crude oil price modeling a macro- economic approach · ing crude oil and petroleum products. the...
TRANSCRIPT
Crude Oil Price Modeling... A Macro- Economic Approach
DTIC QUALITY INSPECTED 2
Student: David Wasberg Advisor: Shapour Vossoughi
DISTRIBUTION STATEMENT A Approved for Public Release
Distribution Unlimited 20000308 147
Contents
Objective PartT 1 Part II 1 Methodology: Hew To Build A-Model 1
Background
The World Petroleum Market Model Early Discoveries Complex Seenarios- Initial Results
3 3 4 4 5 5 5 6 6
Macro-Economics: A Different Approach C±rt/ran_ OTagCOr
Perspective &eopc Reducing Variation
Stage I: Factors Of Relevance Demand-Präe- Cycle 7 Demand Versus Consumption 7 UnnsuatEveiits 7 Price Drivers 8 Pft«n Tit? <■**» 8 1 UpUluLlOIl
Gross Domestic Product 8 8 XjCaUlIlg Et/UllOJIUC UlUlCtuUra
Reducing Complexity 9 Observation- 9 Alternative Energy Sources 10 Balancing-Fefces- 10 Limits On Substitutability 10 Pollution 11 Accounting For Alternative Energy Sources 11 TimeLag- 11 Predictions 11 PastEnergy Usage 12 Magnification 12 OrhTaxes- 12 All Things Considered- 12-
Tax Policy Simplified- Tariffs And Quotas Price-Controls- Climate PoliticaLSpeailatiorL War Or Other Disasters Supply Supply-Price Cycle Crude Production. Assumption Consideration«: And Simplifications Petroleum Reserves Complications Price vs Reserves Price Reserves Dilemma 1 Reserves-Dilemma 7, New Discoveries, Reserves And Price Pressing-Ahead- Reserve To Production Ratio Inventories- Bottlenecks Exploration AnH Drilling OPEC
13 13 13 13 13 14 14 14 15 15 15 16 16 16 16 17 17 17 17 18 18 18 19 19
Stage II: Data KeyEactors- Data Requirements Ftnsfr Petroleum Erfr Second Petroleum Era ThirdPetrnleum Era - Great Transition Stupendous Demand And Change A ShifiingJLepository OLWealthAndPoH Reflection Comparison Conclusion Stable-Or-Net? Changing Technology And Markets ADifferent Petroleum Market Conservation And Environmental Impacts Characterization- Missing & Incomplete Data Data With Different Units Reporting Source Discontinuity Data Gathering Pitfalls Data Approach You Get What You Pay For Reminder Raw Data Data Sources
er_
20 20 21 21 21 21 22 23 23 26 26 27 27 28 28 29 29 29 30 31 31 31 31 32 32-
Stage III: Qualitative Analysis U/fMiU Oil Ajfii-lrr»t WOrlu t«rivlaiKCl 34 Major Events 34 Extraordinary Events - 34 Significant Events 35 I nc iViOQCt 35
Stage IV: Analytical Results Data Randomness" 37 Individual Regression Analysis 37 rriiuipiUlallUil \_flrJnttuiViUlutr/viiuxjMb J\.C£rUHa 38 Multiple Regression Analysis 39 MuMpleRegrcssion Results- 39
Stage V: Conclusion Assumptions And Limitations 41 Results 42 Recommendations 43 Final Thoughts 44
Exhibit I
Exhibit II
Exhibit III
Exhibit IV
Sources
Objective
Examine quantitative and qualitative factors and their correlation to crude oil prices.
Part I: This thesis is Part-I of a two-part evaluation concerning crude oil price prediction and modeling. In
Part I; quantitative factors- which-are suspeeted-te-influence crude oil price wilt be examined by statistically
comparing their effects on crude oil price over time. This will be done on an individual factor basis as well as
using multiple regressions analysis. Additionally, qualitative faetors-which are suspected-of influencing crude
oil prices will be evaluated through historical observation and an application of logical reasoning and analysis.
AH- factors; quantitative- and- qualitative, wMeh-are identified-as-possessing seme significant level- of influence
on crude oil prices, will be incorporated into Part-II of the evaluation. Part-II will be conducted as a separate
thesis-project.
Part II: In Part-II, a statistical model will be developed in which each significant factor identified in Part-I is
assigned an appropriate weight of importance (an plasticity with regards in prjce)-and a probability distribu-
tion indicative of likeliness of occurrence (a measure of risk or frequency). These weighted factors and their
probability distributions. wilLhe adjusted and hypothetical crude oil price results wilLbc-generated and com-
pared to actual price history. The objective of Part-n is to create an empirical model for which all important
factors fneir weighted importance and their chance of occurring are incorporated into a single simulation The
iterative process of adjusting factor weights and probabilities is intended to create a simulation that can pro-
duce predictions which closely coincide with actual historical crude nil price data. Such_a_ simulation _COuld
then be utilized to forecast future crude oil prices given expected circumstances or contingencies.
Methodology: How To Build The Model: When confronted with the task of constructing a mathematical
modeMhere is a temptation tamsh ahead and begin crunching numhers and analyzing results This is a mis-
take. A prudent evaluation calls for a methodical approach. We must first determine WHAT factors are im-
pnrtant to analyze and WHY I This is necessary to ensure that we (1) account for as many significant factors of
influence as possible and (2) have an understanding of why each factor influences crude oil price and should
therefore baincluded in themodel This will also allow us to eliminate factors which may on the surface,seem
to be important but which have no real bearing on our model. We want to analyze the right things ... not eve-
rything_ QnceJwe-haye-Carefully identified the factors of importance we must look to see if data is available If
data is available, terrific, but if not, we may have to choose a proxy-factor which closely mirrors the data we
would really like_tO-USe.- Finally we must examine the reliability consistency and credibility of our data
sources. Although a useful model is never guaranteed, only quality data can hope produce meaningful results.
The approach for Part-lwilLtherefore begin with a rigorous examination of the relevant factors affectingcrude
oil price and seek to develop a reasonable understanding of how that influence comes about.
Prior to 1973, crude oil was relatively inexpensive and few concerned themselves with understanding the
mechanism involved in establishing price. Most believed inappropriately, that oil and oil price behaved like
any other commodity in an open market, depending solely on pressures of supply and demand. The oil em-
bargo of 1973 sent shock waves throughout the industrialized world as OPEC nations effected a modest redac-
tion in worldwide crude oil output. Although the cut-backs were highly selective, targeted at specific western
powers supporting Israel during the Yom Kippur War, the net effect was an "energy crisis" with panic buying
driving up crude oil prices nearly 350% in a single year. l The resulting uncontrolled aid upward spiraling
prices, fuel shortages and long gas lines, and the western world's seeming vulnerability to OPEC decisions de-
manded answers. Governments, industry and academia scrambled to understand the factors controlling crude
oil price and supply. The hope was to regain some measure of control over prices and supply or to develop de-
fensive strategies to limit the degree of price volatility.
The Woilcl Petroleum Market
In 1973, hundreds of separate activities began a quest to understand and model the crude oil market.
