WASTING ASSETSNatural Resources in theNational Income AccountsRobert RepettoWilliam MagrathMichael WellsChristine BeerFabri/io Rossini

W O R L D R E S O U R C E S I N S T I T U T E

WASTING ASSETS:Natural Resources in theNational Income Accounts

Robert RepettoWilliam MagrathMichael WellsChristine BeerFabrizio Rossini

W O R L D R E S O U R C E S I N S T I T U T E

A Center for Policy Research

June 1989

Copyright © 1989 World Resources Institute. All rights reserved.

Library of Congress Cataloging-in-Publication DataWasting assets: natural resources in the national income accounts/

by Robert Repetto.. .[et al.].p. cm.

Bibliography: p.ISBN 0-915825-31-7: $10.001. National income—Accounting. 2. Natural resources—Valuation.

3. Timber—Valuation—Indonesia—Case studies. 4. Petroleum industryand trade—Valuation—Indonesia—Case studies.I. Repetto, Robert C. II. World Resources Institute.HB141.5.W37 1989339.4'9598-dc20 89-5842

CIP

Kathleen CourrierPublications Director

Don StrandbergMarketing Manager

Hyacinth BillingsProduction Manager

FAO/S. PierbattistaCover Photo

Each World Resources Institute Report represents a timely, scientific treatment of a subject of public concern.WRI takes responsibility for choosing the study topics and guaranteeing its authors and researchers freedom ofinquiry. It also solicits and responds to the guidance of advisory panels and expert reviewers. Unless otherwisestated, however, all the interpretation and findings set forth in WRI publications are those of the authors.

Contents

I. The Need for Natural ResourceAccounting 1

A. Overview and Recommendations 1B. Current National Income Accounting . .12

1. Imputations and the Treatment ofDepreciation 12

2. Income Statements and BalanceSheets 12

C. The Scope of Natural ResourceAccounting 14

D. Setting Up Natural Resource Accounts. 171. Physical Accounts 172. Valuation Principles 19

E. Integrating Natural Resources into theNational Accounts 22

Note 25

II. The Indonesian Resource Accounts 27A. Timber Resource Accounts,

1970-1984 271. The Physical Accounts 28

a. Growing Stock 28b. Growth and Reproduction 30c. Harvesting 31d. Deforestation and Degradation .. .31e. Summary of Physical Accounts .. .32

2. The Value Accounts 32a. The Measure of Economic Rent.. .32

B. Indonesian Petroleum ResourceAccounts, 1970-1984 371. The Physical Accounts 382. The Value Accounts 39

C. Indonesia's Soil Account: Java 421. Estimates of Soil Erosion in Physical

Terms 422. Estimates of the Economic Costs of

Erosion 44D. Concluding Remarks 53

Annex 55

Notes 61

References 63

Acknowledgments

T he authors wish to thank those whoassisted in the research process—PedroArens, Tinling Choong, and Richard

Woodward. We also wish to thank KathleenCourrier, Hyacinth Billings, Katherine Harring-ton and Karen Saxon for their able assistance.

Thanks are due to many people who havereviewed our earlier drafts including DennisAnderson, William Baumol, Derek Blades, AlanBrewster, Shanta Devarajan, Mohamed El-Ashry,Tom Fox, Ernesto Hernandez-Cata, KentonMiller, William Moomaw, and Donna Wise.

We also appreciate the interest shown in thestudy by Professor Emil Salim, Minister ofEnvironment and Population, Republic ofIndonesia and Professor Iwan Aziz, Chairmanof the Economics Department, University ofIndonesia.

R.RW.MM.W.C.B.F.R.

111

Foreword

"S ustainable development" hasbeen defined variously as livingon the planet's income instead of

depleting nature's capital, as meeting theneeds of today's population without com-promising the ability of future generations tomeet theirs, and as the management of natural,human, and financial assets so as to increaselong-term wealth and well-being. By whateverdefinition, sustainable development is clearlyan important objective for societies.

National income accounts, the informationframework that countries use to analyze theperformance of their economies and to deter-mine gross and net national product, ought toencompass the concept of sustainability. And,indeed, they do in certain respects. Man-madeassets, including plant and equipment, are val-ued as productive capital, and their deprecia-tion is charged against the value of nationalproduction. But this treatment of capitaldepreciation in national income accountingdoes not extend to natural resource depletion.The result is what Robert Repetto and his co-authors refer to in this report as a "dangerousasymmetry." As he notes, "A country couldexhaust its mineral resources, cut down itsforests, erode its soils, pollute its aquifers, andhunt its wildlife and fisheries to extinction, butmeasured income would not be affected asthese assets disappeared."

loss of natural resources and the services theyprovide, policymakers can get very misleadingsignals, as the results reported here for Indone-sia show. Temporary improvements in con-sumption can be purchased by permanentlosses in wealth and productive capacity.

Few people who aren't economists evenknow what the accounts are, much less howthey are calculated. And yet, whenever quar-terly GNP figures are released, policymakersinvariably find themselves on the line: consti-tuents, reporters, and financial analysts allwant to know why the economy's performanceis up, down, or unchanged. No economic lawsays that natural resources and the servicesthey provide can't be included in nationalincome accounts. Indeed, principles of botheconomics and ecology argue that they shouldbe. But how?

Wasting Assets: Natural Resources in theNational Income Accounts demonstrates that nat-ural resources can be treated similarly to capitalin national accounts, and it argues convincinglythat these accounts should be revised. Repettoand his co-authors don't stop at building atight theoretical case. Using data from Indone-sia, they provide a concrete example of howthe revised accounts would work and what sig-nals the new results would give to those whomake decisions about economic development.

When the index by which we try to measureimprovements in living standards ignores the

This report complements several others thatWRI has conducted that seek to bring economic

and environmental thinking together in a new-synthesis. The Forest for the Trees: GovernmentPolicies and the Misuse of Forest Resources (RobertRepetto, WRI, 1988), Money to Burn? The HighCosts of Energy Subsidies (Mark Kosmo, WRI,1987), Skimming the Water: Rent-Seeking and thePerformance of Public Irrigation Systems (RobertRepetto, WRI, 1986), Paying the Price: PesticideSubsidies in Developing Countries (RobertRepetto, WRI, 1985), and Public Policies and theMisuse of Forest Resources (Robert Repetto andMalcolm Gillis, Cambridge University Press,1988) all show how misguided economic incen-tives cost governments huge sums and distortinvestment decisions while inviting environ-mental abuse and wasting natural resources. In

addition, related work is now under way atWRI on the economics of sustainable agricul-ture and international conservation financingoptions.

For support for all these studies and WRI'sresearch program in general, we are deeplygrateful to the John D. and Catherine T.MacArthur Foundation. Additional much-appreciated support for this study also camefrom the World Bank.

James Gustave SpethPresidentWorld Resources Institute

VI

I. The Need for Natural ResourceAccounting

A. Overview andRecommendations

W hatever their shortcomings, andhowever little their construction isunderstood by the general public,

the national income accounts are undoubtedlyone of the most significant social inventions ofthe twentieth century. Their political and eco-nomic impact can scarcely be overestimated.However inappropriately, they serve to dividethe world into "developed" and "less devel-oped" countries. In the "developed coun-tries," whenever the quarterly gross nationalproduct (GNP) figures emerge, policy-makersstir. Should they be lower, even marginally,than those of the preceding three months, arecession is declared, the strategies and compe-tence of the administration is impugned, andpublic political debate ensues. In the "develop-ing" countries, the rate of growth of GNP isthe principal measure of economic progressand transformation.

The national accounts have become so mucha part of our life that it is hard to rememberthat they are scarcely fifty years old. They werefirst published in the United States in the year1942. It is no coincidence that the period dur-ing which these measures have been available,with all their imperfections, has been theperiod within which governments in all devel-

oped and most developing countries havetaken responsibility for the growth and stabilityof their economies, and during which enor-mous investments of talent and energy havebeen made in understanding how economiescan be better managed. Forecasting the nextfew quarterly estimates of these statistics hasbecome, with no exaggeration, a hundred mil-lion dollar industry.

The aim of national income accounting is toprovide an information framework suitable foranalyzing the performance of the economic sys-tem. The current system of national accountsreflects the Keynesian macroeconomic modelthat was dominant when the system wasdeveloped. The great aggregates of Keynesiananalysis—consumption, savings, investment,and government expenditures—are carefullydefined and measured. But Keynes and hiscontemporaries were preoccupied with theGreat Depression and the business cycle; spe-cifically, with explaining how an economycould remain for long periods of time at lessthan full employment. The least of their wor-ries was a scarcity of natural resources. Unfor-tunately, as Keynesian analysis largely ignoredthe productive role of natural resources, sodoes the current system of national accounts.

In fact, natural resource scarcity played littlepart in 19th century neo-classical economics,from which traditional Keynesian and most

contemporary economic theories are derived.Gone were the dismal predictions of Ricardo,Malthus, Marx, and other earlier classicaleconomists that industrial economies wouldstagnate or collapse because of rising rents andsubsistence wages. In 19th century Europe,steamships and railroads were markedly lower-ing transport costs while foodgrains and rawmaterials were flooding in from North andSouth America, Australia, Russia, and theimperial colonies. What mattered to Englandand other industrializing nations was the paceof investment and technological change. Theclassical economists had regarded income asthe return on three kinds of assets: naturalresources, human resources, and invested capi-tal (land, labor, and capital, in their vocabu-lary). The neo-classical economists virtuallydropped natural resources from their modeland concentrated on labor and invested capital.When these theories were applied after WorldWar II to problems of economic developmentin the Third World, human resources were alsoleft out on the grounds that labor was always"surplus," and development was seen almostentirely as a matter of savings and investmentin physical capital.

their loss entails no debit charge against cur-rent income that would account for thedecrease in potential future production. Acountry could exhaust its mineral resources,cut down its forests, erode its soils, pollute itsaquifers, and hunt its wildlife and fisheries toextinction, but measured income would not beaffected as these assets disappeared. Ironically,low-income countries, which are typically mostdependent on natural resources for employ-ment, revenues, and foreign exchange earningsare instructed to use a system for nationalaccounting and macroeconomic analysis thatalmost completely ignores their principalassets.

A country could exhaust its mineralresources, cut down its forests, erodeits soils, pollute its aquifers, and huntits wildlife and fisheries to extinction,but measured income would not beaffected as these assets disappeared.

There is a dangerous asymmetry todayin the way we measure, and hence, theway we think about, the value of natu-ral resources.

As a result, there is a dangerous asymmetrytoday in the way we measure, and hence, theway we think about, the value of naturalresources. Man-made assets—buildings andequipment, for example—are valued as produc-tive capital, and are written off against thevalue of production as they depreciate. Thispractice recognizes that a consumption levelmaintained by drawing down the stock of capi-tal exceeds the sustainable level of income.Natural resource assets are not so valued, and

Underlying this anomaly is the implicit andinappropriate assumption that naturalresources are so abundant that they have nomarginal value. This is a misunderstanding.Whether they enter the marketplace directly ornot, natural resources make important contri-butions to long-term economic productivity andso are, strictly speaking, economic assets.Many are under increasing pressure fromhuman activities and are deteriorating in quan-tity or quality.

Another misunderstanding underlies the con-tention that natural resources are "free gifts ofnature," so that there are no investment coststo be "written off." The value of an asset isnot its investment cost, but the present valueof its income potential. Many companies val-ued by the stock market as worth many bil-lions of dollars have as their principal assets

the brilliant ideas and inventions of theirfounders: the Polaroid Camera, the AppleComputer, the Lotus Spreadsheet, for example.These inspired inventions are worth vastlymore than any measurable cost to their inven-tors in developing them and could also beregarded as the products of genius—free giftsof nature.

Common formulas for calculating deprecia-tion by "writing off" investment costs (e.g.,straight line depreciation) are just convenientrules of thumb, or artifacts of tax legislation.The true measure of depreciation, whichstatisticians have tried to adopt for fixed capitalin the national accounts, is the capitalizedvalue of the decline in the future incomestream because of an asset's decay or obsoles-cence. (Usher 1980, pp. 104-105) Thus, in thesame sense that a machine depreciates, soilsdepreciate as their fertility is diminished sincethey can produce only at higher costs or loweryields.

77MS difference in the treatment ofnatural resources and other tangibleassets reinforces the false dichotomybetween the economy and the"environment" that leads policymakersto ignore or destroy the latter in thename of economic development.

Codified in the United Nations system ofnational accounts closely followed by mostcountries, this difference in the treatment ofnatural resources and other tangible assets pro-vides false signals to policymakers. It reinforcesthe false dichotomy between the economy andthe "environment" that leads policymakers toignore or destroy the latter in the name ofeconomic development. It confuses the deple-tion of valuable assets with the generation ofincome. Thus it promotes and seems to

validate the idea that rapid rates of economicgrowth can be achieved and sustained byexploiting the resource base. The result can beillusory gains in income and permanent lossesin wealth.

Indeed, natural resource assets are legiti-mately drawn upon to finance economicgrowth, especially in resource-dependent coun-tries. The revenues derived from resourceextraction finance investments in industrialcapacity, infrastructure, and education. Areasonable accounting representation of theprocess, however, would recognize that onekind of asset has been exchanged for another,which is expected to yield a higher return.Should a farmer cut and sell the timber in hiswoods to raise money for a new barn, his pri-vate accounts would reflect the acquisition of anew asset, the barn, and the loss of an oldasset, the timber. He thinks himself better offbecause the barn is worth more to him thanthe timber. In the national accounts, however,income and investment would rise as the barnis built, but income would also rise as thewood is cut. The value of the timber, less thatof any intermediate purchases (e.g., gas and oilfor the chainsaw) would be credited to valueadded in the logging industry. Nowhere is theloss of a valuable asset reflected. This can leadto serious miscalculation of the developmentpotential of resource-dependent economies byconfusing gross and net capital formation.Even worse, should the proceeds of resourcedepletion be used to finance current consump-tion, then the economic path is ultimatelyunsustainable, whatever the national accountssay. If the same farmer used the proceeds fromhis timber sale to finance a winter vacation, hewould be poorer on his return and no longerable to afford the barn, but national incomewould only register a gain, not a loss inwealth.

Many countries now heavily burdened withdebt are resource-dependent: Mexico,Venezuela, and Nigeria are oil exporters, forexample. Their national balance sheets beforethe debt crisis deteriorated substantially as they

drew down natural resource assets and piledup external debt, using the proceeds of both tofinance consumption and subsidize investmentsof little or no economic value. A nationalaccounting system that drew attention to theirdeteriorating asset positions might have alertedpolicy-makers to the need for policy changesand international lenders to the growing risksof further exposure.

The fundamental definition of income encom-passes the notion of sustainability. In account-ing and in economics textbooks, income isdefined as the maximum amount that therecipient could consume in a given periodwithout reducing the amount of possible con-sumption in a future period. (Edwards and Bell1961; Hicks 1946) Business income is defined asthe maximum amount the firm could pay outin current dividends without reducing networth. This income concept encompasses notonly current earnings but also changes in assetpositions: capital gains are a source of income,and capital losses are a reduction in income.The depreciation accounts reflect the fact thatunless the capital stock is maintained andreplaced, future consumption possibilities willinevitably decline. In resource-dependent coun-tries, failure to extend this depreciation conceptto the capital stock embodied in naturalresources, which are such a significant sourceof income and consumption, is a major omis-sion and inconsistency.

This is not academic hairsplitting. Forresource-based economies, evaluations of eco-nomic performance and estimates of macroeco-nomic relationships are seriously distorted byfailure to account for natural resource deprecia-tion. In this report, Indonesia is used as anexample. Over the past 20 years, Indonesia hasdrawn heavily on its considerable naturalresource endowment to finance developmentexpenditures. Revenues from production of oil,gas, hard minerals, timber, and forest productshave offset a large share of governmentdevelopment and routine expenditures. Pri-mary production contributes more than 43 per-cent of gross domestic product, 83 percent of

exports, and 55 percent of total employment.(Table I.I.) Indonesia's economic performanceover this period is generally judged to havebeen successful: per capita GDP growth aver-aging 4.6 percent per year from 1965 to 1986has been exceeded by only a handful of lowand middle-income countries, and is far abovethe average for those groups. Gross domesticinvestment rose from 8 percent of GDP in1965, at the end of the Sukarno era, to 26 per-cent of GDP (also well above average) in 1986,despite low oil prices and a difficult debt situa-tion. (World Bank 1988)

Estimates derived from the Indonesian coun-try case study, presented in more detail in PartII of this report, illustrate how much thisevaluation is affected by "keeping score" morecorrectly. Table 1.2 and Figure 1 compare thegrowth of gross domestic product at constantprices with the growth of "net" domesticproduct, derived by subtracting estimates ofnet natural resource depreciation for only threesectors: petroleum, timber, and soils. It is clearthat conventionally measured gross domesticproduct substantially overstates net income andits growth after accounting for consumption ofnatural resource capital. In fact, while GDPincreased at an average annual rate of 7.1 per-cent from 1971 to 1984, the period covered bythis case study, our estimate of "net" domesticproduct rose by only 4.0 percent per year. If1971, a year of significant additions to petro-leum reserves, is excluded, the respectivegrowth rates from 1972 to 1984 are 6.9 percentand 5.4 percent per year, for gross and netdomestic product.

The overstatement of income and its growthmay actually be considerably more than theseestimates indicate since only petroleum, timber,and soils on Java are covered. Other importantexhaustible resources that have been exploitedover the period, such as natural gas, coal, cop-per, tin, and nickel have not yet been included inthe accounts. The depreciation of other renew-able resources, such as non-timber forest productsand fisheries, is also unaccounted for. Whencomplete depreciation accounts are available,

Table I.I. Direct Contribution of Primary Production (%)

Renewable ResourcesAgriculture

—Food crops—Other crops—Livestock

FishingForestrya

Exhaustible ResourcesOil & natural gasb

Other mining

Total Primary Sectors

a. Logs, sawn timber ancb. Includes crude oil and

products.

Source: Central Bureau of

Share ofGDP

GrowthRate

1983-1987

24.221.3

(14.8)(4.0)(2.5)1.71.2

19.718.50.8

43.9

plywood.condensates,

Statistics and

3.23.5

(2.2)(6.4)(6.6)0.62.5

3.02.95.6

3.1

natural gas, LNG

Bank Indonesia.

Share ofMerchandise

Export1987/88

30.413.7(0.6)

(12.3)(0.8)2.3

14.4

53.347.7

5.6

83.7

and LPG, but excludes

Share ofEmployment

1985

54.6

0.8

55.4

other oil

they will probably show a greater divergencebetween the growth in gross output and netincome.

Other important macroeconomic estimatesare even more badly distorted. Table 1.3 andFigure 2 compare estimates of gross and netdomestic investment, the latter reflectingdepreciation of natural resource capital. Thisstatistic is central to economic planning inresource-based economies. Countries such asIndonesia that are heavily dependent onexhaustible natural resources must diversifytheir asset base to preserve a sustainable long-term growth path. Extraction and sale of natu-ral resources must finance investments in otherproductive capital. It is relevant, therefore, tocompare gross domestic investment with thevalue of natural resource depletion. Should

gross investment be less than resource deple-tion, then, on balance, the country is drawingdown, rather than building up, its asset base,and using its natural resource endowment tofinance current consumption. Should netinvestment be positive but less than requiredto equip new labor force entrants with at leastthe capital per worker of the existing laborforce, then increases in output per worker andincome per capita are unlikely. In fact, theresults from the Indonesian case study showthat the adjustment for natural resource assetchanges is large in many years relative to grossdomestic investment. In 1971 and 1973, theadjustment is positive, due to additions topetroleum reserves.1 In most years during theperiod, however, the depletion adjustment off-sets a good part of gross capital formation. Insome years, net investment was negative. A

Table 1.2.

Year

19711972197319741975197619771978197919801981198219831984AverageAnnualGrowth

Comparison of GDP and "NDP"

GDPa

5,5456,0676,7537,2967,6318,1568,8829,567

10,16511,16912,05512,32512,84213,520

7.1%

a. In constant 1973 Rupiah,b. The flow of resources in

later induring

the text. Positivethe year.

