26 March 2000

Accounting for Renewableand Environmental Resources

, a blue ribbon panel of the National Academy ofSciences’ National Research Council completed a congression-ally mandated review of the work that the Bureau of EconomicAnalysis (BEA) had published on integrated economic and en-vironmental accounts. The panel’s final report commendedBEA for its initial work in producing a set of sound and ob-jective prototype accounts. The November 1999 issue of theSURVEY OF CURRENT BUSINESS contained an article by WilliamD. Nordhaus, the Chair of the Panel, that presented an overviewof the major issues and findings and a reprint of chapter 5,“Overall Appraisal of Environmental Accounting in the UnitedStates.” Chapter 3, “Accounting for Subsoil Mineral Resources”was reprinted in the February 2000 issue; chapter 4, “Accountingfor Renewable and Environmental Resources” is reprinted below.

This article is reprinted with permission from Nature’s Num-bers: Expanding the National Economic Accounts to Includethe Environment. Copyright of the National Academy Press,Washington DC. This is a report of the National Research Coun-cil, prepared by the Panel on Integrated Environmental andEconomic Accounting and edited by William D. Nordhaus andEdward C. Kokkenlenberg.

T chapter reviewed issues involvedin extending the national accounts to include

subsoil assets. This chapter focuses on two otheraspects of environmental accounting: renewableand environmental resources. BEA has proposedcovering these two categories of resources in fu-ture work on integrated accounting. As discussedin Chapter 1, Phase II of that work would focuson different classes of land (e.g., agriculture, for-est, and recreation land), on timber, on fisheries,and on agricultural assets such as grain stocks andlivestock. Phase III would address environmentalresources, including, for example, air, uncultivatedbiological resources, and water.

The general principles set forth in Chapter 2 in-dicate that increasingly severe obstacles are likelyto arise as the national accounts move further fromthe boundaries of the market economy. The dis-cussion in this chapter confirms the premise thatBEA’s Phase III raises the most difficult concep-tual, methodological, and data problems. Thisfinding presents a dilemma that must be faced inexpanding the accounts: Should follow-on effortsfocus on those resources that can be most eas-ily included given existing data and methods, orshould BEA focus on including those resources thatwould have the largest impact on our understand-

ing of the interaction between the U.S. economyand the environment? The panel’s investigation,while based on data that are highly imprecise andin some cases speculative, suggests that the de-velopment of the accounts proposed for PhaseIII would be likely to encompass the most sig-nificant economy-environment interactions. Thisobservation is tempered by the realization that todate nothing approaching adequate comprehen-sive environmental accounting for a country of thecomplexity of the United States has yet been un-dertaken. For BEA or the federal government toprepare a full set of environmental accounts wouldrequire a substantial commitment.

This chapter provides a review of the issuesinvolved in accounting for renewable and envi-ronmental resources. It is not intended to be acomprehensive review of work in this area. Rather,it delineates the issues that are involved in envi-ronmental accounting and presents two importantspecific examples that illustrate these issues. Thefirst section reviews BEA’s efforts in environmentalaccounting to date. Next, we analyze how stocksand flows of residuals from human activities relateto natural sources of residuals, natural resourceassets, stocks, flows, and economic activity. Thethird section examines issues involved in account-ing for renewable and environmental resources.The chapter then turns to general issues associatedwith the physical data requirements of environ-mental accounting and with valuation. We nextinvestigate in greater detail the cases of forests andair quality to illustrate how augmented accountingmight actually be done. The chapter ends with thepanel’s conclusions and recommendations in thearea of accounting for renewable and environmen-tal resources. Appendix B identifies potentiallyuseful sources of data for developing supplementalaccounts identified by the panel in the course of itsinvestigation.

BEA EFFORTS TO DATE INACCOUNTING FOR RENEWABLE AND

ENVIRONMENTAL RESOURCES

This section reviews BEA’s initial design for itssupplemental accounts for natural-resource and

March 2000 • 27

environmental assets. A more complete evalua-tion of BEA’s efforts on forests is included laterin the chapter. As discussed in Chapter 2, acritical issue involved in the development of aug-mented accounts is setting the boundary. Howfar from the boundary of the marketplace should

TABLE 4

This table can serve as an inventory of the estimates available for the IEESA’s. In decrible tangible stock and inventories, from BEA’s national income and product accouand lows of the range based on alternative valuation methods, from the companio(selected rows 23–35 and 42–47) and some environmental assets (selected rows

PRODUCED ASSETSMade assets ........................................................................................................................

Fixed assets .....................................................................................................................Residential structures ..................................................................................................Fixed nonresidential structures and equipment ..........................................................

Natural resource related ..........................................................................................Environmental management ...............................................................................

Conservation and development ......................................................................Water supply facilities .....................................................................................

Pollution abatement ............................................................................................Sanitary services ............................................................................................Air pollution abatement and control ...............................................................Water pollution abatement and control ..........................................................

Other ....................................................................................................................Inventories ........................................................................................................................

Government ..................................................................................................................Nonfarm ........................................................................................................................Farm (harvested crops, and livestock other than cattle and calves) ........................

Corn .........................................................................................................................Soybeans .................................................................................................................All wheat ..................................................................................................................Other ........................................................................................................................

Developed natural assets ..................................................................................................Cultivated biological resources ........................................................................................

Cultivated fixed natural growth assets ........................................................................Livestock for breeding, dairy, draught, etc .............................................................

Cattle ...................................................................................................................Fish stock ............................................................................................................Vineyards, orchards ............................................................................................Trees on timberland ............................................................................................

Work-in-progress on natural growth products ........................................................Livestock raised for slaughter ............................................................................

Cattle ...............................................................................................................Fish stock ........................................................................................................

Calves ..................................................................................................................Crops and other produced plants, not yet harvested .......................................

Proved subsoil assets ......................................................................................................Oil (including natural gas liquids) ................................................................................Gas (including natural gas liquids) .............................................................................Coal ..............................................................................................................................Metals ...........................................................................................................................Other minerals .............................................................................................................

Developed land ................................................................................................................Land underlying structures (private) ...........................................................................Agricultural land (excluding vineyards, orchards) .......................................................

Soil ...........................................................................................................................Recreational land and water (public) ..........................................................................Forests and other wooded land ..................................................................................

NONPRODUCED/ENVIRONMENTAL ASSETSUncultivated biological resources ....................................................................................

Wild fish .......................................................................................................................Timber and other plants and cultivated forests ..........................................................Other uncultivated biological resources ......................................................................

Unproved subsoil assets ..................................................................................................Undeveloped land ............................................................................................................Water (economic effects of changes in stock) ...............................................................Air (economic effects of changes in stock) ....................................................................

n.a. = Not available* The calculated value of the entry was negative.

the purview of the environmental accounts ex-tend? Table 4–1 shows BEA’s tentative decisionson how it proposed to structure its supplemen-tal accounts (BEA’s Integrated Environmental andEconomic Satellite Accounts [IEESA] from Bureauof Economic Analysis, 1994a: Table 1). Phase II

–1 IEESA Asset Account, 1987[Billions of dollars]

easing order of quality, the estimates that have been filled in are as follows: For made assets, estimates of reproduc-nts or based on them, and pollution abatement stock, from BEA estimates (rows 1–21); for subsoil assets, the highsn article (rows 36–41); and best available, or rough-order-of-magnitude, estimates for some developed natural assets48–55) prepared by BEA. The ‘‘n.a.’’—not available—entries represent a research agenda.

Opening Stocks

Change

Total, Net(3+4+5)

Depreciaton,Depletion,

Degradation

CapitalFormation

Revaluationand OtherChanges

ClosingStocks(1+2)

Row(1) (2) (3) (4) (5) (6)

............. 1 11,565.9 667.4 –607.9 905.8 369.4 12,233.3

............. 2 10,535.2 608.2 –607.9 875.8 340.2 11,143.4

............. 3 4,001.6 318.1 –109.8 230.5 197.4 4,319.7

............. 4 6,533.6 290.1 –498.1 645.3 142.9 6,823.7

............. 5 503.7 23.1 –19.2 30.3 12.0 526.8

............. 6 241.3 8.4 –7.0 10.6 4.7 249.6

............. 7 152.7 3.6 –4.4 5.3 2.7 156.4

............. 8 88.5 4.8 –2.5 5.3 2.0 93.3

............. 9 262.4 14.7 –12.2 19.7 7.3 277.1

............. 10 172.9 12.8 –5.6 13.7 4.8 185.8

............. 11 45.3 .6 –4.1 3.5 1.3 45.9

............. 12 44.2 1.3 –2.5 2.6 1.2 45.5

............. 13 6,029.9 267.0 –478.9 615.0 130.9 6,296.9

............. 14 1,030.7 59.3 .......................... 30.1 29.2 1,090.0

............. 15 184.9 6.8 .......................... 2.9 3.8 191.7

............. 16 797.3 62.4 .......................... 32.7 29.7 859.7

............. 17 48.5 –9.9 .......................... –5.5 –4.4 38.6

............. 18 10.2 .3 .......................... –1.1 1.4 10.5

............. 19 5.0 –.1 .......................... –1.0 .9 4.9

............. 20 2.6 0.0 .......................... –.2 .2 2.6

............. 21 30.7 –10.1 .......................... –3.2 –6.9 20.6

............. 22 n.a. n.a. n.a. n.a. n.a. n.a.

............. 23 n.a. n.a. n.a. n.a. n.a. n.a.

............. 24 n.a. n.a. n.a. n.a. n.a. n.a.

............. 25 n.a. n.a. n.a. n.a. n.a. n.a.

............. 26 12.9 2.0 n.a. –.3 2.3 14.9

............. 27 n.a. n.a. n.a. n.a. n.a. n.a.

............. 28 2.0 .2 n.a. 0.0 .2 2.2

............. 29 288.8 47.0 –6.9 9.0 44.9 335.7

............. 30 n.a. n.a. .......................... n.a. n.a. n.a.

............. 31 n.a. n.a. .......................... n.a. n.a. n.a.

............. 32 24.1 7.5 .......................... 0.0 7.5 31.6

............. 33 n.a. n.a. .......................... n.a. n.a. n.a.

............. 34 5.0 .9 .......................... –.5 1.4 5.9

............. 35 1.8 .3 .......................... .1 .2 2.1

............. 36 270.0 - 1,066.9 57.8 - 116.6 –16.7 - 61.6 16.6 - 64.6 58.0 - –119.6 299.4 - 950.3

............. 37 58.2 - 325.9 –22.5 - 84.7 –5.1 - –30.6 5.8 - 34.2 –23.1 - –88.3 35.7 - 241.2

............. 38 42.7 - 259.3 6.6 - 57.2 –5.6 - –20.3 4.1 - 14.9 8.1 - –51.8 49.4 - 202.2

............. 39 140.7 - 207.7 2.2 - 3.4 –5.4 - –7.6 4.4 - 6.3 3.2 - –2.1 143.0 - 204.2

............. 40 (*) - 215.3 67.2 - –29.5 –.2 - –2.2 2.2 - 9.2 65.2 - 22.5 38.5 - 244.8

............. 41 28.4 - 58.7 4.3 - .8 –.4 - –.9 .1 - .0 4.6 - .1 32.8 - 57.9

............. 42 n.a. n.a. n.a. n.a. n.a. n.a.

............. 43 4,053.3 253.0 n.a. n.a. n.a. 4,306.3

............. 44 441.3 42.4 n.a. –2.8 45.2 483.7

............. 45 n.a. n.a. –.5 n.a. n.a. n.a.

............. 46 n.a. n.a. –.9 .9 n.a. n.a.

............. 47 285.8 28.8 n.a. –.6 29.4 314.6

............. 48 n.a. n.a. n.a. n.a. n.a. n.a.

............. 49 n.a. n.a. n.a. n.a. n.a. n.a.

............. 50 n.a. n.a. n.a. n.a. n.a. n.a.

............. 51 n.a. n.a. n.a. n.a. n.a. n.a.

............. 52 n.a. n.a. n.a. n.a. n.a. n.a.

............. 53 n.a. n.a. –19.9 19.9 n.a. n.a.

............. 54 ............................ n.a. –38.7 38.7 n.a. ........................

............. 55 ............................ n.a. –27.1 27.1 n.a. ........................

NOTE: Leaders (...) indicate an entry is not applicable.Source: Bureau of Economic Analysis (1994a) SURVEY OF CURRENT BUSINESS, April 1994. The table has been

slightly simplified for this report.

28 • March 2000

1. All quotations in this section are from the Bureau of Economic Analysis(1994a).

of BEA’s development of supplemental tables fo-cused on assets listed in rows 22–35 and 42–47 ofTable 4–1, while Phase III considers rows 48–55.Because BEA has not completed Phases II and III,actual decisions on what will be included have yetto be made. Each of the following sections of thischapter considers an element of how to draw theline. While an ideal set of accounts would con-tain “everything,” this chapter examines practicalissues that arise in constructing actual accountsbased on available data and tools. As will be seen,the practical is likely to fall far short of the ideal.

