The Genuine Progress Indicator 2006

Full text

1904 Franklin Street Suite 600 / Oakland, CA 94612
T (510) 444.3041 F (510) 444.3191 E info@rprogress.org www.rprogress.org
2007
Date 2006
The Genuine Progress Indicator 2006
A Tool for Sustainable Development
Dr. John Talberth, Clifford Cobb,
and Noah Slattery
© Redefi ning Progress

About the Authors
Dr. John Talberth is Director of the Sustainability Indica-
tors Program at Redefi ning Progress; Cliff Cobb is a senior
fellow at Redefi ning Progress; Noah Slattery is a research
fellow at Redefi ning Progress.
Redefi ning Progress
1904 Franklin Street 6th Floor
Oakland, CA 94612
T (510) 444.3041
E info@rprogress.org
www.rprogress.org
To order more copies of THE GENUINE PROGRESS
INDICATOR 2006: A TOOL FOR SUSTAINABLE
DEVELOPMENT, contact Redefi ning Progress or go to
www.rprogress.org.

The Genuine Progress Indicator 2006
A Tool for Sustainable Development
O
n October 28, 2005 the following headlines
appeared in leading newspapers throughout the
United States:
GDP muscles through
Economy brushes off storms and expands by 3.8 percent
in 3Q, beating estimates.
 e U.S. economy shook off headwinds from
hurricanes Katrina and Rita to grow at a faster-than-
expected 3.8 percent annual rate in the third quarter,
a Commerce Department report showed Friday.
(Reuters, 2005)
Perhaps no headline in recent history does a better job
of illustrating why our nation’s most trusted measure of
economic performance is so woefully out of sync with
people’s everyday experiences. In one fell swoop, these
headlines dismissed the inequitable and catastrophic toll
associated with 1,836 preventable deaths, over 850,000
housing units damaged, destroyed, or left uninhabitable,
disruption of 600,000 jobs, permanent inundation of 118
square miles of marshland, destruction of 1.3 million acres
of forest, and contamination caused by millions of gallons of
fl oodwaters tainted by sewage, oil, heavy metals, pesticides,
and other toxins as irrelevant to the U.S. economy.
1
Few would dispute the fact that gross domestic product
(GDP) fails as a true measure of economic welfare. For
decades, many economists have acknowledged that the
1 For a useful compilation of Hurricane Katrina and Rita damage statistics see:
http://en.wikipedia.org/wiki/Hurricane_Katrina. For wetland loss associated with
the storms see USGS (2006).
Dr. John Talberth, Clifford Cobb, and Noah Slattery
February 2007
CONTENTS
2. Evolution of the Genuine Progress Indicator Framework
3. Theory, Principles and Critiques
8. An Updated GPI Methodology
18. Results and Implications
20. Using the GPI as a Guide to Public Policy
28. Concluding Thoughts and Future Refi nements
29. References
GDP has fundamental shortcomings. “GDP is not a
measure of welfare,” wrote William Nordhaus and James
Tobin, prominent economists at Yale in the early 1970s
(Nordhaus and Tobin, 1972).  e GDP is simply a gross
tally of everything produced in the U.S.—products and
services, good things and bad. In fact, in a 1934 report to
Congress GDP’s chief architect, Simon Kuznets, cautioned
that “[t]he welfare of a nation can scarcely be inferred from
a measurement of national income” (Kuznets, 1934).
Despite these cautions, GDP maintains its prominent
role as a catchall for our collective well being. Perhaps this
is because there has been little consensus on a suitable
replacement. Perhaps, more fundamentally, it is that there
is even less consensus on how well being should really be
measured and if quantitative measurements can be made
at all. Nevertheless, eff orts to fi nd replacements are critical
since GDP forms the basis for important public policy
decisions—i.e. those predicted to increase GDP growth fare
better while those shown to restrict GDP growth are often
killed by political shortsightedness. Recently, GDP growth
was a prominent justifi cation for highly controversial tax
cuts on capital gains while eff orts to secure long overdue
increases in the federal living wage have been thwarted by
persistent gloom and doom forecasts with respect to eff ects
on jobs and economic growth (Foertsch, 2006; Roth, 2005).
In this report, we present an update to the Genuine Progress
Indicator—one of the fi rst alternatives to GDP vetted by
the scientifi c community and used regularly by government
and non-governmental organizations worldwide.  e GPI
is a variant of the Index of Sustainable Economic Welfare
(ISEW) fi rst proposed by Daly and Cobb (1989). Both
the GPI and ISEW use the same personal consumption
data as GDP but make deductions to account for income
inequality and costs of crime, environmental degradation,
and loss of leisure and additions to account for the services
from consumer durables and public infrastructure as
well as the benefi ts of volunteering and housework. By
diff erentiating between economic activity that diminishes
both natural and social capital and activity that enhances

The Genuine Progress Indicator 2006 Redefi ning Progress
2
such capital, the GPI and its variants are designed to
measure sustainable economic welfare rather than economic
activity alone. In particular, if GPI is stable or increasing
in a given year the implication is that stocks of natural and
social capital on which all goods and services fl ows depend
will be at least as great for the next generation while if GPI
is falling it implies that the economic system is eroding
those stocks and limiting the next generation’s prospects.
 e GPI’s structure is grounded in principles set forth in
Natural Step, Hannover, Coalition for Environmentally
Responsible Economies (CERES) and other sustainable
development frameworks that call for no net loss of natural
capital, welfare based accounting, distributional equity, and
throughput minimization.
 e remainder of this report is organized as follows. In
“Evolution of the Genuine Progress Indicator Framework”
(below), we discuss the disconnection between GDP and
true economic welfare and how the GPI responds to these
defects. In “ eory, Principles, and Critiques” (page 3), we
review the GPI’s theoretical underpinnings, place the GPI
in the context of several popular sustainable development
frameworks, and review critiques. In “An Updated GPI
Methodology” (page 8), we explain the new methodology
and rationale for making particular additions or deductions
from personal consumption expenditures. In “Results and
Implications” (page 18) we present results of the 2006
update and key fi ndings. In “Using the GPI as a Guide to
Public Policy” (page 20), we demonstrate how the GPI can
be used to inform public policy debates using globalization,
tax cuts, and sprawl as examples. Concluding thoughts and
directions for future research are set forth in “Concluding
 oughts and Future Refi nements” (page 28).
Evolution of the Genuine Progress Indicator Framework
What’s wrong with GDP as a measure of progress?
During World War II, gross domestic product (then gross
national product) accounts were introduced to measure
wartime production capacity (Cobb et al., 1995). Since
then, GDP has become the world’s most ubiquitous
indicator of economic progress. It is widely used by
policymakers, economists, international agencies and the
media as the primary scorecard of a nation’s economic
health and well-being. Yet, as we know from its creator
Simon Kuznets, the GDP was never intended for this role
(Kuznets, 1934). It is merely a gross tally of products and
services bought and sold, with no distinctions between
transactions that enhance well being and those that
diminish it. Instead of distinguishing costs from benefi ts,
productive activities from destructive ones, or sustainable
ones from unsustainable ones the GDP simply assumes
that every monetary transaction adds to social well-being
by defi nition. In this way, needless expenditures triggered
by crime, accidents, toxic waste contamination, preventable
natural disasters, prisons and corporate fraud count
the same as socially productive investments in housing,
education, healthcare, sanitation, or mass transportation.
It is as if a business tried to assess its fi nancial condition by
simply adding up all “business activity,” thereby lumping
together income and expenses, assets and liabilities.
Moreover, the GDP ignores everything that happens
outside the realm of monetized exchange, regardless
of its importance to well-being.  e crucial economic
functions performed in the household and volunteer
sectors go entirely unnoticed as do ecosystem services such
as fl ood control, water fi ltration, carbon sequestration,
soil formation and maintenance of genetic diversity. As
such, GDP devalues welfare enhancing activities such as
child and elder care, mentoring, or ecological restoration.
In fact, GDP ignores the entire informal, or non-cash
economy—a signifi cant component of the overall exchange
system worldwide and in the United States and made up
of all bartered goods and services. In a 2002 analysis, the
International Monetary Fund reported that worldwide,
the value added by the informal economy had reached
a “remarkably large amount”—up to 44% of GDP in
developing nations, 30% in transition economies, and
16% in Organization for Economic Cooperation and
Development (OECD) economies (Schneider and Enste,
2002). In the United States, the size of the informal
economy is not systematically surveyed, but conservative
estimates place its current size as 9% of offi cial GDP and
involving up to 25 million Americans (Barber, 2003).
Because GDP fails to properly distinguish between welfare
enhancing and welfare degrading expenditures and ignores
non-monetized costs and benefi ts including all informal
sector exchanges, using GDP as a barometer of overall well-
being leads to some perverse results. Consider these:
GDP increases with polluting activities and then again with
clean-ups. Pollution is a double benefi t to the economy
since GDP grows when we manufacture toxic chemicals and
again when we are forced to clean them up.
GDP is boosted by crime. Each year, Americans incur
nearly $40 billion in crime related costs in the form of
lost and damaged property and expenditures on locks,
alarms, and security systems. GDP counts these needless
expenditures as an economic gain, implying that crime is
good for economic growth.

The Genuine Progress Indicator 2006 Redefi ning Progress
3
GDP is oblivious to gross inequality. If a billionaire spends
$10,000 more of her income on aphrodisiacs made from
endangered seals it counts the same as $10,000 spent by a
New Orleans fl ood victim on bare essentials as far as GDP
is concerned. As long as overall expenditures are increasing,
GDP will grow even if the increase is entirely attributable to
conspicuous consumption habits of the wealthy.
GDP plummets as communities become more self reliant.
If a community decided to decrease its reliance on imported
food, energy, and fi nancial markets by expanding rooftop
and community gardens, farmers’ markets, local currencies,
and solar energy and promote social cohesion by expanding
the number of goods and services exchanged by friends and
neighbors, GDP analysts would call for drastic measures to
save the community from impending economic collapse.
GDP grows when we deplete or degrade natural resources.
Clearcutting and sprawl are good for economic growth since
GDP assumes forests, farmland, and wetlands have relatively
little economic value if left alone.
How the GPI attempts to correct these defi ciencies
Beginning with the seminal work of Daly and Cobb (1989)
there have been several attempts to develop alternative
national income accounting systems that address these
defi ciencies. Collectively, these systems measure what is
commonly referred to as “green” GDP. Major objectives of
these green GDP accounting systems are to provide a more
accurate measure of welfare and to gauge whether or not
an economy is on a sustainable time path (Hanley, 2000).
Two of the most popular green GDP systems are the Index
of Sustainable Economic Welfare (ISEW) and the Genuine
Progress Indicator (GPI). Examples of countries with ISEW
data include Austria, Chile, Germany, Italy, the Netherlands,
Scotland, Sweden, and the United Kingdom, while the
United States and Australia off er examples of nations
addressed by proponents of the GPI (Neumayer, 2000).
While methodologies are somewhat diff erent, the ISEW,
GPI, and other green GDP accounting systems all involve
three basic steps (Stockhammer et al., 1997; Neumayer,
2000). Computation usually begins with estimates of
personal consumption expenditures, which are weighted
by an index of the inequality in the distribution of income
to refl ect the social costs of inequality and diminishing
returns to income received by the wealthy. Additions are
made to account for the non-market benefi ts associated with
volunteer time, housework, parenting, and other socially
productive time uses as well as services from both household
capital and public infrastructure. Deductions are then
made to account for purely defensive expenditures such as
pollution related costs or the costs of automobile accidents
as well as costs that refl ect the undesirable side eff ects of
economic progress. Deductions for costs associated with
degradation and depletion of natural capital incurred by
existing and future generations are also made at this stage.
In this way, green GDP systems correct the defi ciencies
of GDP by incorporating aspects of the non-monetized
or non-market economy, separating welfare enhancing
benefi ts from welfare detracting costs, correcting for the
unequal distribution of income, and distinguishing between
sustainable and unsustainable forms of consumption.
Applications of these new accounting systems provide
compelling evidence of a widening gap between traditional
and green GDP, indicating that over time, more and more
economic activity may be self-canceling from a welfare
perspective (Max-Neef, 1995).
For example, the per capita gross domestic product of
Australia nearly tripled between 1950 and 2000, rising
from $10,208 to $29,928 in 2004 dollars. For the period,
the average growth rate was 3.86%. In contrast, per capita
GPI as calculated by Hamilton and Denniss (2000) rose
from $8,074 in 1950 to $14,013, an average growth rate
of just 1.47%. Importantly, the gap between the GDP and
GPI has grown precipitously—from just $2,134 in 1950
to $15,916 in 2000. What this implies is that a decreasing
proportion of economic benefi ts registered by the GDP
count towards improved welfare as time goes on because
such benefi ts are increasingly off set by the costs associated
with growing inequality and deteriorating social and
environmental conditions.
Theory, Principles and Critiques
Theoretical underpinnings
To understand the theoretical foundations for the GPI
it is important to clarify exactly what the GPI is actually
measuring. Summarizing the literature, Asheim (2000)
identifi es three kinds of measurements green GDP accounts
such as the GPI attempt to undertake: (1) welfare equivalent
income; (2) sustainable income, and (3) net social profi t.
Welfare equivalent income refers to the welfare associated
with consumption activities or “psychic” income as fi rst
tagged by Fisher (1906). Paraphrasing Fisher, Lawn (2003,
pg. 111) explains, “[t]he national dividend consists not of
the goods produced in a particular year, but of the services
enjoyed by the ultimate consumers of all human-made
goods.” In recognition of the fact that the economic process
involves many “irksome” activities so that welfare does not
always improve with increasing levels of consumption the
concept of psychic income should be thought of in a net

The Genuine Progress Indicator 2006 Redefi ning Progress
4
sense—i.e. green accounts based on Fisher should measure
not total but net psychic income, which deducts the
harmful aspects of consumption from its welfare enhancing
aspects (Lawn, 2003). To accomplish this, green accounts
fi rst isolate personal consumption expenditures by removing
money spent purchasing, maintaining, or replacing durable
goods and then make a series of additions or deductions to
refl ect both positive and negative externalities associated
with that consumption.
Sustainable income refers to the basic Hicksian notion of
income. In Value and Capital, Sir John Hicks (1948, pg.
179) maintains “we ought to defi ne a man’s income as the
maximum value which he can consume during a week, and
still expect to be as well off at the end of the week as he
was at the beginning.” As such, the very notion of income
is sustainable by defi nition making the term “sustainable
income” a redundancy. To arrive at an adequate measure
of Hicksian income, green accounts deduct from GDP
depreciation of both human built and natural capital stocks
and certain expenditures (i.e. on security systems) made to
defend ourselves from some of the undesirable side eff ects of
economic growth (Daly and Cobb, 1994).
Net social profi t is a measure of policy eff ectiveness. Net
social profi t analysis is simply an expanded form of cost-
benefi t analysis that uses welfare equivalent or sustainable
income rather than GDP.  us, using green accounts in
net social profi t analysis provides a measure of the welfare
or sustainability implications of policy changes (Asheim,
2000). In particular, net social profi t is the diff erence
between green GDP with and without a particular policy
change. Net social profi ts can be positive, indicating that
the proposed policy is welfare enhancing, or negative,
indicating that its social costs exceed benefi ts. Since not all
components of the Fisher and Hicks income concepts are
applicable in any particular policy setting, green accounts
used to calculate net social profi t are not necessarily the
same as either welfare equivalent or sustainable income.
Although the Genuine Progress Indicator has individual
columns that can be of use in calculating welfare equivalent
income, sustainable income, or net social profi t, in
aggregate, it falls squarely under category 1—the Fisherian
concept of welfare equivalent income—because it attempts
to measure the net psychic income households derive from
their consumption activities. However, it only counts the
portion of Fisherian income that is sustainable, or derived
from stable or increasing stocks of human built and natural
capital.  us, the GPI measures the “welfare a nation
enjoys at a particular point in time given the impact of
past and present activities” (Lawn, 2003, pg. 106). While
certainly a more accurate measure of true welfare than
GDP or green GDP accounts rooted in Hicksian notions
of sustainable income, the methodological objectivity of
Fisherian measures such as the GPI is necessarily much less
clear because they necessitate value judgments over what
does and does not constitute welfare enhancing forms of
consumption, what costs and benefi ts are added or deducted
from such consumption, and how these costs and benefi ts
ought to be measured. It is necessary, then, to make explicit
these more subjective aspects of the GPI. We do so by
identifying core principles of sustainable development used
to guide GPI accounting.
Principles of sustainable development
As noted by Hanley (2000), the term sustainable
development has been widely and variously defi ned but a
consensus as to its general implication is that sustainable
development requires a non-declining level of well being
for future generations. Since 1987, when this general
concept was formalized by the World Commission on
Environmental and Development, there have been countless
numbers of processes initiated by non-governmental
organizations, governments, business leaders, and academics
to develop operational principles to guide lifestyle choices,
public policy, and business practices. Such principles are
typically grouped into three core domains: economic,
environmental, and social (Harris, 2000). In fact, a
key meta-principle is “that social, environmental and
economic needs must be met in balance with each other for
sustainable outcomes in the long term.”
2
 is meta-principle is embodied in the GPI. Recognizing
the interdependence of economic well being with the
quality of the natural environment and the quality of our
social relationships, the GPI sub-accounts track progress
in each domain. As explained in detail in “An Updated
GPI Methodology” (page 8), the GPI’s economic domain
is populated by personal consumption expenditures,
consumer durable service fl ows, services from public
infrastructure, net capital investment, and net foreign
borrowing.  e environmental domain assigns costs to air,
noise, and water pollution, lost farmland, wetlands, and
forests, depletion of oil reserves, as well as carbon dioxide
and ozone damages.  e social domain counts the benefi ts
of volunteer work, higher education, and parenting as
well as the costs of crime, inequity, commuting, and auto
accidents.  us, the GPI approximates welfare through a
relatively well balanced set of sub accounts across each of
the major sustainability domains.
2 Taken from the United Nations Conference on Environment and Development
(UNCED) summary of the 1992 Earth Summit at Rio de Janeiro (http://www.
un.org/jsummit/html/basic_info/unced.html).
4

