Sie sind auf Seite 1von 14

The Coupled Cycles of Geopolitics and Oil Prices

MAHMOUD A. EL-GAMAL a and AMY MYERS JAFFE b

ab stract

We analyze the coupled cycles of Middle-East geopolitical violence and oil prices. Building on earlier work that shows that low oil prices are regularly followed by geopolitical strife, and that the latter is usually followed by higher oil prices, due to actual or feared disruption in oil supply, we focus in this paper on one partic- ular factor: Which geopolitical events are most likely to lead to sustained supply disruptions? Using discrete wavelet analysis of oil production at the country level, we find that military conflicts that destroy production installations or disrupt oil transportation networks are the most significant antecedents of sustained, long term, disruptions in oil supply; whereas nonviolent regime change, internal politi- cal strife, and low level geopolitical tensions have more limited sustained impact. We discuss a framework to analyze whether conflict-related disruptions to oil sup- ply could be endogenous to the oil cycle and offer some policy considerations for ameliorating that cycle’s impacts.

https://doi.org/10.5547/2160-5890.7.2.melg

f 1. INTRODUCTION g

The last few decades have been characterized by repeating patterns of global oil-price boom and bust cycles that have wrought negative social, geopolitical and economic consequences on a variety of nations, especially in the Middle East. It is widely accepted in the energy economics literature that the oil boom and bust cycle is connected with the global secular business cycle. In specific, oil demand fluctuations are linked to the expansion and contraction of gross do- mestic product (GDP). The price effects of demand fluctuations during business cycle booms and busts are amplified by the relatively lengthy time scale in the investment cycle for oil production. Periods of strong global economic growth lead naturally to increased demand for oil that cannot be immediately met from existing above-ground inventories and under-ground recoverable reserves that can translate quickly into oil production, as Barsky and Kilian (2004) have observed. In time, higher oil prices incentivize energy companies to search for more oil resources and to invest in new extraction technologies. Actual oil production responds after a time lag, historically as long as five to ten years. 1 During this period of economic expansion and high oil prices, the world economy becomes more susceptible to oil supply shocks, which,

1. The cycle time between first investment and first oil production has been shortened in recent years, with some new tech- niques reducing the oil investment cycle time to 2 years. The investment cycle time for shale oil and gas is far shorter, at 3 to 9 months.

a Rice University, Department of Economics, 6100 Main St., Houston, TX 77005.

b Corresponding author. David M. Rubenstein Senior Fellow for Energy and the Environment Director, Program on Energy Security and Climate Change, Council On Foreign Relations, 58 East 68th Street, New York, NY 10065. E-mail: ajaffe@cfr.org.

Economics of Energy & Environmental Policy, Vol. 7, No. 2. Copyright 2018 by the IAEE. All rights reserved.

2

Economics of Energy & Environmental Policy

in turn, are often coupled with economic recessionary trends, substitution to alternative fuels, and improved demand efficiency. The latter impacts plant the seeds for a new phase of the oil price cycle, as a glut results from increased supply and decelerating or declining demand. In turn, a prolonged period of cheap fuel contributes to the next upswing of the global business cycle, leading to higher demand growth, and so on. El-Gamal and Jaffe (2009) added two further components to this narrative of the cycle in oil prices, arguing that geopolitical events that are often treated as exogenous, as well as debt-driven financial boom and bust cycles, are all part of an endogenous and self-perpetu- ating meta-cycle. Of particular relevance to this paper, they argued that geopolitical events are endogenous to the cycle of oil prices and financial liquidity. Petrodollar recycling during periods of high oil prices takes many forms, including high military spending and buildups in major oil exporting countries. During periods of low oil prices, disgruntled populations and simmering contentious regional relations result in heightened geopolitical risk, which may result in the use of weapons accumulated during periods of plentiful petrodollars. Over the past five decades, it can be argued that this pattern of cyclical military buildups has increased geopolitical risk and thereby eventually contributes to the rise in oil prices during the upswing of the secular business cycle described above. Over the decades, during boom times, as exhibited in the years leading to 1973 and again after 2002, global economic growth and the related increases in oil demand strengthened the hands of oil producers, who reaped massive profits while underinvesting in oil supply capacity. As oil prices rose sharply, oil producer treasuries were filled with a sudden influx of capital that was not easily absorbed by their limited rentier economies. In the Middle East, excess revenues translated, in part, into massive arms purchases, which protect the ruling classes from both external threats and internal challenges, as discussed by Colgan (2013). Proxy groups also benefited, receiving lavish support from oil producing sponsors as oil revenues soared. As we will discuss below, the effect is procyclical, because fears of military conflict or terrorist threats create terror premiums in oil prices, due to the perceived risk of a supply disruption. Using a Global Vector Autoregression model with quarterly data from 1979 to 2017 for 70 countries, Abdel-Latif and El-Gamal (2017) have provided empirical evidence for this three- way coupling of the cycles of oil prices, financial liquidity, and geopolitical risk. In particular, one of their findings is that, other things equal, lower oil prices contribute to significantly higher geopolitical risk, and that higher geopolitical risk, in turn, results in significantly higher oil prices. This is the empirical starting point of our analysis in this paper. This paper aims to provide a framework for considering how geopolitical risk is endog- enous to the oil price cycle. The paper and the coupled cycle framework analysis it surveys contribute to an already extensive literature that seeks to explain turbulent oil price movements over the last several decades. Huntington (1994) provides a good survey of this literature which includes recursive simulation models and optimization models. Many of these models seek to predict long run oil prices by utilizing inputs for world economic growth and oil supply curves based on resource estimates combined with technical constraints that influence rates of devel- opment and depletion but as Huntington notes, “missed projecting oil prices” even during a period free of supply interruptions. More recently, Baumeister and Kilian (2015) contribute to the literature with an enhanced forecast combination approach that includes a measure of re- finers’ crude acquisition costs and changes in inventory levels in addition to measures of global economic activity and oil production. Authors extend these approaches by offering a theoreti-