Pollution Abatement and ControlExpenditures

One particular entry in the environmen-tal accounts—pollution abatement and controlexpenditures—has been the subject of detailed in-vestigation by BEA for many years. These items areshown for 1987 in rows 5–12 of Table 4–1. The Bu-reau of the Census began collecting these data andBEA reporting them in 1972 (with some breaks inthe series); these efforts were suspended in 1995 be-cause of budget cuts. Reporting of these costs doesnot extend the accounts, but rather reorganizes theexisting accounts to provide a better indication ofthe interaction between the environment and theeconomy.

The limitations of these data are well recognizedand were discussed in Chapter 2. Many of the costsincluded in the data overstate the cost of pollu-tion control, while other pollution-reducing costsare omitted because they involve changes in pro-cesses. There is also controversy about the extentto which stringent pollution control regulationsmay have a chilling effect on innovation and tech-nological change. Finally, little thought has beengiven to the appropriate treatment of purchasesof emission permits, which are likely to becomea more important feature of environmental reg-ulation in the future. Despite their limitations,however, data on pollution abatement are likelyto be among the most precise of the data in theenvironmental accounts, and they have been ex-tremely useful for understanding trends and levelsin control costs and for examining how environ-mental programs have affected productivity. Thepanel finds that the data on pollution abatementexpenditures are valuable and, as noted in the finalsection of this chapter, recommends that funds beprovided to improve the design and recommencecollecting these data.

Other Sectors of the Proposed Accounts

As reported by BEA, the quality of actual entriesin published supplemental accounts for Phase IIand III assets ranges from relatively good to con-ceptually defective.1 For Phase II assets, estimateswithin the category “developed land” are describedas “of uneven quality” (p. 45). According toBEA, agricultural land values are “relatively goodand are based on U.S. Department of Agricultureestimates of farm real estate values less BEA’s esti-mates for the value of structures” (p. 45). BEA hasnot attempted to estimate the value of recreationalland, but has entered federal maintenance and re-pair expenditures as an investment (see Table 4–1)and “assumed that these expenditures exactly off-set the degradation/depletion of recreational land”(p. 45). BEA indicates that this assumption ismade only for purposes of illustration and is “notto imply any judgment about the true value ofdegradation/depletion” (p. 45). A more detaileddiscussion of BEA estimates for timber and land inforests is presented later in this chapter.

For Phase III assets, BEA has entered “n.a.”for most of the items, indicating that these esti-mates have not yet been developed. Entries forinvestment in and degradation of water, air, andundeveloped land are included, however. As inthe case of developed recreational land, BEA hasassumed that maintenance exactly offsets degrada-tion, noting that this assumption provides entriesthat “are simply place markers” (p. 46). In thepanel’s view, the use of maintenance expendituresas degradation costs is highly misleading, and thisprocedure should not be followed in the future.Entering “n.a.” would be more accurate. The panelnotes, however, that these estimates do not neces-sarily reflect BEA’s planned approaches, but wereincluded by BEA to show the current state of dataand research.

Regarding future plans, the United Nations Sys-tem of Integrated Environmental and EconomicAccounting (SEEA) “does not recommend that thestock of air—which is truly a global common—orwater be valued; instead it recommends that valu-ation be limited to changes in these assets—theirdegradation and investments in their restoration”(p. 46). It should be emphasized that the entriesfor environmental assets in Table 4–1 are highlyoversimplified. Some components of air quality,such as greenhouse gases and stratospheric ozone,are truly global assets and services; others, such asreductions in urban smog, are local and regional

March 2000 • 29

public goods. Additional dimensions that need tobe incorporated are relations to external events,spatial resolution, and nonlinearities in damages.The discussion of air quality later in this chapterillustrates its multiple dimensions. Similarly, wa-ter quality and quantity, undeveloped land, anduncultivated biological resources are compositesof many different assets and quality characteristicsthat provide multiple goods and services.

BEA’s efforts have focused on the asset accounts.A preliminary table for a production account with-out entries is included in BEA’s report on itsdevelopment of the IEESA (Bureau of EconomicAnalysis, 1994a, 1994b). Production of marketgoods and services from these natural assets—e.g.,timber, agricultural crops, fish—is already in thecore production accounts. Greater attention isneeded to identifying, measuring, and valuing thespecific types of nonmarket goods and servicesproduced by these assets.

POLLUTANT EMISSIONS AND THEIRRELATION TO STOCKS, FLOWS, AND

ECONOMIC ACTIVITY

Before constructing environmental accounts, it isnecessary to determine the interactions betweennatural resources and the environment and eco-nomic activity. It is essential to understand thekey physical flows and stocks and how they affecthumans and economic activities and values. Acomplete accounting requires detailed knowledgeof the physical properties of resources and pollu-tants as described in fate, transport, and impactor damage models, as well as the service flows tomarket and nonmarket sectors.

Figure 4–1 illustrates key relationships amongemissions, stocks of pollutants, natural-resourceassets, and economic activities in different sectors.As the figure shows, economic activities producea variety of uninternalized emissions and resid-

30 • March 2000

uals that find their way into the environment.Many of the pollutants of concern are residuals thatalso have natural sources—sulfur, carbon dioxide,carbon monoxide, nitrogen compounds—and areemitted during volcanic eruptions, produced byforests and wetlands, or released from wildfires.Other residuals of concern—such as chlorofluo-rocarbons (CFCs) and many pesticides used inagriculture—are anthropogenic and have no natu-ral sources. In terms of effects on human activities,the sources of the residuals are not important.What may be important is that human activitieshave increased the levels occurring in the environ-ment, concentrated them to a degree that makesthem dangerous, or relocated them to areas wherepeople or economic activities are exposed to themat high levels.

Whether from natural sources or human activi-ties, environmental variables can affect economicwell-being in three general ways, as illustrated inFigure 4–1: (1) direct effects on consumption orincome of households, industry, and government;(2) accumulation in the environment of stocks ofresiduals that then affect economic activities oreconomic assets; and (3) effects on the serviceflows of economic assets (capital stock, natural re-sources, or human resources), such as recreation,clean air to breathe, and navigable river channelsfree of sedimentary deposits.

Direct Effects

Environmental variables affect human and naturalsystems directly. Urban smog, whose concentra-tions change daily or even hourly, is an obviousexample. Sulfate and nitrate aerosols, pollutantscontributing to acid precipitation, remain in theatmosphere for a matter of days. These pollutantshave short-term health effects, reduce visibility,interfere with recreational activities, affect cropgrowth, and present their own set of problems foraccounting. In many cases, the substances emittedare precursor emissions; that is, they react chem-ically in the atmosphere with other substances toform the substance that is ultimately damaging tohumans or ecosystems. There are also complexnonlinearities because the formation of the dam-aging substance depends on the level of precursoremissions, weather conditions, and the presence ofother substances with which the precursor emis-sions react. All of these processes vary on an hourly,daily, and seasonal basis. Emissions, concentra-tions, and impacts of damaging substances alsovary spatially, and there may be important thresh-old effects as well. Above all, there is the “weed

syndrome”—the fact that the same substance maybe beneficial or harmful depending on where itis, how much of it there is, the time and dura-tion of exposure, and what organism is absorbingit. Virtually every substance on earth, from wa-ter to plutonium, can be an economic good or aneconomic weed depending on the circumstances.

One of the most important difficulties is that thephysical measurements used are often inaccurateindicators of actual human exposures. Averageemissions of the precursor pollutant, average con-centrations over the year, or concentration data forlimited sites are generally not representative of con-centrations to which the population is exposed andmay be a misleading basis for developing damageestimates. For example, tropospheric ozone formsmainly in warm weather. Thus total annual hydro-carbon emissions, the precursor to troposphericozone, are a poor indicator of potential levels oftropospheric ozone. Tropospheric ozone levelsalso vary significantly over the distance of a few cityblocks. One of the major challenges both for betterenvironmental policy and for the construction ofenvironmental accounts is to obtain better meas-ures of direct human exposure to the importantharmful substances among a representative sampleof people.

Accumulation of Stocks

Many environmental problems result from the ac-cumulation of residuals. These substances includemost radiatively active trace gases, which remainin the atmosphere for decades or centuries, andmany radioactive materials, which have half-livesof decades or centuries. Similarly, recovery fromstratospheric ozone depletion is a process requir-ing years or decades. and agricultural chemicalsoften migrate very slowly through soils, contam-inating drinking water only after several years ordecades.

Environmental accounting therefore needs todevelop and include appropriate methods to ac-count for those persistent pollutants, such as heavymetals that accumulate in the environment andlast for many years. Each year’s emissions orproduction of residuals adds to the stock in theenvironment, and it is necessary to understand theprocesses by which these stocks decay or dissipate.In some cases (as with radioactive substances),those processes are easily understood, while inother cases (such as subsoil toxins or the carbon cy-cle), understanding the processes poses enormousscientific challenges. In the economic accounts, thestock-flow dynamics are similar to those of gross

March 2000 • 31

investment and depreciation of capital. Whilethere is a conceptual similarity, however, there isno readily observable market price for these stockchanges. Hence, valuation of a change in stockrequires estimating the value of the impact of addi-tions over the lifetime of the stock, accounting fordissipation, and appropriately discounting futureeffects. It should also be recognized that, with afew exceptions, the stocks are extremely heteroge-neous, so that measuring a simple “environmentalcapital stock” is likely to be extremely difficult.

Effects on Economic Assets

Both short-lived and long-lived residuals can affecteconomic activity over a number of years throughtheir effects on other economic assets, in particu-lar produced capital goods such as buildings andequipment. For example, acid precipitation cancause deterioration of buildings. Accumulatedgreenhouse gases can result in coastal flooding andhigher storm surges, thereby adversely affecting thevalue of existing coastal structures. Pollutants suchas lead can cause long-lasting health consequences,impacts on intellectual functions, and prematuredeath.

ISSUES INVOLVED IN ACCOUNTINGFOR RENEWABLE AND

ENVIRONMENTAL RESOURCES

The previous section addressed the major waysin which natural resources and the environmentinteract with economic activity. Depending onthe intended uses of the data, there are differ-ent approaches to structuring environmental andnatural-resource accounts. The most completeaccounting structure would treat all the relation-ships in Figure 4–1. However, constructing sucha complete set of accounts is infeasible today, andgovernments must choose areas for investigationstrategically in accordance with their national eco-nomic and environmental goals and interests. Thissection delineates some possible approaches to ac-counting for natural and environmental resourcesand activities.

Production and Income Accounts

A complete set of production accounts would iden-tify all the cross-relationships among industry,household, government, and natural sources ofemissions or residuals, as well as the nonmarketedcurrent account input services provided by na-ture and the productive contribution of nature to

final demand. Current-year activities would in-clude production of residuals, just as traditionaleconomic accounts include production accounts.A complete set of accounts would incorporateflows of residuals from abroad, similar to im-ports of goods and services. It would also benecessary to calculate the “price”—negative orpositive—indicating whether the effect was adverseor beneficial. The accounting for current-yearactivities would include final uses of residuals,identifying effects on final consumption, flowsabroad, and contributions to capital stocks, justas traditional accounting frameworks identify finalconsumption of goods and services, exports, andgross capital accumulation.

Accounting for Capital Assets

It is important to measure the volumes and val-ues of the nation’s natural assets for many reasons.One purpose is simply to determine general trends.Another, illustrated in Table 4–1, is to determinethe relative magnitudes of different assets. A fur-ther reason arises in the context of sustainableeconomic growth. As discussed in Chapter 2, onecan calculate measures of sustainable income if onecorrects conventional measures of national incomeby including the value of the change in the stocksof natural and other assets.

For all of these reasons, we would ideally liketo have measures of the value and volume of thenation’s natural assets; thus we must include meas-ures not only of “made assets,” such as housesand computers, but also renewable resources, suchas timber or the fertility of land, and nonrenew-able assets, such as oil and mineral resources.It is important to know whether the economyis generating an ever-growing stock of damagingenvironmental residuals that will pose a large eco-nomic burden on future generations. We want toknow whether the economic value of investmentsin tangible, human, and technological capital ismore than offsetting whatever depletion of naturalassets is occurring.

There is a close connection between the pro-duction accounts and the asset accounts (seeChapter 2). As noted above, measures of compre-hensive income or of sustainable income includenot only current consumption flows, but also thevalue of the change in the stocks of assets. Henceaugmented accounting requires careful and accu-rate measurement of both assets and consumptionflows. Such measurement is currently undertakenwithin the boundary of the marketplace, but aug-mented accounting would require extending that

32 • March 2000

2. Watershed valuation is an example of a holistic approach (see Andersonand Rockel [1991] and Green et al. [1994] as examples).

boundary for both assets and consumption in aconsistent manner. The conceptual basis for as-set valuation in environmental accounts parallelsclosely that in the conventional accounts. De-pletion and degradation of natural resources isconceptually similar to depreciation of producedcapital assets. Stocks of residuals can decay ordissipate, a process that is again conceptually sim-ilar to depreciation of produced assets. Naturalgrowth of biological resources, recharge of ground-water resources, and accumulation of residuals areconceptually similar to gross capital formation orinvestment. Net accumulation of assets is equal tothe value of the change in stocks. Many of the issuesinvolved in constructing chain indexes of valuesand volumes translate directly into measurementof resource and environmental stocks.