In 1973 Rupiah

Net Changein Natural

Petroleum

1,527337407

3,228-787-187

-1,225-1,117-1,200-1,633-1,552-1,158-1,825-1,765

(billions)

Resource SectorsForestry

-312-354-591-533-249-423-405-401-946-965-595-551-974-493

Soil

- 8 9- 8 3- 9 5- 9 0- 8 5- 7 4- 8 1- 8 9- 7 3- 6 5- 6 8- 5 5- 7 1- 7 6

NetChange

1,126-100-2792,605

-1,121-684

-1,711-1,607-2,219-2,663-2,215-1,764-2,870-2,334

billions. From the Indonesian Central Bureau of Statistics.each sector is elaborated in thenumbers imply a growth in the

sections on

NDP

6,6715,9676,4749,9016,5107,4727,1717,9607,9468,5069,840

10,5619,972

11,186

4.0%

the specific resourcephysical reserves of that resource

fuller accounting of natural resource depletionmight conclude that in many years depletionexceeded gross investment, implying that natu-ral resources were being depleted to financecurrent consumption expenditures.

Such an evaluation should flash an unmistak-able warning signal to economic policy-makersthat they were on an unsustainable course. Aneconomic accounting system that does notgenerate and highlight such evaluations is defi-cient as a tool for analysis and policy inresource-based economies and should beamended.

The same holds true with respect to evalua-tion of performance in particular economic

sectors, such as agriculture. Almost three-quarters of the Indonesian population live onthe fertile but overcrowded "inner" islands ofJava, Bali, and Madura, where lowlandirrigated rice paddies are intensively farmed. Inthe highlands, population pressures havebrought steep hillsides into use for cultivationof maize, cassava, and other annual crops. Ashillsides have been cleared of trees, erosion hasincreased, now averaging over 60 tons per hec-tare per year, by our estimates.

Erosion's economic consequences include lossof nutrients and soil fertility from thin soils,and increased downstream sedimentation inreservoirs, harbors, and irrigation systems.Increased silt concentrations affect fisheries and

Figure 1. GDP and "NDP," in Constant 1973 Rupiah

14

13-

11.

1971

I

1972

I

1973 1974 1975 1976 1977 1978

Year

1979 1980 1981 1982 1983 1984

downstream water users. Although crop yieldshave improved in the hills because farmershave used better seed and more fertilizers, theestimates presented in Part II imply that theannual depreciation of soil fertility, calculatedas the value of the lost farm income, is about 4percent of the value of crop production, whichis as large as the annual production increase.In other words, these estimates suggest thatcurrent increases in farm output in Indonesia'suplands are being achieved almost wholly atthe expense of potential future output. Sincethe upland population is unlikely to be smallerin the future than it is now, the process of soilerosion represents a transfer of wealth from

Current increases in farm output inIndonesia's uplands are being achievedalmost wholly at the expense ofpotential future output.

the future to the present. By ignoring thefuture costs of soil erosion, the sectoral incomeaccounts significantly overstate the growth ofagricultural income in Indonesia's highlands.

Table

Year

19711972197319741975197619771978197919801981198219831984

a. In

1.3. Comparison of GDI and '

GDI3

8761,1391,2081,2241,5521,6901,7851,9652,1282,3312,7042,7833,7763,551

ResourceDepletion13

1,126-100-2792,605

-1,121-684

-1,711-1,607-2,219-2,663-2,215-1,764-2,870-2,334

constant 1973 Rupiah, billions.the IndonesianStatistics.

b. In

Central Bureau of

constant 1973 Rupiah, billions.

'NDI"

NDI

2,0021,039

9293,829

4311,006

74358-91

-332489

1,019906

1,217

From

Includes depletion of forests, petro-leum and the cost of erosion on tisland of Java.explained fully

These figures arein Part II.

he

A considerable and growing body of expertopinion has recognized the need to removethis anomaly from the accounting frameworkby accounting for depreciation of naturalresource assets like depreciation of other physi-cal capital. In the words of a recent treatise onthe measurement of economic growth, "Policy-makers need, among other types of informa-tion, a set or sets of accounts which describethe significant dimensions of the system forwhich they are responsible . . . a cogent argu-ment can be made for the view that the pres-ent set of national accounts provides anincreasingly deficient representation of the sub-stantive economic activities taking place withinthe system, and that many of these deficiencies

are capable of being remedied by using avail-able data." (Juster 1973, pp. 26-27)

In June 1985, the member governments ofthe OECD adopted a "Declaration on Environ-ment: Resources for the Future." Theydeclared that they will "ensure that environ-mental considerations are taken fully intoaccount at an early stage in the developmentand implementation of economic and otherpolicies by . . . [inter alia] ... improving themanagement of natural resources, using anintegrated approach, with a view to ensuringlong-term environmental and economic sus-tainability. For this purpose, they will developappropriate mechanisms and techniques,including more accurate resource accounts."(OECD, 1986)

Our Common Future, the 1987 report of theWorld Commission on Environment andDevelopment, stated, "Thus, figuring profitsfrom logging rarely takes full account of thelosses in future revenue incurred throughdegradation of the forest. Similar incompleteaccounting occurs in the exploitation of otherresources, especially in the case of resourcesthat are not capitalized in enterprise or nationalaccounts: air, water, and soil. In all countries,rich or poor, economic development must take fullaccount in its measurements of growth of theimprovement or deterioration in the stock of naturalresources." (Our Common Future, p . 52)

Similarly, academic experts (Stauffer, 1983)and such international agencies as the OECDhave recommended that capital consumptionallowances be extended to natural resourceassets, such as mineral deposits (OECD, 1986).The World Bank and the United NationsEnvironment Programme have emphasized thedeficiencies in the current accounting systemand have sponsored work on improvements.According to a recent Bank publication, "GDPis essentially a short-term measure of economicactivity for which exchange occurs in monetaryterms. It is of limited usefulness to gauge long-term sustainable growth, partly because naturalresource depletion and degradation are being

Figure 2. GDI and "NDI," in Constant 1973 Rupiah

3.5.

3 .

I I I I T I I I I I

1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

Year

ignored under current practices." (Lutz andEl-Sarafy 1988)

A number of OECD member governments,including Canada, France, Netherlands, Aus-tralia, and Norway, have carried on substantialstatistical work programs to compile accountson natural resource stocks and stock changes.France and Norway have made perhaps themost extensive official estimates, France'spatrimony accounts have emphasized thedevelopment of physical accounts. (Weber1983) Norway's resource accounts for energyand other significant economic resources have

stressed integration with macroeconomicmodels and budgets. (Alfsen, Bye and Lorent-sen 1987) There is a detailed estimate of thenational balance sheet of the United States thatincludes values for timber and subsoil assets,and an important study of national balancesheets covering twenty countries by the samescholar. (Goldsmith, 1982, 1985)

Within the last few years, governments indeveloping countries, recognizing their naturalresource dependence, have become interestedin a more adequate accounting framework. TheWorld Resources Institute is collaborating on

pilot studies with government research insti-tutes and statistical agencies in Indonesia,Costa Rica, and the People's Republic ofChina. Other governments considering newwork programs in natural resource accountinginclude Thailand, the Ivory Coast, and Argen-tina. Policy-makers in these countries recognizethe need for a planning tool that more effec-tively integrates economic and ecologicalconsiderations.

In filling this need, the United NationsStatistical Office has an important role to play.The U.N. System of National Accounts pro-vides a standard and model that, at least in itscore flow accounts, is closely followed by mostcountries. The U.N. Statistical Office is also aworldwide source of expertise and guidance inthe development of national income and otherstatistical systems.

The system of national accounts (SNA) pub-lished by the United Nations Statistical Office(United Nations 1968) is more complete withrespect to natural resource accounting than arethe accounting systems actually implementedby most national governments. The SNA pro-vides for balance sheets that record openingand closing stocks, and sources of increase anddecrease. Such accounts are included for repro-ducible tangible assets, such as tree planta-tions, and non-reproducible tangible assets,such as agricultural land and subsoil minerals.The criterion for inclusion in the SNA iswhether the assets are privately owned andused in the commercial production of goodsand services so that economic values can beestablished. Natural resources in the publicdomain, such as surface waters, atmosphere,and wilderness, are excluded on the groundsthat the SNA deals with the market economyand that the economic values of naturalresources outside the market system cannotreadily be established.

For natural resource assets included in theSNA, the accounting framework provides for"reconciliation accounts" that link balancesheet and flow accounts. These revaluation

accounts encompass changes in opening stocksdue to changes in prices during the period,and due to physical changes such as growth,discoveries, depletion, extraction, and naturallosses. The valuation principle endorsed by theUnited Nations for use in these accounts ismarket asset value, when possible. Whendirect asset value cannot be established, theU.N. guidelines endorse the economic asset-valuing principle discussed above: the presentvalue of the expected future income streamobtainable from the resource is the measure ofthe resource's asset value.

The U.N. Statistical Commission, advised bya number of expert working groups, is cur-rently considering changes in the SNA, as itdoes periodically. Dissatisfaction stems frommany inconsistencies and omissions in the cur-rent system. For example, production of goodsand services outside the enterprise sector, nota-bly by households, is largely omitted. Also,along with natural resources, other kinds ofcapital assets, such as knowledge and the stockof skills possessed by the workforce areignored. Furthermore, in the government sec-tor, the goods and services produced are notdirectly measured, but are valued at their fac-tor cost. These and many other deficiencieshave led to a long agenda of suggestedimprovements.

Although deliberations will continue until1991, the U.N. Statistical Commission has evi-dently already reached the decision that thereshould be no fundamental changes in the exist-ing SNA. The existing accounting methodologyis protected, in a sense, by its very inadequacy:wholesale reform is a large task, and improve-ment limited to just one aspect is hard tojustify when so many other problems wouldstill remain. Moreover, both at the national andinternational level, decisions regarding theaccounting system are in the hands of theproducers of statistics, not the users. Thenational income accounts are like sausages:there are many consumers, but few who wantto know how they are put together. Partly forthis reason, decisions are dominated by the

10

concerns of national income statisticians, whoare typically handicapped by shortages of staff,budgets, and raw data. These statisticians areresistant to recommending changes when somuch work remains to be done before the exist-ing SNA can be fully implemented.

National income accounts are likesausages: there are many consumers,but few who want to know how theyare put together.

With respect to depreciation accounts for nat-ural resources, therefore, the expert committeesof the U.N. Statistical Office have taken theposition that countries should be encouraged toimplement balance sheet accounts for reproduc-ible and non-reproducible tangible assets andlink those to conventional national incomemeasures through "satellite accounts," as indi-cated in the present system. In other words,their position is that depletion accounts for nat-ural resources should be calculated, but keptapart from the main tables. The measure ofdepreciation in the national income accountsshould not be extended to include naturalresources, and the present misleading indica-tors of economic performance should bemaintained.

The rationale for this position is pragmatic:until more national statistical offices are capableof estimating depreciation accounts for naturalresource assets, the core national incomeaccounts should not be modified. Any esti-mates of natural resource balance sheets anddepreciation should be displayed in ancillarytables, so that users can make their ownevaluations.

Therefore, from the statistician's perspective,the amount of effort required to implementnatural resource accounts is important. TheIndonesian country case study was implemented

partly to obtain first-hand information aboutthe level of effort needed to prepare numericalestimates. The accounts presented in thisreport were prepared almost entirely by pre-doctoral and master's level graduate students.Enough information to make reasonable esti-mates was found to be already available, sothat compilation and reorganization of datawere the main tasks. In this pilot study, with-out prior experience, working solely with exist-ing data (no fresh field surveys were con-ducted), without the access to data agovernment statistical office would have,researchers spent approximately 12 person-months mostly in the United States. This mod-est input generated estimates that shed sub-stantial new light on Indonesia's growth per-formance over more than a decade.

Only if the basic measures of economicperformance, as codified by the officialnational accounting framework, arebrought into conformity with a validdefinition of income will economicpolicies be influenced towardsustainability.

The importance of bringing such estimatesinto the main national income accounts, ratherthan relegating them to "satellite" or "recon-ciliation" accounts, is demonstrated by eventsof the past decade. While virtually all countriescalculate national income accounts, few haveimplemented the United Nation's recommenda-tions with respect to ancillary tables in theSNA because with limited resources they havehad to "stick to the basics." Similarly, despitetheir recognized deficiencies, politicians, jour-nalists, and even sophisticated economists inofficial agencies continue to use GDP growthas the prime measure of economic perfor-mance. (In the first statistical table of theWorld Bank's annual World Development Report,

11

for example, entitled, "Basic Indicators," theeconomic indicators are GDP, GDP growth percapita, and the rate of inflation.) Only if thebasic measures of economic performance, ascodified by the official national accountingframework, are brought into conformity with avalid definition of income will economic poli-cies be influenced toward sustainability.

There is ample time before the revisions tothe SNA are announced for the U.N. StatisticalOffice to explore fully the implications ofextending the concept of depreciation to naturalresource assets. It should use this time to pre-pare for that change. At the same time, keyinternational economic institutions, such as theWorld Bank, other multilateral developmentbanks, the IMF, and the OECD, should beginto compile, use, and publish revised estimatesof net national product and national income, asthis report has done. All these institutionsshould ready themselves to provide technicalassistance to the growing number of nationalstatistical offices that wish to adopt thesechanges and make such estimates forthemselves.

B. Current National IncomeAccounting1. Imputations and the Treatment ofDepreciation

The market economy—goods and servicesexchanged for financial consideration—broadlylimits the scope of national income accounts.For this reason, intra-household production andexchanges are excluded, except for subsistenceagricultural production. Nonetheless, theaccounts often do impute values to importanteconomic activities that take place without anymarket transaction. For example, the rentalvalue of owner-occupied housing is treated as ifthe owner rented the premises to himself. Thecriteria used to judge whether nonmarket activi-ties should be included in national accounts are(1) whether they are directly comparable toproduction taking place in the market, and

(2) whether their value can be reliably mea-sured, given the statistical resources.

An imputed value of particular concern hereis for the consumption of capital stock. Thevalue of capital goods, such as structures andequipment, declines over time with usebecause of physical wear and obsolescence.This gradual decrease in the future productivepotential of capital goods is reflected in thenational accounts by a depreciation allowancethat amortizes the asset's value over its usefullifetime. There are markets for some used capi-tal goods, such as vehicles, from whichdepreciation factors can be estimated. Other-wise, amortization is a surrogate measure forthe loss of income-generating capacity of olderassets. Straight-line depreciation and other for-mulas are imputations for this loss of value.

Depreciation of tangible reproducible capitalis subtracted from gross national product(GNP) in calculating the net national product(NNP) and national income. A nation mustinvest enough in new capital goods to offsetthe depreciation of existing assets if the futureincome-producing ability of the entire capitalstock is to be preserved. Therefore, accordingto the definition of income given above, thiscapital consumption allowance must beexcluded from total production. However, thisprocedure is applied only to structures andequipment, not to natural resources or othertypes of assets. NNP should provide a moreuseful measure of economic performance thanGNP but generally receives less attention ineconomic policy planning. As currently definedand estimated to include only buildings andequipment assumed to depreciate at fixedrates, gross and net product tend to moveclosely together. However, ignoring orunderestimating the deterioration or depletionof the capital stock can lead to economic policyerrors with serious, long-term consequences.

2. Income Statements and Balance Sheets

A complete system of financial accounts con-sists of two parts, one (the income statement)

12

dealing with transaction flows over a period oftime, and the other (the balance sheet) withstocks of tangible and financial assets at differ-ent points in time. The concepts of production,consumption, revenues and costs relate totransaction flows within accounting periods.The national economic accounts in which theyappear are comparable to income statements inbusiness accounting. In contrast, balance sheetscomprise stocks or levels of assets, liabilitiesand net worth at the end of accountingperiods. Flows and stocks are linked, in thatflows are equal to differences between stocks,and that stocks are equal to accumulated pastflows.

National balance sheets provide a picture of acountry's tangible and financial wealth atdifferent points in time, facilitating intertem-poral and international economic structuralcomparisons. The evaluation of a nation'sfuture potential for sustained income genera-tion can be enhanced by the detailed analysesof national assets and liabilities, through thepreparation of national balance sheets. In theUnited Nations' SNA, the importance of bal-ance sheets and wealth estimates for economicanalysis, are fully recognized, and the SNAincludes models and an explicit recommenda-tion to construct national balance sheets. How-ever, while neither business firms nor house-holds would ignore significant changes in theirbalance sheets, few national governments evencalculate theirs.

At least in concept, the United Nations hasendorsed accounting for certain naturalresources. SNA specifically includes forests andsubsoil assets (e.g., oil and gas reserves) inmodel national balance sheets. Two principalapproaches to valuing assets have beenendorsed for application to natural resources.These are (1) the use of values derived frommarket transactions in assets, and (2) the useof the discounted present value of estimatedfuture income flows derived from the assets tobe valued. For example, the SNA guidelines(United Nations 1977) suggest that the value oftimber tracts should be based upon market

data if available, taking account of timber typeand the situation and character of the land. Ifthere have been insufficient market transactionsin timber to provide estimates, standing timbershould be valued by discounting the futureproceeds of selling the timber at current pricesafter deducting management and harvestingcosts. An identical approach is suggested withrespect to subsoil assets, using as a discountrate a rate of return "expected by investors inmining or quarrying enterprises."

Neither the United Nation's SNA nor thenational income accounts of any country nowintegrates the treatment of natural resourcebetween income and balance sheet accounts.Final sales to consumers are included on theproduct side; on the income side, the valueadded from resource extraction is included inwages and salaries, in rental incomes and incompany profits. In other words, the totalvalue of current production, net of purchasedinputs, is imputed to current income.

There are no accounting entries in the flowaccounts for depletion, growth (in the case offorests), discoveries (in the case of subsoilassets) or asset revaluation due to pricechanges. Only capital investments in durablestructures and equipment used in the industryare subject to depreciation, not the resourcesthemselves. There is no depreciation factor inthe flow accounts to represent the loss offorests, the depletion of minerals, the erosionof soils, or the deterioration of water resources,even though these user costs impair the futureincome-generating capacity of those assets.

The U.N. recommends instead that these bal-ance sheet valuation adjustments should flowthrough reconciliation accounts and not thecurrent income accounts. The SNA guidelinessuggest, for example, that reductions in themarket value of land due to erosion bereflected in the reconciliation accounts. (UnitedNations 1977) An expert group of the UnitedNations has expressed general support for acalculation of the change in the value ofproven subsoil mineral reserves that would

13

include allowances for both depletion and newfinds, as well as the effects of price changes.This group recommended that the resultingadjustments also flow through the reconcilia-tion accounts (United Nations 1980), leavingGNP and NNP unadjusted.

In arguing for keeping such asset revalua-tions in satellite accounts, the U.N. guidelinespointed out that large and sudden revaluationsof subsoil asset values as a result of (1) exten-sive new discoveries; (2) changes in technologyincreasing the range of exploitable reserves; or(3) changes in market conditions couldmarkedly affect estimates of current income ifadmitted into the flow accounts. This positionignores the fact that changes in technology ormarket conditions can equally affect the repro-ducible capital stock. Energy price shocks, forexample, first made most older heavy indus-trial equipment economically worthless becauseat high energy prices those plants could notproduce at a profit. The same fluctuations inenergy markets led to drastic inflation, thendeflation, in real estate values in oil-producingregions, such as Texas. The income accountswere insulated from these changes in assetvalues only because depreciation rates are esti-mated at constant "book" values, a procedureequally applicable to natural resource assets.The impact of capital consumption allowancesfor natural resources on the national incomeaccounts would depend, as it should, entirelyon the importance of natural resources to theparticular economy.

In essence, reconciliation accounts provide ameans of recording changes in the value of netassets between successive measurement dateswithout having to show any effect on the income ofthe intervening period. Recording these adjust-ments in reconciliation accounts is likely tominimize their consideration in national policyanalysis. Therefore, while it is significant thatthe United Nations has specifically endorsedthe principle of valuing natural resource assetsand asset changes in the system of nationalaccounts, the procedure they have recom-mended would still leave the income account

seriously biased as an estimate of economicperformance.

