reprinted with permission from MIT Press

innovations
John P. Holdren TECHNOLOGY | GOVERNANCE | GLOBALIZATION

mitpress.mit.edu/innovations

The Energy Innovation Imperative

Addressing Oil Dependence, Climate Change,
and Other 21st Century Energy Challenges
Society faces many energy challenges in this century, but “running out” of energy
resources in a global or absolute sense is not one of them. The world may be run-
ning out of cheaply extractable and reliably deliverable conventional oil and natu-
ral gas, insofar as (a) these energy forms may continue (with some ups and downs)
to get more costly and less reliable over time and (b) it is unclear for how much
longer the rate at which they are extracted can be increased to meet rising global
demand. But energy resources of other types are immensely larger and capable in
principle of being expanded to multiples of today’s use rates of oil and gas com-
bined: there is 5 to 10 times as much coal as conventional oil and gas; there is 5 -
10 times as much oil shale and unconventional gas as coal; the energy potential of
uranium and thorium resources is larger still; and harnessing even a small percent-
age of the solar energy flow reaching Earth’s land surface could meet multiples of
today’s world energy demand.1
The energy issue is difficult not because of any impending exhaustion of glob-
al resources, then, but for more complicated reasons.
The first of these is the multiplicity and diversity of the economic, environ-
mental, and security aims that energy strategy must serve, many of them in ten-
sion with each other (see box, following page). The desire to limit costs is often at
odds with the aims of increasing reliability, reducing vulnerability, and improving
environmental performance. The historically low costs of oil, natural gas, and
many hydropower projects are not likely to be matched by the more abundant fos-
sil, nuclear, and renewable alternatives. Expanding domestic oil production in
order to limit imports eventually encounters not only rising marginal costs but

John P. Holdren is Teresa and John Heinz Professor of Environmental Policy and
Director of the Program on Science, Technology, and Public Policy at the John F.
Kennedy School of Government at Harvard University, Professor of
Environmental Science and Policy in Harvard's Department of Earth and
Planetary Sciences, and Director of the Woods Hole Research Center. He is also
the current President of the American Association for the Advancement of
Science and the Chair of the Editorial Advisory Board of Innovations.

© 2006 Tagore LLC

innovations / spring 2006 3
John P. Holdren

Aims of a Successful Energy Strategy

Choices about how society obtains and uses energy are crucial determinants
of economic well-being, environmental quality and sustainability, and nation-
al and international security. A successful energy strategy must attend to these
connections by pursuing multiple aims. Among those widely considered
important are:
Economic aims
y ensuring reliable supplies of fuel and electricity for basic needs and eco-
nomic growth;
y limiting the costs of energy to firms and consumers;
y limiting balance-of-payments impacts and macroeconomic vulnerabilities
from energy imports.

Environmental aims
y improving urban and regional air quality;
y limiting impacts of energy development on fragile ecosystems;
y avoiding nuclear-reactor accidents and waste-management mishaps;
y limiting energy-supply contributions to global climate change.

National and international security aims

y minimizing dangers of conflict over oil and gas resources;
y avoiding the spread of nuclear weapons from nuclear-energy technology;
y reducing the vulnerability of energy systems to terrorist attack;
y avoiding energy blunders that perpetuate or create deprivation (which is
one of the most fundamental causes of conflict).

also rising opposition when the remaining domestic resources lie under fragile or
particularly highly prized environments. Replacing conventional oil and gas with
synthetic liquids and gases made from tar sands, oil shales, and coal will sharply
increase the emissions of climate-altering carbon dioxide unless costly carbon cap-
ture and sequestration accompany these conversions. Rapid expansion of nuclear
energy may risk outrunning the capabilities of national and international organi-
zations to manage its risks. And so on.
The second reason the energy issue is so challenging is the fact that no known
energy source is free of significant limitations, liabilities, or uncertainties in rela-
tion to one or more of the important aims. That is, there is no technological “sil-
ver bullet”. Volumes have been written about the character of the major energy
options and the difficulties and uncertainties that cloud their prospects. Here I will
characterize the question marks more tersely:
Conventional Oil and Gas—Not enough resources? This is a matter not only of
global availability, of course, but also of economically and politically challenging

4 innovations / spring 2006

The Energy Innovation Imperative

geographic distribution.

Coal, Tar Sands, Oil Shale—Not enough atmosphere? This refers to the capacity
of the atmosphere to absorb without intolerable consequences the emissions from
mobilizing and burning these immense fossil-fuel resources, above all the carbon
dioxide.
Biofuels—Not enough land? Growing biofuels must compete for land and water
with production of food, fiber, and chemical feedstocks, as well as with the essen-
tial environmental service functions of lightly exploited and unexploited ecosys-
tems.
Wind and Hydropower—Not enough acceptable sites? Wind and hydro are most
economical in places where the respective resources are highly concentrated and
near load centers; not only do costs tend to go up as increasing scale of use drives
society to less attractive sites, but many of the best sites are prized for other pur-
poses and may be placed off limits by politics.
Solar Photovoltaics—Not enough money? Despite decades of remarkable
progress in cost reduction, solar photovoltaic arrays remain several times more
expensive than fossil-fueled, nuclear, and wind electricity generation for grid-con-
nected applications. How much cheaper photovoltaics can get—and how much
costlier their competition might get—remain unclear.
Ocean Energy—Too costly and too disruptive? Harnessing tides by damming
estuaries is both costly and highly disruptive environmentally. Turbines in tidal
straits and devices for harnessing wave energy must be both inexpensive and
robust in the hostile marine environment; the needed combination may be unat-
tainable. Ocean thermal energy conversion must move immense quantities of sea
water in order to extract a tiny fraction of its energy; whether the undersea equip-
ment involved can be made both cheap enough and survivable enough is unclear,
as are the ecological consequences of large-scale use.
Nuclear Fission—Too unforgiving? Nuclear fission is unforgiving of error in
design and operation of reactors, reprocessing plants, fuel-fabrication facilities,
and waste transport and disposal—and unforgiving of malice by those who would
attack the facilities for economic and public-health impact or divert the technolo-
gy and materials for making nuclear weapons. Whether improvements in technol-
ogy and management can outpace the growth in the opportunities for error and
malice as the nuclear enterprise grows is unclear. If nuclear energy were to under-
go large expansion and then prove unacceptable to the public because of high inci-
dence of accidents and/or terrorism in the expanded enterprise, the economic dis-
ruption from shutting it down would be large.
Nuclear Fusion—Too difficult? After more than 50 years of effort and the
expenditure of perhaps 30 billion current U.S. dollars in fusion R&D worldwide,
the best-performing devices aimed at harnessing for power production the process
that powers the stars and hydrogen bombs still require more energy to run than
they produce. The obstacles to ultimate success lie not only in the physics of con-
fining fusion fuel at its ignition temperature of 100 million degrees C or more, but

