ARTICLE IN PRESS

Energy Policy 35 (2007) 2096–2109

www.elsevier.com/locate/enpol

A supply model for crude oil and natural gas in the Middle East
R. Chedida,, M. Kobroslya, R. Ghajarb
a
American University of Beirut, Faculty of Engineering and Architecture, P.O. Box 11-0236, Beirut, Lebanon
b
School of Engineering and Architecture, Lebanese American University, P.O. Box 36, Byblos, Lebanon
Available online 28 August 2006

Abstract

Crude oil (CO) and natural gas (NG) play an important role in the world economy. The Arab countries (ACs) share 64% of the total
oil reserves and 40% of the NG reserves [OPEC, 2003. The Annual Statistical Bulletin]. On the production side, ACs contribute to 30%
and 9% of the world production of CO and NG, respectively. Accurate forecast models are needed to do better planning and create less
risky business environment. In this paper, an econometric model is built to capture the behavior of CO and NG production in the
ESCWA1 countries. The model is used to forecast future production trends of CO and NG, and thus provide a powerful tool for
researchers, planners and investors working in the energy field.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Crude oil; Natural gas; Econometric models

1. Introduction 1983 to 30.4% in 2003 to reach 20,451.6 thousands of

Crude oil (CO) is a major contributor to the world
economy and natural gas (NG) is becoming a direct
competitor for the first due to its environmental advantages
and current and potential future role in electricity
generation. Most of the Arab countries (ACs) in the
ESCWA region are the main participants in the world
energy because of their reserves, supplies and trade
markets. From the Annual Statistical Bulletin of the
Organization of Petroleum Exporting Countries (OPEC)
for the year 2003 (OPEC, 2003), the share of proven CO
reserve of the Middle East (ME) out of the world total
increased from 56% (397,053.3 mb) to 64% (735,866.3 mb)
between 1983 and 2003 (Fig. 1). While the proven NG
reserve in the ME raised from 27.6% (26,405 billion
standard cu m) to 40.4% (72,766 billion standard cu m) out
of world reserves for the same period (Fig. 1). Moreover,
the Arab share of CO production moved up from 21.2% in
Corresponding author.
E-mail addresses: rchedid@aub.edu.lb (R. Chedid),
rghajar@lau.edu.lb (R. Ghajar).
1
ESCWA is the UN Economic and Social Council of West Asia. It deals
with thirteen Arab countries; namely, Lebanon, Syria, Jordan, Palestinian
authority, Iraq, Egypt, Saudi Arabia, Kuwait, Qatar, Bahrain, UAE,
Oman and Yemen.
barrels per day.
The next two decades are expected to see an increase in
energy demand met predominantly by fossil fuels, with oil
set to continue to maintain its major role complemented
with an increasing role of NG mainly in electric power
generation. ACs rely almost entirely on oil and NG to meet
their energy requirements, accounting for 97% of total
energy consumption. According to the 30th Annual Report
of OAPEC in 2003 (OAPEC, 2003), the energy mix in AC
heavily tilts towards oil, accounting for 54.4% of total
energy consumption, followed by NG accounting for
42.6% (Fig. 2).
Many authors have investigated different energy models
to help analysts forecast NG demand. Kaboudan and Liu
(2004) investigated the use of an integrated genetic
programming/multiple regression system of equations to
forecast short-term US demand for NG. Genetic program-
ming is employed to predict the system’s explanatory
variables. The demand equations are used to forecast 4
years of demand for NG. Krichene (2002) examined the use
of a two-stage least-squares method for CO and NG to
obtain short-run estimates of demand and supply. In order
to strengthen the confidence in these estimates, the model is
re-estimated in an Error Correction Model (ECM). Siemek
et al. (2003) estimated the consumption of NG in Poland

0301-4215/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enpol.2006.06.011
 ARTICLE IN PRESS
R. Chedid et al. / Energy Policy 35 (2007) 2096–2109 2097

Reserves of Crude Oil Reserves of Natural Gas

36%
40%

60%
64%

Middle-East Rest of the World Middle-East Rest of the World

Fig. 1. World reserve for CO and NG in 2003. Source: OPEC (2003).

