Working Paper Series

Cristiana Belu Mănescu and Galo Nuño Quantitative effects of

the shale oil revolution

No 1855 / September 2015

Note: This Working Paper should not be reported as representing the views of the European Central Bank (ECB).
The views expressed are those of the authors and do not necessarily reflect those of the ECB
 Abstract

The aim of this paper is to analyze the impact of the so-called “shale oil revolution” on oil
prices and economic growth. We employ a general equilibrium model of the world oil market
in which Saudi Arabia is the dominant firm, with the rest of the producers as a competitive
fringe. Our results suggest that most of the expected increase in US oil supply due to the
shale oil revolution has already been incorporated into oil prices and that it will produce an
additional increase of 0.2 percent in the GDP of oil importers in the period 2010-2018. We
also employ the model to analyse the collapse in oil prices in the second half of 2014 and
conclude that it was mainly due to positive unanticipated supply shocks.
Keywords: Saudi Arabia, general equilibrium, shale oil.

JEL codes: Q41, Q47, E17

ECB Working Paper 1855, September 2015 1

 Non-technical summary

Oil markets have recently undergone a significant transformation with the unexpectedly strong
rise in the US production of shale oil. The combination of horizontal drilling techniques together
with hydraulic fracturing - developed in the 1970s - and rising oil prices have made the exploration
and exploitation of large volumes of shale oil possible. In the United States, the extraction of
shale oil has grown dramatically over the last few years taking the market by surprise. In 2013,
the United States is estimated to have produced 3.5 mb/d of shale oil, which is three times higher
than the amount it produced in 2010 (EIA, 2014). Going forward, the production of shale oil in
the United States is projected to continue to rise, reaching 4.8 mb/d by 2020, equivalent to about
a third of total US oil supply. More generally, from a global perspective, the fast-rising shale oil
production has been a major factor supporting non-OPEC supply growth which, together with
moderating global oil demand, explains the relative stability of Brent oil prices until mid-2014.
Although the Unites States is expected to remain the dominant shale oil producer for the next
decade, there is potential to replicate the US shale oil success elsewhere, as technically recoverable
shale oil resources are estimated to be abundant also outside the United States (five times the size
of US shale oil resources according to EIA, 2013).
This paper assesses the impact of shale oil on oil prices and economic growth, using a general
equilibrium modeling framework that takes account of the oil market structure, and in particular
of the role of Saudi Arabia. The key advantage of our model is that the reaction of Saudi Arabia is
embedded in the model. Saudi Arabia is the largest OPEC producer and also has the largest spare
capacity which gives it substantial power among OPEC members to influence markets. As a result,
the price impact of shale oil will necessarily depend on the response of Saudi Arabia. In contrast
to other studies which typically make an explicit assumption about Saudi Arabia (PWC, 2013), in
our framework the response of Saudi Arabia is determined endogenously. Our paper complements
existing literature that mostly looked into either estimating the size of shale oil production and
resources (e.g. EIA, 2014, IEA, 2013, and earlier studies) or analyzing the factors which caused the
shale oil revolution to occur in the United States and not elsewhere (e.g. Alquist and Guenette,
2014, Maugeri, 2012). Several studies use global econometric models to estimate wider economic
implications of the shale oil revolution (PWC, 2013), whereas a few focus on the impact on oil
prices, usually gauged by relatively simple methods (e.g. PWC, 2013; Sharenow and Worah, 2013).
We assess the impact of the shale oil revolution under three scenarios about US shale oil
projections by 2018: a baseline scenario using the EIA (2014) projections of US shale oil and two
alternative scenarios to capture the uncertainty surrounding the baseline projections. In all the
three cases we assume that Saudi Arabia maximizes its discounted flow of profits. Our results
suggest that (i) most of the shale oil revolution is already priced in, and (ii) even considerable
changes in the scale of the production will have only a small effect on prices. The oil price impact

ECB Working Paper 1855, September 2015 2

of the increase in shale supply under the different scenarios amounts to changes of less than ±USD4
per barrel. The relatively muted impact on prices stems from the anticipated nature of the shock
and from the expected contraction in non-shale world oil supply, largely from Saudi Arabia, which
helps to moderate the change in world oil production and sustain prices. In the baseline scenario,
by 2018, Saudi Arabia is producing 1 mb/d less than in 2014, which largely offsets the increase
in shale oil production from the United States. According to the model, in response to lower
oil prices and a reduced oil market share, Saudi Arabia curbs its investment in new production
capacity which leads to a deterioration in its spare capacity position. Finally, the model suggests
that the shale oil revolution will produce an increase of 0.2 percent in the GDP of oil importers
by 2018, compared to 2010.
If most of the shale oil revolution is already priced-in and therefore its impact on current price
dynamics is negligible, how can we explain the collapse in oil prices in 2014? We discuss three
complementary explanations: (i) the increase in non-US supply that happened in the second half
of 2014; (ii) the global economic slowdown, which depressed most commodity prices; and (iii) the
deviation of Saudi Arabia from its optimal rule as it decided not to reduce its production in order
to accommodate the increase in supply by other producers. These three factors were unanticipated
in contrast to the anticipated nature of the shale oil revolution, at least since 2010-11. We take
the model to the data on oil prices and quantities in order to quantify the relative impact of these
three factors. We conclude that supply factors played the largest role.
The analysis of the 2014 collapse in oil prices complements the existing literature analyzing this
episode, based on structural VARs. Badel and McGillicuddy (2015) and Baumeister and Kilian
(2015) conclude that demand shocks played the major role. In contrast, Baffes et al. (2015) find
that supply shocks accounted for roughly twice as much as demand shocks in explaining the drop
in oil prices during the recent episode.

ECB Working Paper 1855, September 2015 3

1 Introduction
Oil markets have recently undergone a significant transformation with the unexpectedly strong
rise in the US production of shale oil. Shale oil refers to conventional oil (light oil with low sulphur
content) trapped in very low-permeability tight formations known as shales, which makes extrac-
tion difficult. The combination of horizontal drilling techniques together with hydraulic fracturing,
developed in the 1970s, and rising oil prices have made the exploration and exploitation of large
volumes of shale oil possible. Shale oil is only one subset of a broader category of unconventional
oil, known as "tight oil", trapped in a variety of very low-permeability formations, whose extrac-
tion has been made possible by technological advances. However, the convention in the press is to
use the terms "shale oil" and "tight oil" interchangeably when referring to oil extracted from all
low-permeability formations (i.e. not only oil from shale formations). We follow this convention
throughout the paper and refer to the entire tight oil category as "shale oil."1
In the Unites States, the extraction of shale oil has grown dramatically over the last few years
taking the market by surprise. In 2013, the Unites States is estimated to have produced 3.5 mb/d
of shale oil which is three times higher than the amount it produced in 2010 (EIA, 2014). By 2020,
US shale oil is estimated to reach 4.8 mb/d, representing about a third of total US supply. This
is having profound implications for the US economy, in terms of higher US energy independence
and a decline in its trade deficit. More generally, from a global perspective, this has been a major
factor supporting non-OPEC supply growth and contributing to the relative stability of Brent oil
prices until mid-2014. Moreover, although the United States is expected to remain the dominant
shale oil producer for the next decade, technically recoverable shale oil resources are estimated to
be abundant also outside the United States (five times the size of US shale oil resources according
to EIA, 2013).
In response, a rich body of academic and non-academic literature by investment banks, energy
companies, national and international organizations and energy pundits has emerged in an attempt
to assess the magnitude and implications of the shale oil revolution. Most studies concentrate on
the difficult task of estimating the size of shale oil production and resources (e.g. EIA, 2014, IEA,
2013 and earlier studies) or analyze the factors which caused the shale oil revolution to occur in the
United States and not elsewhere (e.g. Alquist and Guenette, 2014, Maugeri, 2012). Several studies
use global econometric models to estimate wider economic implications of the shale oil revolution
(PWC, 2013), whereas a few focus on the impact on oil prices, usually gauged by relatively simple
methods (e.g. Sharenow and Worah, 2013, PWC, 2013). We complement these works by using a
1
Unless otherwise stated, by shale oil we refer to what the Energy Information Administration (EIA) calls "tight
oil" and the International Energy Agency (IEA) calls "light, tight oil." The term tight oil does not have a specific
technical, scientific, or geologic definition. Tight oil is an industry convention that generally refers to oil produced
from very low-permeability shale, sandstone, and carbonate formations. Some of these geologic formations have
been producing low volumes of oil for many decades in limited portions of the formation.

