in summary, choose 2 question from cluster to answer for 3 pages

Question Cluster for ”Will We Ever Stop
Using Fossil Fuels?”

March

1

3

,

2

017

I. Where does the title of the article come from in terms of costs of fossil
fuels?

A. What are the benefits to eliminating auto fossil fuel usage in terms
of traditional pollution sources?

B. What share of greenhouse gas emissions are due to fossil fuels?

C. What two alternatives does the article give for limiting fossil fuels
contribution to carbon?

II. What does the article report in terms of the trends of fossil fuel usage?

A. How does various fossil fuel usage compare to 2005 in the devel-
oped world? What policies contribute to this? What has hap-
pened in the developing world?

B. Why might an ”activist” policy not be needed? What is the ”de-
mand theory” component of this argument?

III. What is the goal of this article in terms of ”supply side” and ”demand
side” forces?

A. What conclusions do the authors draw from figure one? What
trends does this point to in terms of short run fossil fuel usage?

B. Do the authors believe that fossil fuel supplies will be exhausted
in a time frame relevant to global warming?

1

C. What do the authors believe are necessary from a technological
and policy point of view to achieve demand side reductions?

IV. How has the ”supply side” of fossil fuel usage changed in the last 60
years?

A. How much has shale oil increased known fossil fuel reserves? What
was the time frame for that improvement?

B. What can we learn about reserves of fossil fuels from Figure 2?

C. What conclusions do the author’s draw from the Figure about
future usage?

D. What is the theoretical relationship between technological progress
and the success rate of exploring for new oil and natural gas for-
mations?

E. What data does and does not exist on these points? What does
it show about the past? What can we say about the future of
technological progress in exploration?

F. Are there major sources of fossil fuels that have not been developed
yet? What are they? How does this increase known reserves?

V. From the demand side, will low-carbon energy sources undercut the
profitability of extracting carbon from fossil fuels?

A. What are the leading rivals for replacing fossil fuel usage in energy
generation? How have they faired?

B. Explain the term ”levelized cost of energy”. What does it measure
in terms of our model?

C. How does the cost curve for most renewables differ than the one
for fossil fuel over time? What time lag is built into these measures
for non-fossil fuels?

D. What time period do the authors use for measuring the costs of
different types of fuels? Why?

E. From Figure 4, what can we learn about these costs?

1. Why do the authors consider the cost declines from solar en-
ergy to be rapid? What caveats do the authors place on this
conclusion?

2

2. What conclusions do the authors draw about wind power?

3. The authors caveat this with three additional bits of informa-
tion. What are they? How do they affect the measured MC
of fossil fuels? What is the ”Duck Curve”?

F. How do the authors see fossil fuel replace proceeding for automo-
biles?

1. What cost issue(s) exist with electric vehicles which keep them
from being competitive?

2. What other issues exist with electric vehicles along these lines?

3. How do electric vehicles currently fair in terms of carbon re-
duction? Compared to what?

4. How do the authors compare the costs of using electric and
non-electric vehicles? Can you reproduce their calculations?
What conclusion due they draw about competitiveness of the
technologies?

5. How do the authors expect the two above numbers to change
over time? Where do these estimates come from?

6. How does Figure (6) lead the authors to draw conclusions
about competitiveness of these vehicles over the next two
decades?

7. What happens to the above numbers if we add the social cost
of capital”? What is meant by that?

G. What can we learn from this about the future of carbon impacts?

1. How much fossil fuel usage do the authors expect by 2040
without a major policy intervention?

2. How should this effect temperature and welfare? Where do
these estimates come from?

3. What market failures do the authors believe these numbers
show? What is a market failure?

4. What policy do they recommend based on this?

3

Journal of Economic Perspectives—Volume 30, Number 1—Winter 2016—Pages 117–138

F ossil fuels provide substantial economic benefits, but in recent decades, a series of concerns have arisen about their environmental costs. In the United States, for example, the Clean Air Act in 1970 and 1977 addressed concerns
over the emissions of so-called conventional pollutions, notably airborne particulate
matter, by imposing fuel economy standards on vehicles and regulations to reduce
emissions from stationary sources. During the 1980s, concerns mounted about how
the combustion of fossil fuels could lead to acid rain and rising ozone levels. The
Clean Air Act Amendments of 1990 created frameworks to reduce sulfur dioxide
and nitrogen oxide from power plant emissions, as well as from the combustion
of gasoline and diesel fuels in vehicles. However, in many of the world’s largest
cities in the emerging economies around the world, the conventional forms of
air pollution from burning fossil fuels—especially particulates, sulfur oxides,
and nitrogen oxides—are still exacting a heavy toll on human health (Chay and

Will We Ever Stop Using Fossil Fuels?

■ Thomas Covert is Assistant Professor of Microeconomics, Booth School of Business, Univer-
sity of Chicago, Chicago, Illinois. Michael Greenstone is the Milton Friedman Professor in
Economics and the College and Director of the Energy Policy Institute at Chicago, both at the
University of Chicago, Chicago, Illinois. Christopher R. Knittel is William Barton Rogers
Professor of Energy Economics, Sloan School of Management, and Director of the Center for
Energy and Environmental Policy Research, all at the Massachusetts Institute of Technology,
Cambridge, Massachusetts. Greenstone and Knittel are also Research Associates, National
Bureau of Economic Research, Cambridge, Massachusetts. Their email addresses are thomas.
covert@chicagobooth.edu, mgreenst@uchicago.edu, and knittel@mit.edu.
† For supplementary materials such as appendices, datasets, and author disclosure statements, see the
article page at
http://dx.doi.org/10.1257/jep.30.1.117 doi=10.1257/jep.30.1.117

Thomas Covert, Michael Greenstone, and
Christopher R. Knittel

j_covert_301.indd 117 1/20/16 6:57 AM

118 Journal of Economic Perspectives

Greenstone 2003; Chen, Ebenstein, Greenstone, and Li 2013; Knittel, Miller, and
Sanders forthcoming).

By the mid-1990s, concerns about the role of fossil fuels in generating emis-
sions of carbon dioxide and other greenhouse gases gained traction. Approximately
65 percent of global greenhouse gas emissions are generated by fossil fuel combus-
tion.1 Of these emissions, coal is responsible for 45 percent, oil for 35 percent, and
natural gas for 20 percent.2 To reduce carbon dioxide emissions by enough to miti-
gate the chance of disruptive climate change in a substantial way, there would seem to
be only two possible options. One is to find ways to capture carbon from the air and
store it. To a moderate extent this can be done by expanding the size of the world’s
forests, for example, but if carbon capture and storage is to be done at a scale that
will more than counterbalance the burning of fossil fuels, there would need to be very
dramatic developments in the technologies so that it could happen at a cost-effective
industrial level. The other option for reducing emissions of greenhouse gases is to
reduce future consumption of fossil fuels in a drastic manner.

A few developed countries have implemented policies to limit fossil fuel
consumption through a mixture of taxes, fees, or regulation on carbon emissions
(in the case of the European Union, some US states, Japan, Canada, and Australia),
subsidies for energy conservation (the United States and elsewhere), and the devel-
opment of low- or no-carbon energy resources (the United States and elsewhere).
In these and other OECD countries, data in the annual BP Statistical Review of
Energy show that both oil and coal consumption are down about 10 percent, while
natural gas consumption (which has lower carbon emissions) is up 10 percent from
2005 levels.3

There have been few policy responses to limit fossil fuel consumption in
developing countries, even though many of them are experiencing very high and
immediate costs from conventional air pollutants. Moreover, this group of countries
has greatly expanded its use of fossil fuels in this period, with non-OECD coal, oil,
and gas consumption up 46, 33, and 35 percent, respectively, since 2005. As a result,
global consumption of fossil fuels rose 7.5 percent for oil, 24 percent for coal, and
20 percent for natural gas from 2005 to 2014.

