Emerging Trends in The Social and Behavioral Sciences

Environmental Accounting

Item

Title

Environmental Accounting

Author

Muller, Nicholas Z.

Research Area

Social Processes

Topic

Environmental and Climate Change

Abstract

While metrics such as gross domestic product (GDP) constitute important means to gauge the value of production, it is widely recognized that indices that focus exclusively on market production are incomplete. Omitted environmental assets include (i) those that have the capacity to act as a source of valuable inputs to production such as timber, subsurface minerals, or fisheries; and (ii) media that serve as a sink for anthropogenic residuals such as air, water, or soil. The human health impacts depend crucially on two parameters in the integrated asset models (IAMs): the effect that exposures to fine particles have on adult mortality rates and the value attributed to small changes in mortality risk. This essay focuses on recent research that augments standard measures of output to include damages from air and greenhouse gas pollution into national output.

Related Essays

Sociological Theory After the End of Nature (Sociology), Robert J. Brulle

Misinformation and How to Correct It (Psychology), John Cook et al.

Cities and Sustainable Development (Sociology), Christopher Cusack

Theorizing the Death of Cities (Political Science), Peter Eisinger

Architecture of Markets (Sociology), Neil Fligstein and Ryan Calder

Modeling Coal and Natural Gas Markets (Economics), Franziska Holz

Presidential Power (Political Science), William G. Howell

Why Do States Sign Alliances? (Political Science), Brett Ashley Leeds

Money in Politics (Political Science), Jeffrey Milyo

Economics of Renewable Energy Production (Economics), Gregory F. Nemet

The Politics of Disaster Relief (Political Science), Alexander J. Oliver and Andrew Reeves

The Social Science of Sustainability (Political Science), Johannes Urpelainen

Identifier

etrds0430

extracted text

Environmental Accounting
NICHOLAS Z. MULLER

Abstract
While metrics such as gross domestic product (GDP) constitute important means to
gauge the value of production, it is widely recognized that indices that focus exclusively on market production are incomplete. Omitted environmental assets include
(i) those that have the capacity to act as a source of valuable inputs to production
such as timber, subsurface minerals, or fisheries; and (ii) media that serve as a sink
for anthropogenic residuals such as air, water, or soil. The human health impacts
depend crucially on two parameters in the integrated asset models (IAMs): the effect
that exposures to fine particles have on adult mortality rates and the value attributed
to small changes in mortality risk. This essay focuses on recent research that augments standard measures of output to include damages from air and greenhouse
gas pollution into national output.

INTRODUCTION
The development of a standardized set of tools to track national income and
output remains an important achievement in the field of economics.
While the GDP and the rest of the national income accounts may seem to be
arcane concepts, they are truly among the great inventions of the twentieth
century. Samuelson, P.A. and Nordhaus, W.D. (as quoted in Landefeld, 2000).

Some 80 years after their development, tools such as gross domestic product (GDP) are used by nations and governments throughout the world to
estimate economic performance.
The value in such mensuration is manifold. Time series measurement of a
particular nation’s output, expressed in real terms, provides society with a
sense of trends in output. Cross-sectional comparisons on purchasing power
parity bases facilitate assessment of the relative wealth of nations. The
impact of policy interventions, cyclical or episodic downturns, international
or civil conflict, among other factors, is gauged with metrics such as GDP.

Emerging Trends in the Social and Behavioral Sciences.
Robert Scott and Marlis Buchmann (General Editors) with Stephen Kosslyn (Consulting Editor).
© 2017 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.

1

2

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Conversely, policymakers rely on national output statistics when designing
fiscal and monetary policies.
While metrics such as GDP constitute important means to gauge the value
of production, it is widely recognized, and has been for some time, that
indices that focus exclusively on market production are incomplete. In the
most basic sense, GDP focuses on goods transacted in markets. The problem
then is that much economic activity occurs outside of formal markets.
Although Pigou (1932) arguably first tackled the problem of external
economies, Nordhaus and Tobin (1972) were the first to point this out in
the context of the national accounts. Emphasizing aspects of production
outside the purview of GDP such as leisure time, home production, and
environmental goods and services, Nordhaus and Tobin (1972) reported
that in 1965, their proposed (more comprehensive) measure differed from
standard measures by a factor of two. Clearly, they raised an important set
of questions.
More recently, Stiglitz (2009) argued, in addition to recognizing that GDP is
incomplete, that policies that formulate objectives around GDP may be counterproductive.
If we have poor measures, what we strive to do (say, increase GDP) may actually contribute to a worsening of living standards. (Stiglitz, 2009)

Thus, it is the use of GDP to guide national priorities that is problematic.
One of the particular concerns raised by Stiglitz (2009), which was also part of
Nordhaus and Tobin’s proposed extension to GDP, was the value of environmental assets. Using a measure of national output that omits natural capital,
Stiglitz argues that:
[w]e may also be confronted with false choices, seeing trade-offs between output and environmental protection that don’t exist. (2009)

What aspects of the environment are omitted from standard measures of
national output? The literature in this area typically characterizes environmental assets as (i) those that have the capacity to act as a source of valuable
inputs to production such as timber, subsurface minerals, or fisheries; and
(ii) media that serve as a sink for anthropogenic residuals such as air, water,
or soil (Hecht, 2005).
Clearly, one may argue that the value of timber for building or making
paper and minerals such as coal, oil, and gas are embedded in GDP. This is
only partially true. These assets appear in GDP only when privately owned
and when they provide (or may generate in the future) financial benefit to
owners. Publics lands, and resources therein, as well as wild areas and native
species are assets not tracked by the conventional accounts (Hecht, 2005).

Environmental Accounting

3

Conventional measures provide an incomplete glimpse at the value of natural resources.
Are assets such as air and water really valuable? That is, do they belong in
an augmented measure of national output? To explore this important question, consider how a firm disposes of nontoxic solid waste. Typically, the firm
pays a fee to a carting service to transport the waste to either a landfill or an
incinerator. Both destinations typically require fees to dispose of the waste.
The important point here is that there are payments for waste disposal. Why
is that the case?
This is because for a landfill, property rights for land are well defined. Even
without policy constraints, a firm would not be able to simply dump waste
on property that someone else owns. That is a damaging act. When we do
observe illegal dumping it tends to be in vacant lots or other properties that
are not closely monitored by owners, or the ownership of which is ambiguous. Firms must pay to dispose of solid waste. It consumes valuable space. It
is a cost of doing business.
Now consider a firm that produces smoke as it makes its products. In the
absence of regulation, the firm generates smoke free of charge. Note that this
contrasts with the case of solid waste in which a firm must pay for disposal.
Why is there this difference? The problem reduces to one of property rights.
Airborne effluent disperses across space. Particles and gases cross municipal,
state, regional, and national boundaries. These substances may affect multiple parties, raising concerns of public good. They may mingle with pollution
from other firms, making attribution unclear. And, in the case of long-lived
greenhouse gases, damages manifest many years after emission, obfuscating
the link between perpetrator and impact.
So, do environmental assets that serve as a repository for waste belong in a
system of national accounts? Well, if the case of solid waste disposal is used
as a conceptual guide, the answer is clearly yes. Long-term storage of waste
is a valuable service. And when property rights are established, firms (and
households) have revealed a willingness to pay for such services.
The remainder of this essay focuses on recent research that augments
standard measures of output to include damages from air and greenhouse
gas pollution into national output. It highlights aspects of current research
related to this topic and it then proposes new directions for this field of
inquiry.
AIR POLLUTION IN A SYSTEM OF EXTENDED ACCOUNTS
In its 1999 Nature’s Numbers report, the National Research Council of the
National Academies of Science (NAS NRC) identified air pollution damages
as a top priority for extending conventional measures of output:

4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

In the panel’s view, no other area of natural-resource and environmental
accounting would have as great an impact as the potential correction for air
quality. The magnitude of this impact indicates that the development of the
supplemental accounts for air quality is a high priority. (National Academy of
Sciences, National Research Council [NAS NRC], 1999, p. 148)

Part of the motivation for this statement was the finding by the United
States Environmental Protection Agency (USEPA) that the benefits from air
pollution control were on the order of $1 trillion (United States Environmental Protection Agency [USEPA], 1999). Another impetus for the Academies’
position was that the measures reported by USEPA were incomplete (NAS
NRC, 1999).
With the impetus for inclusion of air pollution damage into the NIPAs
established, the question of how to augment the accounts arises. As noted by
Abraham and Mackie (2006), one way in which pollution externalities may
be treated in the NIPAs is as a cost of production. Polluting firms require
disposal of residuals resulting from production. In the absence of property
rights on the atmosphere firms face no cost or fee for this disposal. The role of
an augmented accounting system is to provide a structure in which national
income reflects these costs. Such a framework would deduct these costs from
extant measures of output. In accordance with this treatment, recent work in
this area deducts pollution damage from conventional measures of output
(Bartelmus, 2009; Muller, 2014; Muller, Mendelsohn, & Nordhaus, 2011).
One concern with this approach is that these costs of disposal are already
in the accounts. However, prior research has shown that the vast majority
of air pollution damages are comprised of premature mortality risks, and,
therefore, they lie beyond the market boundary (Muller et al., 2011; USEPA,
1999, 2011).
The fact that firms are often not charged for the use of public waste sinks
such as air, water, and soils means that production, generally, is more costly
than is implied by measures such as GDP. This could, in principle, have two
effects. First, if the adverse consequences of using these sinks manifests in
markets, then it is only the distribution of GDP across sectors that is incorrectly measured. For example, if manufacturing firms producing smoke primarily affect firms producing crops, then the value of manufacturing output
is artificially high while that of crop production is suppressed. If, however,
the deleterious consequences of pollution occur outside of the market boundary, then both the allocation of GDP and its level are wrong. Making the former
correction to GDP alters the perceived productivity of different sectors in the
economy. Making the second deduction also means gross output is lower.
A third issue that was raised conceptually some 15 years ago is how measures of growth are affected by the augmentations discussed above. In terms

Environmental Accounting

5

of pollution, the effects on growth depend on the trajectory of damages. The
NAS NRC panel made the following statement when considering the effects
of a period of improving air quality:
The result might be a substantial increase in the estimate of growth of comprehensive consumption over this period. (NAS NRC, 1999, p. 147)

By the same logic, if air quality conditions deteriorate, augmented measures
of growth will fall relative to their market counterpart.
A final conceptual point worth noting is how total damages are calculated.
In estimating the total value of market output, GDP multiplies the current
price of each good by the quantity produced. In pursuit of a seamless integration of pollution damages into this extant system of accounts, measures
of total damage are typically tabulated by estimating marginal damage (the
extra harm caused by one more unit of emissions) and multiplying that by the
quantity of emissions produced (Abraham & Mackie, 2006; Nordhaus, 2006).
Thus, in the methodological section below, significant emphasis is placed on
the estimation of marginal damages.
HOW ARE DAMAGES MEASURED?
Over the past 15 years, many papers have developed advances in modeling
techniques that facilitate measuring air pollution damages (Banzhaf, Burtraw, & Palmer, 2004; Fann, Fulcher, & Hubbell, 2009; Heo, Adams, & Gao,
2016; Jaramillo & Muller, 2016; Kerl et al., 2015; Levy, Baxter, & Schwartz,
2009; Muller, 2014; Muller & Mendelsohn, 2007, 2009; Tessum, Hill, & Marshall, 2015; USEPA, 1999). Collectively, these papers suggest that (i) air pollution damages comprise a significant share of GDP, (ii) the majority of the air
pollution damages are comprised of adverse consequences on human health,
and (iii) damages hinge critically on two parameters that are discussed below.
Additionally, Muller (2014) finds that the changes in damages have an appreciable effect on measures of growth.
How are these damages calculated? Damages from air pollution are typically computed using integrated assessment models (IAMs). An IAM uses
findings from engineering, atmospheric chemistry and physics, epidemiology, and economics to link the production of residuals to final, monetized
consequences. Although the earliest examples of IAMs date back to the late
1970s and early 1980s (Mendelsohn, 1980), the recent literature has made a
number of advances in these devices (Fann et al., 2009; Heo et al., 2016; Kerl
et al., 2015; Muller & Mendelsohn, 2007, 2009; Tessum et al., 2015).
Most IAMs focusing on air pollution connect emissions of common air pollutants to estimates of ambient concentrations, exposures, physical impacts,