C. The Scope of NaturalResource Accounting

A number of developed countries have pro-posed or set up systems of environmentalaccounts, including Norway, Canada, Japan,the Netherlands, the United States, andFrance. These systems have been reviewed indetail and evaluated for the United NationsEnvironment Programme by Weiller (1983) andFriend (1983). While natural resources take pri-ority in Norway and France, pollution andenvironmental quality have been the focus inthe United States and Japan. The approachesof Canada and the Netherlands combine ele-ments of both approaches.

In both Norway and France, extensive sys-tems of resource accounting have been estab-lished to supplement their economic accounts.The Norwegian system of natural resourceaccounting and the past decade's experiencewith it has recently been described. (Alfsen,Bye, and Lorentsen, 1987; Garnasjordet andSaebo, 1986) Accounts have been compiled for"material" resources, such as fossil fuels andother minerals, such "biotic" resources asforests and fisheries, and such "environmen-tal" resources as land, water, and air. Theaccounts are compiled in physical units ofmeasurement, and not integrated with thenational income accounts. However, resourceaccounts, especially those for petroleum andgas, have been expressed in value terms foruse in macroeconomic planning and projectionmodels maintained by the Central Bureau ofStatistics.

The French natural patrimony accounts areintended as a comprehensive statistical frame-work to provide the authorities with the factsand data they need to monitor the state andchanges in "that subsystem of the terrestrialecosphere that can be quantitatively andqualitatively altered by human activity."

14

(Corniere, 1986) They are conceptually broaderthan the national income accounts: materialand energy flows to and from economic activi-ties form only a subset of the accounts. (Com-mission Interministerielle des Comptes duPatrimoine Naturel, 1983) Methodology andempirical estimates have been under develop-ment since 1971, and they now cover the samerange of resources as Norway's: non-renew-ables, the physical environment, and livingorganisms. The basic accounting units arephysical, with provision for monetary valuationof stocks and flows that are marketed or con-tribute directly to market production. (Weber,1983)

The construction of such frameworks for thecompilation of environmental statistics maywell encourage decision-makers to consider theimpact of specific policies on the national stockof natural resources. However, a physicalaccounting approach by itself has considerableshortcomings. On the one hand, it does notlend itself to useful aggregation. Aggregatingwood from various species of trees in physicalunits (cubic meters) obscures wide differencesin the economic value of different species.Aggregating reserves of a mineral in physicalunits (tons) obscures vast differences in thevalue of different deposits, due to grade andrecovery cost. On the other hand, maintainingphysical accounts in disaggregated detailresults in a mountain of statistics that are noteasily summarized or used.

A further problem is that accounts main-tained in physical units do not enable eco-nomic policy-makers and planners to under-stand the impact of economic policies on anation's natural resources and thereby to inte-grate resource and environmental considera-tions into economic decisions—presumably, themain point of the exercise. While the informa-tion from the physical resource accountsundoubtedly facilitates the assignment ofmonetary values to balances and transactionflows (as will be described in this paper) fromthe perspective of economic policy, it is only anintermediate step. (Theys, 1984) Yet, there is

no conflict between accounting in physical andeconomic units because, as the Indonesian casestudy shows, physical accounts are necessaryprerequisites to economic accounts. If themeasurement of economic depreciation isextended to natural resources, physicalaccounts are inevitable by-products.

There is no conflict between accountingin physical and economic units becausephysical accounts are necessaryprerequisites to economic accounts. Ifthe measurement of economicdepreciation is extended to naturalresources, physical accounts areinevitable by-products.

Notwithstanding these points, there arelimits to monetary valuation, set mainly by theremoteness of the resource in question fromthe market economy. Some resources, such asminerals, enter directly. Others, such as sub-surface water, are extensively used as inputs tomarket production, and although they arerarely bought and sold, values can be readilyimputed. Others, however, such as noncom-mercial wild species, do not contribute directlyto production and can be valued in monetaryterms only through quite roundabout methodsinvolving numerous, somewhat questionable,assumptions. While methodological and empiricalresearch into the economic value of resourcesthat are remote from market processes is to beencouraged, common sense suggests that highlyspeculative values should not be included inofficial accounts.

In industrialized countries experiencingincreasingly acute problems of pollution andcongestion while becoming less dependent onagriculture, mining, and other forms of pri-mary production, the focus of attention has

15

been on "environmental" rather than naturalresource accounting. Since Nordhaus & Tobin(1973) proposed their "measure of economicwelfare" as an alternative to GNP, severaldifferent approaches to the development ofmore comprehensive systems of nationalincome accounting have been described that gowell beyond the scope of natural resourceaccounting described above. An excellentrecent survey of these approaches is available.(Eisner, 1988) Each reflects their authors' par-ticular concerns (e.g., Daly, in press; Hueting1980, 1984; Peskin 1980; Peskin & Peskin 1976,1978). For example, Herfindahl & Kneese (1973)considered how GNP might be modified by thecosts and benefits associated with pollutionand its abatement. Others have proposedgeneral systems to account for the impacts ofeconomic activities on the quality of theenvironment more broadly defined.

Problems with the current framework areobvious since they lead to bizarre anomalies. Iftoxic substances leak from a dumpsite to pol-lute soils and aquifers, measured income doesnot go down, despite possibly severe impair-ment of vital natural resources. If the govern-ment spends millions of dollars to clean up themess, measured income rises, other thingsequal, because such government expendituresare considered to be purchases of final goodsand services. If industry itself undertakes thecleanup, even if under court order, incomedoes not rise because the same expendituresare considered to be intermediate productioncosts if carried out by enterprises. If the site isnot cleaned up, and nearby households sufferincreased medical expenses, measured incomeagain rises because household medicalexpenses are also defined as final consumptionexpenditures in the national income accounts.

Although the system that gives rise to suchresults is widely regarded as faulty, there is littleconsensus on the remedy. Suggested approachescan initially be classified into those involvingphysical accounting and those that attempt toestablish monetary values. The physical approachrests on a straightforward extension of input-

output analysis to keep track of "deliveries" ofvarious material from various resource stocks toproducing and consuming sectors, and "deliver-ies" of materials from producing and consum-ing sectors to various receiving bodies in theenvironment. (Leontief, 1970; Kneese, Ayres, &d'Arge, 1970) Thus, for example, each industrialsector's discharges of waste materials to water,land, and air are estimated, along with eachsector's use of water, primary raw materials,land, and other natural resources.

This approach conceptually straightforwardand empirically feasible, has the virtue ofbringing common economic models of "pro-duction" and "consumption" into approximateaccord with the physical laws of nature. More-over, the data thus organized provides animportant intermediate step toward approachesthat do involve estimation of monetary values.However, the plausible assumption of approxi-mate linearity in the relation of waste genera-tion to production and consumption activitiescannot be carried over to the effects of emis-sions on environmental quality, or to theeffects of environmental quality on human wel-fare. Both of these linkages are often highlynon-linear, due to thresholds and chemical orbiological interactions.

Establishing monetary accounts for changesin environmental quality is by no means sostraightforward. While all would agree in prin-ciple that a good environment yields a continu-ing flow of beneficial goods and services, valu-ing those benefits is complex. For one thing,the existing accounts already reflect some ofthose values, but not others, so that there isthe danger of double-counting along with thatof omitting important elements of income.Agricultural output, yields, and income, forexample, already reflect the environmentalinputs of sunshine and precipitation, whichmake purchased inputs more productive.Increased concentrations of ozone and other airpollutants reduce agricultural yields and thusdiminish measured income in the existingaccounts. Environmental deterioration, insofaras it raises current production costs or reduces

16

productivity, is already reflected in theaccounts of the enterprise sector.

The glaring omission is the direct value ofenvironmental quality or quality changes to thehousehold. In principle, the damages to in-dividuals from increasing pollution, congestion,and noise can be estimated by measuring will-ingness to pay, lost productivity, or neededdefensive expenditures. Despite a large body ofresearch literature on methodological and sta-tistical problems, the task would be formidableif attempted on a national scale and remains inthe realm of research rather than accounting.

The notion of "defensive" expendituresis elusive, since spending on food canbe considered a defence against hunger,clothing a defence against cold, andreligion a defence against sin.

On the other side of the ledger, there are prob-lems—although perhaps not so serious—in improv-ing the accounting of expenditures undertakento prevent or remedy environmental damages.These problems can be brought into focus byassuming that households and enterprises areforced to spend more and more as the economygrows to maintain a constant level of environ-mental quality. (Juster, 1973) One anomalymight be addressed by treating such expendi-tures as intermediate purchases when under-taken by households and governments, as theynow are when undertaken by enterprises. How-ever, this immediately raises the broader ques-tion of treating as intermediate expenses a widerange of outlays by governments and house-holds that have the basic function of maintain-ing productivity (including, for example, trafficcontrol, health maintenance, and so on). Thenotion of "defensive" expenditures is elusive,since spending on food can be considered adefence against hunger, clothing a defenceagainst cold, and religion a defence against sin.

Another difficulty is in establishing theboundary between outlays to maintain environ-mental quality and those undertaken for otherpurposes. A household's purchase of a waterfilter, or a firm's installation of a water treat-ment plant, might be readily identified. How-ever, a household's move to another regionwith a superior environment, or a firm's adop-tion of an intrinsically low-residuals processtechnology would probably not.

There has been little consensus on the princi-ples or quantification of proposals for broaderenvironmental quality accounting so far,though the discussion has helped highlight theimportance of incorporating environmental pro-tection and effective natural resource manage-ment in national economic planning. However,for most developing countries and otherresource-based economies, it is more relevantto think of natural resources as productiveassets than as consumer goods. The first pri-ority is to account for those disappearing assetsin a way that gives due emphasis to the costs.

D. Setting Up Natural ResourceAccounts1. Physical Accounts

Natural resource physical stocks and anychanges in those stocks during an accountingperiod can be recorded in physical unitsappropriate to the particular resource. Thebasic accounting identity is that opening stocksplus all growth, increase or addition less allextraction, destruction, or diminution equalsclosing stocks. Although the following discus-sion refers to oil and gas reserves and timberstocks as examples, the principles are applica-ble to many other resources.

Oil and natural gas resources, the formermeasured in barrels and the latter in barrel-equivalents, consist of identified reserves andother resources and identified reserves can bedivided into proven reserves and probablereserves. Proven reserves are the estimated

17

quantities of oil and gas that geological andengineering data indicate with reasonable cer-tainty to be recoverable from known reservoirsunder existing market and operating condi-tions—that is, prices and costs as of the datethe estimate is made. Probable reserves arequantities of recoverable reserves that are lesscertain than proven reserves. Thus, one limiton the stock of reserves is informational. Addi-tional proven reserves can usually be generatedby drilling additional test wells or undertakingother exploratory investments to reduce uncer-tainty about the extent of known fields. Theboundary between reserves and other resourcesis basically economic. Vast quantities of knownhydrocarbon deposits cannot be extractedprofitably under current conditions. They arethus known resources, but cannot be countedas current reserves, though price increases ortechnological improvements might transformthem into reserves in the future.

For other mining industries, geological char-acteristics tend to be known with more cer-tainty, so there is less distinction betweenproven and probable reserves but a sharp divi-sion between economic reserves and totalresources. Many minerals are present at verylow concentrations in the earth's crust inalmost infinite total amounts. (Goeller & Wein-berg, 1984) Technological changes in miningand refining processes have markedly reducedthe minimum ore concentrations that canprofitably be mined, correspondingly expand-ing mineral reserves.

A similar framework is applicable to sub-soildeposits of water in available aquifers, exceptthat accounting for changes in stocks must takeinto consideration the annual recharge.Accounting for water quality changes encountersproblems that illustrate the limitations of physi-cal accounting. Quality changes can be reflectedin economic valuation rather readily, if theyaffect treatment costs or the economic uses towhich water can be put. However, the numer-ous dimensions of quality, reflecting contamina-tion by many other substances in varying con-centrations and combinations, makes the

construction of discrete physical categoriesdifficult.

Changes in oil and gas stocks may be classi-fied under various headings. Landefeld & Hines(1982) include under additions to reserves: "dis-coveries," the quantity of proven reserves thatexploratory drilling finds in new oil and gasfields or in new reservoirs in oil fields; "exten-sions," increases in proven reserves because ofsubsequent drilling showing that discoveredreservoirs are larger than originally estimated;and, "revisions," increases in proven reservesbecause oil or gas firms acquire new informa-tion on market conditions or new technology.Extensions of and revisions to oil and gasreserves have historically been significantlylarger than new discoveries. Landefeld & Hines(1982) point out that reserve statistics generallyproduce very conservative estimates of the totalresource stocks that will ultimately enter theeconomic system. Soladay (1980) estimated thatactual production from new U.S. fields andreservoirs was over seven times the amountinitially reported as discovered.

Reserve levels fall because of extraction anddownward revisions. In the United States, oiland gas companies are required by the Securi-ties and Exchange Commission to disclose netannual changes in estimated quantities of oiland gas reserves, showing separately openingand closing balances; revisions of previous esti-mates (from new information); improved recov-ery (resulting from improved techniques); pur-chases and sales of minerals in place; extensionsand discoveries; and, production. (FASB 1977)

The accounting framework for timberresources in physical units could be expressedin hectares, in tons of biomass, or in cubicmeters of available wood (Weber 1983), thoughthe last is probably the most important eco-nomic measure. As in the case of minerals, thetotal resource is larger than the economicreserve since a substantial part of the totalstock of standing timber in any country cannotbe profitably harvested and marketed with cur-rent technologies and market conditions.

18

Additions to the timber stock can originatefrom growth and regeneration of the initialstock, and from reforestation and afforestation.Reductions can be classified into production(harvesting); natural degradation (fire, insectinfestations, etc.); and, deforestation by man.Separate accounts might be established fordifferent categories of forests—for example, vir-gin production forests, logged (secondary)forests, protected forests, and plantations. Intemperate forests, where species diversity islimited, timber stocks are further disaggregatedby species.

Physical accounts can be constructed alongsimilar lines for agricultural land. Land and soilmaps and classification systems are used todisaggregate land into productivity categories.Changes in stocks of each land category withina period reflect various phenomena: conversionto non-agricultural uses; conversion to lowerproductivity classes through physical deteriora-tion by erosion, salinization, or waterlogging;and conversions to higher productivity classesthrough physical improvements by irrigation,drainage, and other investments. A set ofphysical accounts for agricultural land wouldrecord stocks of land at each accounting dateby productivity class, and flows among classesand to other land uses according to cause.

Similarly, physical accounts can be set up forother biological resources, such as wildlife orfish populations. The principles are essentiallythose of demography. Additions to initialpopulations are attributed to fertility, estimatedfrom reproduction rates and the size of thebreeding population, and inmigration. Subtrac-tions from stocks are attributed to natural mor-tality, estimated from age-specific or generalmortality rates, harvesting operations, otherspecial sources of mortality, and outmigration.

2. Valuation Principles

The concept of economic rent is central tonatural resource valuation. Economic rent isdefined as the return to any production inputover the minimum amount required to retain it

in its present use. It is broadly equivalent tothe profit that can be derived or earned from afactor of production (for example, a naturalresource stock) beyond its normal supply cost.For example, if a barrel of crude oil can be soldfor $10 and costs a total of $6 to discover,extract, and bring to market, a rent of $4 canbe assigned to each barrel.

Rents to natural resources arise from theirscarcity and from locational and other costadvantages of particular stocks. These rents aredistinct from monopoly rents, which increasereturns to a factor of production beyond itsopportunity cost by restricting supply throughmarket power or government action. In princi-ple, rents can be determined as the interna-tional resource commodity price less all factorcosts incurred in extraction, including a normalreturn to capital but excluding taxes, dutiesand royalties. Thus, the economic rent isequivalent to the net price.

This is the same concept of rent that appearsin a Ricardian scarcity model, which assumesthat resources from different "deposits" will besupplied at a rising incremental cost until profiton the marginal source of supply is completelyexhausted. In this Ricardian model, rents ariseon relatively low-cost, infra-marginal sources ofsupply.

It is also equivalent to a user cost in a Mal-thusian scarcity model, which assumes that ahomogeneous exhaustible resource is exploitedat an economically efficient rate, such that theprofit on the marginal amount brought to mar-ket is equal to the expected return derivedfrom holding the asset in stock for future capi-tal gain. (Hall & Hall 1984) In such a Malthu-sian model, if the resource is being extracted atan efficient rate, the current rent on the lastunit of resources extracted is thus equal to thediscounted present value of future returns froma unit remaining in stock.

As Ward (1982) has pointed out, the grossoperating surplus of the extractive sector in theSNA, represented by the sum of the profits

19

made by all the different enterprises involvedin resource extraction activities, does not repre-sent true rewards to factors of productionalone but also reflects rents from a "one timeonly" irredeemable sale of a non-renewablenatural asset. By failing to measure an appro-priate depletion allowance, conventionalnational accounting procedures allocate a dis-proportionate share of current income flows topresent generations at the expense of futuregenerations. The basic definition of income asthe amount that can be consumed withoutbecoming worse off is clearly being infringedas the value of the asset base declines.

Ward presents the sad exemplary tale ofKiribati, the small atoll republic of the SolomonIslands, which depended throughout the 20thcentury on its phosphate mines for income andgovernment revenues. While the mines ran,gross domestic product was high and rising,but the mining proceeds were treated as cur-rent income rather than as capital consump-tion. When the deposits were mined out in the1970s, income and government revenuesdeclined drastically because far too little hadbeen set aside for investment in other assetsthat would replace the lost revenues.

It would seem reasonable to apply this argu-ment, not only to all soil and subsoil assets,but also to tropical forests which, though theo-retically renewable, are being removed withoutadequate provision being made for theirreplacement in many areas. In forest eco-nomics, the concept of "stumpage value" isvery close to that of economic rent. Stumpagevalue represents timber sale proceeds, less thecosts of logging, transportation, and process-ing. Better quality and more accessible timberstands will command a higher stumpage value.

Asset transactions in natural resources, suchas competitive auction sales of rights to extracttimber or minerals, closely follow estimatedstumpage values or rents, with allowance forrisk. Because holders of those rights canusually hold the resources in stock or bringthem to market immediately, the current rent

or stumpage value tends to reflect the presentvalue of expected future net income that can bederived from them.

This principle is readily extended to otherresources: agricultural land can be valueddirectly on the basis of its current marketworth, or indirectly as the present value of thefuture stream of net income, or annual rent,that can be derived from it. The value of sub-surface irrigation water deposits can be esti-mated from market transactions in "waterrights," or by comparing the value of agricul-tural land overlaying a usable, known, aquiferwith that of otherwise equivalent land withoutsubsurface water. Alternatively, it can be esti-mated as the present value of future rents, cal-culated as the difference between the costs (percubic meter) of supplying the water for irriga-tion and the incremental net farm incomeattributable to the use of the water for irriga-tion. The value of a fishery could be estimated,in principle, as the maximum amount of reve-nue that a government authority could collectin bids from potential fishermen for the rightsto participate in the catch. Alternatively, itcould be estimated as the present value of thenet income fishermen could derive from thecatch under optimal regulation. In a world offrictionless, competitive markets, these valua-tion methods would yield the same results.

If adjustments to national incomeaccounts for natural resource stockchanges are to attain broad acceptance, acredible standard technique for valuingnatural resources must be adopted thatcan be applied to various resources bystatisticians in different countries.

If adjustments to national income accountsfor natural resource stock changes are to attainbroad acceptance, a credible standard technique

20

for valuing natural resources must be adoptedthat can be applied to various resources by sta-tisticians in different countries. That methodmust be as free as possible from speculativeestimates (about future market prices, for exam-ple) and must depend on underlying data thatis reasonably available to statistical agencies.

Landefeld & Hines (1985) have recently com-pared the three principal methods discussedabove for estimating the value of naturalresource stocks: 1) the present value of futurenet revenues; 2) the transaction value of mar-ket purchases and sales of the resource in situ;and 3) the net price, or unit rent, of theresource multiplied by the relevant quantity ofthe reserve.