innovations / spring 2006 5

John P. Holdren

also in the advances in materials science and systems engineering needed to build
a reliable and affordable reactor around a fusion fire.
Hydrogen—Only an energy carrier, not a source. Chemically unbound hydro-
gen does not exist in significant quantities on Earth, and extracting it from the
forms in which it is most abundantly found—water and hydrocarbons—costs
more energy than chemical reactions of the resulting hydrogen can yield. If con-
trolled fusion succeeds, the heavy hydrogen isotopes and perhaps even ordinary
hydrogen would become nuclear fuels…but see above. Unless and until fusion suc-
ceeds, hydrogen will remain merely an energy carrier that, like electricity, may be
prized for its convenience, versatility, and low environmental impact at the point
of end use, but requires the use of a primary energy source for its production.
Improving Energy-End-Use Efficiency—Not enough education? Increasing the
efficiency with which energy is converted into the goods and services that people
want—comfort, mobility, illumination, refrigeration, the powering of industrial
processes, and so on—is equivalent to an energy source, because kilowatt-hours or
liters of fuel saved in one application can be used for another. Such end-use-effi-
ciency improvements are (and are destined to remain for some time to come) the
cheapest, cleanest, surest, most rapidly expandable energy option we have. The
ultimate limits on this option are imposed by thermodynamics, but much more
salient today are the limits imposed on the expandability of end-use-efficiency
improvements by lack of knowledge by firms and consumers about the opportu-
nities that exist and how to exploit them.
Beyond competing goals and the lack of a silver bullet, the third major reason
the energy issue is so challenging is the large embodied capital investment and long
turnover times of the world’s energy-supply and end-use systems, which create
large hurdles to transforming those systems as rapidly as the determinants of what
is desirable and necessary are changing. The replacement cost of today’s global
energy-supply system—all of the power plants, transmission lines, drilling rigs,
pipelines, refineries, coal mines, and so on—is in the range of $12 trillion, and this
immense capital investment turns over with a characteristic time of 30-40 years,
the average operating lifetime of the facilities involved. The stock of energy-using
artifacts—buildings, appliances, cars and trucks, airplanes, industrial machinery—
represents an even larger investment, with turnover times ranging from somewhat
shorter (cars, appliances) than that of energy-supply facilities to considerably
longer (buildings). Adding to the inertia created by these huge investments and
long time scales is the entrenched economic and political power of the organiza-
tions—public as well as private—that achieved their powerful positions by creat-
ing and sustaining the historical and current patterns of energy supply and
demand and are understandably interested in preserving that status quo.
The energy-system inertia that results from these circumstances, combined
with the typical multi-decade time scale for research, development, and demon-
stration to bring a new energy option even to the threshold of competitiveness
with the entrenched approaches, means that it is possible and even likely for prob-

6 innovations / spring 2006

The Energy Innovation Imperative

lems with the status quo to materialize more rapidly than the energy system can
adjust to address them. When this timing mismatch is compounded by additional
time lags in developing a scientific consensus about the harmful phenomena (as in
the case of understanding both the health and the climate impacts of the emissions
from fossil-fuel burning), the chances of being “locked in” to energy-system char-
acteristics that impose higher than expected costs and risks for decades only
increases.

THE TWO MOST DIFFICULT CHALLENGES

Reflection on what makes energy challenges difficult in general and on the partic-
ular characteristics of the challenges faced by the United States and the world in
the early 21st century has led most thoughtful observers to the conclusion that two
of those challenges stand out above all others in their combination of difficulty
and danger. These are:
y how to reduce the macroeconomic vulnerability arising from oil dependence
overall, and the balance-of-payments and foreign policy liabilities associated with
the part that is imported, despite huge and growing liquid-fuel demands from the
transport sector; and
y how to provide the affordable energy needed to sustain prosperity where it now
exists, and to create and sustain it where it now doesn’t, without entraining intol-
erable disruption of global climate by the emissions from fossil-fuel use.
These challenges would be daunting even if only wealthy, industrialized
countries had to face them. But the difficulties (and, in some respects, the dangers)
are larger still for the less developed world. And the solutions require that
everybody get them right, because both the oil market and the machinery of
climate are global.

The Oil Problem

In 2005, the United States used 20.7 million barrels of petroleum products per day
(Mb/d), equivalent in energy content to 19.1 Mb/d of crude petroleum. This oil
accounted for 40% of U.S. primary energy supply and nearly all of the energy used
by the transport sector of the economy. U.S. net imports of crude petroleum and
petroleum products in 2005 were equivalent to 12.5 Mb/d of crude petroleum,
meaning the United States was dependent on imports for 65.4% of its oil.2 This was
the largest percentage and the largest absolute oil-import dependence in U.S. his-
tory. OPEC provided 41% of those imports, the Persian Gulf 17%. The cost of U.S.
oil imports in 2005 was $231 billion, accounting for 30% of the net U.S. trade
deficit in that year.
The “reference projection” of the U.S. Energy Information Administration
(EIA) for 2030 shows an increase of U.S. consumption to 25.3 Mb/d of crude-oil
equivalent, with the transport sector the dominant driver of the growth (as in the
past).3 Import dependence, under the reference projection, reaches 68% by 2030.
OPEC’s share of those imports—and of world petroleum supply—is expected to

innovations / spring 2006 7

John P. Holdren

grow over this period, consistent with the distribution of the world’s remaining
ultimately recoverable resources of conventional oil.
World oil production in 2005 was equivalent to about 80 Mb/d of crude petro-
leum, accounting for 34% of global primary energy supply. Nearly 40% of the oil
production came from OPEC, and 64% of it moved in world trade. The EIA’s 2006
reference forecast for 2030 shows world production reaching nearly 120 Mb/d.
China’s oil imports are forecasted to increase from 3 Mb/d in 2004 to circa 12
Mb/d in 2030 (comparable to U.S. oil imports today), more than half of that com-
ing from the Persian Gulf.
The economic dimension and the international-security dimension of these
oil dependencies are complicated and interconnected. Firstly, a country’s econom-
ic vulnerability to oil-price shocks is proportional to the country’s total depend-
ence on oil, not just on its
import dependence. That is so
because, in a world market, an
Global climate change is economy pays any increase in
increasingly recognized as both the per-barrel price on every
barrel used, not just on the
the most dangerous and the most barrels imported. (Import
intractable of all of energy’s share does matter in terms of
balance of payments, of
environmental impacts—indeed, course.)
The link to conflict arises
the most dangerous and because the extent of a major
intractable of all of civilization’s country’s economic vulnera-
bility in relation to oil—say,
environmental impacts, period. that of the United States or
China—affects the chances
that it will resort to military
action to try to prevent or terminate supply disruptions and the attendant price
shocks. It also affects a country’s freedom of action in how it pursues other aspects
of its foreign-policy agenda (for example, in the case of the United States, the way
in which it pursues its homeland-security/counter-terrorism agenda in its rela-
tions with oil-producing countries, some of which export terrorism as well as oil).
In principle, the dangers of supply disruptions and price shocks can be allevi-
ated in a number of ways: increasing domestic production of conventional oil in
one’s own country; encouraging such increases in other countries in diverse geo-
graphic regions; encouraging increased production of unconventional oil
resources (heavy oils, tar sands, oil shale) in one’s own and other countries;
increasing the production of liquid fuels from coal and from biomass; and reduc-
ing the liquid-fuel intensity of economic activity (energy in liquid fuels divided by
GDP) by a combination of shifting to non-liquid fuels (solids, gases, electricity) in
some applications and increasing the energy efficiency of the remaining liquid-
fueled activities (above all cars, trucks, buses, and aircraft).