The Genuine Progress Indicator 2006 Redefi ning Progress
5
Within each domain, the GPI operationalizes key principles
common to a number of popular sustainability frameworks.
Within the economics domain, Pezzey (1992) groups
such principles into two major categories: (1) ends based
defi nitions, such as non-declining per capita consumption
or utility, and; (2) means based defi nitions, such as a non-
declining stock of human and natural capital from which
future generations can produce well being. By accounting
for the costs of depleting both natural (i.e. farmland) and
human built capital stocks (i.e. net capital investment)
the GPI is closely aligned with frameworks based on the
latter. British Columbia’s Principles for Sustainability is an
example.  is framework contains normative guidance to
promote long term economic development that increases
the benefi ts from a given stock of resources by “living off
the interest of natural resources” and not drawing down
environmental asset stocks (Saunier, 1999).
 is principle is closely related to a common principle
from the environment domain—the principle of strong
sustainability. Strong sustainability assumes a very limited
degree of substitution between human and natural
capital stocks (Pearce et al., 1990; Hanley, 2000). While
some substitution is possible, many natural resource
stocks are presumed to be irreplaceable and provide non-
substitutable services to the economy. Examples include
the natural processes that control the gaseous composition
of the atmosphere, produce soils, or evolve complex
ecological communities such as old growth forests. Strong
sustainability, then, requires a non-declining stock of this
irreplaceable natural capital. In contrast, the principle of
weak sustainability simply requires that capital stocks in
aggregate remain stable or increase on a per capita basis, and
depletion of natural capital is sustainable to the extent that
man-made substitutes can be found and used (Pearce and
Atkinson, 1993). Because the GPI counts costs associated
with lost farmland, wetland, and primary forest rather than
assuming seamless substitutability it is more in line with the
assumption of strong sustainability.
Another key sustainability principle from the environment
domain is the principle of thermodynamic effi ciency. In
the mid to late seventies, and partially in response to the
energy crisis of that period, ecological economists began
to promote an entirely new framework for addressing
the related issues of sustainability and economic growth-
thermodynamics.  e thermodynamic approach, in essence,
calls for a comprehensive “bookkeeping” system to track
the fl ows of energy, matter, and information through the
economy, which is itself an open system embedded within
the closed system of the earth’s biosphere. From a normative
standpoint, the approach calls for recognition of the limits
imposed on the economic system by the fi rst and second
laws of thermodynamics.  e fi rst law of thermodynamics
says that matter and energy can neither be created nor
destroyed.  ey can only be converted from one form
to another.  e second law, also known as entropy law,
states that all physical processes proceed in such a way that
availability of energy involved decreases, i.e. the entropy of
a closed system always increases. Entropy can be understood
as a measure of disorder or energy not available for work.
Implications of the fi rst law for economics are that all
resources are fi nite, and that our use of those resources
generates a fl ow of unusable or harmful residuals into the
environment which, if left unassimilated, generate negative
feedback in the form of pervasive externalities that impede
production and consumption (Ayres 1978; Markandya
and Richardson, 1992). Implications of the second law for
economics are that since complete recycling is impossible,
our current economic system will eventually break down as
shortages of low entropy energy inputs are exhausted, as the
residual high entropy energy and matter ceases to be capable
of being recycled, and as natural resources of all types
become increasingly scarce. Moreover, a greater throughput
of energy and materials will hasten the day where shortages
become acute and any incremental contribution to further
growth is negated by an increase in overall disorder of the
economic system. From the perspective of thermodynamic
effi ciency, a sustainable economic system is one that
concentrates on development, not growth. Growth refers
to the quantitative increase in the physical scale of the
economy, its throughput of matter and energy, and the
stock of human built artifacts while development refers to
largely qualitative improvements in the structure, design,
and composition of physical stocks and fl ows that result
from greater knowledge, both of technique and of purpose
(Folke et al., 1993; Daly and Cobb, 1989). In addition to
being one of the core tenets of the ecological economist
worldview, the notion of thermodynamic effi ciency is
embodied in several popular sustainable development
frameworks including Natural Step (no net increase in
substances produced by society), the World Congress of the
International Union of Architects (eliminate the concept of
waste), and the Hannover Principles (rely on natural energy
fl ows).
In Figures 1 and 2 (page 6), the concept of materials,
energy, and information fl ows is used to describe two
diff erent kinds of economic systems. Figure 1 describes a
less sustainable economy based on maximizing production
and consumption, greater reliance on exhaustible resources
for inputs, and generation of a signifi cant waste stream
that produces a host of negative externalities (such as air

The Genuine Progress Indicator 2006 Redefi ning Progress
Figure 2:
More Sustainable Economy Based on Minimizing Throughput
Low Entropy
Inputs
Sola
r
Ecos
y
stem Services
Resource Extraction
Natural Capital
Investment
High
Entropy
Waste
Waste Stream
Recycling
Cultural Capital
Maintenance
Production and
Consumption
Negative
Externalities
High
Entropy
Waste
Low Entropy
Inputs
Sola
r
Ecos
y
stem Services
Resource Extraction
Figure 1:
Less Sustainable Economy Based on Maximizing Throughput
Production and
Consumption
Maintenance
Cultural Capital
6

The Genuine Progress Indicator 2006 Redefi ning Progress
7
and water pollution) that feed back into the natural world
and impede ecosystem services. Figure 2, on the other
hand, describes a more sustainable economy that depends
more heavily on solar energy and the services provided by
natural ecosystems, that invests more of its resources into
development of cultural capital and knowledge rather than
production and consumption, that recycles a signifi cant
portion of the waste stream, and which invests heavily in
maintaining and restoring natural capital. In Figures 1
and 2, the relative size of arrows and text indicate what is
emphasized or de-emphasized by each economic system.
 e GPI accounts provide a way to measure progress
towards the type of economic system described in Figure
2 by providing at least some of the thermodynamic
bookkeeping needed to fi ll in the market’s inability to
correctly signal scarcities of both low entropy inputs, the
value of building up cultural capital, and the true costs of
environmental externalities associated with air, water, and
noise pollution.
In the realm of social sustainability, one example of
the GPI’s consistency with widely shared principles of
sustainable development is the fact that the GPI makes an
explicit adjustment to personal consumption expenditures
for improvements or declines in distributional equity.
 is adjustment, of course, is based on the widely held
belief that sustainable development must, by defi nition, be
equitable. According to Hanley (2000, pg. 6), “[a] socially
sustainable system must achieve distributional equity…”. A
major goal of the Habitat Agenda Principles is to create “a
more balanced and equitable global system.”  e Natural
Step is concerned that “resources should be used fairly and
effi ciently” (Saunier, 1999).  us, the GPI’s concern with
distributional equity is well grounded within a number of
sustainable development frameworks.
Critiques and limitations
Despite its roots in both economic theory and widely
shared principles of sustainable development, the GPI is
not without its detractors. Criticisms have been leveled at
its theoretical foundations, components, and calculation
methods. Many of the concerns were addressed during
the formative years of the GPI. In their 1994 volume  e
Green National Product: A Proposed Index of Sustainable
Economic Welfare, Cobb and Cobb published a series of
critical essays and described how those criticisms were dealt
with in the revised GPI accounts contained in that volume
(Cobb and Cobb, 1994). It is not our intent to revisit those
debates. Instead, we focus here on lingering criticisms.
Neumayer (1999), Dietz and Neumayer (2006) and Lawn
(2003; 2005) have engaged in the most visible dialogue in
the recent literature.  eoretically, Neumayer and others
argue that it is “not possible to combine an indicator of
current welfare with an indicator of sustainability” because
costs associated with depletion of non-renewable resources
and other forms of natural capital incurred by future
generations make little diff erence to current welfare (Dietz
and Neumayer, 2006, pg. 189). Deductions for natural
capital depletion, then, are inconsistent with the Fisherian
notion of income the GPI purports to measure. In response,
Lawn (2003) maintains that because Fisher’s concept of
income and capital treat the production of replacement
goods as the cost of keeping human made capital intact it is
entirely appropriate to deduct natural capital depletion costs
using the replacement cost method, as described in “An
Updated GPI Methodology,” below.
Critics have also noted the converse—that there are
components of current welfare that have little apparent
link to long term sustainability. Another theoretical fl aw
is the fact that while the GPI purports to be based on
the principle of strong sustainability, it in fact measures
weak sustainability.  is is because the GPI measures the
loss of both natural and human-built capital separately,
so if natural capital is depleted, the costs of doing so
can be masked by substitution of human-built capital of
equal or greater value. According to Neumayer (1999,
pg. 93), “[i]ronically, the ISEW does not measure strong
sustainability, but weak sustainability at best since it assumes
perfect substitutability among diff erent forms of capital.”
In terms of GPI components, the most important critique
is that the GPI is arbitrary in what it includes or implicitly
excludes as contributors to or detractors from welfare
(Neumayer, 1999). For instance, the GPI corrects for
income inequality but does not include corrections for the
degree of political freedom or degree of equality between
the sexes.  e inclusion of almost every disservice item
(i.e. commuting costs, loss of leisure, noise pollution) has
been challenged because it is unclear whether or not these
costs have already been factored into household and worker
decisions (Lawn, 2005; Rymes, 1992). Because the GPI
framework requires a subjective judgment of what does and
does not count towards welfare and what does and does not
properly count as a defensive expenditure, it cannot serve its
desired role as an objective measure of sustainable economic
welfare.
In terms of calculation methods, Dietz and Neumayer
(2006) take issues with four components: (1) the valuation
of the depletion of non-renewable resources; (2) the
cumulative cost of long term environmental damage; (2)
the adjustment of personal consumption expenditures for

The Genuine Progress Indicator 2006 Redefi ning Progress
8
income inequality, and; (4) the deduction of defensive
expenditures.  e critiques here involve the precise
calculation methods, not the basic components. For
example, the GPI uses a replacement cost method to value
depletion of non-renewable resources when Neumayer,
Lawn and others believe a resource rent approach is more
appropriate (Neumayer 1999; Dietz and Neumayer, 2006;
Lawn, 2005).  ere have also been a number of criticisms
made to the sources of data relied upon for calculating
individual GPI sub-accounts. As described by Lawn, the
lack of appropriate data for many GPI components and the
need to “make heroic assumptions ensure the values of these
items are likely to be, at best, distant approximations of
their correct value” (Lawn, 2005, pg. 199).
Despite these lingering theoretical and methodological
issues, the most outspoken recent critic of the GPI and
ISEW has concluded:
…the ISEW’s focus on comprehensive current
welfare is laudable. Indeed, the emerging sustainable
consumption discourse gives the ISEW renewed salience
because, according to some, the task of making society’s
consumption more sustainable is in large part a question
of separating out those things that we consume that make
us “happier” and those that don’t or even make us less
happy.(Dietz and Neumayer, 2006, pg. 190).
In the next section, we present a column by column
explanation of the GPI 2006 update. While we have not
changed the basic theoretical approach, we have made
a number of signifi cant changes to GPI components,
calculation methodologies, and sources of date that seek to
improve upon its overall accuracy.
An Updated GPI Methodology
 e GPI is derived from 26 separate time series data
columns spanning the 1950-2004 period. Due to delays
in government reporting, there is a two year time lag in
publishing GPI accounts. In this section we review the
column by column calculations included in the GPI. We
briefl y describe the rationale for including each column, the
data sources on which we rely, and the general calculation
methodology. We encourage readers to contact the authors
for a more detailed explanation and for the most up to
date reference information for time series data sets.  e
methodology presented here represents a signifi cant update
to the methodology in use at Redefi ning Progress since
the late 1990s as described by Cobb et al. (1998). Many
of the changes are limited to changes in the sources of
information, but several others include changes to the
calculation approach. Unless otherwise noted, all fi gures are
reported in year 2000 dollars.
Column B – Personal Consumption
Personal consumption expenditures on goods and services
are the key driver of the GDP, and are the initial starting
point for the GPI. As noted by Lawn (2005), personal
consumption expenditures are a valid starting point for
the GPI since we are ultimately interested in the welfare
associated with this consumption rather than the monetary
value of production. Accounting for nearly 67% of its
total in 2004, consumer spending contributes far more
to GDP than business investment expenditures (16%)
and government (federal, state, and local) expenditures
on products and services (17%). In 2004, U.S. personal
consumption expenditures amounted to $7.6 trillion,
compared with $1.2 trillion in 1950. On a per capita basis,
personal consumption expenditures have risen steadily
from $7,570 per capita in 1950 to $25,820 in 2004, an
increase of 241 percent. Personal consumption expenditure
data were taken from the National Income and Product
Accounts (NIPA) tables published by the Bureau of
Economic Analysis.
Column C – Income Distribution Index
 ere is strong empirical evidence that rising income
inequality hinders growth in economic welfare (Hsing,
2005). A highly unequal distribution of income can be
detrimental to economic welfare by increasing crime,
reducing worker productivity, and reducing investment.
Moreover, when growth is concentrated in the wealthiest
income brackets it counts less towards improving overall
economic welfare because the social benefi ts of increases
in conspicuous consumption by the wealthy are less
benefi cial than increases in spending by those least well off
(Lawn, 2005).  e GPI accounts for income inequality
by discounting personal consumption expenditures by the
amount of inequality that persists in a given year using the
Gini and income distribution indices (IDI).
 e Gini index is the diff erence between actual distribution
and equal distribution by income quintiles.  e Gini index
ranges from 0, when every household has the same income,
to 1 when one household has all the income.  us the
higher the Gini index the greater the income inequality,
or the greater the portion of aggregate income earned
by the top household income bracket. It incorporates
detailed aggregate income shares data into a single statistic,
which summarizes the dispersion across the entire income
distribution. It compares current income distribution with
an ideal equal distribution of aggregate income, giving equal

The Genuine Progress Indicator 2006 Redefi ning Progress
9
weight to all income levels by calculating the square root
of the sum of the squared diff erences of each quintile from
a 20 percent share.  e Gini index is published regularly
by the U.S. Census Bureau.  e IDI simply measures the
relative change in the Gini index. It is set at a value of 100
in 1968, the year the Gini index was at its lowest value.
As column C indicates, the income distribution index in
the United States is at its most unequal level since 1950
and now stands at 120.10. According to the U.S. Census
Bureau, the richest 20% of U.S. households now receive
nearly 50% of all income, while the poorest 20% receive
just 3.4%.  e Gini index now stands at .464, up from .388
in 1968 (U.S. Census Bureau, 2003). As a result, on a dollar
per dollar basis, personal income expenditures count less
now than they ever have towards genuine progress at any
time since 1950.
Column D – Weighted Personal Consumption
Weighted personal consumption is Column B (personal
consumption expenditures) divided by Column C (income
distribution index) multiplied by 100.  e reason for
dividing rather than multiplying is that larger numbers in
Column B indicate greater inequality. Column C becomes
the base number from which the remaining Columns in
the GPI are either added or subtracted. For 2004, personal
consumption adjusted for income inequality is $6.32 trillion.
Column E – Value of Household Work and Parenting
Work performed in households is more essential than much
of the work done in offi ces, factories, and stores. Yet most of
this goes unaccounted for in the national income accounts.
While the housework and parenting of the stay-at-home
mom or dad counts for nothing in the GDP, commercial
childcare in the monetized “service sector” adds to the
GDP. Other unpaid household labor, such as the physical
maintenance of the housing stock (from cleaning to light
repairs), also constitutes valuable economic activity.
 e calculation of the value of household labor in the
GPI is derived from the work of economist Robert Eisner,
past president of the American Economics Association.
Eisner fi rst derived estimates of the annual hours spent
performing relevant household tasks from time-use studies
conducted by the Michigan Survey Research Center in
1965, 1975 and 1981. He then treated the value of an hour
of housework as equivalent to the amount that a family
would have to pay to hire someone to do equivalent work
in their home.  is then yields an estimate of the total
annual value of household work (Eisner, 1985). Our GPI
update incorporates three new data points: one from the
fi nal Michigan Survey Research Center study in 1985 and
two from the Bureau of Labor Statistics (BLS) American
Time Use Surveys (ATUS) of 2003 and 2004. For the years
in between, we extrapolated using a regression on the years
1981, ’85, ’03 and ’04. Each data point was incorporated
slightly diff erently.
For the 1985 estimate we replicated Eisner’s methodology
as closely as possible. Starting with raw data from the
Michigan survey we calculated the number of hours
of household work performed by each of four groups:
employed men, unemployed men, employed women,
unemployed women. We then multiplied those numbers by
each group’s respective total U.S. population to calculate the
total number of hours of household work performed: 235
billion.  e work was valued at $7.14 per hour, based on
houseworker salaries published by the Bureau of Economic
Analysis. In the 2003 BLS time-use study the number of
household hours for each of the four groups was multiplied
by each group’s respective total U.S. population to calculate
the total number of hours of household work performed:
296 billion.  e work was valued at $8.23 per hour, based
on houseworker wage data from the BLS.
In the 2004 ATUS the data were not only broken down by
sex and employment status, they were further subdivided
by the ages of children in the household. To consolidate
the numbers into the four subgroups we weighted them
using household data from the U.S. Census Bureau’s
Current Population Survey. Otherwise, the methodology
was the same as that used to calculate 2003. Total hours
of household work performed were 304 billion, valued at
$8.34 per hour.  e GPI estimates the value of housework
and parenting at $2.5 trillion in 2004.  is represents the
single most signifi cant positive adjustment to personal
consumption expenditures.  e value of housework and
parenting was roughly 33 percent of personal consumption
expenditures in 2004; in 1950 it was 58 percent. In part,
this refl ects our increasing reliance on the market to provide
services formerly contributed by households.
Column F – Value of Higher Education
 ere has been considerable debate over whether to include
this column at all. Previous editions of the GPI have
omitted the cost of higher education, considering it an
investment. Other studies have considered higher education
to be consumption, while still others have asserted that the
primary value of higher education is as a signaling eff ect, or
queuing mechanism, and it should be considered a defensive
expenditure. While it is clear that the long-term earnings
of college graduates are much higher than those without