60

50

40

30

20

10

0
Petroleum NG Hydroelectricity Coal
Products

Fig. 2. Energy consumption in AC in 2003. Source: OAPEC (2003). Fig. 3. Oman, Emirates, Qatar, Bahrain and Kuwait NG pipeline
network. Source: E/ESCWA/SDPD/2004.

based on the logistic-curve interpretation, where an

adaptation of the Hubbert model is implemented to the
Polish situation based upon the Staizman modification.
Dahl (1994) found that the most commonly used
modeling techniques for aggregate energy demand are the
simple static and the Partial Adjustment Models (PAM).
Erdogan and Dahl (1997) investigated the impact of
income, price and population on the aggregate, industrial,
manufacture and mining sectors of energy in Turkey. Nasr
et al. (2000) built an EM to forecast the electric
consumption in post-war Lebanon. They did include the
impact of the gross domestic product (GDP), the total
imports (TI), and degree days (DD) to formulate an
electricity consumption model over different periods from
1993 to 1997. Eltony and Al-Mutairi (1995) and Eltony
(1996) presented a cointegrating relationship for NG and
gasoline demand in Kuwait, which included price, income
and population as explanatory variables. Later Eltony and Fig. 4. Egypt, Oman, Syria, Lebanon and Turkey NG pipeline network.
Source: E/ESCWA/SDPD/2004.
Hosque (1997) developed cointegrating relationship for
electricity demand in Kuwait.
This paper develops an econometric model for forecast- peculiarities of the selected parameters affecting the model.
ing the supply of CO and NG in AC located in the ESCWA In Section 3, the proposed econometric model is developed
region, and is organized as follows. Section 2 addresses the and used to forecast the production of CO and NG.
 ARTICLE IN PRESS
2098 R. Chedid et al. / Energy Policy 35 (2007) 2096–2109

In Section 4, the model is tested and used to forecast
the production of CO and NG for all studied
ESCWA countries. Finally, Section 5 provides concluding
remarks.
2. Study peculiarities
The analysis of CO and NG supply in AC is influenced
by some peculiarities, and so precautions should be taken
while accounting for the following:
(i) Effectiveness of the chosen parameters: The most
influential parameters in the analysis of CO and NG
supply in ACs are found to be: International selling
price of CO and NG, GDP, population, local
consumption, correlation between the production of
the studied country and world production, and finally,
the correlation between the consumption of the
studied country and world consumption.
(ii) Redundancy and incompleteness of data: A precaution
should be taken to avoid redundancy and data
incompleteness of the chosen parameters. For in-
stance, although Yemen is a producer of CO and not
of NG, it is excluded from the study due to the
inconsistency in its recorded data.
(iii) Characteristics of individual countries: It is noticeable
the potential of NG production which is set at
present to 8% of the world production, and used
almost locally, while, the AC reserves of NG reach 40%
of the total world reserves (OPEC, 2003). NG Arab
producing countries are currently using three ways to
deliver their products (Alexander’s Gas & Oil Connec-
tions, 2005):
(a) Through pipelines that require huge investment.
Usually pipelines are constructed such as they would
pass through a number of countries enabling them to

Table 1
Japan imported NG prices ($/Mmbtu)

Year 1990 1991 1992 1993 1994 1995 1997 1998 1999 2000 2001

Price 3.7 4 3.7 3.6 3.3 3.5 3.9 3 3.1 4.8 4.7
Deviation 0 0.081 0 0.03 0.11 0.05 0.054 0.19 0.16 0.297 0.27

Table 2
CO prices ($/bbl)

Year 80 81 82 83 84 85 86 87

1995 US dollar 21.59 31.77 28.52 26.19 25.88 24.09 12.51 15.4
Year 88 89 90 91 92 93 94 95

1995 US dollar 12.58 15.86 20.03 16.54 15.99 14.25 13.19 14.62

Year 96 97 98 99 00 01 02

1995 US dollar 18.46 17.23 10.87 15.56 26.72 21.84 22.51

Source: InfalationData.com.