An initial sensitivity analysis revealed that crude ott prices were affected by the anticipated factors of
supply and demand.
Early Discoveries: On the Supply side, there werephysical limitations to petroleum discoveries, reserves^and
production rates from various oil fields. These producing, regions were widely dispersed throughout the world
and brought forth a wide range of crude oils which yielded a still wider range of products in varying quantities.
Crude oils and their refined products were then marketed throughout the world-in the regions demanding those
products pr possessing the necessary refining and storage capacity. This entailed a remarkably complex distri-
bution and transportation system with a cost structure which accounted for export taxes, import taxesr tariffs,
quotas and price controls. Researchers discovered that there was no central mechanism for buying and sell-
ing crude oil and petroleum products. The crude, oil market was essentially a collection of contracts between
exporting regions, transporters, refiners and marketers which effectively operated as a decentralized open mar-
ket. 2 This was a holdover from the days when the oil industry was dominated by a few huge, vertically inte-
grated oil companies who manipulated prices and supply according to their own monopolistic desires. On the
Demand side, it was recognized that energy requirements were the driving mechanism and that "speakingabout
energy prices meant in effect speaking about the price of crude oil". 3 Demand for oil was derived from an
overall demand for energy. Energy demand stemmed from population growth and its concomitant effects on
industrialization and energy using equipment. Tins entailed understanding regional population growth, mecha-
nization and the availability of alternate forms of energy and their cost structures relative to petroleum. The
Gross Domestic Product (GDP) became one key measure in capturing the essence of this energy demand. It
was also discovered that international trade balances and currency exchange rates further complicated a re-
gional willingness to consume energy.
Complex Scenarios: Sensitivity analysis alone proved insufficient to understanding the petroleum market and
answering the questions of western leaders. Having seemingly identified the important factors comprising sup-
ply and demand, researches now began to construct scenarios in an attempt to model oil market behavior. In
my research, I located over 2,000 different books or publications within the University of Kansas' libraries
which specifically addressed petroleum price prediction and mathematical modeling. Nearly all of these refer-
ences had attempted to create models based on widely divergent assumptions and modeling templates and were
constrained by unique limitations. To properly review and understand the assumptions, detailed calcula-
tions, methodology employed and result» obtained fry thifr combined effort WQHM take several years and
significantly more resources than those available to perform the Part-I analysis of this study. A partial
review of these sources did provide useful insight however.
Initial Results: The bottom line resulting from the myriad modeling efforts developed by industry and acade-
mia during the 1970s is that the models developed were not effective at predicting petroleum prices. Useable
results from Petroleum price modeling remains elusive. To reduce complexity, most of the models examined a
single country or region of interest within a larger country. These models looked almost exclusively at the
short term; what was expected to happen in the late 1970's and early 1980's based on market conditions of die
late 1960's. Assumptions concerning political factors and/or OPEC policies were made and model predictions
were generated. The initial results were mixed as these models failed-to adequately describe the petroleum
markets, either before or after 1973. Further work was required to achieve adequate results or even create a
consensus among researchers.4 Additional attempts to aggregate the individual. regionaLmodels into a cohe-
sive, worldwide model failed completely. 5'6
Understanding Failure: The reason these scenario based models failed is significant and must be
understood if an effective simulation model attempt is to be made.
♦ First, failure stemmed from die fact tliattheearly models were based on time-series data pertaining-ta a par-
ticular region or country exclusively. Such models were overly simplistic, ignoring many important elements
comprising the dynamic crude oil market as a whole. Additionally, these models only strove to capture the
short-term elasticities in regard to supply and demand. 6 The concept that price changes set in motion cor-
responding supply and demand changes which experienced a significant delay in reaching a new state of equi-
librium was largely ignored. For example, if heating oil prices increase, it is not likely to affect heating oil
consumption during the upcoming winter season. However, higher heating oil prices, sustained over a five to
ten year period, will cause individual and industrial consumers to shift to cheaper alternatives such as natural
gas or to reduce consumption though better insulation or conservation efforts and thereby reduce oil demand
over time. Since these models attempted to capture day-to-day, week-to-week andmonth-to-month price fluc-
tuations, they broke down completely in the long run.
♦ Second, the nature of developing differing sectoral and/or regional models and then attempting to ag-
gregate them was flawed. The structure of energy demand access to transportation anddifferent crude cals
and products, rates of population growth and GDP, and the availability and cost structure of oil substitutes is
profoundly different on a regional basis. Complicate this with politics, disparate tax structures, tariffs andquo-
tas and the ability to meld different models into a single cohesive model becomes an impossible task. Several
models were marginally successful but were limited to only a single product^ like gasoline, within a small geo-
graphic region and applied for only a brief and selective time period. These models could not be generalized to
account for price variations worldwide or in the long-run. 7
♦ Third, limitations in data made working with many small countries or regions nearly impossible. Rarely
was data complete and accurate nor was the data conveniently available and consistently recorded to support
comparable regional models. s'f> Additionally, the data studied frequently contained relatively few data points
which contained a widely varying range in values. Such limited data yielded inconsistent and unreliable
results. 7
♦ Fourth, researchers had basically been attempting to develop detailed models based on micro-economics.
The complexity of the real world made such models impossible to use. 5 For example, there is no single
price of crude oil. Saudi Arabian light is different from the crude oils of Mexico, the North Sea, the Alaskan
Reserves or West Texas. Not only did crude oils of differing quality command different prices, but the mix of
crude oil grades and volumes delivered to various regional customers was in constant flux. The regional mod-
els had in essence attempted to capture this data almost on a transaction-by-transaction basis. The microeco-
nomic approach to such complexities, even with the assistance of modern computers, proved toodifficult.
Macro-Economics: A Different Approach: Thepurpose of Ulis thesis is therefore to attack the crudeoilmar-
ket modeling problem from a macro-economic point of view. The idea is to identify a small number of
large scale, aggregate components to crude oil price and correlate them using available and consistent em-
pirical data gathered over a long period of time. Instead of attempting to validate a theoretical model wc
Mill use real data to construct and calibrate an empirical one. It may not be possible to definitely establish
causality but it is hoped that sufficiently strong correlations can be established to facilitate price predictions.
This will be accomplished in several stages.
Stages: In the first stage, all factors to be evaluated must be identified and discussed in terms of their expected
relevance to crude oil price. This will help identify what data to collect and what numerical factors to seek^to
correlate. In the second stage, appropriate data must be obtained. This data must be complete, as accurate as
possible and cover extensive periods of time. The third stage will review the qualitative aspects of the history
of petroleum to determine if or what events should be incorporated into the final model. In the fourth stage.
the various data will be correlated with, crude oil price over time This will be done on a factor by factor basis
and using a multiple regression analysis of all factors. Those factors which demonstrate a statistically signifi-
cant correlation to crude oil price will be retained for more discrete modeling during Part-IIof this evaluation.