8 innovations / spring 2006

The Energy Innovation Imperative

In practice, most of these approaches suffer from severe limitations. Even if the
United States opens the Arctic National Wildlife Refuge to oil production, it is
unlikely that U.S. domestic production of petroleum can be prevented from con-
tinuing to gradually decline over the next 25 years. Prospects in most other coun-
tries—outside the unstable regions that are more part of the problem than part of
the solution—are not much better. Unconventional oil resources are considerably
more energy-intensive to produce—and more polluting—than conventional oil.
Making synthetic oil from coal is costly, water-intensive and, with current technol-
ogy, polluting.
Shifting from oil to natural gas has already happened to a significant extent in
the electricity-generating, residential-heating, and industrial sectors of the United
States and a number of other countries, but one of the ongoing consequences of
this is the emergence of a global natural-gas market in which, as with oil, an
increasing fraction of the supply seems destined to come from politically unpre-
dictable regions. Continuing a shift to natural gas, therefore, may only succeed in
replicating the problems of excessive dependence and vulnerability from which the
world is trying to escape in the case of oil.
Expanding the use of biofuels and accelerating improvements in oil end-use
efficiency, particularly in the transport sector, are more promising, but they are not
happening rapidly enough to reverse the worldwide trend of increased oil and oil-
import dependence. The EIA reference projection for the United States—which
takes into account current trends, current policies, and current and projected ener-
gy prices and energy-technology characteristics—shows the U.S. share of oil in pri-
mary energy use still at 40% in 2030 while the fraction of oil imported increases.
The EIA’s reference projection for the world as a whole has the share of primary
energy provided by oil increasing over this period, and the fraction of world oil
moved in world trade likewise increases in the reference case. Achieving a signifi-
cant decline in the world’s dependence on oil and oil imports in this period will
require, evidently, far bigger increases in substitutes for oil and in oil-end-use effi-
ciency than currently seem to be in store.

The Climate-Change Problem

Global climate change is increasingly recognized as both the most dangerous and
the most intractable of all of energy’s environmental impacts—indeed, the most
dangerous and intractable of all of civilization’s environmental impacts, period.
It is the most dangerous because climate is the “envelope” within which all other
environmental conditions and processes operate. That envelope is not just a mat-
ter of global-average surface temperature (to which the misleadingly innocuous
term “global warming” applies) but of averages and extremes of hot and cold, wet
and dry, snowpack and snowmelt, wind and storm tracks, and ocean currents and
upwellings; and not just the magnitude and geographic distribution of all of these,
but also the timing. Distortions of this envelope of the magnitude that are under-
way and in prospect are likely to so badly disrupt the environmental conditions

innovations / spring 2006 9

John P. Holdren

and processes influenced by climate as to adversely affect every dimension of

human well-being that is tied to the environment, including:
y the productivity of farms, forests, and fisheries;
y the geography of disease;
y the prevalence of oppressive heat and humidity;
y the damages to be expected from storms, floods, and wildfires;
y the property losses to be expected from sea-level rise;
y the expenditures that must be made on engineered environments (e.g., dams,
dikes, air-conditioned spaces); and
y the distribution and abundance of valued species as well as pests.
Global climate change is the most intractable of environmental problems
because the dominant driver of the disruption—emission of carbon dioxide (CO2)
from fossil-fuel combustion—is a deeply embedded consequence of the way in
which civilization is today acquiring 80 percent of its energy.4 Carbon dioxide is
not a trace contaminant but a major combustion product, generated in immense
quantities (about 27 billion metric tons of CO2 from fossil-fuel burning worldwide
in 2004). The global energy-supply system, as already noted, turns over only slow-
ly, and the fossil-fuel-burning components around which 80% of it is based
include some of the longest-lived components of all: coal-burning power plants
can run for 60 years or more. This would be less problematic if the existing fossil-
fuel burning technologies could be easily and inexpensively retrofitted to capture
the CO2 rather than releasing it to the atmosphere; but it appears that they cannot
be. Capturing CO2 and sequestering it away from the atmosphere is not easy; it is
not cheap; and it appears to be much more difficult and costly in retrofit than for
new technologies designed and located from scratch to have this ability.
While important aspects of the process of human-induced changes in global
climate remain to be fully elucidated scientifically, including especially the exact
timing and geographic distribution in which future impacts will unfold, there is no
longer room for serious doubt about the basic characteristics of what is happen-
ing:
We know—from thermometer records in the atmosphere and the oceans, and
from ice cores, bore holes, tree rings, corals, pollens, sediments, and more—that
Earth’s climate is now changing at a pace far outside the range of expected natural
variation, and in the opposite direction from what the known, natural, cyclic influ-
ences on climate would otherwise be causing at this time.5
We should be cooling, but we are warming up: by ~0.8°C in global average sur-
face temperature in the last 150 years, more over the continents, several times that
over the continents at high latitudes. On a worldwide average, the 12 warmest years
of the last 150 have all occurred since 1990, 20 of the 21 warmest since 1980. The
last 50 years appear to have been the warmest half century in 6000 years.
As expected in a warming world, moreover, observations over recent decades
also show that glaciers are retreating, sea ice is shrinking, Greenland and Antarctic
ice is melting, permafrost is thawing, heat waves and wildfires are multiplying, and