6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

and monetary damage so that the impacts can be incorporated directly into
the national accounts. The air pollutants covered often include nitrogen
oxides (NOx ), sulfur dioxide (SO2 ), fine particulate matter (PM2.5 ), ammonia
(NH3 ), and volatile organic compounds (VOCs). The first step in an IAM
for air pollution relies on emissions data, which reveal where, when, and
in what quantities these substances are produced. In developed economies,
public environmental regulatory agencies typically provide these data.
(Often these agencies are charged with environmental enforcement, which
depends on tracking emissions. So they have the data anyway.)
In the United States, emissions data are reported by the USEPA. In an
IAM, these data are fed through an air quality model that links emissions to
ambient concentrations. Exposures are tabulated by tracking populations of
sensitive receptors (human populations and crops, for example). Physical
impacts, such as reduced crop yields and increased rates of illness, are
estimated through the use of concentration–response functions gleaned
from peer-reviewed publications. Valuation uses either reported market
prices (for impacts on crops) or non-market valuation techniques.
As mentioned above, most environmental accounting exercises tabulate
total damage as the product of marginal damages and emissions. One
of the advances made in the design of IAMs lies in spatial resolution.
Source-specific damages are now computed for particular power plants and
manufacturing facilities. For example, recent work indicates that the per ton
damage from emissions of sulfur dioxide (a common air pollutant produced
when coal is burned) varies by over 100 times within the contiguous United
States (Muller & Mendelsohn, 2009, 2012). Figure 1 shows this stunning
range of impacts. Because the impact of emissions of local air pollution
varies a lot according to where it is emitted, the enhanced resolution is a
critical advance. After all, if total damages are computed as the product
of marginal damage times total emissions, then getting the total damage
figure right depends on having the margins correctly estimated. And,
if the marginal damages vary drastically across space, then an accurate
estimate of total damage depends on having source-specific marginal
damages.
Figure 1 also indicates that the largest damages on a per ton basis are
due to emissions produced in cities. This stems from the importance of
human health effects in pollution damage. The robust finding that the
largest share of damage is due to human health impacts depends crucially
(as noted above) on two parameters in the IAMs: the effect that exposures
to fine particles have on adult mortality rates, and the value attributed
to small changes in mortality risk. Because of the importance of these
parameters in tabulating damages, both are discussed in the following
paragraphs.

Environmental Accounting

7

Canada

Atlantic
Ocean

Damage ($/ton)
0 – 500

2500 – 5000

500 – 1000

5000 – 7500

1000 – 2500

7500 – 15,000

Pacific
Ocean

Figure 1

15,000 – 57,000

Mexico
Gulf of Mexico

SO2 marginal damages. Source: Muller and Mendelsohn (2012).

Traditionally, evidence about the effect of exposure to pollution on human
health is obtained from peer-reviewed research in the epidemiological literature. And, in terms of the effects of exposure to fine particles on premature
mortality in the United States, two studies are most frequently used (Krewski
et al., 2009; Lepeule, Laden, Dockery, & Schwartz, 2012). Damages computed
using these two studies (holding all else in the IAM fixed) differ by more than
a factor of two. How do policymakers or academic researchers conducting
environmental accounting resolve these differences? Typically, both parameters are used in a sensitivity analysis in order to bound or bracket damage
estimates (USEPA, 1999, 2011).
A second point related to this issue is worthy of discussion; recent research
in environmental economics has brought more sophisticated econometric
techniques to the question of how exposure to environmental pollutants
affects mortality rates. This research has raised two critical questions. First,
do the epidemiological studies effectively control for a range of pollutants
such that regulators are confident that risks attributed to particulate matter
are, in fact, due to exposures to that pollutant (Beatty & Shimshack, 2011;
Currie, Neidell, & Schmeider, 2009)? Second, is the magnitude of the effect
that particulate matter has on mortality rates different when researchers use
quasi-experimental models that estimate a causal effect of particulates on
mortality rates (Chay & Greenstone, 2003)? Because mortality effects play

8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

such a large role in the damages from air pollution, new findings on whether
the toxicity of particulate matter holds up to different empirical methods
have potentially significant ramifications for integration of pollution damage
into national accounts.
For a number of years, the most common approach to valuing mortality
risk uses the value of a statistical life (VSL). One common empirical approach
to estimating the VSL uses hedonic wage models to identify the marginal
implicit price of risk of death on the job (Viscusi & Aldy, 2003). The use
of these results to value risks due to air pollution exposure is controversial for several reasons. First, hedonic wage studies rely on a sample of
people from the workplace. Hence, they neglect the elderly and the very
young. This is problematic because most of the premature deaths from air
pollution manifest among the elderly and infants. The VSL is extrapolated
to these non-working populations. Second, risks on the job are typically,
though not always, acute or accidental events. In contrast, risks from air
pollution involve chronic, long-term exposure. Thus, the nature of the risks
differs.
In part because of the concerns related to transferring risk valuation
estimates from hedonic wage studies to the context of air pollution, environmental economists have estimated VSLs in different contexts. For example,
employing variation in vehicle air-bag requirements, researchers have
found a range of VSLs quite similar to those estimated from hedonic wage
models (Rohlfs, Sullivan, & Kneisner, 2015). Other authors explore trade-offs
between risks of death while driving and time savings implied by speed limits (Ashenfelter & Greenstone, 2004). In addition, economists have exploited
data from the Alaskan crab industry to estimate a VSL (Schneir, Horrace,
& Felthoven, 2009). Environmental economists are actively engaged in
exploring new ways to elicit the trade-off between money and risk, which is
very important in the estimation of pollution damage.
RECENT RESULTS AND FUTURE DIRECTIONS
A number of recent papers have contributed to the empirical environmental accounting literature that specifically focuses on air pollution damage
(Bartelmus, 2009; Muller et al., 2011; Muller, 2014, 2016; Jaramillo & Muller,
2016). Bartelmus (2009) estimated the costs due to consumption of environmental capital in various parts of the world for a collection of years between
1990 and 2006. In the European Union environmental costs were under 1% of
GDP. In the United States this figure was under 3%. Intuitively, environmental costs were much higher in developing economies. In China, Bartelmus
estimated environmental impact at up to 12% of GDP in 2006. In Africa as
a whole, this metric was over 25%. These values, while highly uncertain,

Environmental Accounting

9

are illustrative in the cross section and the time series dimensions. In every
region covered, environmental costs rise over time.
While Bartelmus’ work is highly aggregated, a more recent paper drilled
down in great detail in the U.S. economy. Muller et al. (2011) calculated the
air pollution damage for every industry in the U.S. economy. Why should
an economic system be explored at this level of detail? A modern, developed economy consists of a wide range of productive activities: from heavy
manufacturing at iron foundries to financial services and insurance. Accordingly, pollution intensity varies dramatically across sectors; the importance
of augmenting national accounts varies across sectors as well.
In the United States in 2002, pollution damage as a fraction of sector value
added ranged from 0.38 for agriculture/forestry to 0.01 for manufacturing,
and to 0 for sectors such as finance and insurance (Muller et al., 2011).
Within sectors, great variation in pollution intensity was also reported. For
example, while electric power generation using coal produced damages over
two times larger than value-added, natural-gas-based power generation
produced damages less than 10% of value-added (Muller et al., 2011).
What does it mean to have an industry generate greater environmental
pollution damage than its value-added? It may be tempting to say that such
industries are doing more harm than good and therefore should be shut
down. That position is incorrect. If the external costs were internalized by
the firm (perhaps through an efficiently calibrated emission tax) it is quite
likely that output produced by the industry would fall. That is, if no viable
end-of-pipe pollution controls exist (as is the case for CO2 ), then the firm
would either switch to cleaner, more expensive inputs, or they would reduce
emissions by reducing production. If abatement technology exists, then they
would substitute away from some productive inputs toward abatement
devices. In either case output will fall, ceteris paribus. If the entire sector or
industry adopts such responses to internalized costs, then prices will adjust
and value-added will rise. Concurrently damages will fall and the damage
to value-added ratios will fall.
A particularly exciting area of research in this space focuses on measures of
growth in GDP versus growth in measures of output that deduct pollution
damage. Why is this focus on growth? First, the primary value of measures
such as GDP lies in examining changes over time. Academics, policymakers,
and the general public draw inferences regarding national economic performance from changes in GDP rather than levels. As such, a pressing question
in environmental economics is how inclusion of pollution damage into an
augmented system of accounts affects comprehensive growth rates. Second,
from the perspective of public policy, macroeconomic regulatory agencies
often base policy choices on quarterly or annual changes in aggregates such
as GDP. Prior rates of change in GDP are clearly linked to monetary and fiscal

10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

policy; one needs to look no further than the Federal Reserve’s response to
the Great Recession for evidence of this claim. In principle, measures of augmented growth could be used in the same way with one additional dimension. Beyond the standard monetary policy toolkit, macroeconomic regulators basing policy choices on the adjusted output measure could prescribe
reductions to damage as a means to enhance growth. That is, if pollution
damage has a significant effect on augmented measures of growth, as the
NAS NRC panel suggested, does this not merit consideration of pollution
control as a viable lever for growth enhancements?
Recent research using data from the United States suggests that the difference between conventional and environmentally adjusted growth rates is
appreciable. For example, over the period 1999 to 2011, GDP in the United
States expanded at annualized rates of between 1.2% and 2.8%. When corrected for both air pollution and carbon dioxide (CO2 ) damage, national output grew by about 0.3 percentage points faster (Muller, 2014, 2016).
Drilling down further, from 2005 to 2008, the U.S. economy was barreling
toward the Great Recession. Annual GDP growth over this period fell to just
over 1%. However, as conventional measures of growth receded, air pollution
damages did too. Thus, the measure that recognized this reduction in damage
expanded more rapidly. In particular, from 2005 to 2008, growth adjusted to
account for air pollution and CO2 emissions was about 0.4% more rapid than
growth in GDP (Muller, 2016).
The contribution of these results is twofold. First, periods of improvement
in air quality can have an appreciable and positive effect on growth. The
intuition of the NAS NRC panel cited above was confirmed. Second, these
results highlight Stiglitz’ position that focusing on market-centric measures
of output may foster the illusion of a trade-off between investments in environmental quality and growth in market production.
The difference in growth discussed above is largest between 1999 and 2002.
One reason for this finding is that Phase II of the Acid Rain Program began in
2000. Because of this policy, emissions of SO2 from regulated power plants fell
by 2.3 million tons between 1999 and 2002. Importantly, costs of compliance
with the regulation are embedded in GDP, while reduced damages are not.
The finding that the benefits of pollution control are large enough to affect
national growth during this time period suggests the need to amend official
measures of output to include the benefits from environmental policy.
Exciting work in this space is occurring in areas beyond air pollution and
greenhouse gas emissions. For example, research is currently ongoing to construct an IAM for water pollution in the United States with resolution akin
to that of modern air pollution IAMs. Once operational, a model such as
this will facilitate, among other topics, the ability to engage in multimedium
pollution control analyses. For example, in modern economies one central

Environmental Accounting

11

aspect of water pollution control is municipal sewage treatment. While on
the surface this appears to be beneficial, sewage treatment generates copious
amounts of airborne ammonia emissions. Ammonia is an important catalyst
in the formation of secondary fine particulate matter, which elevates mortality risk. What is the value of waste removal from water? What is the damage
caused by airborne emissions from this treatment? What is the optimal mix
of air and water pollution emissions from sewage treatment? Intriguing and
as of yet unanswered questions such as these will form the basis of future
research in the field of environmental economics. And, as an ever greater
range of pollution damages are added to the environmental accounts, our
measures of economic performance will do a better job of measuring welfare.