The Genuine Progress Indicator 2006 Redefi ning Progress
10
a college degree, we sidestepped the debate over how to
address these individual benefi ts by focusing instead on the
benefi ts to society.
Hill et al. (2005) provide an exhaustive list of such benefi ts,
which are both monetary and non-monetary and in the
form of increases in the stock of knowledge, productivity
of workers and capital, civic participation, job market
effi ciency, savings rates, research and development activities,
charitable giving, and health. Based partially on Moretti
(2004) they estimate the total value of this social spillover
eff ect to be $16,000 per year per college-educated worker.
We multiplied this value by the number of people 25
years and older that had completed at least four years of
college as reported in periodic U.S. Census Bureau Current
Population Surveys. In 2004, we estimate the annual social
benefi ts of higher education to be nearly $828 billion.  is
represents the GPI’s second largest addition to personal
consumption expenditures.
Column G – Value of Volunteer Work
Some of the most important work in America is not done
for pay. Such work is not only performed at home, but also
the broader realm of our neighborhoods and communities.
Work done here is the nation’s informal safety net, the
invisible social matrix on which a healthy market economy
depends. Whether each additional lawyer, broker, or
advertising account executive represents a net gain for the
nation is arguable. But there is little question that workers
in the underserved community and volunteer sectors—the
churches and synagogues, civic associations and informal
neighborly eff orts—are doing work that is desperately
needed. Despite its crucial contribution, however, this work
goes entirely unmeasured in the GDP.  e GPI begins to
correct this omission.
First we estimate the total number of hours volunteered
each year. We relied primarily on three Current Population
Surveys conducted by the Bureau of Labor Statistics in
1965, 1974, and 1989 and the American Time Use Surveys
from 2003 and 2004. Intermediate years were interpolated.
Since the questions asked in each survey were not exactly
the same, there are some comparability problems. But the
surveys are close enough to provide a workable estimate
for the purposes of the GPI. Secondly, we applied the
Independent Sector estimate of the value of an hour of
volunteer time in 2000 (since all GPI fi gures are reported
in year 2000 dollars).  at value is $15.68 per hour
(Independent Sector, 2006).  e GPI indicates that the
value of volunteer activities in the United States stood at
$131 billion in 2004 or $447 per capita.  is is signifi cantly
higher than the 1950 value of $202 per capita implying
that over the past few decades, Americans have become
more generous with their time and that this time is of much
greater worth.
Column H – Services of Consumer Durables
 e money spent on durable items, such as cars,
refrigerators, and other appliances is not a good measure
of the actual value consumers receive from them. It is
important to take account, as well, of how long the item
lasts. For example, when you buy a furnace or a dishwasher,
you do not “consume” it in one year.  e appliance (or
“consumer durable”) provides service for a number of years.
Because of this, the GPI treats the services of household
capital as a benefi t and the initial purchase price as a
cost.  is column adds the annual services derived from
consumer durables, which economic theory defi nes as
the sum of the depreciation rate and the interest rate. If a
product lasts eight years, it depreciates at 12.5 percent per
year and thus provides that much of its service each year. At
the same time, if the interest rate is 5 percent, the purchaser
of the product could have received that much interest
by putting the money into the bank instead. Economists
therefore regard the interest rate as part of the monetary
value of the product to the consumer.
Based on an assumed depreciation rate of 15 percent and
an average interest rate of 7.5 percent, the value of services
from household capital is estimated at 22.5 percent of the
value of the net stock of cars, appliances, and furniture
at the end of each year as estimated by the Bureau of
Economic Analysis. To avoid double counting, we make
an adjustment (column M) by subtracting out actual
expenditures on consumer durables. Focusing on annual
services that household appliances and equipment provide
rather than on the purchase price corrects the way the GDP
treats money spent on durables.  e value of services from
consumer durables is treated as a benefi t and is thus an
addition to the GPI account. In 2004, the benefi ts from
household capital amounted to $743.72 billion, making it
the GPI’s third largest addition to personal consumption.
Column I – Services of Highways and Streets
 e GPI does not include most government expenditures
since they are largely defensive in nature; they protect
against erosions in the quality of life, rather than enhancing
it (Leipert 1986, 1989).  is is particularly true of the
government’s largest budgetary item, military spending.
On the other hand, some government activities, such as
transit systems and sewer or water districts, provide services
for a fee in a manner similar to private business.  ese fees

The Genuine Progress Indicator 2006 Redefi ning Progress
11
show up in personal consumption fi gures in the national
income accounts and thus are already included in column
B.  is leaves other government services that could be sold
in theory, but are diffi cult to price with regard to individual
users. Overwhelmingly, the largest item in that category is
the use of streets and highways, which we include here as a
separate GPI category.
 e annual value of services from highways and streets is
derived the Bureau of Economic Analysis fi gures of the net
stock of federal, state, and local government streets and
highways from 1950 to 2004.  e annual value of services
from streets and highways is estimated by taking 7.5 percent
of the net stock value.  is is based on the logic that around
10 percent of the net stock (2.5 percent for depreciation and
7.5 percent for average interest rates) is the estimated annual
value of all services from streets and highways. However,
since we assumed that 25 percent of all vehicle miles are
for commuting (a defensive expenditure), this leaves 75
percent as net benefi ts.  us the GPI assumes the net service
value of streets and highways is 75 percent of 10 percent,
or 7.5 percent of net stock. In 2004 we estimate the value
of services from streets and highways at $111.55 billion, an
addition to the GPI account.
Column J – Cost of Crime
Crime takes a large economic toll on society. Some of
these costs are obvious, such as medical expenses and
lost property. But others are more elusive, because they
are psychological, such as the trauma of being violated,
or are incurred in the form of lost opportunities, such as
activities foregone because people fear the possibility of
theft or violence.  e GPI relies on the Bureau of Justice
Statistics National Crime Survey year to year estimates of
the cost of crime to victims in terms of their out-of-pocket
expenditures or the value of stolen property. Undoubtedly
the full cost of crime is underestimated given the absence of
estimates of the more elusive costs.
We also include other defensive expenditures on locks,
burglar alarms, security devices, and security services.
Most of us would not otherwise purchase these personal,
household, or business security items. In the GPI we
subtract these expenditures on crime prevention because
they represent personal consumption that does not add
to the well-being of our households but merely prevents
its deterioration or violation. Expenditures on locks were
estimated by extrapolating data for locks from Laband
and Sophocleus (1992) while expenditures on alarms were
drawn from regular reports issued by Security Distributing
and Marketing (SDM). Both data sets were extrapolated
forward and backward in time based on security industry
sales data and projections. In 2004, the GPI deducts $34.22
billion from personal consumption expenditures to refl ect
the cost of crime.
Column K – Loss of Leisure Time
 e GDP creates the illusion that the nation is getting
richer, when in fact people are working harder to produce
and buy more and to pay interest on mounting personal
indebtedness. According to Bluestone and Rose (1997)
“since the 1980s people have been saying they work ‘too
hard’—that they are spending too much time on the job,
with too little left for family, chores, or leisure.” A more
accurate measure of genuine progress and well-being would
consider the loss of leisure that went along with increased
output. Accounting for the nation’s well-being ought to
include the value of leisure time lost or gained.
In order to provide a reasonable estimate, the GPI includes
only the value of leisure lost in relation to 1969, the year
with the greatest leisure since 1950.  e number of leisure
hours per year is taken from a study by Leete-Guy and Schor
(1992) who estimated the annual working hours (including
housework) of labor force participants. Estimates from 1969
to 1992 were derived from their fi gures. For 1950 to 1969,
we estimated that annual hours of work declined by 0.3
percent per year. For the period 1993 to 2004 we extrapolated
the trend based on the work of Mishel et al. (1996) who
estimate that annual hours of work have increased an average
5.2 hours per year between 1989 and 1994.
 e number of work hours is then subtracted from 3,650
hours of discretionary time (10 hours per day) to arrive at
an estimate of the total discretionary hours of leisure per
person per year.  e term “discretionary” simply means time
away from work minus time spent sleeping and kindred
maintenance activities. We use 70 hours per week as the
threshold; thus discretionary time is the amount less than
70 hours per week that people work.  e resulting fi gure for
each year is subtracted from the amount in 1969 to derive
an estimate of the hours of leisure per worker.  e change
since 1969 is the basis for estimating the loss of leisure time,
which we value at $13.36 per hour in year 2000 constant
dollars (which is approximately the average real wage rate
for the period 1950 to 2004).  e result is a GPI deduction
of $401.92 billion in 2004.
Column L – Cost of Underemployment
 e GPI does not deal with the eff ects of short-term and
cyclical unemployment. Although such hardships are
not without social consequences and costs, much of the

The Genuine Progress Indicator 2006 Redefi ning Progress
12
fi nancial hardship is mitigated by unemployment insurance
benefi ts. Underemployment is a more inclusive concept
than unemployment. It refers to persons who are either
chronically unemployed, discouraged (gave up looking
for work), involuntary part-time (would prefer full-time
work but are unable to fi nd it), or constrained by other
factors, such as lack of child care or transportation.  e
costs of underemployment fall on the discouraged workers
and their families. But the community and society also
pays a price when limited work opportunities may lead
to frustration, suicide, violence, crime, mental illness, or
alcoholism and other substance abuse.  e GPI treats each
hour of underemployment (the number of unprovided
hours for constrained workers) as a cost, just as leisure time
is considered a benefi t. An hour of leisure time is a desirable
objective whereas an hour of underemployment is a burden.
 e GPI uses the research of Leete-Guy and Schor (1992)
who calculated the number of “unprovided hours” of work
in 1969 and 1989 by constrained workers—people who
want to work more.  ey found that the number of hours
of underemployment in the entire labor force rose from
4.2 billion hours in 1969 to 14.6 billion hours in 1989.
We extrapolate their fi gures from 1950 to 1968 and from
1990 to 2004. We assume the number of unprovided hours
per constrained worker from 1990 to 2004 continues to
increase at the rate of 0.59 percent per year (the rate of
increase between 1969 and 1989).  is approach bypasses
changes in unemployment due to business cycles and
focuses instead on the eff ects of long-term trends.
 e estimates of unprovided hours per constrained worker
are then multiplied by the millions of estimated constrained
or underemployed workers using data from the Economic
Policy Institute and Bureau of Labor Statistics and then by
an average real wage of $13.36 per hour. As with leisure,
this is the average real wage during the accounting period
1950 to 2004.  ese estimates suggest that the cost of
underemployment peaked at $195.09 billion in 1989 and
has since declined to $176.96 billion by 2004.
Column M – Cost of Consumer Durables
 e actual expenditures on consumer durables are a negative
adjustment in the GPI to avoid double counting the value of
their services (column H).  e value of private expenditures
on consumer durables in constant 2000 dollars comes from
the National Income and Products Accounts.  e cost of
consumer durables in 2004 is estimated at $1.09 trillion.
Column N – Cost of Commuting
Urban sprawl has put more cars on the road, exacerbated
traffi c congestion, and increased the time Americans must
spend getting to and from work. According to the U.S.
Department of Transportation, there has been a 66%
increase in the number of vehicles per household and
signifi cant increases in commute times since 1960 (DOT,
2000). While commuting is for most people an unsatisfying
and sometimes frustrating experience, the GDP treats it as
a benefi t to consumers.  e more time and money spent
commuting, the more these regrettable activities contribute
to the GDP. Moreover, GDP does not account for the
opportunity costs of time spent commuting; time that could
be spent freely with family, at leisure, sleeping, or at work.
 e GPI corrects for the shortcoming of the GDP
account by subtracting the cost of commuting.  ere
are two distinct types of costs incurred in commuting.
 e fi rst is the money spent to pay for the vehicle, or for
bus or train fare; the second is the time lost that might
have been spent on other, more enjoyable or productive
activities. In the GPI accounts, the direct (out-of-pocket)
costs of commuting are a function of the portion of non-
commercial vehicle miles used in commuting, the cost of
user operated transport, the cost of depreciation of private
cars, the portion of passenger miles on public transportation
used for commuting, and the price of purchased local
transportation. Data for these variables were taken from the
Statistical Abstract of the United States and BEA’s National
Income and Product Accounts.
 e indirect costs of commuting (i.e., the value of the time
lost) are calculated as the total number of people employed
each year times the estimated annual number of hours
per worker spent commuting times a constant value for
the time. Because some people regard commuting as part
nuisance and part leisure, we assigned a cost of $8.72 per
hour (rather than the $13.36 per hour for lost leisure).
 e number of hours per year was derived from survey
data on time-use by households (Leete-Guy and Schor,
1992) coupled with data from the National Household
Transportation Survey (NHTS) from 1983, 1990, 1995,
and 2001. According to the National Center for Transit
Research (NCTR) at the University of South Florida,
NHTS data show that commuting times have increased
by 29.1% since 1983 (NCTR, 2005).  e estimated cost
of commuting in 2004 was $522.61 billion or $1,778 per
capita. Per capita costs have risen by 91% since 1950.
Column O – Cost of Household Pollution Abatement
One of the costs that pollution imposes on the households
of the nation is the expenditures made for equipment
such as air and water fi lters.  ese defensive expenditures
do not improve the well-being of households, but merely

The Genuine Progress Indicator 2006 Redefi ning Progress
13
compensate for the externalities—that is, pollution—
imposed upon them as a result of economic activity. Such
expenditures merely attempt to restore environmental
quality to a baseline level.
For the period 1972 to 1994, we used data published by
the Bureau of Economic Analysis (Vogan, 1996). For years
prior to 1972, we assumed that personal expenditures on
pollution abatement and control increased by 20 percent
per year according to the trend after 1972. In 1996 the
BEA data series was discontinued, therefore we extrapolated
expenditures based on the average rate of increase from
1991 to 1994. We estimate the cost of household pollution
abatement to be $21.26 billion in 2004.
Column P – Cost 0f Automobile Accidents
 e damage and economic loss due to automobile accidents
represents a real cost of industrialization and increasing
traffi c densities.  e GPI uses fatality and injury statistics
published in the Statistical Abstract and by the National
Center for Statistical Analysis (NCSA, 2004). Economic
losses are based on estimates by the National Safety Council
(NSC, 2004).  e fi gures cover motor vehicle accidents
on and off the road and all injuries regardless of length of
disability and address wage loss, legal, medical, hospital,
and funeral expenses, and insurance administration costs.
Property losses are not included because of signifi cant data
gaps. NSC estimates that on average each motor vehicle
death represents $1,130,000 in economic losses and each
injury $49,700 in 2004 dollars. Economic losses peaked in
1996 at $206.98 billion. In 2004, such losses amounted to
$175.18 billion. NSC attributes this decline to advances in
vehicle safety.
Column Q – Cost 0f Water Pollution
Water is the one of the most precious of all environmental
assets, yet the national income accounts provide neither
an inventory of the quantity or quality of water resources
nor an account for the cost of damage to water quality. In
the GPI framework, the costs of water pollution arise from
(1) damage to water quality and (2) damage from siltation
which reduces the life span of water impoundments or
channels. Although this may involve some double counting
(insofar as siltation also damages water quality), on the
whole the estimates in this column understate damage
because of the lack of data on nonpoint sources of pollution.
 e cost of damage to water quality begins with a 1972
estimate of $12.0 billion, or $39.7 billion in 2000 dollars.
 is is based on the upper range of estimates in three
studies of point source damage to recreation, aesthetics,
ecology, property values, and household and industrial
water supplies (Freeman, 1982). Between 1950 and 1972,
damage from water pollution is assumed to grow 3 percent
per year, from $20.3 billion to $39.7 billion. Between
1972 and 1992, damages are assumed to increase at a rate
corresponding to the per capita increase in spending on
water pollution abatement, which grew from $324 in 1972
to $570 in 1992 (Rutledge and Vogan, 1994). We assume
per capita pollution abatement expenditures are roughly
correlated with the magnitude of actual water quality
damage. After 1992, water pollution abatement data is
no longer available, and pollution damage is assumed to
continue growing at 3% per year from $71.8 billion in
1992 to $102.3 billion in 2004.
Erosion imposes costs in the form of reduced river
navigability, siltation of water impoundments, increased
fl ooding, reduced recreational activities, and degraded
fi sheries. Uri and Lewis (1999) estimated the social cost
of soil erosion to be $17.81 billion in 1997. In that year,
we estimate total erosion from agriculture and forestry
operations to be 2.02 billion tons. Adjusting for infl ation
yields a damage estimate of $8.81 per ton of erosion. As
sources of siltation, we examined erosion from farming
(960 million tons in 2004) and logging (925 million tons in
2004). Tons of cropland erosion comes from the National
Resources Inventory, conducted by the Soil Conservation
Service in conjunction with Iowa State University from
1982 to 2003. From 1950 to 1981, we estimate that erosion
decreased by an average of 1 percent per year, based on the
trend visible in the NRI data.
Tons of logging-related erosion comes from an estimate by
Hagerman (1992) that forest operations contribute 231 tons
of sediment per acre per year. We have assumed Hagerman’s
estimate applies to clear cuts, which are 38 percent of U.S.
harvests (USDA, 2006). We further assumed that selective
cutting contributes only half as much sediment as clear cuts,
or 115.5 tons per acre. To estimate total acreage of forest
operations, we relied on 1950-2002 statistics published by
Adams et al. (2006). Combining damage to water quality
and damage due to siltation we estimate the total cost of
water pollution to be $119.72 billion in 2004.
Column R – Cost 0f Air Pollution
 e annual economic cost of air pollution to households,
infrastructure, the environment, and human health is a
typical example of environmental costs that lie outside the
boundary of the traditional national accounts. It represents a
signifi cant omission from conventional economic indicators
like the GDP.  e GPI relies on Myrick Freeman’s (1982)