10 innovations / spring 2006

The Energy Innovation Imperative

storm and flood damages are soaring.6

A major part of the cause is clearly the well documented, human-caused
buildup of heat-trapping gases (“greenhouse gases” or GHG) in the atmosphere,
most importantly CO2 from fossil-fuel combustion and deforestation but also
methane, nitrous oxide, halocarbons, and tropospheric ozone. A buildup in the
atmospheric concentra-
tion of heat-absorbing
black soot has also been While important aspects of the
important, although the
warming effects of the process of human-induced changes
GHG and the soot have in global climate remain to be fully
been partly cancelled by
the cooling effect of a elucidated scientifically, including
build-up of light-reflect-
ing particles, likewise
especially the exact timing and
caused by human activi- geographic distribution in which
ties.
Causality is estab- future impacts will unfold, there is
lished by the match no longer room for serious doubt
between the magnitudes
and the geographic and about the basic characteristics of
temporal patterns of the what is happening.
observed changes in
atmospheric and oceanic
temperatures with what
theory and models say should result from the known buildup of atmospheric GHG
concentrations that human activities have caused, after the effects of observed
changes in anthropogenic and volcanic particulate matter and best estimates of
changes in solar output are taken into account.
To be credible, the handful of “skeptics” about human causation of current
global climate change would need both to explain what alternative mechanism
could account for the pattern of changes that is being observed and to explain how
it could be that the known human-caused buildup in GHG is not having the effects
predicted for it by the sum of current climate-science knowledge (since, by
assumption, something else is having these effects). No skeptic has met either test.
As for the future of human-caused changes in global climate, mid-range sce-
narios developed by the Intergovernmental Panel on Climate Change (IPCC) and
others project continuing increases of 0.2-0.4°C per decade in global-average sur-
face temperature over the course of the 21st century, hence a 2-4°C increase
between 2000 and 2100.7 Mid-continent warming is expected to be 2 to 3 times
greater, and high-latitude warming over the continents larger still. The Earth is
then likely to be warmer than at any previous time in the tenure of Homo sapiens
on the planet. In this scenario, the carbon dioxide concentration in 2100 will be

innovations / spring 2006 11

John P. Holdren

around 700 parts per million by volume (ppmv), compared to the pre-industrial
level of about 280 ppmv; and CO2 will be accounting for about 60% of the total
warming influences from human changes to the atmosphere (the other main ones
being increases in the concentrations of methane, nitrous oxide, halocarbons, tro-
pospheric ozone, and black soot), compared to 45% in 2000. The increase in sea
level by 2100, assigned a mid-range estimate of about half a meter by the IPCC in
2001, now seems capable of reaching several times that (although the matter
remains highly uncertain).8
Faced with continuing climatic change of this magnitude on a “business as
usual” trajectory, society has three options:
The first is mitigation, which
means measures to reduce the
Human caused climate change pace and the magnitude of
changes in global climate being
is already occurring. Adaptation caused by human activities.
efforts are already taking place Examples of mitigation include
reducing emissions of CO2,
and must be expanded. But other greenhouse gases, and
adaptation becomes costlier and black soot; enhancing “sinks”
for greenhouse gases; and “geo-
less effective as the magnitude engineering” to counteract the
warming effects of increases in
of climate change grows. greenhouse gases and soot that
occur.
The second option is adap-
tation, which means measures to reduce the adverse impacts on human well-being
resulting from the changes in climate that occur. Examples of adaptation include
changing agricultural practices, strengthening defenses against climate-related dis-
ease, and building dams and dikes to control flooding and sea-level rise.
The third option is suffering the adverse impacts that are not avoided by either
mitigation or adaptation.
Clearly, mitigation and adaptation are both essential. Human-caused climate
change is already occurring. Adaptation efforts are already taking place and must
be expanded. But adaptation becomes costlier and less effective as the magnitude
of climate change grows. The greater the amount of mitigation that can be
achieved at affordable cost, the more manageable will be the burdens placed on
adaptation and the smaller will be the suffering that neither mitigation nor adap-
tation succeeds in avoiding.
The question of how much mitigation would be prudent is a crucial one for
energy strategy, given the central role of fossil-fuel derived CO2 in the climate
problem. (Fossil-fuel burning added about 7.3 billion tonnes of C in carbon diox-
ide to the atmosphere in 2004; net deforestation probably contributed 1.5 to 2.5
billion tonnes, and cement manufacture about 0.2 billion tonnes.) Relevant here

12 innovations / spring 2006

The Energy Innovation Imperative

is the goal embodied in the UN Framework Convention on Climate Change of

1992 (which was ratified by the United States and some 180 other countries). It
calls for “stabilization of greenhouse gas concentrations in the atmosphere at a
level that would prevent dangerous anthropogenic interference with the climate
system”. There was no formal consensus in 1992 as to what constitutes “dangerous
anthropogenic interference”, however, or on what level of greenhouse-gas concen-
trations would produce it. And there is still no formal consensus on these points.
It is becoming clear, nonetheless, that the current level of anthropogenic inter-
ference is dangerous. With the increase in global-average surface temperature
above the pre-industrial level amounting only to about 0.8°C, the world is already
experiencing rising incidence of floods, droughts, heat waves, wildfires, coral
bleaching, summer melting of sea ice and permafrost, shrinkage of the mountain
glaciers that stabilize the flows of many of the world’s great rivers, drying out of
rainforests, and category 4 and 5 cyclones, among other impacts. Owing to the
thermal inertia of the oceans and the associated time lag in the climate system’s
coming to equilibrium with greenhouse-gas induced changes in the Earth’s radia-
tion balance, moreover, the global-average surface temperature would increase
another 0.6°C even if the atmospheric concentrations of CO2 and the other
anthropogenic greenhouse gases were stabilized at today’s values. Such instant sta-
bilization cannot be achieved. Even if the growth of emissions could be halted
immediately, the stabilized emissions would produce continuing growth in atmos-
pheric concentrations because of the long residence time of greenhouse gases in the
atmosphere (of the order of a century for CO2).
Models of the global carbon cycle allow specifying the families of emissions
trajectories over time that would lead to stabilization of the atmospheric CO2 con-
centration at any given level above today’s. (There is more than one trajectory for
any given stabilization level, since the key characteristic is the cumulative emis-
sions over century-scale periods; an emissions trajectory that stays on the “busi-
ness as usual” path for some decades and then departs from it very sharply can
have the same cumulative emissions as a trajectory that departs from “business as
usual” sooner but more gradually.) Trajectories that change smoothly and relative-
ly gradually, and that avoid concentrating too much of the burden of reductions
on either the short term or the long term, are likely to be preferable from the stand-
point of minimizing costs and societal disruption.
Emissions curves of this type that correspond to stabilizing atmospheric CO2
at twice its preindustrial level—that is, at about 550 ppmv—start to deflect signif-
icantly from the “business as usual” trajectory around 2020, peak at emissions of
circa 11 billion tonnes of C per year around 2040, decline to about 7 billion tonnes
of C per year by 2100, and continue a gradual decline to about 3 billion tonnes of
C per year in 2200. Under a mid-range estimate of the sensitivity of global-average
surface temperature to CO2 concentration, such a trajectory would produce an
equilibrium increase of 3°C above the pre-industrial value if the other human and
nonhuman warming and cooling influences on climate over this period cancelled