REFERENCES
Abraham, K. G., & Mackie, C. (2006). A framework for nonmarket accounting. In D.
W. Jorgensen, J. S. Landefeld & W. D. Nordhaus (Eds.), A new architecture for the
U.S. national accounts. NBER studies in income and wealth (Vol. 66). Chicago, IL: The
University of Chicago Press.
Ashenfelter, O., & Greenstone, M. (2004). Using mandated speed limits to measure
the value of a statistical life. Journal of Political Economy, 112(2), S226–S267.
Banzhaf, H. S., Burtraw, D., & Palmer, K. (2004). Efficient emission fees in the US
electricity sector. Resource and Energy Economics, 26, 317–341.
Bartelmus, P. (2009). The cost of natural capital consumption: Accounting for a sustainable world economy. Ecological Economics, 68, 1850–1857.
Beatty, T., & Shimshack, J. (2011). School buses, diesel emissions, and respiratory
health. Journal of Health Economics, 30(5), 987–999.
Chay, K. Y., & Greenstone, M. (2003). The impact of air pollution on infant mortality:
Evidence from geographic variation in pollution shocks induced by a recession.
Quarterly Journal of Economics, 118(3), 1121–1167.
Currie, J., Neidell, M., & Schmeider, J. F. (2009). Air pollution and infant health:
Lessons from New Jersey. Journal of Health Economics, 28, 688–703.
Fann, N., Fulcher, C. M., & Hubbell, B. J. (2009). The influence of location, source and
emission type in estimates of the human health benefits of reducing a ton of air
pollution. Air Quality, Atmosphere and Health, 2, 169–176.
Hecht, J. E. (2005). National environmental accounting. Bridging the gap between ecology
and economy. Washington, DC: Resources for the Future Press.
Heo, J., Adams, P. J., & Gao, H. O. (2016). Public Health Costs of Primary PM2.5 and
Inorganic PM2.5 Precursor Emissions in the United States. Environmental Science
& Technology. doi:10.1021/acs.est.5b06125.
Jaramillo, P., & Muller, N. Z. (2016). The air pollution damage from energy production in the U.S.: 2002–2011. Energy Policy, 90, 202–211.
Kerl, P. Y., Zhang, W., Moreno-Cruz, J. B., Nenes, A., Realff, M. J., Russell, A. G.,
… Thomas, V. M. (2015). A new approach for optimal electricity planning and

12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

dispatching with hourly time-scale air quality and health considerations. Proceedings of the National Academy of Sciences, 112(35), 10884–10889.
Krewski, D., Jerrett, M., Burnett, R.T., Ma, R., Hughes, E., Shi, Y., … Tempalski, B.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study
linking particulate air pollution and mortality. HEI Research Report, 140, Health
Effects Institute, Boston, MA.
Landefeld, J. S. (2000). GDP: One of the great inventions of the 20th century. Survey
of Current Business. January, 6–14.
Lepeule, J., Laden, F., Dockery, D., & Schwartz, J. (2012). Chronic exposure to fine
particles and mortality: An extended follow-up of the Harvard Six Cities Study
from 1974 to 2009. Environmental Health Perspectives, 120(7), 965–970.
Levy, J. I., Baxter, L. K., & Schwartz, J. (2009). Uncertainty and variability in
health-related damages from coal-fired power plants in the United States. Risk
Analysis, 29(7), 1000–1014.
Mendelsohn, R. O. (1980). An economic analysis of air pollution from coal-fired
power plants. Journal of Environmental Economics and Management, 7, 30–43.
Muller, N. Z. (2014). Boosting GDP growth by accounting for the environment. Science, 345(6199), 873–874.
Muller, N.Z. (2016). The derivation of discount rates using an augmented measure of
income. NBER WP #22579.
Muller, N. Z., & Mendelsohn, R. O. (2007). Measuring the damages due to air pollution in the United States. Journal of Environmental Economics and Management, 54,
1–14.
Muller, N. Z., & Mendelsohn, R. O. (2009). Efficient pollution regulation: Getting the
prices right. American Economic Review, 99(5), 1714–1739.
Muller, N. Z., & Mendelsohn, R. O. (2012). Efficient pollution regulation: Getting
the prices right: Mortality rate update (Corrigendum). American Economic Review,
102(1), 613–616.
Muller, N. Z., Mendelsohn, R. O., & Nordhaus, W. D. (2011). Environmental accounting for pollution in the U.S. economy. American Economic Review, 101(5), 1649–1675.
National Academy of Sciences, National Research Council (1999). W. D. Nordhaus
& E. C. Kokkelenberg (Eds.), Nature’s numbers: Expanding the U.S. national economic
accounts to include the environment. Washington, D.C.: National Academy Press.
Nordhaus, W. D., & Tobin, J. (1972). Is growth obsolete? Studies in income and wealth
(Vol. 38). New York, NY: NBER.
Nordhaus, W. D. (2006). Principles of national accounting for non-market accounts.
In D. W. Jorgensen, J. S. Landefeld & W. D. Nordhaus (Eds.), A new architecture for
the U.S. National Accounts. NBER Ssudies in income and wealth (Vol. 66). Chicago, IL:
The University of Chicago Press.
Pigou, A. C. (1932). The economics of welfare. London, England: Macmillan and Co.,
Limited.
Rohlfs, C., Sullivan, R., & Kneisner, T. (2015). New estimates of the value of a statistical life using air bag regulations as a quasi-experiment. American Economic Journal:
Economic Policy, 7(1), 331–359.

Environmental Accounting

13

Schneir, K.E. Horrace, W.C., & Felthoven, R.G. (2009). The value of a statistical life:
Pursuing the deadliest catch. http://surface.syr.edu/cpr/49/
Stiglitz, J. E. (2009). GDP fetishism. The Economist’s Voice. September.
Tessum, C. W., Hill, J. D., & Marshall, J. D. (2015). InMAP: A new model
for air pollution interventions. Geoscientific Model Development–Discussions.
doi:10.5194/gmdd-8-9281-2015.
United States Environmental Protection Agency. (1999). The benefits and costs of the
Clean Air Act: 1990–2010. EPA Report to Congress. EPA 410-R-99-001, Office of Air
and Radiation, Office of Policy, Washington, D.C.
United States Environmental Protection Agency. (2011). The benefits and costs of the
Clean Air Act: 1990–2020. Final Report. Office of Air and Radiation, Office of Policy,
Washington, D.C.
Viscusi, W. K., & Aldy, J. E. (2003). The value of a statistical life: A critical review of
market estimates throughout the world. Journal of Risk Uncertainty, 27(1), 5–76.

NICK Z. MULLER SHORT BIOGRAPHY
Nick Z. Muller joined the Middlebury College economics faculty as an
assistant professor in the fall of 2007. He completed his dissertation at Yale
University in May of 2007, where his advisors included Robert Mendelsohn,
William Nordhaus, and Nathaniel Keohane. His dissertation focused on
using integrated assessment models to measure the damages from air
pollution in the United States, and to propose alternative market-based
policies intended to govern the criteria for air pollutants. Dr. Muller also
attended the School of Public and Environmental Affairs where he studied
environmental policy and public finance in pursuit of a master’s degree in
public administration. His current research includes measuring the damages
due to emissions from each industry in the U.S. economy, the design of
market-based environmental policies, and the construction price indices
for air pollution. His research has been published in the American Economic
Review, Science, the Journal of Environmental Economics and Management, and
Environmental Science & Technology, among other outlets.
RELATED ESSAYS
Sociological Theory After the End of Nature (Sociology), Robert J. Brulle
Misinformation and How to Correct It (Psychology), John Cook et al.
Cities and Sustainable Development (Sociology), Christopher Cusack
Theorizing the Death of Cities (Political Science), Peter Eisinger
Architecture of Markets (Sociology), Neil Fligstein and Ryan Calder
Modeling Coal and Natural Gas Markets (Economics), Franziska Holz
Presidential Power (Political Science), William G. Howell
Why Do States Sign Alliances? (Political Science), Brett Ashley Leeds

14

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Money in Politics (Political Science), Jeffrey Milyo
Economics of Renewable Energy Production (Economics), Gregory F. Nemet
The Politics of Disaster Relief (Political Science), Alexander J. Oliver and
Andrew Reeves
The Social Science of Sustainability (Political Science), Johannes Urpelainen



Environmental Accounting
NICHOLAS Z. MULLER

Abstract
While metrics such as gross domestic product (GDP) constitute important means to
gauge the value of production, it is widely recognized that indices that focus exclusively on market production are incomplete. Omitted environmental assets include
(i) those that have the capacity to act as a source of valuable inputs to production
such as timber, subsurface minerals, or fisheries; and (ii) media that serve as a sink
for anthropogenic residuals such as air, water, or soil. The human health impacts
depend crucially on two parameters in the integrated asset models (IAMs): the effect
that exposures to fine particles have on adult mortality rates and the value attributed
to small changes in mortality risk. This essay focuses on recent research that augments standard measures of output to include damages from air and greenhouse
gas pollution into national output.

INTRODUCTION
The development of a standardized set of tools to track national income and
output remains an important achievement in the field of economics.
While the GDP and the rest of the national income accounts may seem to be
arcane concepts, they are truly among the great inventions of the twentieth
century. Samuelson, P.A. and Nordhaus, W.D. (as quoted in Landefeld, 2000).

Some 80 years after their development, tools such as gross domestic product (GDP) are used by nations and governments throughout the world to
estimate economic performance.
The value in such mensuration is manifold. Time series measurement of a
particular nation’s output, expressed in real terms, provides society with a
sense of trends in output. Cross-sectional comparisons on purchasing power
parity bases facilitate assessment of the relative wealth of nations. The
impact of policy interventions, cyclical or episodic downturns, international
or civil conflict, among other factors, is gauged with metrics such as GDP.

Emerging Trends in the Social and Behavioral Sciences.
Robert Scott and Marlis Buchmann (General Editors) with Stephen Kosslyn (Consulting Editor).
© 2017 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.