ECB Working Paper 1855, September 2015 4

state-of-the-art general equilibrium model to assess the impact of shale oil on oil prices and GDP.
We employ the model of Nakov and Nuno (2014) in which the oil market is characterized by the
presence of a dominant supplier with a competitive fringe and which has been shown to fairly well
replicate the dynamics in the oil market.2 The key advantage of our model is that the reaction of
Saudi Arabia is embedded in the model. Saudi Arabia is the largest OPEC producer and also has
the largest spare capacity which gives it substantial power among OPEC members to influence
markets. As a result, the price impact of shale oil will necessarily depend on the response of
Saudi Arabia. In contrast to similar existing studies which typically make an explicit assumption
about Saudi Arabia (PWC, 2013), in our framework the response of Saudi Arabia is endogenously
determined.
We assess the impact of the shale oil revolution under three scenarios for US shale oil projections
until 2018: a baseline scenario using the EIA (2014) projections of US shale oil and two alternative
scenarios to capture the uncertainty surrounding the baseline projection. In all three cases we
assume that Saudi Arabia maximizes its discounted flow of profits. Our results suggest that
(i) most of the shale oil revolution has already been incorporated into oil prices, and (ii) even
considerable anticipated changes in the scale of the production will have only a small effect on
prices. The oil price impact of the increase in shale supply under the different scenarios amounts
to changes of less than ±$4 per barrel (pb). The relatively muted impact on prices onwards stems
mainly from the anticipated nature of the shock and from the contraction in non-shale world oil
supply, largely from Saudi Arabia, which helps to moderate the change in world oil production and
sustain prices. According to the model, in response to lower oil prices and a reduced oil market
share, Saudi Arabia curbs its investment in new production capacity which leads to a deterioration
in its spare capacity position. Finally, the model suggests that the shale oil revolution will produce
an increase of 0.2 percent in the GDP of oil importers by 2018, compared to 2010.
If most of the shale oil revolution is already priced-in and therefore its impact on current price
dynamics is negligible, how can we explain the collapse in oil prices in 2014? We discuss three
complementary explanations: (i) the increase in non-US supply that happened in the second half
of 2014; (ii) the global economic slowdown, which depressed most commodity prices; and (iii) the
deviation of Saudi Arabia from its optimal rule as it decided not to reduce its production in order
to accommodate the increase in supply by other producers. These three factors were unanticipated
in contrast to the anticipated nature of the shale oil revolution, at least since 2010-11. We take
the model to the data on oil prices and quantities in order to quantify the relative impact of these
three factors. We conclude that supply factors played the largest role.
The analysis of the 2014 collapse in oil prices complements the existing literature analyzing this
2
A large portion of the literature suggests that the ‘dominant supplier with competitive fringe’ view of the oil
industry is the one that best describes the empirical evidence. Examples include Mabro (1975), Dahl and Yücel
(1991), Adelman (1995), Alhajji and Huettner (2000a,b) and Brémond, Hache and Mignon (2011).

ECB Working Paper 1855, September 2015 5

episode, based on structural VARs. Badel and McGillicuddy (2015) and Baumeister and Kilian
(2015) conclude that demand shocks played the major role. In contrast, Baffes et al. (2015) find
that supply shocks accounted for roughly twice as much as demand shocks in explaining the drop
in oil prices during the recent episode.
This paper relates to the literature on general equilibrium models that explicitly analyze the
oil market. Recent examples include Backus and Crucini (2000), Blanchard and Gali (2010),
Bodenstein et. al. (2011), Bodenstein and Guerrieri (2011), Bodenstein at al. (2012), De Miguel
and Manzano (2006), Leduc and Sill (2007) and Nakov and Pescatori (2010a,b).
The structure of the paper is as follows. Section 2 describes the magnitude and importance
of the shale oil revolution. Section 3 introduces the model and the shale oil production scenarios
and discusses the main results. Section 4 analyzes the 2014 episode. Section 5 discusses the policy
implications. Finally, section 6 concludes.

2 The transformational impact of US shale oil for the oil

market
2.1 The size of the US shale oil revolution
In the United States, the production of shale oil has grown dramatically over the last few years.
In 2013, the Unites States is estimated to have produced 3.5 mb/d of shale oil, which is three
times as much as in 2010 (EIA, 2014). This accounted for nearly half of US crude oil supply and
almost a third of its total oil supply in 2013 (see Figure 1). As a result, shale oil put an end to
a prolonged period of declining supplies: the Unites States has for the first time produced more
oil in 2013 than in 1985. Looking ahead, US shale-oil production is expected to continue to rise
and peak at 4.8 mb/d in 2020 according to the EIA (see Table 1), before gradually declining to
3.2 mb/d by 2040.
From a global perspective, the rapid rise of US shale oil has been the main driver behind the
increase in non-OPEC supply (see Figure 2). During 2009-2012, US oil supply grew, on average,
at a sustained annual rate of 6.7%, while non-OPEC oil supply produced outside the Unites States
was stable, on average.3 The US share in total non-OPEC supply jumped to more than a quarter
in 2012, compared to just over a fifth in 2008. In 2012, US shale oil accounted for 2.5% of world
oil supply.
Despite its impressive volume, the shale oil revolution was largely unexpected. For example,
in its latest 2014 Annual Energy Outlook, the EIA projects US shale oil production at 4 mb/d in
3
At the time of writing, the latest EIA data for shale oil are for 2012, while data for 2013 onwards are projected
(see EIA, 2014).

ECB Working Paper 1855, September 2015 6

2014, which is nearly three and a half times higher that the amount forecast two years ago (see
Table 1). Similarly, the International Energy Agency (IEA) in its 2013 report revised its estimates
for 2012 up by 30% from those in its 2012 one (see IEA 2012a and 2013). This is not surprising,
however, given that forecasts of shale oil production are critically influenced by assumptions about
the remaining unproved technically recoverable resources, which, with a short history of well data,
can be extremely imprecise. As more information about the well production becomes available,
the size of revisions to shale oil production projections is expected to diminish. Indeed, the size of
forecast revisions was much smaller in 2014 than in 2013 (see Table 1 or EIA, 2014).
The persistent high levels of oil prices in recent years made shale oil exploration economically
viable. However, several factors were in place in the Unites States that allowed the shale oil
revolution to occur first there and not elsewhere: a history of shale gas exploitation, legal incentives
for landowners and an advanced oil production infrastructure (see Alquist and Guenette, 2014).
First, the technology necessary for shale oil exploitation started to be developed in the early 19th
century, when it was aimed at shale gas exploration, and was perfected during the 1980s. Second,
the Unites States has legal and institutional features - such as private ownership rights and easy
access to global capital market funding - that make the economic environment attractive for the
extraction of unconventional oil. Third, the Unites States possesses a unique infrastructure setting
consisting of a large number of state-of-the-art drilling rigs, an extensive pipeline network, and
associated refineries, which is key to ensuring a rapid expansion in production.
Outside the Unites States, shale oil production is expected to take off only slowly, despite the
fact that technically recoverable resources of shale oil are estimated to be abundant worldwide
(five times the size of the ones in the Unites States according to EIA, 2013).4 Companies are
hinting at good prospects in Russia, China and Argentina, which are estimated together to hold
46% of the technically recoverable resources of shale oil outside the Unites States. The primary
reason for this is that technically recoverable resources are different from economically recoverable
resources, owing to three important factors: the cost of production, the quantity produced by any
given well and the market price of oil. As mentioned, the experience with shale oil production
in the Unites States highlights that economic recoverability can be significantly influenced by
above-the-ground factors (property rights, infrastructure, etc.) as well as by geology, which vary
greatly around the globe. This is also reflected in the wide range of the estimated price required to
render shale oil exploitation profitable (at $45-80 pb, according to IEA, 2012a). Nevertheless, the
unconventional technologies used for shale oil extraction could be used to boost the production of
existing conventional oil fields globally.
4
Whereas in the rest of the paper by shale oil we refer to both shale and tight oil, the shale oil reserve estimates
refer exclusively to oil from shale oil formations.