However, some special conceptual difficul-ties arise in measuring stocks of natural assets.Natural-resource assets (like a physical plant orpiece of equipment) are complex systems of com-ponent parts that have value because of the waythey work together. Since produced capital assetsare generally purchased or constructed as modules,they can be valued on the basis of their own marketprices, rather than their synergistic contributionto output. To take an analogy, a baseball player’scontribution to the team is a complex function notonly of hitting, pitching, and fielding, but also oftemperament, teamwork, and verbal abilities; froman accounting perspective, however, the economiccontribution is simply wages and other compen-sation. For environmental assets, determining thevalue will become difficult when the effort extendsbeyond the market boundary. Consider a forest.How can the value of the stumpage in the forest beseparated from the forest’s contribution to erosioncontrol, air quality, and biodiversity?

Even when markets produce evidence of thevalue of a bundle of assets—the composite value ofsoils, timber, nearness to water, and recreation—itmay be difficult to separate out the values of the dif-ferent components without applying complicatedstatistical procedures. Sometimes, the separationis misleading, as when the value of the componentsdepends on their being together. An assembledbicycle is different from a pile of parts; similarly,forests, lakes, rivers, farmland, and coastal es-tuaries are valuable because of the way they areassembled.

One possible way of avoiding this difficulty is toredefine assets in terms of particular functions orcharacteristics, an approach similar to that takenin hedonic valuation, whereby goods are viewed aspackages of characteristics. This approach would

be similar to redefining an automobile as a com-bination of transportation mode, public-healthmenace, and status symbol. Under this approach,an asset is valued in terms of the sum of the val-ues of its various characteristics. In this view,there is little point in trying to analyze the to-tal value of holistic assets such as land or air orclimate; rather, one undertakes the more modesttask of looking at the different functions involved.2

BEA’s treatment of soil erosion is consistent withthis approach; agricultural land is treated as theasset and the soil depth and organic-matter con-tent as characteristics of the land. Other aspectsof land quality—local climate or ambient level ofpollution—can be considered in a similar manner.Identification of the economic effects of erosion onthe value of land makes the resource link explicit.

Thus, a potentially useful alternative to consid-ering the holistic value of assets is to considerhow changes in air quality affect the value of agri-cultural land, forests, residential property, andhuman capital. Thus, fundamental nonhuman as-sets might include forests, lakes, rivers, estuaries,coastal regions, wetlands, farmland, and residen-tial property. This approach has two furtherattractive features: it allows better integration withexisting accounts, since some of these assets (suchas residential property and forests) have an exten-sive existing database; and it allows incrementaldevelopment of a set of valuations, building uponthose in the market sector.

Practical Choices in Expanding theAccounting Framework

A complete accounting system including interac-tions in the production and asset accounts wouldbe a significant undertaking. Deciding on the scaleof augmented accounting and the next steps to betaken will require considerable strategic thought.One question is whether the accounts will beused for scorekeeping or for management (see thediscussion in Chapter 2.

Scorekeeping, which involves developing a bettermeasure of the performance of the economy overtime, is one perspective. It addresses the questionsof trends in the values of environmental assets andwhether current consumption is sustainable. Ifscorekeeping of this type is the purpose of sup-plemental environmental accounts, it will simplifythe enterprise because there will be no need toconsider intermediate interactions between pro-duction sectors. Tracing where pollutants were

March 2000 • 33

produced and how they affect intermediate prod-uct is unnecessary as long as one can measure finalconsumption and changes in assets. For example,a dying forest is a deteriorating asset; whether thedeterioration is caused by acid precipitation, tro-pospheric ozone, or pest infestation is secondaryfrom a scorekeeping perspective. What is im-portant is to measure the deterioration accurately.Similarly, the overall health and skills of humanpopulations is a central issue in measuring whetherthe economy as currently structured is leading to anincrease or decrease in the stock of human capital.Why the change is occurring—whether because ofchanges in health care or education expendituresor reductions in blood lead—is secondary to themeasurement issue. Overall scorekeeping wouldnote the substantial improvements in the healthstatus of Americans over this century rather thandecreases in particular ailments.

The second broad perspective on the functionof environmental accounts is that of environmen-tal management. This perspective focuses onthe sources, transportation, and ultimate disposalof residual pollutants, particularly their contri-butions to outcomes of economic and ecologicalconsequence. Knowing to what extent partic-ular emissions of residuals come from utilities,automobiles, or volcanic eruptions is critical to de-veloping strategies for control. If human sourcesare dwarfed by natural sources, for example, effortsto control human sources may be futile. Simi-larly, knowing that life expectancies have increaseddramatically is not very useful to understandingwhether there are benefits to tightening controlson small particles or ozone. Improvements inhealth care, occupational safety, and traffic safetymay result in increasing life spans and health statusmore than pollutants are shortening life span—butreducing pollution further could extend lives fur-ther. Thus, if the supplemental accounts are meantto support environmental management decisions,knowing the sources of pollutants and the specificcauses of changes in asset quality are essential.

Analogy with Economic Accounts

The discussion in this section has emphasized thecomplexity involved in constructing environmen-tal accounts. It is useful to compare environmentalwith conventional economic accounting. A lit-tle reflection suggests that economic activity has asimilar, almost fractal complexity when one looksunder the surface. It would be just as difficult tomeasure the physical flows in economic life as inenvironmental life, and indeed many of the same

processes come into play. Consider the problemsinvolved in accounting for a simple loaf of bread.Doing so would require measuring and valuinga wide variety of flows of water, fertilizer, pesti-cides, labor, climate, and capital inputs that go intoproducing the wheat; the fuels, transport vehicles,emissions, weather-related delays, induced con-gestion, or floods involved in transportation; themolds, spores, and miscellaneous rodents and theirdroppings that invade the storage silos; the com-plex combination of human skills, equipment, andstructures that go into milling the wheat; the en-trepreneurship of the baker and the software in thecomputer-operated baking and slicing machinery;the complex chemistry and regulatory environ-ment involved in the wrapping materials; and theevolving ecology of the distribution network. Be-hind each of these elements, in addition, is thecomplex general equilibrium of the marketplace,which determines the selection of production pro-cesses by prices, taxes, and locations, along withthe further complexity of needing to unravel theinput-output structure of the inputs into each ofthe steps just described.

It appears unlikely that anyone would try, andsafe to conclude that no one could succeed in,describing the physical flows involved in this lit-tle loaf of bread. Fortunately, however, economicaccounting does not attempt such a Herculeantask. Rather, the national accounts measure allthese activities by the common measuring rod ofdollars. Although the dollar flows are routinelybroken down into different stages—wheat, trans-portation, milling, baking, and distribution—onecould never hope to describe the flows physicallyand then attach dollar values to each physical stage.Yet this is just what would be required for a fulland detailed set of environmental accounts. Theabove comparison may give some sense of whyaccounting for environmental flows outside themarketplace is such a daunting task.

PHYSICAL DATA REQUIREMENTS:GENERAL ISSUES

Some of the analytical questions involved in envi-ronmental accounting have been analyzed in theprevious section. To construct actual accountsrequires both obtaining accurate physical data(discussed in this section) and valuing the flows(discussed in the next section).

Accurate data on physical flows and stocks are aprerequisite for developing any accounting systemand are the focus of national accounting systemsunder development in several European nations.

34 • March 2000

In some areas, ample physical data are available as aby-product of regulatory monitoring and resourcemanagement systems. Appendix B lists a numberof databases identified by the panel that may be ofuse in further work on supplemental accounts.

Three concerns are fundamental to understand-ing data and measurement requirements for thedevelopment of environmental accounts: (1) thedose-response relationship, (2) measurement ofactual doses experienced, and (3) the fate andtransport of residuals in the environment. Thefirst, the dose-response relationship, is the physicalrelationship between the concentration of or expo-sure to an environmental change and the responseof the subject experiencing the dose. The dose-response relationship is applied to many differentsituations, for example, the response of trees andcrops to chemicals such as carbon dioxide, tropo-spheric ozone, or acid deposition and the responseof humans to pollutants such as lead, particulatematter, or radiation.

Dose-response relationships are often difficult todetermine because they may be affected by com-plex interactions and intervening factors. Forexample, there are extensive medical data on causesof death and, less universally, illness. To deter-mine impacts of environmental changes on humanor natural ecosystems requires separating out thedifferent causes of premature death or illness. Insome areas, such as the impact of tobacco or lead,the relationships are relatively well established; inother areas, such as the impact of particulate mat-ter or ozone, much uncertainty persists. For manyof these relationships, average exposure over theyear is rarely the relevant measure. Damage maybe related to extreme levels or to periods in whichthe subject is particularly sensitive to the agent;acute effects may differ from chronic effects relatedto long-term, low-level exposure.

Resolving these uncertainties about dose-response relationships is important for policydecisions, such as the level at which to set pri-mary air-pollution standards. Resolution of theseuncertainties would also allow construction of en-vironmental accounts. The panel’s review of workin this area indicates that the preparation of esti-mates of the economic impacts of air pollution isfeasible today, but there are enormous uncertain-ties at virtually every stage of the effort. WhileBEA or those preparing environmental accountswould not necessarily be involved in preparingdose-response estimates, the accountants will needto work closely with public-health, agricultural,forestry, and ecological experts to use the bestinformation available.

In addition to understanding the dose-responserelationship, national accounting requires regular,statistically valid monitoring of the relevant pop-ulations and the doses they are receiving. A basiclimitation of much of the data currently collectedis that ambient concentration levels in areas whereindividuals, crops, forests, or other relevant en-tities actually reside are poorly measured. Mostmeasurements occur at sites of convenience ratherthan sites of relevance. Air pollution monitors areoften placed with other monitoring devices whereairplanes congregate rather than where people live.

A full account of economic-environmental inter-actions also requires tracking the fate and transportrelationship, or the connection between the emis-sion of a particular pollutant orpollutant precursorat one time and geographic point and the level,time, and location of the pollutant at the pointwhere it affects an economic asset or activity. Theserelationships are generally highly complex andvariable. For air pollutants, wind direction andspeed, temperature, cloudiness, and precipitationall affect how a pollutant is dispersed or concen-trates. Precursor pollutants sometimes do notcreate damage themselves, but react chemically inthe atmosphere to create other agents that are dam-aging. Acid precipitation and tropospheric ozoneare examples. The formation of these pollutantsdepends on the presence of other agents that maylimit, speed, or slow the process. Monitoring ofemissions, concentrations, exposures, and conse-quences would provide the physical foundation fora complete set of environmental accounts, and isalso a critical part of environmental management.

The goals of environmental accounting will dic-tate the assignment of priorities for improveddata. Extensive data on the fate and transportof emissions and concentrations of pollutants area lower priority if the goal is scorekeeping; evendose-response relationships may be secondary tomore direct measurement of consumption flows orchanges in important capital and environmentalassets and human health status. If one is inter-ested primarily in measuring the sustainability ofeconomic activity, understanding the health statusof human and natural systems is more importantthan understanding why conditions have changed.On the other hand, understanding these technicalrelationships is essential if environmental accountsare to serve as a data set to support environmentalmanagement, in which the goals are to understandthe severity and causes of environmental prob-lems, along with remedies needed to mitigate thoseproblems.

March 2000 • 35

3. See Smith (1993) and Braden and Kolstad (1991) for reviews of thetheory and application of these methods.

VALUATION: GENERAL ISSUES

Once appropriate physical data have been de-veloped, the next step in developing integratedaccounts is to value changes in the physical meas-ures. Physical data alone are often interesting anduseful for policy making, and improvements in,physical environmental data could enhance policy-making efforts. Indeed, most countries have notgone beyond developing physical measures andindicators because of the difficulties involved invaluing nonmarket goods. Without valuation,however, physical data alone have serious limi-tations for both scorekeeping and environmentalmanagement. Aggregate physical measures, suchas areas of agricultural land, forest, or wetlands ortons of sulfur, toxic wastes, or particulate emis-sions, provide incomplete second column evidenceon the effects of these chemicals on economic well-being or economic sustainability over time. Forexample, losing 1000 acres of prime Florida Ever-glades would probably impose a greater economicand ecological loss than losing an equivalent areaof frozen wetlands in northern Alaska. Thus an ac-counting entry of “total wetland acres” lost wouldnot be a useful measure. Furthermore, a sim-ple measure of wetland area would fail to captureimprovements in quality that might occur as a re-sult, for example, of current efforts to restore theEverglades as a fully functioning ecosystem.

For many issues, it is necessary to weight thephysical measures by their importance. There areapproaches to weighting physical quantities otherthan valuing all impacts in dollar terms; for ex-ample, different environmental residuals can beweighted by how they affect human mortality.However, such weights would be incomplete be-cause they would exclude impacts on morbidity oron the health of ecosystems. In economic account-ing, the “importance weights” are the economicvalues, usually market prices. The advantageof using economic valuation is that comparisonscan be made across very different environmentaleffects and with goods that are part of the mar-ket economy. While relying on economic valueshas many desirable features, there are a num-ber of difficulties involved in usefully applyingnonmarket valuation studies and techniques to en-vironmental accounting, as discussed below (seealso Chapter 2).