1

2

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Conversely, policymakers rely on national output statistics when designing
fiscal and monetary policies.
While metrics such as GDP constitute important means to gauge the value
of production, it is widely recognized, and has been for some time, that
indices that focus exclusively on market production are incomplete. In the
most basic sense, GDP focuses on goods transacted in markets. The problem
then is that much economic activity occurs outside of formal markets.
Although Pigou (1932) arguably first tackled the problem of external
economies, Nordhaus and Tobin (1972) were the first to point this out in
the context of the national accounts. Emphasizing aspects of production
outside the purview of GDP such as leisure time, home production, and
environmental goods and services, Nordhaus and Tobin (1972) reported
that in 1965, their proposed (more comprehensive) measure differed from
standard measures by a factor of two. Clearly, they raised an important set
of questions.
More recently, Stiglitz (2009) argued, in addition to recognizing that GDP is
incomplete, that policies that formulate objectives around GDP may be counterproductive.
If we have poor measures, what we strive to do (say, increase GDP) may actually contribute to a worsening of living standards. (Stiglitz, 2009)

Thus, it is the use of GDP to guide national priorities that is problematic.
One of the particular concerns raised by Stiglitz (2009), which was also part of
Nordhaus and Tobin’s proposed extension to GDP, was the value of environmental assets. Using a measure of national output that omits natural capital,
Stiglitz argues that:
[w]e may also be confronted with false choices, seeing trade-offs between output and environmental protection that don’t exist. (2009)

What aspects of the environment are omitted from standard measures of
national output? The literature in this area typically characterizes environmental assets as (i) those that have the capacity to act as a source of valuable
inputs to production such as timber, subsurface minerals, or fisheries; and
(ii) media that serve as a sink for anthropogenic residuals such as air, water,
or soil (Hecht, 2005).
Clearly, one may argue that the value of timber for building or making
paper and minerals such as coal, oil, and gas are embedded in GDP. This is
only partially true. These assets appear in GDP only when privately owned
and when they provide (or may generate in the future) financial benefit to
owners. Publics lands, and resources therein, as well as wild areas and native
species are assets not tracked by the conventional accounts (Hecht, 2005).

Environmental Accounting

3

Conventional measures provide an incomplete glimpse at the value of natural resources.
Are assets such as air and water really valuable? That is, do they belong in
an augmented measure of national output? To explore this important question, consider how a firm disposes of nontoxic solid waste. Typically, the firm
pays a fee to a carting service to transport the waste to either a landfill or an
incinerator. Both destinations typically require fees to dispose of the waste.
The important point here is that there are payments for waste disposal. Why
is that the case?
This is because for a landfill, property rights for land are well defined. Even
without policy constraints, a firm would not be able to simply dump waste
on property that someone else owns. That is a damaging act. When we do
observe illegal dumping it tends to be in vacant lots or other properties that
are not closely monitored by owners, or the ownership of which is ambiguous. Firms must pay to dispose of solid waste. It consumes valuable space. It
is a cost of doing business.
Now consider a firm that produces smoke as it makes its products. In the
absence of regulation, the firm generates smoke free of charge. Note that this
contrasts with the case of solid waste in which a firm must pay for disposal.
Why is there this difference? The problem reduces to one of property rights.
Airborne effluent disperses across space. Particles and gases cross municipal,
state, regional, and national boundaries. These substances may affect multiple parties, raising concerns of public good. They may mingle with pollution
from other firms, making attribution unclear. And, in the case of long-lived
greenhouse gases, damages manifest many years after emission, obfuscating
the link between perpetrator and impact.
So, do environmental assets that serve as a repository for waste belong in a
system of national accounts? Well, if the case of solid waste disposal is used
as a conceptual guide, the answer is clearly yes. Long-term storage of waste
is a valuable service. And when property rights are established, firms (and
households) have revealed a willingness to pay for such services.
The remainder of this essay focuses on recent research that augments
standard measures of output to include damages from air and greenhouse
gas pollution into national output. It highlights aspects of current research
related to this topic and it then proposes new directions for this field of
inquiry.
AIR POLLUTION IN A SYSTEM OF EXTENDED ACCOUNTS
In its 1999 Nature’s Numbers report, the National Research Council of the
National Academies of Science (NAS NRC) identified air pollution damages
as a top priority for extending conventional measures of output:

4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

In the panel’s view, no other area of natural-resource and environmental
accounting would have as great an impact as the potential correction for air
quality. The magnitude of this impact indicates that the development of the
supplemental accounts for air quality is a high priority. (National Academy of
Sciences, National Research Council [NAS NRC], 1999, p. 148)

Part of the motivation for this statement was the finding by the United
States Environmental Protection Agency (USEPA) that the benefits from air
pollution control were on the order of $1 trillion (United States Environmental Protection Agency [USEPA], 1999). Another impetus for the Academies’
position was that the measures reported by USEPA were incomplete (NAS
NRC, 1999).
With the impetus for inclusion of air pollution damage into the NIPAs
established, the question of how to augment the accounts arises. As noted by
Abraham and Mackie (2006), one way in which pollution externalities may
be treated in the NIPAs is as a cost of production. Polluting firms require
disposal of residuals resulting from production. In the absence of property
rights on the atmosphere firms face no cost or fee for this disposal. The role of
an augmented accounting system is to provide a structure in which national
income reflects these costs. Such a framework would deduct these costs from
extant measures of output. In accordance with this treatment, recent work in
this area deducts pollution damage from conventional measures of output
(Bartelmus, 2009; Muller, 2014; Muller, Mendelsohn, & Nordhaus, 2011).
One concern with this approach is that these costs of disposal are already
in the accounts. However, prior research has shown that the vast majority
of air pollution damages are comprised of premature mortality risks, and,
therefore, they lie beyond the market boundary (Muller et al., 2011; USEPA,
1999, 2011).
The fact that firms are often not charged for the use of public waste sinks
such as air, water, and soils means that production, generally, is more costly
than is implied by measures such as GDP. This could, in principle, have two
effects. First, if the adverse consequences of using these sinks manifests in
markets, then it is only the distribution of GDP across sectors that is incorrectly measured. For example, if manufacturing firms producing smoke primarily affect firms producing crops, then the value of manufacturing output
is artificially high while that of crop production is suppressed. If, however,
the deleterious consequences of pollution occur outside of the market boundary, then both the allocation of GDP and its level are wrong. Making the former
correction to GDP alters the perceived productivity of different sectors in the
economy. Making the second deduction also means gross output is lower.
A third issue that was raised conceptually some 15 years ago is how measures of growth are affected by the augmentations discussed above. In terms

Environmental Accounting

5

of pollution, the effects on growth depend on the trajectory of damages. The
NAS NRC panel made the following statement when considering the effects
of a period of improving air quality:
The result might be a substantial increase in the estimate of growth of comprehensive consumption over this period. (NAS NRC, 1999, p. 147)

By the same logic, if air quality conditions deteriorate, augmented measures
of growth will fall relative to their market counterpart.
A final conceptual point worth noting is how total damages are calculated.
In estimating the total value of market output, GDP multiplies the current
price of each good by the quantity produced. In pursuit of a seamless integration of pollution damages into this extant system of accounts, measures
of total damage are typically tabulated by estimating marginal damage (the
extra harm caused by one more unit of emissions) and multiplying that by the
quantity of emissions produced (Abraham & Mackie, 2006; Nordhaus, 2006).
Thus, in the methodological section below, significant emphasis is placed on
the estimation of marginal damages.
HOW ARE DAMAGES MEASURED?
Over the past 15 years, many papers have developed advances in modeling
techniques that facilitate measuring air pollution damages (Banzhaf, Burtraw, & Palmer, 2004; Fann, Fulcher, & Hubbell, 2009; Heo, Adams, & Gao,
2016; Jaramillo & Muller, 2016; Kerl et al., 2015; Levy, Baxter, & Schwartz,
2009; Muller, 2014; Muller & Mendelsohn, 2007, 2009; Tessum, Hill, & Marshall, 2015; USEPA, 1999). Collectively, these papers suggest that (i) air pollution damages comprise a significant share of GDP, (ii) the majority of the air
pollution damages are comprised of adverse consequences on human health,
and (iii) damages hinge critically on two parameters that are discussed below.
Additionally, Muller (2014) finds that the changes in damages have an appreciable effect on measures of growth.
How are these damages calculated? Damages from air pollution are typically computed using integrated assessment models (IAMs). An IAM uses
findings from engineering, atmospheric chemistry and physics, epidemiology, and economics to link the production of residuals to final, monetized
consequences. Although the earliest examples of IAMs date back to the late
1970s and early 1980s (Mendelsohn, 1980), the recent literature has made a
number of advances in these devices (Fann et al., 2009; Heo et al., 2016; Kerl
et al., 2015; Muller & Mendelsohn, 2007, 2009; Tessum et al., 2015).
Most IAMs focusing on air pollution connect emissions of common air pollutants to estimates of ambient concentrations, exposures, physical impacts,

6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

and monetary damage so that the impacts can be incorporated directly into
the national accounts. The air pollutants covered often include nitrogen
oxides (NOx ), sulfur dioxide (SO2 ), fine particulate matter (PM2.5 ), ammonia
(NH3 ), and volatile organic compounds (VOCs). The first step in an IAM
for air pollution relies on emissions data, which reveal where, when, and
in what quantities these substances are produced. In developed economies,
public environmental regulatory agencies typically provide these data.
(Often these agencies are charged with environmental enforcement, which
depends on tracking emissions. So they have the data anyway.)
In the United States, emissions data are reported by the USEPA. In an
IAM, these data are fed through an air quality model that links emissions to
ambient concentrations. Exposures are tabulated by tracking populations of
sensitive receptors (human populations and crops, for example). Physical
impacts, such as reduced crop yields and increased rates of illness, are
estimated through the use of concentration–response functions gleaned
from peer-reviewed publications. Valuation uses either reported market
prices (for impacts on crops) or non-market valuation techniques.
As mentioned above, most environmental accounting exercises tabulate
total damage as the product of marginal damages and emissions. One
of the advances made in the design of IAMs lies in spatial resolution.
Source-specific damages are now computed for particular power plants and
manufacturing facilities. For example, recent work indicates that the per ton
damage from emissions of sulfur dioxide (a common air pollutant produced
when coal is burned) varies by over 100 times within the contiguous United
States (Muller & Mendelsohn, 2009, 2012). Figure 1 shows this stunning
range of impacts. Because the impact of emissions of local air pollution
varies a lot according to where it is emitted, the enhanced resolution is a
critical advance. After all, if total damages are computed as the product
of marginal damage times total emissions, then getting the total damage
figure right depends on having the margins correctly estimated. And,
if the marginal damages vary drastically across space, then an accurate
estimate of total damage depends on having source-specific marginal
damages.
Figure 1 also indicates that the largest damages on a per ton basis are
due to emissions produced in cities. This stems from the importance of
human health effects in pollution damage. The robust finding that the
largest share of damage is due to human health impacts depends crucially
(as noted above) on two parameters in the IAMs: the effect that exposures
to fine particles have on adult mortality rates, and the value attributed
to small changes in mortality risk. Because of the importance of these
parameters in tabulating damages, both are discussed in the following
paragraphs.

Environmental Accounting

7

Canada

Atlantic
Ocean

Damage ($/ton)
0 – 500

2500 – 5000

500 – 1000

5000 – 7500

1000 – 2500

7500 – 15,000

Pacific
Ocean

Figure 1

15,000 – 57,000

Mexico
Gulf of Mexico

SO2 marginal damages. Source: Muller and Mendelsohn (2012).