innovations / spring 2006 13

John P. Holdren

each other out (as happened approximately between 1750—the nominal pre-
industrial benchmark—and 2000, and as could happen out to 2100 if reductions
in non-CO2 greenhouse gases are matched by reductions in the emissions of light-
reflecting atmospheric particles and the pollutant gases that lead to formation of
such particles).
Until a few years ago many analysts and groups were suggesting that stabiliza-
tion of atmospheric concentrations at a level corresponding to a 3°C increase was
in fact a suitable target—something of a compromise between the highest level at
which climate-change impacts might be manageable (taking into account the
potential for adaptation) and the lowest level that might be achievable (taking into
account the known mitigation options and their estimated costs). The last few
years of accumulating evidence about impacts already being encountered at only
0.8° C above the pre-industrial average temperature, however, have led many ana-
lysts to argue for a more ambitious target, with some (including the European
Union) settling on 2°C. To have a good chance of holding the average warming to
2°C, the sum of the human influences would need to be held to the equivalent of
CO2’s stabilizing at about 450 ppmv (compared to the 2005 value of 380 ppmv).9
An emissions curve corresponding to stabilizing atmospheric CO2 at 450 ppmv
should depart from “business as usual” much sooner than the curve for 550
ppmv—by about 2012. It should peak no higher than about 9 billion tonnes of C
per year, around 2020, and should be down to about 3.5 billion tonnes per year by
2100 and 2.5 billion tonnes per year by 2200. (Of course, if the non-CO2 influences
on climate add up to a net warming over this period, the CO2 emission curves
would need to be even lower than described here in order for the overall effect to
be equivalent to 550 or 450 ppmv of CO2, respectively.) Mid-range “business as
usual” (BAU) scenarios, by contrast, entail emissions around 20 billion tonnes of
C per year in 2100.
The difference between the BAU path and the stabilization trajectories just
described is immense. Cumulative emissions of C over the 21st century under a
mid-range BAU path would be in the range of 1400 billion tonnes; for the indi-
cated stabilization trajectories, cumulative emissions would be in the range of 500-
800 billion tonnes of C (less if the non-CO2 influences add up to a net warming).
The difference of 600-900 billion tonnes measures the size of the mitigation chal-
lenge to which the world must rise to have a reasonable chance of averting climate-
change disaster.
The types of approach available for this mitigation have already been men-
tioned. In brief, one can (i) reduce the offending emissions, (ii) increase the rate of
removal of the offending substances from the atmosphere, or (iii) try to change
other climate-relevant characteristics of the environment to offset the warming
influences of those substances. Taking these in reverse order…
The third approach is worthy of further study, but the “geo-engineering”
approaches considered so far appear to be afflicted with some combination of high
costs, low leverage, and a high likelihood of serious side effects. Consequently, at

14 innovations / spring 2006

The Energy Innovation Imperative

this juncture, no contribution to mitigation can be counted upon from this direc-
tion.
The second approach—increasing removal rates of GHG and soot from the
atmosphere—has considerable promise, above all in the domain of afforestation
and reforestation (wherein building up the global “standing crop” of trees pulls
CO2 out of the atmosphere and stores it in wood and soil organic matter). The
total carbon currently stored in all the world’s vegetation is estimated at 500-700
billion tonnes of C; increasing this by as much as 20% seems unlikely, and that
would take care of only 100-150 billion tonnes of the 600-900 billion tonne
requirement. (Whether the carbon stocks in soil, as opposed to vegetation, can be
increased at all in a warming world is unclear; the higher temperatures may well
increase decomposition rates on the average, driving carbon out of the soil and
into the atmosphere.)
The preceding two points mean that a heavy share of the mitigation burden
necessarily falls on reducing the offending emissions. In this connection there is
important progress to be made in reducing emissions of the non-CO2 heat-trap-
ping substances, most importantly methane and black soot but also nitrous oxide,
halocarbons, and the precursors of tropospheric ozone. The progress that can be
made with all of these, however, may well not be more than is needed just to coun-
terbalance reductions that are expected in the emissions of light-reflecting parti-
cles and their sulfur-oxide and nitrogen-oxide precursors. (Such reductions, which
by reducing cooling influences will have a warming effect, are motivated by the
desire to reduce the large public-health and acid-rain impacts of these particles.)
The “bottom line” is that a large part of the required mitigation effort must
come in the form of reducing emissions of carbon dioxide. A modest (but still
valuable) piece of this can come from reducing deforestation rates in the tropics,
which today are adding perhaps 1-2 billion tonnes of C per year to the atmosphere.
But the biggest target has to be the over 7 billion tonnes per year of carbon com-
ing from fossil-fuel combustion.
The leverage for reducing the CO2 emissions from fossil fuels can be under-
stood by representing those emissions as a four-fold product:
C emissions = population x GDP/person x energy/GDP x C/energy.
Let us consider each of the contributing factors in turn:
Population. Lower is better for many reasons. If world population were 8 bil-
lion in 2100 rather than the mid-range UN forecast of about 10 billion, holding
down the carbon emissions from the energy to make everybody prosperous would
be that much easier. Fortunately, reduced population growth can be achieved by
measures that are attractive in their own right (notably improving health care,
reproductive rights, and educational opportunities for women).
GDP per person. This is not a lever that most people would want to use to
reduce emissions, because higher GDP/person is generally considered preferable to
lower. People are not getting rich as fast as they think, however, if GDP growth is

innovations / spring 2006 15

John P. Holdren

being achieved at the expense of the environmental underpinnings of well-being.

Internalizing environmental costs of economic growth (including those of climate
change) may slow that growth a bit, but probably not by very much. In addition,
changes in lifestyle in industrial and other urban regions can be envisioned that
would increase quality of life even though they reduced GDP (e.g., shorter com-
mutes to work, or replacing physical commuting with telecommuting).
Energy intensity of GDP. Getting more GDP out of less energy—that is,
increasing energy efficiency—has been a long-term trend. (On a global basis, the
improvement has been averaging about 1% per year for decades.) The process of
achieving this entails deploying more efficient cars, trucks, planes, buildings, appli-
ances, and manufacturing processes, as well as structural shifts in the economy
(toward less energy-intensive forms of economic activity) and increasing the effi-
ciency with which primary energy forms are converted to electricity and portable
fuels. These trends could be substantially accelerated. Indeed, this domain
undoubtedly offers the largest, fastest, cheapest early leverage on carbon emissions.
Carbon intensity of energy supply. The ratio of C emitted in CO2 per unit of
primary energy supplied to the economy has also been falling, but more slowly
than the energy intensity of GDP. The principal leverage for reducing this ratio
more rapidly is in accelerating the introduction of low-carbon and no-carbon
energy-supply options, notably nuclear energy, renewables, and advanced fossil-
fuel technologies that can capture the CO2 for sequestration away from the atmos-
phere.
How much may be demanded of measures to reduce these last two ratios—the
energy intensity of economic activity and the carbon-emissions intensity of ener-
gy supply—can be illustrated by a somewhat oversimplified but still instructive
calculation of the combinations of such reductions that would be needed to get on
the above-described 550-ppmv stabilization trajectory under the assumption that
population growth and growth of GDP per person proceed on a course that main-
tains real economic growth worldwide at 2.4% per year over the course of the 21st
century.10 In round numbers, what the calculation shows is as follows:
If the recent historical rate of reduction of energy of intensity of GDP world-
wide, 1% per year, were to persist over the entire century, the amount of non-car-
bon-emitting energy supply (renewables, nuclear, and advanced fossil-fuel tech-
nologies with carbon capture and sequestration) needed to be on the 550-ppmv
stabilization trajectory would be 800 exajoules per year in 2050 and 1500 exajoules
per year in 2100. (This is to be compared with 100 exajoules from these sources in
2004 and 400 exajoules from fossil-fuels in that year.)
If the historical rate of reduction of energy intensity could be increased by half,
from 1% to 1.5% per year over the whole world and the whole century, the need
for non-carbon-emitting energy supply for the 550-ppmv stabilization trajectory
would still grow to nearly 400 exajoules by 2050 and 800 exajoules by 2100.
Only if the historical rate of reduction of energy intensity could be doubled to
2% per year over the whole world and the whole century could the need for non-