Perspective: To put the Part-I analysis into perspective, an analogy is appropriate. The petroleum price mod-
eling efforts thus far attempted by academia industry and governments alike were extremely complex andat-
tempted to capture every possible effect. This is similar to attempting to understand the physical behavior of
every molecule of gas in a room full of air. The historical models tried to calculate in effect, the mass, volume,
charge, velocity, trajectory, kinetic energy, momentum, inter-attractive forces, etc. of each individual
*.!*#
gas molecule as it interacted with the others and then aggregate the results to predict how the entire room of gas
would behave. The complexities involved with capturing millions of pieces of data and accounting for millions
of interactions among the molecules is frankly, impossible, andthat is why these modeling-efforts failed The
goal of the Part-I analysis is to develop macro-factors which can be correlated with real data to provide useful
results. This would be analogous to measuring- temperature (I)„ pressure- (P) and volume (V) of the gas in a
room and developing a simple model such as PV = nRT , to explain the behavior of the system. R is basi-
cally a correlation "fudge factor" derived from empirical data, WithsucLarelationship,. we can measure a-few
simple parameters (such as P or V), and easily predict T without having to understand or measure the true in-
teractive complexity of the system. The IdeaL Gas. Law relationship has provedusefuLinpredicting the behav-
ior of gas systems and it is hoped that our analysis can identify a few appropriate macro-measures and associ-
ated correlation coefficients (fudge factors) to develop a similar model which can be used to effectively predict
petroleum prices. In effect we will try to construct an eclectic model which seeks to include a wide range vari-
ables employed by numerous other models andthen identify a. macro-variable which is simpler to rueasiue^and
use.
Scope: If macro-variables are to be identified and woven into apractical model a decision has to be made con-
cerning the scope of our efforts. In the Part-I analysis, all factors potentially effecting crude oil price will be
analyzed on two macro-economic levels. The first level will assume a single, unified Global Petroleum
Market. Under such an assumption, important factors t& be correlated such as GDP, Population, Proven Re-
serves, etc. will be global aggregates. The second level will assume the United States is a single, unified,
stand-alone petroleum market and U. S. aggregate data will be used. All relevant factors will be analyzed on
each of these two levels.
Reducing Variation: Throughout the history of the U.S. and World oil markets, other than free market forces
have occasionally been at work. The effects of noncompetitive pricing- and supply controls were typically par-
geted at specific companies or countries. This often distorted the open market supply-demand-price relation-
ships regionally. However, any cut-back in supply and corresponding price increase in one region was typi-
cally offset by a surplus and lower prices elsewhere. Utilizing Global (U.S.) aggregate data in a long-term
analysis should provide a smoothing out effect for this phenomenon. It should also be noted Üiat these events
were of a short duration as both rapidly increasing energy demand coupled with an ever expanding number of
new discoveries, producers and technologies has continued to drive conditions towards a free market. 8
One useful fact from past price-modeling efforts was the recognition that market fundamentals, worldwide eco-
nomics, and business cycle forces consistently overwhelm the efforts of monopolies, governments and cartels
(such as OPEC) which attempt to manipulate petroleum supply and price. 9 Basically, market fundamentals
are stronger than politics in the long-run.
Stage Ii Factors Of Relevance
Demand-Price Cycle: As demand for energy increases, relative energy shortages
cause energy prices to increase. Increasing prices have both short-term and long-term
effects. Such effects relate to energy demand through possible conservation and reductions in energy use or by
consumers turning to less expensive^ more, abundant substitute sources of energy. Such-actions, although slow
to occur, eventually reduce energy demand (or the growth in demand) and create a relative surplus of energy in
the market. This excess energy results in a- decrease in prices in general which in turn stimulates economic
growth. Increasing economic growth and prosperity eventually stimulate an increased demand for energy. The
cycle then repeats itself It should be noted that the short-term effects on energy consumption patterns, have
proved relatively unresponsive to price. In the long-term, however, changes in these consumption patterns, for
example, towards use of more natural gas and coal and less oil, will have a profound influence on worjd en-
ergy markets. 7 The component factors of energy demand which effect the cost of energy use are numerous
and intricately linked together. Each of these factors will now be addressed in turn.
Demand versus Consumption: Consumption is the amount of energy actually used by residential and com-
mercial activity. In terms of petroleum, consumption includes crudeoil produced from reservoirs (production),
draw-down of inventory stocks and additional petroleum supplied by refining operations (cracking), Petroleum
Gas Liquids (PGL) and coal-gasification, and shale oil processes. The vast majority of petroleum consumed,
however, comes from crude oil production with the other elements making minor contributions only. Basi-
cally, what is consumed each year is exactly what is produced from reservoirs with little exception.
Unusual Events: More oil cannot be consumed than is supplied, therefore the ratio of petroleum produced to
petroleum consurnedis essential 1.000. Afew minor deviations to this rule have been observedandare linked
to major inventory build-ups or draw-downs. For example, following the 1973-1974 oil crisis, the United
States embarked on a massive, crude oil stockpiling project called the. Strategic Petroleum. Reserve (SPR).
The SPR was intended to be a large crude oil resource which could be set aside and utilized to maintain oil
supplies and prices for the U.S. economy during, any future unforeseen production cut-backs in the world mar-
ket. The purpose of the SPR was to reduce the volatility of U.S. crude oil supplies. During the SPR build-up,
more crude oil was purchased (supplied) than was actually used worldwide, with the difference going into un-
derground storage. In this case, the production to consumption ratio was slightly greater than 1.000. If the
SPR is ever used to augment worldwide petroleum supplies^ more oil will be consumed than is produced and
the production/consumption ratio will be slightly less than 1.000. Since either of these events is extremely spo-
radic and of a very limited duration, they will be ignored in any long-term modeling-of petroleum-markets.
Price Drivers: Why discuss nuances in the definitions of production and consumption? We do this to distin-
guish demand from consumption and to show that consumption data is irrelevant to an understanding of
crude oil market price. Demand i& actually the sum of various economic pressures placed oivthe energy in-
dustry to provide energy resources. When demand isn't met, prices and production increase until the pressures
are reduced. At any given time, what is produced equals what is consumed. Consumers, however, may have
been willing to use and pay for even more resources, thereby making energy demands over and above that actu-
ally used. Production and consumption will be equal at some market clearing price in an open market.
Changes in demand, however, are what drives changes in price. Consumption itself is superfluous. We
now turn to those specific components of pressure which change energy demand.
Population: The effects of a growing population on. energy demand are- fairly obvious. More people require
more energy and the rate at which a given population is growing is- proportional to the rate of growth in energy
demanded. Population growth varies widely around the world, and is particularly different between highly de-
veloped, industrialized nations and developing countries. Any crude oil-price model must evaluate the impor-
tance of both population and population growth rate on price.