The present value method requires thatfuture prices, operating costs, productionlevels, and interest rates be forecast over thelife of for example, a given oil field, after itsdiscovery. The present value of the stream ofnet revenue is then calculated, net revenuerepresenting the total revenue from theresource less all extraction costs. Soladay (1980)extends the present value method by attempt-ing to take into account the upward revisons inestimates of reserves that typically occur subse-quent to the initial discovery. The UnitedNations Statistical Office has recommended useof the present value method when marketvalues for transactions in resource stock are notavailable. (United Nations 1979)

The net price method applies the prevailingaverage net price per unit of the resource (cur-rent revenues less current production costs) tothe physical quantities of proved reserves andchanges in the levels of proved reserves. Lan-defeld and Hines make the important pointthat while the net price method requires onlycurrent data on prices and costs, it will beequivalent to the other two methods if outputprices behave in accordance with long-runcompetitive market equilibrium. "Equilibriumin natural resource markets (where the netprice rises in accordance with the rate of returnon alternative investments) produces the

interesting result that depletion as measured bychanges in the present value of the resourceequals depletion as measured by the net pricemethod." (Landefeld & Hines 1985, p. 14) Theassumption here is derived from the theory ofoptimal depletion of exhaustible resources, thatresource owners will tend to arbitrage returnsfrom holding the stock into future periods withreturns from bringing it immediately to market,adjusting current and future supplies untilprice changes equate those returns. (Dasgupta,1982) When expected future increases in thenet price take place at a rate equal to thereturn on alternative investments, theseincreases would therefore be eliminated in thecalculation of the net present value of futurecash flows. (Miller & Upton, 1985)

A number of recent studies (Boskin et al.1985; Landefeld & Hines 1982, 1985; Soladay1980; Ward 1982; Lutz and El-Sarafy 1988,Devarajan and Wiener 1988) have consideredthe issues associated with valuing the discov-ery and use of depletion of exhaustible naturalresources in measures of national income andwealth.

In the private sector, financial accounting andreporting for petroleum- and mineral- produc-ing companies has been debated for manyyears in the United States by the accountingprofession, regulatory agencies, industry groups,and the companies themselves. The U.S. Secu-rities and Exchange Commission (SEC) andFinancial Accounting Standards Board (FASB)have given extensive consideration to theappropriate accounting and financial reportingfor publicly traded corporations involved inextractive activities. The debate was initiallyfocussed on the two widely different methodsof reserve valuation used by companies, thefull cost method and the successful effortsmethod. (FASB 1977) Each was based upon thecosts of exploration and development actuallyincurred, but without reflecting the marketvalue of reserves or annual changes in reservesto which the company has rights of ownership.Believing with ample justification that neithermethod provided sufficient information to

21

stockholders, the SEC proposed a new methodof reserve recognition accounting (SEC 1978)that valued proven oil and gas reserves accord-ing to the discounted present value of thestream of future income at current prices andcosts. Following further debate on the issue,however, SEC abandoned this method ofaccounting since the burden of producing theinformation was considered to outweigh itsusefulness to users of financial statements.(SEC 1981)

The FASB recently considered means bywhich companies could provide informationabout future cash flows from oil and gasreserves as supplemental information, outsidethe financial statements. (FASB 1982) Theyevaluated the alternatives of fair market value,discounted future net cash flows, and a "stan-dardized measure" of discounted net cashflows. Fair market value was rejected on thebasis that relatively few exchanges of oil andgas mineral interests take place, and the geo-logical characteristics of each property areunique and thus incomparable. The use of dis-counted future net cash flows, based on esti-mated future prices and costs, production tim-ing, and an enterprise-specific discount rate asa surrogate for fair market value, was alsorejected since such subjectively based calcula-tions could not provide sufficiently comparableand verifiable information for financial report-ing. The Board settled on a standardized mea-sure of discounted net cash flows. Future netcash flows result from subtracting futuredevelopment and production costs (and taxexpenses) from future cash inflows relating toproved oil and gas reserves, using prices andunit costs as of the end of the reporting year.A discount rate of 10 percent is specified. TheFASB points out that the standardized measurecannot be considered an estimate of fair marketvalue but should reflect some of the key varia-bles that affect fair market value—such aschanges in reserve quantities, selling prices,production rates, and tax rates. Thus, the pri-vate accounting profession, after lengthy con-sideration, has adopted a valuation methodbased on the net price approach.

E. Integrating Natural Resourcesinto the National Accounts

Income accounts for natural resources can bedeveloped directly from accounts expressed inphysical units by assigning appropriate mone-tary values to stock levels and changes. Netchanges in the value of stocks are attributed tocurrent year additions (discoveries, net revi-sions, extensions, growth or reproduction) lessdeductions (degradation, deforestation, ordepletion) plus any price changes of theresource during the year, as illustrated in Table1.4. This framework is applicable with suitablespecification to a wide variety of resources.

If the primary objective were only thenational balance sheet presentation of naturalresource accounts, the example shown in Table1.4 would be relatively straightforward. The netvalue of the resource increased by $155($255-$100) during the year and net nationalwealth also increased by $155.

To adjust gross national product to a netbasis, economists have a number of options. Ifthe only desired adjustment to income were toreflect resource depletion, then net nationalproduct would be reduced by $32, using theaverage valuation rate of $1.60 per barrel. If allthe physical changes in the resource base werenetted, yielding a decrease of 15 physical units,NNP would be reduced by $24, at the sameaverage valuation. However, if the gain invalue of the opening stock due to pricechanges were also treated as current income,NNP would be increased by $155, the differ-ence between the two balance sheets.

In other words, alternative adjustments arepossible, depending on the objective. Treatingunrealized capital gains and losses due to pricechanges as income is consistent with the defi-nition of income given above, since the capitalgain during the year could be consumed with-out reducing future potential consumptionbelow what it would have been at the originalprice level. However, accounting conventions

22

Table 1.4. Example

Opening Stock

Additions:DiscoveriesRevisions (Net)ExtensionsGrowth*Reproduction*

Reductions:ProductionDeforestation*Degradation*

Net Change

Revaluations:Opening StockTransactions

Closing Stock

Note: Example of aThe resource

of Resource Inclusive National Acounting System

PhysicalUnits

100

20(30)1500

(20)00

(15)

——

85

natural resource account c

UnitValue

1.00

1.601.601.601.601.60

1.601.601.60

1.60

——

3.00

Value Basis of($) Calculation

100

32(48)2400

(32)00

(24)

200 100 x (3.00-1.00)J21) 15 x (3.00-1.60)

255

is it might appear in national economic accounts.unit value (based on international c

incurred in extraction or production) is assumed:ommodity prices less factor coststo be $1.00 at the beginning of the year,

$3.00 at the end of the year, and to average $1.60 during the year.The total increase in the value of the

($255-$100).resource

The methodology recommended inadjustment to net national product of ($24). The$179 ($155 + $24 or $200 - $21) wouldimpact on income of the current period

Items marked * are specific to forest isubsoil minerals, e.g., oil and gas.

over the period shown is equal to $155this paper would result in a downwardremaining net change in total value of

be recorded in a revaluation reserve and have no

"esources, all other categories are applicable to

now in use for physical plant and equipment resource stocks due to new finds, price changes,value assets at "book value" rather than and depletion should be excluded from thereplacement cost: they do not reflect changes in income accounts. At the opposite extreme,asset values in current income accounts. Eisner (1980, 1985) has argued that capital

revaluations in excess of those generated byThe United Nations (1977, 1980) has sug- general price rises should be included in mea-

gested that all changes in the value of natural sures of income and capital accumulation.

23

Accordingly, the money value of all capitalgains in excess of those necessary to keep thereal value of capital intact should be includedin income. Eisner (1980) extends this argumentto propose the inclusion of capital gains arisingfrom the discovery of new resources in income,and the exclusion of resource depletion fromincome.

International resource commodity prices aresubject to dramatic fluctuations over compara-tively short time periods because price elastici-ties of demand and supply are often small inthe short run. Including unrealized capitalgains from natural resource price changes incurrent income could lead to significant swingsin income between successive periods inresource-dependent countries. However, natu-ral resource price swings (such as the energyprice shocks) also markedly affect the value ofplant, equipment, and real estate that arespecific to those natural resource sectors.

The procedures illustrated here, which incor-porate the net price method (Landefeld &Hines 1982, 1985), include only the value ofphysical resource stock changes in nationalincome. This procedure is consistent with cur-rent asset-accounting practices. In addition, itis more readily implemented since, for mostnatural resources, information on stock changesdue to extraction or discovery is more accuratethan information on the size and compositionof the total stock. At the end of each account-ing period, the physical units comprising theopening balance of natural resource stockshave been revalued at the net price prevailingat the end of the period. The revaluationadjustment (which, in the example shown inTable 1.4, equals $200) has been recorded in arevaluation reserve and therefore has no effect onthe current period's income. The net physical

change (15 units) is valued at the average netprice prevailing during the period and is usedto adjust NPP downwards by $24. The remain-ing revaluation adjustment, which arises fromthe difference between average and closingprices applied to the net physical change inresource stocks ($21) is recorded in the revalua-tion reserve as an unrealized capital gain, withno impact on income.

If national accounts are adjusted to showincome at constant prices, thereby eliminatingthe effects of general inflation, the adjustmentto income ($24) should be deflated by anappropriate price index. As a result, only realwealth increases or decreases will be reflectedin measured national income.

Preliminary resource accounts in physical andvalue terms for tropical timber, petroleum, andsoil resources in Indonesia from 1970 to 1984illustrate this methodology. The net changes inresource values from physical sources (forexample, excluding price revaluations) impliedby these tables were reflected in the summarytables and figures presented earlier to illustratethe usefulness of such calculations in macro-economic evaluation. The resource accounts arepreliminary, in that they have not beenendorsed by official Indonesian statistical oreconomic agencies, but represent a non-govern-mental research effort that drew on publishedand some unpublished statistical sources.Efforts are currently under way in cooperationwith the Ministry for Environment and Popula-tion and a consortium of universities inIndonesia to revise these accounts and toextend the methodology to other resourcesectors.

Details of the estimates are provided in PartII of this report.

24

Note

1. It may seem anomalous that in 1971 and1973 depreciation was a negative number,that is, net capital consumption was added togross domestic product and investment. Thereason for this is that the value of additionsto petroleum reserves in these years wereconsiderably larger than all categories ofdepletion combined, leading to "negative"depreciation.

One way of resolving this apparentanomaly would be to account separately foradditions and subtractions from naturalresource assets. Real capital gains (as distinctfrom those resulting from price changes) can

be accounted for as gross income and grosscapital formation. This is consistent with ourearlier definition of income, because addi-tions to resources during the current yearaugment the amount that could be consumedcurrently without reducing potential con-sumption in future years. This is obvious inthe case of forest growth, but less obviousfor mineral discoveries, since current discov-eries may leave less to be discovered lateron. However, insofar as additions to mineralreserves reflect advances in the technologyof exploration or extraction, the total poten-tial resource base will have expanded.

25

II. The Indonesian ResourceAccounts

A. Timber Resource Accounts,1970-1984

P reliminary accounts in physical andvalue terms for Indonesia's timberresources were estimated using the

methodology explained in Part I. The accountsdo not represent the full value of Indonesia'sforest resources, which yield such importantnon-timber commodities as rattan, oils, resins,foodstuffs, and pharmaceutical products, andwhich also provide important ecological services.In principle, the values of non-timber forestcommodities enter into gross domestic product,though in practice they are greatly underesti-mated. Exports alone of non-timber productsreached U.S. $120 million in 1982. Full account-ing for deforestation would include the presentvalue of foregone future income from thesenon-timber forest products and services. Indo-nesian forests are mostly equatorial rainforestsand tropical semi-deciduous (dipterocarp andmixed dipterocarp) forests, but also includeswamp and mangrove forests along the coastsof Sumatra and Kalimantan and small areas ofpeat forests. Before World War II, timber pro-duction was concentrated on Javanese teakplantations, but after 1967 timber extractionincreased rapidly from extensive primary low-land rainforests in Kalimantan, Sumatra,Sulewesi, and Irian Jaya. Indonesia joinedMalaysia and the Philippines as Southeast

Asia's leading log exporters, accounting togetherfor 80 percent of world tropical hardwoodexports. Indonesia's share of world exports rosefrom 1 percent in 1964-66 to a peak of 31 per-cent in 1979-81, when timber was the country'ssecond largest export commodity in terms ofgross receipts. After 1980, government restric-tions to promote domestic processing reducedlog exports, and replaced them with increasingvolumes of lumber and plywood.

All Indonesian natural forests are state-owned and administered by the Ministry(formerly Directorate General) of Forestry.While the basic forestry law acknowledges thetraditional rights of indigenous communities, inpractice these adat rights are honored more inthe breach than in the observance. On theOuter Islands, management and harvesting ofmost tracts are contracted to private com-panies, subject to regulation under the BasicForestry Act of 1967, which prescribes goodtimbering and ecological practices. Up till now,the forestry agencies have been unable toenforce these prescriptions effectively. Virtuallyall concessionaires have nominally adopted theIndonesian Selective Cutting System, whichspecifies minimum tree sizes harvested andnumbers, spacing, and size classes of residualtrees per hectare, along with the allowable cut,but few actually follow it faithfully.1 Forestmanagement and policy in Indonesia haveposed difficult development problems.2

27

1. The Physical Accounts

While stock estimates at different points intime provide consistency checks, an estimate ofthe physical growing timber stock was essentialfor only one benchmark year during theperiod. Stocks for the remaining years werecomputed from the respective annual net addi-tions and reductions, for which estimates (ofvarying quality) were available.

Estimates of the total forest land area varyconsiderably among sources. An estimate for1985 by the Directorate General of Forestry,presented in Table II. 1, puts the total forestarea at 143 million hectares, nearly three-quarters of Indonesia's total land area. How-ever, this estimate included land within classi-fied forest boundaries designated as "conver-sion forest," much of which had already been

deforested. A 1981 FAO study (Table II.2) givesa figure of 158.2 million hectares, of which113.9 million are closed forests.3 A more recentassessment by the Land Resources Develop-ment Center and the Ministry of Transmigra-tion, using aerial photographs dating mostlyfrom the early 1980s, roughly agree with FAOtotals for the islands already covered. Aregional breakdown shows that only 1 percentof forest land is in the densely populated InnerIslands of Java and Bali, while Sumatra,Kalimantan, Irian Jaya, and Sulewesi accountfor 90 percent.

a. Growing Stock

Estimates of the total growing stock are de-rived from concession surveys carried out inthe 1970s and from more recent provincial sur-veys. Although such inventories, carried out at

Table II.1. Indonesian Forest Resources: Department of Forestry Classification 1985(million hectares)

1. Total Land Area

2. Total Forest Area *

3. Elements of Forest Area *a. Protection Forestb. Nature Conservation and Tourism Forestc. Production Forest (available for commercial harvesting)

i) Limited Productionii) Permanent Production

d. Production Forest that may be converted to non-forest purposes

4. Area Awarded to Concessionaires or in Process of Awarda. Area under Concession (holders of Forest Exploitation rights)b. Areas under Forestry Agreements (last step prior to award of rights)

193.6

143.0

30.319.064.0(30)(34)30.0

65.452.213.2

* Total Rain Forest AreaTotal Swamp Forest AreaTotal Secondary Forest AreaOther Forest Area

82.212.014.634.2

Sources: Forest Area: 1985 Departamen Kehutanan, Draft Long-Term Forest Plan (Jakarta, January1985, p. 17). Concession Areas: P.T. Data Consult, 1983.

28

Table II.2. Indonesian Forest Resources FAO Classification, 1980 and 1985 (million hectares)

Total Area, Natural Woody Vegetation

A. Closed Forestsa

1. Productive Forestsb

a. Undisturbed Forestb. Logged Forest

2. Unproductive Forestc

a. For Physical Reasonsb. For Statutory Reasons

(parks, reserves)

B. Open Forestd

C. Fallowse

D. Shrub Formations 23.9 23.9 n.a.

1980

158.2

113.973.738.934.8

40.234.75.4

3.0

17.4

1985

157.1

110.967.733.034.7

43.2n.a.n.a.

2.8

19.5

Indonesiaas % of allSoutheastAsia, 1985

n.a.

61.858.556.668.6

19.7n.a.n.a.

n.a.

n.a.

a. Closed forests are those that have not been recently cleared for shifting cultivation or heavilyexploited. In closed forest formations, tree crowns, underlayer and undergrowth combine toclose off most of the ground from light so that continuous grass cover cannot develop.

b. Productive forests are those from which it is both physically and legally possible to producewood for industry.

c. In unproductive forests, timber is not exploited for statutory reasons, or because harvestingis infeasible due to difficult terrain or stand conditions.

d. Open forest formations are marked by continuous grass cover on the ground.e. Fallow refers to secondary vegetation following the clearing of forests.

Source: Food and Agriculture Organization of the United Nations, Forest Resources of TropicalAsia, Rome, FAO, Tropical Forest Assessment, 1981, p. 40, pp. 211-237, 277-313,391-416.

different times and by different methods, crown point of first main branch) of all livingaren't fully comparable, they have been made trees more than 10 centimeters in diameter atas consistent as possible through cross- breast height.checking and adjustment. The measure ofstocking volume used is 'volume over bark' Regional VOB values were obtained in the FAO(VOB): gross volume in m3 per hectare over study by comparing results from sample areasbark of free bole (from stump or buttresses to with more complete data from the Malaysian

29

national inventory.4 Stocking rate estimates of323 m3/ha for virgin forests (49 percent), 204m3/ha for logged forests (17 percent), and 198m3/ha for unproductive forests (34 percent)were applied to Sumatra and Kalimantan.These estimates were reduced by 15 percent forSulawesi, Maluku, and Nusantenggara, and by25 percent for Irian Jaya. Estimated stockingrates are then applied to data on forested areasby island and category. The total growing stockin natural forests at the end of 1980 comes to24,248 million cubic meters, an average stock-ing rate of 212.9 m3 per hectare.

A 1980 closing stock estimate of 24,248 mil-lion m3, reflecting the VOB measure, has beenincluded in the timber resource accounts fornatural forest. This includes 10,311 million m3

of virgin forests, 6,911 million m3 of loggedforests, and 7,026 million m3 of unproductiveforests.

Another measure, 'volume actually commercial-ized' (VAC), describes the "volume under barkof logs commercially exploitable actuallyextracted from the forest," and it has been esti-mated only for virgin productive forests wherethe volume extracted per hectare is generallywell-known. The average volume of commer-cial timber remaining in logged-over forests isdifficult to estimate. VAC has been estimatedat 20 m3/ha for Irian Jaya, 25 m3/ha forSulewesi, and 45 m3/ha for Sumatra andKalimantan. Compared to VAC measures inMalaysia and the Philippines, these rates arelow, reflecting the more selective logging prac-tices in Indonesian forests. The value actuallycommercialized depends in part on the systemof forest taxes and royalties, which influencesthe degree to which concessionaires limit har-vesting to the most valuable trees. The averageannual commercial production during the years1974-80 was actually 40 m3/ha.

Government-sponsored plantation programsimplemented by the Directorate for Reforesta-tion, the parastatal timber company, and tim-ber concessionaires now cover significant areasin most provinces. Plantations established by

concessionaires in response to financial incen-tives are largely pine, and many are of ques-tionable commercial value. The Directorate'splantations include fodder and fruit trees andhave had a lower survival rate.

Estimates of the area and volume of forestplantations are uncertain, partly due to unrelia-bly reported survival rates on planted areas.The United Nations Food and AgricultureOrganization estimated that "successfullyestablished and reasonably stocked" planta-tions had a total estimated area of 1,918,000 ha(1,446,000 industrial and 472,000 non-industrial)in 1980.5 The stocking of plantations has beenassumed to be at an average rate of 100 m3/ha,yielding 192 million m3 of growing stock at theend of 1980. This figure has been added to thenatural forest stocks to derive the 1980 resourceaccount closing balance of 24,440 million m3 ofstanding timber.

b. Growth and Reproduction

An average annual increment in volume ofall trees in the forest can be expected onlyfrom disturbed or managed forests since undis-turbed forests have reached their climax equi-librium. No detailed information on growthrates of disturbed natural forests is available forIndonesia, but estimates for dipterocarp forestselsewhere in the region suggest annual growthin commercial species between 1 and 2 m3/ha/year.6 Commercial growth in the forests ofSulewesi and Irian Jaya, where stocking ratesare relatively low, must be lower. AnotherFAO study indicates an annual net incrementof 1.3 m3/ha.7 The 34.6 million ha of loggedforests were estimated to carry a timber vol-ume of 6,911 million m3 at an average stockinglevel of 200 m3/ha. This study assumes agrowth rate of 1.5 m3/ha for these forests,yielding an annual increase in volume of 51.9million m3, which corresponds to an annualbiomass increment of 0.75 percent. This figurehas been used in the timber resource accounts.