The Genuine Progress Indicator 2006 Redefi ning Progress
14
analysis of the cost of air pollution. His fi gure of $30
billion in 1972 dollars is converted to $99.34 billion in
year 2000 dollars.  e damage estimate includes damage
to agricultural vegetation, materials damage (paint, metals,
rubber), costs of cleaning soiled goods, acid rain damage
(aquatic and forest), urban disamenities (reduced property
values and wage diff erentials), and aesthetics.
We estimate the annual cost of air pollution for years
other than 1970 by extrapolating the $99.34 billion fi gure
according to the relative change in air pollution levels. To
do so, we measure the relative change in air quality using an
index of ambient air pollution levels based on 1975-1996
data from EPA (EPA, 1998). For earlier years, ambient air
conditions are assumed to have deteriorated by 1 percent
per year in the 1950s and by 2.4 percent per year in the
1960s, and to have improved by 3.0 percent per year from
1971 to 2004 (as a result of the Clean Air Act of 1970).  e
2004 fi gures for NOX, SO2, and particulates are projected
based on the trend 1990–1996. Indices were created for
ambient levels of particulates (PM), sulfur dioxide (SOX),
and nitrogen dioxide (NOX). In each case, the year 1975
(the year the EPA began collecting the data) is set equal
to 100. A single index number of ambient air pollution is
created for each year by averaging these three indexes. A
value greater than 100 implies an increase in air pollution,
while a value less than 100 signifi es a decline in air
pollution. To calculate the cost of air pollution, we divide
the ambient air pollution index of the given year by the
index for 1970 and multiply the result by our estimate for
the cost of air pollution in 1970 ($99.34 billion).
Since 1975, the decline in absolute emissions of sulfur
dioxide and particulates (which outweigh the small increase
in nitrogen dioxide emissions) suggests a decreasing
economic cost of air pollution for these three emissions.  e
GPI account estimates the cost of air pollution to be $40.05
billion in 2004, signifi cantly less than the all time high of
$99.34 billion in 1970.
Column S – Cost 0f Noise Pollution
While the U.S. has noise pollution regulations, there
are no offi cial inventories of its extent or severity.  e
damage caused by noise pollution in the U.S. in 1972 was
estimated at $4 billion by the World Health Organization
(Congressional Quarterly, Inc. 1972). Starting with that
estimate, we assumed that the quality of the auditory
environment declined by 3 percent per year from 1950
to 1972, based on industrialization and increased noise
emissions from motor vehicles and airplanes. From 1972
to 1994, noise abatement regulations are assumed to have
reduced the rate of deterioration to 1 percent per year,
but not to have improved it. With no new noise pollution
data since the 1995 GPI estimates, we assume a constant
rate of decline in the auditory environment at 1 percent
per annum.  e GPI account estimates the cost of noise
pollution in 2004 at $18.21 billion.
Column T – Loss of Wetlands
Wetlands contain some of the most productive habitat in
the world. Yet their value is not represented in economic
accounts because the benefi ts—such as regulating and
purifying water and providing habitat for fi sh and
waterfowl—are generally “public goods,” for which there is
no overt price. When a farmer drains and fi lls a marsh, the
GDP rises by the increased output of the farm. However,
the loss of services from the wetland goes uncounted.  e
GPI rectifi es this by estimating the value of the services
that are given up when wetlands acreage is converted to
other purposes. To do this, multiply wetland loss in each
year by $914, the value of an acre of wetland as estimated
by a meta-analysis of wetland valuation studies reviewed
by Woodward and Wui (2000). We add this value to an
assumed baseline of wetland loss prior to 1950, since we
continue to incur the cost of not having these wetlands
present to perform essential services such as water fi ltration.
 e U.S. Fish and Wildlife Service (USFWS) estimates that
136 million acres of wetlands were fi lled in North America
from the colonial period to 1950. Acreage declined from
an original 395 million (including the contiguous lower 48
states and Alaska) in the 1780s to about 259 million acres in
1950—a loss amounting to 60 acres an hour for 200 years
(USFWS, 1997). Our estimates of acres of wetland loss
are based on USFWS data published in Status and Trends
of Wetlands in the Conterminous United States (USFWS,
1997).  eir most recent study estimated the loss of
wetlands at 462,000 acres per year through 1975, 294,000
acres per year from 1976 to 1984, and 121,000 acres per
year in subsequent years. Each of these fi gures includes
4,000 acres per year lost in Alaska while the remaining acres
were lost in the lower 48 states. We extrapolate the loss
fi gures since 1995 by using the rate of change from 1985
to 1995.  e GPI estimates the value of ecological services
lost due to the accumulated loss of wetlands in 2004 to be
$53.26 billion.
Column U – Loss of Farmland
Loss of either natural or human-built capital generates
costs to both present and future generations in the form
of lost services from that capital. By destroying farmland,
we are losing a vital ecosystem service - sustainable food

The Genuine Progress Indicator 2006 Redefi ning Progress
15
supply. Farmland losses also generate costs in the form of
lost scenic, aesthetic, and historic values, increased fl ooding,
deterioration in water quality, and degradation of wildlife
habitat. In the GPI accounts, we address farmland losses
resulting from urbanization and lost productivity.
Obtaining accurate time series data on farmland loss is
a surprisingly diffi cult task. Variations in time periods
studied, how farmland is defi ned, and how acreage is
counted are considerable. For this reason, we combined
data from a number of sources including the American
Farmland Trust, the National Agricultural Statistics Service,
the USDA’s National Agricultural Lands Study and the
Farm Information Center. Using these data sets, we
estimate the average annual conversion of prime farmland
to urbanization to average nearly 400,000 acres per year
since 1950.
To put a price tag on this loss, we added the average
value ($5,459) from three contingent valuation studies
summarized by Ready et al. (1997) that considered lost
amenity values to the Costanza et al. (1997) fi gure of
$41.34 per acre for lost ecosystem services. We then
multiplied the resulting value ($5,501 in year 2000 dollars)
by an index that defl ates this value in years before 2000
and infl ates it after to account for relative scarcity. By 2004,
the GPI accounts assign a cost of $6,203 for every acre of
farmland lost to urbanization.  e cumulative loss fi gure
is obtained by multiplying each year’s value per acre by the
acres lost in that year, then adding it to the previous year’s
loss. As with wetlands, the reason for tracking cumulative,
and not marginal losses, is the fact that we are still incurring
the costs of farmland lost in 1950, 1960, etc. because we
are no longer receiving the stream of benefi ts these lands
once conferred (and still could if they are restored).  e GPI
assumes that the initial pre-1950 loss was roughly $3.31
billion, a fi gure that has grown to $91.19 billion in 2004.
Urbanization removes the productive potential of farmland
in a highly visible way. But it may not be as serious in
the long run as the deterioration of soil due to poor
management.  e decline of soil quality over the past
forty years has been masked by higher inputs of fertilizer,
pesticides, and fuel. In addition, soil depletion is not
necessarily linear. It may not show up gradually in yield
reductions, but rather in a sudden and irreversible decline.
Agricultural productivity losses from erosion have been
estimated at $1.3 billion per year, or $2.5 billion in 2000
dollars (USDA, 1985). In 1985, erosion calculations from
column Q show 2.9 million tons of cropland erosion in that
year, which translates into roughly $.86 per ton. We assume
the cumulative damage prior to 1950 was $16.3 billion, and
add to that by multiplying the $.86 fi gure by the annual
erosion estimated from column Q.
 e damage to soil from compaction by heavy machinery
in 1980 was estimated at $3.0 billion in 1980 dollars
(Sampson, 1981), or $5.5 billion in 2000 dollars. We
assumed a 3 percent increase per year in the losses due to
compaction prior to and following 1980.  e 2004 estimate
of the cost of soil compaction is $11.27 billion.  e total
economic costs of the loss of farmland to urbanization,
soil erosion, and soil compaction in the GPI is estimated
at $263.86 billion in 2004 having risen steadily from an
estimated $25.80 billion in 1950.
Column V – Loss of Primary Forests and Damage from
Logging Roads
Whenever native, or primary forest land is cut for timber,
converted into tree plantations, or cleared to build a road,
that forest’s ability to control fl oods, purify air and water,
maintain biological and genetic diversity, provide habitat
for sensitive species, produce non-timber forest products or
provide scenic, recreational, and aesthetic values to nearby
communities is impaired or lost forever.  e GPI accounts
measure this loss by assigning a price tag to year by year
estimates of key primary forest losses and adding such
losses to the cumulative damage from previous years. In
particular, we assign costs to the loss of longleaf pine forests
in the southeastern U.S., old growth forests in the Pacifi c
Northwest, Sierras, and southeast Alaska, and inventoried
roadless areas on national forests.
While certainly debatable, we assume relatively little
overlap in the damage assigned to loss of roadless areas
and old growth forest largely because roadless areas tend
to be located in higher, less productive areas not typically
included in inventories of low elevation, high productivity
old growth stands. While there are other critical forest types
lost in the United States each year, these primary forest
types are particularly rich in biological diversity, have been
extensively studied, and have reasonable estimates of both
extent and value on which GPI accounts can be based. We
also incorporate costs associated with national forest logging
roads, which are continuing sources of sedimentation,
landslides, fi res, and habitat fragmentation.
For longleaf pine, data points for original extent, 1935,
1955, 1985, and 2003 as well as rate of loss in this period
are drawn from Outcalt and Sheffi ed (1996) and the
USFWS (2003). Out of an original extent of 60 million
acres, only 2.9 million remain in 2004. In the Pacifi c
Northwest, the Forest Service estimates that between 60

The Genuine Progress Indicator 2006 Redefi ning Progress
16
and 70% (65% as a mid point) or 19.57 million acres of
forests within the range of northern spotted owl were in
late successional/old growth condition during the pre-
industrial era (USDA, 2005). In 1950, we assume that
most old growth on private lands had been taken and that
national forest boundaries provide a crude proxy for what
remained. In 1994, the Forest Service found that only 7.87
million acres remained. Previous years assume a rate of loss
of 180,000 acres per year back to 1950. Post 1994 fi gures
are based on losses due to logging and fi res reported by the
USDA (2005).
In the Sierras, data points for 1945 and 1993 were
estimated by Beardsley, et al. (1999). Remaining points were
interpolated. In Alaska, we assume that nearly all timber
harvests on the Tongass National Forest back to 1950
involved the clearing of old growth temperate rainforest.
Harvest data were taken from spreadsheets provided by the
Tongass National Forest. For inventoried roadless areas,
we assume an original extent equivalent to the extent of
national forest system lands in the western United States
(167 million acres). In 1979, the Forest Service inventoried
62.02 million (USFS, 1980). In 2000, that fi gure fell to
58.51. For intervening years, we incorporated a variety of
Forest Service data points on new road construction and
multiplied these fi gures by the amount of roadless area loss
per mile of new road construction (26.44 acres per mile).
Taken together, GPI accounts show a cumulative primary
forest loss of 74.56 million acres in 2004. To assign a cost,
we take the Costanza et al. (1997) fi gure of $134 per acre
for ecosystem services not including raw materials and
climate regulation (since young forests also provide these
functions) plus 3 times that amount for passive use values
as estimated by numerous studies including Vincent, et al.
(1995). An example of passive use values is the willingness
to pay for preservation of old growth forest habitat
critical to the northern and Mexican spotted owls, a value
determined through contingent valuation surveys. In 2004,
the GPI accounts estimate the magnitude of costs associated
with primary forest loss to be $39.89 billion.
 e calculation of losses due to national forest logging
roads is based on the total stock of roads in any given
year. A mile of forest road with a 60-foot right-of-way
covers approximately 7 acres of land. If the impacts such
as noise, edge eff ects, and runoff are included, a mile of
road aff ects at least 500 acres of land.  is provides a
very rough estimate of the environmental costs because
the damage caused by roads depends on many factors
including age, location, slope, the quality of construction,
and the frequency of maintenance. Nevertheless, even the
best roads cause some continuing ecological disruption by
breaking up the landscape, raising erosion levels, disturbing
downstream fi sheries, and generally increasing the level of
human activity. Estimates of total miles of forest roads are
taken from twelve separate Forest Service point estimates
from 1955–2004. In the 1995 GPI, we assumed that the
cost of damages to forests caused by roads from 1950 to
1959 was $10,000 per mile in 1982 dollars.  at fi gure is
here converted to 2000 dollars, or $15,939 per mile. From
1960 to 1979, the cost per mile is assumed to decline on a
straight-line basis to $7,500 ($11,954 in year 2000 dollars)
per mile due to improvements in road standards. We
estimate the cost of ecological damage due to roads at $4.62
billion in 2004. Added together, the GPI accounts show
that the loss of primary forest and damage from logging
roads amounts to $50.64 billion in 2004.
Column W – Depletion of Nonrenewable Energy Resources
 e depletion of nonrenewable resources is a cost shifted
to future generations that should be borne in the present.
Nonrenewable natural capital cannot be increased, it can
only be diminished. As Herman Daly notes (1996) in
Beyond Economic Growth, for nonrenewable capital the
question is not how to invest, but how best to liquidate
the inventory and what to do with the net fi nancial wealth
realized from that liquidation. Our current accounting
system counts this liquidation of natural capital wealth
as income “which is clearly wrong, because it is not a
permanent or sustainable source of consumption” (Daly,
1996). A prudent approach to sustaining the income
and well-being of America’s households would require
investment of a portion of the net rents derived from
mining nonrenewable natural capital into sustainable
renewable energy and productivity or energy effi ciency
gains. In this vein, the GPI uses estimates of renewable
energy replacement costs as an approximation for the costs
of depleting nonrenewable energy reserves.
To calculate replacement costs, we rely on the costs of
biomass fuel production. While this approach is debatable,
we believe it is both intuitive and reasonable, since biomass
fuel was the largest share (47%) of the renewable energy
market in 2004 according to the most recent annual data
compiled by the Energy Information Administration. We
assume a nominal replacement cost of $99.10 per barrel
based on a USDA (1988) study that took into account the
eff ects of subsidies and increasing marginal costs as biomass
demand and production increase. To account for scarcity,
we decrease that cost by 3% per year prior to 1988, and
increase by the same rate in subsequent years. We convert
annual nonrenewable energy consumption in quadrillion
BTUs to equivalent barrels of oil, and then multiply by the
adjusted annual replacement cost fi gure.

The Genuine Progress Indicator 2006 Redefi ning Progress
17
Using this methodology, the GPI accounts show the cost
of replacing nonrenewable energy production to be $1.76
trillion in 2004.  is represents the largest cost included
in the GPI account.  e fact that, after almost fi fty years
of nonrenewable energy liquidation, renewable energy
makes up just 6.12% percent of total energy consumption
in 2004 suggests insuffi cient investment of nonrenewable
resource rents into sustainable energy substitutes for
the well-being of future Americans.  e longer we defer
investment in renewable energy resources, the greater the
economic impact on the well-being of current and future
American households.
Column X – Carbon Dioxide Emissions Damage
Few scientists dispute the link between carbon dioxide
emissions and global warming or the link between global
warming and increasing incidence and severity of damaging
storms, fl oods, and droughts. And as hurricane Katrina
illustrated all too well, this erratic weather is exacting an
enormous economic toll each year on our households,
infrastructure, and natural capital. As the incidence of severe
weather events escalate the costs in insurance payouts and
replacing lost or damaged homes, buildings, livestock, and
other household resources mount. Ironically, these natural
disturbances result in a positive feedback loop whereby
increasing frequency and intensity of storms and other
severe weather leads to increasing use of natural capital
resources as we rebuild shattered homes and infrastructure
in the aftermath. Yet neither the cost of our impacts on the
Earth’s climate, nor the increasing costs of cleaning up after
the storm, nor the increased depletion of nature’s capital is
accounted for by GDP.  e GPI attempts to address this
oversight by assigning costs to carbon emissions.
 ere are many ongoing studies that attempt to calculate
economic damages per ton of carbon emitted into the
atmosphere through our burning of fossil fuels. In one
recent meta-analysis of 103 separate studies, Tol (2005)
found a mean of $93 per metric tonne, or $89.57 in year
2000 dollars.  ough hotly debated, we adopt this fi gure
as a conservative starting point for incorporating carbon
emissions damage into GPI accounts.
 e GPI relies on carbon emissions data reported by the
Oak Ridge National Laboratory. We assume that only
excess emissions are contributing to global warming and
deduct the portion of these emissions sequestered by the
world’s terrestrial and aquatic ecosystems. Globally, the
Intergovernmental Panel on Climate Change estimates the
Earth’s carbon sequestration capacity to be 3 gigatonnes
(Gt) carbon per year (IPCC, 2000). Worldwide, overshoot
of this sequestration capacity began in 1964 (not counting
natural sources of carbon dioxide), and has now risen to
58%, or roughly 4 Gt. In the GPI accounts, we assign costs
to a percentage of U.S. emissions identical to the global
overshoot percentage. We also assume that, due to positive
feedback eff ects, marginal damage increases over time. To
account for this, we taper the marginal damage costs down
from $89.57 in 2004 to just over zero in 1964—the fi rst
year of carbon overshoot. Finally, we assume that marginal
damage from carbon emissions are cumulative so that costs
incurred in one year continue to be incurred the next year.
Using this approach, we estimate carbon emissions damage
to be $1.18 trillion in 2004.  is is the second largest cost
included in the GPI, arguably, as it should be. After all,
global warming is a phenomenon that threatens hundreds
of millions of lives, entire cities, and the planetary economic
system like no other threat in human history and the United
States is by far the single greatest source of carbon emissions
implicated in that warming.
Column Y – Cost of Ozone Depletion
While annual production of CFCs may have declined
dramatically, the cumulative impacts on the depletion of
the earth’s ozone layer continues. According to NOAA’s
Climate Prediction Center, “[e]xtensive ozone depletion
was again observed over Antarctica during the Southern
Hemisphere winter-spring of 2005, with widespread total
ozone anomalies of 45 percent or more below the 1979-
1986 base period” (NOAA, 2006). In September 2005,
the area covered by extremely low total ozone values of less
than 220 Dobson Units, defi ned as the Antarctic “ozone
hole” area reached maximum size of 25 million square
kilometers, with an average size of more than 22 million
square miles, among the largest sizes of recent years.  ere
are no defi nitive studies showing the combined health and
ecological consequences of ozone depletion over the next
half century. However, scientists warn that the ozone loss
could result in increased exposure to harmful solar radiation
that can destroy plants and cause cataracts and skin cancer
in humans. Given the potentially catastrophic eff ects on all
forms of life, the GPI includes an estimate refl ecting our
expectation of the economic costs associated with this long-
term environmental problem - $49,669 per tonne.
 e calculation for the cost of ozone depletion involves
multiplying the U.S. share of cumulative world production
of CFCs 11, 12, 113,114 and 115 by $49,669 per metric
tonne in year 2000 dollars. To calculate the U.S. share, we
combined data sets from the Alternative Fluorocarbons
Environmental Acceptability Study (
www.afeas.org), the