Two primary market forces could moderate the need for an activist policy
response to rising fossil fuel consumption. First, the ongoing consumption of fossil
fuels could cause the marginal cost of extracting additional fuel to rise to where the
marginal barrel of oil (or ton of coal or cubic meter of natural gas) will be more
costly than clean energy technologies (for examples of this argument, see Rutledge

1 US Environmental Protection Agency at http://www3.epa.gov/climatechange/ghgemissions/
global.html.
2 Carbon Dioxide Information Analysis Center at http://cdiac.ornl.gov/ftp/trends/co2_emis/
Preliminary_CO2_emissions_2012.xlsx.
3 All country-level fuel consumption data in this article come from the BP Statistical Review of World Energy
2015, available at http://www.bp.com/en/global/corporate/energy-economics/statistical-review-of
-world-energy.html.

j_covert_301.indd 118 1/20/16 6:57 AM

Thomas Covert, Michael Greenstone, and Christopher R. Knittel 119

2013; Murray and King 2012). We label this “the supply theory”—that is, the world
will run out of inexpensive fossil fuels.

Second, scientific advances could improve the energy efficiency of existing
technologies and develop newer, cheaper carbon-free technologies (for example,
see the “McKinsey curve” in McKinsey 2009). We label this “the demand theory”—
that is, the economy will stop demanding fossil fuels as alternatives become more
cost-competitive. This line of thinking is appealing. After all, who wouldn’t prefer to
consume energy on our current path and gradually switch to cleaner technologies
as they become less expensive than fossil fuels? But the desirability of this outcome
doesn’t assure that it will actually occur—or even that it will be possible.

In this article, we look back at the historical record to assess the power of natural
supply-side and demand-side forces (forces unrelated to future policy intervention) to
achieve significant reductions in fossil fuel consumption. To gain some initial perspec-
tive, consider the ratio of current proven fossil fuel reserves divided by current annual
consumption. Proven reserves are defined as fossil fuels that are technically and
economically recoverable at current prices, and so this ratio represents the number
of years of current consumption that can be economically supplied by known fossil
fuel deposits. The total amount of any category of fossil fuels that is in the ground
and recoverable at any cost is fixed, of course, but reserves can increase or decrease
depending on how extraction technologies and prices evolve.

Figure 1 plots reserves-to-consumption for oil, natural gas, and coal at the global
level. The available data for oil and natural gas span 1980 to 2014; the data for coal begin
in 1994. The striking feature of this graph is how constant the reserve-to-consumption
ratio is for both oil and natural gas. It is an empirical regularity that, for both oil and
natural gas at any point in the last 30 years, the world has 50 years of reserves in the
ground. The corollary, obviously, is that we discover new reserves, each year, roughly
equal to that year’s consumption. This phenomenon seems to be independent of the
enormous variation in fossil fuel price changes over the last 30 years. Coal, on
the other hand, shows a dramatic decrease in its reserves-to-production ratio, but
there remain many more years of coal reserves, at current consumption, than for
either oil or natural gas. Furthermore, the decline in the reserve-to-consumption
ratio for coal flattened in the early 2000s and more recently the ratio has even ticked
upward a bit. The stability of the reserve-to-consumption ratio provides the first piece
of evidence against the idea that the supply of or demand for fossil fuels are likely to
“run out” in the medium term.

In the next section of the paper, we analyze the supply behavior for fossil fuels
over the past three decades through the lenses of reserve growth and exploration
success. The story seems clear: we should not expect the unfettered market to lead
to rapid reductions in the supply of fossil fuels. Technical progress in our ability to
extract new sources of fossil fuels has marched upward steadily over time. If the
advance of technology continues, there is a nearly limitless amount of fossil fuel
deposits—at least over the time scale that matters for climate change—that, while
they are not yet economical to extract at current prices, could become economical
in the future.

j_covert_301.indd 119 1/20/16 6:57 AM

120 Journal of Economic Perspectives

If we cannot rely on market-driven shifts in supply to reduce our consumption
of fossil fuels, can we expect the demand for fossil fuels to fall? In the following
section, we investigate the prospects for low-carbon alternatives to fossil fuels to
become cost-competitive. Hydroelectric, solar, wind, and nuclear are obvious
substitutes for fossil fuels in electric power generation. Reducing our demand
for petroleum will also require low-carbon sources of transportation, potentially
through the large-scale adoption of electric vehicles. We analyze trends in produc-
tion costs for these cleaner technologies and find it implausible that these trends
alone will sufficiently reduce fossil fuel consumption for the world as a whole.

Thus, we are driven to the conclusion that activist and aggressive policy choices
are necessary to drive reductions in the consumption of fossil fuels and greenhouse
gas emissions. We end the paper with a peek into the future of risks if we were to
continue our heavy consumption of known deposits of fossil fuels without capturing
and sequestering the emissions. The picture is alarming.

Supply: Peak Oil, Natural Gas, and Coal?

In the aftermath of the oil price shocks of 1973–74, a number of market
observers began to warn of ever-dwindling oil resources. A robust debate ensued
with geologists and environmentalists often citing Hubbert’s (1956) theory of “peak
oil” as the basis for the concern. President Jimmy Carter (1977) reflected these

Figure 1
Ratio of Proven Fossil Fuel Reserves to Production

Source: BP Statistical Review of World Energy, 2015.

0

50

100

150

200

250

1980 1985 1990 1995 2000 2005 2010

P
ro

ve
n

r
es

er
ve

s
to

p
ro

d
u

ct
io

n
r

at
io

(
in

y
ea

rs
)

Oil

Gas

Coal

j_covert_301.indd 120 1/20/16 6:57 AM

Will We Ever Stop Using Fossil Fuels? 121

concerns when he told the nation in a televised speech: “World consumption of oil
is still going up. If it were possible to keep it rising during the 1970’s and 1980’s by
5 percent a year, as it has in the past, we could use up all the proven reserves of oil
in the entire world by the end of the next decade.”

Economists, on the other hand, stressed the ongoing tension between consump-
tion and technological progress. Yes, taken literally, there is a finite amount of any
one fossil fuel. Each barrel of oil or cubic foot of natural gas that is taken out of the
ground cannot be replaced in a relevant time scale. However, over time we are able to
extract more fossil fuels out of the ground as technology improves, and the ultimate
resources of planet Earth are both highly uncertain and very large. This observation
is perhaps most closely connected with the work of Morris Adelman (1993).

Indeed, two enormous sources of modern oil production—oil from tar sands
and oil (as well as gas) from shale deposits—only recently were categorized as
“reserves,” having previously not been on the radar as economically relevant energy
sources. Though these two “unconventional” sources of hydrocarbons currently
represent a substantial fraction of total production in the US and Canada and
approximately 10 percent of world oil and gas reserves, their economically useful
existence was driven entirely by recent technological advances.

Canadian geologists first studied the possibility of production of oil from
bituminous sands (also referred to as tar sands or oil sands) in the 19th century
(Atkins and MacFadyen 2008). Though the commercial potential for this viscous
mixture of heavy oil, sand, and clay had long been recognized, it took scientists
several decades to figure out how to economically mine the mixture, separate out
the heavy hydrocarbons and “upgrade” them to light, sweet crude oil. It wasn’t
until 1967 that a small-scale commercial production and upgrading facility began
operation (10,000 barrels, hereafter “bbls,” per day) and it took until 1999 for
Canadian energy authorities to recognize the growing number of tar sands proj-
ects as reserves. This decision increased total Canadian oil reserves by 130 billion
bbls and total world reserves by about 10 percent. Now, the Canadian Association
of Petroleum Producers estimates tar sands production in 2014 was more than
2 million bbls/day.4

A similar pattern occurred in the development of oil and gas from shale and other
low-permeability rock formations in the United States. These resources had long been
known to contain tremendous quantities of hydrocarbons (Shellberger, Nordhaus,
Trembath, and Jenkins 2012). However, their low permeability inhibited the rate at
which oil and gas molecules could flow out of conventionally designed vertical wells
drilled into them. In the 1980s, engineers working in the Barnett natural gas shale
formation in Texas began experimenting with hydraulic fracturing and horizontal
drilling as tools to solve the permeability problem. By the early 2000s, thousands of
wells had been successfully drilled and “fracked” into the Barnett, and the technique

4 This figure includes synthetic “upgraded” crude oil as well as raw bitumen production, according to
data from the Canadian Association of Petroleum Producers at http://statshbnew.capp.ca/SHB/Sheet.
asp?SectionID=3&SheetID=233.

j_covert_301.indd 121 1/20/16 6:57 AM

122 Journal of Economic Perspectives

later spread to the Marcellus natural gas shale formation in Pennsylvania and the
Bakken oil shale formation in North Dakota. As a result, US oil and gas reserves
expanded 59 and 94 percent, respectively, between 2000 and 2014. This technology-
driven growth in reserves also caused increases in production. In 2014, US natural gas
production was greater than ever before and oil production was at 97 percent of the
peak reached in 1970.