16 innovations / spring 2006

The Energy Innovation Imperative

carbon-emitting energy supply by 2100 be held to “merely” the 400 exajoules per
year being provided by fossil fuels in 2004.
The requirements for non-carbon-emitting energy supply would be signifi-
cantly larger still, of course, for the 450-ppmv stabilization trajectory.
What emerges from these figures is the clear finding that success in address-
ing the climate-change challenge is likely to require enormous efforts both on
increasing the pace of energy-intensity reductions worldwide and on accelerating
the deployment of non-carbon-emitting energy sources in place of the conven-
tional fossil-fuel technologies that dominate today’s global energy system.
Nothing remotely like the needed scale of effort on either front is happening today.
(After all, the “business as usual” projection—a prescription for climate-change
catastrophe—was constructed by assuming continuation of more or less what is
currently going on.)

Interaction of the Oil-Dependence and Climate-Change Problems

If the oil-dependence problem is a 600-pound gorilla already in the room, the cli-
mate-change problem is an 800-pound gorilla in the process of beating down the
door. Society does not have the luxury of concentrating its efforts on the first prob-
lem with the expectation of turning to the second one later. The two must be
addressed together, with attention to the ways in which they interact.
The good news in this respect is that there are a number of “win-win”
approaches that can reduce both oil-dependence and climate-change risks at
modest cost—or even, as Amory Lovins has long been energetically pointing out
and an increasing number of corporations have been demonstrating, at a profit.11
Sharply increasing passenger-vehicle fuel economy offers large potential in this
category, as do advanced biofuels and a wide range of technical improvements in
energy-intensive industrial processes.
The bad news, on the other hand, is that some approaches to reducing oil
dependence would make the climate-change problem worse. Conspicuous in this
category is the production of synthetic petroleum substitutes from tar sands, oil
shales, and coal-to-liquid technologies, all of which entail large increases in CO2
emissions per liter of delivered liquid fuel unless approaches are chosen that cap-
ture and sequester much of the CO2 that otherwise is emitted during production
of such liquids.
The “peak oil” debate—about when world production of conventional oil will
peak and begin to decline, as well as about the consequences to be expected when
this occurs—is thought by some to be germane to the climate issue in the sense
that passing the oil-production peak might mark the beginning of a transition
away from civilization’s dependence on fossil fuels overall and thereby an amelio-
ration of the CO2 emission driver of global climate change. But there is little
agreement among specialists about whether peak oil is 5 years away or 50, and
even less agreement about whether its occurrence will precipitate a shift away
from fossil fuels or just a shift among them—with production of liquids from

innovations / spring 2006 17

John P. Holdren

solid fossil fuels as described above (or from unconventional natural gas
resources) taking over the burden from conventional petroleum.
I suggest that for purposes of energy-policy planning today it does not really
matter very much who is right about peak oil. The economic and security perils
of the world’s current and growing dependence on oil tell us that we need to move
smartly to reduce that dependence no matter whether peak oil is close or far away.
And the looming danger of unmanageable climate change tells us that we must
choose ways to do this that reduce rather than increase the energy sector’s emis-
sions of CO2.

THE ROLE OF INNOVATION

The multiplicity of challenges at the intersection of energy with the economy, the
environment, and international security—led by the oil-dependence and climate-
change challenges just described—add up to a need for policies designed for two
ends:
y to help society find and implement a satisfactory compromise among compet-
ing economic, environmental and security objectives—which includes trying to
leave the biggest margins of safety against the biggest dangers—given the resources
and technologies available at any given time, and
y to accelerate the processes of energy-technology innovation that, over time, can
reduce the limitations of existing energy options, can bring new options to
fruition, and thereby can reduce the tensions among energy-policy objectives and
enable faster progress on the most critical ones.
These ends cannot be achieved by markets alone, without supplementary poli-
cies, because many of the goals relate to public goods (such as national security
and meeting the basic energy needs of society’s poorest members) and externali-
ties (such as air pollution and greenhouse gases) that are not priced in markets
unless policies achieve this.
A further implication of the characteristics of today’s energy challenges is that
society will do better to pursue a broad portfolio of improved energy-supply and
end-use options, rather than putting its eggs in too few baskets. The merits of such
diversity are manifold: it provides flexibility to respond to changing conditions
and new information (an “insurance policy” for an uncertain world), including
providing the possibility of discarding options that ultimately prove unsuitable; it
takes into account that, even after all plausible technological improvements, there
comes a point in the expansion of any energy option where rising marginal costs
and/or risks make further expansion unattractive (meaning a broad portfolio is
likely to have lower costs and risks overall than a narrower set of options wherein
each has to bear too much of the load); and by combining the growth of multiple
new or improved options—each drawing on different types of material resources,
skills, and firms—it can replace status quo technologies more rapidly than would
be possible by one or two new options alone.
The need for deployment of technologies of energy supply and end-use better

18 innovations / spring 2006

The Energy Innovation Imperative

than those that now dominate the energy system is acute. Without an accelerated
transition to improved technologies, societies will find it increasingly difficult—
and in the end probably impossible—either to limit oil imports and oil depend-
ence overall without incurring excessive economic and environmental costs or to
provide the affordable energy needed for sustainable prosperity everywhere with-
out intolerably disrupting the Earth’s climate. They will not be able to improve
urban air quality while meeting growing
demands for personal transportation;
not be able to use their abundant coal Major innovations in both
resources without intolerable impacts
on regional air quality and acid rain; not technology and policy are
be able to expand the use of nuclear
energy enough to make a difference for
urgently needed but not
climate change and oil and gas depend- currently materializing at
ence while still reducing the risks of
accidents, nuclear terrorism, and the pace that is required.
nuclear-weapon proliferation.
In this context, the needed process
of innovation in energy technology must be understood as not consisting only of
research and development (R&D), but also of at least equal emphasis and
resources devoted to demonstration at commercial scale and in diverse contexts of
the technological improvements that R&D have made possible and to mechanisms
to promote accelerated deployment of those demonstrated options that offer the
greatest leverage for reducing important externalities and enhancing important
public goods.12 The energy-technology-innovation “pipeline” is full of potentially
valuable—even potentially crucial—technologies at every stage of development,
and it is no less important to push along toward full commercialization those that
are already close to that threshold than to be doing the applied research and early
development needed to move forward the more “far out” possibilities. Indeed, the
need for rapid response to the linked oil-dependence and climate-change chal-
lenges means that the world cannot afford to wait for such long-term possibilities
as fuel-cell-powered vehicles and fusion to come to fruition. This is not to say that
that investment in such long-term options is not essential, for it is; but it should
not replace or come at the expense of the needed efforts to move nearer-term, oil-
sparing, climate-friendly options into the marketplace.