Copyright © 2018 by the IAEE. All rights reserved.

The Coupled Cycles of Geopolitics and Oil Prices

3

cal framework to consider additional channels that can be used to demonstrate how these links between demand cycles and oil supply trends are also influenced endogenously by geopolitical events. Using Discrete Wavelet Transform analysis, authors explore how war-related events can create redirection in oil production trends that deviate from projected geological depletion curves. Our survey also explains a model for how to link passive and active military spending to fluctuating oil revenue cycles. Pulling these various approaches together, we offer a unique framework that suggests that oil price movements are not only connected to the rise and fall of demand related secular business cycles, but also to a more pernicious channel that posits the accumulation of armaments during expansionary economic and higher oil price periods that then increase the chances of military conflict during contractionary economic, low oil price environments, perpetuating a pattern of repeating crises. Authors plan to extend this frame- work with additional quantitative studies as part of the larger research agenda. Specifically, Abdel-Latif and El-Gamal will expand work using Global Vector Auto-regression modeling to include recent geopolitical events and additional observations about petrodollar liquidity while El-Gamal and Jaffe plan to extend discreet wavelet transform analysis to update earlier findings with a more extensive focus on war related events and to include data on U.S. shale resources. In presenting our proposed framework for the links between the oil-price cycle and geo- politics, it is useful to revisit the geopolitical trends during the latest phase of the oil boom and bust cycle between 2008 and 2017, as presented by the authors at the 2017 U.S. Association for Energy Economics Annual Meeting in Houston (Jaffe, 2017). Following the peak in oil prices to $147 a barrel and the subsequent Great Recession in 2008, crude prices plunged well below $50 per barrel. They began to recover again in 2011 after the Arab Spring amplified perceived risks of disruption in Middle East oil supply, especially as demonstrations spread to oil producing regions such as Libya and the oil rich Eastern Province of Saudi Arabia. Brent prices led the increase, bringing both Brent and WTI prices well above $100 per barrel, and the positive Brent-WTI spread remained significantly robust. High crude prices persisted de- spite mounting inventories, as measured by United States crude oil stocks excluding Strategic Petroleum Reserve -- which serves as a traditional measure of trends in global oil storage due to poor data on storage outside the United States, c.f. Kilian and Murphy (2014). The end of the Arab Spring, as marked by the ascent of President Abdel Fattah El-Sisi of Egypt, coincided with a moderation in oil prices. In March 2015, the decline in oil prices was stemmed, and partially reversed. That price change coincided with Saudi Arabia’s war effort in neighboring Yemen to oust the Houthi forces that had led a successful coup to take power during the previous month. Commentators noted at the time that the market’s reaction to the Yemen War was limited because potential disruption of oil flow through the Strait of Bab El-Mandab was deemed a remote possibility. However, as conflicts across the Middle East escalated at the end of 2017, the Brent-WTI price gap widened as both prices rose, again led by Brent. Brent prices ended the year at close to $67, up 17% for the year, following another increase of more than 50% in 2016. The latest price in- creases coincided with missile attacks on Saudi Arabia from Yemen (in the fall of 2017), which have escalated tensions between Riyadh and Tehran, who is reported to support the Houthi rebels. The oil price response was linked to fears that a geographic spread of the proxy wars between Iran and Saudi Arabia could potentially result in significant disruption of oil supply. In the meantime, military conflict over control of oil fields in Northern Iraq led to actual oil supply disruption during the autumn of 2017.

Copyright © 2018 by the IAEE. All rights reserved.