Traditionally, evidence about the effect of exposure to pollution on human
health is obtained from peer-reviewed research in the epidemiological literature. And, in terms of the effects of exposure to fine particles on premature
mortality in the United States, two studies are most frequently used (Krewski
et al., 2009; Lepeule, Laden, Dockery, & Schwartz, 2012). Damages computed
using these two studies (holding all else in the IAM fixed) differ by more than
a factor of two. How do policymakers or academic researchers conducting
environmental accounting resolve these differences? Typically, both parameters are used in a sensitivity analysis in order to bound or bracket damage
estimates (USEPA, 1999, 2011).
A second point related to this issue is worthy of discussion; recent research
in environmental economics has brought more sophisticated econometric
techniques to the question of how exposure to environmental pollutants
affects mortality rates. This research has raised two critical questions. First,
do the epidemiological studies effectively control for a range of pollutants
such that regulators are confident that risks attributed to particulate matter
are, in fact, due to exposures to that pollutant (Beatty & Shimshack, 2011;
Currie, Neidell, & Schmeider, 2009)? Second, is the magnitude of the effect
that particulate matter has on mortality rates different when researchers use
quasi-experimental models that estimate a causal effect of particulates on
mortality rates (Chay & Greenstone, 2003)? Because mortality effects play

8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

such a large role in the damages from air pollution, new findings on whether
the toxicity of particulate matter holds up to different empirical methods
have potentially significant ramifications for integration of pollution damage
into national accounts.
For a number of years, the most common approach to valuing mortality
risk uses the value of a statistical life (VSL). One common empirical approach
to estimating the VSL uses hedonic wage models to identify the marginal
implicit price of risk of death on the job (Viscusi & Aldy, 2003). The use
of these results to value risks due to air pollution exposure is controversial for several reasons. First, hedonic wage studies rely on a sample of
people from the workplace. Hence, they neglect the elderly and the very
young. This is problematic because most of the premature deaths from air
pollution manifest among the elderly and infants. The VSL is extrapolated
to these non-working populations. Second, risks on the job are typically,
though not always, acute or accidental events. In contrast, risks from air
pollution involve chronic, long-term exposure. Thus, the nature of the risks
differs.
In part because of the concerns related to transferring risk valuation
estimates from hedonic wage studies to the context of air pollution, environmental economists have estimated VSLs in different contexts. For example,
employing variation in vehicle air-bag requirements, researchers have
found a range of VSLs quite similar to those estimated from hedonic wage
models (Rohlfs, Sullivan, & Kneisner, 2015). Other authors explore trade-offs
between risks of death while driving and time savings implied by speed limits (Ashenfelter & Greenstone, 2004). In addition, economists have exploited
data from the Alaskan crab industry to estimate a VSL (Schneir, Horrace,
& Felthoven, 2009). Environmental economists are actively engaged in
exploring new ways to elicit the trade-off between money and risk, which is
very important in the estimation of pollution damage.
RECENT RESULTS AND FUTURE DIRECTIONS
A number of recent papers have contributed to the empirical environmental accounting literature that specifically focuses on air pollution damage
(Bartelmus, 2009; Muller et al., 2011; Muller, 2014, 2016; Jaramillo & Muller,
2016). Bartelmus (2009) estimated the costs due to consumption of environmental capital in various parts of the world for a collection of years between
1990 and 2006. In the European Union environmental costs were under 1% of
GDP. In the United States this figure was under 3%. Intuitively, environmental costs were much higher in developing economies. In China, Bartelmus
estimated environmental impact at up to 12% of GDP in 2006. In Africa as
a whole, this metric was over 25%. These values, while highly uncertain,

Environmental Accounting

9

are illustrative in the cross section and the time series dimensions. In every
region covered, environmental costs rise over time.
While Bartelmus’ work is highly aggregated, a more recent paper drilled
down in great detail in the U.S. economy. Muller et al. (2011) calculated the
air pollution damage for every industry in the U.S. economy. Why should
an economic system be explored at this level of detail? A modern, developed economy consists of a wide range of productive activities: from heavy
manufacturing at iron foundries to financial services and insurance. Accordingly, pollution intensity varies dramatically across sectors; the importance
of augmenting national accounts varies across sectors as well.
In the United States in 2002, pollution damage as a fraction of sector value
added ranged from 0.38 for agriculture/forestry to 0.01 for manufacturing,
and to 0 for sectors such as finance and insurance (Muller et al., 2011).
Within sectors, great variation in pollution intensity was also reported. For
example, while electric power generation using coal produced damages over
two times larger than value-added, natural-gas-based power generation
produced damages less than 10% of value-added (Muller et al., 2011).
What does it mean to have an industry generate greater environmental
pollution damage than its value-added? It may be tempting to say that such
industries are doing more harm than good and therefore should be shut
down. That position is incorrect. If the external costs were internalized by
the firm (perhaps through an efficiently calibrated emission tax) it is quite
likely that output produced by the industry would fall. That is, if no viable
end-of-pipe pollution controls exist (as is the case for CO2 ), then the firm
would either switch to cleaner, more expensive inputs, or they would reduce
emissions by reducing production. If abatement technology exists, then they
would substitute away from some productive inputs toward abatement
devices. In either case output will fall, ceteris paribus. If the entire sector or
industry adopts such responses to internalized costs, then prices will adjust
and value-added will rise. Concurrently damages will fall and the damage
to value-added ratios will fall.
A particularly exciting area of research in this space focuses on measures of
growth in GDP versus growth in measures of output that deduct pollution
damage. Why is this focus on growth? First, the primary value of measures
such as GDP lies in examining changes over time. Academics, policymakers,
and the general public draw inferences regarding national economic performance from changes in GDP rather than levels. As such, a pressing question
in environmental economics is how inclusion of pollution damage into an
augmented system of accounts affects comprehensive growth rates. Second,
from the perspective of public policy, macroeconomic regulatory agencies
often base policy choices on quarterly or annual changes in aggregates such
as GDP. Prior rates of change in GDP are clearly linked to monetary and fiscal

10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

policy; one needs to look no further than the Federal Reserve’s response to
the Great Recession for evidence of this claim. In principle, measures of augmented growth could be used in the same way with one additional dimension. Beyond the standard monetary policy toolkit, macroeconomic regulators basing policy choices on the adjusted output measure could prescribe
reductions to damage as a means to enhance growth. That is, if pollution
damage has a significant effect on augmented measures of growth, as the
NAS NRC panel suggested, does this not merit consideration of pollution
control as a viable lever for growth enhancements?
Recent research using data from the United States suggests that the difference between conventional and environmentally adjusted growth rates is
appreciable. For example, over the period 1999 to 2011, GDP in the United
States expanded at annualized rates of between 1.2% and 2.8%. When corrected for both air pollution and carbon dioxide (CO2 ) damage, national output grew by about 0.3 percentage points faster (Muller, 2014, 2016).
Drilling down further, from 2005 to 2008, the U.S. economy was barreling
toward the Great Recession. Annual GDP growth over this period fell to just
over 1%. However, as conventional measures of growth receded, air pollution
damages did too. Thus, the measure that recognized this reduction in damage
expanded more rapidly. In particular, from 2005 to 2008, growth adjusted to
account for air pollution and CO2 emissions was about 0.4% more rapid than
growth in GDP (Muller, 2016).
The contribution of these results is twofold. First, periods of improvement
in air quality can have an appreciable and positive effect on growth. The
intuition of the NAS NRC panel cited above was confirmed. Second, these
results highlight Stiglitz’ position that focusing on market-centric measures
of output may foster the illusion of a trade-off between investments in environmental quality and growth in market production.
The difference in growth discussed above is largest between 1999 and 2002.
One reason for this finding is that Phase II of the Acid Rain Program began in
2000. Because of this policy, emissions of SO2 from regulated power plants fell
by 2.3 million tons between 1999 and 2002. Importantly, costs of compliance
with the regulation are embedded in GDP, while reduced damages are not.
The finding that the benefits of pollution control are large enough to affect
national growth during this time period suggests the need to amend official
measures of output to include the benefits from environmental policy.
Exciting work in this space is occurring in areas beyond air pollution and
greenhouse gas emissions. For example, research is currently ongoing to construct an IAM for water pollution in the United States with resolution akin
to that of modern air pollution IAMs. Once operational, a model such as
this will facilitate, among other topics, the ability to engage in multimedium
pollution control analyses. For example, in modern economies one central

Environmental Accounting

11

aspect of water pollution control is municipal sewage treatment. While on
the surface this appears to be beneficial, sewage treatment generates copious
amounts of airborne ammonia emissions. Ammonia is an important catalyst
in the formation of secondary fine particulate matter, which elevates mortality risk. What is the value of waste removal from water? What is the damage
caused by airborne emissions from this treatment? What is the optimal mix
of air and water pollution emissions from sewage treatment? Intriguing and
as of yet unanswered questions such as these will form the basis of future
research in the field of environmental economics. And, as an ever greater
range of pollution damages are added to the environmental accounts, our
measures of economic performance will do a better job of measuring welfare.

REFERENCES
Abraham, K. G., & Mackie, C. (2006). A framework for nonmarket accounting. In D.
W. Jorgensen, J. S. Landefeld & W. D. Nordhaus (Eds.), A new architecture for the
U.S. national accounts. NBER studies in income and wealth (Vol. 66). Chicago, IL: The
University of Chicago Press.
Ashenfelter, O., & Greenstone, M. (2004). Using mandated speed limits to measure
the value of a statistical life. Journal of Political Economy, 112(2), S226–S267.
Banzhaf, H. S., Burtraw, D., & Palmer, K. (2004). Efficient emission fees in the US
electricity sector. Resource and Energy Economics, 26, 317–341.
Bartelmus, P. (2009). The cost of natural capital consumption: Accounting for a sustainable world economy. Ecological Economics, 68, 1850–1857.
Beatty, T., & Shimshack, J. (2011). School buses, diesel emissions, and respiratory
health. Journal of Health Economics, 30(5), 987–999.
Chay, K. Y., & Greenstone, M. (2003). The impact of air pollution on infant mortality:
Evidence from geographic variation in pollution shocks induced by a recession.
Quarterly Journal of Economics, 118(3), 1121–1167.
Currie, J., Neidell, M., & Schmeider, J. F. (2009). Air pollution and infant health:
Lessons from New Jersey. Journal of Health Economics, 28, 688–703.
Fann, N., Fulcher, C. M., & Hubbell, B. J. (2009). The influence of location, source and
emission type in estimates of the human health benefits of reducing a ton of air
pollution. Air Quality, Atmosphere and Health, 2, 169–176.
Hecht, J. E. (2005). National environmental accounting. Bridging the gap between ecology
and economy. Washington, DC: Resources for the Future Press.
Heo, J., Adams, P. J., & Gao, H. O. (2016). Public Health Costs of Primary PM2.5 and
Inorganic PM2.5 Precursor Emissions in the United States. Environmental Science
& Technology. doi:10.1021/acs.est.5b06125.
Jaramillo, P., & Muller, N. Z. (2016). The air pollution damage from energy production in the U.S.: 2002–2011. Energy Policy, 90, 202–211.
Kerl, P. Y., Zhang, W., Moreno-Cruz, J. B., Nenes, A., Realff, M. J., Russell, A. G.,
… Thomas, V. M. (2015). A new approach for optimal electricity planning and

12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

dispatching with hourly time-scale air quality and health considerations. Proceedings of the National Academy of Sciences, 112(35), 10884–10889.
Krewski, D., Jerrett, M., Burnett, R.T., Ma, R., Hughes, E., Shi, Y., … Tempalski, B.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study
linking particulate air pollution and mortality. HEI Research Report, 140, Health
Effects Institute, Boston, MA.
Landefeld, J. S. (2000). GDP: One of the great inventions of the 20th century. Survey
of Current Business. January, 6–14.
Lepeule, J., Laden, F., Dockery, D., & Schwartz, J. (2012). Chronic exposure to fine
particles and mortality: An extended follow-up of the Harvard Six Cities Study
from 1974 to 2009. Environmental Health Perspectives, 120(7), 965–970.
Levy, J. I., Baxter, L. K., & Schwartz, J. (2009). Uncertainty and variability in
health-related damages from coal-fired power plants in the United States. Risk
Analysis, 29(7), 1000–1014.
Mendelsohn, R. O. (1980). An economic analysis of air pollution from coal-fired
power plants. Journal of Environmental Economics and Management, 7, 30–43.
Muller, N. Z. (2014). Boosting GDP growth by accounting for the environment. Science, 345(6199), 873–874.
Muller, N.Z. (2016). The derivation of discount rates using an augmented measure of
income. NBER WP #22579.
Muller, N. Z., & Mendelsohn, R. O. (2007). Measuring the damages due to air pollution in the United States. Journal of Environmental Economics and Management, 54,
1–14.
Muller, N. Z., & Mendelsohn, R. O. (2009). Efficient pollution regulation: Getting the
prices right. American Economic Review, 99(5), 1714–1739.
Muller, N. Z., & Mendelsohn, R. O. (2012). Efficient pollution regulation: Getting
the prices right: Mortality rate update (Corrigendum). American Economic Review,
102(1), 613–616.
Muller, N. Z., Mendelsohn, R. O., & Nordhaus, W. D. (2011). Environmental accounting for pollution in the U.S. economy. American Economic Review, 101(5), 1649–1675.
National Academy of Sciences, National Research Council (1999). W. D. Nordhaus
& E. C. Kokkelenberg (Eds.), Nature’s numbers: Expanding the U.S. national economic
accounts to include the environment. Washington, D.C.: National Academy Press.
Nordhaus, W. D., & Tobin, J. (1972). Is growth obsolete? Studies in income and wealth
(Vol. 38). New York, NY: NBER.
Nordhaus, W. D. (2006). Principles of national accounting for non-market accounts.
In D. W. Jorgensen, J. S. Landefeld & W. D. Nordhaus (Eds.), A new architecture for
the U.S. National Accounts. NBER Ssudies in income and wealth (Vol. 66). Chicago, IL:
The University of Chicago Press.
Pigou, A. C. (1932). The economics of welfare. London, England: Macmillan and Co.,
Limited.
Rohlfs, C., Sullivan, R., & Kneisner, T. (2015). New estimates of the value of a statistical life using air bag regulations as a quasi-experiment. American Economic Journal:
Economic Policy, 7(1), 331–359.