INADEQUACY OF ENERGY INNOVATION EFFORTS

Current efforts in energy-technology research, development, and demonstration

(RD&D)—and in accelerated deployment of the best options that such RD&D
produces—are woefully inadequate in relation to the scale of the challenge and the
size of the opportunities.13 U.S. public and private spending on energy-technology
RD&D totals only $5-6 billion per year, less than one percent of what this country
spends for electricity and fuels. The situation in other industrialized countries

innovations / spring 2006 19

John P. Holdren

(with the conspicuous exception of Japan) is no better, and in developing countries

it is worse.
Around the world, the energy sector’s ratio of RD&D investments to total rev-
enues is well below that for any other high-tech sector of the economy. In a $45
trillion world economy (calculated using purchasing power parities), fueled by
circa $3 trillion worth of energy, total public and private investments in energy
RD&D appear to be in the range of $15-20 billion, hence something like half a per-
cent of energy expenditures and 0.03 percent of world GDP. These investments will
need to be boosted at least 2-3-fold if the world is to meet the energy challenges it
faces in the decades immediately ahead.
In principle, such an increase should not be difficult to achieve, given the mod-
est sums involved in relation to the scale of the energy enterprise. In the United
States, for example, a tripling of the federal government’s expenditures on energy
technology RD&D could be financed with a increase of about 2 cents per gallon in
the federal tax on gasoline. In practice, however, governments have proven
extremely reluctant to increase energy RD&D expenditures, even when their rhet-
oric would appear to call for such increases. Again the case of the United States
provides an instructive example: despite the Bush Administration’s consistent
rhetoric (including very conspicuously in President Bush’s January 2006 State of
the Union address) to the effect that advances in energy technology will hold the
key to addressing the oil-dependence and climate-change challenges, the appropri-
ations for energy RD&D have been essentially level in real terms since Fiscal Year
2001, and the President’s FY2007 budget request is for less money for this purpose
than the Congress appropriated for FY2006.14
Private-sector expenditures for energy RD&D are more difficult to track in
detail because of lack of comprehensive and consistent data, but such analyses as
are available for the case of the United States suggest that these investments have
been falling overall.15 The reasons usually adduced for such a decline include gen-
erally low oil and gas prices (until recently) and a corporate financial environment
that has placed particular emphasis on short-term rates of return (to which R&D
investments contribute little). Government tax incentives for corporations to
undertake more R&D are mostly modest, and the incentive of recent higher oil and
gas prices for RD&D on alternatives is weakened by industry uncertainty about
whether these recent prices will persist. Perhaps most importantly in the context of
the character of energy challenges as elaborated in this article, companies are like-
ly to continue to under-invest in developing and deploying low- and no-carbon
energy options until there is a stronger marketplace incentive for such action,
either in the form of a substantial carbon tax or its practical equivalent in the form
of economy-wide emissions caps implemented through tradable permits.

CONCLUSIONS: WHAT SHOULD BE DONE?

Accelerated improvement of the energy-supply and energy-end-use technologies

available and propagating in the marketplace will be essential if the suite of ener-

20 innovations / spring 2006

The Energy Innovation Imperative

gy challenges confronting the world in this century—most compellingly the cou-

pled problems of oil dependence and climate change—are to be successfully
addressed. Bringing about the needed pace of energy-technology innovation will
require major innovations in policy and management, aimed at
y providing the scale, continuity, and coordination of effort in energy research,
development, and demonstration needed to bring an appropriate portfolio of
improved options to the threshold of commercialization in a timely way;
y promoting and financing early deployment of the most promising options to
emerge from the RD&D process, in order to accelerate their progress down the
learning curve toward market competitiveness;
y ensuring that improved energy technologies not only diffuse rapidly through the
industrialized countries and the relatively prosperous urban sectors of developing
ones, but also reach the least developed countries and sectors;
y devising and implementing an adequate, equitable, and achievable global frame-
work for limiting global emissions of greenhouse gases;
y more effectively mobilizing the power of partnerships—among countries, levels
of government, and the public, private, academic, and NGO sectors—in achieving
all of the preceding ends; and
y more effectively communicating to the broad public the reasons all this must be
done and the benefits to be gained and dangers averted by doing it, in order to
develop and sustain the needed political support.
Numerous major, high-level, multi-sectoral studies conducted in the United
States over the past decade—and similar efforts internationally—have arrived at
more or less the same recommendations and have elaborated what carrying them
out would entail.16 Insights about what to do and how to do it, then, are not lack-
ing. What has been and remains missing is political leadership at the level needed
to make it happen. Let us hope this changes soon.

We invite reader comments. Email <[email protected]>.

1. Goldemberg, J., editor, World Energy Assessment: Energy and the Challenge of Sustainability (New
York: UN Development Programme, UN Department of Economic and Social Affairs, and World
Energy Council), 2000, <http://www.undp.org/seed/eap/activities/wea/drafts-frame.html>.
2. U.S. Energy Information Administration, Monthly Energy Review (Washington, DC: EIA),
March 2006, <http://www.eia.doe.gov/emeu/mer/contents.html>. Note that a smaller import-
dependence figure of 12.35 Mb/d / 20.66 Mb/d = 59.8% is shown for 2005 in this reference's Table
1.7, based on summing crude-oil and petroleum-product net imports and dividing by total con-
sumption of petroleum products. Because crude oil and petroleum products have differing energy
content per barrel, combining barrels per day of crude and products can be misleading in energy
terms; calculating import dependence based on energy content should be preferred.
3. U.S. Energy Information Administration, Annual Energy Outlook (Washington, DC: EIA),
February 2006, <http://www.eia.doe.gov/oiaf/aeo/index.html>.