GDP at Market Exchange Rates GDP at Market Exchange Rates

20
160
Billions of 1995 U.S. Dollars

18 Bahrain Kuwait
Billions of U.S. Dollars

Iraq 140 SA
16 Jordan
Syria
UAE
14 Oman 120 Egypt
Qatar
12 100
10 80
8 60
6
40
4
2 20
0 0
1980 1983 1986 1989 1992 1995 1998 2001 1980 1983 1986 1989 1992 1995 1998 2001
Year Year

Fig. 5. GDP in billions of 1995 US dollars. Source: EIA (2005).

 ARTICLE IN PRESS
R. Chedid et al. / Energy Policy 35 (2007) 2096–2109 2099

benefit from the product, and as a result making it less 3. Mathematical modeling
expensive.
(b) Transferring NG through special containers after The mathematical econometric model given in Eq. (1) is
liquefying and freezing the NG. As a result, the cost widely accepted in energy supply studies (Liu et al., 1991
can go beyond that of CO for the same energy content.
(c) Using Fischer–Tropsch technology to convert NG into
liquid fuels (GTL) of excellent technical and environ-
mental qualities, which supersede those obtained from
77000
ordinary refinery products. Qatar is a pioneer among World CO Production
the ESCWA countries in the introduction of GTL World PetroleumConsumption

Thousand Barrels per Day

technology. 72000

67000
Enormous investments (US$160 billion) are spent in the
development of pipeline networks to transport NG 62000
through the region to the European continent. Figs. 3
and 4 show, respectively, the pipeline network in the Gulf 57000
region and in the northern part of the ESCWA region.
Parts of these networks (Oman–Emirates, Jordan–Egypt,
52000
Kuwait–Qatar and Kuwait–Iraq) have been completed and 1980 1983 1986 1989 1992 1995 1998 2001
made operational, and works on other parts are either Year
ongoing or pending. Most of this gas will go on power Fig. 7. World consumption of petroleum and the world CO production.
generation in the ESCWA region since the installed Source: EIA (2005).
capacity of the combined cycle technology is projected to
increase from 38,324 MW in 2001 to 131,692 MW by 2010.
Oil and oil products will always be the major contributor
in the ACs’ economy. Petroleum export revenues of the World Natural Gas Production
Organization of Arab Petroleum Exporting Countries 95
(OAPEC) Member Countries, which are a vital source
for financing various development projects, increased by 90
14% from $142 billion in 2002 to $162 billion in 2003 at
Trillion Cubic Feet

current prices (OAPEC, 2003). Increases in oil revenues

vary among member countries, ranging from as low as 85
7.6% in the UAE to high of 41.2% in Bahrain. Oil
revenues were up by 8.9% in both Qatar and Syria. In 80
Saudi Arabia, revenues increased by 18.3%, and were up
by 17.8% in Egypt and by 20% in Kuwait. Iraq was the
75
only exception with revenues dropping down due to the
unstable situation in this country. As a result, the overall
increase in petroleum revenues of OAPEC member 70
countries was reflected in the increase of per-capita 1990 1992 1994 1996 1998 2000
Year
revenues ranging from 1.9% and 36.5% (excluding Iraq).
Fig. 8. World NG production. Source: EIA (2005).

Population Population
80 6
Bahrain
70 Jordan
5 Kuwait
60 Oman
Iraq
50 4 Qatar
Millions
Millions

SA
UAE
40 Egypt 3
Syria
30
2
20
10 1
0 0
1980 1983 1986 1989 1992 1995 1998 2001 1980 1983 1985 1989 1992 1995 1998 2001
Year Year

Fig. 6. Population in millions. Source: EIA (2005).

