Explosive oil prices, oil eld development and the role of speculation

 
 
Explosive oil prices, oil field development and the
                  role of speculation
                                  Marc Gronwald∗

                                    October 2012



                                       Abstract

           This paper, first, investigates whether or not oil prices are char-
       acterized by transitory phases of explosiveness. The application of a
       recently proposed recursive unit root test shows that phases with explo-
       sive behavior were present 1990/1991, 2005/2006 as well as 2007/2008.
       Second, it is shown that these transitory oil price hikes affect the de-
       cision on when to develop an oil field: the profitability of the project
       increases and, thus, the fields are developed earlier. As this leads to
       an increase in current oil production, current carbon emissions also
       increase. Finally, the underlying causes of oil price hikes are discussed.
       It is important to note that the effect described here holds irrespec-
       tive of whether oil price hikes are fundamentally driven or caused by
       “speculation”. The ongoing debate on oil market speculation, however,
       yields at least some evidence of speculative influences. Thus, oil mar-
       ket speculation seems indeed to be a harmful, but in a manner which
       so far appears to have been overlooked.
       Keywords: Oil prices, Explosiveness, Uncertainty, Real options, Cli-
       mate change, Speculation
       JEL-Classification: C12, C58, D81, Q30


   ∗
    ifo Institute - Leibniz Institute for Economic Research and CESifo, Poschinger Strasse
5, 81679 Munich, email: gronwald@ifo.de. The author is grateful for the hospitality of the
University of California, Berkeley as well as Resources for the Future, Washington, DC,
while working on the paper. Financial support by the Fritz Thyssen Foundation for these
research visits is gratefully acknowledged.
1   Introduction

The global oil market is heavily disrupted on a regular basis. The two
oil crises as well as the OPEC collapse are good examples in this regard.
Each of these disruptions attracted considerable attention - from the general
public as well as from academia. Most recently, the 2007/2008 oil price hike
joined this list of incidents: in July 2008 oil prices reached the record high
of more than 140 USD per barrel. Subsequently, a heated debate emerged
regarding the causes and the consequences of this oil shock. Whether or not
it contributed to the recession observed after 2008 ([4]) as well as whether
“speculative” behavior or fundamentals explain it ([2]) are among the themes
under discussion. What is more, there is also a vast literature which is
concerned with characterizing the idiosyncratic behavior of oil prices - in
both the long-run and the short-run. With regard to the former, it is still
subject of debate whether deterministic ([13]; [9]) or stochastic trends ([14])
are present in oil prices. On the short-run behavior front, a number of recent
studies deal with jumps in oil prices ([5]; [3]). Among the most prominent
candidates for explaining this behavior are low short-run demand and supply
elasticities as well as political influences, see e.g. [15]. Finally, the issue of
the optimal taxation of exhaustible resources is one of the main fields in
the area of resource economics. The fact that crude oil is a fossil resource
and, thus, the usage of which is directly linked to the problem of climate
change makes this research even more important. In a nutshell, it does not
need much emphasis that these any research effort invested in understanding
crude oil prices is more than justified.
    Although considerable efforts have already been invested, it appears to
be the case that the currently existing models are not able to capture oil
price hikes like the one observed 2007/2008. Consulting Figure 1’s plot of
daily oil prices from 1983 to 2011 shows that initially more of a horizontal
movement is present. After a structural break in the late 1990s, however, oil
prices started to increase. This upward movement still seems to be intact and
currently oil prices are at about 100 USD per barrel. This visual inspection
certainly does not allow one to discard either of the two existing long-run