The Genuine Progress Indicator 2006 Redefi ning Progress
18
EPA, the United Nations Environmental Programme, and
the U.S. Congress.  e GPI account estimates the cost of
ozone depletion in 2004 at $478.92 billion. Since CFC
production in the U.S. has all but halted, this cost fi gure has
remained basically unchanged since 1995.
Column Z – Net Capital Investment
For an economy to prosper over time, the supply of capital
(buildings, machinery, and other infrastructure) must be
maintained and increased to meet the demands of increased
population. If this does not occur, the society is consuming
its capital as income.  us, one element of economic
sustainability is constant or increasing quantities of capital
available for each worker.  e GPI calculates changes
in the stock of capital (or net capital growth) by adding
the amount of new capital stock (increases in net stock
of private nonresidential fi xed reproducible capital) and
subtracting the capital requirement, which is the amount
necessary to maintain the same level of capital per worker.
 e aim of this column is to estimate increases in the stock
of capital available per worker.
 e capital requirement is estimated by multiplying the
percent change in the labor force by the stock of capital
from the previous year. Labor force statistics are provided
by the Bureau of Labor Statistics while capital stock fi gures
are taken from the Bureau of Economic Analysis. A fi ve-year
rolling average of changes in labor force and capital is used
to smooth out year to year fl uctuations.  e GPI considers
an increase in the capital stock available to workers or
households as a positive adjustment in the GPI account. In
2004 growth in the net capital stock was $388.3 billion,
down from its peak of $490.29 billion in 2001.
Column AA – Net Foreign Borrowing
 e economic sustainability of a nation is also aff ected by
the extent to which it relies on foreign funding to fi nance its
current consumption. A nation that borrows from abroad to
pay for a spending spree will feel rich for a short time. But
the illusion of wealth will vanish when the debt comes due
or when the value of the currency drops as foreign investors
lose confi dence in that nation’s ability to repay its loans.
 is column measures the amount that Americans invest
overseas minus the amount foreigners invest in the United
States, or the net change in our international investment
position.  e annual change indicates whether the U.S. is
moving in the direction of net lending (if positive) or net
borrowing (if negative). If the change is positive, the U.S.
has in eff ect increased its capital assets. If it is negative,
part of U.S. capital formation is in fact based on wealth
borrowed from abroad that must eventually be repaid with
interest. We have thus included annual changes in the net
international position as a measure of the long-term viability
of our economy.
 e annual fi gures for the market value of the U.S. net
international investment position (NIP) from the Bureau
of Economic Analysis show a rapid deterioration through
the 1980s through 2004. From a net lending position of
$257 billion 1983, the U.S. has slipped to a net borrowing
position of $2.54 trillion in 2004.  e GPI accounts track
the change in the fi ve year rolling average of NIP and add
or subtract this change depending on its sign. In 2004, the
GPI deducts $254 billion.
Column AB – The Genuine Progress Indicator
 e Genuine Progress Indicator (GPI) starts with personal
consumption adjusted for income inequality (column D),
adds fi ve columns (E through I), subtracts sixteen columns
(J through Y), and adds two columns (Z and AA).  e
result is a more honest account of the genuine economic
progress of the U.S. economy and the state of its households
than GDP because it takes into account the benefi ts of
non-market activities, education, and services from capital
and the costs associated with inequality, environmental
degradation, and a weakening international position. While
incomplete, the GPI demonstrates the value of services
derived from real wealth and assets that one could argue
are more meaningful in defi ning the well-being of the
nation’s households than those tallied by the GDP.  e
GPI accounting exercise demonstrates the complexity of
accounting for real wealth. If as many economists and
statisticians were devoted to this more complete accounting
of the state of the economy as they are to GDP we might
be empowered with better information to manage the
collective well being of the nation more prudently.
Column AC – Per Capita GPI
Per capita GPI is calculated by dividing the GPI by the U.S.
population. Annual population fi gures are taken from the
Economic Report of the President.
Column AD – Per Capita GDP
 e value of the GDP also comes from the Economic
Report of the President. Per capita GDP is the GDP divided
by the population.
Results and Implications
In “An Updated GPI Methodology,” we discussed column
by column results and some implications drawn from those
results. Here, we present the GPI results in aggregate. Table
1 (page 21) provides a detailed year by year accounting of

The Genuine Progress Indicator 2006 Redefi ning Progress
19
all GPI columns for the 1950 to 2004 period. In Figure 3
below, we show GPI and GDP side by side. As shown in
Table 1 and Figure 3, real GPI has increased from $1.31
trillion in 1950 to $4.42 trillion in 2004.  is corresponds
to an average growth rate of 4% for the period. By
comparison, GDP grew steadily from $1.78 trillion in 1950
to $10.76 trillion in 2004, an average annual growth rate of
roughly 9%.
Of course, these fi gures mask the eff ects of increasing
population.  us, it is important to look at both GPI and
GDP fi gures in per capita terms. As shown in Table 1 and
Figure 4 (page 20), GPI per capita has barely moved since
1978, remaining near $15,000 since that time. Over the
period 1950–2004, GPI grew at an extremely sluggish
rate of just 1.33%. In contrast, GDP per capita rose
precipitously from $11,672 in 1950 to $36,596 in 2004—
an annual growth rate of 3.81%. It is also critical to look at
annual growth rates for each year so that important trends
within particular time periods are not overshadowed by the
full time series. Figure 5 (page 23) compares annual GDP
and GPI per capita growth rates using a rolling three year
average to smooth out year to year fl uctuations. Here, we
fi nd a rather striking trend: while GDP growth rates have
more or less fl uctuated within a positive range GPI growth
rates fall into two distinct periods. In the fi rst period,
spanning 1950 to 1980, GPI per capita growth rates more
or less match those of the GDP and are generally positive,
ranging as high as 4%. Beginning in 1980, GPI growth
rates are commonly negative, bottoming out at -1.64% in
1994. GPI per capita has more or less stagnated since 1978
when it surpassed $15,000 for the fi rst time. Importantly,
what this implies is that since 1980 or so the marginal
benefi ts associated with growth in personal consumption
expenditures, non-market time, and capital services have
been off set by the marginal costs associated with income
inequality, natural capital depletion, consumer durable
expenditures, defensive expenditures, undesirable side eff ects
of growth, and net foreign borrowing.  is trend, found
in many of the GPI and ISEW studies completed over the
past fi fteen years or so has been put forth as evidence of a
“threshold” eff ect. According to Max-Neef (1995):
For every society there seems to be a period in which
economic growth brings about an improvement in the
quality of life, but only up to a point—the threshold
point—beyond which, if there is more economic growth,
quality of life may begin to deteriorate (Max-Neef,
1995, pg. 117).
Dietz and Neumayer (2006) argue that the threshold
eff ect found in most GPI and ISEW studies is less a true
refl ection of welfare growth and decline and probably no
more than an artifact of methodological fl aws. As a case
in point, they argue that assumptions made about growth
in nonrenewable resource depletion costs and long term
environmental damage make the threshold eff ect all but
certain. While their criticisms certainly have merit and
warrant closer inspection of the relationship between the
threshold eff ect and actual column by column assumptions,
we believe this update has at least partially remedied some of
those concerns. For instance, in the calculation of long term
environmental damage, we have discarded any assumptions
about growth in this damage and, instead, tied damage
calculations to actual carbon emissions and the estimated
marginal social costs of those emissions. In several other
columns, assumed growth rates were replaced by actual data
so it remains unclear the extent to which the “hard wired”
threshold eff ect hypothesis Dietz and Neumayer (2006)
suggest still applies.
Figures 6 and 7 (page 24) show the growth and relative
importance of GPI contributions and GPI deductions over
time. Following Lawn (2005) and for the sake of graphical
clarity, we have condensed GPI columns into several groups.
On the contributions side, we have left weighted personal
consumption expenditures alone, and grouped columns E
through I into two categories: non-market time (columns E,
F, and G) and capital services (columns H and I). Figure 6
charts trends in these three categories of GPI contributions.
While the absolute magnitude of each has grown steadily,
the relative contribution of personal consumption
expenditures and non-market time have changed. In 1950,
personal consumption expenditures accounted for 51% of
all positive contributions to the GPI. In 2004, that share
had risen to 59%.  e increasing relevance of personal
consumption expenditures has been accompanied by a
corresponding decrease in the relevance of non-market

The Genuine Progress Indicator 2006 Redefi ning Progress
20
time spent on volunteer activities, parenting, and higher
education.  is share has fallen from 41.21% in 1950 to
32.80% in 2004.
As briefl y noted in the discussion of column E, this may
refl ect an increasing reliance on the market to provide
services formerly contributed by households (such as
home cooking) and a general decrease in our availability
to volunteer, extend our formal or informal education,
or participate in civic activities. Spending more money
for more goods and services each year is seen as a sign
of a healthy economy and a well-to-do society—at least
so the GDP account tells us.  e fact that the GDP has
risen relentlessly and per capita personal consumption
expenditures have more than tripled since 1950 would
suggest that America is becoming more prosperous.
 ere is little doubt that we have achieved unprecedented
material gains and improved living standards. Yet the
GPI account indicates that while per capita personal
consumption of goods and services continues to rise,
average real hourly wages have declined, personal
indebtedness has risen, personal savings rates have
fallen, and quality time with our families, participating
constructively in civic aff airs, or pursuing self betterment
has steadily eroded. Yet according to the key yardstick of
the economy, the GDP, all is well with the households of
the nation.  e declining share of non-market time in the
GPI accounts is worrisome trend indicating that while
our affl uence may be on the rise, both our personal and
collective sense of well being may be suff ering.
As for GPI deductions, one signifi cant trend that jumps
out dramatically in Figure 7—GPI deductions—is the
growing relevance of costs associated with depletion of and
damage to natural capital.  is share, which includes loss
of wetlands, farmland, and primary forest, depletion of
oil reserves, carbon dioxide and ozone damage rose from
35.45% of GPI deductions in 1950 to 59.32% in 2004.
 e largest component of this $3.8 trillion dollar cost is
the $1.18 trillion cost associated with excess carbon dioxide
emissions. One reason why this cost is so large is simply the
fact that the damage is assumed to be cumulative. In other
words, the GPI assumes that we are still incurring the cost
of excess carbon emissions from 1950 and later. Dietz and
Neumayer (2006) take issue with this and argue, instead,
for counting only the marginal, not cumulative social cost
of carbon emissions. In support of their argument, they
point out that most marginal cost values incorporate the
present value of future costs so tracking cumulative instead
of marginal costs involves double counting.
However, global warming is replete with positive feedback
loops. For example, warming induces greater carbon
emissions by way of increasing forest fi re extent and severity
and thawing of the arctic tundra which leads to even more
warming. Ice sheet melting diminishes the albedo eff ect
which, in turn, leads to greater oceanic warming. Given the
existence of these positive feedback eff ects it would clearly
be inaccurate to assume constant marginal costs or somehow
neglect the importance of atmospheric thresholds for
carbon dioxide beyond which catastrophic eff ects are more
likely. To their credit, Dietz and Neumayer (2006, pg. 200)
suggest increasing the marginal damage fi gure over time in
recognition of the fact that “the marginal social cost of each
tonne of emissions is a positive function of the accumulated
stock of carbon in the atmosphere.” So something beyond
constant marginal cost accounting is appropriate, but it is
not clear what that is. Currently, the GPI treats the cost
of carbon emissions as cumulative, and increasing over
time, but reduces the magnitude of such costs by counting
only excess emissions over and above the Earth’s ability to
sequester those emissions. Given the ongoing murkiness
over exactly how to deal with carbon emissions, we suggest
that the methodology presented in this 2006 GPI update
be viewed as simply one approach among many potential
approaches that should be properly vetted in the years ahead.
Using the GPI as a Guide to Public Policy
Given the subjective aspects of the GPI and lingering
doubts as to its methodological rigor, some have argued its
policy irrelevance (Neumayer, 1999). For example, Carson
and Young (1994, pg. 112) have suggested:
…a single, dimension, aggregate measure of sustainable
welfare will be of little direct use in guiding, shaping, or
choosing among government policies because the factors
determining welfare cannot be reduced and combined
into a single measure that would command widespread
agreement and acceptance.