 ARTICLE IN PRESS
2100 R. Chedid et al. / Energy Policy 35 (2007) 2096–2109

National University of Singapore): Egypt (NG Production)

lnðY Þ ¼ b0 þ b1 lnðX 1 Þ þ b2 lnðX 2 Þ 0.95

þ b3 lnðX 3 Þ þ          þ b6 lnðX 6 Þ, ð1Þ 0.85

Tested
Real
where Y is the production of CO (thousand barrels per Estimated
0.75

Trillion Cubic Feet

day) and of NG (trillion cubic feet) (see Appendix B), X1
the international selling price in US$/barrel for CO and 0.65
US$/Mmbtu for NG, X2 the GDP (billions of 1995 US
dollars), X3 the population (millions), X4 the consumption 0.55
of CO (thousand barrels per day) (see Appendix B), X5 the
correlation factor between the local and the world 0.45
production(%), X6 the correlation factor between the local
and the world consumption (%), b0, b1, b2, b3,yy, b6 the 0.35
coefficients to be estimated using historical data of total
annual ‘‘oil\NG’’ supply for past years. 0.25
1990 1992 1994 1996 1998 2000
Year
Qatar (CO Production)
Fig. 11. Egypt NG production. Source: Matlab Result.
730

680 Tested
Real Oman (NG Production)
630
Thousand of Barrels per Day

Estimated 0.49
580

530 Tested
0.39
Real
480
Trillion Cubic Feet

Estimated
430
0.29
380

330
0.19
280
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002
Year

Fig. 9. Qatar CO production. Source: Matlab Result. 0.09

1990 1992 1994 1996 1998 2000
Year

SA (CO Production) Fig. 12. Oman NG production. Source: Matlab Result.

10

9
Projection for NG prices
7
Billion of Barrels per Day

8
6.5
7 Tested 6
$/Mmbtu

Real
6 5.5
Estimated

5 5

4.5
4
4
3 2002 2004 2006 2008 2010
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 Year
Year
Fig. 13. Projection for NG price ($/Mmbtu). Source: Annual Energy
Fig. 10. SA CO production. Source: Matlab Result. Outlook (2005).
 ARTICLE IN PRESS
R. Chedid et al. / Energy Policy 35 (2007) 2096–2109 2101

Correlation for NG production The objective of the model is to simulate the past trend
2.5
Bahrain Oman of CO and NG production of AC (Appendix A), and make
Egypt Qatar
Iraq SA extrapolation for future prediction. This extrapolation is
Jordan Syria
2 Kuwait UAE strongly trusted after the testing phase is realized.
In the model of NG, the consumption factor (X4) can not
be used and will be excluded due to the fact that most of
1.5 NG production has been consumed locally especially in the
Percentage

early period of study. For the same reason, the NG

consumption correlation factor (X6) will be replaced by the
1 production correlation factor of CO due to the competi-
tiveness and effect of CO on the NG sector. Furthermore
the CO sector is always used as a reference energy source to
0.5 compare with.
The study for CO will be for a period of 23 years
extending from 1980 till 2002, whereas for NG, the period
0
1994 1995 1996 1997 1998 1999 2000 2001 of examination is only for 12 years extending from 1990 to
Year 2001.
The following section details the data used and the way it
Fig. 14. Correlation for NG production. Source: Matlab Result.
is treated in the model:
Correlation for CO production
16.0
 Price: Most of the AC export NG to East-Asia and
14.0 mainly to Japan. As the price of AC exported NG is
12.0
documented in the literature for certain years (OAPEC,
Bahrain
Egypt
2002), the price variation in Japan imported NG
10.0 Iraq (Table 1) is used to establish correlations with NG
Percentage

Kuwait
Oman price variation in each of NG Arab producing country
8.0 Qatar
SA so as to develop a time series variation of those prices
Syria
6.0 UAE for the whole period of study.
It is worth mentioning that the analysis of NG prices is
4.0 done because there is no commodity market for NG,
2.0 and so prices are determined according to individual
agreements made between the suppliers and consumers.
0.0 As for CO, the prices are recorded from international
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year market data (InfalationData.com) as shown in Table 2.
 GDP and population: GDP and population data (EIA,
Fig. 15. Correlation for CO production. Source: Matlab Result. 2005) for AC are plotted in Figs. 5 and 6. As shown in

Kuwait CO production
2.5
Thousand Barrels per Day *1000

1.5
Kuwait CO estimated

Kuwait CO real
1
Kuwait CO forecasted

0.5

0
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year

Fig. 16. Kuwait CO future projection. Source: Matlab Result.