Gross Domestic Product. While population levels and growth rates are indicative of a society's energy .needs
they don't capture the-economic state within that society. Industrializedregions possess significantly more en-
ergy using (oil burning) equipment and consume far more energy per capita than do less industrial regions of an
equal population. Various methods to capture this economic state, include computing-energy consumed per per-
son, energy consumed per GDP, cataloging the number and capacity of energy using equipment, or measuring
the total energy consumed The most accepted measure of the economic state of a region (specifically coun-
tries) is the Gross Domestic Product (GDP). The GDP is representative of the consumptive and productive
capacity of a. country as well as that county' & relative wealth. The greater the- GDP, the greater the- wealth and
consumptive energy demand. Because the GDP is widely computed and utilized it is both an appropriate and
convenient measure of the industrialized element of demand we seek to analyze. We will use GDP in our
analysis. It should be noted that population growth is imbedded into GDP growth and explains about one-third
of changes in GDP, while two-thirds of GDP is explained by productivity factors. If both population/
population growth rate and GDP/GDP growth rate show a strong correlation to crude oil price, it may be pru-
dent to use only GDP related data to prevent over emphasizing population contributions.
Leading Economic Indicators: The U.S. Department of Commerce, Securi-
ties and Exchange Commission and the National Industrial Conference Board
collate and publish various time series data for a range of economic indica-
tors. These indicators range from production hours worked, to interest rates,
to stock prices, to building permits, to the number of unemployed These
indicators are used by industry and government analysts who attempt to pre-
dict imminent and long-tenn changes in the overall economy or in specific
industrial sectors from the signals these indexes send. More often, composite indexes (weighted averages of
individual index components) are used to smooth out random fluctuations in an individual index and reduce
false signaling. The composite indexes are grouped into Leadings Coincident andLagging indicators depend-
ing on when an index is expected to react to specific economic change. It may be possible to correlate these
economic indexes, which are related to GDP, to crude oiiprice. Since GDP conveniently assimilates an infi-
nite number of variables into a single, useable number, it is the preferred factor to use in our macro-
economic model. Attempting to correlate the many economic indexes available to crude oil price begins^ to
lead us towards the more detailed, micro-economic analysis which have historically proven unsuccessful. If we
find only weak correlations of crude oil price to GDP data, however, it may necessitate turning to the more in-
volved process of determining crude oil price correlations for the many economic indicators available in hopes
of finding.a stronger correlation.
Reducing Complexity: To perform an analysis of Population and GDP as they relate to crude oil Price-
Demand & decision has to be made concerning which, populations and GDP's to measure. The micro-
economic approach requires that distinct regions be identified with their individual population and GDP statis-
tics to develop demand data on a region by region basis. This approach is extremely complex and data, inten-
sive. The simple process of identifying a discrete region as requiring substantially different treatment from its
neighbors is actually not so simple considering how dynamic population growth and industrial growth have
been since the discovery of petroleum in 1859. Any given region would be expected to change its characteris-
tics regarding, population growth rates and degree of industrialization (as well as many other factors) several
times during a given interval under long-term analysis. The continual shifting of these regional characteristics
is almost impossible to capture, especially since these changes occur gradually. There is also the problem that
regions of a similar nature requiring comparable macro-economic treatment will not conveniently fit them-
selves within national borders. Additionally, the time-history data supporting such an analysis simply doesn't
exist for the large number of affected regions comprising a worldwide petroleum market. If we can't analyze
individual regions, then what? We are simply left with a macro-economic approach in which large scale^ ag-
gregate population and GDP data will be used for the long-term price analysis. A short-term analysis of popu-
lation or GDP effects is not prudent as these have historically been quite unpredictable. 10 Long-term, aggre-
gate data is more readily available and we are relieved of the complex task of melding every changing regional
circumstance into a cohesive model.
Observation: Reflecting on the Demand-Price cycle discussed earlier, it appears that population and GDP/
mechanization are always increasing andthat therefore energy demand must always increase. In practice ris-
ing energy prices often cause economic slow downs and recessions and a reduction in GDP (or growth in
GDP) which often counter-balances the energy demand increases associated with population growth alone. "
f%
Alternative Energy Sources: The crude oil and petroleum products
market does not exist in. a vacunrtL It is only one component in the
entire energy market. The availability and cost of alternative sources
of energy play a vital role in_ determining the price of crude oiL The
primary sources of worldwide energy are oil, natural gas, coal, hydroe-
lectric, and nuclear power. Additional minor contributions are. .made
by solar, wind and geothermal sources, wood and trash burning and
biomass energy. The. mix of type, and quantity of energy resources
used in a given region is highly dependent on the resources available
locally. Hydroelectric power isn't available where there is no. water.
One cannot burn coal unless coal is available. A country must possess
the technological andmonetary resources to construct a nuclear power
industry before nuclear power is available for use.
Balancing Forces: Some regions have few indigenous energy resources while others have both abundance and
choice. Some are net importers of energy while others arenet exporters. An equilibrium of energy exchange
exists in which energy is produced, distributed and consumed based on the relative cost of each type of energy
as well as the relative energy requirements of various regions. The technological, production, refining and
transportation costs associated with bringing each energy type into a region determines the overall cost of
that type of energy by region. For regions rich in coal andnatural gas but poor in petroleum, the relative cc^st
of oil energy will be substantially higher than for coal or natural gas energy. Such a region would be expected
to utilize relatively more coal and gas and less oil. This reduced use of oil would be an indicator of relatively
higher oil prices.
Limits On Substitutability: There are.limits on the substitutability between energy sources. While both coal
and oil can be burned to produce electricity, only oil canbe used to power automobilesandairplanesv Different
types of fuel and fuel burning equipment have different thermal efficiencies and different environmental im-
pacts. Attempts te capture thermal efficiencies for fuels, in the form of BTU's per energy dollar, have been
relatively unsuccessful.
10
Pollution: Societal response to varying degrees of pollution related to energy use remain difficult to quantify
but is significant. For example, rigorous economic analysis concludes that nuclear reactors are much more
economical than coal or oil for producing electricity. Nuclear power, however, has continued to make limited
contributions to the overall energy mix because of political and environmental concerns. n Additionally, burn-
ing of both coal and oil are coming under increasing environmental pressures for their contributions to air pol-
lution and possible global warming effects. This has led to a push to use more "clean" energies such as natural
gas, hydroelectric- and solar power.
Accounting For Alternative Energy Sources: An obvious method would be to compute a ratio of the
amount of energy consumed (by type) to the total energy consumed; oil BTU's/Total Energy BTU's for exam-
ple . If this ratio is relatively high, it indicates a greater relative importance as well as a higher degree of de-
pendence on that type of energy. We could therefore use the proportion of oil energy to total market energy
consumed as an indicator of the value and availability of oil and hence price. Such a measure would implicitly
include the cost structure and influence of alternative energy sources.
Time Lag: In the short run, energy consumption is not likely to respond much to price fluctuations-or political
or environmental pressures. A region will consume what it always has consumed In the long-run, however,
these pressures will drive energy consumers towards cheaper alternatives. 7 What this means is that a low en-
ergy market share is indicative of relatively more expensive energy costs. The difficulty lies in assessing the
time lag associated with pricing pressures and a subsequent change in eonsunnrtion pattern». Several
studies indicate that changes to energy consumption patterns aren't evident until 5-10 years after the pricing/
political/ environmental pressures begin. Comparable changes in production patterns occur much more rapidly,
and especially with regard to idle reserves and excess capacity which can be brought on line quickly. 10 Add to
this, the uncertainty in measuring the degree of change in terms of both conservation measures and use of sub-
stitute products, and the problem worsens. What measure then can we correlate with energy prices?