Valuation Techniques

Markets provide the conventional valuation formarket goods and services. A variety of meth-ods for valuing nonmarket goods and services has

been developed. Table 4–2 indicates the poten-tial and actual uses of various valuation methodsfor many environmental problems, including thedose-response method discussed above. Thesemethods have been developed over a number ofyears and have been applied to many specificproblems.3

The dose-response method, as a valuation methodin and of itself, is directed toward converting expo-sure to a specified dose of a substance, from whichis calculated a physical response for which a di-rect market price can be observed. For example,exposure to ozone or particulate matter results inwheat-yield loss or lost work-days due to respira-tory illness; using the market price of wheat or oflabor, an estimate of economic value can be made.The valuation techniques in this approach are con-sistent with prices used in the economic accounts.Incomparability or additional uncertainties are in-troduced only through imputation of output byuse of the dose-response relationship, which con-verts the environmental effects into market-goodterms.

Travel-cost and hedonic methods also use behav-ior and observed market transactions as a basis forestimating values, but the activities involve timeuse and expenditures on goods and services relatedto use of the environmental or natural-resourcegood, rather than on the resource itself. For ex-ample, a recreational site might be valued usingthe travel-cost method by estimating the time andout-of-pocket costs involved in reaching the site.

Hedonic methods use statistical techniques toexplain variations in market prices based on thebundle of characteristics of a good. This ap-proach is currently used in the national accounts.Computers, for example, are considered bundlesof attributes such as speed, memory, and ran-dom access memory (RAM), and the value of thecomputer is a weighted sum of the values of itsattributes.

For resource and environment valuation pur-poses, hedonic methods are used to explainvariations in land values that reflect natural-resource or environmental characteristics. Suchestimates are based on observed price differencesof land with different amenities or disamenitiessuch as noise, pollution, and crime. Hedonic wagestudies—looking at the wage premiums of high-risk jobs—are currently the standard approach toestimating the value of workplace hazards; the re-sults are often used as estimates of the value of

36 • March 2000

life-threatening effects due to such causes as airpollution or traffic accidents.

Contingent value (CV) methods are survey tech-niques that ask people directly what they wouldpay for goods and services. Applications in thearea of environment and natural resources include,for example, asking individuals what they wouldbe willing to pay to reduce smog, to increase visi-bility in places such as the front range of Colorado,and to clean up an oil spill in a coastal area. CVmethods differ from the other methods discussedabove in that there are no budget constraints or be-havioral observations involved; the results reflectrespondents’ estimates of the value of a hypotheti-cal change, rather than a dollar or time cost actually

TABLE 4–2 Methods for E

Pollution Type of Effect ImpactHedo

Air pollutionConventional pollutants: Respiratory WLD

(total suspensed illness RADparticulate [TSP], Medicalsulfur dioxide [SO2], sufferingnitrous oxides [NOX])

Respiratory Deathillness

Aesthetics Visual,sensory

Recreation Visits, especially toforests

Materials Maintenance/repairVegetation Crop losses

Water pollutionConventional pollutants Recreation Visit behavior

(e.g., biochemical (e.g., fishing,oxygen demand [BOD]) boating)

Commercialfisheries

Stock losses

Aesthetics Turbidity, odor,unsightliness

Ecosystem Habitat andspecies loss

Trace concentrations Drinking Illness,water mortality

Fisheries Stock losses

Toxic substancesAir (benzene, Illness, WLD

polychlorinated mortality RADBiphenyls [PCBs], Medical expenses

pesticides) Pain andsuffering

Chemicals hazardous to land Aesthetics UnsightlinessEcosystem Anxiety, ecosystem

losses

Radiation Illness, WLDmortality RAD

Lives lost

Marine pollutionOil, radioactive

substances,sewageAesthetics Unsightliness

Swimming Visit behaviorIllness

Fish/livestock losses

Noise Nuisance Annoyance

U = Used technique; Poss = Not developed, but possible; X = Inapplicable technique; WLD= Work loss days; L = Very limited applications; RAD = Resource activity days.

incurred. While widely used for environmentalvaluation, CV is highly controversial because it of-ten fails elementary tests of consistency and scalingand is subject to a wide variety of potential responseerrors if not carefully constructed.

The overriding problem with all these methodsis that they require voluminous data and statisti-cal analysis and can hardly be used routinely for alarge number of products in constructing environ-mental accounts. Where existing CV studies areused for environmental or natural-resource val-uation, they often employ valuation approachesthat are inappropriate for national accounts. Forexample, many estimates used in environmentalmanagement rely on average value (including con-

nvironmental Valuation

Techniques for estimation impacts

nic Property HedonicWages Travel Cost Contingent

ValuationDose

Response

L L X U U

L and U U X X U

U L X U X

L X U U X

X X Poss Poss UL X X X U

L X U U X

X X X X U

U X L U X

X X X U U

X X X Poss U

X X X X U

U X U U U

X X X U U

Poss U X L U

U X U U U

U X X U L

Source: Adapted from Organization for Economic Cooperation and Development (1989), asappearing in Costanza (1997).

March 2000 • 37

5. This discussion greatly simplifies the discussion of public goods. Thereare further distinctions among public goods that are central to many issuesinvolved in environmental accounting, particularly as regards valuation meth-ods. One such distinction is whether consumption is excludable; in the caseof global warming, for example, no coastal nation can exclude itself from therising seas. Another distinction is between pure and congestible public goods.Congestible public goods are those whose consumption is neither completelyrival nor nonrival; one person using a beach does not preclude others from do-ing so, but most people find crowded beaches less enjoyable than deserted ones(see Cornes and Sandler, 1986). Crowding of this sort means that even withopen access, the marginal value of use of these sites is greater than zero. A finaldistinction is between those goods whose use affects market activities or marketvalues and those that are completely independent of the market. Public goodswithout traces in markets are frequently referred to as “nonuse values.” Nonusevalues include values people derive from knowing that a species exists, natural

sumer surplus), rather than the prices or marginalvalues that are the convention in national incomeaccounting.4 In a competitive economy, marketprices measure both the incremental value to theeconomy of consuming another unit of the goodand the incremental cost to the economy of pro-ducing that unit. Therefore, prices are a usefulbenchmark for valuation.

In one sense, the market value underestimatesthe total value of goods and services to consumers.Because consumers pay the price of the last ormarginal unit for all units consumed, they enjoya surplus of total satisfaction over total cost. Theterm used for the extra utility consumers receiveover what they pay for a commodity is consumersurplus (see also Chapter 2). Consumer surplusintroduces a complication in comparing marketprices with nonmarket values. For goods withoutmarkets, value is often measured by total willing-ness to pay for the good. Such values are notdirectly comparable to market prices because thevalues include the consumer surplus. In otherwords, when nonmarket goods are valued accord-ing to total willingness to pay, the value of thosegoods is overstated relative to the market value ofmarketed goods. For example, travel costs canprovide the average value of a recreational serv-ice, but the marginal value of the resource for anopen-access beach or forest with no fee may bezero. This discussion illustrates the importance ofensuring comparability in estimating values in theconstruction of nonmarket economic accounts.

Classes of Economic Goods

The valuation of environmental goods and serv-ices raises an issue that is largely overlooked inconventional accounting—the distinction betweenprivate and public goods. These deceptively com-mon terms are used in a specialized sense here (seeSamuelson, 1954, 1955). Private goods are onesthat can be divided up and provided separately todifferent individuals, with no external benefits orcosts to others. An example is bread. Ten loavesof bread can be divided up in many ways amongindividuals, and what one person eats cannot beeaten by others. Public goods, by contrast, are oneswhose benefits are indivisibly spread among theentire community, whether or not individuals de-sire to purchase them. An example is smallpoxeradication. It matters not at all whether one is

4. Marginal costs and marginal values are central concepts in determiningeconomic efficiency. For example, knowing the marginal value of reductionsin atmospheric lead is more useful to the policy maker than knowing theaverage value of all reductions. Marginal cost and marginal value are definedin Appendix D.

old or young, rich or poor, American scientist orAfrican farmer—one will benefit from the eradi-cation whether one wants to or not. The exampleof smallpox eradication is a dramatic case of apublic good. The economy is replete with activ-ities, such as pollution abatement, new scientificknowledge, national defense, and zoning, that havepublic-good characteristics.5

The distinction between public and privategoods is central for many nonmarket and environ-mental commodities. In a perfectly competitivemarket, the price of a marketed private good is themarginal value of consumption to the consumer.Similarly, while observed prices do not exist fornonmarket private goods, the marginal value of theconsumption of such goods is conceptually equiv-alent to a market price. The national accountsvalue food produced and consumed on farms, eventhough it is not marketed, the same way food soldin the marketplace is valued.

Valuation of public goods is an especially difficultproblem because their value to all consumers mustbe reckoned with. For example, improvements inair quality affect everyone. Conceptually, there-fore, one should value public goods by adding upthe marginal values of changes to the entire affectedpopulation. Doing so poses severe measurementdifficulties for two reasons. First, the “personalprices” or marginal values of the public good aresure to vary across people—some may be signif-icantly affected and therefore place a high valueon air quality, while others may be relatively indif-ferent. Second, determining the values of publicgoods is extremely difficult because people makefew decisions that reveal their preferences in this re-gard. People cannot choose how much defense orsmallpox eradication they would like to consume;these decisions are made collectively. Since peo-ple cannot choose different levels of a public good,

wonders remain, or natural systems survive intact beyond any specific use towhich they might be put (see Randall and Stoll, 1983). When Congress createdYellowstone National Park in 1872, for example, no member of Congress hadever been there, and its value as a natural wonderland was largely a “nonusevalue” imagined on the basis of photographs of William Henry Jackson anddrawings of Thomas Moran.

38 • March 2000

TA

Type of goods

Market BrCaReHo

Nonmarket Ho

LeTeGRe

there are no behavioral traces of their preferencesor personal prices.

For the above reasons, constructing environ-mental accounts will necessarily be different forprivate and public goods. For private goods,particularly near-market goods that have close rel-atives in the market economy, valuation appearsfeasible and has a level of reliability that approachesthat of the current national income accounts. Mostpublic goods, by contrast, present greater measure-ment and conceptual problems. Table 4–3 showsexamples of each type of goods that have thesedifferent characteristics.

Strategies for Valuation

Near-market natural-resource and environmentalgoods (which are largely private goods) offer themost promise for valuation and inclusion in theaccounts. Often there are markets for comparablegoods that provide direct evidence of the value ofthe nonmarketed goods or services. This approachis consistent with the use of market prices usedelsewhere in the accounts and has precedent inthe valuation of owner-occupied housing services.Thus, the methods for including these near-marketgoods have already been established. A potentialsource of error in using this approach is that thequality may differ for goods or services produced orprovided in the household and those produced inthe market. It would be appropriate to undertake amodest research program to investigate the adjust-ments necessary to make market and near-marketactivities comparable.

Two basic types of near-market goods are of in-terest. The first is the service flow from a naturalresource. Here, as in the case of timber from forestsor crops from farmland, the service flow is alreadyin the core accounts, and the returns to these assetsappear as profits and/or returns to other assets, butthe accounting is incomplete because it omits thenonmarket activities. The second case is a good

BLE 4–3 Classes of Goods and Services

Private (examples)Public (examples)

Related to Markets Independent of Markets

eadrsstaurant mealsusing rentals

Knowledge and innova-tions that are patentedand copyrighted

Pollutants with tradeablepermits

None

usehold preparedmealsisure timelevision viewing

roundwater for drinkingntal values of owner-used assets

Air and water qualityClimateMosquito control

Passive or nonuse value(e.g., knowledge of theexistence of species,unique national treas-ures such as Yellow-stone National Park)

not currently in the accounts, such as recreationservices enjoyed by households; in this case, thevalue that is attributable to the service is equal tothe value of household labor and capital services,plus a service flow from a natural resource.

Public goods that affect markets offer oppor-tunities for using observations of actual markettransactions to generate valuation estimates. Anexample would be concessionaire activity withina national park. The hedonic property and wagetechniques can be explored as a basis for develop-ing values or imputing how changes in these publicgoods affect markets. There are some potentiallysound ways to make the links between these pub-lic goods and the market explicit in the accounts,but there is not yet a consensus on how to in-clude them, and each provides a challenge for datadevelopment and estimation of values.

Other classes of public goods, particularly thosethat are national or global in nature and do notleave behavioral traces of individual preferences,are currently problematic for the national ac-counts. Most of these public goods, such as thoseinvolving nonuse values of natural-resource andenvironmental assets, can be valued only with CVmethods. Some reviews have conveyed cautiousapproval for use of these methods in limited cir-cumstances. For example, a panel convened bythe National Oceanic and Atmospheric Adminis-tration to review CV methods for use in federalcompensation decisions identified “a number ofstringent guidelines for the conduct of CV studies”that, when followed, allow “CV studies [to] conveyuseful information” (see Arrow et al., 1993:4610).However, the accuracy of the values developed withthese methods remains controversial among thosein the economics profession (see Portney, 1994;Hanemann, 1994; Mitchell and Carson, 1989; andDiamond and Hausman, 1994).