Environmental Accounting

13

Schneir, K.E. Horrace, W.C., & Felthoven, R.G. (2009). The value of a statistical life:
Pursuing the deadliest catch. http://surface.syr.edu/cpr/49/
Stiglitz, J. E. (2009). GDP fetishism. The Economist’s Voice. September.
Tessum, C. W., Hill, J. D., & Marshall, J. D. (2015). InMAP: A new model
for air pollution interventions. Geoscientific Model Development–Discussions.
doi:10.5194/gmdd-8-9281-2015.
United States Environmental Protection Agency. (1999). The benefits and costs of the
Clean Air Act: 1990–2010. EPA Report to Congress. EPA 410-R-99-001, Office of Air
and Radiation, Office of Policy, Washington, D.C.
United States Environmental Protection Agency. (2011). The benefits and costs of the
Clean Air Act: 1990–2020. Final Report. Office of Air and Radiation, Office of Policy,
Washington, D.C.
Viscusi, W. K., & Aldy, J. E. (2003). The value of a statistical life: A critical review of
market estimates throughout the world. Journal of Risk Uncertainty, 27(1), 5–76.

NICK Z. MULLER SHORT BIOGRAPHY
Nick Z. Muller joined the Middlebury College economics faculty as an
assistant professor in the fall of 2007. He completed his dissertation at Yale
University in May of 2007, where his advisors included Robert Mendelsohn,
William Nordhaus, and Nathaniel Keohane. His dissertation focused on
using integrated assessment models to measure the damages from air
pollution in the United States, and to propose alternative market-based
policies intended to govern the criteria for air pollutants. Dr. Muller also
attended the School of Public and Environmental Affairs where he studied
environmental policy and public finance in pursuit of a master’s degree in
public administration. His current research includes measuring the damages
due to emissions from each industry in the U.S. economy, the design of
market-based environmental policies, and the construction price indices
for air pollution. His research has been published in the American Economic
Review, Science, the Journal of Environmental Economics and Management, and
Environmental Science & Technology, among other outlets.
RELATED ESSAYS
Sociological Theory After the End of Nature (Sociology), Robert J. Brulle
Misinformation and How to Correct It (Psychology), John Cook et al.
Cities and Sustainable Development (Sociology), Christopher Cusack
Theorizing the Death of Cities (Political Science), Peter Eisinger
Architecture of Markets (Sociology), Neil Fligstein and Ryan Calder
Modeling Coal and Natural Gas Markets (Economics), Franziska Holz
Presidential Power (Political Science), William G. Howell
Why Do States Sign Alliances? (Political Science), Brett Ashley Leeds

14

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Money in Politics (Political Science), Jeffrey Milyo
Economics of Renewable Energy Production (Economics), Gregory F. Nemet
The Politics of Disaster Relief (Political Science), Alexander J. Oliver and
Andrew Reeves
The Social Science of Sustainability (Political Science), Johannes Urpelainen


Environmental Accounting
NICHOLAS Z. MULLER

Abstract
While metrics such as gross domestic product (GDP) constitute important means to
gauge the value of production, it is widely recognized that indices that focus exclusively on market production are incomplete. Omitted environmental assets include
(i) those that have the capacity to act as a source of valuable inputs to production
such as timber, subsurface minerals, or fisheries; and (ii) media that serve as a sink
for anthropogenic residuals such as air, water, or soil. The human health impacts
depend crucially on two parameters in the integrated asset models (IAMs): the effect
that exposures to fine particles have on adult mortality rates and the value attributed
to small changes in mortality risk. This essay focuses on recent research that augments standard measures of output to include damages from air and greenhouse
gas pollution into national output.

INTRODUCTION
The development of a standardized set of tools to track national income and
output remains an important achievement in the field of economics.
While the GDP and the rest of the national income accounts may seem to be
arcane concepts, they are truly among the great inventions of the twentieth
century. Samuelson, P.A. and Nordhaus, W.D. (as quoted in Landefeld, 2000).

Some 80 years after their development, tools such as gross domestic product (GDP) are used by nations and governments throughout the world to
estimate economic performance.
The value in such mensuration is manifold. Time series measurement of a
particular nation’s output, expressed in real terms, provides society with a
sense of trends in output. Cross-sectional comparisons on purchasing power
parity bases facilitate assessment of the relative wealth of nations. The
impact of policy interventions, cyclical or episodic downturns, international
or civil conflict, among other factors, is gauged with metrics such as GDP.

Emerging Trends in the Social and Behavioral Sciences.
Robert Scott and Marlis Buchmann (General Editors) with Stephen Kosslyn (Consulting Editor).
© 2017 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.

1

2

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Conversely, policymakers rely on national output statistics when designing
fiscal and monetary policies.
While metrics such as GDP constitute important means to gauge the value
of production, it is widely recognized, and has been for some time, that
indices that focus exclusively on market production are incomplete. In the
most basic sense, GDP focuses on goods transacted in markets. The problem
then is that much economic activity occurs outside of formal markets.
Although Pigou (1932) arguably first tackled the problem of external
economies, Nordhaus and Tobin (1972) were the first to point this out in
the context of the national accounts. Emphasizing aspects of production
outside the purview of GDP such as leisure time, home production, and
environmental goods and services, Nordhaus and Tobin (1972) reported
that in 1965, their proposed (more comprehensive) measure differed from
standard measures by a factor of two. Clearly, they raised an important set
of questions.
More recently, Stiglitz (2009) argued, in addition to recognizing that GDP is
incomplete, that policies that formulate objectives around GDP may be counterproductive.
If we have poor measures, what we strive to do (say, increase GDP) may actually contribute to a worsening of living standards. (Stiglitz, 2009)

Thus, it is the use of GDP to guide national priorities that is problematic.
One of the particular concerns raised by Stiglitz (2009), which was also part of
Nordhaus and Tobin’s proposed extension to GDP, was the value of environmental assets. Using a measure of national output that omits natural capital,
Stiglitz argues that:
[w]e may also be confronted with false choices, seeing trade-offs between output and environmental protection that don’t exist. (2009)

What aspects of the environment are omitted from standard measures of
national output? The literature in this area typically characterizes environmental assets as (i) those that have the capacity to act as a source of valuable
inputs to production such as timber, subsurface minerals, or fisheries; and
(ii) media that serve as a sink for anthropogenic residuals such as air, water,
or soil (Hecht, 2005).
Clearly, one may argue that the value of timber for building or making
paper and minerals such as coal, oil, and gas are embedded in GDP. This is
only partially true. These assets appear in GDP only when privately owned
and when they provide (or may generate in the future) financial benefit to
owners. Publics lands, and resources therein, as well as wild areas and native
species are assets not tracked by the conventional accounts (Hecht, 2005).

Environmental Accounting

3

Conventional measures provide an incomplete glimpse at the value of natural resources.
Are assets such as air and water really valuable? That is, do they belong in
an augmented measure of national output? To explore this important question, consider how a firm disposes of nontoxic solid waste. Typically, the firm
pays a fee to a carting service to transport the waste to either a landfill or an
incinerator. Both destinations typically require fees to dispose of the waste.
The important point here is that there are payments for waste disposal. Why
is that the case?
This is because for a landfill, property rights for land are well defined. Even
without policy constraints, a firm would not be able to simply dump waste
on property that someone else owns. That is a damaging act. When we do
observe illegal dumping it tends to be in vacant lots or other properties that
are not closely monitored by owners, or the ownership of which is ambiguous. Firms must pay to dispose of solid waste. It consumes valuable space. It
is a cost of doing business.
Now consider a firm that produces smoke as it makes its products. In the
absence of regulation, the firm generates smoke free of charge. Note that this
contrasts with the case of solid waste in which a firm must pay for disposal.
Why is there this difference? The problem reduces to one of property rights.
Airborne effluent disperses across space. Particles and gases cross municipal,
state, regional, and national boundaries. These substances may affect multiple parties, raising concerns of public good. They may mingle with pollution
from other firms, making attribution unclear. And, in the case of long-lived
greenhouse gases, damages manifest many years after emission, obfuscating
the link between perpetrator and impact.
So, do environmental assets that serve as a repository for waste belong in a
system of national accounts? Well, if the case of solid waste disposal is used
as a conceptual guide, the answer is clearly yes. Long-term storage of waste
is a valuable service. And when property rights are established, firms (and
households) have revealed a willingness to pay for such services.
The remainder of this essay focuses on recent research that augments
standard measures of output to include damages from air and greenhouse
gas pollution into national output. It highlights aspects of current research
related to this topic and it then proposes new directions for this field of
inquiry.
AIR POLLUTION IN A SYSTEM OF EXTENDED ACCOUNTS
In its 1999 Nature’s Numbers report, the National Research Council of the
National Academies of Science (NAS NRC) identified air pollution damages
as a top priority for extending conventional measures of output:

4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

In the panel’s view, no other area of natural-resource and environmental
accounting would have as great an impact as the potential correction for air
quality. The magnitude of this impact indicates that the development of the
supplemental accounts for air quality is a high priority. (National Academy of
Sciences, National Research Council [NAS NRC], 1999, p. 148)

Part of the motivation for this statement was the finding by the United
States Environmental Protection Agency (USEPA) that the benefits from air
pollution control were on the order of $1 trillion (United States Environmental Protection Agency [USEPA], 1999). Another impetus for the Academies’
position was that the measures reported by USEPA were incomplete (NAS
NRC, 1999).
With the impetus for inclusion of air pollution damage into the NIPAs
established, the question of how to augment the accounts arises. As noted by
Abraham and Mackie (2006), one way in which pollution externalities may
be treated in the NIPAs is as a cost of production. Polluting firms require
disposal of residuals resulting from production. In the absence of property
rights on the atmosphere firms face no cost or fee for this disposal. The role of
an augmented accounting system is to provide a structure in which national
income reflects these costs. Such a framework would deduct these costs from
extant measures of output. In accordance with this treatment, recent work in
this area deducts pollution damage from conventional measures of output
(Bartelmus, 2009; Muller, 2014; Muller, Mendelsohn, & Nordhaus, 2011).
One concern with this approach is that these costs of disposal are already
in the accounts. However, prior research has shown that the vast majority
of air pollution damages are comprised of premature mortality risks, and,
therefore, they lie beyond the market boundary (Muller et al., 2011; USEPA,
1999, 2011).
The fact that firms are often not charged for the use of public waste sinks
such as air, water, and soils means that production, generally, is more costly
than is implied by measures such as GDP. This could, in principle, have two
effects. First, if the adverse consequences of using these sinks manifests in
markets, then it is only the distribution of GDP across sectors that is incorrectly measured. For example, if manufacturing firms producing smoke primarily affect firms producing crops, then the value of manufacturing output
is artificially high while that of crop production is suppressed. If, however,
the deleterious consequences of pollution occur outside of the market boundary, then both the allocation of GDP and its level are wrong. Making the former
correction to GDP alters the perceived productivity of different sectors in the
economy. Making the second deduction also means gross output is lower.
A third issue that was raised conceptually some 15 years ago is how measures of growth are affected by the augmentations discussed above. In terms