Predictions: Following the above discussion, a high proportion of current energy use is indicative of relatively
lower costs in the past (5-10 years previously). A correspondingly low proportion of use today indicates rela-
tively higher energy costs in the past. Although we may be able to correlate the rate and direction at which the
proportion of energy use changes in relation to past pricing pressures (accounting for the time. lag), this doesn't
help us predict future prices. It may be possible to make some reasonable future usage predictions (5-10 years
ahead) to model current prices. The real difficulty lies in projecting energy consumption needs 20 years out in
order to reflect pricing pressures 10-15 years out. This simply isn't reasonable. There are also significant
assumptions concerning the proportion of energy which will be supplied by petroleum, political and econoniic
stability, etc. which make such predictions highly suspect. The bottom line is that there simply isn't a clear,
convenient method to account for energy substitutes explicitly. For now, we must understand that there are
relatively un-measurable forces shaping consumption patterns and that adequate micro or macro-economic
11
variables are not always available to capture these effects. We acknowledge this limitation in any subsequent
correlation or model developed. It is hoped that any correlation coefficients (fudge factors) developed by the
use of an energy use ratio (BTU's oil/Total Energy BTU's) will implicitly and adequately capture, these uncer-
tainties.
Past Energy Usage: We have already hinted at this factor in our discussion of alternative, sources of energy.
Basically, population and the degree of industrialization de not change dramatically from year to year. Addi-
tionally, changes in consumption patterns in response to changes in energy costs occur slowly. It takes time for
conservation efforts to gain momentum and become-effective. Likewise, converting domestic and commercial
energy using equipment to use an alternate energy source requires time and money, and must be accomplished
in stages. This degree of past energy use is alse implicitly tied to a country's short-term dependence on that
energy resource. Energy use simply can't change much from year to year. The past energy use is analogous to
an object moving at a constant velocity. Rising prices tend to slow down usage while falling prices tend to
speed up usage. Price then provides the accelerating forces necessary to accomplish a change in demand.
Magnification: The more of a particular type of energy a region uses the more dependent on
that type of resource the region is. For a region that is heavily dependent on petroleum, for ex-
ample, it is unlikely that consumption will vary much from year to year. On the other hand, a
region which only gets a small percent of its energy from oil would be able do without this
source relatively easier and demand would be extremely price sensitive What this means is that the propor-
tion of energy use acts like a magnifying force to price fluctuations. When the proportion of petroleum to
total energy consumption is high, the price of petroleum will be quite sensitive to changes in supply or demand.
In other words, the more dependent on a particular energy source you are, the more volatile prices will be. The
ratio of energy use, discussed earlier, may be able to capture this price sensitivity and the energy dependence as
it relates to energy alternatives.
Oil Taxes: Oil Taxes take several forms. There are export taxes which are designed to provide revenues to
exporting governments or to deter the export of oil products. There are also import taxes which are designed to
provide revenues to importing governments, reduce dependence on foreign oil and protect domestic oil provid-
ers from global competitors. These taxes exist in varying degrees with regard to every exporter and importer of
oil energy in the world. The net effect is that consumers pay higher prices.
All Things Considered: The contribution to price made by taxes, although large, should be considered part
of the distribution cost of acquiring oil and oil products and not explicitly analyzed as a factor of pr^ce
movement. It is incorporated into the various energy costs and mix of energy resources used in a region as dis-
cussed previously. Another way of looking at this is if taxes were reduced or eliminated in one region new
demand and supply patterns would be established as part of a worldwide distribution equilibrium. If another
12
ß
region wasn't satisfied with the share of oil it received, it could reduce taxes (and price locally) and once again
realign demand in the worldwide distribution equilibrium.
Tax Policy Simplified: The fact that we are committed to analyzing price and demand-on a macro-economic
basis means that we have to ignore the individaat,, regional tax differences. The degree of complexity and
lack of detailed regional tax data on the myriad variations over time make a micro-economic analysis hopeless
anyway. Taken with the observation that the tax structures of most exporters and importers remains fairly con-
stant, individual regional tax anomalies are not likely to affect global oil markets.
Tariffs And Quotas: Tariffs and import quotas have essentially the same effect as taxes; they add an addi-
tional component cost to imported oil ancLoil products. Currency exchange rates are also part of the cost of do-
ing business. Using the same logic as applied to taxes, these incremental costs are simply jpart of the overall
energy distribution cost matrix and shouldnotbe analyzed separately.
Price Controls: Price controls are different. They are political in nature and designed to protect the domestic
oil industry or stimulate the economy. Price controls are employed to gamer political support and are not a
consequence of open market forces. The United States has been the primary region employing price controls
to any degree.13 During the 1960's, when crude oil was plentiful and world prices were relatively lowr theJJ.
S. artificially held oil prices high to protect domestic producers. In the 1970's, when world oil supplies were
constricted and prices were high, domestic prices were held somewhat lower. In both of these periods, U^.S.
consumption patterns relative to the global open market price of oil were distorted. It is difficult to know how
to quantitatively address this factor. The basic result was that there was a price off-set. The magnitude of
these off-sets generally remained about the same and in effect, became part of the overall energy distribu-
tion cost matrix encompassing taxes, tariffs and import quotas. Using previously described logic, individual
price off-sets will not be analyzed as a discrete factor affecting demand or price.
Climate: In the short run, climate can play a role in oil price. An unusually
long and cold-winter may increase regional demand for heating oils, driving up
__!».*. * iiL.AA» V local prices. Such imexpected deviations from established energy distribution ■'/////-' 7/7/7///' ■ //'/>//'/ ''• Wfi' vr///'/■'' ■' '7777/7//' patterns are extremely short-lived and should be of little consequence in con-
^ ■//■'/■'■■'■//')W/ '//'> ''//'/ 7, ■'77'///////'/////' structing a long-term oil market simulation model. Numerous models have
attempted to account for weather through the use of widely reported cooling degree/heating degree environ-
mental indexes. These efforts have proved unsuccessful and we will therefore ignore weather considerations.
Political Speculation: Throughout the twentieth century, uncertainty about the continued availability jof oil
supplies has prompted individual governments andindnstry to conduct speculative, oil buying-to assure them-
selves of adequate supplies. These speculative purchases were made without the use of credible economic indi-
cations of shortage or impending shortage, wereirraiional. and were therefore not part of open market forces.
13
In spite of this, they had definite and pronounced effects on the availability and price of oil for brief periods.
The short but vigorous price response to these speculative actions makes them more suited to the qualitative
analysis stage rather than to make an attempt to account for them in a long-term quantitative model.
War or other Disasters: War, regional conflict, economic depression, natural disasters or political collapse
can have definite impacts on the supply, demand and price of oil. The sporadic and relatively unpredictable
nature of these events makes an analysis of their impact better suited to the qualitative analysis stage.