These two technological advances are at least partially responsible for a more
general long-term pattern of worldwide reserve growth. Figure 2 plots annual
reserve estimates in absolute terms for oil, natural gas, and coal. The steady rise of
proven oil and natural gas reserves is striking. The average growth rate in reserves is
2.7 percent for both oil and natural gas. In only one year did total proven oil reserves
fall, and this year was immediately followed by a growth in reserves of 12.2 percent
the following year. Natural gas reserves fell in only two years, but by less than half of
1 percent in both cases. Coal reserves, on the other hand, fell consistently through
the late 1990s to 2008 but have since shown fairly constant positive growth.

A potential concern about the consistency of this pattern is that individual
countries do not consistently scale back reserve estimates when low prices cause
existing discoveries to be unprofitable. While there is no uniform standard that all
countries use in calculating reserves, there are at least two reasons to believe that
these data are informative about the scale of fossil fuel resources readily available
in the future. First, securities regulators in developed countries heavily regulate

Figure 2
Proven Reserves of Oil, Natural Gas, and Coal over Time

Source: BP Statistical Review of World Energy, 2015.
Note: The figure plots annual reserve estimates in absolute terms for oil, natural gas, and coal.

0

5,000

10,000

15,000

20,000

25,000

30,000

1980 1985 1990 1995 2000 2005 2010
P
ro
ve
n
r
es
er
ve

s
(Q

u
ad

ri
ll

io
n

B
T

U
)

Oil
Gas
Coal

j_covert_301.indd 122 1/20/16 6:57 AM

Thomas Covert, Michael Greenstone, and Christopher R. Knittel 123

reserve estimates published by publicly traded oil and gas companies, and these
numbers are regularly revised downward following oil price crashes. For example,
in 1986 and 1998, when oil prices fell 46 and 30 percent, respectively, US oil reserves
fell by 3 and 6 percent, respectively. Second, even in the absence of truth-telling
regulation in reserve estimates (for example, in the case of oil owned by national
oil companies), the country-level data in the BP Statistical Review still shows a mean-
ingful number of downward revisions in a typical year. In the history of oil reserve
changes in the BP Statistical Review, a full quarter of countries report decreases in
reserve estimates in the average year. However, the converse of this statement is
what is important for understanding how hydrocarbon reserves may change in the
future. The BP data show that most countries seem to find significantly more oil
(and gas) than they consume in most years.

If the past 35 years is any guide, not only should we not expect to run out of
fossil fuels any time soon, we should not expect to have less fossil fuels in the future
than we do now. In short, the world is likely to be awash in fossil fuels for decades
and perhaps even centuries to come.

An alternative measure of technological progress in the exploration and exploi-
tation of fossil fuel resources is the success rate associated with exploring for new oil
and natural gas formations. Without technological progress in this area, present-value
considerations will cause firms to drill lower-risk prospects before higher-risk pros-
pects. As a result, the probability of successful exploration falls over time as the “easy”
wells are exhausted and oil and gas prices rise to equilibrate supply and demand.
In contrast, if technology advances and lowers exploration costs and/or risks, it is
possible for the probability of successful exploration to stay constant, or even rise over
time, independent of the path of prices.

We are unaware of any systematic data on the history of exploration and devel-
opment costs for oil or gas, given that it is difficult to observe private company costs
for seismic studies, drilling, and other inputs to the exploration process. However,
the US Energy Information Administration and the data and consulting firm IHS
do publish data on the number of successful and failed exploration and develop-
ment wells in the United States, so it is possible to measure changes in risk over
time. Figure 3 plots the fractions of successful exploration and development wells
in each year from 1949 to 2014. The probability of a successful exploratory well did
drift downward from about 20 percent in 1949 to 16 percent in the late 1960s. But
in 1968, the highly successful Alaska North Slope field was discovered, leading to a
near doubling in the probability of exploratory drilling success by 1979.

Similar technological events preceded other periods of growing drilling
success. These include the development of ultra-deep water fields in the Gulf of
Mexico during the 1980s, hydraulic fracturing technology for natural gas forma-
tions in the 1990s, and the same technology for oil formations in the 2000s. By 2007,
69 percent of exploratory wells yielded successful oil or gas production. Though the
probability of successful exploration has drifted down to about 50 percent in recent
years, it is still markedly higher than in much of the history of US fossil fuel explora-
tion. Figure 3 also shows that the fraction of successful development wells has also

j_covert_301.indd 123 1/20/16 6:57 AM

124 Journal of Economic Perspectives

grown over time. Though these wells are drilled into formations already known to
contain oil or gas, there is still risk that a development well faces technical difficul-
ties and produces no output. Growth in this number is also important, since there
have been 10 to 20 times as many development wells compared to exploratory wells
in recent years. In the United States, at least, it appears that technical progress has
consistently helped increase the supply of fossil fuels, in spite of price volatility and
the exhaustion of existing fossil fuel formations.

The supply curve for fossil fuels has constantly shifted out over large stretches
of time, both because of new discoveries (like the large-scale development of new oil
sources or oil and gas from Alaska and from the North Sea in the 1970s) as well as
from new techniques like deep-water drilling, hydraulic fracturing, and extracting
oil from tar sands. What might be next on the horizon?

Besides measuring fossil fuel reserves, geologists also measure fossil fuel
“resources”—that is, fossil fuel deposits that are known to exist but are not currently
economical to extract. McGlade and Ekins (2015) summarize reported resources
by the Federal Institute for Geosciences and Natural Resources, the International
Energy Agency, and the Global Energy Assessment. The range of oil resources
is from 4.2 to 6.0  trillion barrels, which is 2.8 to 4 times larger than the existing

Figure 3
Fraction of Development and Exploratory Wells that are Successful in the United
States

Source: US Energy Information Administration and IHS.
Notes: Development wells are drilled into formations that have already been explored and thus are known
to contain oil or gas. Exploratory wells are drilled into formations that have not yet been explored and
which might not contain oil or gas.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1949 1954 1959 1964 1969 1974 1979 1984 1989 1994 1999 2004 2009

2014

Exploratory wells

Development wells

j_covert_301.indd 124 1/20/16 6:57 AM

Will We Ever Stop Using Fossil Fuels? 125

reserves of roughly 1.5 trillion barrels. The range of natural gas resources is also
immense, ranging from 28,000 trillion cubic feet to over 410,000 trillion cubic feet.
For comparison, current global reserves of natural gas are roughly 7,000 trillion
cubic feet (according to the US Energy Information Administration’s International
Energy Statistics).5 Finally, the estimate range for coal resources is from 14,000 to
23,500 gigatons, compared to existing reserves of around 1,000 gigatons.

These existing “resources” for currently exploited fossil fuel technologies could
potentially, if technological progress continues to advance, allow the supply curves
for fossil fuels to continue to shift outward for quite some time. In addition, two
notable additional resources have not yet been commercially developed, but are
known to exist in large quantities: oil shale and methane hydrates.