An estimate of annual increase in plantationtimber volume can be developed from the

30

plantation species' growth rates (expressed asthe mean annual increment at rotation age)and the distribution of industrial plantationareas by species reported by the FAO, asfollows:8

Species

Tectona grandisPinus merkusiiOthers

Area ofEstablishedIndustrialPlantation

(1980)(thousand ha)

861390195

1,446

Mean AnnualIncrement(M.A.I.)

(m3/ha/year)

8.518.0n.a.11.5*

* Approximate weighted average M.A.I, basedon Tectona grandis and Pinus merkusii only.

FAO (1981) estimates that 542,000 ha of theindustrial plantation area of 1,446,000 ha (i.e.,37.5 percent) was established during 1976-80.The physical stock of industrial plantations atthe end of 1976 can therefore be estimated asthe remaining 904,000 ha. Assuming the samerate of increase in non-industrial plantations,the 1976 total plantation physical stock can beestimated as 1,199,000 ha (62.5 percent of1,918,000). The growth in plantation area forthe remaining years during the study period(1970-82) has been estimated by assuming alinear growth rate, and the volume change byapplying the annual growth rate of 11.5m3/ha/year calculated above.

c. Harvesting

Figures for the total log harvest are reportedby the Directorate General of Forestry in theannual report on Indonesian forestry statistics.They may be underestimates, due in part toconsiderable log smuggling and underinvoicingof exports to avoid export restrictions, taxes,and royalties. The recorded harvest rises to apeak of more than 25 million m3 in 1979 and1980, then declines due to export restrictions

and domestic processing requirements. Alterna-tive World Bank estimates, not incorporatedhere, place the annual harvest in the 1980s atabout 25 million m3 per year. The log output iscomposed of meranti, kerning, ramin, teak, and afew other species, extracted predominantlyfrom Kalimantan and Sumatra during theperiod reviewed. These harvest figures areentered directly into the physical accounts.

d. Deforestation and Degradation

Deforestation denotes transfers of forestlands to other uses, including shifting and per-manent cultivation, reservoirs, and other infras-tructure. The area deforested annually inIndonesia is the highest in the region, duemainly to agricultural conversion. The nationaltransmigration program settled 50,000 familiesfrom the Inner Islands on Sumatra andKalimantan between 1974 and 1978, each on 5hectares of land, and moved about 300,000households between 1979 and 1984. Between1980 and 1986, the government cleared about800,000 hectares of land for transmigrants, ofwhich perhaps 600,000 was logged or secon-dary forest. In addition, about 330,000 hectaresof land still forested were allocated to trans-migrants and will probably have been clearedby the end of the decade. (World Bank, 1987)Land clearance by spontaneous migrants isthought to be of the same order of magnitudeas clearance by sponsored transmigrants. Otherplanned deforestation, largely in designated"conversion forests" for estate crops and otherdevelopment projects, is estimated at about100,000 hectares per year in the 1980s. (WorldBank, 1987)

An estimated 10 to 12 million people on theOuter Islands subsist by shifting cultivation,largely on Kalimantan, Sumatra, Sulewesi, andMalaku. Most of the area affected has beenreduced to poor secondary forest and scrub orconverted to grassland. An estimated 20 per-cent of the total land area in Kalimantan hasbeen affected by shifting cultivation, and 14percent in Irian Jaya. Expansion of shiftingcultivation is most rapid in logged productive

31

forest and slowest in unproductive closed for-est, due to differences in accessibility.

Taking into account all causes, an FAO studyestimated that,9.27 million hectares weredeforested between 1950 and 1977, at annualrates of 550,000 ha/year during the 1970s andrates of 600,000 in 1981 and 1982.9 Estimatesfor the mid-1980s compiled by a recent WorldBank assessment suggest a higher figure ofover 700,000 hectares annually, but the moreconservative FAO figure is used in the timberresource accounts. If a stocking rate of 200m3/ha (the average stocking rate for secondaryforests) is used, annual volume losses of 110and 120 million m3 are implied for the 1970sand 1980s respectively.

Degradation refers to forest deterioration dueto such natural disasters as fires, earthquakes,and pests, and due to destructive exploitationof forest resources in logging operations, graz-ing, and fuelwood collection. Intensive loggingin Indonesia has been estimated to damage upto 40 percent of the residual trees. Loggingdamage has been estimated through a residualbalance equation that equates the differencebetween stocking rates on virgin and loggedforest to harvest removals and logging damage.The resulting estimate of 79 m3 per hectare isconsistent with the figure of 40 percent of vol-ume remaining after harvest. This calculationyields a ratio of 1.98 m3 damaged for everycubic meter harvested, and this ratio isassumed in the accounts to hold in each year.

Fires are also an important factor. The ElNino perturbation in 1982-1983 provokedsevere drought, and led to disastrous forestfires in Kalimantan and neighboring Sabah in1983-84, especially in logged-over areas litteredwith dead trees and branches. The damagedarea in Kalimantan has been estimated at800,000 ha of primary lowland rainforest,550,000 ha of peat forest, and 1,200,000 ha ofselectively logged primary forest. In addition,750,000 ha of swidden area was affected, bring-ing the total to 3,700,000 ha. (Prance 1986)Sampled mortality rates in the burned area

averaged 60 percent for small trees and 25 per-cent for those greater than 30 cm./dbh. Makingthe most conservative assumption that only 25percent of timber resources in the areas burnedwere lost, the fire still cost over U.S. $3.5 bil-lion in timber assets.

Based on results of a consultant study, firesin the preceding five years consumed an aver-age of 60,000 hectares, mostly in secondaryforests. Taking this rate to represent normalfire losses in other years adds 14 million m3 inannual timber losses to the accounts.

e. Summary of Physical Accounts

These categories constitute the principalsources of increase and decrease in Indonesia'sforest resources. Together with the 1980 bench-mark estimates of growing stock, they permitconstruction of the physical accounts presentedin Table II.3 for the years 1970 to 1984. Theyimply a cumulative net decrease in growingstock of 1,866 million m3, about 7.2 percent ofthe total standing timber resource in 1970.Losses due to deforestation and degradationappear to have been several times greater thanthose due directly to timber harvests, but itmust be remembered that logging roadsincreased access for settlers and accelerated for-est conversion. Throughout the period, directharvest volume appears to be less than totalannual growth, but when associated loggingdamages are also considered timber operationshave resulted in losses that exceeded growth.Since the growth of commercial species is esti-mated to be less than 1 m3 per hectare, selec-tive cutting has unambiguously reduced theforest in value.

2. The Value Accounts

a. The Measure of Economic Rent

The relevant measure of economic value tobe applied to these changes in the physicalresource base is the value of the standing tim-ber prior to any value added by processing.Timber's economic rent corresponds to its

32

Table II.3. Forestry

PHYSICAL UNITS

Opening StockAdditions

GrowthReforestation

ReductionsHarvestingDeforestationLogging DamageFire Damage

Net Change

Closing Stock

Accounts 1970-1984

(million cubic meters)1970

25,773.1

51.91.3

10.0110.019.814.0

-100.6

25,672.5

UNIT VALUES (US$/m3)FOB Export Price 10.90Costs 4.90Primary Rent 6.00Secondary Rent 3.00

MONETARY ACOUNTS (US$ millions)Opening Stock — :Additions

Growth 155.7Reforestation 0

ReductionsHarvestingDeforestationLogging DamageFire Damage

Net Change

Revaluation8

Closing Stock

60.0330.059.442.0

-335.7

—

108,335.7 1

197125,672.5

51.93.4

13.8110.027.314.0

-109.8

25,562.7

15.106.808.304.15

1.08,335.7

215.40

114.5456.5113.358.1

-527.0

41,537.9

L49,346.6

a. The Revaluation category accounts for changesdue only to differences in the rental rates.

197225,562.7

51.95.5

16.9110.033.414.0

-116.9

25,445.8

17.107.909.204.60

149,346.6

238.70

155.5506.0153.664.4

-640.8

16,204.5

164,910.4

in the value of

197325,445.8

51.97.6

26.3110.051.914.0

-142.7

25,303.1

29.3013.1816.128.06

164,910.4

418.30

424.0886.6418.3112.8

-1,423.4

124,059.3

287,546.3

the overall stock

197425,303.1

51.99.7

23.3110.046.014.0

-131.7

25,171.4

41.6018.7222.8811.44

287,546.3

593.70

533.11,258.4

526.2160.2

-1,884.2

120,609.6

406,271.7

which are

33

Table II.3. (cont.) Indonesian Forestry

PHYSICAL ACCOUNTS

Opening Stock

AdditionsGrowthReforestation

ReductionsHarvestingDeforestationLogging DamageFire Damage

Net Change

Closing Stock

UNIT VALUES (US$/m3)FOB Export PriceCostsPrimary RentSecondary Rent

(million cubic1975

25,171.5

51.911.8

16.3110.032.214.0

-108.8

25,062.6

26.4011.8814.527.26

Accounts

meters)1976

25,062.6

51.913.8

21.4110.042.314.0

-122.0

24,940.6

44.7020.1224.5812.29

MONETARY ACCOUNTS (US$ millions)Opening Stock 406,271.7 256,848.6

AdditionsGrowthReforestation

ReductionsHarvestingDeforestationLogging DamageFire Damage

Net Change

Revaluation3

Closing Stock

376.80

236.7798.6233.8101.6

-993.9

148,429.2

256,848.6

637.90

526.01,351.9

519.9172.1

-1,932.0

177,981.9

432,898.5

197727,940.6

51.915.9

22.2110.043.814.0

-122.2

24,818.4

47.5021.3826.1213.06

432,898.5

677.80

579.91,436.6

572.0182.8

-2,093.5

27,151.5

457,956.5

197824,818.4

51.918.0

24.2110.047.814.0

-126.1

24,692.2

46.7021.0525.6512.82

457,956.5

665.60

620.71,410.8

613.0179.6

-2,158.5

-8,211.7

447,586.3

197924,692.3

51.920.1

25.3110.050.014.0

-127.3

24,565.0

85.2129.8455.3727.68

447,586.3

1,436.90

1,400.93,045.31,384.3

387.6

-4,781.3

518,669.0

961,474.0

34

Table II.3. (cont.) Indonesian Forestry

PHYSICAL ACCOUNTS

Opening Stock

AdditionsGrowthReforestation

ReductionsHarvestingDeforestationLogging DamageFire Damage

Net Change

Closing Stock

UNIT VALUES (US$/m3)FOB Export PriceCostsPrimary RentSecondary Rent

(million cubic198024,565.0

51.922.1

25.2110.049.814.0

-125.0

24,440.0

106.9334.2472.6936.34

Accounts

meters)198124,440.0

51.924.2

16.0120.031.614.0

-105.5

24,334.5

95.8437.9357.9128.95

MONETARY ACCOUNTS (US$ millions)Opening Stock 961,474.0 1,256,046.9

AdditionsGrowthReforestation

ReductionsHarvestingDeforestationLogging DamageFire Damage

Net Change

Revaluation"

Closing Stock 1

1,886.30

1,831.83,998.01,810.0

508.8

-6,262.3

300,835.2

256,046.9

b. The value for fire damage in 1983 isassuming mortality rates of 25%.

1,502.80

926.63,474.6

915.0405.4

-4,218.8

-255,561.3

996,266.8

198224,334.5

51.926.3

13.4120.026.514.0

-95.7

24,238.8

100.5941.0059.5929.79

996,266.8

1,546.40

798.53,575.4

789.6417.1

-4,034.2

28,727.0

1,020,959.6

made up entirely of estimates

198324,238.8

51.929.6

15.2120.030.0

153.8b

-237.5

24,001.3

78.7543.3135.4417.72

1,020,959.6

919.70

538.72,126.4

531.63,870.9

-6,148.3

-413,762.7

601,049.0

198424,001.3

51.935.3

16.0120.031.614.0

-94.4

23,906.9

93.1551.2341.9220.96

601,049.0

1,087.80

670.72,515.2

662.3293.4

-3,053.8

109,775.0

707,770.2

for the Kalimantan fire,

35

"stumpage value," the market value of stand-ing trees. With full knowledge of the resourceand competitive bidding, this is the maximumamount potential concessionaires would pay forharvesting rights. Since the Indonesian govern-ment has not administered its forest resourceso as to recover a large fraction of these rentsfrom concessionaires, stumpage value must beestimated by the net price method describedearlier—by subtracting costs of extraction andtransportation to the port from the exportvalue of the timber.

The export value has been measured directlyby the free on board (f.o.b.) export unit value,which is simply the ratio between gross exportreceipts and the volume of log exports. Thisfigure is a conservative estimate of timbervalue because log exports were considerablyunderinvoiced throughout the period to avoidexport taxes. This value is applied to timberextracted for domestic processing as well sinceit is the relevant measure of economic opportu-nity cost.

Published information on extraction andtransportation costs were sparse. Averageproduction cost estimates were available for1973 and for 1980.10 Deducting these total costestimates from the respective weighted averagef.o.b. price per m3 of exported logs (the 'exportunit value') yielded estimates of average stum-page values or rents per m3 for those two yearsfor logs actually harvested. In 1973, thismethod produced a rent figure of $16.15/m3,based on an f.o.b. unit export value of $30.34less a production cost estimate of $14.19 (for E.Kalimantan). These calculations are given inTable II.3. For both benchmark years, unitrents equal 53-55 percent of f.o.b. export unitvalues. Rents for the period 1970-78 and1983-84 have been estimated by assuming thatthe same relation held for other years in thoseperiods. Thus, unit rents are assumed to be 55percent of f.o.b. value in each year. The1979-82 f.o.b. values, production costs, andrents have been taken directly from a detailedstudy of rents and rent capture covering thoseyears.11

Domestic processing of roundlogs into sawn-wood and plywood for export dissipated poten-tial rent from the roundlog harvest during1979-83 because restrictions on log exports pro-tected inefficiencies in Indonesia's wood-processing industries. The potential rentobtainable from roundwood at the time of har-vesting, equivalent to that on log exports, istherefore a more valid economic measure ofdepletion costs for the timber resource than therents actually earned.

The rental value of harvested timber andmature virgin forest stands from which futurecommercial extraction can be anticipated cannotbe applied without modification to the remain-ing elements of the timber accounts—reforesta-tion/afforestation, deforestation, and degrada-tion—which refer only to secondary forests.The value of each cubic meter of timber ini-tially harvested from an area of virgin forestmay be anticipated to exceed that of theremaining timber and of subsequent harvestsfrom the logged-over forest. This implies thatlower rent values should be assigned tochanges in timber resource levels that arisefrom growth, deforestation, or degradation insecondary forests.

However, subsequent harvests will typicallynot exhaust the stumpage value in secondaryforests, in part because trees worth less thanthe royalties and taxes that would be due onthem will never be harvested. For example, ifsuch charges total $20 per cubic meter, rationalconcessionaires will never deliberately harvesttrees with a lower stumpage value, eventhough those trees are not economicallyworthless.

For commercial species, one indicator ofstumpage values in secondary forests can bederived from the rotation period between suc-cessive harvests, which is set to allow stands toregenerate. In the Indonesian selective cuttingsystem, the prescribed period between harvestsis thirty-five years. Immediately after logging,the present value method implies that theresource rent on the residual stand is the

36

discounted present value of the income fromthe next harvest thirty-five years in the future.Since secondary forests contain stands last har-vested varying numbers of years ago (from oneto thirty-five), an average of such present valuesyields an estimate of the resource rents for com-mercial species in secondary forests. This esti-mate must then be adjusted for the distributionof trees between commercial and non-commer-cial species. Applying this reasoning to theIndonesian case in the absence of extensivedata, results in an estimate of an averageresource rent in secondary forests equal toapproximately one-half that of the timber har-vest. This estimate has been applied to thephysical accounts for stocks, regeneration,deforestation, and degradation of secondaryforests.

The value of changes in plantation timberlevels (reforestation/afforestation and harvesting)has been estimated as zero because plantationinvestments in Indonesia have not been shownto yield more than a normal rate of return oninvestment. In fact, a normal rate of return maybe a generous estimate of their profitability todate. Further, the proportion of the total har-vest originating from plantations is small andhas been provisionally estimated as zero. Insummary, physical stocks have been valued asfollows:

Virgin forests Primary rent (PR)Logged forests Secondary rent

(PR x 0.50)Protection forests Secondary rent

(PR x 0.50)Plantations Zero

B. Indonesian PetroleumResource Accounts, 1970-1984

Indonesia's geological situation at the intersec-tion of several continental plates may accountfor the vast reserves of oil and natural gas inthe region. In 1849, there were reports of oilseepages in West Java, but it wasn't until 1871that exploratory wells were drilled and 1885 thatcommercial production began. Between 1880

and 1930, seven of the fourteen major basinswere discovered, and by 1965 three more largefields were found, including the Minas field inCentral Sumatra—the region's largest.

After 1966, technological developments andpolitical stabilization encouraged exploitation ofoffshore oil resources. Rising production andworld oil prices in the early 1970s led to boomconditions, which were dampened by the reve-lation in 1974 of serious financial mismanage-ment by the state oil company, PERTAMINA.After several years of reduced explorationwhile foreign oil company contracts wererenegotiated and Indonesia's oil-related busi-nesses were reorganized, developmentresumed in 1978 and accelerated during thesecond oil boom. After 1982, falling oil pricesled to a downward trend in investment.

Most experts believe that all major fieldshave been discovered and further explorationwill yield no great surprises. Future productionprospects are apparently mediocre, given thatproven reserves of 9.7 billion barrels in 1984were enough for only another 18 years at the1984 production rate of about 500 million bar-rels per year. Production is already declining inmost major oilfields, and estimates of undis-covered reserves range from 10 to 40 billionbarrels.12 Significant natural gas productionbegan only in the 1970s, however, after theMinas field was found, and production sincethen has grown sharply. Indonesia's vast gasresources are found in more remote regions ofthe archipelago and remain largely unmapped.Current proven reserves exceed 12.7 billionbarrels of oil equivalent (BOE) and possiblereserves are estimated at 45 billion barrels. Thehuge Arun gas field alone has estimatedreserves of 2.7 billion BOE, and the recentlydiscovered Natuna field may have twice thatmuch.13 (Resource accounts for natural gas arenot included in this report, due to paucity ofdata on output, reserves, and production costs,but will be constructed for future analysis.)

Oil has accounted for more than 50 percentof export earnings and government revenues

37

since 1967, and it has financed rapid growth ininvestment and consumption expenditures. Inthe 1980s, falling oil prices and growingdomestic consumption sharply reduced exportreceipts. The Indonesian government hasresponded by devaluing the currency to pro-mote other exports and by reducing domesticpetroleum subsidies to restrain domestic con-sumption and improve government finances.

PERTAMINA remains responsible for oil andgas development, but with powers morelimited than they were before 1975. Over 90percent of exploration and production is con-tracted to foreign oil companies. Early contractsexchanged exploration and production conces-sions for royalty payments. In the early 1960s,new "contract of work" agreements wereintroduced, under which the government holdstitle to the oil and collects a share of profitsrather than royalties. Most current productionis under production-sharing contracts thatrequire the contractor to pay a bonus when theagreement is signed, spend a specific amounton exploration within a stipulated period, sup-ply 25 percent of output to PERTAMINA atcost plus $0.20 (formerly $0.30) per barrel, andpay an additional amount such that the networth of the oil is split according to negotiatedratios between PERTAMINA and the oil com-pany. PERTAMINA's share ranges from 65 to88 percent.

1. The Physical Accounts

Petroleum resources are divided into identi-fied and undiscovered reserves. Identifiedreserves are subdivided into proven reserves,those that can be recovered under current eco-nomic and technical conditions, and probablereserves, those estimated to exist on the basis ofengineering and geological data that areobtained with current operating practices.Undiscovered reserves are surmised to exist onthe basis of broad geologic theory and experi-ence. By definition, since only proven reserveshave a positive rental value (their net priceexceeds their estimated recovery cost), onlyproven reserves enter the resource accounts.