Sup ply: Total crude oil supply is derived from discovered reserves, unproved reserves, inventories, crude oil
production, production gains from refining-cracking operations and petroleum-gas liquids (PGL). The assump-
tion is made that an increase in demand will result in a proportional increase in supply. There is some diffi-
culty in specifying supply or demand, for with numerous production andrefining processes, differing grades of
crude and a highly variable demand for light and heavy finished products, supply or demand could be inter-
preted in many ways (i.e., the supply of gasoline or the supply of Saudi Light Crude). The cost/price of these
various products varies widely. For our purposes, energy supply and demand will be equated to aggregate
crude oil supply anddemand, anda single^ aggregate, crude oil price wilL be used to represent energy prices
in general I4 This is in line with our intent to model the petroleum market as either a single global market or
as the U.S. market in isolation.
Supply-Price Cycle: When energy demand increases, demand for crude oil increases and crude oil prices rise
as consumers compete for scarce resources. Production and refining- facilities increase output to meet the..in-
creased demand. Petroleum reservoirs, due to their physical arrangement, have optimal extraction rates and
thus have production limits. Additionally, existing tankers, pipelines and refineries, with their tremendous
capital costs, are typically operated at or near capacity and little reserve surge capability exists. Pre-positioned
inventories can alleviate short-term shortages, but any sustained energy demand quickly overwhelms existing
capacity. The only remaining solution lies in locating and developing new petroleum reserves. Higher prices
result in increased exploratory drilling and increased production from existing reserves when feasible. 1S This
pressure leads to new discoveries and the construction of new production, transportation and refining capacity.
As may be imagined the time from a sustained increase in energy demand and crude oil price, to the time when
new resources are brought to market, is measured in years and in billions of dollars in capital investment. As
new sources of supply ramp up production output, the increased demand becomes satisfied and the relative
scarcity dissipates. As production capacity exceeds demand, energy/oil prices drop, making additional explo-
ration and drilling unprofitable. As exploration and drilling activity wane, the petroleum industry settles out at
a new level of optimal production output. Eventually, the pressures of increasing demand begin working until
a significant rise in prices is once again experienced. The great capital expense and significant time-lag in-
volved with locating and developing petroleum reserves precludes the petroleum industry from mirroring the
relatively linear increase in demand and therefore condemns the industry to a constant cycle of overshooting
14
JtM*
and undershooting needed supply.
Crude Production: The first and most important element of the supply
cycle is gross crude oil production. One might argue that Total Supply,
which included the products of refining-cracking petroleum gas liquids
(PGL), and coal-gas and shale-oil operations, in addition to gross crude
oil production, would be more appropriate since we are attempting to
capture the relationship of total supply and total demand in our model.
The difficulty in calculating total supply in this fashion is that reliable
data for production volumes for these secondary sources is not readily
available or consistently reported throughout the world. It is reasonable
to assume that if demand for crude oil were to increase by 10%, that
each of the components of total supply (gross crude production, crack-
ing. PGL, etc.) would each increase by that same 10% to meet the de-
mand It is also reasonable to state that the vast proportion (>70%) of
Total Supply is accounted for by gross crude oil production and that any
measurable shift in production would be completely representative
of a shift in Total Supply. Since we are trying to correlate a change in
Total Supply with price, we should be able to use gross crude oil pro-
duction as a reasonable proxy for Total Supply in our model.
Assumption: A key assumption in the use of gross crude oil production as a proxy for Total Supply is that the
secondary sources remain exactly that, secondary. If energy prices and demand increase to the point where
coal-gas and shale-oil operations and thermal and catalytic cracking and reforming operations become signifi-
cant and independent contributors to Total Supply, gross crude oil production alone may prove to be unrepre-
sentative by itself. For the purposes of this study, the assumption that gross crude oil production is an excellent
proxy for Total Supply is presumed to hold.
Considerations And Simplifications: In attempting to understand gross crude oil production, there are several
factors to consider. The level of production is constrained by the- physical size of a given reservoir and outer
geologic factors such as rock porosity, reservoir pressure and temperature, viscosity of the crude present, the
amount of gas and water present, etc. Crude oil production may be enhanced by polymer, steam, water or gas
injection processes. Crude oil production may be constrained by the number of wells which can be drilled due
to climatic or geographic hardship such as those involved with arctic or off-shore production facilities. It may
also be constrained by a limited distribution or storage system whereby available tankers or pipelines simply
cannot remove product to refineries or markets any faster. Price is also a significant factor. Unless the crude
oil can be produced profitably, oil will not flow. These variables are further complicated because they vary
15
considerably from reservoir to reservoir and change for individual reservoirs over time. There is no meaningful
way to capture the vast array of differences in production on a reservoir-by-reservoir, country-by-country or
region-by-region basis. Our only hope is to use aggregate Global (U.S.) gross crude oil production data in our
attempts to explain price. Once again we will ignore the micro-economic complexities in favor of a macro-
economic measure in our model.
Petroleum Reserves: Another important element related to crude oil price is the availability of known petro-
leum reserves. Reserves are an estimate of when supplies will run out given current product withdrawal rates.
If reserves are deemed low and petroleum is relatively more scarce, petroleum prices are expected to be higher
and visa versa. However, the use of petroleum reserve data to help explain price is not without complications.
Complications: Reserves are estimates and as it turns out, extremely conservative estimates. For a particular
reservoir, as deeper wells are drilled and as actual production and depletion rate information is compiled, a bet-
ter picture develops as to the total economic production potential of a given petroleum formation. For large
formations expected to produce for 50-100 years or more, an accurate picture is decades in forming. Thus, pro-
duction and reserve estimates for any given reservoir are revised extensively over the producing life of a given
reservoir. Early calculations tend to significantly underestimate the actual recoverable volume present and
thus, reserve estimates are constantly revised upwards.16
Price vs Reserves: Price also plays a major role in establishing the level of proved reserves. Proved reserves
should perhaps better be called reserves which ean be economically recovered at this time If supply ex-
ceeds demand and prices are low then only the most efficient (least costly) producing reservoirs will be allowed
to produce because only the cheapest oil can compete effectively in the energy marketplace. Such reservoirs
would typically have substantial natural pressure driving petroleum towards recovery wells (primary recovery)
and would tend to not involve significant and costly pumping, injection or secondary/enhanced recovery opera-
tions. The volume of reserves which could be economically recovered at this price would be understandably
smaller. If prices were to rise significantly, then less efficient reservoirs/wells could be brought online, secon-
dary and enhanced recovery techniques could be employed and even coal-gas and shale-oil operations begin to
make contributions. The-volume of reserves which could be economically recovered at higher prices is under-
standably higher. The problem is that the amount of oil in the ground or the technological capacity to recover
the oil hasn't changed, only the price. This means that price affects the level of reported reserve» and the
level of reported reserves affects price making it difficult to assign meaning to reserves estimates in our
model.
Price: To account for the circular influences of price on various elements which themselves influence price, a
method of relating past and current prices must be used. The best accepted method for doing this is to use a
price-to-price ratio, such as the ratio of prices from 1982 to 1981 and correlate this ratio with the 1982 price.