 ARTICLE IN PRESS
2102 R. Chedid et al. / Energy Policy 35 (2007) 2096–2109

SA CO production

10

9
Thousand Barrels per Day *1000
8

7
SA CO estimated

6 SA CO real

SA CO forecasted
5

3
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year

Fig. 17. SA CO future projection. Source: Matlab Result.

these figures there is a steady increase in both GDP and UAE NG production
population. This annual increase is essential to forecast 2.2
the future production of CO and NG. 2 UAE NG estimated
 Correlation factors: Fig. 7 plots the trend of world NG
Trillion Cubic Feet

1.8 UAE NG real

production, and Fig. 8 shows the development of world 1.6 UAE NG forecasted
CO production and consumption. Based on this figure,
1.4
the following remarks can be made:
1.2
(i) In the late 80s there was a drop in CO production due
to the war between Iraq and Iran, which are major 1
contributors to the world production. 0.8
(ii) Referring to Hubbert’s prediction (Deffeyes and 0.6
Kenneth, 2001), the world CO production will take a 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
peak in the first part of the 21st century then will Year
decrease; as the AC have the largest reserves, their Fig. 18. UAE NG future projection. Source: Matlab Result.
participation in the world production proved not to be
affected by this phenomena. AC will keep a stable
position in the production of CO and their contribution
may increase. This will be used later in order to forecast respectively. It can be noticed that the model provides
the future production for CO while as for NG the good matching with the measured data over the
Hubbert’s peak will also occur but a few decades later. same years.

4.2. Forecasting based on the model

4. Results
4.1. Model testing
The mathematical model proposed was built for the
period 1980 to 1997 and later tested versus real data for the
period 1998 to 2002 (for CO). For NG it was constructed
for the period 1990 to 1998 and later tested for years
1999–2001. After gaining confidence in the model, a future
forecast is done.
Below is a sample of the results of the testing phase for
CO and NG production. Figs. 9 and 10 show CO
production for Qatar and Saudi Arabia respectively, while,
Figs. 11 and 12 show NG production for Egypt and Oman,
Having shown the strength of the EM in the testing
phase, the objective now is to use this model to do
forecasting for the future. Figs. 5–8 suggest that GDP,
population and the NG production correlation factor are
increasing functions, thus the yearly increase of these
functions—calculated independently—will be used in the
forecast which is set for 8 years. From the Annual Energy
Outlook (2005), CO price is projected to reach $24.50 per
barrel by 2010, and the future trend of NG price will be as
shown in Fig. 13.
The rate of change of the correlation factor of CO
production is obviously a constant except for Iraq because
of the unstable situation there. Similarly, the correlation
 ARTICLE IN PRESS
R. Chedid et al. / Energy Policy 35 (2007) 2096–2109 2103

Qatar NG production
2.5

2
Qatar NG estimated
Trillion Cubic Feet Qatar NG real
1.5
Qatar NG forecasted

0.5

0
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year

Fig. 19. Qatar NG future projection. Source: Matlab Result.

Table 3
Results for crude oil production function

Country Bahrain Egypt Iraq Kuwait Oman Qatar SA Syria UAE

R2 0.986784 0.999917 0.999850 0.999853 0.999852 0.999848 0.999889 0.999835 0.999908

R 0.993370 0.999958 0.999925 0.999926 0.999926 0.999924 0.999944 0.999917 0.999954
F-test 117.8 258.8 759.8 403.2 1935.5 485.4 234.5 3118.3 3029.9

Source Matlab Result.