                                       2
160

              140

              120

              100

               80

               60

               40

               20

                0
                    1985   1990    1995         2000   2005   2010

                                          Oil prices




                             Figure 1: Oil prices


descriptions. However, there are temporary phases in which deviations from
the long-run behavior are apparent: certainly during the 2007/2008 oil price
hike, but also 1990/1991 and in the early 2000s.
   Motivated by this observation, this paper deals with three issues. First,
the idiosyncratic behavior of oil prices by is modeled by testing whether or
not oil prices exhibit temporary phases of explosive behavior. The appli-
cation of a recently proposed recursive application of unit root tests shows
that oil prices are indeed marked by temporary phases of explosiveness:
1990/1991, 2005/2006 as well as 2007/2008. This empirical feature has so
far not been discussed in the oil price behavior literature.
   Second, the paper adds to the literature on the consequences of oil price
hikes. However, rather than focussing on macroeconomic impacts, it is con-
cerned with consequences with respect to the decision when to extract the
fossil resource oil. Phases with transitory high oil prices like those identified
in this paper lead to an increase in the value of undeveloped oil fields. This,
in turn, is a strong incentive to begin oil extraction earlier. As this implies
that current carbon emissions increase, the current atmospheric concentra-


                                          3
tion of carbon also increases - this has obviously negative impacts for climate
change.
   Third, the underlying causes of transitory oil price hikes are discussed.
It is important to note that the effect on the oil field development decision
occurs regardless of this cause. This implies in particular that it is not
relevant whether the oil price hike is fundamentally driven or caused by
“speculation” in oil markets. The heated debate on this issue - see [11],
[4], [7] as well as [2] - however, yields at least some evidence that there are
speculative influences in the oil market. Thus, there appear to be harmful
consequences of speculation which so far seem to have been overlooked.
   The empirical strategy of this paper consists of the forward recursive
application of an augmented Dickey-Fuller unit root test. In each step, the
null of a unit root is tested against the alternative of explosiveness. This
procedure yields a sequence of test statistics which is used to identify phases
with explosive and non-explosive behavior. Monthly as well as daily oil price
data spanning from 1986-2011 and 1982-2010, respectively, is used in this
study. This procedure is derived from a test for periodically collapsing
bubbles recently proposed by Phillips et al. (2011) as further-development
of cointegration-based tests for the existence of bubbles.
   The investigation of the consequences of the observed oil price behavior
on the resource extraction decision is based a reconsideration of [8]’s study
on the relationship between transitory oil price hikes and the development
of an offshore oil field. This real option application extends a model on
investment under uncertain oil prices discussed in [12]. In the basic version
of the model it is assumed that oil prices follow a continuous Brownian mo-
tion with drift. Motivated by the oil price hike associated with the Gulf
war 1990/91, [8] augment the model by using a jump component and study
the effect of non-permanent oil price increases. They find that also this
type of increase leads to development of oil fields, but it needs to be four
times as large as a permanent increase in order to have the same effect. [8],
however, assume that transitory oil price hikes are associated with wars, oc-
cur relatively rarely (every 20 years) and are relatively short (3-6 months).
This paper’s empirical findings, however, suggest that they can have var-

                                      4
ious causes, occur much more often and can last longer. In consequence,
investment is more responsive to transitory hikes. Thus, there is an increase
in current production of crude oil. The consequences have been outlined
above.
    The driving force behind this effect is an increase in the value of an
undeveloped oil field. Thus, it emerges regardless of whether or not the
observed oil price increase is fundamentally driven. Motivated by the 2008
oil price hike, various papers attempt to shed light on its underlying causes.
One important issue is the apparent “financialization” of oil futures markets,
see e.g. [7]. Papers such as [11] provide evidence that capital inflows in the
oil futures market affected also oil spot prices and, thus, contributed to the
2008 oil price hike. This finding, however, is not undisputed, see e.g. [4] or
[2]. This latter paper, however, provides strong evidence that the 1990/91
oil price hike is partly driven by speculative demand. In a nutshell, there
is at least some evidence of speculative influences in the oil market, but
additional research is required.
    The remainder of the paper is organized as follows: Section 2 outlines
the empirical method employed in this paper and Section 3 presents the
empirical results. Section 4 discusses the consequences of the empirical
behavior of oil prices; Section 5 provides some comments on the recent
debate on oil price bubbles. Section 6 finally concludes.