ECB Working Paper 1855, September 2015 7

2.2 The short-term impact of the shale oil revolution
The surge in US unconventional oil production has had a major impact on the US economy.
First, a major importer of oil and natural gas since the 1970s, the Unites States is projected to
significantly reduce its energy import dependency. This would entail a reduction in the US trade
deficit. The US self-sufficiency ratio, defined as the ratio of indigenous energy production divided
by total primary energy demand, is expected to rise from just under 80 percent in 2010 to 97
percent by 2035 which stands in contrast to developments in other regions (China, India, and the
European Union) which will become less self-sufficient. Overall, the IEA estimates that US net
energy import volumes could fall from 9.5 mb/d in 2011 to 3.4 mb/d in 2035.5 At the same time,
a reduction in US net oil imports would also affect global oil supply indirectly, as supply usually
directed to the Unites States would become available for other buyers.
Second, the rising oil supply led to a collapse in the spread between the WTI, the US crude
oil price benchmark, and international benchmarks. Prices of the WTI and Brent crude oil bench-
marks have historically been related, with the WTI trading at a slight premium ($1-2 pb) owing
to higher quality and stronger product demand. This historical relationship collapsed in 2011.
Limitations in the US crude pipeline system and increasing oil flows to the WTI benchmark pric-
ing hub at Cushing, Oklahoma, helped push WTI to a discount of $16 pb to Brent forward crude
on average during 2011 which persisted in 2012. More recently, however, increased pipeline and
rail capacity eased the transport bottlenecks in the landlocked US midcontinent, narrowing the
discount to Brent to less than $8 pb in May 2014.
At the global level, the surge in supply from US shale sources was an important factor con-
tributing to the relative stability of Brent oil prices observed from 2011 until mid-2014. Despite
substantial and repeated supply disruptions in major OPEC producers (e.g. Libya, Iran) that tem-
porarily spiked oil prices, Brent oil prices remained remarkably resilient in the range of $100-115
pb during that period.6 In addition to the mitigating impact of the fast-rising shale oil supplies,
estimated to have added 3.9% to total oil supply in 2013, demand-side factors also played a role.
The slow economic recovery in advanced economies after the financial crisis, coupled with the slow-
down in emerging economies, reduced the growth in oil demand. Indeed, according to the IEA,
global oil demand increased only 1.1% per annum on average between 2011 and 2013, compared
to a strong 1.5% per annum during the period 2005-2007.
5
See IEA (2012b).
6
For example, Libya lost a significant share of its supply both in 2013 and 2011 due to political uncertainty (1.3
mb/d in 2011 and 0.5 mb/d in 2013), while sanctions by Western nations on Iranian oil exports reduced Iranian
supply by 0.7 mb/d in 2012.

ECB Working Paper 1855, September 2015 8

3 The long-term impact of the shale oil revolution: a DSGE
model of the global oil market
The long-term impact on oil prices of further rising shale oil supplies is, however, far from clear and
will crucially depend on the reaction of Saudi Arabia. Saudi Arabia is one of the largest players in
the global oil market: it produces more than a tenth of the world’s oil output and owns a quarter
of the world’s proven reserves. The Kingdom is also a key OPEC member, typically playing a
central role in OPEC’s decision-making. Saudi Arabia’s spare capacity is much larger than the
aggregate spare capacity of the rest of the oil producers. In addition, as discussed in Nakov and
Nuño (2014), it typically acts as a “swing producer” in the oil market, increasing its production
in the face of supply disruptions in other producers or rises in demand, and reducing it in the
opposite case.
General equilibrium effects are therefore key to understanding the effect of the increase in
shale oil production in the United States; in particular, the response of Saudi Arabia is especially
relevant. In this paper we employ the model introduced in Nakov and Nuño (2014) to analyze
the oil market. The model comprises three regions: one oil-importing and two oil-exporting. The
oil-importing region imports oil for use in consumption, and employs labor and capital in the
production of final goods, some of which are consumed domestically, the rest being exported to
the two oil-producing regions. The model is summarized in the Appendix. Oil is a homogeneous
commodity supplied by a dominant supplier and a fringe of competitive oil producers. The fringe
take the oil price as given when choosing their production level. The dominant supplier faces a
downward sloping residual demand curve and picks the profit-maximizing points on that curve at
each point in time and in each state of nature. Oil exporters produce oil only and their revenue is
recirculated to the oil-importing region in the form of demand for final consumption and investment
goods.
The dominant supplier is aware that it can manipulate à la Stackelberg the choices both of the
competitive fringe, and of the oil importer. For example, the dominant supplier understands that
a change in its own oil supply will have an impact on oil demand, oil supply by fringe producers,
and the oil price. Except for the difference in market power, which is founded on a technological
gap between the two types of oil producers, the competitive fringe and the dominant supplier are
modelled symmetrically.

3.1 Scenarios
In order to introduce the different scenarios into the model, we calibrate the shocks for the oil
production technology of the fringe producers in order to replicate the increase in supply due

ECB Working Paper 1855, September 2015 9

to the shale oil revolution. These shocks are assumed to be anticipated since the beginning of
the period (i.e. 2010). It implies that the agents in the model know in advance what the rate
of production of shale oil will be in the coming years and that they may take their decisions
accordingly in a forward-looking way.
We consider one baseline and two alternative scenarios for the size of shale oil supply, based
on the EIA projections published in its Annual Energy Outlook (EIA, 2014). We rely on EIA
estimates of US shale oil supply because, unlike other international institutions such as the IEA,
the EIA also provides detailed historical data on the production of US shale oil, which facilitates
the understanding of recent developments from a historical perspective.
The baseline scenario is calibrated so that shale oil production in the Unites States increases
from roughly zero in 2010 to 2.7 mb/d in 2014 and reaches a peak of 4 mb/d in 2018.7 This scenario
assumes that current law and regulations affecting the energy sector remain in place throughout
the projection horizon. It also assumes that the production of shale oil remains concentrated
in the Unites States and does not extend significantly to other countries before 2018. Second,
in the lower shale production scenario we consider the case in which technical or environmental
constraints limit US shale production which remains broadly flat at 3mb/d during the 2014-2018
period. Finally, in the higher shale production scenario we assume that technological progress and
economic incentives push US shale production up, to 6 mb/d by 2018.
The choice of the alternative scenarios is, to some extent, arbitrary because essentially they
are meant only to suggest the possible range of the price impact in the event that future shale
oil supplies turn out to be different from the baseline. While the magnitude of the alternative
scenarios does not necessarily correspond to alternative scenarios proposed in the literature, the
asymmetric bounds were chosen on purpose to reflect the idea that, in light of historical revisions,
there is a higher chance that shale oil supply will surprise on the upside than on the downside.

3.2 Results
Figure 3 shows the difference between the long-run path followed by oil production and prices in
the three scenarios and the counterfactual path in the case of no shale oil production. The period
of interest is 2014-2018, but we assume, in line with the data, that shale oil production began in
2010. In the upper left panel we display the alternative scenarios for the production of shale oil
described above.
The model analysis suggests that oil prices in 2014 already incorporated most of the reduction
due to the increase in the oil supply. In the upper right panel of the figure, we display the evolution
of oil prices (relative to the counterfactual path). In the baseline scenario, most of the impact of
7
Notice that we are subtracting 0.8 mb/d in all years in order to remove the “pre-revolution” residual production
of shale oil.