“Oil shale” is defined by the US Geological Survey as “fine-grained sedimen-
tary rock containing organic matter that yields substantial amounts of oil and
combustible gas upon destructive distillation” (Dyni 2006). Despite the similarity
in nomenclature, “oil shale” is fundamentally different from the earlier-mentioned
“shale oil” which is currently being extracted using hydraulic fracturing techniques
in North Dakota, Texas, and Oklahoma. The production technology for oil shale is
similar to that of oil sands. In both cases, a heavy hydrocarbon (bitumen in the case
of oil sands and kerogen in the case of oil shale) exists in sand, clay, or sedimen-
tary rock. Heat is used to separate the hydrocarbon from the surrounding material.
The resulting bitumen or kerogen then goes through additional refining steps to
become a final consumer product. Oil shale resources are enormous. A US Geolog-
ical Survey report from 2006 estimates that 2.8 trillion barrels of oil shale exist (that
is, almost 50 percent of the high-end estimate of oil resources), and some private
estimates are much larger. If oil shale became economical in the near future, it
could cause oil reserves to nearly triple. Because of the use of heat in the extraction
phase and the extra steps required for refining, carbon emissions for producing
oil from oil shale are greater than for conventional sources of oil. Some estimates
suggest that the emissions are 21–47 percent greater per unit of energy produced
compared to conventional sources (Brandt 2008).

Methane hydrates are a solid mixture of natural gas and water that forms
in low-temperature and high-pressure environments, usually beneath seafloors.
Geologists have recognized methane hydrates as a potential source of hydrocar-
bons since the 1960s, and testing on whether the resource might be commercially
viable dates back to the 1970s. The technology to extract methane hydrates at a
commercial price does not currently exist, although a number of countries are
actively pursuing technology in this area (Boswell et al. 2014). If this technology
eventually achieves commercial success, the potential scale of methane hydrate
resources ranges from 10,000 trillion cubic feet to more than 100,000 trillion
cubic feet, according to US Geological Survey estimates (US Energy Information

5 The page of the International Energy Statistics on which this data is found is: http://www.eia.gov/
cfapps/ipdbproject/iedindex3.cfm?tid=3&pid=3&aid=6&cid=ww,&syid=2011&eyid=2015&unit=TCF.

j_covert_301.indd 125 1/20/16 6:57 AM

126 Journal of Economic Perspectives

Administration 2012). For comparison, current global reserves of natural gas are
roughly 7,000 trillion cubic feet.

The status of oil shale and methane hydrates today is similar to that of oil sands
and shale gas in the 1980s. Geologists knew of their existence, but oil and gas compa-
nies did not yet know how to recover them in a cost-effective manner. The remarkably
successful history of innovation in oil and gas exploration makes it seem more than
possible that oil shale and methane hydrates will become commercially developed.

Even without the emergence of new technology for oil shale and methane
hydrates, the use of existing technology to develop shale gas and shale oil resources
outside of the United States is just now starting. The US Energy Information Admin-
istration (2013) estimates that 93 percent of shale oil and 90 percent of shale gas
resources exist outside of the United States. Further, these resources represent 1

0

and 32 percent, respectively, of total world resources for oil and gas. As shale tech-
nology spreads across the world, these resources are likely to become economically
productive reserves.

The policy implications of this ongoing expansion of fossil fuel resources are
potentially profound. Even if countries were to enact policies that raised the cost of
fossil fuels, like a carbon tax or a cap-and-trade system for carbon emissions, history
suggests that technology will work in the opposite direction by reducing the costs of
extracting fossil fuels and shifting their supply curves out.

Demand: Will Low-Carbon Energy Sources Knock Fossil Fuels Out
of the Money?

Absent large upward shifts of the supply curve for fossil fuels, deep cuts in
fossil fuel consumption will have to come from inward shifts in their respective
demand curves.6 In this section, we analyze the recent changes in the relative prices
of carbon-free and fossil-fuel–based energy technologies and characterize what
future relative prices are necessary to reduce demand for fossil fuels in a substantial
manner. We focus on the electricity and transportation sectors, which are major
users of fossil fuels. In the United States, for example, the electricity sector consumes
over 90 percent of total coal consumption and 30 percent of natural gas, while the
transportation sector consumes over 60 percent of total US oil consumption.7

6 We admittedly are confounding shifts in the supply curve with changes in the slope. That is, a shift
would represent the addition of zero-marginal-cost supplies. In practice, the marginal cost of many of
the new resources is nonzero, implying that the supply curve shifts, or becomes flatter, at some strictly
positive price.
7 US statistics on the end-uses of fossil fuels are readily available at the website of the US Energy Informa-
tion Administration. For example, coal statistics are available at from the Quarterly Coal Report at http://
www.eia.gov/coal/production/quarterly; natural gas statistics from the “Natural Gas Consumption by
End-Use” page at http://www.eia.gov/dnav/ng/ng_cons_sum_dcu_nus_a.htm; and oil statistics at the
“Petroleum and Other Liquids” webpage at http://www.eia.gov/dnav/pet/pet_cons_psup_dc_nus_
mbbl_a.htm.

j_covert_301.indd 126 1/20/16 6:57 AM

Thomas Covert, Michael Greenstone, and Christopher R. Knittel 127

Replacing Natural Gas and Coal in Electricity Generation
Solar photovoltaics, wind turbines, and nuclear fission power plants are the

current leading candidates to replace coal and natural gas in electricity genera-
tion. However, nuclear fission has already been commercially exploited for almost
75 years. In spite of this maturity, the rate of new nuclear power plant construction
has significantly slowed down since the 1980s, and its share of power generation
in most countries has actually fallen over the last decade due to decreasing cost
competitiveness (Deutch et al. 2009; in this journal, Davis 2012; see also US EIA
International Energy Statistics). Thus, we focus here on the recent history of cost
improvements in solar photovoltaics and wind turbines.

We compare costs using levelized cost of energy estimates across different technol-
ogies. The levelized cost of energy is the present discounted value of costs associated
with an energy technology divided by the present discounted value of produc-
tion—that is, it is a measure of the long-run average cost of the energy source.8
This measure offers a way of adjusting for the fact that renewable energy sources
and fossil fuel plants have a different profile of costs over time. For both renewable
plants and fossil fuel plants, the single largest cost occurs in the first year a plant
is built, representing the up-front capital cost of the plant. However, for renew-
able technologies, the operating costs in the remaining years are small, as there
are no fuel inputs necessary, only maintenance. In contrast, fossil fuel technologies
like natural gas and coal require ongoing fuel costs. Thus, a comparison between
the levelized cost of energy for a renewable energy technology and a fossil fuel
technology hinges on the difference in up-front capital costs between the two tech-
nologies relative to expected fuel costs for the fossil fuel technology.

We report forecasts of the levelized cost of electricity generation published
by the US Energy Information Administration during the last two decades. For
each year in our data, these forecasts report a projected levelized cost of energy
for coal, natural gas, nuclear, and wind for electricity generation plants to be built
5–10 years in the future. The forecast for solar is on a similar timeframe, but for
a specific year (usually 4–7 years in the future).9 We prefer this set of estimates of
levelized costs of energy to others that are available for a number of reasons: they
are produced annually; the methodology is clearly explained and documented; a
wide range of electricity-generating technologies is compared; and the time series
is relatively long.

8 Specifically, the levelized cost of energy cost of energy is calculated as:

LCOE = ∑ t=0
T δ t C t ________

∑ t=0 T δ t q t
,

where δ is the discount factor, T is the lifetime of a generating technology, and C and q are the technol-
ogy’s per-period costs and production, respectively.
9 For 2007, the projection is for 2014; for 2008 through 2011, it is for 2016; for 2012 it is for 2017; 2013
for 2018; and, 2014 for 2019. Therefore, the average across all of these years is a six-year forecast, while
the last four years report a five-year forecast.

j_covert_301.indd 127 1/20/16 6:57 AM

128 Journal of Economic Perspectives

One could certainly argue that a longer-term cost outlook would be more
appropriate. We focus on estimates 5–10 years out for at least two reasons. First,
many analysts believe that it is important to reduce CO2 emissions in the next decade
to mitigate the odds of disruptive climate change. Second, while longer-term costs
estimates might exist (although we are unaware of a reasonably long series), they
are believed to be very imprecise.