4

Economics of Energy & Environmental Policy

f 2. THE PETRODOLLAR-FUELED MIDDLE EAST ARMS RACE g

We focus in this paper on the effects of militarization of oil revenues, especially with re- gard to how it endogenously amplifies the oil-price cycle. In this regard, there are two primary factors that have contributed to massive arms buildup in Middle East oil-rentier states. The first is that arms purchases, primarily from the U.S. and Western-European countries, but also from China and other sources, have served as effective means of recycling the massive petro- dollar flows to major Middle East oil-exporting countries during periods of high oil prices. The second and closely related factor is harmonization of regional security arrangements with U.S. and NATO partners, which provides further non-pecuniary incentives for Western powers to encourage the military channel of petrodollar recycling. Other things equal, military buildups are ostensibly expected to protect the status quo for ruling elites in oil exporting countries against internal as well as external threats. Recognizing this potential to increase precautionary demand for its exports during a global slump, and building on the standard dynamic models reviewed in Deger and Sen (1995), El- Gamal (2016) formulated a dynamic model that distinguishes between passive military expen- diture (the acquisition of weapons during periods of high oil prices) and active military spend- ing (using the accumulated weapons during periods of low oil prices). He argued that major oil producers likely recognize that active military expenditure may be of value not only to secure their own sovereignty and resources, but also to increase their oil proceeds by boosting precautionary demand (in anticipation of potential supply disruption) and preventing prices from falling further. The formal model in that paper, therefore, explains why an oil exporting country might be incentivized to time its military activism during periods of price slump, in order to boost its revenues, thus converting its military capital into civilian capital. The same model also helps to explain another anomaly in the behavior of major oil exporters: excessive spending of resources (that could otherwise be earmarked for civilian capital) on building their military arsenals during periods of plentiful oil revenue. Of course, episodes of high oil prices eventually result in the global pursuit of new sources of conventional and unconventional oil supply, as well as improved efficiency of energy de- mand, which may also include switching from oil to alternative fuels. When the inevitable downturn in the cycle returns and oil revenues shrink, the resulting unemployment and re- duced rent redistributions to local populations and interest groups feed anger, just when the rentier states’ abilities to spend on security and population appeasement is most constrained by large fiscal deficits. During such episodes of increased geopolitical risks, the large arsenals accu- mulated during years of plentiful petrodollars magnify the risk of widespread violence and de- struction, as oil exporters are incentivized to use their weapons actively to address those threats. We have seen evidence of this pattern, most recently, during the expanded military activities of Saudi Arabia, Qatar, and the United Arab Emirates in conflicts that span various corners of the Middle East, from Yemen to Syria to Libya. Other direct and indirect military participants in these conflicts, most notably Russia and Iran, have also financed their military expenditures in large part from oil revenues. This increased military activity pared significantly price declines seen in 2014–2015 by enhancing the geopolitical risk premium in oil prices, which were also buttressed by coordinated efforts of OPEC and Russia to constrain production. A significant part of the Arab countries’ military equipment used in these conflicts was accumulated during the plentiful petrodollar years following the Iraq invasion, 2003–2007, as well as the period of Arab Spring uprisings 2011–2014. In this regard, military spending in Saudi Arabia has been particularly large, both in absolute and relative terms. According to

Copyright © 2018 by the IAEE. All rights reserved.

The Coupled Cycles of Geopolitics and Oil Prices

5

TABLE 1 Military spending 2016 (absolute and as % of GDP)

2016 Military Spending

As Percentage of GDP

United States

$611 billion

3.3%

China

$215 billion

1.9%

Russia

$69 billion

5.3%

Saudi Arabia

$64 billion

10.4%

Israel

$18 billion

5.8%

Iran

$13 billion

3.0%

Source: Stockholm International Peace Research Institute (SIPRI).