As discussed above, the hypothetical nature ofthe valuation makes these methods quite differ-ent from other methods that are based on actualmarket transactions. For these reasons, while CVis sometimes useful for other purposes, the panelhas determined that it is currently of limited valuefor environmental accounting. This means that,for many important environmental assets, envi-ronmental accounts will omit a portion of thevalue of the assets. That is, it appears to be fea-sible to work toward accounting for goods suchas recreation activities associated with the FloridaEverglades, Yellowstone National Park, and similarsites. However, it is beyond the ability of currenttechniques to provide reliable measures of the valueof the public-goods services provided by these as-

March 2000 • 39

6. The following discussion focuses primarily on issues pertinent to theUnited States. A significant issue in natural-resource accounting for manydeveloping countries is deforestation. For example, a major concern in thenational accounts of developing countries such as Indonesia is that harvestingof forests is contributing to rapid growth in current consumption at the expenseof the stock of forest assets. In the late 1800s, the deforestation rate in theUnited States equaled or exceeded that found in many tropical countries today,but deforestation is no longer significant on a national scale, and the generaltrend since the 1950s has been a net growth in the forest stock of the UnitedStates.

7. Because of the decision not to use markets in allocating such resources,but typically to provide them through collective decisions, common usagesometimes refers to such goods and services as “public goods.” This reportfollows the conventional definitions of public and private goods discussed inthe previous section.

sets, even though we may suspect that these servicesare precious to the nation.

In the remaining sections we explore the issuesraised in the preceding sections in far more detailfor the cases of forests and air quality.

FORESTS: A RENEWABLE NATURALRESOURCE

Forests are a prime example of renewable natural-resource assets. They present many of the samenational economic accounting issues as otherrenewable natural-resource assets, such as agricul-tural land, fisheries, and coastal and freshwaterresources. Many of the products derived fromnatural-resource assets are included in the pro-duction accounts of the existing core NIPA. Butthese assets are not generally included in nationalasset accounts, and the production accounts them-selves exclude any nonmarket goods and servicesderived from these natural-resource assets. Forestsare a useful example because much effort has beendevoted internationally to forest accounting.

While the NIPA as currently structured are notintended to include the full range of forest values,regular reports of economic activity as measured bythe NIPA are widely noted and interpreted as mea-suring important aspects of economic well-being.It is logical to try to capture in these accountsmore of the important relationship between forestsand humans. Forests support human material andspiritual welfare in countless ways. They harbormany important species of plants and animals.They form an aesthetically pleasing backdrop forrecreation and for everyday life. They filter andregulate the flow of much of the U.S. water sup-ply. They have been a reservoir for land availablefor conversion to agriculture and other developedactivities. Wood is one of the world’s most im-portant industrial raw materials and a ubiquitoussource of energy. And worldwide, literally millionsof indigenous people call forests home.

This section examines, in five parts, method-ological and practical issues that arise with regardto including forests in national economic accounts.It begins with a discussion of the nature of theeconomics of forest values, providing a generalframework for assessing those values. The secondsubsection translates this general discussion into amore precise statement of how forest values mightbe incorporated in the U.S. economic accounts.Given this context, the third subsection commentson BEA’s work to date and provides a brief dis-cussion of the extensive international literature onforest accounting. This is followed by discussion

of a recommended approach for measuring the netaccumulation of timber. The section ends with thepanel’s conclusions on forest resources.

The Nature of Forest Values

Forests produce economic value through threeprincipal classes of economic goods: private goodstraded in markets, private goods not traded in mar-kets, and public goods. These goods can affectboth the national asset accounts and the NIPA.6

These three classes of forest goods and services arediscussed in decreasing order of availability of dataand of accepted analysis required to include themin the national economic accounts.

Private, market-related activities. Some forest-based market-related activities are already includedin the national income accounts; examples areall forest products used in manufacturing (log-ging, lumber production, the manufacture ofpaper, wooden furniture, and musical instru-ments). Some fuel wood production would fallinto this category; the part that flows through themarket economy would enter the accounts, whilethe part that is produced for own consumptionwould not.

The major issue in the current treatment of pri-vate, marketed forest-based goods and services isthe failure to account for changes in the value of thestanding timber. Most of the conceptual problemsinvolved in doing so have been fully consideredand developed, as discussed below. Accounting forchanges in the timber inventory would address oneof the major shortcomings of the existing forestaccounts.

Private goods not traded in markets. Forestsproduce many private goods and services that—for reasons of custom, law, or economics—societyhas elected not to allocate through markets.7 Forexample, the water flowing from forested water-sheds has considerable economic value. Indeed,the rationale for forest conservation in the latenineteenth century related primarily to protec-tion of forested upland watersheds. Protection

40 • March 2000

8. The discussion in this section draws heavily on the recent comprehensivetreatment of the subject by Vincent and Hartwick (1997).

of navigation was the explicit constitutional ba-sis for creation of the eastern national forests,and congressional agricultural interests concernedabout irrigation provided the principal support forwithdrawing the national forests from the west-ern public-domain lands. A study by Bowes et al.(1984) of the Front Range of the Rockies aroundDenver and informal estimates for the QuabbinWatershed servicing Boston demonstrate that insome locations, the value of the water producedfrom a forest may far exceed the value of thetimber production. Changes in forest attributescan affect stream flow and therefore the valueof water “produced.” Interestingly, Bowes et al.(1984) demonstrate that when water is valuable,it is optimal to keep timber stocks low to reduceevapotranspiration and therefore increase runoff.

Public goods. Public goods are ones for whichconsumption by one individual does not reduce theamount available for others to consume. Forestsproduce many public goods, including aestheti-cally pleasing landscapes, a carbon sink, and astore of biological diversity. Given data on changesin forest inventories, it may be possible to valuesome of these services (e.g., the value of carbonsequestration), although the uncertainties of suchvaluation should not be underestimated. In othercases, the valuation problems go far beyond theresults of current research.

The interactions among these three sourcesof forest value—private marketed goods, privatenonmarketed goods, and public goods—can becomplex. For example, cutting trees leads to in-creases in manufacturing activity. This in turnmight cause an increase in water yields and therebyreduce the costs of industrial and household pro-duction. It might also cause a shift of speciesdiversity away from late-seral-stage organisms,such as spotted owls, and toward early-seral-stageones, such as elk. It would lead to an imme-diate release of carbon associated with loggingand forest products manufacturing, but mightresult in a long-term increase in carbon seques-tration with forest growth if the wood productswere sequestered in long-lived furniture or houses.Given the site-specific nature of such productionrelationships and the lack of current scientific un-derstanding of many of the underlying ecologicalprocesses, there is currently an insufficient scien-tific basis for specifying a full set of such linkagesin supplemental accounts.

Incorporation of Forest Values in theNational Economic Accounts8

To be most useful, the economic accounts wouldidentify the separable contributions of forests tothe national economy. It is convenient to discussthe problems involved in incorporating forest val-ues in the U.S. national economic accounts firstfor the production accounts and then for the assetaccounts.

Adjustments to Production Accounts

A full treatment of forests in the production ac-counts would involve the following adjustments tonational income and product.

Timber income. Sales of timber are already in-cluded, although some are recorded as part ofpersonal income, some as part of manufacturingincome, and some as part of government receipts.The principal difficulty is ascribing these incomestreams to the forest sector; in this respect, theissues are very similar to those encountered inthe treatment of mineral incomes discussed inChapter 3. Ordinary production costs associatedwith forest production activities are similarly cov-ered by the current NIPA, but may not be easilyassociated with the forests themselves, rather thanforest-products manufacturing. Problems remainwith the allocation of joint costs. For example,forest roads are a costly input to the productionof many forest products, including timber, mi-nor forest products, and recreation. Yet standardaccounting practices, especially for the nationalforests, attribute the full cost of these roads to thetimber program. As currently constructed, theNIPA include the costs of road construction, butexclude the benefits produced by the road.

Near-market forest products. To the extent thatnear-market forest products, such as fuel wood,berries, mushrooms, and Christmas trees, are pro-duced by households but not purchased throughmarkets, they would be included in the forestaccounts.

Contributions to household production (e.g.,recreation). The accounts would include the valueof household production of activities such as hik-ing, hunting, and fishing. However, if thereis uncongested, open access to the forest-basedinputs needed for household production, the con-tribution of these inputs to household value on themargin is zero. Current practice often uses averagerather than marginal values, so care must be taken,

March 2000 • 41

9. USDA Forest Service (1995) also present estimates based on fees collected(which show much lower value overall and relatively less for recreation andwildlife); willingness to pay, including consumer surplus (which show higheroverall values and greater importance for recreation and wildlife); and incomegenerated, including that generated by downstream activities such as lodgingand equipment rentals related to forestland recreation (which show the highestoverall value). From the perspective of comparability with the current nationaleconomic accounts, the methods associated with the discussion in the text arepreferable to the other three methods.

particularly for open-access forests, to ensure con-sistent valuation in order to prevent overvaluationof nonmarket activities.

Environmental services used by other industries(e.g., watershed protection, domestic/industrial wa-ter supply). Some of the impacts of forests arealready included in the NIPA. For example, ifforests moderate water flows and reduce the costof agricultural production, this benefit is fully in-corporated in the NIPA. Ascribing the benefit tothe forest sector, while a difficult task, would berequired for a full accounting.

Public goods (e.g., carbon sequestration, bio-diversity, species preservation). At present, theonly public goods that have been the subject ofwidespread attempts at valuation are those asso-ciated with carbon sequestration (Brown, 1996).While quantitative data on carbon sequestrationare available, valuation is still highly uncertain.Moreover, because valuation of carbon sequestra-tion is based on global benefits, the issue of howsuch benefits would be incorporated in a singlenation’s accounts is unresolved.

There are few comprehensive studies of the totalvalue of forest products. Recent work on goods andservices produced on public lands managed by theU.S. Forest Service indicates that more forestlandvalue is due to recreational and wildlife servicesthan to timber, mineral, and range goods (U.S.Department of Agriculture Forest Service, 1995).For example, of the estimated total $9 billion valueof forest goods and services in 1993 (valued atmarket prices), recreational and wildlife servicesaccounted for 80 percent, whereas the productionof minerals and timber and grazing range servicesaccounted for just 20 percent.

While the above estimates illustrate the impor-tance of nonmarket production, they should beinterpreted with caution. First, they include onlyland managed by the U.S. Forest Service, which isnot representative of all forestland. By contrast,on private lands that are intensively managed fortimber production, much of the value is due totimber harvesting. Second, these estimates do notinclude all nonmarket values; for example, theyomit the potential value of carbon sequestration.A recent estimate is that U.S. forests sequestered211 million metric tons of carbon in 1992 (Birdseyand Heath, 1995). At $10 per ton, a value consis-tent with the Intergovernmental Panel on ClimateChange (IPCC) estimates of the marginal value ofemission reductions (see Bruce et al., 1996), theannual value of carbon sequestration in all U.S.forests would be $2.1 billion; the numbers couldbe an order of magnitude larger if the U.S. adopted

stringent emission controls under the Kyoto Pro-tocol of 1997. Third, the Forest Service presentsdifferent types of estimates for the value of forestservices, market-clearing prices being only one ofthese.9

Forests Asset Accounting

A key conceptual problem with the present NIPAis the lack of any accounting for changes in assetvalues of U.S. forests. Accomplishing this task waspart of the Phase II work outlined by BEA (seeChapter 2). We address this issue in some detailfor two reasons. First, from a conceptual stand-point, natural-resource assets should be treatedconsistently with produced capital assets, addingnet accumulation or subtracting net decumulationfrom gross domestic product (GDP) to arrive ata measure of net national product (NNP) moreclosely associated with a sustainable-income con-cept. Second, the capacity exists to rectify thisomission with respect to the value of forests that islinked to marketed production.

While adjustments in an asset account are con-ceptually similar to net investment of “madeassets,” for forests it is more precise to call thechange in asset values net accumulation to reflectthe fact that, even at constant prices, the assetvalue of a forest can either increase or decrease.Most generally, net accumulation is defined as thechange in an asset value from one period to thenext. Because asset values cannot generally beinferred, economists infer the value of the assetfrom assumptions about timber markets. A fullanalysis of this issue is presented in Appendix C.Three major alternative approaches to accountingfor changes in asset values of forests are describedbelow.

Hotelling model. The first approach is analogousto the literature on nonrenewable resources dis-cussed in Chapter 3. In a sense, this approach treatsthe exploitation of primary, old-growth forests astimber mining. Since it is generally uneconomic toreplace primary forests with forests of a similarlyold age, this analogy is not as odd as it might ap-pear. Under these circumstances, the change in thevalue is the volume of the harvest times the differ-ence between the price and the marginal extractioncost. This model of net accumulation is called the

42 • March 2000

Hotelling model to emphasize the connection be-tween mining old growth that will not be replacedand mining minerals that cannot be replaced.