Additions to reserves in the physical accountsconsist of discoveries (reserves found in newreservoirs by exploratory drilling) and upwardadjustments of reserve estimates in existingreservoirs because of new information orchanged technological and economic condi-tions. Subtractions from reserves are attributableto extractions, downward adjustments toproven reserves, and other depletion losses,such as oil spills. (See Table 17.4.)

Within this framework, data on flow itemsare more reliable than stock estimates becauseIndonesia's exploration and extraction is closelymonitored by other members of OPEC and bythe international oil community. Data onIndonesia's proven reserves is sketchy and oflimited reliability, in part because such infor-mation is treated as sensitive by the govern-ment. Reported revisions to reserves seem alsoto be influenced by companies' strategicinterests. In most years, they closely parallelreported production to keep total reserves sta-ble. While significant discoveries were reportedduring the period, upward and downwardrevisions of reserve figures were negligible.This differs significantly from typical experiencein other oil-producing regions, where, as dis-cussed earlier, upward revisions add substan-tially to initially reported reserve figures. More-over, in years of sharp oil price hikes, whichought to have made more oil economicallyrecoverable, reported reserves did not increase.However, in 1974, in the wake of changes inU.S. tax law and Indonesian contracts favor-able to exploration activities, reported reservesincreased sharply.

For these reasons, the net changes inresources that correspond to resource depletionwithin the national income accounting frame-work are more useful than the valuation andrevaluation of total resource stocks. Even datafor the flow items are not without problems.Production data are probably understated,since Indonesia, as a member of the OPEC car-tel, has been obliged to limit production belowthe amount it would wish to sell. Undeclaredproduction is primarily in the form of condensate

38

Table II.4. Petroleum Accounts 1970-1984

PHYSICAL ACCOUNTS (million barrels)1970

Opening Stock

AdditionsDiscoveriesUpward Revisions

Depletions

Net Change

Closing Stock

9,000

1,2690

312

957

9,957

UNIT VALUES (US$/barrel)FOB Export Price 1.70Production Costs 0.50Rent/barrel 1.20

MONETARY ACCOUNTS (million US$)Opening Stock —

AdditionsDiscoveriesUpward Revisions

Depletions

Net Change

Revaluation

Closing Stock

1,522.80

374.4

1,148.4

—

11,948.4

19719,957

2,1430

326

1,817

11,774

2.210.791.42

11,948.4

3,043.10

462.9

2,580.2

2,190.5

16,719.1

197211,774

6760

396

280

12,054

2.960.782.18

16,719.1

1,473.70

863.3

610.4

8,948.2

26,277.7

197312,054

8240

489

335

12,389

3.730.802.93

26,277.7

2,414.30

1,432.8

981.5

9,040.6

36,299.8

197412,389

1,7620

502

1,260

13,649

10.801.749.06

36,299.8

15,963.70

4,548.1

11,415.6

75,944.6

123,659.9

oil, which is extracted at the rate of about depletion, but have been negative through the100,000 barrels a day. latter part of the period.

The physical accounts show the net resource 2. The Value Accountsflow, or change in the petroleum reserve, inmillions of barrels per year. Depletions, essen- Petroleum resources are valued by the nettially extraction, peaked in 1977 and 1978. price method, defined as the market price lessNet resource flows were positive during the all factor costs of extracting the resource andearly 1970s, as reported discoveries exceeded bringing to the point of sale. The alternative

39

Table II.4. (cont.) Indonesian Petroleum Accounts

PHYSICAL ACCOUNTS (million barrels)1975

Opening Stock

AdditionsDiscoveriesUpward Revisions

Depletions

Net Change

Closing Stock

13,649

1700

477

-307

13,342

UNIT VALUES (US$/barrel)FOB Export Price 12.60Production Costs 2.36Rent/barrel 10.24

MONETARY ACCOUNTS (million US$)Opening Stock 123,659.9

AdditionsDiscoveriesUpward Revisions

Depletions

Net Change

Revaluation

Closing Stock

1,740.80

4,884.5

-3,143.7

16,105.8

136,622.1

197613,342

4690

550

- 8 1

13,261

12.702.14

10.56

136,622.1

4,952.60

5,808.0

-855.4

4,269.5

140,036.2

197713,261

942

615

-519

12,742

13.631.49

12.14

140,036.2

1,141.224.3

7,466.1

-6,300.6

20,952.3

154,687.9

197812,742

1010

597

-496

12,246

13.631.52

12.11

154,687.9

1,223.10

7,229.7

-6,006.6

-382.2

148,299.1

197912,246

761

581

-504

11,742

13.981.96

12.02

148,299.1

913.512.0

6,983.6

-6,058.1

-1,102.1

141,138.8

method—estimating the present value of futurenet income from the field—requires estimatesof recoverable reserves, production costs,future output prices, and interest rates that arenot available. As with timber, the market priceis measured as the f.o.b. export price, which isalso the opportunity cost of sales on thedomestic market. The unit cost panel in theresource accounts gives the f.o.b. export pricefor 1970-84 in U.S.$ per barrel.

The factor costs of developing, extracting,and transporting a barrel of oil are estimatedfor the same time period by dividing the totalannual expenditures for exploration anddevelopment of the contracting companies bytheir total annual production.14 More detaileddata on costs per barrel are calculated by thecompanies and submitted to PERTAMINA inaccordance with contract provisions, but aresensitive and not publicly available. The cost

40

Table II.4. (cont.) Indonesian Petroleum

PHYSICAL ACCOUNTS (million barrels)1980

Opening Stock

AdditionsDiscoveriesUpward Revisions

Depletions

Net Change

Closing Stock

11,742

1410

577

-436

11,306

UNIT VALUES (US$/barrel)FOB Export Price 28.11Production Costs 3.80Rent/barrel 24.31

MONETARY ACCOUNTS (million US$)Opening Stock 141,138.8

AdditionsDiscoveriesUpward Revisions

Depletions

Net Change

Revaluation

Closing Stock

3,427.70

14,026.9

-10,599.2

144,309.2

274,848.9

Accounts

198111,306

2230

586

-363

10,943

35.835.50

30.33

274,848.9

6,763.60

17,773.4

-11,009.8

68,062.1

331,901.2

198210,943

1720

484

-312

10,631

35.748.59

27.15

331,901.2

4,669.80

13,140.6

-8,470.8

-34,798.7

288,631.6

198310,631

710

521

-450

10,181

34.759.15

25.60

288,631.6

1,817.60

13,337.6

-11,520.0

-16,478.1

260,633.6

198410,181

670

517

-450

9,731

31.947.64

24.30

260,633.6

1,628.10

12,563.1

-10,935.0

-13,235.3

236,463.3

estimates used in this exercise are reasonableapproximations, according to industry experts.PERTAMINA, which accounts for less than 10percent of production, was excluded for lack ofdata, so the implicit assumption is that PERTA-MINA's cost structure equals the average forthe industry. Exploration costs are treated ascurrent production costs in this exercise whiletaxes and royalties are not treated as produc-tion costs.

The difference between unit revenues andcosts gives the resource rent per barrel of oil. Itrises sharply over the period in response toincreases in petroleum prices. Since 1985, aver-age rents per barrel have declined even moresharply than world oil prices have. This rent isdivided between the contractors and thegovernment in accordance with the terms ofthe various production contracts in force. Thereduction in government rental receipts from

41

petroleum has forced a sharp curtailment indevelopment expenditures.

The final monetary accounts presents thevalues of stocks and flows in current dollarfigures. They are analogous to the valueaccounts given for forest resources: they valuethe net additions to and subtractions from theresource base in terms of the relevant economicrent. In the petroleum sector, the sharp swingfrom positive to negative value flows stemsfrom the fact that extractions began to exceedapparent additions to reserves at the timewhen the unit value of the resource was risingsharply. From 1980 to 1984, the annual netresource depletion on petroleum accountexceeded U.S. $10 billion, an amount that issignificant relative to annual GDP growth andannual gross fixed capital formation. Treatingthis net depletion of a limited natural resourceasset as current income rather than assetdepreciation must seriously distort perceptionand analysis of Indonesia's economicperformance.

C. Indonesia's Soil Account:Java

Soil erosion has both physical and economiceffects. Removing part of the topsoil anddepositing it elsewhere lowers the agriculturalpotential and economic value of the erodedland. The loss of potential future farm incomeis equivalent to the depreciation of an eco-nomic asset. Besides the on-site costs of soilerosion are off-site or downstream costs, suchas siltation of reservoirs and irrigation systems,harbors, and other waterways.

1. Estimates of Soil Erosion in PhysicalTerms

Comprehensive data on soil erosion inIndonesia are not available, so estimates werebased on erosion models. (The relevant deter-minants of soil erosion are the topography,climate, soil characteristics, and land uses ofthe specific areas affected.) Among the data

available with which to estimate erosion ratesare maps at the scale 1:1,000,000 of three varia-bles that play a major role in determining ero-sion rates—soil types and slope, rainfall ero-sivity and land use.

The soil map used for this study (FAO, 1959)combines soil types with topography to create25 soil classes:

—five classes of soils on level to undulatingland, with dominant slopes under 8 per-cent (units 01-05);

—eleven classes of soils on rolling to hillyland, with dominant slopes from 8-30 per-cent (units 06-16); and

—nine classes of soils on hilly to moun-tainous land, with dominant slopes over 30percent (units 17-25).

(Areas of soil types by province are included inAnnex Table A.I.)

The kinetic energy released as raindrops strikethe ground contributes to soil erosion. Bols(1978) has prepared a map of Java based oncorrelations of a measure of the kinetic energyof storms with annual rainfall data, which isavailable for most of Java over an extendedperiod. Eleven rainfall erosivity classes aremapped at a scale of 1:1,000,000. (Area esti-mates for each erosivity class are shown inAnnex Table A.2.) In 1985 the Ministry of For-estry produced a land use map of Java, whichdistinguishes five types of land use (or vegeta-tion cover) that influence erosion rates:

—Areas of sawah (irrigated ricefields), includ-ing fishponds. These areas are character-ized by low erosion rates; in fact, in largeareas sedimentation prevails over erosion;

—Areas of legal (dryland farming), mostly onsloping uplands where erosion rates arevery high;

—Areas of natural and planted forest, includ-ing perennial plantation crops where ero-sion is slight;

42

—Degraded forest areas, including areas ofshifting cultivation and degraded pekarangan(home gardens) where erosion is moderateto high; and

—Wetlands, where erosion is low.(See Table II. 5.)

Aggregate land use data of questionable relia-bility are also available for Java from the Cen-tral Bureau of Statistics. (See Table 11.6.) Themapped areas of sawah exceed the CentralBureau of Statistics figures for every province,totalling about one third more land for thewhole of Java. The Forestry Ministry map isbased in part on aerial photos that can measuresawah area accurately. Provincial discrepanciesfor tegal, on which erosion is more severe,range from 80 to 177 percent of CBS estimates,but for Java as a whole average only 11percent.

Given these discrepancies, the estimates ofper hectare erosion rates were based on theforestry map, which could be matched spatiallywith the other elements of the soil erosionmodel. However, because the Central Bureauof Statistics estimates for land uses other thansawah appear to be somewhat more reliable inthe aggregate, this data is used in the final eco-nomic calculations.

The three maps described above were con-verted to digital form and analyzed using theGeobased System by the World Bank'sEnvironmental Operations and Strategy Divi-sion.15 Essentially, the procedure overlays thethree maps to estimate land areas by 1,375 pos-sible combinations of slope and soil type, ero-sivity, and land use. The analysis also dividedJava along provincial boundaries to generate5,500 possible combinations.16

The estimates of actual erosion rates cor-responding to each of the possible combina-tions were based on actual measurementsunder given conditions of plant cover or crop-ping when possible, supplemented by judg-ments based on erosion elsewhere under com-parable conditions. Several recent projects onJava have yielded valuable data on actual ero-sion of uplands. These include the successiveUNDP/FAO Projects in the upper Solo water-shed, the USAID Project in the Citanduywatershed, the Dutch-sponsored projects in theupper Brantas (Kali Konto), and the UplandAgricultural Projects of Jogyakarta and theJratunseluna watersheds financed by USAIDand the World Bank. Other erosion data havebeen collected by the Soils Department of theAgricultural University in Bogor, by the SoilResearch Centre in Bogor, and by the Water-shed Management Centre in Solo.

Table II.5. Land Use on Java ('000

Land Use

SawahForestDegraded ForestWetlandsTegal

TOTAL

Source: Calculated

West Java

1,350542299

2,546

4,737

from Ministry of

ha)

Central Java

1,3807313429

1,127

3,301

Forestry (1985)

Jogyakarta

1214

210

335

East Java

1,7521,222

53103

1,401

4,531

Java

4,6032,499

386132

5,283

12,903

43

Table II.6. Comparison of Land

Sawah Area Estimates

West Javaa

Central JavaJogyakartaEast JavaJAVA

Tegal Area Estimates13

West JavaCentral JavaJogyakartaEast JavaJAVA

Use Estimates ('000

CBS

1,2151,023

641,1993,501

1,4401,366

1961,7444,747

ha)

Ministry ofForestry

1,3501,380

1211,7524,603

2,5461,127

2101,4015,283

a. Including D.K.I. Jakarta,b. House compound and surroundings and bareland/border/shifting cultivation.

Model asPercent

110135191146131

17782

10780

111

Estimates of annual soil loss by soil type andland use for Java's four provinces are pre-sented in Annex Tables A.3-A.6. Table II.7shows tegal suffers by far the greatest per hec-tare and total soil loss. On a per hectare basis,soil loss is highest on tegal land on West Java,followed by tegal on Central Java. The soils ofEast Java are least subject to erosion.

If it is assumed that geologic erosion, the rateof soil loss that occurs without human inter-vention, is similar to that which occurs underforest cover, incremental erosion due to humanintervention can be estimated by the differencebetween per-hectare loss on forestland and ontegal. On average, each hectare that isdeforested and brought into agriculturalproduction causes the loss of an additional 133metric tons annually. In the calculations thatfollow, no attempt is made to segregate thecosts of man-made erosion.

2. Estimates of the Economic Costs of Erosion

Erosion reduces the availability and concen-tration of plant nutrients and alters soil struc-ture in ways that affect water availability androot growth. Subsoil weathering may partiallyreplace these soil elements over the longterm.17 Erosion's impacts on productivitydepend on soil type and crop. Some soils con-tain most of their organic matter in the top fewcentimeters. In other soils, nutrients are dis-persed over the whole soil profile. In addition,such demanding crops as tobacco suffer moredrastically from nutrient loss than non-demanding crops—cassava, for example. Thisstudy distinguishes two groups of rainfed foodcrops:

—sensitive crops (maize, soybeans, ground-nuts, green beans, and dryland rice)

—insensitive crops (cassava).

44

Table H.7.

Tegal

Predictedthousand

Forest Land

Degraded

Sawah

TOTAL

Forest

Tegal

Soil Loss Bymetric tons)

per

168

10

100

0

98

Forest Land

Westha

.1

.3

.3

.8

.1

Degraded Forest

Sawah

TOTAL

Region anc

Javatotal

4,279

56

300

11

4,647

. Land Use (metric tons

Central Javaper ha total

145

5

38

0

52

East Javaper ha total

87.2

4.4

50.9

0.3

29.5

1,221

54

27

6

1,308

8

3

2

4

0

1,643

39

13

6

1,701

per

139

6

88

0

61

per hectare and hundred

Jogyakartaper ha total

108.1

5

0

0.4

68.0

JAVAha total

.5

.0

.3

.5

.7

7,370

150

341

23

7,883

227

0.2

0

0.5

227

Few studies of erosion effects on yields areavailable for Indonesia. From scanty experi-mental data, yield-erosion relationships havebeen estimated for the study's 25 soil typesand two crop groups, as shown in Tables II.8and II.9. Soil losses of less than 15 tons/ha/yrare estimated to result in no yield loss.18

Applying these estimates of productivity loss tothe areas of the different soil types under tegalyields estimates of the extent and severity ofphysical yield loss. (See Tables 11,10 and 11.11.)This procedure predicts average yield losses of6.8 percent per year for sensitive crops andlosses of 4.3 percent per year for insensitivecrops. Among provinces, Jogyakarta is themost severely affected, followed in descendingorder by West, East, and Central Java.

These predicted yield declines can only cau-tiously be compared with actual yield trendsfor dryland crops, which have consistentlyrisen, despite erosion, because of continuedintensification of farming practices. From 1972through 1983, upland rice, maize, and cassavayields on Java increased on average by 4.3, 4.7,and 2.8 percent per year, respectively. (Roche1987) However, fertilizer use on maizeincreased from 38 kg/ha to nearly 106 kg/haand on cassava from 8 kg/ha to more than 16kg/ha. (Central Bureau of Statistics) Labor costshave also been rising on upland crops. (Roche1987) The release and rapid adoption of high-yielding maize varieties may also have maskeddeclines in the productivity of the resourcebase.19

45

Table II.8. Productivity Loss Estimates as a Result of Soil Erosion for Major Soils of Java

I. For Maize, Soybeans, Groundnuts

Soil Loss(tons/ha/year)

0-1515-6060-250

250-600Over 600

1,17

0.000.020.030.040.05

2, 3, 4,6, 9, 16

0.000.030.050.070.09

Soil Types5, 8, 10, 11, 12,

15, 18, 20, 21, 25

0.000.050.080.100.12

7, 13, 14, 19,22, 23, 24

0.000.070.100.120.15

As the results of erosion, farm output andincome have fallen in some regions withoutmajor changes in farm practices; some farmershave been induced to change cropping patternsand input use; and, in extreme cases, land hasbeen withdrawn from cultivation. Mclntoshand Effendi (1983) cite the example of theupper Citanduy Watershed, where farmersgrow corn, upland rice, and cassava on bettersoils. As erosion becomes more severe, rice isreplaced by peanuts, and on nearly depletedsoils only cassava is grown.

Whatever the response, farm revenuesdecline as crop output falls, but costs may not.Erosion may lead some farmers to work harder

and purchase more fertilizer to make up forproductivity losses, while costs for harvestlabor, crop transport, and other inputs mightdecrease. Available farm budget data suggestthat costs that would fall along with outputaccount for a small share of farm productioncosts, so erosion lowers net farm income andeventually leads to the adoption of less profit-able crops.

To account for adjustments in cropping sys-tems, a variety of farm level data for Java'sprovinces were used to develop sets ofrepresentative farm budgets.20 The budgetspublished by Roche were updated to 1985prices, adjusted to reflect yield changes by

Table H.9. Productivity

II. For Cassava

Soil Loss(tons/ha/year)

0-1515-6060-250

250-600Over 600

Loss

1,17

0.000.010.020.030.04

Estimates as a Result of

2, 3, 4,6, 9, 16

0.000.020.030.050.07

Soil Erosion for Major Soils

5, 8, 10, 11, 12,15, 18, 20, 21, 25

0.000.030.050.070.10

of Java

7, 13, 14, 19,22, 23, 24

0.000.050.060.080.12

46

Table 11.10. Area and Severity of

West JavaCentral JavaJogyakartaEast JavaJAVA

West JavaCentral JavaJogyakartaEast JavaJAVA

a. Productivity

Note: All values

Annual0%

512190

19194914

Annual

8%

42916847

4811,125

Estimated Erosion-Induced Productivity Losses on Tegal on Java

Productivity Loss2%

310

4549

Productivity Loss

10%

802366118408

1,693

loss based on maize.

as a Percent of3%

Area ('000 ha)27

1202691

264

as a Percent of

12%

Area ('000 ha)351

6100

412

do not sum exactly due to rounding.

Current Total5%

417216

0128762

Current Total

TotalArea

2,5631,126

2091,4135,312

Productivity8

7%

2230

6792

ProductivityAverage

ProductivityLoss (%)

7.06.47.86.66.8

using the Central Bureau of Statistics and Malangdata, and then used to estimate the effects ofyield losses on net farm incomes. Insofar as canbe determined, the farm budgets are consistentwith land values and rental rates for tegal.