Environmental Accounting

5

of pollution, the effects on growth depend on the trajectory of damages. The
NAS NRC panel made the following statement when considering the effects
of a period of improving air quality:
The result might be a substantial increase in the estimate of growth of comprehensive consumption over this period. (NAS NRC, 1999, p. 147)

By the same logic, if air quality conditions deteriorate, augmented measures
of growth will fall relative to their market counterpart.
A final conceptual point worth noting is how total damages are calculated.
In estimating the total value of market output, GDP multiplies the current
price of each good by the quantity produced. In pursuit of a seamless integration of pollution damages into this extant system of accounts, measures
of total damage are typically tabulated by estimating marginal damage (the
extra harm caused by one more unit of emissions) and multiplying that by the
quantity of emissions produced (Abraham & Mackie, 2006; Nordhaus, 2006).
Thus, in the methodological section below, significant emphasis is placed on
the estimation of marginal damages.
HOW ARE DAMAGES MEASURED?
Over the past 15 years, many papers have developed advances in modeling
techniques that facilitate measuring air pollution damages (Banzhaf, Burtraw, & Palmer, 2004; Fann, Fulcher, & Hubbell, 2009; Heo, Adams, & Gao,
2016; Jaramillo & Muller, 2016; Kerl et al., 2015; Levy, Baxter, & Schwartz,
2009; Muller, 2014; Muller & Mendelsohn, 2007, 2009; Tessum, Hill, & Marshall, 2015; USEPA, 1999). Collectively, these papers suggest that (i) air pollution damages comprise a significant share of GDP, (ii) the majority of the air
pollution damages are comprised of adverse consequences on human health,
and (iii) damages hinge critically on two parameters that are discussed below.
Additionally, Muller (2014) finds that the changes in damages have an appreciable effect on measures of growth.
How are these damages calculated? Damages from air pollution are typically computed using integrated assessment models (IAMs). An IAM uses
findings from engineering, atmospheric chemistry and physics, epidemiology, and economics to link the production of residuals to final, monetized
consequences. Although the earliest examples of IAMs date back to the late
1970s and early 1980s (Mendelsohn, 1980), the recent literature has made a
number of advances in these devices (Fann et al., 2009; Heo et al., 2016; Kerl
et al., 2015; Muller & Mendelsohn, 2007, 2009; Tessum et al., 2015).
Most IAMs focusing on air pollution connect emissions of common air pollutants to estimates of ambient concentrations, exposures, physical impacts,

6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

and monetary damage so that the impacts can be incorporated directly into
the national accounts. The air pollutants covered often include nitrogen
oxides (NOx ), sulfur dioxide (SO2 ), fine particulate matter (PM2.5 ), ammonia
(NH3 ), and volatile organic compounds (VOCs). The first step in an IAM
for air pollution relies on emissions data, which reveal where, when, and
in what quantities these substances are produced. In developed economies,
public environmental regulatory agencies typically provide these data.
(Often these agencies are charged with environmental enforcement, which
depends on tracking emissions. So they have the data anyway.)
In the United States, emissions data are reported by the USEPA. In an
IAM, these data are fed through an air quality model that links emissions to
ambient concentrations. Exposures are tabulated by tracking populations of
sensitive receptors (human populations and crops, for example). Physical
impacts, such as reduced crop yields and increased rates of illness, are
estimated through the use of concentration–response functions gleaned
from peer-reviewed publications. Valuation uses either reported market
prices (for impacts on crops) or non-market valuation techniques.
As mentioned above, most environmental accounting exercises tabulate
total damage as the product of marginal damages and emissions. One
of the advances made in the design of IAMs lies in spatial resolution.
Source-specific damages are now computed for particular power plants and
manufacturing facilities. For example, recent work indicates that the per ton
damage from emissions of sulfur dioxide (a common air pollutant produced
when coal is burned) varies by over 100 times within the contiguous United
States (Muller & Mendelsohn, 2009, 2012). Figure 1 shows this stunning
range of impacts. Because the impact of emissions of local air pollution
varies a lot according to where it is emitted, the enhanced resolution is a
critical advance. After all, if total damages are computed as the product
of marginal damage times total emissions, then getting the total damage
figure right depends on having the margins correctly estimated. And,
if the marginal damages vary drastically across space, then an accurate
estimate of total damage depends on having source-specific marginal
damages.
Figure 1 also indicates that the largest damages on a per ton basis are
due to emissions produced in cities. This stems from the importance of
human health effects in pollution damage. The robust finding that the
largest share of damage is due to human health impacts depends crucially
(as noted above) on two parameters in the IAMs: the effect that exposures
to fine particles have on adult mortality rates, and the value attributed
to small changes in mortality risk. Because of the importance of these
parameters in tabulating damages, both are discussed in the following
paragraphs.

Environmental Accounting

7

Canada

Atlantic
Ocean

Damage ($/ton)
0 – 500

2500 – 5000

500 – 1000

5000 – 7500

1000 – 2500

7500 – 15,000

Pacific
Ocean

Figure 1

15,000 – 57,000

Mexico
Gulf of Mexico

SO2 marginal damages. Source: Muller and Mendelsohn (2012).

Traditionally, evidence about the effect of exposure to pollution on human
health is obtained from peer-reviewed research in the epidemiological literature. And, in terms of the effects of exposure to fine particles on premature
mortality in the United States, two studies are most frequently used (Krewski
et al., 2009; Lepeule, Laden, Dockery, & Schwartz, 2012). Damages computed
using these two studies (holding all else in the IAM fixed) differ by more than
a factor of two. How do policymakers or academic researchers conducting
environmental accounting resolve these differences? Typically, both parameters are used in a sensitivity analysis in order to bound or bracket damage
estimates (USEPA, 1999, 2011).
A second point related to this issue is worthy of discussion; recent research
in environmental economics has brought more sophisticated econometric
techniques to the question of how exposure to environmental pollutants
affects mortality rates. This research has raised two critical questions. First,
do the epidemiological studies effectively control for a range of pollutants
such that regulators are confident that risks attributed to particulate matter
are, in fact, due to exposures to that pollutant (Beatty & Shimshack, 2011;
Currie, Neidell, & Schmeider, 2009)? Second, is the magnitude of the effect
that particulate matter has on mortality rates different when researchers use
quasi-experimental models that estimate a causal effect of particulates on
mortality rates (Chay & Greenstone, 2003)? Because mortality effects play

8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

such a large role in the damages from air pollution, new findings on whether
the toxicity of particulate matter holds up to different empirical methods
have potentially significant ramifications for integration of pollution damage
into national accounts.
For a number of years, the most common approach to valuing mortality
risk uses the value of a statistical life (VSL). One common empirical approach
to estimating the VSL uses hedonic wage models to identify the marginal
implicit price of risk of death on the job (Viscusi & Aldy, 2003). The use
of these results to value risks due to air pollution exposure is controversial for several reasons. First, hedonic wage studies rely on a sample of
people from the workplace. Hence, they neglect the elderly and the very
young. This is problematic because most of the premature deaths from air
pollution manifest among the elderly and infants. The VSL is extrapolated
to these non-working populations. Second, risks on the job are typically,
though not always, acute or accidental events. In contrast, risks from air
pollution involve chronic, long-term exposure. Thus, the nature of the risks
differs.
In part because of the concerns related to transferring risk valuation
estimates from hedonic wage studies to the context of air pollution, environmental economists have estimated VSLs in different contexts. For example,
employing variation in vehicle air-bag requirements, researchers have
found a range of VSLs quite similar to those estimated from hedonic wage
models (Rohlfs, Sullivan, & Kneisner, 2015). Other authors explore trade-offs
between risks of death while driving and time savings implied by speed limits (Ashenfelter & Greenstone, 2004). In addition, economists have exploited
data from the Alaskan crab industry to estimate a VSL (Schneir, Horrace,
& Felthoven, 2009). Environmental economists are actively engaged in
exploring new ways to elicit the trade-off between money and risk, which is
very important in the estimation of pollution damage.
RECENT RESULTS AND FUTURE DIRECTIONS
A number of recent papers have contributed to the empirical environmental accounting literature that specifically focuses on air pollution damage
(Bartelmus, 2009; Muller et al., 2011; Muller, 2014, 2016; Jaramillo & Muller,
2016). Bartelmus (2009) estimated the costs due to consumption of environmental capital in various parts of the world for a collection of years between
1990 and 2006. In the European Union environmental costs were under 1% of
GDP. In the United States this figure was under 3%. Intuitively, environmental costs were much higher in developing economies. In China, Bartelmus
estimated environmental impact at up to 12% of GDP in 2006. In Africa as
a whole, this metric was over 25%. These values, while highly uncertain,

Environmental Accounting

9

are illustrative in the cross section and the time series dimensions. In every
region covered, environmental costs rise over time.
While Bartelmus’ work is highly aggregated, a more recent paper drilled
down in great detail in the U.S. economy. Muller et al. (2011) calculated the
air pollution damage for every industry in the U.S. economy. Why should
an economic system be explored at this level of detail? A modern, developed economy consists of a wide range of productive activities: from heavy
manufacturing at iron foundries to financial services and insurance. Accordingly, pollution intensity varies dramatically across sectors; the importance
of augmenting national accounts varies across sectors as well.
In the United States in 2002, pollution damage as a fraction of sector value
added ranged from 0.38 for agriculture/forestry to 0.01 for manufacturing,
and to 0 for sectors such as finance and insurance (Muller et al., 2011).
Within sectors, great variation in pollution intensity was also reported. For
example, while electric power generation using coal produced damages over
two times larger than value-added, natural-gas-based power generation
produced damages less than 10% of value-added (Muller et al., 2011).
What does it mean to have an industry generate greater environmental
pollution damage than its value-added? It may be tempting to say that such
industries are doing more harm than good and therefore should be shut
down. That position is incorrect. If the external costs were internalized by
the firm (perhaps through an efficiently calibrated emission tax) it is quite
likely that output produced by the industry would fall. That is, if no viable
end-of-pipe pollution controls exist (as is the case for CO2 ), then the firm
would either switch to cleaner, more expensive inputs, or they would reduce
emissions by reducing production. If abatement technology exists, then they
would substitute away from some productive inputs toward abatement
devices. In either case output will fall, ceteris paribus. If the entire sector or
industry adopts such responses to internalized costs, then prices will adjust
and value-added will rise. Concurrently damages will fall and the damage
to value-added ratios will fall.
A particularly exciting area of research in this space focuses on measures of
growth in GDP versus growth in measures of output that deduct pollution
damage. Why is this focus on growth? First, the primary value of measures
such as GDP lies in examining changes over time. Academics, policymakers,
and the general public draw inferences regarding national economic performance from changes in GDP rather than levels. As such, a pressing question
in environmental economics is how inclusion of pollution damage into an
augmented system of accounts affects comprehensive growth rates. Second,
from the perspective of public policy, macroeconomic regulatory agencies
often base policy choices on quarterly or annual changes in aggregates such
as GDP. Prior rates of change in GDP are clearly linked to monetary and fiscal