The Genuine Progress Indicator 2006 Redefi ning Progress
Column A Column B + Column C +/- Column D + Column E + Column F + Column G + Column H + Column I + Column J - Column K - Column L - Column M - Column N - Column O -
Year
Personal
consumption
Income
distribution
index
Weighted
personal
consumption
Value of
housework and
parenting
Value of higher
education
Value of volunteer
work
Services of
consumer
durables
Services of
highways
Costs of
crime
Loss of
leisure time
Costs of under-
employment
Cost of consumer
durables
Cost of
commuting
Cost of household
pollution abatement
1950 1,152.80 107.97 1,067.73 749.48 84.35 30.72 133.85 32.01 8.82 12.07 15.88 77.08 141.84 0.02
1951 1,171.20 103.59 1,130.56 771.11 91.12 30.82 138.50 33.76 9.18 11.39 16.97 70.40 141.96 0.03
1952 1,208.20 105.04 1,150.21 793.36 97.89 30.93 144.39 34.85 9.49 10.76 18.14 68.34 141.08 0.05
1953 1,265.70 102.53 1,234.42 816.26 101.26 31.03 151.87 31.77 9.77 10.26 19.39 76.89 145.63 0.07
1954 1,291.40 106.08 1,217.39 839.82 104.63 31.14 154.24 32.55 10.09 9.70 20.72 76.69 142.09 0.10
1955 1,385.50 103.84 1,334.26 864.05 108.01 31.24 164.69 34.70 10.37 9.23 22.14 93.63 151.36 0.14
1956 1,425.40 102.43 1,391.59 888.99 111.38 31.35 174.37 37.72 10.68 8.76 23.66 89.75 152.23 0.20
1957 1,460.70 100.47 1,453.94 914.65 114.75 31.46 179.30 37.21 11.07 8.12 25.29 90.74 153.62 0.29
1958 1,472.30 101.33 1,452.96 941.05 119.25 31.57 179.18 39.15 11.47 7.52 27.02 83.26 148.60 0.41
1959 1,554.60 103.38 1,503.84 968.21 123.74 31.67 184.34 39.19 11.80 6.90 28.88 93.51 155.05 0.58
1960 1,597.40 104.24 1,532.40 996.15 121.87 31.78 186.35 40.40 12.20 6.31 30.86 95.28 158.31 0.83
1961 1,630.30 107.19 1,521.00 1,024.90 132.95 31.89 187.21 42.23 12.62 5.67 32.98 91.77 156.69 1.18
1962 1,711.10 103.85 1,647.60 1,054.48 144.03 32.00 191.17 44.87 13.03 4.97 35.25 102.50 161.89 1.69
1963 1,781.60 103.85 1,715.49 1,084.91 146.78 32.11 199.35 47.40 13.49 4.32 37.67 112.38 166.77 1.88
1964 1,888.40 103.57 1,823.34 1,116.22 149.52 32.22 206.76 48.68 13.94 3.66 40.26 122.88 171.55 2.09
1965 2,007.70 102.15 1,965.38 1,148.43 155.87 32.33 215.30 52.02 14.44 2.98 43.02 138.45 178.77 2.32
1966 2,121.80 100.18 2,117.92 1,181.58 163.39 37.20 229.75 55.53 14.96 2.27 45.98 150.05 182.99 2.59
1967 2,185.00 102.84 2,124.76 1,215.68 168.80 42.80 244.98 58.41 15.53 1.54 49.14 152.29 185.63 2.89
1968 2,310.50 100.00 2,310.50 1,250.76 178.74 49.25 262.07 60.10 16.09 0.78 52.51 169.22 193.74 3.22
1969 2,396.40 100.77 2,378.01 1,286.86 184.56 56.66 273.60 63.85 16.75 0.00 56.12 175.07 199.51 3.61
1970 2,451.90 101.55 2,414.56 1,324.00 192.99 65.20 280.82 68.89 17.44 0.00 59.73 169.50 198.85 4.05
1971 2,545.50 102.06 2,494.08 1,362.21 201.79 75.02 286.66 68.30 18.08 0.00 63.57 186.44 206.72 4.50
1972 2,701.30 103.35 2,613.73 1,401.53 213.82 86.32 300.26 68.97 18.69 0.00 67.65 210.16 217.71 4.73
1973 2,833.80 102.32 2,769.56 1,441.98 227.65 99.33 314.97 75.77 19.47 11.86 72.00 231.70 226.00 6.21
1974 2,812.30 101.80 2,762.46 1,483.59 244.80 114.29 328.85 91.40 20.26 12.04 76.63 215.79 219.17 7.02
1975 2,876.90 102.32 2,811.68 1,526.41 259.90 114.68 334.42 82.46 21.10 12.13 81.56 215.87 218.37 9.03
1976 3,035.50 102.58 2,959.23 1,570.46 279.94 115.07 347.09 76.44 21.88 12.29 86.80 243.42 233.50 10.03
1977 3,164.10 103.61 3,053.91 1,615.79 298.03 115.46 362.08 72.50 22.67 12.50 92.38 265.97 247.63 10.75
1978 3,303.10 103.61 3,188.07 1,662.42 309.31 115.85 381.92 71.99 23.57 12.75 98.32 279.99 260.71 11.21
1979 3,383.40 104.12 3,249.40 1,710.40 329.26 116.24 394.31 76.76 24.82 139.86 104.64 279.06 262.32 11.73
1980 3,374.10 103.87 3,248.51 1,759.76 355.09 116.64 393.25 83.46 26.18 146.34 111.36 257.21 255.24 12.78
1981 3,422.20 104.64 3,270.48 1,810.55 362.78 117.03 388.08 86.99 26.35 152.63 118.52 260.24 259.79 14.45
1982 3,470.30 106.19 3,268.15 1,837.46 384.80 117.43 383.21 85.06 26.87 158.78 126.14 260.07 262.76 13.96
1983 3,668.60 106.70 3,438.20 1,864.78 414.64 117.83 392.41 78.86 27.13 163.30 134.25 298.15 276.91 15.64
1984 3,863.30 106.96 3,611.95 1,892.50 429.79 118.23 410.07 74.92 27.84 168.06 142.88 341.71 298.28 17.04
1985 4,064.00 107.99 3,763.32 1,920.63 444.93 118.63 431.91 76.15 28.64 175.14 152.06 376.22 314.66 18.17
1986 4,228.90 109.54 3,860.74 1,949.18 455.82 119.04 466.81 80.76 29.08 183.08 161.83 412.55 334.02 18.50
1987 4,369.80 109.79 3,980.01 1,978.16 474.19 119.44 489.28 83.49 29.81 190.67 172.24 419.75 343.25 15.98
1988 4,546.90 110.05 4,131.61 2,007.56 492.59 119.85 511.38 82.91 30.48 198.07 183.31 445.05 358.47 16.93
1989 4,675.00 111.08 4,208.58 2,037.40 521.04 120.25 524.73 83.64 31.52 210.86 195.09 454.89 367.76 14.41
1990 4,770.30 110.31 4,324.48 2,067.69 532.66 118.56 530.85 84.47 32.21 220.28 189.23 453.52 372.45 11.59
1991 4,778.40 110.31 4,331.82 2,098.43 544.42 116.87 531.89 83.56 32.69 229.11 182.70 427.88 365.51 8.90
1992 4,934.80 111.86 4,411.76 2,129.62 549.39 116.07 535.89 83.45 33.07 240.23 177.58 453.00 376.51 9.20
1993 5,099.80 117.01 4,358.42 2,161.28 569.44 115.26 550.66 84.22 33.68 250.39 171.26 488.41 389.74 9.51
1994 5,290.70 117.53 4,501.74 2,193.41 584.70 117.83 569.59 87.56 35.97 271.52 167.14 529.38 406.98 10.88
1995 5,433.50 115.98 4,684.88 2,226.01 611.62 120.39 582.47 91.17 34.70 284.80 157.85 552.62 416.64 11.64
1996 5,619.40 117.27 4,791.93 2,259.10 634.69 120.66 593.68 93.44 33.73 297.98 154.71 595.94 429.03 12.44
1997 5,831.80 118.30 4,929.71 2,292.68 651.15 120.92 605.93 98.58 35.35 314.50 145.96 646.97 446.95 13.30
1998 6,125.80 117.53 5,212.30 2,326.77 671.57 121.19 628.95 100.69 34.00 329.61 134.00 720.29 467.88 14.23
1999 6,438.60 118.04 5,454.53 2,361.35 700.85 123.15 653.73 104.38 33.16 344.65 127.37 804.52 484.54 15.21
2000 6,739.40 119.07 5,659.93 2,396.46 717.52 125.10 678.35 107.80 31.04 363.30 124.48 863.30 495.19 16.26
2001 6,910.40 120.10 5,753.72 2,432.08 755.65 126.70 692.93 109.74 32.49 370.90 145.13 900.69 504.53 17.39
2002 7,099.30 119.07 5,962.18 2,468.23 779.14 128.20 711.23 111.50 34.64 376.93 171.83 964.75 512.03 18.60
2003 7,306.60 119.59 6,109.83 2,504.92 806.13 129.70 721.40 110.34 35.05 388.05 184.07 1,028.56 518.32 19.88
2004 7,588.60 120.10 6,318.41 2,542.16 827.98 131.30 743.72 111.55 34.22 401.92 176.96 1,089.91 522.61 21.26
Table 1
Genuine Progress Indicator 2006 Update
21

The Genuine Progress Indicator 2006 Redefi ning Progress
Column P - Column Q - Column R - Column S - Column T - Column U - Column V - Column W - Column X - Column Y - Column Z +/- Column AA +/- Column AB Column AC Column AD
Cost of auto
accidents
Cost of water
pollution
Cost of air
pollution
Cost of noise
pollution
Loss of
wetlands
Loss of
farmland
Loss of primary
forests
Resource
depletion
Carbon dioxide
emissions
damage
Cost of ozone
depletion
Net capital
investment
Net foreign
borrowing
Genuine Progress
Indicator
GPI per capita
($2000)
GDP per capita
($2000)
135.37 45.82 71.47 6.78 38.56 25.80 35.10 174.82 8.63 11.25 0.01 1,311.33 8,611.81 11,671.95
137.69 46.17 72.20 6.99 38.98 29.60 35.52 198.10 10.43 10.92 0.51 1,381.69 8,921.13 12,364.57
140.06 46.65 72.93 7.20 39.41 33.41 35.95 199.57 12.45 23.66 0.00 1,439.80 9,138.54 12,619.88
142.40 47.13 73.66 7.43 39.83 37.25 36.38 207.94 14.99 29.10 0.00 1,526.71 9,530.97 12,981.95
144.93 47.59 74.41 7.65 40.25 41.12 36.81 204.75 17.78 30.81 0.13 1,536.03 9,421.99 12,669.14
147.51 48.25 75.16 7.89 40.67 45.02 37.25 233.59 21.63 30.48 0.11 1,623.68 9,785.29 13,335.66
150.16 48.90 75.92 8.14 41.10 48.93 37.66 256.33 26.16 28.07 1.44 1,686.33 9,984.03 13,355.59
152.89 51.31 76.68 8.39 41.52 52.90 38.08 266.67 30.97 21.88 1.39 1,746.04 10,152.34 13,379.73
155.47 51.69 77.46 8.65 41.94 56.92 38.50 254.94 35.58 22.41 1.35 1,787.48 10,221.04 13,032.79
158.09 52.74 78.24 8.91 42.36 60.55 38.93 275.75 40.99 23.58 1.34 1,822.62 10,249.25 13,728.28
160.62 52.90 79.03 9.19 42.79 64.59 39.35 290.30 46.83 10.40 1.32 1,831.28 10,135.99 13,847.27
163.30 53.48 79.83 9.47 43.21 68.72 39.68 302.67 53.24 17.73 1.55 1,844.94 10,043.69 13,936.45
165.83 54.30 81.79 9.77 43.63 72.89 40.02 323.11 61.20 25.98 1.66 1,969.93 10,560.46 14,555.75
168.24 55.16 83.80 10.07 44.05 77.00 40.35 351.58 69.68 27.30 1.66 2,018.53 10,666.40 14,975.53
170.59 56.04 85.86 10.38 44.48 80.85 40.68 376.87 0.00 78.55 34.47 1.63 2,114.14 11,017.49 15,626.74
172.74 57.16 87.98 10.70 44.90 84.81 41.02 400.82 0.10 88.91 50.45 1.59 2,252.26 11,591.51 16,423.32
174.74 58.14 90.14 11.03 45.32 88.91 41.30 437.87 0.51 99.57 65.78 -3.71 2,401.05 12,215.34 17,292.94
176.66 58.89 92.36 11.37 45.75 93.93 41.57 474.72 1.34 111.01 67.59 -3.65 2,404.76 12,101.74 17,535.93
178.43 60.14 94.63 11.73 46.17 98.94 41.84 505.77 2.91 123.29 79.88 -3.60 2,588.27 12,895.85 18,199.26
180.18 61.12 96.95 12.09 46.59 103.80 42.12 540.48 5.55 137.29 84.34 -3.53 2,647.13 13,060.85 18,578.33
182.29 62.13 99.34 12.46 47.01 108.21 42.39 586.68 9.66 153.07 82.46 -3.45 2,672.68 13,034.16 18,394.85
184.61 63.07 96.36 12.85 47.44 112.64 42.72 594.94 14.96 169.42 76.67 2.75 2,749.16 13,238.71 18,773.87
186.60 64.36 93.47 13.25 47.86 116.92 43.06 624.54 21.87 187.83 76.56 2.56 2,845.05 13,554.58 19,557.30
188.39 65.46 90.66 13.38 48.28 121.03 43.40 639.85 30.85 208.10 90.16 2.49 3,005.25 14,181.81 20,487.57
190.12 66.22 87.94 13.51 48.70 124.98 43.75 642.37 40.48 228.44 124.52 2.30 3,114.78 14,565.00 20,198.83
189.20 66.62 80.31 13.65 49.13 134.94 44.09 651.09 50.55 244.94 76.14 2.12 3,125.25 14,470.55 19,961.75
191.89 68.47 82.48 13.78 49.55 139.43 44.43 673.61 63.39 260.98 68.14 2.53 3,222.96 14,781.85 20,826.47
195.60 69.96 78.95 13.92 49.82 144.23 44.77 706.27 78.14 275.16 64.00 -7.15 3,265.89 14,828.86 21,569.75
199.79 71.60 76.95 14.06 50.09 148.05 45.11 731.84 94.90 288.15 53.08 -0.52 3,375.03 15,162.87 22,530.72
201.93 73.29 69.61 14.20 50.36 151.81 45.45 789.32 114.47 299.71 35.40 5.58 3,284.80 14,595.54 22,987.27
213.42 74.17 68.65 14.34 50.62 155.68 45.85 826.66 134.22 311.57 99.48 2.57 3,354.46 14,730.24 22,666.27
196.00 75.26 64.26 14.49 50.89 160.04 46.25 849.33 153.58 324.40 99.89 7.12 3,376.43 14,682.31 23,010.79
192.41 76.97 57.96 14.63 51.16 164.93 46.63 862.11 172.81 335.03 85.15 80.32 3,418.34 14,722.31 22,349.56
192.95 78.94 57.30 14.78 51.43 169.12 47.01 845.20 193.15 347.37 97.31 77.45 3,568.86 15,231.57 23,148.26
195.85 80.61 59.23 14.92 51.70 173.67 47.40 943.67 216.36 361.81 98.05 32.21 3,526.68 14,921.55 24,597.63
197.09 81.92 56.28 15.07 51.97 178.26 47.78 960.58 241.79 375.21 98.27 23.54 3,606.54 15,123.92 25,386.01
190.62 84.06 55.88 15.22 52.24 183.15 48.05 979.70 269.44 390.66 98.41 16.48 3,639.17 15,122.17 26,027.73
192.32 85.58 55.96 15.38 52.29 187.61 48.32 1,025.99 300.17 405.72 110.31 -61.41 3,632.41 14,960.25 26,668.01
192.29 87.05 56.61 15.53 52.35 191.98 48.60 1,084.05 335.40 424.82 105.44 -76.17 3,654.19 14,913.77 27,518.87
184.72 88.40 55.94 15.69 52.41 196.24 48.88 1,109.68 373.22 440.73 98.43 -51.59 3,702.06 14,967.38 28,225.70
191.67 89.70 52.29 15.84 52.47 200.46 49.16 1,171.29 412.34 450.65 99.72 -68.10 3,725.17 14,892.80 28,434.99
187.13 91.22 52.15 16.00 52.52 204.59 49.45 1,199.31 453.66 459.20 89.75 -90.05 3,694.66 14,575.01 28,010.64
188.42 92.40 48.88 16.16 52.58 209.62 49.74 1,231.78 495.73 466.79 118.54 -118.50 3,684.52 14,342.57 28,558.86
189.38 93.85 47.99 16.32 52.64 214.61 50.03 1,230.66 541.52 473.13 138.55 -35.41 3,689.31 14,175.75 28,943.54
195.43 95.81 48.56 16.49 52.69 219.57 50.16 1,318.57 590.73 477.01 151.71 -18.01 3,701.64 14,051.40 29,743.47
205.88 97.92 44.36 16.65 52.75 224.57 50.22 1,355.38 643.22 478.74 178.32 -26.09 3,840.85 14,409.10 30,131.27
206.98 99.79 43.76 16.82 52.81 229.62 50.27 1,414.48 699.34 478.77 250.39 -14.95 3,912.45 14,508.46 30,885.87
200.63 101.87 42.60 16.99 52.87 234.73 50.33 1,456.64 761.08 478.81 301.02 -67.76 3,932.67 14,410.04 31,891.23
192.81 104.18 42.22 17.16 52.92 238.74 50.39 1,521.39 824.47 478.82 369.35 -189.34 4,018.36 14,553.23 32,837.40
195.09 106.60 42.06 17.33 52.98 245.91 50.44 1,539.69 889.51 478.84 446.63 -182.04 4,234.69 15,162.06 33,907.88
193.14 109.09 40.58 17.50 53.04 251.69 50.48 1,585.89 960.07 478.87 475.60 -249.80 4,277.03 15,145.93 34,764.23
186.14 111.21 40.40 17.68 53.09 255.26 50.52 1,669.58 1,033.53 478.89 490.29 -380.20 4,113.48 14,417.04 34,665.17
182.01 113.82 40.22 17.85 53.15 258.10 50.56 1,677.57 1,110.76 478.90 455.49 -298.81 4,255.44 14,765.33 34,866.85
180.15 116.57 40.05 18.03 53.21 260.97 50.60 1,701.30 1,146.79 478.91 372.14 -224.33 4,309.61 14,807.16 35,460.01
175.18 119.72 40.05 18.21 53.26 263.86 50.64 1,761.27 1,182.82 478.92 388.80 -254.02 4,419.08 15,035.65 36,595.59
Table 1 (Continued)
22

The Genuine Progress Indicator 2006 Redefi ning Progress
23
Others, including Daly (1996) point out that using
GDP growth as a policy target is a fundamentally fl awed
approach and that even the “poorest approximation” of
welfare would do a better job of policy guidance. Anielski
(2001, pg. 43) goes quite a bit further by asserting that
GPI accounts “provide vital information for holistic and
integrated policy decision making, covering virtually every
area of government policy.” Of course, what information
policy makers choose to rely upon in making their decisions
is often more a function of their political orientations,
beliefs, and personal relationships and so regardless of
concerns about the GPI’s accuracy and rigor, leaders within
government and non-governmental organizations (NGOs)
have used the GPI and its variants as a basis for advocacy.
For example, in Alberta, the Pembina Institute has been
publishing GPI accounts since 2001 as a way to persuade
the provincial government to adopt a more comprehensive
accounting framework that is “capable of assessing the full
benefi ts and full costs of all forms of capital in Alberta
—human, social, natural and built.”
3
In Nova Scotia, the
organization GPI Atlantic reported that the provincial
government had created an Offi ce of Health Promotion
3 See “Alberta could lead the way in sustainable progress indicators,” posted May
16, 2006 at http://www.fi scallygreen.ca/gpi/news.php.
responsible for all matters relating to health promotion,
wellness and addiction services in part based on GPI sub-
accounts documenting the enormous toll ($3 billion)
of largely preventable chronic diseases. As a result, they
conclude “[t]he signifi cance of this cannot be understated:
GPI Atlantic is having an impact on public policy.”
4
In
the San Francisco Bay Area, the quasi-governmental Bay
Area Alliance for Sustainable Communities adopted a local
variant of the U.S. GPI as a means for tracking progress in
achieving the policy objective of a “diversifi ed, sustainable,
and competitive economy” (BAA, 2004, pg. 12).
 e policy relevance of green GDP indicators such as the
GPI and ISEW has also been demonstrated by dozens of
peer reviewed studies. As we previously noted in “ eories,
Principles, and Critiques,” Asheim (2000) found green GDP
indicators useful as measures of welfare equivalent income,
sustainable income, and net social profi t. Hanley (2000)
concludes that the ISEW can be used in tandem with more
traditional economic indicators to generate useful insights
for policy-makers seeking to implement broad sustainability
goals such as those included in Agenda 21. More recently,
Clarke and Islam (2004) estimated an ISEW for  ailand
that further reinforced the threshold hypothesis and
4 See GPI Atlantic Newsletter #14, April 2003, available online at: http://www.
gpiatlantic.org/gpinews/gpinews14.pdf.