Figure 4 plots the forecasts from the US Energy Information Administration
for the last 18 years. Perhaps the most striking finding shown in the figure is the
dramatic fall in the cost of solar energy during the past five years. The 2009 forecast
for the near-term levelized cost of solar power was nearly $450/MWh of electricity
generated, while the 2014 number is under $150/MWh. The speed of these reduc-
tions appears to have subsided, but the downward trend continues.

The decline in the costs of solar energy is indeed rapid, but it seems plau-
sible. For example, the current cost of solar energy can be inferred from auctions
of photovoltaic installations—in which the prices imply long-run average cost per
megawatt-hour of electricity generated. For example, in November 2014, Dubai’s
state utility held an auction for 100 MWs of photovoltaic power over a 25-year period.
The lowest bid was $59.8/MWh. More generally, Bollinger and Seel (2015) docu-
ment that utilities in the southwestern parts of the United States are now routinely
acquiring power from new solar photovoltaic projects at prices in the range of
$40–$50/MWh. Though these new projects receive a federal investment tax credit

Figure 4
Levelized Cost of Energy (LCOE) Forecasts from the US Energy Information
Administration

Source: EIA Annual Energy Outlook reports from 1997 to 2014.

0
50
100
150
200
250

300

350

400

450

500

1997
1998

1999
2000

2001
2002

2003
2004

2005
2006

2007
2008

2009
2010

2011
2012

2013
2014

L
C

O
E

(
$/

M
W

h
)

Forecast year

Coal: 5–10 year forecast
Nuclear: 5–10 year forecast
Wind: 5–10 year forecast
Gas: 5–10 year forecast
Solar: Short term forecast

j_covert_301.indd 128 1/20/16 6:57 AM

Will We Ever Stop Using Fossil Fuels? 129

equal to 30 percent of construction cost, their implied “real” costs are still near or
below the levelized cost of energy estimates for natural gas generation in Figure 4.

However, these examples of highly competitive prices for solar energy are from
specific locations that are exceptionally well-suited for generating solar energy. In
contrast, Figure 4 reports the average forecast across locations, including locations
where solar exposure may be limited or wind speeds may be slow. Thus, the average
levelized cost of solar—even given that more favorable sites are being chosen for
solar—is still twice that of natural gas in the United States. Wind power seems close
to cost competitive with fossil fuel generation in many locations. As with solar, it is
complicated to calculate the exact underlying costs of wind power, due to the avail-
ability of investment and production tax credits and support from state renewable
portfolio standards.

The levelized cost for an electricity technology is only one dimension to be
taken into account in making comparisons. Three additional challenges exist. First,
both the intensity of sunlight and the speed of the wind vary tremendously across
space,10 meaning that the same solar panel or wind turbine installed in one loca-
tion will generate vastly different amounts of electricity than if it were installed in
another location. This implies that the long-run marginal cost of solar and wind will
be upward sloping and large-scale deployment in some areas is likely to be infeasible.

Second, solar and wind energy are inevitably intermittent, which requires either
increases in backup generation (often supplied by natural gas generators) or increases
in energy storage that aren’t typically reflected in the numerator of the basic levelized
cost of energy calculation. These costs will depend upon a variety of factors, such
as the level of penetration, the degree of variation in generation from the renew-
able resources, and the correlation in generation across renewable resources. While
more research on magnitude of these costs is needed, some estimates exist; these
estimates are likely to be very site-specific since they depend on the variability in solar
and wind availability. The bulk of this research simulates electricity systems under
varying penetration levels of renewables assuming a specific location. Mills and Wiser
(2010) simulate the costs of operating an electricity system under different penetra-
tion of renewables using detailed simulated output from Midwestern solar and wind
installations. They find that for a 10 percent penetration of solar photovoltaic power,
intermittency can add as much as $39 per MWh, when the solar is installed in one
location, to as little as $3 per MWh when it is installed across 25 different sites. The
intermittency costs of a 10 percent penetration of wind across 25 sites are simulated
to be below $2 per MWh. In contrast, Wolak (2015) suggests that the costs of inter-
mittency in California are likely to be high because the benefits from diversifying site
locations are small.

Finally, because the generation from solar (and wind) resources in a given
area tend to be positively correlated, large-scale penetration of either resource will

10 It is also possible for fossil fuel technologies to have different localized cost of energy in different
places as a result of fuel transportation constraints. For example, natural gas prices and coal prices vary
significantly across the United States (as shown by US Energy Information Administration data).

j_covert_301.indd 129 1/20/16 6:57 AM

130 Journal of Economic Perspectives

inevitably reduce the value of incremental capacity additions. The impact of this on
net demand for electricity (after netting out the supply of solar resources) within
California has generated what is being referred to as the California Independent
System Operator’s (CAISO) “Duck Curve,” represented in Figure 5. The CAISO has
forecast demand for electricity, net of renewable generation, in each year through
2020, when the required amount of generation from renewables hits 30 percent.
Figure 5 shows these forecasts for March 31. The figure illustrates that as progres-
sively more and more renewables hit the market, net demand will be lowest during
daytime hours and prices during those hours will obviously fall, making additional
investments in renewables less valuable.

Intermittency and the large reductions in net demand during peak generation
periods imply that, absent economical storage technologies, solar and wind power
are ill-suited for baseload generation which is currently covered by coal, natural gas,
nuclear, and hydroelectric power.

The levelized costs for fossil fuel technologies presented above also ignore
externalities. How much would pricing the externalities associated with carbon

Figure 5
California ISO’s “Duck Curve”
(Net load–March 31)

Source: Figure 2 from “Fast Facts” published by California ISO, available at: https://www.caiso.com/
Documents/FlexibleResourcesHelpRenewables_FastFacts .
Notes: This graph shows net electricity demand (load), across the hours of the day on March 31, as California
approaches penetration of 30 percent renewables by 2020. Because the renewables will predominately be
solar, and solar generation peaks during the day, net demand will continue to fall during the day.

28,000

26,000

24,000

22,000

20,000

18,000

16,000

14,000

12,000

10,000
0

M
eg

aw
at

ts

12am 6am 9am 12pm 3pm 6pm 9pm

2014

2017

2 0 1 3 (actual)

Hour

2020

2016
2015

2012
(actual)

2019
2018

ramp need
~13,000 MW
in three hours

overgeneration
risk

3am

j_covert_301.indd 130 1/20/16 6:57 AM

Thomas Covert, Michael Greenstone, and Christopher R. Knittel 131

dioxide emissions change these conclusions? Based on Greenstone, Kopits, and
Wolverton (2013), the US government applies a social cost of carbon of $43 per
metric ton of carbon dioxide in 2015 dollars. Using this value, the carbon dioxide
externality for the typical natural gas plant is $20 per MWh, while the externality for
the typical coal plant is $40 per MWh. But the US Energy Information Administra-
tion forecasts that the gap between average levelized costs for solar and natural gas
will still be about $50 per MWh in the near-term.

Replacing Oil Usage in Motor Vehicles
While there are many substitutes for fossil fuels in electricity generation, the pri-

mary path for moving away from them in the transportation sector is the use of battery
powered electric vehicles, which in turn requires several technology breakthroughs to
occur. First, even if oil prices were at $100 per barrel, the price of batteries that store
the energy necessary to power these vehicles needs to decrease by a factor of three.
Second, the time needed for these batteries to charge must be shortened. Third, the
electricity that is fueling these cars will need to have sufficiently lower carbon content
than petroleum. Otherwise, we could transition from oil-based transportation with
moderately high carbon emissions to coal-fired-electricity-based transportation
with even higher carbon emissions. As noted in Graff Zivin, Kotchen, and Mansur
(2014), effective carbon emissions from electric vehicles that are powered by the exist-
ing US power plant fleet are generally higher than emissions from high-efficiency
gasoline-powered vehicles. Only 12 percent of fossil-fueled power plants have low
enough carbon emissions that electric vehicles powered by them would have lower
emissions than a Toyota Prius.