Table 4
Results for natural gas production function

Country Bahrain Egypt Iraq Kuwait Oman Qatar SA Syria UAE

2
R 0.997615 0.999811 0.999981 0.995521 0.998050 0.999738 0.999943 0.999865 0.999726
R 0.998807 0.999905 0.999991 0.997758 0.999024 0.999869 0.999971 0.999933 0.999863
F-test 416.7 2358.7 1241.9 1125.5 1231.8 1799.7 769.8 1052.4 189.9

Source Matlab Result.

factor of NG shows an increasing trend except for Iraq for
the same reason. Thus an average correlation factor for
each country will be calculated and used for future
forecasting of both CO and NG. (Figs. 14 and 15).
The results of the forecast for CO and NG for selected
countries are shown in Figs. 16–19.
R2 is the proportion of the total variation in the n
observed values of the dependent variable that is explained
by the overall regression model (Tables 3 and 4). The more
the value of R approaches the unity, the more the
confidence in the model is. Relying on presented results,
NG will play an important role in the near future. Most
of the ACs are switching from CO to NG due to its
environmental advantages. This will put the NG in a
favorable position to compete with CO in the energy
market. In addition the infrastructure for NG is being
developed both internally (within individual AC) and
externally within the region as mentioned before. The
results show a big potential for some AC in NG though still
not fully profited from (i.e., Qatar and UAE). For CO, the
forecast shows a less rate of increase in supply with a small
increase till 2010. All the profits from NG and CO sectors
are not limited to the producers only, but exceed the limits
to the near non-producing AC, such as the case of the NG
project between Lebanon and Syria and many other
projects going on.
 ARTICLE IN PRESS
2104 R. Chedid et al. / Energy Policy 35 (2007) 2096–2109

5. Conclusion bk x 1 ¼ ½ b0 b1 b2  bk  0 , (A.4)

In this paper, some aspects of natural gas (NG) and and

crude oil (CO) supply sectors in ESCWA countries have
been discussed. An analysis of recorded data revealed that E ¼ ½ E 1980 E 1981 E 1982  E 2002 0 . (A.5)
the most important parameters affecting the forecast of the
supply of CO and NG were GDP, population, interna- The aim is to estimate b using least squares. We want to
tional prices and correlation factors for production and find the value of b that minimizes the sum of squared
consumption of both CO and NG. An EM has been built residuals:
to forecast the supply of CO and NG in the AC located in
the ESCWA region. The testing results show good X
n
correlation with real data and thus provide good back- SSEðbÞ ¼ ðE i Þ2 ¼ E 0 E. (A.6)
ground for accurate forecast for the future. The forecast i¼1
reveals that the supply of NG will double in about 10 years,
but the supply of CO will only slightly increase by about Now for this value of b, substituting
5–10% during the same time interval. Such results high-
light the importance of the oil sector in general and the NG E^ ¼ Y  Xb, (A.7)
sector in particular in future economic development of the
ESCWA region. we can write SSE (b) as
However, because of political instability in the ME
which greatly influences the oil and gas markets, the model SSEðbÞ ¼ ðy  X  bÞ0 ðy  X  bÞ. (A.8)
should be upgraded in the future to reflect the effects of
such external factors. Now, taking the partial derivatives with respect to b0, b1 ,b2
y bk and setting then equal to zero, we obtain the system
Acknowledgments of equations:

The first two authors would like to thank the University qSSEðbÞ h qSSE qSSE qSSE

i
qSSE 0
¼ qb0 qb1 qb2 qbk
Research Board of the American University of Beirut for qb
their financial support of this work. set ½ 0 0 0    0 0 . ðA:9Þ