ECB Working Paper 1855, September 2015 10

the expected increase in oil production was already incorporated into 2014 prices, which were $5
pb below the counterfactual path. There is only an additional decline of less than $1 pb before
the end of 2015 and then prices slowly revert. This means that, from the perspective of 2014 oil
prices, the anticipated increase of 1 mb/d until 2018 has almost no effect on long-run oil prices.
This effect is more striking in the case of the lower production scenario, where prices revert
back to around -$3 pb by 2018. In contrast, in the higher production scenario there is still room
for price decreases, reaching levels below -$9 pb by 2017. These bounds based on our alternative
scenarios are roughly ±$4 pb around the baseline.
The relatively muted impact on prices from 2014 onwards stems from the anticipated nature
of the shocks and from the contraction in non-shale world oil supply, primarily from Saudi Arabia,
which helps to moderate the change in overall world oil production and sustain prices. The lower
panels of Figure 3 display the total world production and the production of Saudi Arabia. In
2014, of almost 3 mb/d of shale oil, around 1 mb/d were expected to be offset by Saudi Arabia
and another 0.5 md/d by the rest of producers due to the lower oil prices. The result was a net
expansion of just 1.5 mb/d compared to the no shale oil counterfactual scenario, since half of
the increase in oil production in the Unites States was offset either by Saudi Arabia or by higher
marginal cost producers being expelled from the market due to the $5 fall in prices.
The amount of the reduction in oil supply by Saudi Arabia increases with time as this is
its profit-maximizing path. In the baseline scenario, by 2018, Saudi Arabia should be producing
around 1 mb/d less than in 2014, which would completely offset the increase in shale oil production.
In the lower and higher shale production scenarios, Saudi Arabia should also reduce its production
by a further 0.5 mb/d and 1.5 mb/d, respectively, by 2018.
Apart from the impact on prices and quantities, the shale oil revolution also affects Saudi
Arabia’s spare capacity. Spare capacity provides a buffer to smooth supply and demand shocks.
Our baseline model simulation indicates that spare capacity should have risen slightly in 2014 as
a result of the reduction in Saudi Arabia’s production (see Figure 4). Thereafter, however, spare
capacity declines. The explanation is that lower oil prices and a reduction in Saudi Arabia’s oil
market share lower its capital investment in new production capacity, which eventually leads to a
gradual decline in its spare capacity. In conclusion, the spare capacity of Saudi Arabia is projected
to decline in response to lower oil prices and lower capital investment. Only in the higher shale
scenario is there still room for an additional increase in capacity, peaking in mid-2016.
The shale oil revolution also impacts the GDP of oil importers. As discussed in the Appendix,
the model considers an aggregate block of "oil importers" without distinguishing between advanced
and emerging economies. We compute a cumulative impact equivalent to 0.2 percent of 2010 GDP
by 2018. This amount is the result of three different effects. The first one is the direct impact of
a change in the terms-of-trade between importers and exporters due to the decline in oil prices.

ECB Working Paper 1855, September 2015 11

This effect is clearly growth-enhancing. The second effect is the fall in the imports of oil exporters
due to the fall in their demand. This second effect is growth-reducing for oil importers. Finally,
the third effect is related to the impact that lower energy prices may have on the consumption
and leisure decisions of households, and its impact on growth is much more difficult to predict.
It should be stressed that alternative assumptions about the structure of financial markets or the
consumption-leisure elasticities may increase the impact on GDP, so that this result should be
read as a lower bound of the true impact in the real world.

3.3 Robustness to the demand elasticity

We also consider a robustness check. As discussed in the calibration part of the Appendix, the
P ∂O
short-run price elasticity of oil demand in the model is O ∂P
= η1 = 0.25, consistent with Kilian and
Murphy’s (2014) estimates. However other papers, such as Smith’s (2009), provide an estimate
of the price elasticity of oil demand of about 0.05. We denote this second case as "low demand
elasticity". Here we assess how sensitive our results are to changes in the demand elasticity.8
Figure 5 displays the results. In this case the anticipated effect of the increase in shale supply
is still larger, reaching almost $9 pb in 2014 and forcing Saudi Arabia to reduce its production
by 2 mb/d. However, looking ahead the dynamics described in the baseline case are even more
patent. Saudi Arabia should again offset more than 1 mb/d between 2014 and 2018 so that total
oil production decreases slightly during the period. Oil prices increase by $2 pb. Therefore, it
seems that the main conclusions are reinforced in the case of low demand elasticity.

4 The collapse in oil prices in 2014

Brent oil prices collapsed from a level of $105 pb in June 2014 to less than $60 pb in December.
Given this major decline, the conclusion that most of the current and expected increase in US
supply due to the shale oil revolution had already been incorporated into oil prices may seem
surprising. However, several additional factors to the shale oil revolution may account for this
decline:

1. Unanticipated supply shocks. Several major producers experienced large increases in

their production in the second half of the year. In Libya, for example, rebel militias an-
nounced in July an end to their 11-month blockade of vital oil terminals that had crippled
oil production.9 Iraq increased its production by almost 0.5 mb/d between June and De-
8
To this end, we set η = 21 and recalibrate the supply shocks in the baseline case to 0.0052 from January 2010
to December 2017 and 0.0045 thereafter. See the Appendix for the baseline calibration.
9
Libya’s state-owned NOC on 6 July lifted force majeure - a waiver of contractual obligations - at the eastern
ports of Ras Lanuf and Es Sider.

ECB Working Paper 1855, September 2015 12

 cember despite the conflict with the so-called Islamic State for Iraq and Syria (ISIS). Other
non-OPEC producers such as Russia, Canada, Norway and the UK also increased their out-
put during the period. The total increase in oil supply from June to December for all the
producers bar the Unites States was roughly 1.5 mb/d (see Figure 6), compared to a US
increase of 0.6 mb/d.

2. Unanticipated demand shocks. During 2014 there were a number of negative surprises
to the global economic growth rate which forced the IEA to revise downwards its oil demand
forecasts. This global slowdown produced a decline in the prices of many commodities,
including metals and foodstuffs, during the second half of the year. In addition, the strong
appreciation of the US dollar increased the real price of oil in other currencies, thus reducing
the demand for oil.

3. Surprise Saudi Arabia reaction. As discussed above, in a context of positive supply

and negative demand shocks, the optimal profit maximizing strategy for Saudi Arabia is to
partially reduce its production in order to prevent the collapse in oil prices.10 This time
however Saudi Arabia strongly signalled its reluctance to sacrifice market share in order to
defend prices. At its 166th meeting on 27 November, OPEC as a whole decided to maintain
the quotas unchanged due to the adamant position of Saudi Arabia, despite rising pressures
from other members such as Iran or Venezuela to reduce them. Figure 6 displays how Saudi
Arabia only reduced its output by 0.05 mb/d from June to December. Apparently, Saudi
Arabia is trying to avoid repeating the events in the early 1980s, when it systematically lost
output share in order to defend prices amid a generalized increase in oil production.11 It could
be the case that Saudi Arabia has decided that the current situation with shale oil is similar
to the one in the 1980s and that it may be beneficial to deviate from its profit maximizing
strategy in order to adopt a constant share rule. This would expel some shale oil producers
from the market, avoiding demand substitution by renewable sources and weakening the
position of regional rivals such as Iran.
10
See also Nakov and Nuño (2014).
11
Higher prices during the 1970s and early 1980s, following the first and second oil shocks, caused oil supply in
non-OPEC countries to increase significantly (e.g. in the Soviet Union, Norway, Mexico, China and the UK). In
response to this, OPEC decided to lower its production via a quota system, in an attempt to defend prices. This
notwithstanding, oil prices continued to fall. Moreover, Saudi Arabia had to absorb most of these reductions as
many OPEC members were still producing above their quotas. Indeed, at its lowest in 1983, Saudi Arabia’s output
was nearly 8 mb/d down from the beginning of 1981. The combination of lower prices and declining production
substantially reduced Saudi Arabia’s revenues, which fell by more than 80% from January 1981. In September 1985,
Saudi Arabia decided to raise production in an effort to discipline the other producers. Over the next year, its
output increased by more than 4 mb/d with dramatic effects on oil prices, which declined more than 50% over the
period (known as the third oil shock), from $28 to $12 pb. Thus, the sharp fall in oil prices fully offset the rise in
output, which kept Saudi oil revenues constant.