Stockholm International Peace Research Institute (SIPRI) data, Saudi Arabia’s military spend- ing in 2016 (at $64 billion) was fourth only to the U.S. (at $611 billion), China (at $215 billion), and Russia (at $69 billion). Table 1 summarizes Saudi Arabia’s military spending as a percentage of GDP compared to other major countries. Oil dealers’ fears that Middle East conflicts might spread or escalate further appeared jus- tified at the close of 2017. The festering tensions between Riyadh and Doha remained a hot spot that included a militarily imposed blockade of Qatar, and the Saudi-Yemen war has not only created a devastating humanitarian crisis but also increased the chances of more direct conflict between Saudi Arabia and Iran that could lead to bombardment of critical infrastruc- ture. Already, a ballistic missile launched from Yemen has made a near miss to both the Riyadh airport and a key royal palace. And indeed, Saudi Arabia stepped up security at its offshore oil fields that border Iran amid war rhetoric between Saudi and Iranian leaders. Arab separatists, widely rumored to be funded by Saudi Arabia, have sabotaged oil production and refining installations in the oil rich Khuzestan province of Iran. An oil pipeline to Bahrain was also sub- ject to recent sabotage. Iranian-Saudi Arabian geopolitical rivalry has also manifested itself in Lebanon where Saudi Arabia tried unsuccessfully to disrupt the country’s political leadership model, with an almost explicit goal of reducing the role of Iranian-supported Hezbollah. The festering conflict, in addition to more complex conflicts in Syria, Iraq and Lebanon, contrib- uted to popular uprisings in Iran at year end, as citizens expressed concerns about the excessive burden on the Iranian economy from these external interventions. In Iraq, a Kurdish initiative to vote on independence resulted in military action to contest possession of the Northern Iraqi regions’ major oil fields around Kirkuk. Those fields were wrested away from the Kurdish Regional Government (KRG) in a battle that involved Irani- an-backed militias fighting together with official Iraqi government forces. Since the hostilities that took place in the autumn of 2017, oil production from the Kirkuk region has been prof- fered to Iran for coordinated oil transport and export strategies by the Iraqi government. These events, coming on the heels of a major down cycle that began in 2014, have helped to raise oil prices and signal that markets are more sensitive again to geopolitical risk factors. This supply-risk factor is particularly important now that global economic growth is also re- turning, with a promise of accelerated oil demand, at least in the short to medium term. El-Gamal and Jaffe (2013) consider the oil price risk premium during the Arab Spring uprisings period, from 2011 to 2014. The authors specifically consider the context of geopo- litical risk premiums, ostensibly following then-feared regime changes that could disrupt oil supplies, much like the fall of the Shah of Iran in 1979. In this regard, El-Gamal and Jaffe (2013) find that the prevailing oil prices during that period were unreasonably high absent any

Copyright © 2018 by the IAEE. All rights reserved.

6

Economics of Energy & Environmental Policy

major geopolitical event such as outright war that destroys production infrastructure or ob- structs transport routes. They argue, based on careful analysis of Venezuela and other countries, summarized below, that regime change by itself has not resulted in significant oil production disruptions in any major oil-exporting countries. Therefore, they predicted that if no signifi- cant war would erupt in the aftermath and/or as a result of the Arab Spring uprisings, oil prices would have to fall precipitously within the medium term, as they have done in fact.

f 3. WHAT CAUSES ACTUAL OIL PRODUCTION DISRUPTION? g

3.1 Methodology

In this section, we review some of the main results in El-Gamal and Jaffe (2013), which used discrete wavelet transform analysis to study events associated with upward or downward jumps in oil production trajectories in major oil producers. Discrete wavelet transform anal- ysis is a method originally developed in signal processing and adapted recently to the study of economic and financial time series. The methodology allows econometricians more easily to identify with more exactitude when there is a significant break in pattern at a particular frequency of a data series. The advantage of this method for our study is its ability to allow us to isolate changes in the oil production trajectory at a particular time frame (for our case the medium term, defined as 3 to 5 years), by subtracting more transitory production fluctuations (for example, deviations due to short-lived weather events, worker strikes, or technical glitches) that disappear in a shorter time frame and do not signal significant changes in a long range trend. Percival and Walden (2000) offer an excellent exposition of the use of wavelet methods in time series analysis, and Gencay et al. (2002) offer a summary of how those methods have been used in economic and financial analyses. El-Gamal and Jaffe (2013) use discrete wavelet transform analysis to extend the use of Hubbert curve oil production trajectories to analyze events that cause significant deviations from the technically-oriented predicted oil production path. The Hubbert curve is a geolog- ically driven curve fitting technique developed in 1956 by geologist Marion King Hubbert (1962). The technique postulated that production of a finite resource, when viewed over time, will resemble an inverted U, or a bell-curve (Deffeyes 2001). The theoretical basis of the bell curve follows from the technical limits of exploitation, where the estimated parameters of the curve determine the rate of ascent and descent before and after the peak. The Hubbert curve was widely accepted as a method of predicting long term oil production pathways but led many analysts to predict incorrectly a peak in oil supply would take place in the 2010s. El-Gamal and Jaffe (2013) extrapolate from this technique to show that geopolitical events and technology disruptions can and have altered the production curve for resources over time. The authors’ methodology offers explanatory power for changes in long term production that cannot be explained solely by geology-oriented Hubbert style bell curves for a particular resource. Rather the El-Gamal and Jaffe (2013) wavelet analysis method tests whether geopolitical events can contribute additional analytical elucidation. Earlier literature, such as Toft (2011), has linked disruptions to oil production and periods of interstate conflict. Toft (2011) found that oil supply disruptions resulted in roughly half of these instances. Toft’s approach does not consider the duration of outages or differences in the relative impacts of those disruptions on oil markets. El-Gamal and Jaffe (2013) pursued a different strategy that is primarily data driven. Instead of testing for disruptions only at specific moments, they used the discrete wavelet transform to examine monthly oil production ab-

Copyright © 2018 by the IAEE. All rights reserved.