A multivariable calculus argument shows that the

Appendix A value of b that solves this minimization problem is
given by
In fitting a multiple linear regression model, particularly
when the number of variable exceeds two, knowledge of
matrix theory can facilitate the mathematical manipulation b ¼ ðX 0 X Þ1 X 0 Y . (A.10)
considerably. Having the model explained before, it can be
written as This is called the least-square estimator of b. In order to
compare the effectiveness of the results, a measure criterion
Y ¼ X  b þ E, (A.1) for the curve generated will be adopted. Finding SSE, the
P
mean value for the production (MV ¼ Y/number of
2 3
ln Y ð1980Þ year) andP the sum of squares about the mean
6 ln
6 Y ð1981Þ 7
7 (SST ¼ [(Yiyiav)42]) the fitness factor R can be
6 7 calculated as follows:
6 ln Y ð1982Þ 7
Y ¼6 7, (A.2)
6 .. 7
6 . 7 R ¼ 1  ðSSE=SSTÞ. (A.11)
4 5
ln Y ð2002Þ
The more R approaches to the unit, the more the data
2 3 calculated coincide with the real one which is our objective.
1 ln X 1ð1980Þ ln X 2ð1980Þ ln X 3ð1980Þ  ln X kð1980Þ
61 ln X 1ð1981Þ ln X 2ð1981Þ ln X 3ð1981Þ  ln X kð1981Þ 7
6 7
6 7
61 ln X 1ð1982Þ ln X 2ð1982Þ ln X 3ð1982Þ  ln X kð1982Þ 7 Appendix B
X ¼6 7,
6. 7
6 .. 7
4 5
1 ln X 1ð2002Þ ln X 2ð2002Þ ln X 3ð2002Þ  ln X kð2002Þ
Crude oil production, petroleum consumption and dry
natural gas consumption have been explained in
(A.3) Figs. B1–B8.
 ARTICLE IN PRESS
R. Chedid et al. / Energy Policy 35 (2007) 2096–2109 2105

Crude Oil Production

1200

1000

Thousand Barrel per day

800

600

400
Egypt
Om an
200
Qata r
Syr ia
0
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002
Yea r

Fig. B1. Crude oil production (part 1) Source: OPEC (2003).

Crude Oil Production

12000

10000
Thousand Barrel per day

8000

6000 Iraq
Kuwait
SA
4000 UAE

2000

0
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Yea r

Fig. B2. Crude oil production (part 2) Source: OPEC (2003).

 ARTICLE IN PRESS
2106 R. Chedid et al. / Energy Policy 35 (2007) 2096–2109

Petroleum Consumption
1,600
Egyp t Iraq Kuwait
1,400 SA Syria UAE

Thousand Barrels per day 1,200

1,000

800

600

400

200

0
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Year

Fig. B3. Petroleum consumption (part 1) Source: OPEC (2003).

Petroleum Consumption
120
Bahrain
Jordan
100 Oman
Qatar
Thousand Barrels per day

80

60

40

20

0
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002
Year

Fig. B4. Petroleum consumption (part 2) Source: OPEC (2003).

 ARTICLE IN PRESS
R. Chedid et al. / Energy Policy 35 (2007) 2096–2109 2107

D ry Natural Gas Production

2.5
Oman
Qat a r

2.0
SA
Trillion Cubic Feet UA E
1.5
Egyp t
1.0
0.5
0.0

1990 1992 1994 1996 1998 2000 2 002

Year

Fig. B5. Dry natural gas production (part 1) Source: OPEC (2003).

Dry Na tural G as Production

0.40
Bah rain
0.35 Ira q
Jord a n
0.30
Ku wait
Trillion Cubic Feet

Syria
0.25

0.20

0.15

0.10

0.05

0.00
1990 1992 1994 1996 1998 2000 2002
Ye a r

Fig. B6. Dry natural gas production (part 2) Source: OPEC (2003).
 ARTICLE IN PRESS
2108 R. Chedid et al. / Energy Policy 35 (2007) 2096–2109

Dry Na tural G as C ons umption

Bahrain Iraq Kuw a it

600
Oman Qatar Syria

500

400
Billion Cubic Feet

300

200

100

0
1990 1992 1994 1996 1998 2000 2002
Year

Fig. B7. Dry natural gas consumption (part 1) Source: OPEC (2003).