ECB Working Paper 1855, September 2015 13

 In order to assess the relative importance of each of these factors in the decline in prices, we
employ the DSGE model. We augment the model in order to include unanticipated supply and
demand shocks as well as deviations from Saudi Arabia’s optimal strategy.12 The model is solved
by first order perturbation methods and the resulting linear state space model is embedded into
a Kalman filter. In order to compute the realizations of the three shocks (supply, demand and
Saudi Arabia) we consider three observed series: (i) growth in real oil prices; (ii) growth in world
oil production; and (iii) Saudi Arabia’s share in oil production. We employ monthly data from
January 1973 to December 2014. The Kalman filter allows us to identify the most likely paths of
the shocks in order to generate the observed time series. Then we employ each of these shocks to
compute the relative contributions to real oil prices.
Figure 7 displays the results. The three factors discussed above each contributed to the decline
in prices, with the largest contribution stemming from supply shocks. Demand shocks also play
an important role, whereas the reaction of Saudi Arabia seems to have played a minor role in
this particular episode. These results suggest that the recent decline in prices was mainly due to
the large increase in oil production by non-US producers (as the increase in shale oil was already
priced-in) against a background of demand slowdown.

5 Policy implications
The shale oil revolution and the recent decline in oil prices affect economic growth in a country-
specific way, and thus the policy prescriptions should vary accordingly.
The losers are the traditional oil exporters. Figure 8 displays the share of oil exports in GDP for
a sample of major oil exporters. A decline by half in the price of oil may produce GDP contractions
as large as 20-25% for countries such as Kuwait and Venezuela. Some of these countries, such as
Saudi Arabia and Kuwait, have large sovereign wealth funds (SWFs) that may help to cushion the
fall in prices but others are more exposed to this negative shock.
Once again, the important role of Saudi Arabia in the future evolution of oil prices should be
highlighted. In 2014, in contrast with the long-run predictions of the model, Saudi Arabia has
decided not to accommodate the increase in oil supply, in an attempt to defend market share and
to expel some high cost producers from the market. This is a strategy aimed at curbing new
investment in exploration, avoiding demand substitution towards alternative energy sources and
improving its relative position vis-à-vis some of its geopolitical competitors such as Iran. However,
in the short-run the Saudis are facing a major loss of revenues that, if long-lasting enough, may
destroy all the wealth accumulated during the boom years.
12
Due to their negligible short-run effect, we ignore the shale oil shocks. The extension is explained in the
Appendix.

ECB Working Paper 1855, September 2015 14

 The main winner is naturally the United States. The US shale oil supply is estimated to be
around 4 mb/d in 2014. Assuming an average price of $100 pb, this amounts to more than $146
billion in oil revenues or roughly 0.8 percent of US GDP. This is a major transfer of resources from
the rest of the world to the US economy that not only boosts economic growth but also helps to
reduce the US current account deficit. Beyond the direct impact, there is a full range of second-
round effects on the US economy associated with the development of new ancillary industries and
the increase in employment associated, directly or indirectly, with the shale oil sector.
Oil-importing countries such as Japan, India, China and most European countries are also
indirect winners of the shale revolution and the recent drop in prices. The decrease in oil prices
will help to reduce energy imports and to increase GDP growth, as highlighted by Figure 8.
However, two important caveats should be mentioned. The first one relates to the recycling of
petrodollars. Many oil exporters have large amounts of accumulated wealth in real and financial
assets in oil-importers through SWFs. The decline in oil revenues will probably reduce the exposure
to such assets, reducing their price. In a context of high private and public debt levels, a fall in
debt prices (a rise in yields) may have contractionary effects due to the increased costs of debt
servicing. The second point relates to the current situation in many advanced economies, where
short-term nominal interest rates are close to the zero lower bound (ZLB) and inflation rates are
very low or even negative. A decline in oil prices may introduce additional deflationary pressures,
producing a hike in real interest rates that may hurt economic growth.13 Therefore, central banks
in oil-importing countries should be ready to act in the event that further declines in oil prices
that may threaten the economic recovery.

6 Conclusions
Oil markets have recently undergone a significant transformation with the rise in the US production
of shale oil, which is expected to pick up and to reach around 4.8 mb/d by 2020, i.e. about a third
of total US supply, according to the EIA. This paper assesses the impact of the “US shale oil
revolution” on global oil prices and quantities, which ultimately depends on two key factors: 1)
the quantity of oil that is eventually produced, subject to environmental and technological risks
and 2) the response of Saudi Arabia to sustain prices.
To capture the current uncertainty regarding the size of future US shale oil supply, we consider
two alternative scenarios, i.e. a lower and a higher shale oil production scenario; this results in
some bounds around our baseline projections, which are based on EIA forecasts. Regarding the
response by Saudi Arabia to sustain prices, we employ a general equilibrium model that rationalizes
13
The impact of oil prices on inflation and economic activity at the ZLB has been discussed in recent papers such
as Bodenstein et al. (2013) and Wieland (2014).

ECB Working Paper 1855, September 2015 15

the behavior of Saudi Arabia - the core OPEC producer - as that of a dominant producer with a
competitive fringe. The model allows for an endogenous response by Saudi Arabia, thus avoiding
having to make specific assumptions on the OPEC response to the surge in global oil supply.
Results suggest that most of the expected increase in oil supply due to the shale oil revolution
was already incorporated into oil prices in 2014 and that it will produce an additional increase
in the GDP of oil importers of 0.2 percent in 2010-2018. The oil price impact of the increase
in shale supply under the different scenarios by 2018 amounts to changes of less than ±$4 pb.
The relatively muted impact on prices stems from the anticipated nature of the shock and from
the expected contraction in non-shale world oil supply, largely from Saudi Arabia, which helps to
moderate the change in world oil production and sustain prices.
Finally, the paper sheds some light on the causes of the collapse in oil prices in the second
half of 2014. Our results suggest that the decline was not due to the anticipated increase in shale
oil, but to unexpected increases in oil production in a number of major non-US producers against
a background of negative demand surprises and unwillingness by Saudi Arabia to accommodate
the increase in supply. In particular, we show that non-US supply increased by roughly 1.5 mb/d
between June and December. Therefore, this paper underlines the weak connection between the
shale oil revolution, which is an anticipated long-term phenomenon, and the recent collapse in
prices, which is the consequence of unanticipated shocks in the market.

ECB Working Paper 1855, September 2015 16

References
[1] Adelman, M.: 1995. The Genie Out of the Bottle: World Oil Since 1970, The MIT Press.

[2] Alhajji, A. and Huettner, D.: 2000a, OPEC and other commodity cartels: a comparison,
Energy Policy, vol. 28, pp. 1151—1164.

[3] Alhajji, A. and Huettner, D.: 2000b, OPEC and world crude oil markets from 1973 to 1994:
cartel, oligopoly or competitive?, Energy Journal vol. 21(3), pp. 31—60.

[4] Alquist, R., and Guenett, J.D.: 2014, A blessing in disguise: The implications of high global
oil prices for the North American market, Energy Policy, vol. 64(C), pp. 49-57..

[5] Backus, D. and Crucini, M.: 2000, Oil prices and the terms of trade, Journal of International
Economics 50(1), 185—213.

[6] Badel, A. and J. McGillicuddy: 2015, Oil Prices: Is Supply or Demand behind the Slump?,
On the Economy, Federal Reserve Bank of St. Louis

[7] Baffes, J., M. A. Kose, F. Ohnsorge and M. Stocker: 2015, The Great Plunge in Oil Prices -
Causes, Consequences, and Policy Responses, Policy Research Note No.1, World Bank.

[8] Baumeister, C. and L. Kilian: 2015, Understanding the Decline in the Price of Oil Since June
2014, CEPR Discussion Paper 10404

[9] Blanchard, O. J. and J. Gali: 2010, The Macroeconomic Effects of Oil Shocks: Why are the
2000s So Different from the 1970s?, in J. Gali and M. Gertler (eds) International Dimensions
of Monetary Policy, University of Chicago Press, Chicago, 373-421

[10] Bodenstein, M., and L. Guerrieri: 2011, “Oil Efficiency, Demand and Prices: A Tale of Ups
and Downs,” International Finance Discussion Papers No. 1031, Board of Governors of the
Federal Reserve System.