The Genuine Progress Indicator 2006 Redefi ning Progress
24
underscored the need for welfare enhancing interventions
by governments of developing nations seeking to off set the
deleterious impacts of pursuing economic growth.
Talberth and Bohara (2006) were among the fi rst to use GPI
and ISEW time series data to analyze the welfare impacts
of policy change by focusing on the eff ects of greater trade
openness. Using panel data from eight countries with GPI
and ISEW accounts and an aggregate production function
model, they found a strong negative correlation between
openness and green GDP and a strong positive correlation
between openness and the gap between traditional and green
GDP.  e eff ects, however, were non-linear, implying that up
to a point, greater openness is benefi cial. Below, we partially
update their analysis using the new U.S. GPI accounts
presented here and extend their analysis to policy variables of
interest to the debates over tax cuts and urban sprawl.
Economic openness
 e debate over the eff ects of economic openness or
globalization has regularly captured headlines since
the World Trade Organization began its attempts to
signifi cantly increase the pace of trade liberalization in the
early 1990s. Empirical studies on the eff ects of openness
fall into two distinct camps. A number of studies have
reported on the benefi cial aspects of more open trade
regimes, noting, for instance, that export expansion raises
the rate of economic growth by way of its impact on
total factor productivity (Dar and Amirkhalkhali, 2003).
Other studies link greater openness to deteriorating social
and environmental conditions, such as increased income
inequality or greater emissions of greenhouse gases (Baten
and Fraunholz, 2004; Managi, 2004). Of course, what is
actually being measured in these studies has a signifi cant
bearing on the outcome.
Studies relating openness to higher economic growth rates
rely almost exclusively on GDP and related measures,
while studies which document the immiserating eff ects of
openness rely on measures outside the realm of traditional
growth models.  us, Talberth and Bohara (2006) suggest
that conducting growth studies using green GDP can help
bridge this divide because green GDP is a more accurate
measure of welfare that explicitly addresses factors of
paramount concern to GDP critics while maintaining
components (i.e. personal consumption expenditures) that
are more consistent with traditional notions of economic
growth.  us, they present a model of growth in green GDP
using data sets spanning 30 – 50 years from eight countries:
Australia, Austria, Brazil, Italy, the Netherlands, Sweden, the
United Kingdom, and the United States. In their growth
model, economic openness was considered along with
measures of human and physical capital typically included
in models of aggregate production functions.
In standard economic models, economic growth is
assumed to be a function of changes in a nation’s stock of
both physical and human capital as well as other factors
that may aff ect the productivity of these inputs such as
economic openness (Solow, 1956; 1957). In their model,
Talberth and Bohara (2006) used changes in the percent
of GDP represented by gross fi xed capital formation, the
age dependency ratio, and economic openness.  e use of
gross fi xed capital formation is standard variable measuring
a nation’s stock of physical capital.  e age dependency ratio
is a ratio of the non-working age to working age population,
and is considered relevant to economic growth because the
size of the dependent population may constrain productivity
enhancing investments (Holtz-Eakin et al., 2004).
Economic openness is the ratio of trade activity (imports
and exports) to GDP.

The Genuine Progress Indicator 2006 Redefi ning Progress
25
Here, we replicate and update the Talberth and Bohara
(2006) analysis with respect to the United States. Time
series data for gross fi xed capital formation and the age
dependency ratio were taken from the World Development
Indicators data set. Time series data for economic openness
were taken from the Penn World Tables. GPI data were
taken from Table 1. Following Talberth and Bohara (2006)
we tested:
[1] GGPI
t
=a
0
+a
1
DGFCFpct
1
+a
2
DOPEN
t
+a
3
DOPEN
2
t
+a
4
DADR
t
+u
t
In equation 1, GGPI is the growth rate of the GPI in year
t, DGFCFpct, DOPEN, and DADR are the year-to-year
changes in the ratio of gross fi xed capital formation to GDP,
economic openness, and the age dependency ratio, and u
is the error term. In recognition of the potential non-linear
eff ects of openness, we have included the square of the
openness term as well (DOPEN
2
). In fact, non-linear eff ects
are strongly suggested by Figure 8 (this page), which plots
the relationship between the openness index and per capita
GPI. In Figure 8, per capita GPI rises strongly when the
openness index is below 25 or so, and then stabilizes when
openness exceeds this level.  e  terms are parameters
estimated by the model. We are interested in the sign,
magnitude, and signifi cance of these terms. Table 2 (page
26) reports the results.
Validating Talberth and Bohara (2006), our modeling
suggests a signifi cant negative non-linear correlation
between growth in the U.S. GPI and economic openness,
a positive relationship with changes in gross fi xed capital
formation, and a negative relationship with the age
dependency ratio.  e results provide some empirical
support for the burgeoning literature associating greater
openness with environmental degradation, income
inequality, and an increase in economic activity that may
be self canceling from a welfare perspective.  ey also
suggest a cautionary approach to trade liberalization policy
that is cognizant of the fact that liberalization may be
counterproductive past a particular threshold.
Tax cuts
Tax cuts have been one of the most visible economic
policy debates since the Bush Administration took offi ce
in 2001.  e debate has been a bone of contention in both
policy and academic circles. In the context of standard
growth theory, tax cuts can stimulate long term economic
growth through six main channels depending on the type
and incidence of the particular tax involved: (1) they can
encourage productivity-enhancing investments in the
capital stock; (2) encourage growth in both the quality
and quantity of the labor force; (3) stimulate research and
development; (4) steer capital investment to sectors with
higher productivity, and (5) steer workers towards sectors
with higher social productivity (Engen and Skinner, 1996).
Additionally, in the short run, tax cuts can lead to increases
in consumer spending.
On the other hand, tax cuts can harm economic growth if
not matched by a commensurate decrease in government
spending; otherwise, they will raise defi cits and interest
rates. If tax cuts disproportionately benefi t the wealthy, the
resulting “windfall gains” on asset holders may undermine
incentives for new investments (Gale and Orszag, 2005).
Tax cuts may also reduce labor force participation if the
incentive to work more hours at higher pay is more than
off set by the incentive to work less and keep income
constant (Gale and Orszag, 2005). Finally, if tax cuts are
matched with decreases in government programs, the socio-
economic benefi ts of those programs are sacrifi ced.
Empirical studies relating tax cuts to economic growth
are also ambiguous. Hashemzadeh and Wayne (2004, pg.
112) assert that “[f ]rom an historical perspective, there is
scarce evidence of a consistent relationship between income
taxes and economic growth.”  ey also note that periods of
high economic growth in output have correlated quite well
with higher taxes. On the other hand, Engen and Skinner
(1996) predict a .2 to .3% boost in economic growth rates
associated with a 5% cut in marginal tax rates. Recently,
Diamond (2005) predicted that extending the 2001
and 2003 income tax cuts would stimulate investment,
employment, and output.
As with the debate over economic openness, both
proponents and opponents of tax cuts have almost

The Genuine Progress Indicator 2006 Redefi ning Progress
26
exclusively argued their points from a single perspective—
economic growth as traditionally defi ned rather than
from the standpoint of more comprehensive measures of
welfare like the GPI. Given the empirical and theoretical
ambiguity of the debate and given the paucity of studies
relating taxation and welfare, a correlation between GPI
and taxes may be a useful exercise.  ere are a number of
ways GPI and tax cuts may be related. If tax cuts exacerbate
income inequalities, the GPI will fall. If tax cuts cause
reductions in benefi cial government programs (i.e. for
farmland conservation, renewable energy, or water quality
improvements) the GPI may also fall.  e GPI may also fall
because tax cuts often induce an infl ux of foreign capital
(Gale and Orszag, 2005). If this capital is used to fi nance
current consumption (see discussion under “An Updated
GPI Methodology,” column AA) the GPI will fall. On
the other hand, it tax cuts boost personal consumption or
participation in volunteer work or educational activities,
GPI could be expected to rise. GPI may also rise if tax cuts
stimulate greater capital investment.
As a preliminary investigation, we modify equation [1]
by adding a tax variable. In particular, we incorporate
tax collection time series data from the National Income
and Product Accounts tables published by the Bureau of
Economic Analysis. Conceptually, adding a tax collection
variable to the aggregate production function framework
embodied by equation [1] is complicated by the fact that
the causation may run in the opposite direction—growth
may induce greater tax collections, and not vice versa. Of
course, it is not clear if the causality concern is as relevant
to GPI as it is to growth of GDP. In addition, we rely—as
with openness - on growth rates as suggested by Engen and
Skinner (1996) rather than absolute GPI and tax collection
values. We also rely on per capita tax collection fi gures,
not totals. Finally, we lag the tax collection variable so
that we are testing the correlation between the change in
tax collections between 1963 and 1964 on the growth in
GPI between 1964 and 1965; a modifi cation that makes
intuitive sense if we are testing the proposition that reduced
government spending aff ects welfare. By adding a lagged tax
collection variable, our GPI growth model becomes:
[2] GGPI
t
=a
0
+a
1
DGFCFpct
t
+a
2
DOPEN
t
+a
3
DOPEN
2
t
+a
4
DA
DR
t
+a
5
DTAXCOL
t-1
+u
t
In equation 2, DTAXCOL is the change in per capita tax
collections in year t-1. All other variables are as before.
 e results are displayed in Table 2, column B. As shown,
we fi nd a strong positive correlation between the change
in per capita tax collections and growth of the GPI.  is
fi nding is consistent with the historical relationship between
higher taxes and high economic growth (as measured
by GDP) noted by Hashemzadeh and Wayne (2004).
A full investigation of these fi ndings to determine the
exact channel by which changes in taxes infl uence GPI
growth is beyond the scope of this report. Nonetheless,
as with openness, we have demonstrated the potential use
of GPI data to inform the debate over tax cuts and other
adjustments to tax policy.
Growth in urbanization
In our discussion of openness and tax cuts, we relied on
the aggregate production function framework to examine
the impacts of policy variables on GPI growth. Another
potentially useful approach is to explore the impacts of
policy variables on the gap between GDP and GPI. By
looking at the gap, we can simultaneously address economic
changes in economic growth (GDP) and welfare (GPI). In
particular, in years when the gap is widening, the costs of
TABLE 2: Models of U.S. GPI Growth (GGPI)
DOPEN
DOPEN
2
-1.00***
(-3.12)
6.13*
(1.84)
-1.28***
(-4.31)
7.55**
(2.54)
Model 1
Openness
Model 2
Tax Cuts
DGFCFpct
DDADR
1.14**
(2.33)
-9.00***
(-3.03)
.21
(0.696)
-7.48***
(-2.80)
F-statistic
R-squared
(adj)
Observations
5.35***
.3383
35
7.38***
.4841
35
DTAX
Constant
0.03***
(6.73)
0.65***
(3.08)
0.03***
(6.74)
Independent
Variables
Numbers in parentheses are t-statistics. *, **,
and *** denote signifi cance at the .10, .05, and
.01 levels.

The Genuine Progress Indicator 2006 Redefi ning Progress
27
economic growth are more than off set by the deleterious
social and environmental welfare costs of that growth. In
years when the gap is closing, positive contributions to GPI
overshadow these costs and economic growth is welfare
enhancing. In their model, Talberth and Bohara (2006)
modeled the eff ects of changes in economic openness, the
growth rate of carbon dioxide emissions
5
and livestock
production on the gap and found each to have a signifi cant,
positive infl uence on the rate of gap growth. Here, we
adopt that model and substitute a variable of interest to the
debate over urban growth for the livestock variable – degree
of urbanization, measured in terms of urban land area per
capita. Specifi cally, we test:
[3] GGAP
t
=a
1
+a
2
DURBAN
t
+a
3
DCO2grw
t
+a
4
DOPEN
t
+
a
5
DOPEN
2
+u
t
In equation 3, GGAP is the growth rate of the gap between
GDP and the GPI, DURBAN is the change in urban
land area per capita as measured by Census Bureau data,
DCO2grw is the change in the growth rate in per capita
carbon dioxide emissions.  e openness variables are as
before. We are particularly interested in the urbanization
variable, which is a good proxy for urban sprawl since it
measures the amount of urban land per person. According
to the General Accounting Offi ce, urban sprawl is
“sprawling, low density, fragmented, automobile-dependent
development.” (GAO, 1999).
 ere is little dispute that public policy has a direct
infl uence on the extent of urban sprawl. According to
the Environmental Protection Agency (EPA), a number
of federal urban growth and development programs
“intentionally or unintentionally accelerated the spread of
low density development and businesses at greater distances
from towns and cities.”
6
 e question is whether or not
urban sprawl enhances or detracts from welfare. Despite the
negative connotation associated with the term, there are at
least two channels by which the GDP–GPI gap can improve
with more sprawl, again, defi ned here as more urban land
area per person.
First, it is important to note that urban sprawl is partially
driven by the need to accommodate high volume, low cost
5 Because carbon dioxide emissions are indirectly included in the GPI calculations,
care must be taken to avoid spurious regression results. To do this, Talberth and
Bohara (2006) look at changes in the growth rate of emissions and not the level of
emissions. Policy variables aff ecting this growth rate may be changes in CAFE stan-
dards, speed limits, regulations governing oil and gas development, or fossil fuel
subsidies. Here, however, we are focusing attention on the eff ects of urbanization,
and leave the debate over carbon policy for another time.
6 See “About Smart Growth,” U.S. EPA, online at http://www.epa.gov/smart-
growth/about_sg.htm#fedrole.
retail “big box” stores such as Wal-Mart, Home Depot,
and Costco, who bring an unprecedented volume and
variety of low cost consumer goods to the public in a
single location. Importantly, if more sprawl is associated
with a greater abundance and easier access to these low
cost consumer goods, the GPI will likely increase since
it is based on personal consumption expenditures. But
GDP also includes personal consumption expenditures,
so this eff ect will have little impact on the GDP–GPI gap.
However, to the extent that concentrated retail centers free
up time otherwise spent shopping in multiple locations the
GDP–GPI gap may improve if there is a corresponding
increase in time spent volunteering, in educational
activities, parenting, or housekeeping, the value of which
is overlooked by GDP. Indeed, time savings have always
been one of the most important benefi ts associated with
concentrated retail centers:
Back in the city, the search for goods, whether pleasurable
or not, consumes a great deal of time. Shopping competes
with other activities and the geography of retailing has
always been driven, in part, by the need to economize
on time. Minimizing procurement time underlies the
existence of retailers in the fi rst place…Convenience, one
of the most enduring themes of retailing, thus has driven
the geographic arrangement of stores through cities and
suburbs (Campbell, 1996).
In addition, because the GPI counts the services yielded by
pubic streets and highways, sprawl no doubt enhances this
GPI contribution since by defi nition, more sprawl means
more streets and highways per person. And since these
services are not counted in GDP, sprawl may help close
the GDP–GPI gap. On the other hand, the GPI deducts
costs associated with longer commutes, auto-accidents,
carbon emissions, and lost farmland. None of these costs are
included in GDP, and so the gap will widen as these costs
escalate.  us, the net eff ects are ambiguous, and worth
exploring in a more systematic fashion. Equation [3] does
just that.
 e results are displayed in Table 3 (page 28). First, we
note that our results corroborate earlier fi ndings of Talberth
and Bohara (2006) by demonstrating a positive non-linear
relationship between openness and growth of the GDP–
GPI gap, and a positive relationship between changes in
the growth rate of carbon dioxide emissions and the gap.
Secondly, we note the positive relationship between growth
in urban land area per capita and the gap.  is suggests that
on balance, the personal consumption, time savings, and
public infrastructure benefi ts from sprawl are more than
off set by the costs associated with traffi c congestion, auto-
accidents, carbon emissions, and lost farmland.

The Genuine Progress Indicator 2006 Redefi ning Progress
28
Concluding Thoughts and Future Refi nements
 e Genuine Progress Indicator (GPI) and its variants such
as the Index of Sustainable Economic Welfare (ISEW)
were conceived as a way to measure changes in national
economic welfare with a single, aggregate index.  e
GPI considers households as the basic building block of
a nation’s welfare, and thus begins its accounting exercise
with personal consumption expenditures. To this the GPI
adds benefi ts associated with welfare enhancing activities
such as parenting, housework, volunteering and higher
education as well as the services which fl ow from household
capital and public infrastructure.  e GPI then deducts
costs associated with pollution, loss of leisure time, auto
accidents, destruction or degradation of natural capital,
international debt and resource depletion.  e end result is
an index that attempts to measure our collective welfare in
terms of principles of sustainable development drawn from
the economic, social, and environmental domains.
In this report, we presented an updated methodology for
the U.S. GPI and a new set of accounts current through
2004. Our updates are the fi rst signifi cant changes to the
GPI methodology since 1998, and incorporated a wealth of
new studies and sources of information that have evolved
since that time.  e accounts suggest that while the U.S.
economy has grown steadily since 1950, our collective
welfare may have peaked in the late 1970s and stagnated
ever since as the benefi ts of economic growth since that
time have been more and more off set by costs associated
with income inequality, loss of time spent on non-market
activities, and environmental degradation.  e costs of
climate change are becoming an increasingly large share,
as demonstrated all too well by the disasters in the Gulf of
Mexico in the summer of 2005.
While some dispute the GPI’s ability to measure sustainable
welfare or take issue with its methodological soundness,
it has, nonetheless, prompted government and non-
governmental organizations throughout the world to use
it as a tool for promoting sustainable policies and for
demonstrating the fallacy of relying on gross domestic
product (GDP) as a welfare measure. And because the
GPI accounts yield historical data going back 54 years, it
is readily adaptable for use by researchers seeking to test
the infl uence of past policy changes on welfare growth. In
this report, we demonstrated how GPI time series data can
be incorporated into standard economic growth models to
inform policy debates involving economic openness, tax
cuts, and urban sprawl.
While future refi nements to the GPI will attempt to address
some of its outstanding theoretical challenges—such as
relating future impacts to current welfare—the bulk of these
new refi nements will be focused on developing new sources
of information and more precise calculation methodologies.
 e GPI accounts would be well served by a new set of
valuation studies addressing time use, natural capital
depletion, and costs associated with disservices such as air
and water pollution, since many of the sources underlying
these GPI columns are somewhat dated.
 ere are a number of changes to calculation methodologies
that could be made in response to the latest round of vetting
in the literature. For example, Lawn (2005) expresses
wholehearted agreement with Neumayer’s (2000) critique
regarding the methods used to calculate resource depletion,
and there is no reason why future GPI iterations could not
adopt their recommendations. Taken together, these changes
will make the GPI a more accurate, robust, and widely
endorsed tool for promoting sustainable development in the
decades ahead.
TABLE 3: Models of the GDP-GPI Gap(GGAP)
and urbanization
DOPEN
DOPEN
2
4.42***
(4.68)
-35.12***
(-3.94)
Urbanization
Model
DGFCFgrw
DURBAN
0.60***
(3.48)
88.41**
(2.53)
F-statistic
R-squared (adj)
Observations
9.87***
.4249
49
Constant
-0.08**
(-2.19)
Independent
Variables
Numbers in parentheses are t-statistics. *, **,
and *** denote signifi cance at the .10, .05, and
.01 levels.