16
Constructing a model in this way will allow any coefficient to capture the price rate of change and price mo-
mentum effects as well as implicitly linking past price influences into the model. At this point there is no rea-
son to suspect one particular time4ag ratio wer another ( year-to-year minus one ratio vs year-to-year minus
five ratio, for example) and the current and previous year pricing data will be used in this analysis.
Reserves Dilemma 1: Over time, there has been a constant threat of depleting proved reserves and running out
of oil. Such speculation has boosted prices and led to a subsequent increase in estimated reserves. Proved re-
serves alone ignore the dependence of recoverabilitv on price. Annual revisions and extensions to proved
reserves have historically amounted to two-thirds of the total annual production in that given year. " Essen-
tially, there seems to always be a concern that supplies will run out, but as price increases, new sources, now
economically recoverable, are added in and we actually never run out! The bottom line is that proved reserves
are an unreliable estimate of actual recoverable petroleum.
Reserves Dilemma 2: Another problem in evaluating reserves is that reservoirs are not conveniently located
within geographic borders or in readily accessible areas. It is not always possible to adequately map out a
given reservoir given these constraints. There may also be a problem of double-counting if adjacent countries
report reserves for a reservoir shared by both countries. Strategic Reserves, such as the U.S. Strategic Petro-
leum Reserve (SPR), when compared to global reserves and are not significant and are not a factor in this
study.
New Discoveries, Reserves And Price: Another component affecting reserves was alluded to in the discus-
sion of the Supply-Price Cycle. Following a price increase, the amount of exploration and drilling increase and
previously undiscovered reservoirs are added to the global reserve total or estimates for existing reserves are
revised upwards. Studies indicate that the number and size of new discoveries are linked to price, the number
and size of past discoveries, and to the degree of new discovery in specific geographic regions. Studies attempt-
ing to link the degree of new discoveries (new reserves) to price increases in specific geographic regions has
not produced useable results. 4 The primary problem is that data is only available for regions encompassed by
national boundaries and that what is needed is data based on areas of similar geologic conditions. Data is sim-
ply not available to support a model to this degree of detail.
Pressing Ahead: Despite the significant shortcomings of petroleum reserves they do have an impact on
price. The best method for dampening out or smoothing over the regional volatility is to use Global proved
reserves (U.S. proved reserves) in our attempt to correlate reserves and price. We know that the absolute
value of proved reserves is inaccurate, but since the petroleum industry has consistently used this conserva-
tive approach in its estimates, we may vet find a meaningful contribution towards price over time in a mul-
tiple regression analysis including proved reserves.
17
Reserve To Production Ratio: Another factor which we may be able to relate to price is the Reserve/
Production ratio. Since such a ratio incorporates both production and proved reserves and is a genuine indica-
tor of the relative scarcity of crude oH given current consumption rates, it may prove better than either Pro-
duction or Proved Reserves in explaining changes in price. Since this ratio incorporates production and proved
reserves it carries with it the assumptions and uncertainties associated with its components. A similar regional
problem exists in attempting to capture Reserve/Production ratios for they vary from 100:1 down to 3:1 de-
pending on the particular reservoir or geographic region. I8 For this reason, only a Global (U.S.) Reserve/
Production ratio will be used, calculated from corresponding production and proved reserve data.
Inventories: In addition to Production, Proved Reserves and the Reserves/Production ratio, several additional
factors should be considered in attempting to capture the effects of petroleum supply on price. One might want
to consider inventories an important component but in reality they are extremely small given aggregate world
wide demand. Inventories primarily provide a small cushion ... a buffering volume allowing producers, refin-
ers, transporters and retailers to make optimal, cost effective use of facilities and distribution channels. Inven-
tories (with the exception of the SPR) should be viewed analogously to a cash drawer at a-bank, necessary to
accomplish daily transactions but insignificant when compared to the bank's true inventory of financial assets.
Bottlenecks: Experts have also attempted to relate refining capac-
ity to price. This makes sense in theory, for if refining capacity be-
came the limiting factor in providing adequate supplies of petroleum
products, this should be linked to price. However, if refining capac-
ity consistently exceeded refining demand, it would never prove to be
a bottle neck and therefore have no substantial impact on changes in
price. In fact, refiners, both large and small are quite numerous and
exist in a highly competitive market. Such a market earns only a
normal return on the capital invested and refiners are extremely effi-
cient at achieving just enough refining capacity to meet demand
without enduring meaningful periods of over or under capacity. i4
Similarly, the tanker and pipeline distribution industry is highly com-
petitive and does not subject the oil industry to unusual price fluctua-
tions. Basically, transportation (freight), refining (processing) and oil
taxes are simply added on to the base crude oil export price. Although these elements make up a portion of the
final crude oil price they are rarely involved in causally changing the price and therefore will not be examined
in this analysis. Although this is the rational approach, markets don't always think rationally. If production
and consumption levels increase relative to total refining capacity, then the market gets worried and prices tend
to rise. We can capture this factor by correlating price to a ratio of Production/Refining Capacity.
18
Exploration And Drilling: Experts have attempted to correlate petroleum price with the number of wells
drilled in a given time period (drilling rate) and the rate of new reservoir discoveries. As discussed earlier, the
number of wells drilled and the rate of discovery of new sources is linked to price fluctuations. This relation-
ship is somewhat time-lagged in that drilling efforts increase after prices start to rise and taper off after prices
start to fall. Both of these measures are a result of changes in price and not the other way around however.
More drilling does not necessarily result in lower prices although one would suspect that more drilling would
lead to the discovery of new reserves and new production sources, eventually lowering price. The time lag-b^-
tween exploratory wells drilled and their impact on proved reserves and price may be captured by correlating
price to the number of exploratory wells drilled some number of years previously.
OPEC: One final topic to address when considering supply issues is OPEC. In 1973, OPEC was able to
briefly reduce the availability of global petroleum supplies and influence price. It was able to do this at a time
when the quantity of reserves and excess production capacity was greatly diminished and when an increasing
demand was about to precipitate significant price increases anyway. 19 The degree to which OPEC was able
to alter prices is still highly debated. Over time, OPEC lias suffered from lack of cohesion and discipline as
a cartel and its production cut-backs have had relatively arbitrary effects on petroleum price in the long-run.
With individual OPEC members constantly cheating on production limits and with the vast increase in non-
OPEC discoveries, reserves and production capacity since the 1970's, OPEC can be considered to have a
loose influence on petroleum price at best and at worst negligible effects on price. 20 A confounding factor for
OPEC and an understanding of OPEC's influence is that if OPEC reduces production to raise prices it suffers
a reduced petroleum market share- and erodes its leveL of influence. The- more OPEC squeezes supply, the
more customers turn to alternative suppliers or meet energy needs differently and thereby deprive OPEC of its
desired influence. The important point is that any component in the supply chain which can restrict or be per-
ceived to restrict the flow of petroleum can act as a bottleneck and influence price. One quantitative method of
determining the significance of OPEC ou oil price is to calculate OPEC's share of total crude oil production
(OPEC's market share) to see if it has any correlation to price.
19
Stage II; Data
Key Factors: Through a reasonable application of logic, twelve key factors were identified and are suspected
to have a measurable influence on crude oil price.