[11] Bodenstein, M., L. Guerrieri, Luca and C. Gust: 2013, Oil shocks and the zero bound on
nominal interest rates," Journal of International Money and Finance, vol. 32(C), pp. 941-967.

[12] Bodenstein, M., Erceg, C. J. and Guerrieri, L.: 2011, Oil shocks and external adjustment,
Journal of International Economics, 83(2), 168-184.

[13] Bodenstein, M., L. Kilian, and L. Guerrieri: 2012, Monetary Policy Responses to Oil Price
Fluctuations, IMF Economic Review, vol. 60(4), pp. 470-504.

ECB Working Paper 1855, September 2015 17

[14] Brémond, V., E. Hache and V. Mignon: (2011). ‘Does OPEC still exist as a cartel? An
empirical investigation,’ EconomiX Working Papers 2011-5.

[15] Dahl, C. and M. Yucel: (1991). ‘Testing alternative hypotheses of oil producer behavior,’ The
Energy Journal, vol. 12(4), pp. 117-138.

[16] De Miguel, C. and B. Manzano: 2006. Optimal Oil Taxation in a Small Open Economy,
Review of Economic Dynamics, 9(3), 438-454.

[17] Energy Information Administration: 2013, Technically recoverable shale oil and shale gas
resources: An assessment of 137 shale formations in 41 countries outside the United States,
Washington.

[18] Energy Information Administration: 2014, Annual Energy Outlook, Washington.

[19] International Energy Agency: 2008, World energy outlook, OECD, Paris.

[20] International Energy Agency: 2009, World energy outlook, OECD, Paris.

[21] International Energy Agency: 2012a, Medium-term oil market report, OECD, Paris.

[22] International Energy Agency: 2012b, World energy outlook, OECD, Paris.

[23] International Energy Agency: 2013, Medium-term oil market report, OECD, Paris.

[24] Kilian, L., and D. P. Murphy: 2014, The Role of Inventories and Speculative Trading in the
Global Market for Crude Oil, Journal of Applied Econometrics, 29(3), 454-478.

[25] Leduc, S. and Sill, K.: 2007, Monetary policy, oil shocks, and TFP: Accounting for the decline
in US volatility, Review of Economic Dynamics 10(4), 595—614.

[26] Mabro, R.: 1975, ‘Can OPEC hold the line?,’ in OPEC and the World Oil Market: The
Genesis of the 1986 Price Crisis.

[27] Maugeri, L.:2012. The Shale Oil Boom: A U.S. Phenomenon, Discussion Paper 2013-05, Belfer
Center for Science and International Affairs, Harvard Kennedy School.

[28] PWC: 2013, Shale oil: the next energy revolution http://www.pwc.com/en_GX/gx/oil-gas-
energy/publications/pdfs/pwc-shale-oil.pdf

[29] Sharenow, G. E. and M. P. Worah: 2013. Shale Oil: A Deep Dive Into Implications for the
Global Economy and Commodity Investors http://www.pimco.com/en/insights/pages/shale-
oil-a-deep-dive-into-implications-for-the-global-economy-and-commodity-investors.aspx

ECB Working Paper 1855, September 2015 18

[30] Nakov, A. and G. Nuño: 2014, Saudi Arabia and the Oil Market, Economic Journal, 132 (12),
1333-1362.

[31] Nakov, A. and Pescatori, A.: 2010a, Monetary policy tradeoffs with a dominant oil producer,
Journal of Money, Credit and Banking, 42(1), 1-32.

[32] Nakov, A. and Pescatori, A.: 2010b, Oil and the Great Moderation, Economic Journal, 120
(543), 131-156.

[33] Smith, J.: 2009, World oil: Market or mayhem?, Journal of Economic Perspectives 23(3), 145—
164.

[34] Wieland, J. F.: 2014, Are Negative Supply Shocks Expansionary at the Zero Lower Bound?,
mimeo.

ECB Working Paper 1855, September 2015 19

Appendix. Model description
This Appendix is strongly based on Nakov and Nuño (2014): we sketch their model and discuss
the particular calibration.

Oil-importing region
A representative household has a period utility function which depends on consumption, Ct , oil
Ot , and labor lt , and takes the form

U (C, O, l) = log(C) + ν t O1−η /(1 − η) − l1+ω /(1 + ω),

where η and ω are oil demand and labor supply elasticities.

The household faces the period budget constraint

Ct + It + Bt + Pt Ot = wt lt + rtk Kt−1 + rt−1 Bt−1 , (1)

which equates income from labor, wt lt , capital, rtk Kt−1 , and bonds, rt−1 Bt−1 , to outlays on con-
sumption, Ct , capital investment, It , new one-period bonds, Bt , and oil, Pt Ot ; Pt denotes the
real price of oil, while wt denotes the real wage. All oil must be consumed within the period of
production. Capital is accumulated according to

Kt = (1 − δ) Kt−1 + It , (2)

where δ is the depreciation rate of installed capital.

The household chooses Ct , Ot , Bt , It , Kt , and lt , to maximize expected present discounted
utility
∞
max Eo β t U (Ct , Ot , lt )
Ct ,Ot ,Bt ,It ,Kt ,lt t=0

subject to the budget constraint (1), where 0 < β < 1 is a time preference parameter.
Final goods are produced with labor and capital by a representative price-taking firm, according
to
Yt = (Zt lt )α Kt−1
1−α
, (3)

with 0 < α < 1. Aggregate total factor productivity Zt evolves according to a deterministic
z
exponential process: Zt = Zt−1 eg , where g z is a constant.
Utility maximization by households implies the following oil demand curve

ν t Ct = Pt Otη , (4)

ECB Working Paper 1855, September 2015 20

the labor supply curve
Ct ltω = wt , (5)

and the Euler equations  

Ct  k 
1 = βEt r + (1 − δ) , (6)
Ct+1 t+1
and  
Ct
1 = βEt rt . (7)
Ct+1
Profit maximization by price-taking firms implies the following labor demand

wt lt = αYt , (8)

and capital demand

rtk Kt−1 = (1 − α)Yt . (9)

The region’s resource constraint states that total output, Yt , must equal the sum of consump-
tion, Ct , investment, It , and the current account, (Pt Ot + Bt − rt−1 Bt−1 ) .

Yt = Ct + It + (Pt Ot + Bt − rt−1 Bt−1 ) . (10)

Competitive fringe of oil exporters

A representative household maximizes the present discounted flow of utility from consumption,14
∞
max Eo β t log(C̃t )
C̃t ,I˜t ,B̃t ,K̃t t=0

subject to the period budget constraint

C̃t + I˜t + B̃t = r̃tk K̃t−1 + D̃t + r̃t−1 B̃t−1 ,

where consumption, C̃t , and investment, I˜t , are both purchased from the oil-importing region, r̃tk
is the rental price of capital K̃t rented out by the household to oil firms, B̃t are one-period bonds
that pay an interest of r̃t and D̃t are oil firm dividends rebated lump sum to the household.
The household invests in capital according to

K̃t = (1 − δ) K̃t−1 + I˜t .

14
We denote variables belonging to the competitive fringe by tildes (e.g. X̃t ), and variables belonging to the
dominant firm by hats (e.g. X̂t ).