The large-scale adoption of electric vehicles, instead of petroleum-based
internal combustion engine vehicles, seems likely to require all three of these
events. To date, none of them have happened. The previous subsection discussed
the prospects for greening the electrical grid. Refueling times, as well as the absence
of an abundant refueling infrastructure, remain challenges. First we will explore the
necessary innovation in batteries.

What is the magnitude of the necessary improvement in storage technologies?
Here, we describe some back-of-the-envelope calculations that provide a sense of the
present discounted value of operating an internal combustion engine, compared
with the present discounted cost of operating an electric vehicle. The bottom line
from these calculations is that truly dramatic improvements in battery technology
are necessary to bring these technologies into cost parity.

We want to compare the operating costs of an electric vehicle to an internal
combustion engine. We assume a 3,000-pound vehicle. (For comparison, the
2015 Honda Accord has four doors and a four-cylinder engine, weighs about
3,200  pounds, and gets a combined-fuel economy of 33 miles per gallon.11) We

11 The official fuel economy rating is 31 miles per gallon, while the user-reported fuel economy is
33.8 miles per gallon. See the US Department of Energy website at https://www.fueleconomy.gov/feg/
PowerSearch.do?action=noform&path=7&year=2015&make=Honda&model=Accord&srchtyp=ymm.

j_covert_301.indd 131 1/20/16 6:57 AM

132 Journal of Economic Perspectives

assume both the electric car and the internal combustion engine car are driven
15,000 miles per year, and we use a discount rate of 5 percent.

For the internal combustion engine, we assume that it presently gets
30  miles per gallon. For the electric vehicle, we consider a battery size for a
range of 250  miles, which of course is still shorter than driving range of most
current internal combustion vehicles. We use a price for purchasing electricity of
12.2 cents per kWh, which is consistent with the average US retail price in 2014.
The electric vehicle is assumed to consume 0.3 kilowatt-hours electricity per mile.
Finally, because internal combustion engines tend to be more costly than electric
motors, we “credit” electric vehicles by $1,000 (Peterson and Michalek 2013).

Current battery costs for an electric vehicle are roughly $325 per kWh. This esti-
mate is consistent with the cost of Tesla’s Powerwall home battery, which retails for a
price of $350 per kWh for the 10 kWh model (and does not include the price of an
inverter for use in the home). This cost estimate may be lower than an average battery
cost. For example, Tesla charges $3,000 for an extra 5 kWhs of battery capacity in
its Model S, which is $600 per extra kWh; however, this incremental price may also
include some level of price discrimination on the part of Tesla. In a 2014 “EV Every-
where Grand Challenge” study, the US Department of Energy finds that the current
battery cost is $325 per kWh.12 At a battery cost of $325 per kWh, the price of oil would
need to exceed $350 per barrel before the electric vehicle was cheaper to operate.13
During 2015, the average price of oil was approximately $49 per barrel. At present,
the costs of batteries make large-scale penetration of electric vehicles unlikely.

How is this comparison between the operating costs of an internal combus-
tion engine and an electric vehicle likely to evolve into the future? In these kinds
of comparisons, a common mistake is to compare future costs of an electric vehicle
with current costs of an internal combustion engine. But this is the wrong test.
Future electric vehicles will be competing against future combustion engine vehi-
cles, not current ones. There will be technological progress in all areas, not only for
low-carbon technologies.

The historical record suggests that we should expect persistent innovation
in the efficiency of combustion engine vehicles, just as it suggests that there will
be continued innovation in the extraction of fossil fuels. For the internal combus-
tion engine, we assume that the fuel economy grows at 2 percent per year, which is
consistent with Knittel (2011).

As a basis for estimating the future price trajectory of batteries, Nykvist and
Nilsson (2015) survey 85 peer-reviewed estimates of current and future battery costs
published since 2008. They also report battery cost estimates for the Nissan Leaf,

12 A presentation connected to the study gives a range of $325–500 per kWh. http://energy.gov/sites/
prod/files/2014/03/f8/5_howell_b .
13 To connect the price of gasoline to the price of oil, we regress the log of historic gasoline prices on
the log of oil prices to capture the nonlinear relationship between gasoline and oil prices. We use data
from the US Energy Information Administration from April 1993 to July 2015. The estimated intercept is
−1.57; the slope is 0.6046. The R 2 is 0.98. The general conclusions are unchanged if we instead assume
a linear relationship between gasoline and oil prices.

j_covert_301.indd 132 1/20/16 6:57 AM

Will We Ever Stop Using Fossil Fuels? 133

Tesla S, and other electric vehicles. They find large reductions in battery costs over
the past 10 years. However, their data also suggest battery costs are predicted to level
off between $150 to $300 per kWh over the next 15 years. The same US Department
of Energy (2014) “EV Everywhere” presentation discussed above defines a “target”
battery cost of $125 per kWh by 2022.

We can extend the calculation above for a variety of oil-price–battery-cost pairs
by calculating the indifference price of oil for a given battery cost. Figure 6 plots this
relationship across battery costs of 0 to $400 per kWh. For this graph, we compare
the technologies in 2020 allowing for the efficiency of internal combustion engine
vehicles to increase by 2 percent per year. Even at the US Department of Energy
target price for 2020, oil prices would have to rise to $115 per barrel for electric

Figure 6
Break-even Oil and Battery Costs

Notes: This graph plots the relationship between oil prices and battery costs such that the present dis-
counted value of owning and operating an internal combustion engine vehicle equals the present
discounted value of owning and operating an electric vehicle assuming the year is 2020. We assume the
vehicle is driven 15,000 miles per year. We use the average retail price of electricity of 12.2 cents per
kWh to charge the electric vehicle. We estimate the relationship between oil prices and gasoline prices
using monthly data from 1993 to 2015 and assume a log–log relationship. We size the battery such that
the electric vehicle has a range of 250 miles, assuming electricity consumption of 0.3 kWh per mile. The
cost of the electric vehicle is reduced by $1,000 to reflect the lower costs associated with electric motors
relative to the internal combustion engine vehicle. The fuel economy of internal combustion engines is
assumed to grow at an annual rate of 2 percent, consistent with Knittel (2011).

0

100

200

300

400

500

600

700

800

0 50 100 150 200 250 300 350 400 450

P
ri

ce
o

f
o

il
i

n
2

02
0

($
/

b
b

l)

Cost of batteries in 2020 ($/KWh)

Equal costs, Electric and Internal Combustion Engine Vehicles (EVs and ICEs)

DOE target cost of batteries
in 2020: $125

Break-even price of oil: $115

December 2020 WTI
oil futures price: $55

Break-even cost of
batteries: $64

EVs are cheaper than ICEs

ICEs are cheaper than EVs

DOE estimated current price of
batteries: $325

Break-even price of oil in 2020:
$420

Price of 10-kWh Tesla
Powerwall: $350

Break-even price of
oil in 2020: $470

0
100
200
300
400
500
600
700
800
0 50 100 150 200 250 300 350 400 450
P
ri
ce
o
f
o
il
i
n
2
02
0
($
/
b
b
l)

Cost of batteries in 2020 ($/KWh)
Equal costs, Electric and Internal Combustion Engine Vehicles (EVs and ICEs)
DOE target cost of batteries
in 2020: $125
Break-even price of oil: $115
December 2020 WTI
oil futures price: $55
Break-even cost of
batteries: $64
EVs are cheaper than ICEs
ICEs are cheaper than EVs
DOE estimated current price of
batteries: $325

Break-even price of oil in 2020:
$420
Price of 10-kWh Tesla
Powerwall: $350

Break-even price of
oil in 2020: $470

j_covert_301.indd 133 1/20/16 6:57 AM

134 Journal of Economic Perspectives

vehicles to be cost-competitive with internal combustion engines under the assump-
tions discussed above. If battery costs remain at $325 per kWh, oil prices would have
to exceed $420 per barrel. For comparison, the current December 2020 oil futures
price using the West Texas Intermediate benchmark (observed on December 18,
2015) was $55/barrel, requiring a battery cost that would fall to $64 per kWh.