The Coupled Cycles of Geopolitics and Oil Prices

7

stracting from short-term disruptions (defined as any change in production that does not per- sist for 3 or more years) that may be caused by weather, technical and other kinds of political conditions. Because El-Gamal and Jaffe (2013)’s primary focus was on medium-term effects, corresponding to time horizons of three to five years, they tested for any breaks in smoothed production at that frequency using monthly data for production in all countries that had ever produced more than one million barrels per day since the early 1970s. These medium term effects characterize the kinds of lasting oil supply disruptions that we theorize are part of the coupled oil price cycles described above. El-Gamal and Jaffe (2013) have thus investigated disruptions in oil production for each major oil-producing country in terms of potential jumps from one Hubbert curve to a higher one (usually due to technical innovations spurred by high prices) or lower one (usually as a result of war, as shown below). Using the test methodology of Wang (1995), they tested for significant alteration of Hubbert curve trajectories in various countries’ oil production using a wavelet analysis that allows them to focus on significant changes at moderate frequencies (corresponding to the medium term 3–5 years period), abstracting from high frequency moves that are caused by transient weather, technical, political, or financial drivers. The Hubbert curve came to be seen as representing the expected or normal trajectory of any given country’s oil production over time after Marion King Hubbert’s 1956 curve fitting approach correctly predicted a medium-term peak in U.S. oil production by 1970, c.f. Hub- bert (1962) and Cleveland and Kaufmann (1991). Ensuing hysteria about indefinitely rising oil prices resurfaced in the mid-2000s, spurred by what was then dubbed the “peak oil” hypoth- esis, predicting that Middle East oil production would also peak in the near future based on Hubbert curve estimates. Of course, we now know that both demand and supply responded to high-price pressures in the early 1980s to prevent any permanent peaking in oil supply—with demand declining due to recession and improved efficiency, and supply increasing through unexpected technology improvements used in North Sea and other areas. The same pattern was likewise repeated following the great recession and eventual collapse in oil prices after the Arab Spring ended. The narrative of “peaking” has changed so much in recent years that most analysts today debate the projected date for peak oil-demand rather than supply. Our wavelet analysis methods allow for consideration of geopolitical events and technological disruptions in studying the oil supply trajectories of major oil producing countries.

3.2 Application (1): Venezuela oil production

We begin with Figure 1, which illustrates how the discrete wavelet transform method works using Venezuela’s monthly oil production data. The actual production level, in thou- sands of barrels per day, is shown in the top left panel of the Figure. The discrete wavelet transform decomposes this series into the sum of multiple series, in our case six, depending on the time horizon of primary interest. The bottom right panel of Figure 1 shows very fast fluctuations in production, which were reversed within a time horizon of two to four months. Once this component of the series is subtracted from the monthly data, all remaining fluc- tuations have a time horizon of four or more months. Next, we extract the second wavelet detail, shown in the bottom left panel of the Figure, which corresponds to fluctuations that are reversed within a four to eight month time frame. Notice that the effect of the workers’ strike in 2003, for example, appears at both details at the two-to-four and four-to-eight month lev- els. Having subtracted both wavelet details 1 and 2 from the series of Venezuelan production, the remaining series only has fluctuations that are reversed over the course of eight or more

Copyright © 2018 by the IAEE. All rights reserved.

8

Economics of Energy & Environmental Policy

8 Economics of Energy & Environmental Policy FIGURE 1 Discrete Wavelet Transform Decomposition of Venezuelan Oil

FIGURE 1 Discrete Wavelet Transform Decomposition of Venezuelan Oil Production.

Source: Own illustration.

months. We continue with this procedure for level 3 (which captures fluctuations that reverse in eight-to-sixteen months), level 4 (which captures fluctuations that reverse in sixteen-to-thir- ty-two months), and finally level 5 (which captures fluctuations that reverse in thirty-two-to- sixty-four months), which captures our target “medium term” of three to five years. Once we have subtracted discrete wavelet details 1 through 5 from the original series of Venezuelan oil production, what remains is the “smooth” curve at level 5 (left panel, second from the top in Figure 1), which has only fluctuations that are reversed in 64 months or longer. Adding detail level 5 back to the smooth curve (which is the same as the “smooth” level 4, but we show detail 5 separately for emphasis), we have obtained our object of study: The original series of Venezuelan production less all fluctuations that were reversed within 32 months or less. For each country’s series of monthly production, we perform the same exercise to ob- tain the corresponding object of study that contains only fluctuations that are reversed in 32 months or longer. Only significant alterations in the trajectory of this curve will be considered

Copyright © 2018 by the IAEE. All rights reserved.