Based on historical studies, this approach ap-pears to be a reasonable approximation of em-pirical trends in forest development (see Berck,1979; Lyon, 1981; Sedjo and Lyon, 1990; and Sedjo,1990). In the early stages of development, netgrowth of the forest is nil: photosynthesis just bal-ances the death of plant tissues and entire trees.Because growth is nil, any harvest at all exceeds thegrowth of the forest. Since the harvest is greaterthan the growth, the timber inventory declines. Asthe inventory of old-growth timber declines, tim-ber becomes more scarce, and timber prices rise.In addition, harvesting costs increase as loggingextends into increasingly remote sites. Prices riseuntil the purposeful husbandry of second-growthtimber and the use of nonwood substitutes (stone,concrete, and steel for construction; fossil fuels, so-lar energy, and conservation for energy) becomeseconomic. This analysis is broadly consistent withthe development of the forest sector in the UnitedStates. Harvest exceeded growth until the 1950s.Timber prices rose at a real rate of about 4.6 percentper year between 1910 and World War II and 3.1percent per year from that period to the mid–1980s(Clawson, 1979; Sedjo, 1990; and Binkley andVincent, 1988).

Transition models. While the Hotelling modelmay be appropriate for the case of pure depre-ciation under the assumption of perfect capitalmarkets,10 it misses several important aspects ofthe forest sector, including (1) “discovery” of newold-growth forest stocks (e.g., the rapid expansionof logging in the British Columbia interior to serveU.S. markets once U.S. prices had risen to the pointthat accessing this comparatively remote regionbecame economic), and (2) the fact that the old-growth forests were replaced with faster-growingsecond-growth forests. Both effects attenuate priceincreases, causing the ordinary Hotelling modelto overstate forest depreciation. These effects arethe forest analog of mineral deposits analyzed inChapter 3.

Transition models account in part for theseproblems by recognizing that forest growth offsetsharvests. Assuming constant prices and a forestinventory recognized only by total net growth, thismodel suggests net accumulation is given by the

10. The Hotelling model assumes perfect capital markets in which the rate ofreturn in the mining or old-forest sector equals the rate of return in alternativeeconomic activities. In countries, especially developing countries, where bothforest and mining activities earn disproportionally high returns because ofspecial favors and licenses, the Hotelling model is not appropriate. It greatlyoverstates the true decline in the value of these stocks as they are mined.

difference between price and marginal harvestingcost times growth minus harvesting (rather thansimply minus harvesting in the Hotelling model).By recognizing forest growth, such a formulationimproves on the ordinary Hotelling approach, butstill suffers the defects of (1) ignoring endogenousprice changes in the sector, and (2) characteriz-ing the forest only by net growth and not its morecomplex underlying age-class structure.

Managed second-growth forests. Economic the-ory suggests that, once the transition betweenold- and second-growth forests is complete, tim-ber prices will stabilize, and the economic returnto holding forests will arise solely from forestgrowth. Vincent (1997) has analyzed this case anddeveloped the appropriate measures of net accu-mulation for optimally managed second-growthforests. The appropriate estimate of the value ofasset accumulation is more complicated here (seeAppendix C for a full discussion). Accumula-tion depends on the forest age structure, discountrate, timber-yield function, and economically op-timal rotation age. While this approach improveson both the Hotelling and transition approaches,certain shortcomings remain. In particular, thisapproach assumes that forest owners cut their treesat the economically optimal time and that timberprices grow at a constant rate. This theory of for-est valuation can be used to formulate a practicalapproach to measuring the economic depreciationof forests. Before turning to that recommendedapproach, it is useful to examine BEA’s work onforests and the international literature in this field.

BEA’s Approach and InternationalComparisons

As noted, forests are part of Phase II of BEA’s IEESAeffort. As a consequence, BEA’s work on foreststo date has not been extensive and may need re-finement (see Howell, 1996). In its current work,BEA separates forestland from the timber inven-tory. “Forests and other wooded land” are valuedat the average value of agricultural land. In general,edaphic and geomorphologic factors make forest-land less valuable than agricultural lands, and therate of change in forestland prices is uncorrelatedwith the rate of change in farmland prices (seeWashburn, 1990). BEA updated their estimatesof the timber inventory each period using sepa-rate Forest Service estimates in physical terms ofgrowth and removals. Starting with physical in-ventory estimates, BEA added physical estimatesof growth (additions) and removals (depletion) toderive closing stocks. Each year’s closing stock es-

March 2000 • 43

timate became the following year’s opening stocks(except in the Forest Service inventory years, wheninventory estimates of standing timber were used).Opening and closing stocks, additions, and deple-tions were then valued at the stumpage prices; thedifference between the opening stocks plus addi-tions less depletion and closing stocks, in monetaryterms, was placed in revaluations.

BEA uses the Hotelling model to value the tim-ber stock in each period. Timber is valued at thenational average stumpage rate, with species di-vided into two categories, softwood and hardwood.When measured at a national level, marginal ex-traction costs are probably nonzero (productionincreases are accomplished by turning to increas-ingly costly regions). There is some evidence thatextraction costs are constant within regions, how-ever (Adams, 1997). One conceptual flaw in BEA’scurrent approach is that it measures the deprecia-tion of recreational land on the basis of the costsof repair and maintenance of federal governmentexpenditures for parks. The panel has noted innumerous places the flaw in this approach. Hav-ing accounted for one of the costs of providingrecreational services, BEA does not adjust nationalincome to reflect the benefits. BEA recognizes the

TABLE 4–4 Summary of Forest Accounting Studies

Study Area Reference

Valuation Method

NetPrice

ElSerafy NPV Other

Global World Bank (1997) T UAsia Vincent and Castaneda

(1996)G

Australia I Young (1993) UAustralia II Skinner (1995), Joisce

(1996)H U U

Austria Sekot et al. (1996) H U UCanada I Anielski (1992a, 1992b,

1994, 1996)T

Canada lI Statistics Canada (1997),Baumgarten (1996)

H U

Chile Claude and Pizarro (n.d.) ? ? ? ?China Li (1993) TCosta Rica I Repetto et al. (1991) ? ? ? ?Costa Rica II Aguirre (1996) TEcuador Kellenberg (1995) T UFinland I Koltolla and Mukkonen

(1996)T

Finland II Hoffren (1996) TIndonesia Repetto et al. (1989) TMalaysia I Vincent et al. (1993) TMalaysia II Vincent (1997), Vincent et

al. (1997)G

Mexico van Tongeren et al. (1993) T UNepal Katila (1995) TNew Guinea Bartelmus et al. (1992,

1993), Bartelmus (1994)X X X X

New Zealand Bigsby (1995) HPhilippines I IRG et al. (1991, 1992) T UPhilippines II Cruz and Repetto (1992) TSweden I Hulkrantz (1992) TSweden II Eliasson (1996) TTanzania Peskin (1989a) X X XThailand Sadoff (1993, 1995) T UUnited States Howell (1996) HZimbabwe Crowards (1996) T

Key: H = Hotelling approach; T = transition approach; G = generalized El Serafy approach(elasticity of marginal cost not infinity); X = no timber valuation performed; ? = no information;U = used technique; NPV = net present value.

Source: Vincent and Hartwick (1997). References in original.

criticisms of this approach and plans to use otherapproaches in the future. BEA publishes a full ac-count for 1987, although it produces data on thevalue of timber stocks for 1952–1992. Using BEA’sdata, the net accumulation of timber in 1987 was$2.1 billion at 1987 prices and $47.0 billion if pricechanges are included.

While BEA’s methods can and should be refinedas the environmental accounts are developed, theyare consistent with current international practice.Table 4–4 provides a summary of 29 studies fromaround the world that have attempted to extend thetreatment of forests in national income and prod-uct accounts. Most of these efforts use variantsof the so-called “net price” approach (see equa-tions C.3 and C.4 in Appendix C). Many fail todistinguish marginal and average extraction costs.Accounting for net timber accumulation is well es-tablished in the international literature. None ofthe studies appears to use the third method de-scribed in the previous subsection of a managedsecond-growth forest.

A Recommended Approach for MeasuringNet Accumulation of Timber

The three alternative approaches to accounting forchanges in asset values of forests discussed aboveincorporate many restrictive assumptions. Thepanel investigated other alternatives and identifiedone (developed by Vincent [1997]) that is similarto the second-growth forests approach, but allowsfor the possibility that forest managers may devi-ate from ideal wealth-maximizing behavior. Thisapproach is described in detail in Appendix C. Areview of available data indicates that the approachcan be readily implemented for the United Statesusing data maintained by the U.S. Forest Service.

Conclusions on Forest Resources

BEA has initiated a useful effort to recognize theeconomic contributions of forests in the NIPA.Doing so is consistent with a wide internationalinterest in such accounts. The data and methodsemployed by BEA to date are reasonably consis-tent with the body of international work in thisarea. At the same time, data are available for U.S.forestlands that can enable much more completeestimates of net timber accumulation than eitherthose developed to date by BEA or those availablein the literature for other countries. BEA couldfruitfully work with the U.S. Forest Service in devel-oping annual estimates of net timber accumulationusing these data.

44 • March 2000

This work could also be related to other im-portant values of the forest, particularly recreationand other nonmarket activities. While the dataand analytical methods are not yet adequate toprovide precise estimates of the value of all forest-sector flows to the economy, nonmarket forestvalues for the nation as a whole appear to ex-ceed the value of timber by a substantial amount.Many of these forest values (such as recreationor self-produced fuel wood) are best understoodconceptually in the context of household produc-tion. The household combines specific aspectsof the forest resource with household capital andlabor to produce valuable nonmarket goods andservices. Viewed in this context, forests presentmany of the same challenges for national account-ing as do such important products and servicesas home-cooked meals and in-home education orchildcare. It is therefore logical for BEA to considerthese aspects of environmental accounting as partof the larger problem of valuing the contributionsof nonmarket activity to economic well-being.

In conclusion, constructing a set of forestaccounts is a natural next step in developing inte-grated economic and environmental accounts. Atthe same time, it must be recognized that thereare many thorny problems involved in forest ac-counting. Given the available data and methods,the panel concludes that this accounting is a usefulnext step in developing the IEESA.

AIR QUALITY:A PUBLIC ENVIRONMENTAL GOOD

Air quality is one of the most important exam-ples of a public environmental good and thusshould be among the top priorities for inclusionin environmental accounts. It also presents issuesfor environmental accounting similar to those en-countered with other environmental assets, suchas water quality and climate change. Severely de-graded air quality in many cities of the UnitedStates in the 1960s generated a number of fed-eral regulations during the early 1970s designedto reduce emissions of pollutants that contributedto this degradation. Air quality has many di-mensions, and early regulations focused on someof the more obvious and easily addressed prob-lems. As scientific research further illuminated theless immediately obvious impacts of degraded airquality, such as chronic effects on health, these ear-lier controls were tightened, and new regulationsaddressed a wider range of pollutants.

The first subsection below examines the variousmarket and nonmarket impacts of air quality. The

second reviews some major pollutants that result indegradation of air quality and their primary phys-ical effects. This is followed by review of a recentattempt to estimate comprehensively the benefitsassociated with improvements in air quality. Thefourth subsection addresses the relevance of thesedamage estimates to environmental accounting.The section ends with the panel’s conclusions onaccounting for air quality.

Air Quality Impacts on Market andNonmarket Activities

Degraded air quality can have a harmful effect onboth market activities (e.g., reduced crop yieldsor lost work-days) and nonmarket activities (e.g.,losses due to illness beyond those related to paidlabor, such as those to retired persons, and reducedamenities in recreational facilities). These air qual-ity effects belong in the production accounts ofenvironmental accounts. Moreover, degraded airquality can affect the value of natural-resource as-sets (e.g., acid deposition damage to forests), cancause deterioration of physical capital (e.g., dam-age to the exterior of buildings), and has long-termhealth impacts that affect human capital (e.g., pre-mature death and effects of lead on measured IQof children). Such effects might be included in theasset component of environmental accounts. Withassets as with production, there are both marketand nonmarket effects: market impacts includecapital asset deterioration and forest timber loss,while nonmarket impacts include lost value due todamaged landmarks or degradation of forests forrecreational purposes.

Major Air Pollutants and Their Health andEcological Effects

Table 4–5 lists some important health and ecologi-cal effects of exposure to six air pollutants for whichthe U.S. Environmental Protection Agency (EPA)has established National Air Quality Standards—carbon monoxide,ground-level ozone, lead, nitro-gen dioxide, particulate matter, and sulfur dioxide.These chemicals are sometimes referred to as “cri-teria pollutants.” In addition, there are many otherconstituents of the atmosphere that may have im-pacts of economic consequence. Table 4–6 listssome other components of air pollutants, includ-ing air toxins (e.g., benzene), stratospheric ozonedepletors (e.g., CFCs), and greenhouse gases (e.g.,carbon dioxide and methane). As indicated, EPAhas identified 188 air toxins alone.