Table 11.12 summarizes the cropping systemsfor each region and provides an estimate oftheir relative occurrence. These farming sys-tems appeared to be marked by a large propor-tion of fixed costs. Costs categories in theCentral Bureau of Statistics that seem mostlikely to vary with output are harvesting laborand transportation. These variable costs wereassumed to decline in proportion to cassavayield declines, while yields of maize and othermore sensitive intercropped cultivars declinedfurther. Consequently, farm income declines

linearly as erosion increases, at rates that varyby cropping system and by region.

The estimated loss in farm income from a1-percent decline in yield depends on both thebasic profitability of the cropping system andthe importance of fixed production costs. Onthe assumption that the farming systems aredistributed independently of rates of produc-tivity decline, in Table 11.12 the costs of a1-percent decline in productivity for each crop-ping system and the predicted weighted aver-age yield declines are applied to the tegal areasallocated to each cropping system. These costsare for only a single year. But, the appropriateeconomic measure of soil depletion is the pres-ent value of losses in farm income in currentand future years.

47

Table 11.11. Area and Severity of

West JavaCentral JavaJogyakartaEast JavaJAVA

West JavaCentral JavaJogyakartaEast JavaJAVA

a. Productivity loss

Note: All values do

Annual0%

512190

19194914

Annual

6%

659201118394

1,371

Estimated Erosion-Induced Productivity Losses on Tegal on Java

Productivity1%

310

4549

Productivity

7%

144165

014

322

based on cassava.

Loss as a Percent of2%

Area ('000 ha)27

1202691

264

Loss as a Percent of

8%

Area ('000 ha)351

6100

412

not sum exactly due to rounding.

Current Total3%

417216

0128762

Current Total

TotalArea

2,5631,126

2091,4135,312

Productivity2

5%

45117147

5481,217

ProductivityAverage

ProductivityLoss (%)

4.44.14.74.14.3

If soil loss is recurrent and exceeds soil for-mation, productivity losses occur with eachsuccessive net loss of soil depth. The correctmeasure of the cost of the initial episode oferosion is the capitalized value of the infinitestream of productivity losses associated withthat episode. Loss of productivity associatedwith future erosion should be charged againstincome when it occurs.

On Java, erosion is clearly a recurrentphenomenon, and productivity losses are per-manent. As productivity falls, land eventuallygoes out of production, and its productionvalue falls to zero. Current and future technicalchange that raises farm productivity has noeffect on these losses unless technical change is

faster on good soils or, on the contrary, isdriven to compensate for erosion losses. If theformer holds true, as is likely, the cost of ero-sion is larger than estimated above.

The one-year costs of erosion have beencapitalized to obtain a total present value offuture losses of Rp 539 billion (U.S. $484 mil-lion). To put this figure into perspective, Table11.13 shows the approximate value of output ofsix major rainfed crops at 1983/84 prices. Theone-year costs of erosion are about 4 percent ofthe annual value of dryland farm output, andthey are of the same order of magnitude asannual recorded growth in agricultural produc-tion in the uplands. Thus, despite apparentlyhealthy growth, upland farming on Java has

48

Table 11.12. Costs Due to Soil

EstimatedPropor-

Cropping tion ofSystem Crops Tegal (%)

West JavaI Cassava,

CornUplandRice &Legumes 58

II Cassava,Corn &UplandRice 27

III PureStandCassava 15

Total Tegal 100

Central JavaI Intercropped

Corn &Cassava 57

II IntercroppedCorn,Cassava &Legumes 43

Total Tegal 100

Erosion

Areaa

('000 ha)

835

389

216

1,440

779

587

1,366

for Various

EstimatedCurrent

NetIncome(Rp/ha)b

139,496

49,531

1,279

95,039

6,698

10,183

8,196

Cropping Systems on

WeightedProductionLoss (%)c

4.4

4.4

4.4

4.4

4.1

4.1

4.1

a. Based on Central Bureau of Statistics. See Table II.6.b. Net income equal to returns to land and management,c. Based conservatively on rates for land cultivated in Cassava

sensitive crops ranges from 6.8 to 7.8 percent.

Source: Adapted from Roche 1984, Central BureauAgroeconomic Survey, Bogor. See Magrath,

Java

Annual Costof a One SinglePercent Year

Productivity CostDecline (million(Rp/ha) Rp)

4,309

3,616

1,563

3,718

800

937

859

. Annual

of Statistics, and dataArens, 1987.

15,831

6,186

1,485

23,508

2,555

2,255

4,810

jroductivity

provided b>

CapitalizedCost

(millionRp)

158,310

61,860

14,850

235,080

25,550

22,550

48,100

loss for

r the

49

Table 11.12 (cont.)

EstimatedPropor-

Cropping tion ofSystem Crops Tegal (%)

JogyakartaI Intercropped

Corn &Cassava 57

II IntercroppedCorn,Cassava &Legumes 43

Total Tegal 100

East JavaI Intercropped

Corn &CassavaLevel Tegal 30

II IntercroppedCorn &CassavaTerracedHillsides 30

III Pure StandCassavaLevel Tegal 20

IV Pure StandCassavaTerracedHillsides 20

Total Tegal 100

TOTAL TEGAL

Areaa

('000 ha)

112

84

196

523

523

349

349

1,744

4,747

EstimatedCurrent

NetIncome(Rp/ha)b

8,220

11,279

9,531

298,327

58,130

145,005

27,806

141,499

83,649

WeightedProductionLoss (%)c

4.7

4.7

4.7

4.1

4.1

4.1

4.1

4.1

4.3

Annual Costof a OnePercent

ProductivityDecline(Rp/ha)

1,011

1,047

1,026

4,926

2,876

3,746

1,816

3,453

2,686

SingleYearCost

(millionRp)

532

416

948

10,567

6,169

5,357

2,597

24,690

53,956

CapitalizedCost

(millionRp)

5,320

4,160

9,480

105,670

61,690

53,570

25,970

246,900

539,560

50

Table 11.13. Comparison of the ValueErosion (million rupiah)

Dry Rice

Maize

Cassava

Sweet Potatoes

Peanuts

Soybeans

Total

Cost of Single YearErosion Loss

Capitalized Value ofErosion Losses

Single-year ErosionCost as a Fractionof Value of Agricul-tural Output

WestJava

46,533

21,809

81,041

22,191

44,916

17,807

234,297

23,508

235,080

0.10

of Output of

CentralJava

18,194

123,596

109,148

12,131

56,475

45,398

364,942

4,810

48,100

0.01

Six Major Rainfed

Jogyakarta

12,682

15,061

22,410

542

18,340

37,664

106,699

948

9,480

0.01

Crops to the

EastJava

26,358

262,981

134,962

15,331

74,615

124,171

638,418

24,690

246,900

0.04

Cost of

JAVA

103,767

423,447

347,561

50,195

194,346

225,040

1,344,356

53,956

539,560

0.04

been on a treadmill: each current increment inproduction is offset by an equal but unrecordedloss in soil productivity.

The capitalized losses in future productivityare approximately 40 percent of the annualvalue of upland farm production. If erosionlosses are regarded as the cost of obtaining thecurrent year's livelihood from vulnerableupland soils, then these estimates show thebargain to be harsh. Nearly 40 cents in futureincome is sacrificed to obtain each dollar forcurrent consumption. Whether such a bargaincan be sustained is open to question, butignoring the heavy costs of current farming

practices unquestionably overstates drylandagricultural income.

The methodology and data used in estimat-ing erosion costs produced results for a singleyear, 1985. Benchmark data for other yearswere not available. To extrapolate the resultscrudely to other years in the period underreview, a double indexation procedure wasused. First, physical erosion rates were indexedto the area under legal. Since such other factorsas topography, soil type, and climate remainedconstant throughout the period or varied ran-domly, physical erosion rates varied systemati-cally only with changes in land use, of which

51

conversion to annual cropping is the mostimportant. (In fact, the area in upland cropschanged little.) Then, the costs of given ratesof erosion were indexed to dryland crop priceson the assumptions that 1) cropping patternsand practices changed little, and 2) net farmincome remained a constant proportion of farmrevenues. While these assumptions cannot bereadily verified with existing data, indexationdoes at least correct for the general inflationaryrise in farm prices during the period. (See Table11.14.)

Erosion simply moves soil particles from oneplace to another. The deposition of sediment onsawahs renews their fertility. More commonly,

however, the off-site effects of soil erosion arenegative. Silt clogs irrigation channels andports, and it lowers the capacity of water-storage reservoirs.

Only a crude attempt was made to estimatethe magnitudes of such costs as the increasedexpenditures needed to dredge waterways andclean irrigation channels.21 These additionalannual costs due to upstream erosion appear tobe in the range of U.S. $15-50 million, anorder of magnitude less than on-site produc-tivity losses. Moreover, such costs alreadyenter the national income and product accountsas additional government expenditures. Thisillustrates another anomaly of the current

Table

Year

197119721973197419751976197719781979198019811982198319841985

a. Inof

11.14. Estimates

TotalTegal in

Javaa

4,3773,9884,7774,4844,2323,6423,9824,5224,1114,1234,3563,3194,0814,4164,747

of Erosion Losses 1971-1985

Per ha.Cost of a

1% Loss inProductivity

312.42354.90471.77692.40781.63894.66

1,019.921,100.441,288.321,485.141,610.401,843.012,308.242,540.852,686.00

AverageProductivity

Loss onCultivatedArea (%)

4.3%4.34.34.34.34.34.34.34.34.34.34.34.34.34.3

thousands of hectares. Based on estimates of drylandStatistics, Jakarta

b. Current rupiah per1972-1984.

hectare. The 1985 value1971-1984 are derived using indices of cropthe ratio between revenue and the cost of a

c. Value does not sun

is based on (prices faced

Single-yearCost ofErosion

(mill. Rp.)

5,8806,0869,691

13,35014,22414,01117,46421,39822,77426,33030,16426,30340,50648,24853,956C

crops in Java from

CapitalizedCost ofErosion

(mill. Rp.)

58,80060,86096,910

133,500142,240140,110174,640213,980227,740263,330301,640263,030405,060482,480539,560

Central Bureau

detailed budget analysis. Values forby farmers and the

1% productivity loss remainedi exactly due to averaging.

assumption thatconstant.

52

income-accounting system since these erosioncosts enter with a positive sign—as additions tonational income. Although the expenditures aremade to prevent even greater damages fromsiltation, they are entered as additions toincome and the production of goods and ser-vices because the expenditures are incurred byhouseholds and the government and are there-fore defined as final expenditures. Were such"defensive" expenditures subtracted from thevalue of final output, Indonesian nationalincome would be roughly $30-$100 millionlower in each year.

D. Concluding Remarks

Three general points will suffice here:

First, these estimates were prepared with amodest expenditure of time and money, draw-ing entirely on data source and informationalready available, mostly in published sources.Estimation required some interpolationbetween benchmark years and extrapolationfrom samples of limited coverage, but such

techniques are already common in nationalincome accounting.

Second, the results require a significant reas-sessment of Indonesia's economic performanceduring the period, and they bring to lightaspects of the sustainability of Indonesia's eco-nomic growth strategy that would not be read-ily apparent from the conventional nationalincome accounting framework.

Third, efforts to improve the accuracy andcoverage of such resource accounts are entirelycomplementary to efforts to improve the infor-mation base for better resource management.For example, the Government of Indonesia,with external assistance, is embarking on anew inventory of timber resources that willincrease the accuracy of forest resourceaccounts and also provide better guidance inallocating timber concessions, delineating pro-tected forest areas, siting transmigrationprojects, and other resource management deci-sions. The same kinds of data needed forresource accounting are essential for effectiveresource management.

Robert Repetto is Director of the Program in Economics and Institutions at the World ResourcesInstitute. Formerly, he was an associate professor of economics in the School of Public Health atHarvard University and a member of the economics faculty at Harvard's Center for PopulationStudies. William Magrath is a natural resource economist in the World Bank's environment depart-ment. Formerly, he was an associate at the World Resources Institute and on the staff at CornellUniversity. He holds graduate degrees in natural resources and economics from the University ofMichigan. Michael Wells is a doctoral candidate at the University of British Columbia and a consul-tant to the World Bank environment department. He was recently a visiting scholar in the eco-nomics department of the University of Indonesia in Jakarta and has eight years experience in theUnited States and Europe with Touche Ross, the international accounting and consulting firm.Christine Beer is presently the Co-Program Manager in Gaza, Israel, for Save the Children Interna-tional. She holds a Masters in International Affairs from the Johns Hopkins School of AdvancedInternational Studies and has worked as a consultant for Energy Development International inKhartoum. Fabrizio Rossini worked on this paper while completing his Masters degree at the JohnsHopkins School of Advanced International Studies.

53

Annex

ANNEX A.I. Soils

SoilType

123456789

10111213141516171819202122232425

TOTAL

Source: Calculated

on Java Based on Slope

WestJava

87126136

1714511

64667

223174308

116

537149582722

6160

4,737

from FAO

CentralJava

4956220

25927

349

164400

10

36200

1757

384198

131202128

3,301

(1959).

and Soil Characteristics

Jogyakarta

4153

2218

823

60

3466

335

('000 ha)

EastJava

2971

14210

54211

13149190

4

11115326

259451

12309227

196608

4,170

JAVA

572,323

775402977

75324

4001242

9123717473835921

609481

131,210

571582850425979

12,868

55

ANNEX A.2.

ErosivityLevel

123456789

1011

TOTAL

Areas of Java Subject

WestJava

31464475777187267057521

4,625

Source: Calculated from Bols (1978)

to Alternative

CentralJava

1513779026964593022491114313

3,304

Levels of Erosivity

Jogyakarta

10170894817

334

('000 ha)

EastJava

1151,1782,178

43636719696473

4,612

JAVA

1041,6323,3602,2171,8821,546

9958671353613

12,788

56

ANNEX A.3.

SoilType

123456789

10111213141516171819202122232425

TOTAL

Predicted Soil

WestJava

063249

11400

9,48288

44511,2221,2514,156

94840,122

00

2,12300

47,14621,22793,831

159,7166,372

28,939

427,863

Losses From Tegal

CentralJava

2039840

19400

11,7760

4246,672

0910

3,4849,724

07,594

3090

52,02430,495

06,317

21,61813,095

164,274

By Region and

Jogyakarta

10381890

352178

000000

7,25900

856000000

5,0528,906

22,668

Soil Type ('000 metric

EastJava

053119

1530

8915,278

0953412

080

19,9114,985

2943,894

11,006592

20,7904,694

00

14,96732,754

122,132

tons)

JAVA

201,599

126469

01,242

26,71588

1,82118,3061,2514,255

97570,77514,710

29414,46711,315

592119,96056,41693,831

166,03348,00883,695

736,963

57

ANNEX A.4.

SoilType

123456789

10111213141516171819202122232425

TOTAL

Predicted Soil

WestJava

000000

2931

71115

111000

12800

24463

164112

0148

1,091

Loss from Sawah

CentralJava

000000

3009

430001

1080

9960

196505

5721

580

by Region and Soil

Jogyakarta

0000024000000020

4100000003

51

Type ('000 metric

EastJava

0000007053000

30370

91132

095800

9183

582

tons)

JAVA

000002

703

15118115

11131

1470

358138

053576

164117148255

2,304

58

ANNEX A.5.

SoilType

123456789

10111213141516171819202122232425

TOTAL

Predicted Soilmetric tons)

WestJava

020000000

8600000000000

2,7052,851

00

5,644

Losses from Forest

CentralJava

030100

8106

780200

400

4770

431857

0257

1,860261

3,931

Land on Java by

Jogyakarta

000000000000010000000000

27

28

Region and Soil Type

EastJava

000002

18227000

176194

017817417

7791,404

002

2,419

5,376

('000

JAVA

050102

9928

171020

177234

022518117

1,2102,2612,7053,1081,8622,708

14,979

59

ANNEX A.6.

SoilType

123456789

10111213141516171819202122232425

TOTAL

Predicted Soil Losses('000 metric tons)

WestJava

000000000

62600000000000

15,89913,033

0482

30,041

from Degraded

CentralJava

200010

588000

14000000000

539520000

1,333

Forest on Java

Jogyakarta

0000000000000000000000000

0

by Region and

EastJava

11000000

140000

895000

620

1,000277

00

90347

2,688

Soil Type

JAVA

310100

5880

14766

000

895000

620

1,539339

15,89913,033

90829

34,062

60

Notes

1. A. Soemitro, Foreign Investment in the ForestBased Sector of Indonesia: Increasing Its Contri-bution to Indonesian Development (Jakarta:Gadjah Mada University, 1975).

2. Malcolm Gillis, "MNCs, Environment, andResource Management in Indonesia's Tropi-cal Forests," Charles Pearson, (ed.), Mul-tinational Corporations, Environment and theThird World: Business Matters, Duke Univer-sity Press, Durham, N.C. 1987.

3. FAO, Tropical Forest Resources Assessment Pro-ject: Forest Resources of Tropical Asia, preparedin conjunction with the United NationsEnvironment Programme, Rome, 1981.

4. Adrian Sommer, "Assessment of WorldTropical Resources," Unasylva, no. 28, 1976.

5. FAO, op. cit.

6. FAO, Forest Resources in Asia and the FarEast Region, Rome, 1976.

7. FAO, 1981, op. cit.

8. Ibid.

9. Ibid.

10. I. Ruzicka, "Rent Appropriation in IndonesianLogging: East Kalimantan 1972/3-1976/7,"Bulletin of Indonesian Economic Studies, vol.XV, no. 2, July 1979, p. 54; V. Beunaflor,

"Forestry and Forest Product Developmentin Indonesia: Logging and Transportation,"Working Paper no. 10, UNDP/FAO, Bogor,Indonesia, March, 1981.

11. Malcolm Gillis, "Indonesia: Public Policies,Resource Management, and the TropicalForest," in Robert Repetto and MalcolmGillis, (eds.), Public Policies and the Misuse ofForest Resources, Cambridge UniversityPress, 1988.

12. World Bank, Joint UNDP/World BankEnergy Assessment Program, Indonesia:Issues and Options in the Energy Sector,Washington, D.C., 1981.

13. U.S. Dept. of State, Indonesia's Petroleum Sec-tor, 1984, Jakarta, U.S. Embassy, July 1984.

14. Ibid.

15. Rounding errors in the Geobased Systemprogram result in minor discrepancies inarea estimates. Consequently, columns androws may not add exactly. The errorsintroduced in this way are insignificant.

16. West Java, Central Java, D.I. Jogyakarta,and East Java, D.K.I. Jakarta was includedin West Java.

17. For discussion of the impact of erosion onvarious dimensions of productivity seePierce and others (1983).

61

18. It also takes, at least partially, into accountthe omission of plant cover and conserva-tion practices in the erosion model.

19. For authoritative treatments of maize andcassava production systems in Indonesia,see, respectively, Mink, Dorosh and Perry(1987) and Roche (1984).

20. Crop budgets for many rainfed crops andyears are compiled by the Central Bureauof Statistics (CBS) from large sample sur-veys. They omit family labor, which typi-cally exceeds hired labor use on Java, butprobably best depict the aggregate structureof production cost. Because they are availa-ble for current years, they were used toidentify variable and fixed costs. Data fromthe Survey Agro Ekonomi (Agro-economicSurvey) was also compared with thebudgets prepared by Roche (1983, 1984).Roche's budgets, based on detail surveys of

small samples of farmers throughout Java,include information on family and hiredlabor, purchased inputs and yields. Datafrom the Malang Institute for Food Crops(MARIF) (Brotonegoro, Laumans andStavern 1986) were used to adjust Roche'sbudget to make it more representative ofEast Java as a whole. The MARIF datashows that Kediri Kabupaten has yieldsbetween 40 and 175 percent higher,depending on crop, than the average forEast Java. In addition fertilizer use in Kediriis almost double that of the rest of EastJava.

21. More detail is available in W.B. Magrathand P.L. Arens, "The Costs of Soil Erosionon Java: A Natural Resource AccountingApproach," unpublished paper, WorldResources Institute, November 1987. (Forth-coming as World Bank EnvironmentDepartment Working Paper).