These basic calculations make it clear that at least for the next decade or two,
electric vehicles face an uphill battle. Not only are large continuing decreases in
the price of batteries necessary, but oil prices would have to increase by more than
financial markets currently predict.

Other barriers to widespread electric vehicle adoption are not reflected in these
calculations. First, the estimates are built on a battery with a range of 250  miles,
which for some drivers would not be enough. Second, we assume that there is no
disutility associated with longer recharging times of electric vehicles relative to
fueling times for internal combustion engines. Finally, oil prices are endogenous, so
substantial penetration of electric vehicles would reduce demand for oil. Provided
that the supply curve for oil is upward sloping (as it is in almost all markets), this
drop in demand would translate into lower oil prices, making gasoline vehicles
more attractive.

How does this comparison change if we include the social cost of carbon dioxide
emissions? Assume that 20 pounds of carbon dioxide is emitted into the atmosphere
for every gallon of gasoline burned. (Most of the weight of the carbon dioxide arises
when the car emissions of carbon atoms in the gasoline combine with oxygen in the air,
which is why a gallon of gasoline weighing about 6 pounds can produce 20 pounds of
carbon dioxide.) As we noted above, a battery price of $125 kWh implies $115/barrel
of oil as a break-even price for an internal combustion engine. If we account for the
social cost of carbon and make the extreme assumption that electricity for the electric
vehicle is carbon free, the break-even price for oil falls to $90 per barrel.

Our emphasis in this section has been on the scope for battery technologies to
unseat oil in transportation as well as solar and wind resources as noncarbon methods
of generating electricity. There are other noncarbon alternatives for producing
energy. In some parts of the world, including Africa and South America, some propor-
tion of the rising demand for electricity might be met by hydroelectric power. Certain
regions may have possibilities for electricity generated by geothermal energy or ocean
thermal gradients. We mentioned earlier that we were setting aside any discussion
of nuclear power in this paper, given the high and rapidly rising costs of nuclear
power generation in recent decades. We are generally supportive of research and
development into all of these other noncarbon methods of generating electricity.
However, the International Energy Administration Agency (2015) projects that fossil
fuels will account for 79 percent of total energy supply in 2040 under the current,
business-as-usual policies, which already takes into account some rise in these alterna-
tive noncarbon energy production technologies. In the medium-run of the next few
decades, none of these alternatives seem to have the potential based on their produc-
tion costs (that is, without government policies to raise the costs of carbon emissions)
to reduce the use of fossil fuels dramatically below these projections.

j_covert_301.indd 134 1/20/16 6:57 AM

Thomas Covert, Michael Greenstone, and Christopher R. Knittel 135

Discussion

Our conclusion is that in the absence of substantial greenhouse gas policies,
the US and the global economy are unlikely to stop relying on fossil fuels as the
primary source of energy. The physical supply of fossil fuels is highly unlikely to run
out, especially if future technological change makes major new sources like oil shale
and methane hydrates commercially viable. Alternative sources of clean energy like
solar and wind power, which can be used both to generate electricity and to fuel
electric vehicles, have seen substantial progress in reducing costs, but at least in the
short- and middle-term, they are unlikely to play a major role in base-load electrical
capacity or in replacing petroleum-fueled internal combustion engines. Thus, the
current, business-as-usual combination of markets and policies doesn’t seem likely
to diminish greenhouse gases on their own.

What are the consequences of a continued reliance on fossil fuels? We
conducted some back-of-the-envelope calculations of the potential warming asso-
ciated with using all available fossil fuels. This requires estimates of total reserves
and resources of each fossil fuel, carbon conversion factors, estimates of historical
emissions, and a model to convert carbon dioxide emissions into temperature
changes. It is important to note that this exercise is based on high levels of green-
house gas emissions for many decades beyond 2100. For example, our calculations
are based on total carbon emissions ranging from 12,744 to 17,407 gigatons of CO2.
For comparison, the business-as-usual scenario from IPCC (2013) has cumulative
emissions of 6,180  gigatons of CO2 between 2012 and 2100. The calculations are
described in detail in a web appendix.14 Our headline finding is that the combus-
tion of currently known fossil fuels would increase global average temperatures by
10°F to 15°F, depending on the choice of carbon conversion factors and model.
Such scenarios imply difficult-to-imagine change in the planet and dramatic threats
to human well-being in many parts of the world. Further, these estimates do not
account for advances in fossil fuel extraction techniques that could make other
deposits economically accessible; for example, the use of oil shale and methane
hydrate deposits would add another 1.5°F to 6.2°F of warming. Research on the
economic consequences of such changes in temperatures is an important area that
is rapidly advancing (for example, Deschênes and Greenstone 2011).

What are the prospects for avoiding this dystopian future? At a high level, there
are two market failures—greenhouse gas emissions are not priced adequately, and
basic or appropriable research and development is too often underfunded—
and the corresponding solutions of pricing emissions and subsidizing basic
research and development are easy to identify. However, the politics of imple-
menting such policies are complex, particularly because energy consumption is
projected to grow the most in low- and middle-income countries during the coming

14 This appendix is available with this paper at http://e-jep.org. Alternatively, see the following link
for details on these calculations: https://epic.uchicago.edu/sites/default/files/Calculating%20the%20
Temperature%20Potential%20of%20Fossil%20Fuels .

j_covert_301.indd 135 1/20/16 6:57 AM

136 Journal of Economic Perspectives

decades, and thus the majority of emissions cuts would need to take place in those
countries. Without direct compensation, global emissions cuts will require the
poorer countries to use and pay for more expensive energy sources. The last sev-
eral decades have seen limited global progress in tackling these policy problems.

The Conference of Parties (COP21) climate conference in Paris in December
2015 has set out the broad outlines of what could constitute a dramatic change
in global climate policy. Whether this high-level voluntary agreement leads to the
climate policies necessary to correct the market failures related to greenhouse gases
around the globe will be determined in the coming years and decades. Ultimately,
their enactment would greatly reduce the probability that the world will have to
contend with disruptive climate change. The alternative is to hope that the fickle
finger of fate will point the way to low-carbon energy sources that rapidly become
cheaper than the abundant fossil fuels on their own. But hope is too infrequently a
successful strategy.

References

Adelman, Morris A. 1993. The Economics of Petro-
leum Supply: Papers by M. A. Adelman, 1962–1993.
MIT Press.

Atkins, Frank J., and Alan J. MacFadyen. 2008.
“A Resource Whose Time Has Come? The Alberta
Oil Sands as an Economic Resource.” Energy
Journal 29(special issue): 77–98.

Bolinger, Brian, and Joachim Seel. 2015.
“Utility-Scale Solar 2014: An Empirical Analysis of
Project Cost, Performance, and Pricing Trends in
the United States.” Lawrence Berkeley National
Laboratory report LBNL-1000917.

Boswell, Ray, Koji Yamamoto, Sung-Rock Lee,
Timothy Collett, Pushpendra Kumar, and Scott
Dallimore. 2014. “Methane Hydrates.” Chap. 8 in
Future Energy, 2nd edition. Elsevier.

BP. 2015. BP Statistical Review of World Energy
2015. Available at: http://www.bp.com/en/
global/corporate/energy-economics/statistical
-review-of-world-energy.html.

Brandt, Adam R. 2008. “Converting Oil Shale
to Liquid Fuels: Energy Inputs and Greenhouse
Gas Emissions of the Shell in Situ Conversion
Process.” Environmental Science and Technology
42(19): 7489–95.

Carbon Dioxide Information Analysis Center.
No date. http://cdiac.ornl.gov/ftp/trends/co2
_emis/Preliminary_CO2_emissions_2012.xlsx.