March 2000 • 45

Exposure to air pollution has a wide range of im-pacts, including respiratory illnesses (which resultfrom ground-level ozone, sulfur dioxide, nitrogendioxide, particulate matter, and air toxins); childIQ loss, infant mortality, strokes, and heart attacks(which result from lead); skin cancer (which is theindirect consequence of stratospheric ozone de-pletors); and increased mortality (resulting fromparticulate matter, lead, and air toxins) (see Pearceet al., 1996). Ecological effects include impactson agricultural, forest, and aquatic ecosystems.Airborne chemicals have both positive and nega-tive effects on production of marketed goods andservices. Ground-level ozone harms crops, whilenitrogen deposition and carbon dioxide enhanceplant and timber growth. Ground-level ozoneand sulfur dioxide reduce crop yields and tim-ber growth, while air toxins and sulfur dioxidereduce freshwater fish yields. In other cases, at-mospheric trace gases have subtle effects that willoccur far in the future affecting biological diver-sity (for greenhouse gases) or ocean food webstresses, and ultimately causing severe sight dam-age for many mammals (for stratospheric ozonedepletors).

Table 4–5 also shows the change in emissions andsampled concentrations of EPA’s six criteria pollu-

TABLE 4–5 Environmental Protection Agency’s Six CriteriaAir Pollutants

Pollutant Trends(1986–1995) Major Effects Leading Source

Ground-level ozone (O 3)Concentration –6%Emissions –9%

Respiratory illness/lungdamage

Crop/forest damageBuilding/material damageVisibility problems

Transportation* (37%)Solvent utilization (28%)

Carbon monoxide (CO)Concentration –37%Emissions –16%

Reduced oxygenation ofblood

Heart damage

Transportation (81%)

Sulfur dioxide (SO 2)Concentration –37%Emissions –18%

Respiratory illnessBuilding/material damage

(acid rain)Crop/forest damageVisibility problems

Electric utilities (66%)

Nitrogen dioxide (NO 2)Concentration –14%Emissions –3%

Respiratory illness/lungdamage

Building/material damage(acid rain)

Crop/forest damageVisibility problems

Transportation (49%)Electric utilities (29%)

Lead (Pb)Concentration –78%Emissions –32%

Infant mortalityReduced birth weightChildhood IQ lossHypertensionHeart attacks

Metals processing (smelt-ers, battery plants)(39%)

Transportation (31%)

Particulate matter (PM-10)Concentration –22%Emissions –17%

Lung diseaseMortality

Fugitive dust (68%)Agriculture and forestry

(20%)

* Based on volatile organic compounds (VOC) emissions.Source: U.S. Environmental Protection Agency (1996).

tants from 1986 to 1995.11 Primarily as a result ofthe Clean Air Act and the Clean Air Act Amend-ments, emissions of the six primary pollutants havedecreased substantially. For example, installingscrubbers and switching to low-sulfur coal causeda 19 percent decline in emissions from coal utilityplants, which in turn resulted in an overall 18 per-cent decline in sulfur dioxide emissions from 1986to 1995. A 16 percent decline in carbon monoxideemissions during the same period resulted primar-ily from a 20 percent decline in carbon monoxideemissions of on-road motor vehicles. Similarly, a32 percent decline in lead emissions was primarilya result of the ban on leaded gasoline.

Declines in nitrogen dioxide (14 percent) andground-level ozone emissions (6 percent) were lessdramatic, but are expected to become more pro-nounced as the Clean Air Act Amendments of 1990become effective. For example, reformulated fuelrequirements (for oxygen and volatility) for on-road vehicles are likely to reduce carbon monoxideand ground-level ozone emissions. Similarly, theAcid Rain Program (Title IV) requires a 40 percentreduction in sulfur dioxide and a 10 percent re-duction in nitrogen dioxide emissions from 1980to 2010. Particulate matter may be more diffi-cult to control given that almost 70 percent ofanthropogenic-related emissions result from fugi-tive dust (e.g., unpaved roads), with an additional20 percent coming from agriculture and forestry.

The declines in emissions are, of course, linked tolower concentrations of the six primary pollutants.Whereas emissions are estimated on the basis of

11. Data prior to 1986 exist, but cannot be directly compared with datacollected from 1986 on because of changes in data collection (see U.S.Environmental Protection Agency, 1996, for more details).

TABLE 4–6 Other Pollutants of Air Quality Identified byEnvironmental Protection Agency

Pollutant Major Effects Leading Source

Air toxins (188 in total,e.g., dioxins, benzene,arsenic, beryllium,mercury, vinyl chloride)

Thought to cause canceror other serious healtheffects, such as birthdefects or reproductiveeffects

Ecosystem damage (par-ticularly freshwaterfish)

Transportation, woodcombustion, chemicalplants, oil refineries,aerospace, manufac-tures, dry cleaners

Stratospheric ozonedepleters (e.g.,chlorofluorocarbons[CFCs], halons, carbontetrachloride, methylchloroform)

Skin cancerCataractsSuppression of the im-

mune systemOcean food chain

stresses

Fossil fuel, industrialcleaners

Greenhouse gases (e.g.,carbon dioxide,methane, halogenatedfluorocarbons [HFCs])

Broad-scale changes intemperature and pre-cipitation affecting agri-culture, health, waterresources, recreation,ecosystems

Sea level rise

Fossil fuel, combustion,landfills

Source: U.S. Environmental Protection Agency (1996).

46 • March 2000

industrial activity, technology, fuel consumption,and vehicle miles traveled, concentrations of pol-lutants are measured at selected monitoring sitesacross the country. Based on these measurements,estimated airborne concentrations of lead havefallen by 78 percent since 1986, while concentra-tions of airborne carbon monoxide, sulfur dioxide,and particulate matter have fallen by 37, 37, and 22percent, respectively. Smaller declines occurred forground-level ozone and nitrogen dioxide (6 and 14percent, respectively).

Data on other air chemicals vary widely. Excel-lent data are available on emissions and concentra-tions of many of the greenhouse gases (particularlycarbon dioxide) and stratospheric ozone destroy-ers. EPA presently monitors national ambientconcentrations for few of the 188 air toxins identi-fied in the Clean Air Act Amendments. Rather, theagency sets technology-based performance stan-dards to control emissions of these substances. Asa result, EPA has only begun developing a NationalToxins Inventory.

Monetized Benefits of Clean Air Regulations

Although a great deal of work has been done onvaluing components of air quality, there is cur-rently no comprehensive measure of the economicimpacts of air pollution for the United States.However, a recent EPA study evaluating the eco-nomic costs and benefits of clean air regulationsprovides a useful benchmark that sheds light onthis issue (U.S. Environmental Protection Agency,1997). The estimates given are subject to manyuncertainties due to the difficulty of estimatingexposure and the incidence of effects related to ex-posure and valuing the effects. In addition, dataon air toxins have only recently become available,making it difficult to develop comparable estimatesfor these pollutants. The EPA study includes nophysical or monetary assessments of the impacts ofchanges in air quality on ecosystem health, physi-cal capital, or global public goods, such as slowingclimate change and preventing ozone depletion.Moreover, many of the estimates of benefits, par-ticularly those involving the valuation of healthbenefits and the discount rate, have been the sub-ject of major criticism (see Clean Air Act Councilon Compliance, 1997).

Notwithstanding these limitations, the EPAstudy provides an indication of the overall eco-nomic importance of changes in air quality, as wellas a sense of the relative importance of the var-ious air pollutants and the impacts on differentsectors. The study estimates the economic benefit

of actual air pollution relative to a counterfactualbaseline that assumes no controls imposed after1970; roughly speaking, the counterfactual is foremissions to grow with the economy, rather thandeclining as described above. The major resultpresented is that the economic benefits of reducedair pollution in 1990 are estimated to be worth$1,248 billion. Reduced mortality benefits ($1,004billion) account for 80 percent of this total; to-gether, avoided human health effects account for99 percent of the total. In addition, benefits ofimproved visibility are estimated at $3.4 billion,those of reduced household soiling at $4.0 billion,and those of increased agricultural income fromreduced yield losses due to ozone at about $1.0billion. With regard to specific pollutants, mostof the benefits are attributed to reductions in par-ticulate matter (PM–10) and lead; the benefits ofozone reduction are estimated to be only on theorder of $2 billion.

Caution is warranted in drawing too many con-clusions from these estimates and comparisons.Certain assumptions might have had the effect ofexaggerating the economic benefits, and there aremajor uncertainties about the health impacts, par-ticularly because of weaknesses in human exposuredata. Moreover, the study omits some of the ma-jor effects of acid deposition on forests, lakes, andbuildings, and the impact of tropospheric ozoneon ecosystems is not valued. The figures presentedshould therefore be viewed as order-of-magnitudeestimates. Even with all these qualifications, how-ever, it appears that the economic impacts of airquality on human health are highly significant.

Air Quality Benefits and SupplementalAccounts

The estimates of the benefits of pollution con-trol just discussed reflect the value of changes inthe level of air pollutants resulting from proposedregulations. They are relevant for regulatory orcost-benefit purposes, but they are not the appro-priate values for economic accounts. Productionaccounts should measure the damages associatedwith remaining levels of pollution, in terms of bothproduction accounts and change in asset values.This difference between abatement and residualdamage can be quantitatively large. For example,ozone concentrations fell only 6 percent between1986 and 1995. As a result, regardless of the bene-fits of preventing higher levels of ozone than thoseof 1986, the value of changes in ozone concentra-tions over this period would be relatively small. Incontrast, lead and PM–10 concentrations fell 78

March 2000 • 47

and 22 percent, respectively, over the same period,and consequently the damages from these chemi-cals would be much smaller in 1995 than in 1986.In other words, whereas comprehensive consump-tion would have a substantial negative entry due tolead and PM–10 in 1986, the negative values wouldbe of much smaller magnitude in 1995. The resultmight be a substantial increase in the estimate ofgrowth of comprehensive consumption over thisperiod.

As discussed earlier, air pollution affects pro-duction activities, assets, and nonmarket activities.Most of the estimates from the EPA study referto the production accounts: days of work lost,shortness of breath and acute bronchitis, loss ofvisibility, and crop losses are effects on produc-tion activities. Crop losses and the output lossesfrom lost work-days are already included implic-itly in the accounts because these relate to marketactivities. Supplemental accounts that would iden-tify these losses separately would serve to connectthem specifically to air pollution. The estimates forshortness of breath and acute bronchitis includeboth damages that may already be reflected in theproduction accounts (i.e., reduced worker produc-tivity while on the job) and damages that wouldbe reflected only if the accounts were expandedto include household production (e.g., impacts ontennis and jogging). Many of the effects not es-timated by EPA, such as those of acid depositionon forest health, freshwater quality, or ecosystemfunction, would also include effects on both mar-ket activities already in the accounts, such as timberor commercial fishing, and nonmarket goods, suchas recreation.

Asset effects present greater complexity, as wasseen above for the case of forests. Some impacts,such as those on soil or fish farms, would be re-flected in the market value of these assets. Others,such as mortality and chronic bronchitis, are long-term effects on human resources. These effectswould require adjustments in the asset accounts ifa full set of asset accounts for human health andcapital were constructed.

One particular concern arises if the accounts areto include the impact of air pollution on humanhealth. The impact of air pollution and other en-vironmental activities on human health is oftentaken out of the context of other health-related ac-tivities. If one were to track environmental trendsalone, it might be concluded that until the 1970s,growing environmental problems were leading toa deterioration in the health status of Americans.This conclusion is, in fact, incorrect. Activitiesoutside the environmental arena—including im-

proved sanitation, vaccinations, and public-healthmeasures—led to improved life expectancy overthe first seven decades of this century. It wouldtherefore be misleading to enter only a large healthnegative into a set of augmented income accounts.The positives and negatives in the environmentalentry in a set of health accounts would have to beplaced in the context of the vast changes in healthstatus of the American population.

Conclusions on Air Quality

The basic finding emerging from the above dis-cussion is that air quality is likely to be a majornonmarket effect. While EPA’s estimates of benefitsof $1.2 trillion per year due to reduced air pollu-tion are highly uncertain, do not include all effects,and measure a somewhat different concept thanwould be appropriate for the accounts, it is likelythat a realistic assessment of reduced damages dueto improved air quality would yield a much largerfigure than the $27.1 billion in air pollution controlexpenditures used by BEA as a placeholder. In thepanel’s view, no other area of natural-resource andenvironmental accounting would have as great animpact as the potential correction from air qual-ity. The magnitude of this impact indicates thatthe development of supplemental accounts for airquality is a high priority. Indeed, the overall reviewof augmented accounting in Chapter 2 reveals onlya few areas close in importance, such as the value ofleisure, health status, and nonmarket educationalinvestments.

At the same time, air quality is a most elusiveconcept since it has so many different compo-nents. To include these effects in the accounts,several data and measurement obstacles must beovercome. First, determination of the physicalimpacts of changes in air quality, generally esti-mated through dose-response functions, should befocused on the effects of actual human exposureto air pollution. Second, the damage estimatesmust separate the market effects of changes in airquality that are currently captured in the accounts(lost productivity) from the nonmarket effects thatare not currently captured (lost leisure activities).Third, there is a need for reliable and objectivephysical and monetary damage estimates associ-ated with exposure to air pollutants, includingair toxins, ozone depletors, and greenhouse gases.Fourth, significant data gaps with respect to theimpacts of air pollution and changes in air qualityon ecosystem health must be filled. And finally,the estimates must represent year-to-year changes,

48 • March 2000

rather than changes from a hypothetical level ofpollution without regulations.