2   Testing for explosiveness

The statistical properties of monthly as well as daily oil prices are inves-
tigated here using a forward recursive application of a augmented Dicker
Fuller unit root test. The null of a unit root is tested against the alternative
of an explosive root. Thus, the following equation is estimated:

                             ∑
                             J
         xt = µx + δxt−1 +         ϕj ∆xt−j + ϵx,t ,   ϵx,t ∼ NID(0, σx2 ).   (1)
                             j=1

    The hypothesis H0 : δ = 1 is tested against the alternative H1 : δ > 1.
Initially, a subset of the sample with τ0 = nr0 observations is used. In each


                                         5
subsequent regression this subset is supplemented by successive observations
giving a sample of size τ = nr for r0 ≤ r ≤ 1. This procedure yields a
sequence of t-statistics with corresponding p-values. These sequences are
used in order to identify origination r̂e and collapse dates r̂f of explosive
behavior in the data:

                      r̂e = infs≥r0 {s : ADFs > cvadf
                                                  βn (s)}




                      r̂f = infs≥r̂e {s : ADFs < cvadf
                                                   βn (s)}


     This procedure has been derived from the test for periodically collapsing
bubbles recently proposed by [6].
     This paper uses nominal monthly oil prices spanning from 1982-2010
as well as nominal daily oil prices from 1986-2010 (WTI) in order to test
for explosiveness in oil prices. The following section presents the empirical
results.


3    Results

This section presents the results obtained from applying the test procedure
outlined above on monthly as well as daily oil prices. Initially, the results
for monthly data are considered. Figure 2 displays oil prices as well as
the sequence of p-values. p-values below 5 % indicate rejection of the null
hypothesis. As explained above, for periods in which the null of a unit root is
rejected, oil prices are said to exhibit explosive behavior. This is the case in
particular 2005-2006 as well as 2007-2008, but also the price hike associated
with Gulf war at the end of 1990 is classified as explosive. While the earlier
phase is of relative short duration, the two later ones are about year-long.
The analysis of daily oil prices generally confirms these results, see Figure
3.
     Having presented this paper’s empirical results, now the consequences of
explosive oil prices are discussed.



                                       6
160


                                                                                        120


                                                                                        80


                 1.0                                                                    40

                 0.8                                                                    0

                 0.6

                 0.4

                 0.2

                 0.0
                           1990         1995           2000          2005        2010

                                  Monthly oil prices          p-values      5%




                       Figure 2: Explosiveness of monthly oil prices


4       Explosive oil prices and oil field development

When it comes to studying the economic consequences of oil price shocks,
predominantly consequences for macroeconomic activity are considered. This
paper, however, considers the decision when to optimally extract the fossil
resource oil.
        The vehicle for studying this issue is [8]’s analysis of the relationship
between transitory oil price hikes and oil field development. The point of
departure of their analysis is a situation in which a firm discovers an offshore
oil field of a certain size.1 Developing this oil field involves investing a lump
sum irreversible charge. Once the field is developed it is assumed that the
extraction rate is geologically determined. The firm decides when to develop
the field. A real option model is used in order to analyze the influence of
uncertain oil prices on this irreversible investment decision. Initially, oil
    1
    For a detailed exposition of the model, the reader is referred to the original publication.
For a general introduction to the investment decision under uncertainty, see [12].




