, Shane. 1996. Commodity Policy and . 1983. Statistical Yearbook:1982. New Analysis Unit, International Economics York. Department, World Bank. Personal . 1985. Statistical Yearbook :1983/84. New Communication. York. The Economist Intelligence Unit. 1989. Country . 1986. Industrial Statistics 'Yearbook 1984. Profile: USSR, 1988-89. London: The Volume II: Commodity Productioni Statistics. Economist. Dept. of International Economic and Social ______. 1991. Country Profile: USSR, 1990-91. Affairs, Statistical Office of the United London: The Economist. Nations. New York. ______. 1992. Country Profile: USSR, 1991-92. . 1987. Statistical Yearbook:1985/86. New London: The Economist. York. Tiratsoo, E. N. 1979. Natural Gas. 3rd ed., . 1988. Statistical Yearbook :1987. New Beaconsfield, England: Scientific Press Ltd. York. Tretyakova, Albina and Meredith Heinemeier. .1990. Statistical Yearbook 1988/89. New 1986. Cost Estimatesfor the Soviet Coal Industry: 1970 to 1990. Washington, DC: York. Center for International Research, Bureau of . 1992. Statistical Yearbook 1990/91. New the Census, US Dept. of Commerce. York. 42 Environment Department Papers References . 1995. 1993 Industrial Commodity World Bank. 1992. Arab Republic of Egypt: An Statistics Yearbook: Production and Agricultural Strategyfor the 1990s. Sector Consumption Statistics. Department for Report Number SR-11083. World Bank, Economic and Social Information and Policy Agriculture Operations Division, Country Analysis, Statistical Division. New York. Department II, Middle East and North Valais, M. and M. Hiegel. 1977 L'industrie du gaz Africa Region. Washington f EC. dans le monde. Departement Economie, . 1993a. Republic of Yemen: Agricultural Institut Francais du Petrole. Paris: Editions Sector Study: Strategyfor Sustainable Technip. Agricultural Production. Vol. 1 and 2. Sector Report Number SR-11126. World Bank, van Buren, Ariane. 1990. The Woodfuel Market in Agriculture Operations Division, Country Nicaragua: The Economics, Sociology and Department II, Middle East and North Management of a Natural Energy Resource. Africa Region. Washington DC. Amersterdam: Centrum voor Studie en . 1993b. Commodity Trade and Price Documentatie van Latijns Amerika. Trends. 1989-91 edition. Baltimore: Johns Varzi, Medhi, ed. 1983. World Gas Survey. Hopkins University Press. London: Grieveson, Grant and Co. . 1994. Market Outlookfor Major Energy Investment Research. Products, Metals, and Minerals. IECIT Vincent, J., 1996, Resource Depletion and division. Washington, DC. Economic Sustainability in Malaysia, . 1995. Mongolia: Prospectsfor Wheat Development Discussion Paper No. 542, July Production. Sector Report Number SR-13882. 1996, Harvard Institute for International World Bank, Agriculture Operations Development, Harvard University. Division, China and Mongolia Department, Wagner, Lore A. and Audrey A. Soja. 1995. East Asia and Pacific Regional Office. International Mining Cost Indexation System. Washington DC. Denver, Colorado: US Dept. of the Interior, . 1996. World Development Report 1996: US Bureau of Mines Minerals Availability From Plan to Market. Oxford University Field Office. Press: New York. West, Jim, ed. 1996. International Petroleum . 1997a. Five Years after Rio: Innovations in Encyclopedia. Tulsa, Oklahoma: PennWell Environmental Policy. Environmentally Publishing Co. Sustainable Development Studies and Monographs Series No. 17. World Bank: Whitehead, Harry. 1983. An A-Z of Offshore Oil Washington DC. & Gas. Second edition. Houston: Gulf . 1997b. Expanding the Measure of Wealth: Publishing Co. Indicators of Environmentally Sustainable World Bank. 1979. Coal Development Potential Development. Environmentally Sustainable and Prospects in the Developing Countries. Development Studies and Monographs Washington, DC. Series No. 17. World Bank: Washington DC. . 1989. Republic of Mexico Mining Sector World Bureau of Metal Statistics. 1995. World Review. Trade, Finance and Industry Metal Statistics Year Book 1995. London. Division, Country Department II. Latin World Commission on Environment and America and the Caribbean Office. Development. 1987. Our Common Future. Washington, DC. Oxford: oxford University Press. . 1991. Tin Sector Strategy Report. World Resources Institute. 1992. World Resources Industry & Energy Dept., Operations 1992-93. Oxford University Press: New York, Division. Washington, DC. Oxford. Environmental Economics Series 43 Estimating National Wealth: Metiodology and Results ______. 1995. World Resources, 1994-95. New . 1996. World Resources 1996-97. Oxford York: Oxford University Press. University Press: New York, Oxford. _____. 1996. World Resources Database 1996- 1997. Magnetic media. Washington, DC. 44 Environment Department Papers : . . - ; : - s Environment Department The Woricl Bank 1818 H Street, N.W. - - - - - . Washington, D. C. 20433. - -- . . - -- . - w 202 473 3641 - - 202 477 05O5 FAX - -: . _ . . . .  - - - . . . . . . . . . . - .. . . . . . . . . . . .. . . * X * . . . . ... : . .. . . . . < .. . . . .- . ... . .- . . . . . , . . . - .- - . - * . . ,0 . - : -,- -.: .. . . . . . . . . . , ; .. , . . . . . , . . . . .. . <, . . .. . . . . . . .. . ,. '-, , , . . . , . . . . . - . . . . . . . . . . .. . / r . ... . . . . . . . . . . , . . .. . . . . . . . . . !. .. . . . . . . .. . . _ . _ . :, ... . . ._ .' , ' . , ' , ,. - ' .., " '. . . . . . - . . ' . , . . . .. . . . .. . . 63 Printecl on lP0% post-consumer recycled poper - - - . . . . . . A , .  . , r . ' . . . . .. _ . _