ECB Working Paper 1855, September 2015 21

 A representative fringe firm, owned by the household, maximizes period profits
 
D̃t = max Pt Õt − X̃t − r̃t K̃t−1
X̃t ,K̃t−1

subject to the production technology

Õt = Z̃t ũt K̃t−1 , (11)

 
taking the oil price as given. Z̃t is the productivity of oil production and ũt K̃t−1 is the effective
capital in the oil industry, defined as the product of capital utilization ũt and installed capital K̃t−1 .
Productivity Z̃t = ãt Z̃0 exp(g z̃ t) follows an exogenous process with a secular trend component
Z̃0 exp(g z̃ t), with g z̃ < 0, and a stationary AR(1) component in logs, log(ãt ) = ρã log(ãt−1 ) + εãt
with persistence ρã and a Gaussian innovation εãt with mean zero and variance σ 2ã . The downward
trending component of oil productivity, Z̃0 exp(g z̃ t), is introduced to capture the fact that, over
time, the cheaper sources of oil production are exhausted before the more expensive ones.
There is variable capital utilization. In order to operate the installed capital, i.e., to have a
capital utilization rate above zero, the representative fringe firm employs the intermediate good
X̃t purchased from the oil-importing region and rents the capital K̃t from the household given a
decreasing returns to scale function. The capacity utilization rate of installed capital is given by

ũ(X, K) ≡ (X/K)γ̃ , (12)

so that the oil production technology results in

Õt = Z̃t X̃tγ̃ K̃t−1

1−γ̃
. (13)

Optimal oil supply by competitive fringe producers implies

γ̃Pt = X̃t /Õt , (14)

while optimal capital accumulation is given by


C̃t Õt+1
1 = βEt (1 − γ̃) Pt+1 + (1 − δ) ,
C̃t+1 K̃t

and optimal bond-holding yields

C̃t
1 = βEt r̃t .
C̃t+1

ECB Working Paper 1855, September 2015 22

 Combining (13) and (14), the short-run supply curve is

1/γ̃−1
Õt
γ̃Pt = 1/γ̃ 1/γ̃−1
. (15)
Z̃t K̃t−1

The region’s resource constraint is

Pt Õt = C̃t + I˜t + X̃t + B̃t − r̃t−1 B̃t−1 . (16)

Dominant oil exporter

The dominant producer’s economy has a structure symmetric to that of fringe producers, except
that there is a single oil firm. The firm produces oil, Ôt , according to

Ôt = Ẑt X̂tγ̂ K̂t−1

1−γ̂
, (17)

using an imported intermediate good X̂t and capital K̂t−1 . The dominant supplier’s technology
evolves deterministically according to Ẑt = Ẑ0 exp(g ẑ t) where g ẑ = g z̃ , that is, unlike the fringe, the
dominant supplier’s output is not directly affected by productivity shocks. Capital is accumulated
by purchasing Iˆt units of the investment good from the oil-importing region and the representative
household receives a stream of log utility from consumption Ĉt .
The substantial difference with the competitive fringe is that the dominant oil supplier has
market power: it is aware of the dependence of fringe oil supply, of oil demand, and of the
equilibrium oil price on its supply decision. The dominant supplier chooses a state-contingent plan
which maximizes the expected present discounted utility of its owners, subject to the demand by
the oil-importing region, the supply of competitive fringe producers, and clearing in the oil market.
Thus, the decision problem of the dominant oil producer is to
∞
max Eo β t log(Ĉt )
Ĉt ,B̂t ,X̂t ,K̂t ,Pt t=0

subject to oil demand (4), fringe oil supply (15), its production technology (17), the oil market
clearing condition
Ot = Ôt + Õt (18)

and its resource constraint

Pt Ôt = Ĉt + K̂t − (1 − δ)K̂t−1 + X̂t + B̂t − r̂t−1 B̂t−1 . (19)

This amounts to assuming that while Saudi Arabia acts as a monopolist supplier of its residual

ECB Working Paper 1855, September 2015 23

demand in the short-run, it lacks information regarding the long-run (typically secret) investment
plans of its competitors and the effect that its decisions may have on labor demand or capital
accumulation in the oil-importing region.
The solution to the above problem under commitment yields the following first-order conditions:

Ĉt
1 = βEt r̂t ,
Ĉt+1

γ̂(Pt + Λt ) = X̂t /Ôt ,


Ĉt Ôt+1
1 = βEt (1 − γ̂) (Pt+1 + Λt+1 ) + (1 − δ) ,
Ĉt+1 K̂t
where Λt is the Lagrange multiplier associated with equation (18)

Pt Ôt
Λt = −  . (20)
1 γ̃
O
η t
+ Õ
1−γ̃ t

Notice that equation (20) can be also expressed as

Ôt
Λt = −  ,
Ot Õt
Φ
Pt D
+ Pt S
Φ

γ̃
where ΦD ≡ − OPtt ∂Ot
∂Pt
= η1 is the short-run oil demand elasticity from (4) and ΦS ≡ ÕPt ∂∂PÕtt = 1−γ̃ is
t
the short-run fringe supply elasticity from (15).
The dominant oil supplier extracts a pure rent by picking the profit-maximizing point on the
residual demand curve, where marginal revenue equals his marginal cost.

Market clearing and balanced growth path

In equilibrium, all markets clear. There is financial autarky and thus

Bt = B̃t = B̂t = 0. (21)

Oil market clearing is

Ot = Ôt + Õt . (22)

ECB Working Paper 1855, September 2015 24

 And goods market clearing is

Yt = Ct + It + C̃t + I˜t + X̃t + Ĉt + Iˆt + X̂t . (23)

The model incorporates secular trends in the growth rate of final goods technology (Zt ), oil
production technology (Z̃t and Ẑt ), and oil efficiency (ν t ). Stationarity of Saudi Arabia’s market
share requires g z̃ = g ẑ . In a steady-state with balanced growth, the ratio Pt Ot /Yt must remain
stationary. Given g z̃ = g ẑ , since Ot grows at rate g z + g z̃ = g z + g ẑ , while Yt grows at rate g z , for
the ratio Pt Ot /Yt to remain stable, Pt should grow at rate −g z̃ .

Calibration
We employ the same calibration as in Nakov and Nuño (2014). The working frequency of the
model is monthly. The secular growth rate of technology of the oil-importing region is set to
exp(g z ) = 1.031/12 , consistent with an average world output growth rate of 3 percent per year for
the period from 1973 to 2009. Given this, the time preference parameter is set to β = 1.01−1/12
equivalent to an average real interest rate of 4 percent per year. Based on the stationary market
share of Saudi Arabia in the data, we impose equality between the growth rates of the dominant
oil producer and the fringe, g ẑ = g z̃ . The average growth rate of total oil production is 0.8 percent
per year in the data. In the model this must equal the sum of the growth rate of the inputs
of oil production, g z , and the growth rate of the oil production technology itself, g z̃ , implying a
value for exp(g z̃ ) = 0.9982 on a monthly basis. The latter implies that the real oil price must
grow at an annual rate of 2.2 percent, which is consistent with the average growth rate actually
observed in the data. Second, the three parameters governing the oil-importing region’s labor
disutility, technology, and capital depreciation are set to their typical values in the RBC literature:
the inverse Frisch elasticity is set to ω = 1, the labor share in the production of final goods is
set to α = 0.67; and the depreciation rate is set to δ = 1.101/12 − 1, consistent with 10% annual
P ∂O
depreciation of installed capital. Third, the price elasticity of oil demand in the model is O ∂P
= η1 .
We set η = 4, consistent with Kilian and Murphy’s (2014) estimates of the short-run price elasticity
of oil demand of about 0.25. Fourth, the parameters of the two oil production technologies and the
efficiency (Z̃0 , Ẑ0 , γ̃, γ̂, ν 0 ) are calibrated as follows: γ̃ = 0.4, Ẑ0 = 1.8740, γ̂ = 0.5, ν 0 = 0.0041.
These values allow the model to match an average “oil share in spending” of 5 percent of GDP,
as well as the following three averages for Saudi Arabia: a global market share of 12.3%, capacity
utilization of 75%, and a price mark-up of 25%.
Regarding the scenarios, the baseline scenario is calibrated so that shale oil production in the
Unites States increases from roughly zero in 2010 to 2.7 mb/d in 2014 and reaches a peak of 4
mb/d in 2018 in line with EIA projections. This scenario assumes that current law and regulations

ECB Working Paper 1855, September 2015 25

affecting the energy sector remain in place throughout the projection horizon. It also assumes
that the production of shale oil remains concentrated in the Unites States and does not extend
significantly to other countries before 2020. This implies setting the value of the fringe supply shock
εãt to 0.0025 from January 2010 until December 2012, then raising it to 0.0040 until December 2017
and reducing it to 0.0030 until December 2019.
Second, in the lower shale production scenario we consider the case in which technical or
environmental constraints limit US shale production which remains broadly flat at 4mb/d during
the 2014-2018 period. This implies setting the value of the fringe supply shock to 0.0025 from
January 2010 to December 2012, then raising it to 0.0040 until November 2013, reducing it to
0.0025 from December 2013 until December 2017 and reducing it again to 0.0030 until December
2019.
In the higher shale production scenario we assume that technological progress and economic
incentives push up US shale production, which reaches 6 mb/d by the end of the decade. This
implies setting the value of the fringe supply shock to 0.0025 from January 2010 to December 2012,
then raising it to 0.0040 until November 2013, increasing it again to 0.0064 from December 2013
until December 2017 and reducing it to 0.0030 until December 2019.
Finally, in order to analyze the 2014 drop in prices we modify the model in a number of
dimensions:

1. Supply shocks. We recalibrate the process log(ãt ) = ρã log(ãt−1 ) + εãt , with persistence
ρã = 0.8745 (which is equivalent to 0.2 on an annual basis) and a Gaussian innovation εãt
with mean zero and variance σ 2ã = 0.052 .