Dry Na tural G as C ons umption

2,500

2,000 SA UA E Egypt
Billion Cubic Feet

1,500

1,000

500

0
1990 1992 1994 1996 1998 2000 2002
Year

Fig. B8. Dry natural gas consumption (part 2) Source: OPEC (2003).
 ARTICLE IN PRESS
R. Chedid et al. / Energy Policy 35 (2007) 2096–2109
References
Alexander’s Gas & Oil Connections, 2005. News & Trends: Middle East.
Natural gas and its role in world economy, vol. 10(6). /www.
gasandoil.comS.
Annual Energy Outlook, 2005. Energy Information Administration.
Official energy statistics from the U.S. Government. Forecasts/Annual
Energy Outlook 2005/Market Trends—Natural Gas Demand and
Supply /www.eia.doe.govS.
Dahl, C., 1994. A survey of energy demand elasticities for the developing
world. Journal of Energy Development, 1–47.
Deffeyes, Kenneth, S., 2001. Hubbert’s Peak: The Impending World Oil
Shortage. Princeton University Press, Princeton, NJ.
Eltony, M.N., 1996. Demand for natural gas in Kuwait: an empirical
analysis using two econometric models. Energy Research 20, 957–963.
Eltony, M.N., Al-Mutairi, N.H., 1995. Demand for gasoline in Kuwait.
Energy Economics 17 (3), 249–253.
Eltony, M.N., Hosque, A., 1997. A cointegrating relationship in the
demand for energy: the case of electricity in Kuwait. Journal of Energy
and Development 21 (2), 293–301.
Energy Information Administration, 2005. /www.eia.doe.govS.
Erdogan, M., Dahl, C., 1997. Energy demand in Turkey. Journal of
Energy and Development 21 (2), 173–187.
E/ESCWA/SDPD/2004. United Nations Publication. A study update
of the current status of selected energy sectors in ESCWA countries.
Part I. Natural Gas Sector, 2004.
Kaboudan, M., Liu, W.Q., 2004. Forecasting Quarterly US Demand for
Natural Gas. Information Technology in Economics and Management
(ITEM).
Krichene, N., 2002. World crude oil and natural gas: a demand and supply
model. Energy Economics 24 (6), 557–576.
Liu, X.Q., Ang, B.W., Goh, T.N., 1991. Forecasting of electricity
consumption: A comparison between an econometric model and a
2109
neural network model. Department of Industrial and Systems
Engineering, National University of Singapore.
Matlab Result. A computer program written by the authors.
Nasr, G.E., Badr, E.A., Dibeh, U.G., 2000. Econometric modeling of
electricity consumption in post-war Lebanon. Energy Economics 22,
627–640.
OAPEC (Organization of Arab Petroleum Exporting Countries), 2002.
The Secretary General’s Twenty-ninth Annual Report.
OAPEC (Organization of Arab Petroleum Exporting Countries), 2003.
The Secretary General’s 30th Annual Report.
OPEC (Organization of Petroleum Exporting Countries), 2003. The
Annual Statistical Bulletin.
Siemek, J., Nagy, S., Rychlicki, S., 2003. Estimation of natural gas
consumption in Poland based on the logistic-curve interpretation.
Applied Energy 75 (1–2).
www.InfalationData.com.
Further readings
Eckbo, P.L., 1979. A Basin Development Model of Oil Supply. Advances
in the Economics of Energy and Resources, vol. 2. JAI Press Inc.,
Greenwich, CT, pp. 39–58.
Foat, K.D., MacFadyen, 1983. Modeling success in Alberta’s oil plays.
Study No. 19. Canadian Energy Research Institute (CERI), Calgary
Alberta.
NRCan, 1997. Canada’s Energy Outlook: 1996–2020. Ottawa, Canada.
Persaud, A.J., Kumar, Uma, 2001. An eclectic approach in energy
forecasting: a case of Natural Resources Canada’s (NRCan’s) oil and
gas outlook. Journal of Energy Policy 29, 303–313.
Uhler, R.S., 1979. The rate of petroleum exploration and extraction. In:
Pindyck, R.S. (Ed.), Advances in Economics of Energy and Resources.
JAI Press, Greenwich, Connecticut pp. 93–118.