10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

policy; one needs to look no further than the Federal Reserve’s response to
the Great Recession for evidence of this claim. In principle, measures of augmented growth could be used in the same way with one additional dimension. Beyond the standard monetary policy toolkit, macroeconomic regulators basing policy choices on the adjusted output measure could prescribe
reductions to damage as a means to enhance growth. That is, if pollution
damage has a significant effect on augmented measures of growth, as the
NAS NRC panel suggested, does this not merit consideration of pollution
control as a viable lever for growth enhancements?
Recent research using data from the United States suggests that the difference between conventional and environmentally adjusted growth rates is
appreciable. For example, over the period 1999 to 2011, GDP in the United
States expanded at annualized rates of between 1.2% and 2.8%. When corrected for both air pollution and carbon dioxide (CO2 ) damage, national output grew by about 0.3 percentage points faster (Muller, 2014, 2016).
Drilling down further, from 2005 to 2008, the U.S. economy was barreling
toward the Great Recession. Annual GDP growth over this period fell to just
over 1%. However, as conventional measures of growth receded, air pollution
damages did too. Thus, the measure that recognized this reduction in damage
expanded more rapidly. In particular, from 2005 to 2008, growth adjusted to
account for air pollution and CO2 emissions was about 0.4% more rapid than
growth in GDP (Muller, 2016).
The contribution of these results is twofold. First, periods of improvement
in air quality can have an appreciable and positive effect on growth. The
intuition of the NAS NRC panel cited above was confirmed. Second, these
results highlight Stiglitz’ position that focusing on market-centric measures
of output may foster the illusion of a trade-off between investments in environmental quality and growth in market production.
The difference in growth discussed above is largest between 1999 and 2002.
One reason for this finding is that Phase II of the Acid Rain Program began in
2000. Because of this policy, emissions of SO2 from regulated power plants fell
by 2.3 million tons between 1999 and 2002. Importantly, costs of compliance
with the regulation are embedded in GDP, while reduced damages are not.
The finding that the benefits of pollution control are large enough to affect
national growth during this time period suggests the need to amend official
measures of output to include the benefits from environmental policy.
Exciting work in this space is occurring in areas beyond air pollution and
greenhouse gas emissions. For example, research is currently ongoing to construct an IAM for water pollution in the United States with resolution akin
to that of modern air pollution IAMs. Once operational, a model such as
this will facilitate, among other topics, the ability to engage in multimedium
pollution control analyses. For example, in modern economies one central

Environmental Accounting

11

aspect of water pollution control is municipal sewage treatment. While on
the surface this appears to be beneficial, sewage treatment generates copious
amounts of airborne ammonia emissions. Ammonia is an important catalyst
in the formation of secondary fine particulate matter, which elevates mortality risk. What is the value of waste removal from water? What is the damage
caused by airborne emissions from this treatment? What is the optimal mix
of air and water pollution emissions from sewage treatment? Intriguing and
as of yet unanswered questions such as these will form the basis of future
research in the field of environmental economics. And, as an ever greater
range of pollution damages are added to the environmental accounts, our
measures of economic performance will do a better job of measuring welfare.

REFERENCES
Abraham, K. G., & Mackie, C. (2006). A framework for nonmarket accounting. In D.
W. Jorgensen, J. S. Landefeld & W. D. Nordhaus (Eds.), A new architecture for the
U.S. national accounts. NBER studies in income and wealth (Vol. 66). Chicago, IL: The
University of Chicago Press.
Ashenfelter, O., & Greenstone, M. (2004). Using mandated speed limits to measure
the value of a statistical life. Journal of Political Economy, 112(2), S226–S267.
Banzhaf, H. S., Burtraw, D., & Palmer, K. (2004). Efficient emission fees in the US
electricity sector. Resource and Energy Economics, 26, 317–341.
Bartelmus, P. (2009). The cost of natural capital consumption: Accounting for a sustainable world economy. Ecological Economics, 68, 1850–1857.
Beatty, T., & Shimshack, J. (2011). School buses, diesel emissions, and respiratory
health. Journal of Health Economics, 30(5), 987–999.
Chay, K. Y., & Greenstone, M. (2003). The impact of air pollution on infant mortality:
Evidence from geographic variation in pollution shocks induced by a recession.
Quarterly Journal of Economics, 118(3), 1121–1167.
Currie, J., Neidell, M., & Schmeider, J. F. (2009). Air pollution and infant health:
Lessons from New Jersey. Journal of Health Economics, 28, 688–703.
Fann, N., Fulcher, C. M., & Hubbell, B. J. (2009). The influence of location, source and
emission type in estimates of the human health benefits of reducing a ton of air
pollution. Air Quality, Atmosphere and Health, 2, 169–176.
Hecht, J. E. (2005). National environmental accounting. Bridging the gap between ecology
and economy. Washington, DC: Resources for the Future Press.
Heo, J., Adams, P. J., & Gao, H. O. (2016). Public Health Costs of Primary PM2.5 and
Inorganic PM2.5 Precursor Emissions in the United States. Environmental Science
& Technology. doi:10.1021/acs.est.5b06125.
Jaramillo, P., & Muller, N. Z. (2016). The air pollution damage from energy production in the U.S.: 2002–2011. Energy Policy, 90, 202–211.
Kerl, P. Y., Zhang, W., Moreno-Cruz, J. B., Nenes, A., Realff, M. J., Russell, A. G.,
… Thomas, V. M. (2015). A new approach for optimal electricity planning and

12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

dispatching with hourly time-scale air quality and health considerations. Proceedings of the National Academy of Sciences, 112(35), 10884–10889.
Krewski, D., Jerrett, M., Burnett, R.T., Ma, R., Hughes, E., Shi, Y., … Tempalski, B.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study
linking particulate air pollution and mortality. HEI Research Report, 140, Health
Effects Institute, Boston, MA.
Landefeld, J. S. (2000). GDP: One of the great inventions of the 20th century. Survey
of Current Business. January, 6–14.
Lepeule, J., Laden, F., Dockery, D., & Schwartz, J. (2012). Chronic exposure to fine
particles and mortality: An extended follow-up of the Harvard Six Cities Study
from 1974 to 2009. Environmental Health Perspectives, 120(7), 965–970.
Levy, J. I., Baxter, L. K., & Schwartz, J. (2009). Uncertainty and variability in
health-related damages from coal-fired power plants in the United States. Risk
Analysis, 29(7), 1000–1014.
Mendelsohn, R. O. (1980). An economic analysis of air pollution from coal-fired
power plants. Journal of Environmental Economics and Management, 7, 30–43.
Muller, N. Z. (2014). Boosting GDP growth by accounting for the environment. Science, 345(6199), 873–874.
Muller, N.Z. (2016). The derivation of discount rates using an augmented measure of
income. NBER WP #22579.
Muller, N. Z., & Mendelsohn, R. O. (2007). Measuring the damages due to air pollution in the United States. Journal of Environmental Economics and Management, 54,
1–14.
Muller, N. Z., & Mendelsohn, R. O. (2009). Efficient pollution regulation: Getting the
prices right. American Economic Review, 99(5), 1714–1739.
Muller, N. Z., & Mendelsohn, R. O. (2012). Efficient pollution regulation: Getting
the prices right: Mortality rate update (Corrigendum). American Economic Review,
102(1), 613–616.
Muller, N. Z., Mendelsohn, R. O., & Nordhaus, W. D. (2011). Environmental accounting for pollution in the U.S. economy. American Economic Review, 101(5), 1649–1675.
National Academy of Sciences, National Research Council (1999). W. D. Nordhaus
& E. C. Kokkelenberg (Eds.), Nature’s numbers: Expanding the U.S. national economic
accounts to include the environment. Washington, D.C.: National Academy Press.
Nordhaus, W. D., & Tobin, J. (1972). Is growth obsolete? Studies in income and wealth
(Vol. 38). New York, NY: NBER.
Nordhaus, W. D. (2006). Principles of national accounting for non-market accounts.
In D. W. Jorgensen, J. S. Landefeld & W. D. Nordhaus (Eds.), A new architecture for
the U.S. National Accounts. NBER Ssudies in income and wealth (Vol. 66). Chicago, IL:
The University of Chicago Press.
Pigou, A. C. (1932). The economics of welfare. London, England: Macmillan and Co.,
Limited.
Rohlfs, C., Sullivan, R., & Kneisner, T. (2015). New estimates of the value of a statistical life using air bag regulations as a quasi-experiment. American Economic Journal:
Economic Policy, 7(1), 331–359.

Environmental Accounting

13

Schneir, K.E. Horrace, W.C., & Felthoven, R.G. (2009). The value of a statistical life:
Pursuing the deadliest catch. http://surface.syr.edu/cpr/49/
Stiglitz, J. E. (2009). GDP fetishism. The Economist’s Voice. September.
Tessum, C. W., Hill, J. D., & Marshall, J. D. (2015). InMAP: A new model
for air pollution interventions. Geoscientific Model Development–Discussions.
doi:10.5194/gmdd-8-9281-2015.
United States Environmental Protection Agency. (1999). The benefits and costs of the
Clean Air Act: 1990–2010. EPA Report to Congress. EPA 410-R-99-001, Office of Air
and Radiation, Office of Policy, Washington, D.C.
United States Environmental Protection Agency. (2011). The benefits and costs of the
Clean Air Act: 1990–2020. Final Report. Office of Air and Radiation, Office of Policy,
Washington, D.C.
Viscusi, W. K., & Aldy, J. E. (2003). The value of a statistical life: A critical review of
market estimates throughout the world. Journal of Risk Uncertainty, 27(1), 5–76.

NICK Z. MULLER SHORT BIOGRAPHY
Nick Z. Muller joined the Middlebury College economics faculty as an
assistant professor in the fall of 2007. He completed his dissertation at Yale
University in May of 2007, where his advisors included Robert Mendelsohn,
William Nordhaus, and Nathaniel Keohane. His dissertation focused on
using integrated assessment models to measure the damages from air
pollution in the United States, and to propose alternative market-based
policies intended to govern the criteria for air pollutants. Dr. Muller also
attended the School of Public and Environmental Affairs where he studied
environmental policy and public finance in pursuit of a master’s degree in
public administration. His current research includes measuring the damages
due to emissions from each industry in the U.S. economy, the design of
market-based environmental policies, and the construction price indices
for air pollution. His research has been published in the American Economic
Review, Science, the Journal of Environmental Economics and Management, and
Environmental Science & Technology, among other outlets.
RELATED ESSAYS
Sociological Theory After the End of Nature (Sociology), Robert J. Brulle
Misinformation and How to Correct It (Psychology), John Cook et al.
Cities and Sustainable Development (Sociology), Christopher Cusack
Theorizing the Death of Cities (Political Science), Peter Eisinger
Architecture of Markets (Sociology), Neil Fligstein and Ryan Calder
Modeling Coal and Natural Gas Markets (Economics), Franziska Holz
Presidential Power (Political Science), William G. Howell
Why Do States Sign Alliances? (Political Science), Brett Ashley Leeds

14

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Money in Politics (Political Science), Jeffrey Milyo
Economics of Renewable Energy Production (Economics), Gregory F. Nemet
The Politics of Disaster Relief (Political Science), Alexander J. Oliver and
Andrew Reeves
The Social Science of Sustainability (Political Science), Johannes Urpelainen



Media

Environmental Accounting