The Genuine Progress Indicator 2006 Redefi ning Progress
29
References
Adams, Darius, R. Haynes, and A. Daigneault. 2006. Estimated Timber Harvest
by U.S. Region and Ownership, 1950-2002. Gen Tech. Rpt. PNW-GTR-659.
Portland, Oregon: USDA Forest Service, PNW Research Station.
AFEAS (Alternative Fluorocarbons Environmental Acceptability Study). 1998.
Production, Sales, and Atmospheric Release of Fluorocarbons through 1996.
Washington, D.C.: AFEAS.
Anielski, Mark. 2001. Measuring the Sustainability of Nations:  e Genuine
Progress Indicator System of Sustainable Well Being Accounts. Edmonton,
Canada: Pembina Institute for Appropriate Development.
Asheim, Geir B. 2000. “Green national accounting: why and how?” Environment
and Development Economics 5(2000): 25048.
Ayres, R. U. 1978. Resources, Environment, and Economics: Applications of the
Materials/ Energy Balance Principle. New York, New York: John Wiley and Sons.
Barber, D.A. 2003. “ e ‘new’ economy,” Tuscon Weekly, January 8th, 2003.
Baten, J. and Fraunholz, U., 2004. “Did partial globalization increase inequality?
 e case of the Latin American periphery, 1950-2000.” CESifo Economic Studies
50(1), 45-84.
Bay Area Alliance for Sustainable Communities. 2004. State of the Bay Area:
A Regional Report. Oakland, California: Bay Area Alliance for Sustainable
Communities.
Beardsley, Debby, Charles Bolsinger, and Ralph Warbington. 1999. Old Growth
Forests in the Sierra Nevada: By Type in 1945 and 1993 and Ownership in 1993.
Research Paper PNW-RP-516. Portland, OR: USDA Forest Service, Pacifi c
Northwest Research Station.
Bluestone, Barry, and Stephen Rose. 1997. “Overworked and underemployed:
unraveling an economic enigma.” American Prospect 31(March–April):58–69.
Campbell, John. 1996. Time to Shop:  e Geography of Retailing. Boston:
Federal Reserve Bank of Boston.
Carson, Carol S. and Allan H. Young. 1994. “ e ISEW from a national
accounting perspective,” In Cliff ord W. Cobb and John B. Cobb (eds).  e Green
National Product: A Proposed Index of Sustainable Economic Welfare. Lanham:
University Press of America.
Clarke, Matthew and Sardar M.N. Islam. 2004. “Diminishing and negative welfare
returns of economic growth: an index of sustainable economic welfare (ISEW) for
 ailand.” Ecological Economics 54(1): 81-93.
Cobb, Cliff ord, Ted Halstead, and Jonathan Rowe. 1995. “If the GDP is up, why
is America down,” Atlantic Monthly: October, 1995.
Cobb, Cliff ord W. and John B. Cobb, Jr. 1994.  e Green National Product: A
Proposed Index of Sustainable Economic Welfare. Lanham, MD: University Press
of America.
Dietz, Simon and Eric Neumayer. 2006. “Some constructive criticisms of the
Index of Sustainable Economic Welfare.” In Philip Lawn (ed.): Sustainable
Development Indicators in Ecological Economics. Chetltenham: Edward Elgar
Publishing.
Congressional Quarterly, Inc. 1972. Congressional Quarterly Almanac.
Washington, D.C.: Congressional Quarterly, Inc.
Costanza, Robert, Ralph d’Arge, Rudolf de Groot, Stephen Farber, Monica Grasso,
Bruce Hannon, Karin Limburg, Shahid Naeem, Robert V. O’Neill, Jose Paruelo,
Robert G. Raskin, Paul Sutton, and Marjan van den Belt. 1997. “ e value of the
world’s ecosystem services and natural capital.” Nature 387(15 May):253–59, p.
256.
Daly, Herman and John B. Cobb. 1989. For the Common Good: Redirecting
the Economy Toward Community, the Environment, and a Sustainable Future.
Boston: Beacon Press.
Daly, Herman. 1996. Beyond Economic Growth:  e Economics of Sustainable
Development. Beacon Press, Boston.
Dar, A. and Amirkhalkhali, S., 2003. “On the impact of trade openness on
growth: further evidence from OECD countries.” Applied Economics 35(2003),
1761-1766.
Diamond, John W. 2005. “Dynamic eff ects of extending the 2001 and 2003
income tax cuts.” International Tax and Public Finance 12(2): 165-192.
Engen, Eric and Jonathan Skinner. 1996. “Taxation and economic growth.”
National Tax Journal 49: 617-642.
Fisher, I. 1906. Nature of Capital and Income. New York: A.M. Kelly.
Foertsch, Tracy L. 2006. A Victory for Taxpayers and the Economy. Heritage
Foundation WebMemo #1082. Washington:  e Heritage Foundation. (www.
heritage.org/Research/Taxes/wm1082.cfm).
Folke, Carl, Monica Hammer, Robert Costanza, and AnnMari Jansson. 1993.
“Investing in natural capital—why, what, and how?” In AnnMari Jansson et al.
(eds). 1993. Investing in Natural Capital:  e Ecological Economics Approach to
Sustainability. Washington, D.C. Island Press.
Freeman, Myrick. 1982. Air and Water Pollution Control: A Benefi t-Cost
Assessment. New York: John Wiley and Sons.
Gale, William G. and Peter R. Orszag. 2005. “Economic eff ects of making the
2001 and 2003 tax cuts permanent.” International Tax and Public Finance 12(2):
193-232.
General Accounting Offi ce. 1999. Community Development: Extent of Federal
Infl uence on Urban Sprawl is Unclear. Washington, D.C.: General Accounting
Offi ce.
Hagerman, J.R. 1992. Upper Little Tennessee River Aerial Inventory of Land Uses
and Nonpoint Pollution Sources. TVA/WR-92/10. Chattanooga, TN: Tennessee
Valley Authority, Water Quality Department.
Hamilton, Clive and Richard Denniss. 2000. Tracking Well Being in Australia:
 e Genuine Progress Indicator. 2000. Canberra, Australia: Australia Institute.
Hanley, Nick. 2000. “Macroeconomic measures of sustainability.” Journal of
Economic Surveys 14 (1): 1– 30.
Harris, Jonathan. 2000. Basic Principles of Sustainable Development. Medford,
MA: Global Development and Environment Institute, Tufts University.
Hashemzadeh, Nozar and Wayne Saubert. 2004. “ e eff ects of Bush’s tax cuts on
income distribution and economic growth in the United States.” Problems and
Perspectives in Management 0(3): 111-120.
Heston, A., Summers, R. and Aten, Bettina, 2002. Penn World Table Version 6.1,
Center for International Comparisons at the University of Pennsylvania (CICUP).
Hicks, John. 1947. Value and Capital, Second Edition. London: Clarendon.
Hill, Kent, Dennis Hoff man, and Tom R. Rex. 2005. 
e Value of Higher
Education: Individual and Societal Benefi ts. Tempe, Arizona: Arizona State
University, W.P. Carey School of Business.

The Genuine Progress Indicator 2006 Redefi ning Progress
30
Holtz-Eakin, D., Lovely, M. and Tosun, M., 2004. Generational confl ict, fi scal
policy, and economic growth. Journal of Macroeconomics 26, 1-24.
Hsing, Yu. 2005. “Economic growth and income inequality: the case of the U.S.”
International Journal of Social Economics 32(7): 639-47.
Independent Sector. 2006. Independent Sector’s Giving and Volunteering in the
United States, Annual Update. Washington D.C.: Independent Sector.
Intergovernmental Panel on Climate Change (IPCC). 2000. Land Use, Land-Use
Change, and Forestry. Geneva: United Nations Environmental Program and the
World Meteorological Organization.
Kuznets, Simon. 1934. National Income, 1929-1932. Senate document no. 124,
73d Congress, 2d session, 1934.
Laband, David N., and John P. Sophocleus. 1992. “An estimate of resource
expenditures on transfer activity in the United States.” Quarterly Journal of
Economics (August):959–83.
Lawn, Philip A. 2003. “A theoretical foundation to support the Index of
Sustainable Economic Welfare (ISEW), Genuine Progress Indicator (GPI), and
other related indexes.” Ecological Economics 44(2003): 105-118.
Lawn, Philip. 2005. “An assessment of the valuation methods used to calculate
the Index of Sustainable Economic Welfare (ISEW), Genuine Progress Indicator
(GPI), and Sustainable Net Benefi t Index (SNBI).” Environment, Development,
and Sustainability 2005(7): 185-208.
Leete-Guy, Laura, and Juliet B. Schor. 1992.  e Great American Time Squeeze:
Trends in Work and Leisure, 1969–1989. Washington, D.C.: Economic Policy
Institute.
Leipert, Christian. 1989. “National income and economic growth: the conceptual
side of defensive expenditures.” Journal of Economic Issues 23(3): 843-856.
Managi, S., 2004. “Trade liberalization and the environment: carbon dioxide for
1960-1999.” Economics Bulletin 17(1), 1-5.
Max-Neef, M. 1995. Economic growth and quality of life: a threshold hypothesis.
Ecological Economics 15: 115– 118.
Markandya, A. and J. Richardson. 1992. Environmental Economics: A Reader.
New York, New York: St. Martin’s Press.
Mishel, Lawrence, Jared Bernstein, and John Schmitt. 1996.  e State of Working
America 1995–96. Armonk, N.Y.: M.E. Sharpe.
Moretti, E. 2004. “Estimating the social return to higher education: evidence from
longitudinal and repeated cross-sectional data.” Journal of Econometrics, July/
August 2004: 175-212.
National Center for Statistics and Analysis (NCSA). 2003. Traffi c Safety Facts,
2003 Data. Washington, D.C.: National Center for Statistics and Analysis.
National Center for Transit Research (NCTR). 2005. Public Transit in America:
Results from the 2001 National Household Transit Survey. Tampa: National
Center for Transit Research, Center for Urban Transportation Research, University
of South Florida.
National Oceanic and Atmospheric Administration (NOAA). 2006. Southern
Hemisphere Winter Summary 2005. Washington D.C.: U.S. Department
of Commerce, National Oceanic and Atmospheric Administration, Climate
Prediction Center (www.cpc.ncep.noaa.gov).
National Safety Council. 2004. Estimating the Costs of Unintentional Injuries,
2004. Itasca, IL: National Safety Council.
National Safety Council. 1998. Accident Facts. Itasca, IL: National Safety Council.
Neumayer, Eric. 2000. “On the methodology of ISEW, GPI and related measures:
some constructive suggestions and some doubt on the ‘threshold’ hypothesis.”
Ecological Economics 34: 347–361.
Neumayer, Eric. 1999. “ e ISEW—not an index of sustainable economic
welfare.” Social Indicators Research 48(1): 77-101.
Nordhaus, William and James Tobin. 1972. “Is growth obsolete,” in Economic
Growth, National Bureau of Economic Research Series No. 96E. New York:
Columbia University Press.
Outcalt, Kenneth and Raymond M. Sheffi eld. 1996.  e Longleaf Pine Forest:
Trends and Current Conditions. Resource Bulletin SRS-9. Asheville, NC:
Southern Research Station.
Pearce, D. and G. Atkinson. 1993. “Capital theory and the measurement
of sustainable development: an indicator of weak sustainability.” Ecological
Economics 8(2): 103-108.
Pearce, D., A. Markandya, and E. Barbier. 1990. Sustainable Development: Policy
and Analysis in the  ird World. Cheltenham: Edward Elgar Publishing.
Pezzey, J.C.V. 1992. “Sustainability: an interdisciplinary guide.” Environmental
Values 1: 321-362.
Ready, Richard C., Mark C. Berger, and Glenn C. Blomquist. 1997. “Measuring
amenity benefi ts from farmland: hedonic pricing vs. contingent valuation.”
Growth and Change 28(Fall):438-458.
Roth, Andrew. 2005.  e Bridge to Nowhere and the Minimum Wage.
Washington:  e Club for Growth. (www.clubforgrowth.org/blog/archives/
o26323.php).
Rutledge, Gary L. and Christine R. Vogan. 1994. “Pollution abatement and
control expenditures, 1972–92, estimates for 1992, revised estimates for 1972–
91.” Survey of Current Business, May 1994: 36.
Rymes, T. 1992. Some  eoretcial Problems in Accounting for Sustainable
Consumption. Carleton Economic Papers, 92-02.
Sampson, R. Neil. 1981. Farmland or Wasteland. Emmaus, Penn.: Rodale Press.
Saunier, Richard E. 1999. Perceptions of Sustainability: A Framework for the
21st Century. Washington, D.C.: Executive Secretariat for Integral Development
(SEDI), Organization of American States (OAS).
Schneider, Friedrich and Dominik Enste. 2002. “Hiding in the shadows: the
growth of the underground economy,” Economic Issues 30 (March 2002).
Brussels: International Monetary Fund.
Solow, R. M., 1956. “A contribution to the theory of economic growth.” Quarterly
Journal of Economics 70(1956), 65-94.
Solow, R. M., 1957. “Technical change and the aggregate production function.”
Review of Economics and Statistics 39(1957), 312-320.
Stockhammer, E., H. Hochreiter, B. Obermayr and K. Steiner, K. 1997. “ e
index of sustainable economic welfare (ISEW) as an alternative to GDP in
measuring economic welfare: the results of the Austrian (revised) ISEW calculation
1955–1992.” Ecological Economics 21 (1997): 19– 34.
Talberth, John and Alok Bohara. 2006. “Economic openness and green GDP.”
Ecological Economics 58: 743-758.

The Genuine Progress Indicator 2006 Redefi ning Progress
31
 e World Bank Group (TWBG), 2004. World Development Indicators 2004.
Washington, D.C.
Tol, Richard S.J. 2005. “ e marginal damage costs of carbon dioxide emissions:
an assessment of the uncertainties.” Energy Policy 33(2005): 2064-2074.
Uri, Noel D. and James A. Lewis. 1999. “Agriculture dynamics and soil erosion in
the United States.” Journal of Sustainable Agriculture 14(2/3): 63-75.
U.S. Census Bureau. 2003. Income, Poverty, and Health Insurance Coverage in
the United States: 2003. Washington, D.C.: U.S. Census Bureau.
U.S. Department of Agriculture. 1988. Ethanol: Economic and Policy Tradeoff s.
Washington, D.C.: U.S. Department of Agriculture.
U.S. Department of Agriculture. 2005. Status and Trend of Late Successional and
Old Growth Forest. General Technical Report PNW-GTR-646. Portland, OR:
USDA Forest Service, Pacifi c Northwest Research Station.
U.S. Department of Agriculture. 2006. U.S. Forest Facts and Historical Trends.
Washington, D.C.: USDA Forest Service.
U.S. Department of Labor. Bureau of Labor Statistics. 1990. “ irty-eight million
persons do volunteer work.” News 90–154(29 March). Washington, D.C.: U.S.
Department of Labor.
U.S. Department of Transportation (DOT). 2000. Journey to Work Trends in
the United States and its Major Metropolitan Areas 1960–2000, Publication No.
FHWA -EP-03-058. Washington, D.C.: U.S. Department of Transportation,
Federal Highway Administration.
U.S. Fish and Wildlife Service (USFWS). 2003. Recovery Plan for the Red-
cockaded Woodpecker (Picoides borealis) Second Revision. Atlanta, GA: U.S. Fish
and Wildlife Service, Southeast Region.
U.S. Fish and Wildlife Service (USFWS). 1997. Status and Trends of Wetlands in
the Conterminous United Status and Trends of Wetland: Projected Trends 1985 to
1995. Washington, D.C.: U.S. Government Printing Offi ce.
U.S. Forest Service. 1980. Final Environmental Impact Statement: Roadless Area
Review and Evaluation (RARE II). Washington, D.C.: USDA Forest Service.
United States Geologcical Survey. 2006. USGS Reports Latest Land-Water
Changes for Southeastern Louisiana. USDI/ USGS. (www.lacoast.gov).
Vincent, James W, Daniel A. Hagen, Patrick Welle, Kole Swanser. 1995. Passive
Use Values of Public Forestlands: A Survey of the Literature. St. Paul, MN:
University of St. Paul, Department of Economics.
Vogan, Christine R. 1996. “Pollution abatement and control expenditures, 1972–
94.” Survey of Current Business, September 1996: 48–67.
Woodward, Richard T., Yong-Suhk Wui. 2001. “ e economic value of wetland
services: a meta-analysis.” Ecological Economics 37:257-270.
The nation’s leading policy
institute dedicated to smart
economics, Redefi ning Progress
develops solutions that help
people, protect the environment,
and grow the economy.
At Redefi ning Progress, we believe we do not have to
choose between a strong economy, a healthy environ-
ment, and social justice. Redefi ning Progress makes a
difference through:
Unbiased research. We conduct objective research
about how economic policies and business practices
impact people’s lives.
Innovative tools. We pioneer practical tools and
metrics—like the Ecological Footprint and the Genuine
Progress Indicator—to help policymakers, government
agencies, businesses, and individuals understand the
impact of their practices on the environment and social
equity.
Smart solutions. We identify and champion smart,
forward-looking policy solutions to grow the economy,
empower people, and strengthen communities.
Redefi ning Progress works with a broad array of part
ners
to shift the economy toward sustainable growth. Our
partners include grassroots communities, labor unions,
policymakers, academics and businesses. Our efforts
are largely focused on the United States because rede
fi ning
progress here will have enormous global reverberations.
•
•
•
1904 Franklin St, Suite 600
Oakland, CA 94612
T: 510.444.3041
F: 510.444.3191
www.redefi ningprogress.org