The Coupled Cycles of Geopolitics and Oil Prices

9

The Coupled Cycles of Geopolitics and Oil Prices 9 FIGURE 2 Venezuelan Monthly Crude Production (thin

FIGURE 2 Venezuelan Monthly Crude Production (thin solid curve), DWT Smoothed production (bold solid curve), and number of active rigs (grey dashed curve, right hand scale).

Source: Own illustration. Vertical bold dot-dash lines indicate statistically significant breaks in wavelet-smoothed series.

a meaningful disruption in the medium term (or longer). Hence, we use the statistical method

of Wang (1995) to test for kinks or other sudden but persistent changes in this trajectory. In fact, as we can see in Figure 2, the only significant cusp or jump in Venezuelan production was observed around the time of the first price surge in 1973, when the country ramped up in- vestment in production infrastructure quite substantially. A second wave of investment, as mea- sured by the number of active rigs, also showed in Figure 2, took place between 1988 and 1998 during Venezuela’s apertura campaign to entice foreign investors to the country, but this period did not result in a detected significant cusp or jump to a new Hubbert curve trajectory. Most interestingly for our analysis, regime change at the onset of Chavez’s assention to power in 1998 was similarly not associated with a significant downward jump or cusp to a lower Hubbert curve. Investment clearly declined, preventing Venezuela from reaching its full production potential, but this was merely detected by the medium-term smoothed production curve as continuation along a less ambitious Hubbert curve. In other words, regime change by itself (1998), as well as events as major as the workers’ strike in 2003, were not associated with significant breaks in the medium-term Hubbert curve estimated by our discrete wavelet transform.

3.3 Application (2): Monthly crude production in Iraq and Iran

By contrast, as we can see clearly in Figure 3, very significant output reductions (jumps to

a lower Hubbert curve) in Iraq were detected, coinciding with the onset and late stages of the

Copyright © 2018 by the IAEE. All rights reserved.

10

Economics of Energy & Environmental Policy

10 Economics of Energy & Environmental Policy FIGURE 3 Iraqi Monthly Crude Production (thin solid curve),

FIGURE 3 Iraqi Monthly Crude Production (thin solid curve), DWT Smoothed production (bold solid curve), and number of active rigs (grey dashed curve, right hand scale).

Source: Own illustration. Vertical bold dot-dash lines indicate statistically significant breaks in wavelet-smoothed series.

country’s 1980 to 1988 war with Iran. During that war, oil facilities were intentionally target- ted and destroyed in air bombings, and this, indeed, led to significant downward discountinu- ities in Iraq’s oil production profile which should have followed a Hubbert curve very similar to Saudi Arabia’s, given their similar prolific geologies. Likewise, Iran’s long-term production exhibits a significant downward break following the Iranian Revolution of 1979. In this regard, because Iran’s jump to a lower Hubbert curve has persisted throughout the sample, dating the exact moment of the jump is more difficult, be- cause the effects of the Iran-Iraq war of 1980–1988 as well as the Tanker war (1986–7) seem to be subsumed by the decline in production that began prior to 1979. These results support our thesis that geopolitical conflicts contribute to significant discontinuities in oil supply trends from major producing countries and thereby can effectively contribute to rising oil price cycle dynamics that are more structural and less fleeting when they happen in periods of high de- mand during periods of global economic expansion.

3.4 Application (3): UK monthly crude production

Of course, jumps from one Hubbert curve trajectory to another may also occur in the opposite (positive) direction, especially through technical innovation as response to prolonged periods of high oil prices. In this regard, U.S. production has exhibited a remarkable pattern in the aftermath of high oil prices during the mid-2000s. The latter was made possible by a

Copyright © 2018 by the IAEE. All rights reserved.

The Coupled Cycles of Geopolitics and Oil Prices

11

The Coupled Cycles of Geopolitics and Oil Prices 11 FIGURE 4 Iranian Monthly Crude Production (thin

FIGURE 4 Iranian Monthly Crude Production (thin solid curve), DWT Smoothed production (bold solid curve), and number of active rigs (grey dashed curve, right hand scale).

Source: Own illustration. Vertical bold dot-dash lines indicate statistically significant breaks in wavelet-smoothed series.