Developing a set of accounts in this area, alongwith the associated physical measures and val-uations to apply to those measures, is a majorlong-run task for the nation. This task far tran-scends the scope and budget of BEA, and much ofthe necessary work lies outside BEA’s specializedexpertise. The task for the short run, therefore, isto continue basic research on the underlying sci-ence and economics of estimating the benefits ofpublic goods such as clean air. Many years of con-certed research are likely to be required before thematerials for a set of augmented accounts in thisarea are available. But the payoff from the researchwould be large, not only in producing the raw ma-terials for improved environmental accounts, butmore important in providing the data and analysisneeded for improved public policy concerning theenvironment. In short, the task of constructing en-vironmental accounts for important public goodsshould be part of a more general goal of improvingthe nation’s information and analytical systems inthis area.

CONCLUSIONS ANDRECOMMENDATIONS ON RENEWABLE

AND ENVIRONMENTAL RESOURCES

General Approach

4.1 The panel recommends that BEA continue itswork toward accounting for changes in natural-resource assets and for the flow of services from theseassets.

Environmental variables affect economic well-being in three major ways: direct effects on con-sumption or income of households, industry, andgovernment; accumulation in the environment ofstocks of residuals that then affect economic activ-ities or economic assets; and effects on the serviceflows of economic assets, including capital stock,natural resources, and human resources. The mainvalue of natural-resource accounting is in provid-ing a complete picture of the role these resourcesplay in the economy. Sometimes this informationcan be used to judge the overall sustainability of theuse of resources, while at other times it can be usedto manage natural and environmental resourcesand to inform public policy choices.

Valuation

4.2 For valuation, the panel recommends that BEArely primarily on market values or proxies of market

values that are based on actual behavior. Contin-gent valuation, while sometimes useful for otherpurposes, is currently of limited value for environ-mental accounting in the context of the economicaccounts.

Valuing environmental goods and services re-quires distinguishing between private and publicgoods. Market prices provide the marginal valua-tions for private goods, but determining the valueof public goods requires the summation of individ-ual values. Moreover, there may be no behavioraltraces for individual valuation of public goods.

Price data are relatively reliable for private mar-ket goods produced from forest and agriculturalassets, such as timber stumpage, livestock, andland use and quality. Values for near-marketgoods—those that have direct counterparts in themarket—can be constructed by comparing thenear-market goods with their market counterparts,adjusting for quality as necessary. Techniques forvaluation of public goods are still under devel-opment. Some techniques—such as hedonic ortravel-cost studies—rely on behavioral or market-based estimates; while these estimates are subject tosignificant measurement errors, they are concep-tually appropriate in economic accounts. Othertechniques, such as contingent valuation, are notbased on actual behavior, are highly controversial,and are subject to potential response errors.

Quantitative Data

4.3 Quantitative data on many natural-resource as-sets are currently relatively adequate. However,the data on many environmental variables are atpresent poorly designed for the construction of en-vironmental accounts. The panel recommends thatgreater emphasis be placed on measuring effects asdirectly as possible. Of particular importance aremeasures of actual human exposure to air and waterpollutants, rather than modeled measures of expo-sure based on ambient pollutant levels at currentmonitoring sites.

Quantitative data for natural resources are oftenof high quality relative to the other quantitativedata in the NIPA because there are well-establishedunits of measure for many natural resources.Quantitative data on near-market activities such asfuel wood for own use are conceptually straight-forward, and many of these data are currentlycollected by federal agencies. Measurement ofnonmarket goods and services and explicit ac-counting for quality changes, particularly for thosethat have public-good characteristics, are currentlysubject to severe methodological difficulties and

March 2000 • 49

insufficient data. There are relatively good dataon emissions of many residuals from industrialand human activities, but for most harmful pol-lutants except lead there is very little systematicmonitoring of human exposures.

Inclusion of Public Goods

4.4 The panel finds that more work will be neededon techniques for establishing production flows andvalues for the assets and services of public goods toplace them on a comparable basis with the prices andquantities used in the core accounts.

True public goods, for example biodiversity,species preservation, and national treasures suchas the Florida Everglades and Yellowstone NationalPark, present severe conceptual and measurementissues for incorporation into a national accountingsystem.

Data Collection

4.5 The panel encourages BEA to help mount aconcerted federal effort to identify the data neededfor measuring changes in the quantity and qualityof natural-resource and environmental assets andassociated nonmarket service flows.

Many different federal agencies collect data orhave expertise that will be essential to BEA, particu-larly as its efforts expand to include Phase III assetsand associated flows. BEA already cooperates withother agencies in collecting data for the core ac-counts; supplemental environmental accounts willrequire cooperation with, for example, the Envi-ronmental Protection Agency, the Department ofAgriculture, the Department of the Interior, theBureau of Labor Statistics, the Bureau of the Cen-sus, the Energy Information Administration, theNational Institute of Environmental Health Sci-ences, and the Department of Health and HumanServices.

Regional Resolution

4.6 The panel recommends BEA focus on developingsupplemental accounts for the nation as a whole asa first priority. At the same time, BEA should pre-serve regionaldetail where it exists so that these dataare available for analysts interested in developingaccounts at the regional level.

The development of national estimates willrequire sampling, measurement, and valuationtechniques that reflect the fact that the qualityand value of natural-resource assets and associatedflows vary geographically. While some assets and

flows may not be important to the national econ-omy, they could be far more important to regionaland local economies.

Next Steps

4.7 The panel recommends that funds be providedto reinitiate and improve the design of the collec-tion of data on pollution control and abatementexpenditures.

4.8 As BEA further develops its natural-resourceand environmentalaccounts, an important step is toincorporate near-market goods and services—thosethat have close counterparts in marketed goods andservices. There is a clear basis here for measur-ing quantities and establishing values in a mannercomparable to that used for the core accounts.

4.9 Construction of a set of forest accounts isa natural step in developing integrated economic-environmental accounts. The United States hasmuch of the data needed for such an effort, and theanalytical techniques are relatively well developed.

4.10 Based on available information, the economicimpacts of air quality are likely to be the mostsignificant element in the environmental accounts;development of such accounts is a central task for en-vironmental accounting. At the same time, becauseof the unresolved conceptual issues and the need forappropriate physical measures, the development ofstock and flow accounts for air quality and otherimportant public goods poses awesome difficulties.This task far transcends the scope, budget, and ex-pertise of BEA. A major goal for the near term isto continue basic research on the underlying scienceand economics in this area.

References

Anderson, R., and M. Rockel. 1991. EconomicValuation of Wetlands. Discussion Paper #65,American Petroleum Institute. Washington, DC.

Arrow, K., R. Solow, P. Portney, E. Leamer,R. Radner, and H. Schuman. 1993. Reportof the National Oceanographic and AtmosphericAdministration Panel on Contingent Valuation.Federal Register (January) 58:4601–4614.

Berck, P. 1979. The economics of timber: Arenewable resource in the long run. Bell Journal ofEconomics 10:447–462.

Binkley, C.S., and J.R. Vincent. 1988. Timberprices in the U.S. South: Past trends and outlookfor the future. Southern Journal of Applied Forestry12:15–18.

Birdsey; R.A., and L.S. Heath. 1995. Carbonchanges in U.S. forests. 70 pp. in L. Joyce, ed., Pro-ductivity of American Forests and Climate Change.

50 • March 2000

General Technical Report RM–271, USDA FS Pub.,Rocky Mountain Forest and Range ExperimentalStation, Fort Collins, CO.

Bowes, M., J. Krutilla, and T. Stockton. 1984.Forest management for increased timber and wateryields. Water Resources Research 20:655–663.

Braden, J.B., and C.D. Kolstad, eds. 1991.Measuring the Demand for Environmental Quality.Amsterdam: North-Holland, Elsevier PublishersB.V.

Brown, S. 1996. Managing forests for mitigationof greenhouse gases. Pp. 773–778 in R.T. Wat-son, M.C. Zinyowera, and R.M. Moss, eds., ClimateChange 1995: Impacts, Adaptations, and Mitigationof Climate: Scientific and Technical Analyses. Pub-lished for the Intergovernmental Panel on ClimateChange. Cambridge: Cambridge University Press.

Bruce, J.P., H. Lee, and E.H. Haites, eds. 1996.Climate Change 1995: Economic and Social Dimen-sions. Published for the Intergovernmental Panelon Climate Change. Cambridge, U.K.: CambridgeUniversity Press.

Bureau of Economic Analysis. 1994a. Integratedeconomic and environmental satellite accounts.SURVEY OF CURRENT BUSINESS April: 33–49.

Bureau of Economic Analysis. 1994b. Ac-counting for mineral resources. SURVEY OF CURRENT

BUSINESS April: 50–72.Clawson, M. 1979. Forests in the long sweep of

American history. Science 204:1168–1174.Clean Air Act Council on Compliance. 1997.

Report of the Clean Air Act Council on Com-pliance. Letter report to the EPA Administrator,private communication.

Cornes, R., and T. Sandler. 1986. The Theory ofExternalities, Public Goods, and Club Goods. NewYork: Cambridge University Press.

Costanza, R., R. d’Arge, R. de Groot, F. Farber,M. Grasso, B. Hannon, K. Limburg, S. Naeem, R.V.O’Neill, J. Pervello, R.Q. Raskin, P. Sutton, andM. Van den Belt. 1997. The value of the world’secosystem services and natural capital. Nature(May) 387:253–260.

Diamond, P.A., and J.A. Hausman. 1994. Con-tingent valuation: Is some number better thanno number? Journal of Economic Perspectives8(4):45–64.

Green, I.M., C. Folke, K. Turner, and I. Mate-man. 1994. Primary and secondary valuation ofwetland ecosystems. Environmental and ResourceEconomics 4:55–74.

Hanemann, W.M. 1994. Valuing the environ-ment through contingent valuation. Journal ofEconomic Perspectives 8(4):19–43.

Howell, S.L. 1996. A review of the concep-tual and methodological issues in accounting forforests. In Proceedings of Third Meeting of the Lon-don Group on Natural Resource and EnvironmentalAccounting. Stockholm: Statistics Sweden.

Lyon, K.S. 1981. Mining of the forest andthe time path of the price of timber. Jour-nal of Environmental Economics and Management8:330–344.

Mitchell, R.C., and R.T. Carson. 1989. UsingSurveys To Value Public Goods: The Contingent Val-uation Method. Washington, D.C.: Resources forthe Future.

Pearce, D.W., W.R. Cline, A.N. Achanta, S.Fankhauser, R.K. Pachauri, R.S.J. Tol, and P. Vel-linga. 1996. The social costs of climate change:Greenhouse damage and the benefits of control.Pp. 145–178 in J.P. Bruce, H. Lee, and E.H. Haites,eds., Climate Change 1995: Economic and Social Di-mensions. Cambridge, U.K: Cambridge UniversityPress for the Intergovernmental Panel on ClimateChange.

Portney, P.R. 1994. The contingent valuationdebate: Why economists should care. Journal ofEconomic Perspectives 8(4):3–17.

Randall, A., and J. Stoll. 1983. Existence valuein a total valuation framework. In R. Rowe and L.Chestnut, eds.,

Managing Air Quality and Scenic Resources atNational Parks and Wilderness Areas. Boulder, CO:Westview Press.

Samuelson, P.A. 1954. The pure theory of pub-lic goods. The Review of Economics and Statistics36(4):387–389.

Samuelson, P.A. 1955. Diagrammatic expositionof a theory of public expenditure. The Review ofEconomics and Statistics 37(4):350–356.

Sedjo, R. 1990. The Nation’s Forest Resources.Discussion Paper ENR 90–07. Washington, DC:Resources for the Future.

Sedjo, R.A., and K.S. Lyon. 1990. The Long-TermAdequacy of World Timber Supply. Washington,DC: Johns Hopkins University Press.

Smith, V.K. 1993. Nonmarket valuation of en-vironmental resources: An interpretive appraisal.Land Economics 69(1):1–26.

U.S. Department of Agriculture Forest Service.1995. Chapter 4, Socioeconomic Effects and Im-plications of the Proposed Program, 1995 draftRPA program.<http://www.fs.fed.us/land/RPA/chapt4.htm>.

U.S. Environmental Protection Agency. 1996.National Air Quality and Emissions Trend Report,1995. EPA 454–R–96–015 Office of Air Quality

March 2000 • 51

Planning and Standards, Research Triangle Park,NC, October.

U.S. Environmental Protection Agency. 1997.The Benefits and Costs of the Clean Air Act, 1970to 1990. Draft, April. Office of Air and Radia-tion/Office of Policy Analysis and Review/Office ofPolicy, Planning, and Evaluation.

Vincent, J. 1997. Net Accumulation of Tim-ber Resources. Manuscript, Harvard Insti-

tute for International Development, Cambridge,MA.

Vincent, J., and J.M. Hartwick. 1997. Account-ing for the Benefits of Forest Resources: Conceptsand Experience. Draft, July 10. FAO ForestryDepartment.

Washburn, C.L. 1990. The Determinants of For-est Value in the U.S. South. Ph.D. thesis, YaleUniversity.