2. Demand shocks. We introduce a demand shock as a shock to the TFP of oil importers.
z z
We have now thatZt = Zt−1 e(g +ḡt ) , where ḡtz is a stochastic process

log(ḡtz ) = ρz log(ḡt−1
z
) + εzt ,

with persistence ρz = 0.9913 (which is equivalent to 0.9 on an annual basis) and a Gaussian
innovation εzt with mean zero and variance σ 2z = 0.012 .

3. Saudi Arabia shocks. In order to analyze the deviation of Saudi Arabia from its optimal
rule, we modify the objective function of the dominant oil producer
∞
max Eo β t ât log(Ĉt ),
Ĉt ,B̂t ,X̂t ,K̂t ,Pt t=0

by introducing a stationary AR(1) preference shock ât , log(ât ) = ρâ log(ât−1 ) + εât with
persistence ρâ = 0.8254 (which is equivalent to 0.1 on an annual basis) and a Gaussian

ECB Working Paper 1855, September 2015 26

 innovation εtâ with mean zero and variance σ 2â = 0.052 . This shock introduces a change in
the preferences of the dominant supplier and may be interpreted as a deviation of Saudi
Arabia from its profit-maximizing rule.

ECB Working Paper 1855, September 2015 27

 2012

14

12 Shale oil

10

6 Other oil

0
1985 1990 1995 2000 2005 2010 2015 2020

Figure 1: US total oil supply (millions of barrels per day). Source: EIA and authors’ calculations.

Tables and Figures

Table 1: Shale oil projections in 2014 and the size of their revisions with respect to the 2013 and
2012 projections.

Projections AEO 2014 Revision since AEO 2013 Revision since AEO 2012
(mb/d) (%) (%)
2014 4.07 62 343
2015 4.49 70 361
2016 4.67 72 355
2020 4.79 70 300

Source: EIA Annual Energy Outlook reports 2012, 2013 and 2014. Note that for consistency we label the table as
shale oil, although the numbers refer to tight oil projections in the EIA reports.

ECB Working Paper 1855, September 2015 28

 4

1
%

-1

Rest of non-OPEC
-2 US
Total non-OPEC

-3
1980 1985 1990 1995 2000 2005 2010 2015

Figure 2: US shale oil contribution to non-OPEC supply growth (annual, y-o-y growth rates).
Source: EIA. Note: the last datapoint referes to 2012; cut-off date 2014H1.

Shale oil production Oil price difference

6 -2

5
-4
USD per barrel

4
mbd

3 -6

2 baseline
-8
1 higher shale production
lower shale production
0 -10
2014 2015 2016 2017 2018 2014 2015 2016 2017 2018

World oil production difference Saudi Arabia production difference

3 -0.5

2.5 -1

2 -1.5
mbd

mbd

1.5 -2

1 -2.5

0.5 -3
2014 2015 2016 2017 2018 2014 2015 2016 2017 2018

Figure 3: Differential effect of different shale shale production scenarios. All the figures represent
differences with respect to the counterfactual scenario of no shale production.

ECB Working Paper 1855, September 2015 29

 Spare capacity
29.5

29

28.5
Baseline
higher shale production
28
% total capacity

lower shale production

27.5

27

26.5

26

25.5
2014 2014.5 2015 2015.5 2016 2016.5 2017 2017.5 2018

Figure 4: Evolution of the spare capacity of Saudi Arabia (share of total capacity).

Shale oil production Oil price difference

4.5 -5
USD per barrel

4 -6
mbd

3.5 -7

3 -8

2.5 -9
2014 2015 2016 2017 2018 2014 2015 2016 2017 2018

World oil production difference Saudi Arabia production difference

2 -1

1.5
-2
mbd

mbd

1
-3
0.5 baseline
low demand elasticity
0 -4
2014 2015 2016 2017 2018 2014 2015 2016 2017 2018

Figure 5: Differential effect of alternative elasticity assumptions. All the figures represent differ-
ences with respect to the counterfactual scenario of no shale production.

ECB Working Paper 1855, September 2015 30

 WTI prices World oil production
110 80

100

million barrels per day

79
USD per barrel

90

80 78

70
77
60

50 76
2014 2014.2 2014.4 2014.6 2014.8 2014 2014.2 2014.4 2014.6 2014.8

Saudi Arabia oil production US oil production

10 10
million barrels per day

million barrels per day

9.5 9.5

9 9

8.5 8.5

8 8
2014 2014.2 2014.4 2014.6 2014.8 2014 2014.2 2014.4 2014.6 2014.8

Figure 6: Evolution of oil prices and oil production in 2014. Source: EIA and Federal Reserve
Bank of St. Louis.

10

-5
m-o-m %

-10

-15

Demand shocks
-20 Supply shocks
Saudi Arabia shocks

-25
2014 2014.1 2014.2 2014.3 2014.4 2014.5 2014.6 2014.7 2014.8 2014.9

Figure 7: Contributions to the growth rate of oil prices. Source: EIA, Federal Reserve Bank of St.
Louis and authors’ calculations.

ECB Working Paper 1855, September 2015 31

 Oil exporters Oil importers

Canada 8 Brazil 8

Russia 7 US 7

Algeria 6 UK 6

Saudi Arabia 5 Euro area 5

Iraq 4 China 4

Venezuela 3 Turkey 3

Angola 2 Japan 2

Kuwait 1 India 1

0 10 20 30 40 50 60 0 1 2 3 4 5 6
Net oil exports (% of GDP) Net oil imports (% of GDP)

Figure 8: Net oil exports/imports for some of the major exporters/importers. Source: JODI
database and IMF. The value of exports/imports is computed for an average oil price of USD100.

ECB Working Paper 1855, September 2015 32

Acknowledgements
The views expressed in this manuscript are those of the authors and do not necessarily represent the views of the European
Commission, the Bank of Spain or the Eurosystem. The authors are very grateful to Anton Nakov, Livio Stracca, Frauke Skudelny, Ine
Van Robays, three anonymous referees, participants at the CAMP Workshop 2014 on "Commodity price dynamics and financialization"
in Oslo and seminar participants at the European Central Bank and the Bank of Spain for helpful comments and suggestions. All
remaining errors are ours.

Cristiana Belu Mănescu

European Commission, Brussels, Belgium;
e-mail: cristiana.manescu@ec.europa.eu

Galo Nuño
Banco de España, Madrid, Spain;
e-mail: galo.nuno@bde.es

© European Central Bank, 2015

Postal address 60640 Frankfurt am Main, Germany

Telephone +49 69 1344 0
Website www.ecb.europa.eu

All rights reserved. Any reproduction, publication and reprint in the form of a different publication, whether printed or produced
electronically, in whole or in part, is permitted only with the explicit written authorisation of the ECB or the authors.

This paper can be downloaded without charge from www.ecb.europa.eu, from the Social Science Research Network electronic library at
http://ssrn.com or from RePEc: Research Papers in Economics at https://ideas.repec.org/s/ecb/ecbwps.html.

Information on all of the papers published in the ECB Working Paper Series can be found on the ECB’s website,
http://www.ecb.europa.eu/pub/scientific/wps/date/html/index.en.html.

ISSN 1725-2806 (online)

ISBN 978-92-899-1668-4
DOI 10.2866/964931
EU catalogue No QB-AR-15-095-EN-N