62

References

Aaheim, Asbjorn, 1986. The Resource Rent ofOil and Gas Production. Oslo, Norway:Research Department, Central Bureau ofStatistics.

Alfsen, Knut H., Torstein Bye and LorentsLorentsen, 1987. Natural Resource Account-ing and Analysis: The Norwegian Experience1978-1986. Oslo, Norway: Central Bureau ofStatistics of Norway.

Blades, Derek W. Environment Statistics andthe National Accounts. Paper presented atJoint UNEP-World Bank workshop onEnvironmental Accounting. Washington,D.C., November 17-18, 1986.

Bols, P.L., 1978. The iso-erodent map of Javaand Madura, scale 1:1,000,000. Bogor,Indonesia: Soil Research Institute.

Boskin, Michael J., Marc S. Robinson, TerranceO'Reilly, and Praveen Kumar, 1985. NewEstimates of the Value of Federal MineralRights and Land. American Economic Review(December): 923-936.

Brotonegoro, Soetarjo, QJ. Laumans, and J. PhVan Stavern, 1986. Palawija Food CropsOther Than Rice in East Java Agriculture.Malang: Malang Research Institute for FoodCrops.

Burman, P. ed., 1980. Red Soils in Indonesia.Agricultural Research Report No. 889.

Wageningen: Centre for Agricultural Publish-ing and Documentation.

Carson, Brian and Wani Hadi Utomo, (n.d.).Erosion and Sedimentation Processes in Java.Jakarta: The Ford Foundation/KEPAS.

Central Bureau of Statistics. Statistik Indonesia,Jakarta (various years).

Central Bureau of Statistics. Cost Structure ofFarms Paddy and Palawija, Jakarta (variousyears).

Clark, Edwin H., Jennifer Haverkamp, andWilliam Chapman, 1985. Eroding Soils; TheOff-farm Impacts. Washington, D.C.: Conser-vation Foundation.

Commission Interministerielle des Comptes duPatrimoine Naturel, 1983. Les Comptes duPatrimoine Naturel. Paris: Institut National dela Statistique et des Etudes Economiques.

Corniere, P., 1986. Natural Resource Accountsin France. An Example: Inland Water. InOrganisation for Economic Cooperation andDevelopment, Information and NaturalResources. Paris: OECD

Daly, Herman E., 1986. On Sustainable Devel-opment and National Accounts. In D. Col-lard, D. Pearce and D. Ulph, eds., Economicsand Sustainable Environments: Essays in Honourof Richard Lecomber. New York: Macmillan.

63

Dasgupta, Partha, 1982. The Control ofResources. Cambridge: Harvard UniversityPress.

Department of Public Works, DirectorateGeneral of Water Resources Development,Directorate of Irrigation, General Informationon: Irrigation Operation and MaintenanceActivities in Indonesia, Jakarta, July 1984.

Devarajan, Shantayanan and Robert J. Weiner,1988. Natural Resource Depletion andNational Income Accounts. Unpublished.Harvard University, John F. Kennedy Schoolof Government, Cambridge, Massachusetts.

Directorate General of Reforestation and LandRehabilitation, 1985. Final Report, SecondPhase, The Kali Konto Upper WatershedRegional Development, Trends and Issues.Malang: Government of Indonesia.

Drechsler, Laszlc, 1976. Problems of RecordingEnvironmental Phenomena in NationalAccounting Aggregates. The Review of Incomeand Wealth. Vol. 22, No. 3: 239-252.

Edwards, E.O. and P.W. Bell, 1961. The The-ory and Measurement of Business Income.Berkeley: University of California Press.

Eisner, R., 1980. Capital gains and income: realchanges in the value of capital in the UnitedStates, 1946-77. In Usher, D. (ed.), The Mea-surement of Capital. Studies in Income andWealth, Vol. 45. Chicago: University ofChicago Press.

Eisner, R., 1985. The total incomes system ofaccounts. Survey of Current Business. Vol. 65(1): 24-28.

Fagi, Achmal M. and Cynthia Mackie, 1987.Watershed Management in Upland Java, PastExperience and Future Directions. Paperpresented at the Conference on Soil andWater Conservation on Steep Lands, SoilConservation Society of America. San Juan,Puerto Rico, March 22-27, 1987.

Falcon, Walter, William O. Jones and Scott R.Pearson, 1984. The Cassava Economy of Java.Stanford: Stanford University Press.

Fish, II, 1983. Reservoir Sedimentation Study,Selorejo, East Java, Indonesia. Wallingford:Hydraulics Research Station.

Food and Agriculture Organization, 1981. TropicalForest Resources Assessment Project—Forests ofTropical Asia. In Cooperation with the UnitedNations Environment Programme. Rome: FAO.

Financial Accounting Standards Board, 1977.Financial accounting and reporting by oil andgas companies. Statement of financialaccounting standards no. 19. AccountingStandards Board. Stamford, Connecticut.

Friend, Anthony. Natural Resource Accounting:International Experience. Paper presented atConsultative Meeting of United NationsEnvironment Programme. Geneva, February23-25, 1983.

Gadjah Mada University, (1985). Study ofRegional Capability to Finance the O+MCosts for Irrigation Systems in the ProsidaProjects in the Pemali-Comal Area, CentralJava and in the Bantimurung and Lanrae Pro-ject Areas, South Sulawesi. Jakarta: Ministryof Public Works, Directorate General of WaterResources Development.

Garnasjordet, P. A. and H. Viggo Saebo, 1986.A System of Natural Resource Accounts inNorway. In Organisation for EconomicCooperation and Development, Information andnatural resources. Paris: OECD.

Gauchon, M.J., 1976. Some Aspects of Water-shed Management Economics. Upper SoloWatershed Management and Upland Develop-ment Project. Rome: Food and AgricultureOrganization.

Goeller, H.E. and A. Zucker, 1984. Infiniteresources: the ultimate strategy. Science. Vol.223, (February): 456-62.

64

Goldberg, J., 1980. Indicative Economics of SoilConservation Works. Washington, D.C.:Memorandum December 22, World Bank.

Goldsmith, R.W., 1982. The national balancesheets of the United States, 1952-1975. Chicago:University of Chicago Press.

Goldsmith, R.W., 1985. Comparative nationalbalance sheets. A study of twenty countries,1688-1978. Chicago: University of ChicagoPress.

Government of Indonesia and U.S. Agency forInternational Development, May 1983. Com-posite Report of the Watershed Assessment Team.Appendix II. Washington, D.C.: U.S. Agencyfor International Development.

Hall, Darwin C. and Jane V. Hall, 1984. Con-cepts and Measures of Natural ResourceScarcity with a Summary of Recent Trends.Journal of Environmental Economics and Manage-ment. Vol 11: 363-379.

Herfindahl, O.C. & A.Y. Kneese, 1973. Meas-uring social and economic change: benefitsand costs of environmental pollution. In: M.Moss (ed.), The measurement of economic andsocial performance. Studies in Income and Wealth.No. 38. National Bureau of EconomicResearch, New York.

Hicks, John R., 1946. Value and Capital: An In-quiry into Some Fundamental Principles of Eco-nomic Theory. Oxford: Oxford University Press.

Hueting, R., 1980. New Scarcity and EconomicGrowth. Amsterdam: North-Holla PublishingCompany.

Hueting, Roefie, 1984. Economic Aspects of En-vironmental Accounting. Paper prepared forthe Environmental Accounting Workshop,organized by UNEP and hosted by the WorldBank, Washington, D.C., 5-8 November, 1984.

Juster, F. Thomas, 1973. The framework for themeasurement of economic and social perfor-

mance. In: M. Moss (ed.), The Measurement ofEconomic and Social Performance. Studies inIncome and Wealth. Vol. 38. National Bureauof Economic Research, New York.

Kneese, A. Ayres, R. & d'Arge, R., 1970. Eco-nomics and the Environment: A MaterialsBalance Approach. Resources for the Future,Washington, D.C.: Johns Hopkins Press.

Lai, Rattan, 1987. Effects of Soil Erosion onCrop Productivity. CRC Critical Reviews inPlant Sciences 5:4: 303-367.

Landefeld, J.S. & J.M. Hines 1982. Valuingnon-renewable natural resources: the miningindustries. In: Measuring Nonmarket EconomicActivity: BEA Working Papers, Bureau of Eco-nomic Analysis.

Landefeld, J. Steven and James R. Hines, 1985.National Accounting for Non-renewable Nat-ural Resources in the Mining Industries.Review of Income and Wealth, Vol. 31, No. 1(March): 1-20.

Leighton, Mark and Nengah Wirawan, 1986.Catastrophic Drought and Fire in BorneoTropical Rainforest Associated with the1982-1983 El Nino Southern OscillationEvent. In Prance, Ghillean T. (ed.), TropicalRainforests and the World Atmosphere. AAASSelected Symposium 101. Colorado: West-view Press.

Leipert, Christian, 1986. Social Costs of Eco-nomic Growth as a Growth Stimulus: Theneed for a new indicator of net production.Paper presented at the Conference for a NewEconomics, The Other Economic Summit,Bedford College, London, 16-19th April, 1985.

Leontief, W., 1970. Environmental Repercus-sions and the Economic Structure. Review ofEconomics and Statistics, August.

Lone, Oyvind, 1987. Natural ResourceAccounting and Budgeting: A Short Historyof and Some Critical Reflections on the

65

Norwegian Experience 1975-1987. Paris,France: Organisation for Economic Coopera-tion and Development Environment Direc-torate, November.

Lutz, Ernst and Salah El Serafy. Environmentaland Resource Accounting: An Overview.Environment Department Working Paper No.6, World Bank. Washington, D.C.: June 1988.

MacDonald, Sir M. and Partner Asia, 1987.East Java Irrigation Project, Notes RegardingEfficient O+M. Jakarta: Ministry of PublicWorks, Directorate General of WaterResources Development, February.

Magrath, William B. and Margaret Grosh. AnEconomic Approach to Watershed Manage-ment with Emphasis on Soil Erosion. Paperpresented to the World Forestry Congress.Mexico City, Mexico, August 1985.

Marin, Carlos, 1984. Water Resource Assess-ment: A System of National Water Accounts.Unpublished paper.

McCauley, David S., 1985. Upland CultivationSystems in Densely Populated Watersheds ofthe Humid Tropics. Opportunities and Con-straints Relating to Soil Conservation: A Casefrom Java, Indonesia. Honolulu: East-WestEnvironment and Policy Institute, WorkingPaper.

Mclntosh, J. and Suryata Effendi, 1983. Farm-ing Systems Research and Development. In:Government of Indonesia and U.S. Agencyfor International Development, CompositeReport of the Watershed Assessment Team.Appendix II. Washington, D.C.: U.S. Agencyfor International Development.

Miller, Merton H. and Charles W. Upton, 1985.A Test of the Hotelling Valuation Principle.Journal of Political Economy. Vol. 93, No. 1:1-25.

Mink, Stephen D. and Paul A. Dorosh, 1987.An Overview of Corn Production. In:

C. Peter Timmer, (ed.), The Corn Economy ofJava. Ithaca, N.Y.: Cornell University Press.

Mink, Stephen D., Paul A. Dorosh and DouglasH. Perry, 1987. Corn Production Systems. In:C. Peter Timmer (ed.), The Corn Economy ofJava. Ithaca, N.Y.: Cornell University Press.

Molster, H.C., 1978. Methods of Estimated Fer-tilizer Response with an Application to UreaUse on Rice in Jogjakarta, Indonesia. Agricul-tural Research Report No. 877. Wageningen:Centre for Agricultural Publishing andDocumentation.

Montgomery, Roger, 1981. Maize Yield Increasedin East Java. Bulletin of Indonesian EconomicStudies, Vol. 18 No. 3 (November): 74-85.

Muryadi, Amir. Catchment Management andits Role in Water Resources Development inIndonesia. Paper presented for the ESCAPSeminar on Catchment Management forOptimum use of Land and Water Resources.Hamilton, New Zealand, March 15-19, 1982.

Nordhaus, W.D. & J. Tobin 1973. Is growthobsolete? In: M. Moss (ed.), The Measurementof Economic and Social Performance. Studies inIncome and Wealth. No. 38. National Bureau ofEconomic Research, New York.

Organisation for Economic Cooperation andDevelopment, 1986. Declaration on environ-ment resources for the future. In OECD,OECD and the Environment. Paris: OECD.

Peskin, Henry M. and Janice Peskin, 1978. TheValuation of Nonmarket Activities in IncomeAccounting. Review of Income and Wealth. Vol.24, March: 41-70.

Peskin, Henry M., (n.d.). National Accountsand the Development Process (Illustrationwith Tanzania). Unpublished paper. Wash-ington, D.C.: Resources for the Future.

Peskin, Henry M., 1976. A National Account-ing Framework for Environmental Assets.

66

Journal of Environmental Economics and Manage-ment. Vol. 2: 255-262.

Peskin, Henry M., 1980. Two papers onnational accounting and the environment.Discussion Paper D-71, Resources for theFuture, Washington, D.C., October.

Pierce, F.J., W.E. Larson, R.H. Dowdy andW.A.P. Graham, 1987. Productivity ofSoils—Assessing Long Term Changes Due toErosion. Journal of Soil and Water Conservation.Vol. 38 no. 1, January-February: 39-44.

Rampubulon, S.M.H., and B. Saragih, 1976.'Model Farm' Upland Farming Technology inthe Citanduy River Basin: A State of the Art.Ciamis: USESE.

Roche, Frederick C , 1983. Cassava ProductionSystems in Java and Madura. Ph.D. disserta-tion. Palo Alto, CA: Stanford University.

Roche, Frederick C , 1984. Production Systems.In: Falcon and others, The Cassava Economy ofJava. Stanford: Stanford University Press.

Roche, Frederick C , 1987. Sustainable Farm De-velopment in Java's Critical Lands: Is a "GreenRevolution" Really Necessary? Unpublishedmanuscript. Ithaca: Cornell University.

Roedjito, D.M. and H. Soenaro. SedimentManagement. Paper prepared for the Work-shop on Integrated River Basin Developmentand Watershed Management. Jakarta, March22-29, 1986.

Ruzicka, I., 1978. Rent Appropriation in Indo-nesian Logging: East Kalimantan 1972/3-1976/7. Bulletin of Indonesian Economic Studies.Vol. 15, No. 2 (July): 49-56.

Securities and Exchange Commission, 1978.Adoption of requirements for financialaccounting and reporting practices for oil andgas producing activities. ASR No. 253. Secu-rities and Exchange Commission, Washing-ton, D.C.

Securities and Exchange Commission, 1981.Financial reporting by oil and gas producers.ASR No. 259, Securities and Exchange Com-mission, Washington, D.C.

Sfeir-Younis, Alfredo, 1985. Soil Conservationin Developing Countries, A BackgroundReport. Unpublished manuscript. Washing-ton, D.C: World Bank.

Soladay, J., 1980. Measurement of Income andProduct in the Oil and Gas Industry. In: D.Usher (ed.), The Measurement of Capital.Studies in Income and Wealth. Vol. 45. Univer-sity of Chicago Press.

Southgate, Douglas. The Off-Farm Benefits ofSoil Conservation in a Hydroelectric Water-shed. Paper presented at the Annual Meetingof the American Agricultural EconomicsAssociation. Reno, Nevada, July 1986.

Stauffer, Thomas R. Accounting for "wastingassets": Income Measurement for Oil andMineral-Exporting Rentier States. Paperpresented to Eighteenth General ConferenceInternational Association for Research inIncome and Wealth, Luxembourg, 1983.

Sunarno and Sutadji, 1982. Reservoir-Sedimen-tation Technical and Environmental Effect.Proceedings of the International Commissionon Large Dams, Rio de Janeiro, pp. 489-508.

Theys, J., 1984. Environmental accounting andits use in development policy. Proposalsbased on the French experience. Unpublishedmanuscript.

Timmer, C. Peter, ed., 1987. The Corn Economyof Java. Ithaca, N.Y.: Cornell University Press.

Turner, R. Eugene, 1975. The Sagara ArakanReclamation Project: The Impact on Commer-cial Fisheries. Reports to Engineering Consul-tants, Inc., Denver, Colorado.

United Nations, Department of Economic andSocial Affairs, 1968. A System of National

67

Accounts. Statistical Papers, Series F, No. 2,Rev. 3. United Nations, New York.

United Nations, Department of Economic andSocial Affairs, 1977. Provisional internationalguidelines on the national and sectoralbalance-sheet and reconciliation accounts ofthe system of national accounts. StatisticalPapers, Series M, No. 60. United Nations,New York.

United Nations, Statistical Office 1979. FutureDirections for Work on the System of NationalAccounts. United Nations, New York.

United Nations, Economic and Social Council,Statistical Commission, 1980. Future Direc-tions for Work on System of NationalAccounts.

Usher, Dan, 1980. The Measurement of EconomicGrowth. New York: Columbia UniversityPress.

Ward, Michael, 1982. Accounting for theDepletion of Natural Resources in theNational Accounts of Developing Economies.Paris, France: Development Centre, Organi-sation for Economic Cooperation andDevelopment, June.

Ward, Michael. The Impact of Resource Deple-tion on the National Accounts. Paperpresented to Conference on "NationalAccounts and their Uses in DevelopmentPlanning in the Arab Countries," Kuwait,May, 1984.

Weber, J.L., 1983. The French NaturalPatrimony Accounts. Statistical Journal of theUnited Nations. ECI 1:419-444.

Weiller, Edward, 1983. The Use of Environ-mental Accounting for Development Plan-ning. Report to the United Nations Environ-ment Programme, New York, January.

World Bank, 1987. Indonesia, The OuterIslands: Issues in the Sustainable Use ofLand and Forest Resources. Main Report.Unpublished paper.

World Bank, 1988. World Development Report,1988. Oxford: Oxford University Press.

World Commission on Environment andDevelopment, 1987. Our Common Future.Oxford: Oxford University Press.

68

World Resources Institute

1709 New York Avenue, N.W.Washington, D.C. 20006 U.S.A.

WRI's Board of Directors:Matthew NimetzChairmanJohnE. CantlonVice ChairmanJohnH. AdamsRobert O. AndersonRobert O. BlakeJohnE. BrysonWardB. ChamberlinRichard M. ClarkeAlice F. EmersonJohn FirorJose GoldembergMichio HashimotoCurtis A. HesslerMartin HoldgateJames A. JosephIan K. MacGregorAlan R. McFarlandRobert S. McNamaraPaulo Nogueira-NetoThomas R. OdhiamboRuth PatrickJames Gustave SpethMaurice F. StrongM.S. SwaminathanMostafa K. TolbaRussell E. TrainAlvaro UmanaGeorge M. Woodwell

James Gustave SpethPresidentMohamed T. El-AshryVice President for Research and Policy AffairsJ. Alan BrewsterVice President for Administration and FinanceJessica T. MathewsVice PresidentWallace D. BowmanSecretary-Treasurer

The World Resources Institute (WRI) is a policyresearch center created in late 1982 to helpgovernments, international organizations, and privatebusiness address a fundamental question: How cansocieties meet basic human needs and nurtureeconomic growth without undermining the naturalresources and environmental integrity on which life,economic vitality, and international security depend?

Two dominant concerns influence WRI's choice ofprojects and other activities:

The destructive effects of poor resource managementon economic development and the alleviation ofpoverty in developing countries; and

The new generation of globally importantenvironmental and resource problems that threatenthe economic and environmental interests of theUnited States and other industrial countries and thathave not been addressed with authority in their laws.

The Institute's current areas of policy research includetropical forests, biological diversity, sustainableagriculture, energy, climate change, atmosphericpollution, economic incentives for sustainabledevelopment, and resource and environmentalinformation.

WRI's research is aimed at providing accurateinformation about global resources and population,identifying emerging issues and developing politicallyand economically workable proposals.

In developing countries WRI provides field servicesand technical program support for governments andnon-governmental organizations trying to managenatural resources sustainably.

WRI's work is carried out by an interdisciplinary staffof scientists and experts augmented by a network offormal advisors, collaborators, and cooperatinginstitutions in 50 countries.

WRI is funded by private foundations, UnitedNations and governmental agencies, corporations, andconcerned individuals.

c/1Dcp

5?o

S"

CD

—

a; =c

a;

Z <

=