Carter, Jimmy. 1977. “Address to the Nation on
Energy.” April 18. The American Presidency Project.
http://www.presidency.ucsb.edu/ws/index
.php?pid=7369.

Chay, Kenneth Y., and Michael Greenstone.
2003. “The Impact of Air Pollution on Infant
Mortality: Evidence from Geographic Variation
in Pollution Shocks Induced by a Recession.”
Quarterly Journal of Economics 118(3): 1121–67.

Chen, Yuyu, Avraham Ebenstein, Michael
Greenstone, and Hongbin Li. 2013. “Evidence on
the Impact of Sustained Exposure to Air Pollu-
tion on Life Expectancy from China’s Huai River
Policy.” Proceedings of the National Academy of Sciences
110(32): 12936–41.

Davis, Lucas W. 2012. “Prospects for Nuclear
Power.” Journal of Economic Perspectives 26(1): 49–66.

Deschênes, Olivier, and Michael Greenstone.
2011. “Climate Change, Mortality, and Adaptation:
Evidence from Annual Fluctuations in Weather
in the US.” American Economic Journal: Applied
Economics 3(4): 152–85.

j_covert_301.indd 136 1/20/16 6:57 AM

http://pubs.aeaweb.org/action/showLinks?system=10.1257%2Fapp.3.4.152

http://pubs.aeaweb.org/action/showLinks?crossref=10.1021%2Fes800531f

http://pubs.aeaweb.org/action/showLinks?crossref=10.1162%2F00335530360698513

http://pubs.aeaweb.org/action/showLinks?crossref=10.1073%2Fpnas.1300018110

http://pubs.aeaweb.org/action/showLinks?system=10.1257%2Fjep.26.1.49

Will We Ever Stop Using Fossil Fuels? 137

Deutch, John M., Charles W. Forsberg, Andrew
C. Kadak, Mujid S. Kazimi, Ernest J. Moniz, and
John E. Parsons. 2009. Update of the MIT 2003 Future
of Nuclear Power. An Interdisciplinary MIT Study.
http://web.mit.edu/nuclearpower/pdf/nuclear
power-update2009 .

Dyni, John R. 2006. “Geology and Resources of
Some World Oil-Shale Deposits.” USGS Scientific
Investigations Report 2005-5294. http://pubs.usgs.
gov/sir/2005/5294/pdf/sir5294_508 .

Graff Zivin, Joshua S., Matthew J. Kotchen,
and Erin T. Mansur. 2014. “Spatial and Temporal
Heterogeneity of Marginal Emissions: Implications
for Electric Cars and Other Electricity-Shifting Poli-
cies.” Journal of Economic Behavior and Organization
107(Part A): 248–68.

Greenstone, Michael, Elizabeth Kopits, and
Ann Wolverton. 2013. “Developing a Social Cost
of Carbon for US Regulatory Analysis: A Method-
ology and Interpretation.” Review of Environmental
Economics and Policy 7(1): 23–46.

Hubbert, M. King. 1956. “Nuclear Energy and
the Fossil Fuels.” Publication no. 95. Shell Devel-
opment Company, Exploration Production and
Research Division. http://www.hubbertpeak.com/
hubbert/1956/1956 .

Intergovernmental Panel on Climate Change
(IPCC). 2013. “Summary for Policymakers.” In:
Climate Change 2013: The Physical Science Basis,
edited by T. F. Stocker et al. Contribution of
Working Group I to the Fifth Assessment Report
of the Intergovernmental Panel on Climate
Change. Cambridge University Press.

International Energy Agency. 2015. World Energy
Outlook 2015. http://www.worldenergyoutlook.org/
weo2015/.

Knittel, Christopher R. 2011. “Automobiles
on Steroids: Product Attribute Trade-offs and
Technological Progress in the Automobile Sector.”
American Economic Review 101(7): 3368–99.

Knittel, Christopher R., Douglas L. Miller,
and Nicholas J. Sanders. Forthcoming. “Caution,
Drivers! Children Present: Traffic, Pollution, and
Infant Health.” Review of Economics and Statistics.

McGlade, Christophe, and Paul Ekins. 2015.
“The Geographical Distribution of Fossil Fuels
Unused When Limiting Global Warming to 2°C.”
Nature 517( January 8): 187–90.

McKinsey & Company. 2009. Pathways to a
Low-Carbon Economy. http://www.mckinsey.com/
client_service/sustainability/latest_thinking/
pathways_to_a_low_carbon_economy.

Mills Andrew, and Ryan Wiser. 2010. “Impli-
cations of Wide-Area Geographic Diversity
for Short-Term Variability of Solar Power.”
LBNL-3884E, Lawrence Berkeley National Labora-
tory, Berkeley, California.

Murray, James, and David King. 2012. “Climate
Policy: Oil’s Tipping Point Has Passed.” Nature
481( January 26): 433–35.

Nykvist, Björn and Måns Nilsson. 2015. “Rapidly
Falling Costs of Battery Packs for Electric Vehicles.”
Nature Climate Change 5: 329–332.

Peterson, Scott B., and Jeremy J. Michalek.
2013. “Cost-Effectiveness of Plug-in Hybrid
Electric Vehicle Battery Capacity and Charging
Infrastructure Investment for Reducing US Gaso-
line Consumption.” Energy Policy 52: 429–38.

Rutledge, David B. 2013. “Projections for
Ultimate World Coal Production from Produc-
tion Histories Through 2012.” Presentation to
the 125th Anniversary Annual Meeting and Expo
of the Geological Society of America, Annual
Meeting, Denver, CO, October 2013.

Shellberger, Michael, Ted Nordhaus, Alex
Trembath, and Jesse Jenkins. 2012. “Where the
Shale Gas Revolution Came From.” Technical
Report, The Breakthrough Institute, May.

US Department of Energy. 2014. “EV Every-
where Grand Challenge.” DOE/EE-1024. http://
energy.gov/sites/prod/files/2014/02/f8/
eveverywhere_road_to_success .

US Energy Information Administration (EIA).
2012. “Potential of Gas Hydrates is Great, But Prac-
tical Development is Far Off.” November 7. http://
www.eia.gov/todayinenergy/detail.cfm?id=8690.

US Energy Information Administration (EIA).
2013. Technically Recoverable Shale Oil and Shale Gas
Resources: An Assessment of 137 Shale Formations in
41 Countries Outside the United States. June. http://
www.eia.gov/analysis/studies/worldshalegas/pdf/
fullreport .

US Energy Information Administration (EIA).
No date. “International Energy Statistics.” Database.
http://www.eia.gov/cfapps/ipdbproject/IED
Index3.cfm.

US Environmental Protection Agency. No date.
Global Greenhouse Gas Emissions Data. http://
www3.epa.gov/climatechange/ghgemissions/
global.html.

Wolak, Frank. 2015. “Mean versus Standard
Deviation Trade-offs in Wind and Solar Energy
Investments: The Case of California.” Unpublished
paper, Stanford University.

j_covert_301.indd 137 1/20/16 6:57 AM

http://pubs.aeaweb.org/action/showLinks?system=10.1257%2Faer.101.7.3368

http://pubs.aeaweb.org/action/showLinks?crossref=10.1016%2Fj.enpol.2012.09.059

http://pubs.aeaweb.org/action/showLinks?crossref=10.1016%2Fj.jebo.2014.03.010

http://pubs.aeaweb.org/action/showLinks?crossref=10.1093%2Freep%2Fres015

http://pubs.aeaweb.org/action/showLinks?crossref=10.1038%2Fnclimate2564

138 Journal of Economic Perspectives

j_covert_301.indd 138 1/20/16 6:57 AM

Will We Ever Stop Using Fossil Fuels?
Supply: Peak Oil, Natural Gas, and Coal?
Demand: Will Low-Carbon Energy Sources Knock Fossil Fuels Out of the Money?
Replacing Natural Gas and Coal in Electricity Generation
Replacing Oil Usage in Motor Vehicles
Discussion
REFERENCES