combination of healthy global GDP growth, boosting demand during a period that witnessed supply disruptions from conflicts in Iraq, Libya, and the Delta region of Nigeria. As a result of the ensuing shale revolution, the U.S. Energy Information Administration is currently project- ing that U.S. oil production will reach 10 million barrels per day in 2018, up 780,000 barrels per day from 2016. This is a remarkable doubling of U.S. production, which averaged roughly 5 million barrels per day in 2006. Some analysts predict even faster growth in U.S. oil produc- tion. Analysts Cornerstone Macro, for example, project in their November 2017 private report “Introducing Shale Model 1.0,” that U.S. tight oil production could rise by 960,000 barrels in 2018 (excluding natural gas liquids), and an additional 770,000 barrels per day in 2019, if prices hold around $60. This would bring total U.S. crude oil production above 11 million barrels per day over the next two years. Production growth is projected to be even larger in a $65 oil price environment. A similar dramatic technology-driven supply response was observed in an earlier cycle pe- riod after the dramatic rise of oil prices during 1973–1979. Our medium-term discrete wavelet analysis of U.K. production is shown in Figure 5, highlighting a clear jump to a higher Hubbert curve around 1979, which corresponded to North Sea production enabled by improvements in deep-water drilling that reached its full potential in the early 1980s. Even though the first exploration and production licenses were granted by the U.K. and Norway in 1964, the incen- tive to develop new technology for deep water drilling was only present after oil prices increased dramatically during the 1970s, in a market development that mirrors today’s shale boom.

Copyright © 2018 by the IAEE. All rights reserved.

12

Economics of Energy & Environmental Policy

12 Economics of Energy & Environmental Policy FIGURE 5 UK Monthly Crude Production (thin solid curve),

FIGURE 5 UK Monthly Crude Production (thin solid curve), DWT Smoothed production (bold solid curve), and number of active rigs (grey dashed curve, right hand scale).

Source: Own illustration. Vertical bold dot-dash lines indicate statistically significant breaks in wavelet-smoothed series.

f 4. CONCLUDING REMARKS g

Authors have introduced the idea that the oil boom and bust cycle is coupled not only with the secular global business cycle but also endogenously with geopolitical cycles in mil- itary spending followed by military adventurism. Once again, high oil prices from 2011 to 2014 were accompanied by sharp increases in military expenditure in oil producing countries across the Middle East, which eventually has translated into escalating conflicts once oil prices receded (2014–2016). At the end of 2017, markets were caught between upward oil price momentum, driven by risk premiums related to these escalating conflicts, and the potential for higher competition among oil producers as technology advances has driven new oil supplies. As in past cycles, rising oil prices invite efficiency gains and wider adoption of alternative fuels. In light of rising geopolitical risks across the Middle East, including key populations that are restive in Iran and Saudi Arabia, we revisit our prior work on regime change and oil supply disruptions. We survey our findings on potential medium-term production disruptions that would affect global supply adversely over periods of three to five years. We suggest that a longer lasting structural shift back to upward cycles could be expected if the proxy wars between Iran and Saudi Arabia in Yemen and the Levant intensify into overt war similar in magnitude to the Iran-Iraq war of the 1980s, wherein oil-production facilities were intentionally targeted and bombarded.

Copyright © 2018 by the IAEE. All rights reserved.

The Coupled Cycles of Geopolitics and Oil Prices

13

Of course, such a scenario would be catastrophic both globally, as it might cause oil prices to jump to high triple-digits, causing a deep global recession, and to the major exporters themselves, as the quantity of oil they sell drops dramatically, reducing the total receipts even as the price rises. The long-term effects on major Middle East oil exporters are likely to be more catastrophic during this episode, because past cycles have left markets better prepared to respond to oil price changes. New shale supplies may eliminate the need for Middle East oil more permanently while the switch to alternative fuels and more efficient energy consumption could be accelerated by higher prices. The latter holds the key to trying to ameliorate future cycles. Western governments need to stay the course on efficiency and energy substitution programs that can ease the negative impacts of the oil-price cycle on the global economy and on international security. Governments should also consider creating expedited leasing and permitting procedures for new domestic oil production in areas that already have extensive production and transport infrastructure to speed the time lag from investment response to first production in the aftermath of major international supply outages. In addition, rather than increase arms sales as oil rentier states seek to externalize their problems, major economies such as the United States, China, Japan and Europe multilaterally and through international agencies should encourage the acceleration of economic reforms in oil producing countries to improve their resiliency to the commodity cycle, while stepping up energy efficiency programs that can prevent demand driven amplifications of the cycle. Forty years of military buildups have failed to bring peace and economic prosperity to the Mid- dle East region. New digital technologies are enabling simultaneously revolutionary improve- ments in oil and gas production outside the Middle East and dramatic cost improvements in alternative energy. Governments need to support these efforts to minimize the impact of the endogenous geopolitical aspects to the oil cycle. However, Western governments should simultaneously consider the significant long-term difficulties that Gulf oil economies will face, and shift emphasis in trade and investment to goods and services that will ensure a long-term transition away from dependence on oil revenue. All countries have an interest in fashioning a soft landing for the region’s oil states. While it is very unlikely that the Middle East major oil exporters will intentionally escalate their regional proxy wars in a manner that leads to destruction of oil facilities, the oil industries of Syria and Yemen have already been decimated by recent geopolitical conflicts. Violence has also broken out in Northern Iraq, causing disrupted production and exports from the Kirkuk region. The nature of war is unpredictable and irrational, hence explaining the return of a geopolitical risk premium in Brent prices, which also pulls WTI prices higher. Unlike the Arab Spring, which was short lived, the status-quo of low to medium level violence in the Middle East can persist for years, or even decades. Therefore, unlike our prediction in 2012 that if outright war were not to materialize, oil prices would have to plunge in the medium term of three to five years, oil markets could remain under geopolitical pressure for the coming year or more until the inevitable acceleration in new supply and substitution gains critical mass again. While a sudden regime change in 2018 might not push oil back to a crisis situation, the market is now poised such that a war event that involves destruction of facilities in a major producing country could precipitate a new oil crisis and related economic tremors. In light of this self-perpetuating cycle, industrialized governments would benefit from revisiting trigger mechanisms for coordinated use of strategic oil stockpiles.