innovations / spring 2006 21

John P. Holdren
4. Global primary energy supply in 2004 amounted to about 500 exajoules, of which 34% came
from oil, 25% from coal, 21% from natural gas, 13% from renewables (including traditional bio-
mass fuels, sometimes excluded from official tabulations), and 6% from nuclear. (Total adds to
99% rather than 100% because of rounding. 1 exajoule = 10^18 joules = 0.95 quadrillion Btus.)
See International Energy Agency, Key World Energy Statistics 2005 (Paris: IEA), 2005,
<http://www.iea.org/textbase/nppdf/free/2005/key2005.pdf>; and BP Amoco, Annual Review of
World Energy 2005 (London: BP), June 2005, <http://www.bp.com/>.
5. On this and subsequent points on climate-change science not otherwise referenced, please see
Intergovernmental Panel on Climate Change, Climate Change 2001: Synthesis Report—Summary
for Policymakers (Geneva: IPCC), 2001, <http://www.ipcc.ch/pub/un/syreng/spm.pdf>; National
Academy of Sciences, Climate Change Science: An Analysis of Some Key Questions (Washington,
DC: National Academy Press), 2001,
<http://books.nap.edu/html/climatechange/climatechange.pdf>; The Arctic Council, Arctic
Climate Impact Assessment ( Moscow: Arctic Council Secretariat), 2005.
<http://www.acia.uaf.edu>; Millennium Ecosystem Assessment, Ecosystems and Human Well-
Being: Synthesis (Nairobi: UN Environment Programme), 2005, <http://www.millenniumassess-
ment.org/en/index.aspx>; James Hansen et al., Earth’s energy imbalance: confirmation and impli-
cations, Science, vol. 308, pp 1431-35, 3 June 2005,
<http://www.columbia.edu/~jeh1/hansen_imbalance.pdf>; and the websites of the U.S. National
Climate Data Center <http://www.ncdc.noaa.gov/oa/ncdc.html> and the Hadley Centre of the UK
Meteorological Office <http://www.met-office.gov.uk/research/hadleycentre/index.html>.
6. Some of the increase in damages from storms and floods is due to the growth of population and
infrastructure in the vulnerable regions, but increasingly persuasive data and analysis relating to
rising trends in extreme precipitation events and the most powerful tropical cyclones (hurricanes
and typhoons) indicate that global climate change is also a factor.
7. These figures correspond to a scenario that the IPCC called IS92a, a mid-range trajectory
among many studied by the IPCC in an extensive modeling exercise using a range of assumptions
about the future size and composition of economic activity and about technological change. A
similar result for the dominant CO2-emission driver of climate change can be obtained much
more simply by assuming that real economic growth worldwide averages 2.4% per year over the
21st century, the energy intensity of economic activity declines at 1.0% per year, and the carbon
intensity of energy supply (carbon emitted in CO2 per unit of primary energy supplied to the
economy) declines at 0.2% per year. These rates of decline of energy intensity and carbon intensi-
ty correspond to recent global experience and so are often characterized as “business as usual” val-
ues; the economic-growth figure is a mid-range value among projections found in the literature.
8. James Hansen, A slippery slope: How much global warming constitutes “dangerous anthro-
pogenic interference”? Climatic Change, vol. 68, pp 269-279, February 2005, <http://www.colum-
bia.edu/~jeh1/hansen_slippery.pdf>.
9. This criterion needs to be stated in terms of probabilities because of remaining uncertainties in
the relation between the CO2 concentration and the size of the global-average surface temperature
increase. This relation is called the “sensitivity” of the climate and is expressed as the size of the
temperature increase corresponding to a doubling of the pre-industrial CO2 concentration.
Sensitivity is thought to be between 1.5 and 4.5°C, with a “best estimate” in the middle of this
range at around 3°C.
10. Such a calculation is described in detail in John P. Holdren, U.S. Climate Policy Post-Kyoto, in
The Convergence of U.S. National Security and the Global Environment, The Aspen Institute
Congressional Program, vol. 18, no. 3, 2003, pp 7-24, <http://bcsia.ksg.harvard.edu/BCSIA_con-
tent/documents/ClimatePostKyoto.pdf>.
11. See, e.g., Amory B. Lovins, E. Kyle Datta, Odd-Even Bustnes, Jonathan G. Koomey, and Nathan
J. Glasgow Winning the Oil End-Game (Old Snowmass, CO: Rocky Mountain Institute), 2005
<http://www.rmi.org/images/other/WtOE/WtOEg_72dpi.pdf>; Michael Northrop and David
Sassoon, The mythology of economic peril, Environmental Finance, June 2005, pp 18-19; and The

22 innovations / spring 2006

The Energy Innovation Imperative
Climate Group, Carbon Down / Profits Up, Second Edition, 2005, <http://www.rbf.org/publica-
tions/CDPU.pdf>.
12. For elaboration on this and related points about the process of innovation, see Kelly Sims
Gallagher, John P. Holdren, and Ambuj Sagar, Energy-technology innovation, Annual Review of
Resources and Environment, 2006 (forthcoming).
13. See Gallagher et al. (Note 12) and World Energy Council, ERD&D Study Group, Energy
Technologies for the 21st Century, August 2001,
<http://www.worldenergy.org/wec-geis/publications/reports/et21/introduction/introduction.asp.>
14. See Kelly Sims Gallagher, “Making sense of President Bush’s new advanced energy initiative,”
February 2006, <http://bcsia.ksg.harvard.edu/BCSIA_content/documents/Makingsense.pdf>.
15. See, e.g., J. Dooley, “U.S. National Investment in Energy R&D, Battelle PNWD-3108, July 2001,
<http://www.globalchange.umd.edu/publications/PNWD-3108.pdf>; and Daniel M. Kammen and
Gregory F. Nemet, “Reversing the incredible shrinking energy R&D budget,” Issues in Science and
Technology, Fall 2005, pp 84-88, <http://rael.berkeley.edu/EnergyRDIssues2005.pdf>.
16. See, e.g., President’s Committee of Advisors on Science and Technology (PCAST), Federal
Energy R&D for the Challenges of the 21st Century (Washington, DC: Executive Office of the
President of the United States), November 1997,
<http://bcsia.ksg.harvard.edu/publication.cfm?program=STPP&ctype=book&item_id=136>;
PCAST, Powerful Partnerships: The Federal Role in International Cooperation on Energy Innovation
(Washington, DC: Executive Office of the President of the United States), July 1999,
<http://bcsia.ksg.harvard.edu/publication.cfm?program=STPP&ctype=book&item_id=201>;
National Commission on Energy Policy, Breaking the Energy Stalemate: A Bipartisan Strategy to
Meet America’s Energy Challenges, December 2004, <http://www.energycommission.org/>; and
Stephen Byers and Olympia Snowe, Co-Chairs,
Meeting the Climate Challenge: Recommendations of the International Climate Change Taskforce
Institute for Public Policy Research, Center for American Progress, and Australia Institute,
January 2005.
<http://www.tai.org.au/Publications_Files/Papers&Sub_Files/Meeting%20the%20Climate%20Cha
llenge%20FV.pdf>.

innovations / spring 2006 23