                                                       7
160

                                                                                  120

                                                                                  80

                                                                                  40
             1.0
                                                                                  0
             0.8

             0.6

             0.4

             0.2

             0.0
                     1990        1995           2000           2005        2010

                             Daily oil prices       p-values          5%




                   Figure 3: Explosiveness of daily oil prices


prices are assumed to follow a continuous Brownian motion with drift:

                              dPt = αPt dt + σPt dBt

Based on this assumption, [8] show that the oil field is developed once oil
prices exceeds a certain trigger price. This trigger price is higher than the
one in absence of uncertainty.
   Motivated by the oil price hike associated with the Gulf War 1990/1991
they investigate whether or not also transitory oil price hikes lead to oil
field development. Therefore, they augment the price process by a jump
component so that oil prices are now assumed to follow a Brownian motion
mixed with a jump component. In peace time, prices are assumed to jump
up by a certain portion:

                            dPt
                                = αdt + σdBt + ϕdqtN .
                             Pt

Once the war started, it is assumed that oil prices will go down again once


                                                8
war is over:
                       dPt                 ϕ
                           = αdt + σdBt −    dq N
                        Pt                1+ϕ t
Their main finding is that not only permanent, but also transitory oil price
hikes lead to development of oil fields. These transitory hikes, however, need
to be 4 times as large as a permanent one in order to have the same effect.
    In the following it is assumed that the periods of explosive oil price
behavior and transitory oil price hikes are different ways to capture the
same phenomenon. Thus, this paper’s empirical results can be used as basis
for a reconsideration of [8]. As asserted above [8]’s model is motivated by the
1990/1991 Gulf War and it is assumed that a war occurs once in 20 years and
lasts on average 6 months. This paper’s empirical results, however, clearly
indicate that, first, transitory oil price hikes can have other causes than just
wars. Second, due to this finding, it would be implausible to assume that
they occur only once in 20 years. Finally, they can last considerably longer
than just 6 months. Carrying forward the comparative statics in [8], it can
be shown that these different features lead to a higher responsiveness of
investments in oil field developments to jumps.


5   Oil market speculation?

It is important to note that the effect of transitory oil price hike on the
oil field development decision holds regardless of whether the oil price hike
is explained by fundamental factors or driven by “speculation”. However,
various recent papers attempt to shed light on the influence of “speculation”
as well as the “financialization” of oil futures markets. This section briefly
summarizes this debate.
    The concerted research efforts on “oil price bubbles” is epitomized by
the following papers: [1] argue that there is a fundamentally driven long-
term increase in oil prices, which, however, is “exacerbated by speculators.”
In the same vein, [10] show that there is a “change in the relationship be-
tween real oil prices and real stock prices which may suggest the presence
of several stock market and/or oil price bubbles.” [11] also conclude that



                                      9
“financial activity appears to have exacerbated gyrations in the oil mar-
ket, particularly in 2007-2009.” These findings, however, are far from being
undisputed: The survey paper by [7] reviews a number of recent studies and
comes to the conclusion that “the co-movements between spot and futures
prices reflect common economic fundamentals rather than the financialisz-
tion of oil futures markets.” Moreover, [4] argues that “a low price elasticity
of demand, and the failure of physical production to increase, rather than
speculation per se, should be construed as the primary cause of the oil shock
of 2007-08.” [2], finally, shows that the 2003-08 oil price surge “was caused by
unexpected increases in world oil consumption driven by the global business
cycle.” This same paper, however, provides strong evidence that the 1990/91
oil price hike is partly driven by speculative demand. [2] use a structural
VAR model with the variables global crude oil production, a measure of
global real activity, the real price of crude oil and the change in oil invento-
ries above the ground. Four different types of shocks are identified: an oil
flow supply shock, an oil flow demand shock, a residual oil demand shock
and, most importantly, a speculative demand shock. This last shock is de-
fined as a shock to the demand for “above-ground oil inventories arising
from forward-looking behavior not otherwise captured by the model”.
    [7] trenchantly assert that “one of the problems in this literature and,
more importantly, in the public debate about speculation is that it is rarely
clear how speculation is defined and why it is considered harmful to the
economy.” This paper’s empirical findings in combination with the identi-
fied influence of transitory oil price hikes on the oil field development de-
cision indicate that there is a ground for concluding that there are indeed
undesirable effects of speculation.