Copyright © 2018 by the IAEE. All rights reserved.

14

Economics of Energy & Environmental Policy

References

Abdel-Latif, H. and M. El-Gamal (2017). “Financial Liquidity, Geopolitics, and Oil Prices.” manuscript, presented at Swansea University, November 2017. https://doi.org/10.2139/ssrn.3095787. Barsky, R. and L. Kilian (2004). “Oil and the Macroeconomy since the 1970s.” Journal of Economic Perspectives 18(4): 115–34. https://doi.org/10.1257/0895330042632708. Baumeister, C. and L. Kilian (2015). “Forecasting the Real Price of Oil in a Changing World: A Forecast Combi- nation Approach” Journal of Business and Economic Statistics 33(3): 238–252. https://doi.org/10.1080/0735001

5.2014.949342.

Cleveland, C. and R. Kaufmann (1991). “Forecasting Ultimate Oil Recovery and its Rate of Production: Incorpo- rating Economic Forces in Models of M. King Hubbert.” The Energy Journal 12: 17–46. https://doi.org/10.5547/

ISSN0195-6574-EJ-Vol12-No2-3.

Colgan, J. (2013). “Fueling The Fire: Pathways from Oil to War,” International Security 38(2): 147–180. https://

doi.org/10.1162/ISEC_a_00135.

Deffeyes, K. (2001). Hubbert’s Peak: The Impending World Oil Shortage. Princeton, NJ: Princeton University Press. Deger, S. and S. Sen (1995). “Military Expenditure and Developing Countries,” in Handbook of Defense Economics, Vol. 1, edited by K. Hartley and T. Sandler, 275–307. Amsterdam: Elsevier. El-Gamal, M. (2016). “Winter Is Coming: Controlled Conflicts and The Oil-Price Geopolitical-Risk Premium.” Texas: Center for Energy Studies, Baker Institute for Public Policy. El-Gamal, M. and A. Jaffe (2009). Oil, Dollars, Debt, and Crises: The Global Curse of Black Gold. New York: Cam- bridge University Press. https://doi.org/10.1017/CBO9780511812323. ——— (2013). “Oil Demand, Supply, and Medium-Term Price Prospects.” California: Institute for Transportation Studies, University of California, Davis. Gencay, R., F. Selcuk, and B. Whitcher (2002). An Introduction to Wavelets and Other Filtering Methods in Finance and Economics. San Diego: Academic Press. Hubbert, M. (1962). Energy Resources: A Report to the Committee on Natural Resources, National Academy of Scienc- es Pub. 1000-D. Washington, D.C.: The National Academies Press. Huntington, Hillard G. (1994). “Oil Price Forecasting in the 1980s: What went wrong?” The Energy Journal 15(2):

1–22. https://doi.org/10.5547/ISSN0195-6574-EJ-Vol15-No2-1. Jaffe, A. (2017). “Geopolitics and the Oil Cycle.” Plenary presentation at the session “The U.S. as an International Oil and Natural Gas Market Maker—Markets, Geopolitics and U.S. Energy Policy.” Houston, TX, U.S.: Asso- ciation for Energy Economics Annual Meeting. Kilian, L., and D. Murphy (2014). “The Role of Inventories and Speculative Trading in the Global Market for Crude Oil.” Journal of Applied Econometrics 29: 454–78. https://doi.org/10.1002/jae.2322. Percival, D., and A. Walden (2000). Wavelet Methods for Time Series Analysis. New York: Cambridge University Press. https://doi.org/10.1017/CBO9780511841040. Toft, P. (2011). “Intrastate Conflict in Oil Producing States: A Threat to Global Oil Supply?” Energy Policy 39(11):

7265–74. https://doi.org/10.1016/j.enpol.2011.08.048. Wang, Y. (1995). “Jump and Sharp Cusp Detection by Wavelets.” Biometrika 82: 385–397. https://doi.org/10.1093/

biomet/82.2.385.

Copyright © 2018 by the IAEE. All rights reserved.