6   Conclusions

Academic studies on crude oil are anything but scarce. To some extent this
literature evolves wavelike: The oil crises witnessed in the 1970s sparked
enormous efforts to investigate macroeconomic consequences of oil price
shocks. Moreover, the emergence of various resource economic studies which


                                      10
focus on the scarcity of resources can also be explained by these incidents.
A recent offshoot of this literature is motivated by the increasing awareness
of the challenge of climate change. Interest in studying the behavior of oil
prices, however, seems to have been present continuously.
    Motivated by the oil price hikes witnessed at the end of the past decade,
this paper contributes to the literature in three ways. First, the behavior of
oil prices is investigated by testing whether or not oil prices are explosive.
Based on this empirical exercise, second, this paper deals with the conse-
quences of the observed behavior on the resource extraction decision. Third,
the role of speculation in this context is discussed.
    The key findings that emerge from this study are that oil prices are
marked by transitory explosive behavior. Phases in which oil prices exhibit
this kind of behavior are 1990/1991, 2005/2006 and 2007/2008. These find-
ings point to the fact that transitory oil price hikes can occur relatively
often. A reconsideration of existing studies on the relationship between this
type of oil price hikes and oil field development shows that the response of
investments in oil field developments is stronger than originally assumed.
These results hold irrespective of whether the observed oil price hike is fun-
damentally driven or caused by speculative influences. However, as there is
at least some evidence that oil market speculation can play a role during
oil price hikes, it seems to be the case that oil market speculation can have
harmful consequences which previously appear to have been overlooked.

References

 [1] R.K.Kaufmann B.Ullman. Oil prices, speculation, and fundamentals: Interpreting
     causal relations among spot and futures prices. Energy Economics, 31:550–558, 2009.

 [2] L.Kilian D.P.Murphy. The role of inventories and speculative trading in the global
     market for crude oil. Working Paper, 2011.

 [3] M. Gronwald. A characterization of oil price behavior - evidence from jump models.
     Energy Economics, 34:1310–1317, 2012.

 [4] J. Hamilton. Causes and consequences of the oil shock of 2007-08. Brookings Papers
     on Economic Activity, Spring:215–261, 2009.



                                          11
[5] Y.-H.Lee H.-N.Hu J.-C.Chiou. Jump dynamics with structural breaks for crude oil
     prices. Energy Economics, 32:343–350, 2010.

 [6] P.C.B.Phillips Y.Wu J.Yu. Explosive behavior in the 1990s nasdaq: When did exu-
     berance escalate asset values. International Economic Review, 52(1):201–226, 2011.

 [7] B.Fattouh L.Kilian L.Mahadeva. The role of speculation in oil markets: What have
     we learned so far? mimeo, 2011.

 [8] M.Miller L.Zhang. Oil price hikes and development triggers in peace and war. The
     Economic Journal, 106:445–457, 1996.

 [9] J.Lee J.A.List M.C.Strazicich. Non-renewable resouce prices: Deterministic or
     stochastic trends. Journal of Environmental Economics and Management, 51:354–
     370, 2006.

[10] J.I.Miller R.A.Ratti. Crude oil and stock markets: Stability, instability and bubbles.
     Energy Economics, 31:559–568, 2009.

[11] M.J.Lombardi I.van Robays. Do financial investors destabilize the oil price. Ghent
     University Working Paper, 760, 2011.

[12] A.K.Dixit R.S.Pindyck. Investment under uncertainty. Princeton University Press,
     1996.

[13] M. Slade. Trends in natural-resource commodity prices: An analysis of the time
     domain. Journal of Environmental Economics and Management, 9:122–137, 1982.

[14] M. Slade. Grade selection under uncertainty: Least cost last and other anomalies.
     Journal of Environmental Economics and Management, 15:189–205, 1988.

[15] J.L. Smith. World oil: Market or mayhem. Journal of Economic Perspectives,
     23(3):145–164, 2009.




                                            12
You can also read
Next part ... Cancel