Directly Proportional Inversely Proportional Population Gross Crude Oil Production Population Growth Rate Proyed Reserves Gross Domestic Product (GDP) Reserve/Production Ratio GDP Growth Rate Exploratory Wells Ratio of Energy Use Production/Refining Capacity Price/Price Ratio OPEC Market Share
The eight factors in the left column would be expected to increase price with an increase in the factor. We will
attempt to correlate these factors directly with price. The four factors in the right column would be expected to
decrease price with an increase in the factor. We will attempt to correlate the reciprocal of these factors (L /
Gross Crude Oil Production for example) to price. The correlations we obtain by relating price to a factor or 1/
factor is the obvious place to start, however, this tends to only capture linear relationships. It is also necessary
to attempt price correlations with non-linear derivatives for each of these key factors. In this analysis, we will
calculate the logarithm and the square for each, factor anicorrelate those calculated quantities with price as
well. In this way, we can determine if a non-linear relationship exists which may be stronger than or comple-
ment any corresponding linear correlation. These factors were selected as the best macro-economic variables
for which data is readily available. Our goal is to attempt to correlate these factors to crude oil price over a
long period of time to give our model viability. In assembling a. model from any correlations we uncover, it
must be remembered that most of the macro-economic factors and economic indicators we seek to use are often
reported time late due to the necessity to collect, collate and publish data This stipulates that our model will be
useful to the extent of predicting crude oil price levels and expected changes to those levels on perhaps a quar-
terly, semi-annual or annual basis. It will not be possible to predict daily or weekly price variations using the
macro-economic variables in our model. Our attention must now turn to identifying an appropriate period
(number of inclusive years) for which to obtain that data.
Data Requirements: An ideal period on which to base our model should be reasonably long (>20 years), have
consistent data available for the factors we seek to study and be homogeneous. Homogeneity is crucial in thai
we want our model to be contemporary, that is, to work now and for the next decade or two. Since oil emerged
into the market place in 1859, the petroleum industry has passed through many distinct phases as it evolved, to
the industry we know today. The current petroleum industry is what we seek to model and we must there-
fore exclude any periods for which the characteristics of the oil industry markedly differ from the con-
temporary system.
20
First Petroleum Era: Clearly the era prior to World War I is completely dissimilar to today's market. Dis-
coveries were based on blind luck, the industry was rocked by violent booms and busts and wild growth and
experimentation were the norm. Monopolies, such as Standard Oil, dominated the market and the independent
producers. The environment was never considered and the kerosene product was used almost exclusively to
light lamps. Several technological breakthroughs radically changed this infant oil industry. The invention of
the electric light bulb in 1882 paved the way for electricity to replace oil as the source of light for residential
and commercial customers. The invention of the internal combustion engine in 1896 revived the oil industry
and generated a revolutionary transformation in which the automobile replaced the horse.
Second Petroleum Era: The period between World Wars was largely a
transition phase. Anti-trust efforts broke up the monopolies and the large
oil companies sought to extend their influence into international markets.
Oil was discovered in new regions such as Persia, Mexico, Venezuela, Iraq
and Saudi Arabia. Foreign oil concessions became the dominant theme as
oil moved to the center of the energy picture. Oil burning ships, airplanes,
and power plants joined the automobile in their demand for more oil.
Economically, the world experienced the high of the Roaring Twenties and the low of the Great Depression.
The United States was a relatively passive player in world affairs. The difficulties in communication and trans-
port left large portions of the world isolated. During World War II, the United States emerged as the dominant
economic, military and political influence on world affairs. In the aftermath of WW II, much of the world was
rebuilt or dominated by democracy. The U.S. oil reserves and oil industry were critical to achieving this sud-
den transformation and economic growth. The power struggle for global dominance clearly shifted to a strug-
gle for dominion over oil.
mfc"r ;::"'. ;;r;::7::-''. .;;!■;:;:,';':, ..„...,.- Third Petroleum Era - Great Transition. Following WW II, the U.S. and
£' H iTwe t£ '% - the world began building roads and automobiles at record pace. The petro-
FjH IMMM' chemical and plastics industries emerged along with technologically ad-
^^*™fW^^^"PW^P?"^V vanced methods for the location and production of crude oil. Telephones,
,.:.,^^..:,,,.<^^^^5iw.,,,;,-;.,:J.-,:-: < televisions and jet airplanes closed the communication gap and more closely
linked disparate global economies. Superior economic growth resulted from massive infusions of oil energy
used to construct a modern industrial base.
Stupendous Demand And Change: Between 1948-1971, U.S. oil consumption rose 300%, European con-
sumption rose 1,500% and Japanese consumption rose over 13,000%. During this same period, the number of
cars in the U.S. increased from 45 million to 119 million. Outside America, the number of cars increased from
18.9 million to 161 million as the global automobile industry flourished.21 New roadways and a migration of
people to the suburbs was behind the ever increasing number of cars. By 1976, nearly 90% of population
growth in the U.S. had settled in suburbs, necessitating commuting, and thus, more automobiles.
21
The Interstate Highway Bill of 1956 provided for the construction of over 42,000 miles of new roads. An en-
tire industry of hotels, restaurants and service stations exploded across the landscape to support an increasingly
mobile population.22
A Shifting Repository Of Wealth And Power: Between 1948-1972, world crude oil production increased
from 8.7 to 42 MMbbls/day. The U. S. went from producing 65% of the worlds total crude oil to less than 22%
as developing countries in the middle east joined the production frenzy. The status of proved oil reserves
shifted from the U.S. controlling 33% of 64 billion barrels of crude in 1948 to controlling less than 7% of
nearly 666 billion barrels available by 1972. This was a time of great economic growth and of significant insta-
bility and transition for the petroleum industry. Concessions still dominated the foreign oil industry and price
was largely controlled by exploitive, long-term contracts. Crude oil import quotas were imposed by the United
States from 1959-1973 and kept U.S. oil prices excessively high to protect domestic oil producers. The oil in-
dustry was still far from an open market. Events, such as the 1956 Suez Crisis, in which Egypt nationalized
this critical trade choke point, set the stage for the downfall of concessions and for the emergence of an qpen
petroleum market. H
Crude Oil Price History
$80.00
$70.00
$60.00
^™*Crude Oil Price-Norrrtnat ($/bbl)
""—Crude Oil Price -1999 Dotlars ($/bblft
1860 1880 1900 1920 1940 1960 1980 2000
Years
Figure I: Source: Energy Information Administration, 1997 & www.orst.edu
22
Reflection: Figure I is illustrative in helping define the different petroleum eras. Crude oil price, both Nomi-
nal and inflation adjusted to 1999 dollars, is plotted over time. It is obvious that the period from 1860-1890 is
quite different from the 1890-1945 period. It is equally obvious that crude oil price patterns during the period
1945-1970 were substantially different from those between 1970-2000. 1970 corresponds to the moment when
OPEC overtook the U.S. and gained a 50% market share of the world petroleum production output. M
Comparison: Figure II depicts a comparison of aggregate world crude oil production since 1945. Aggregate
world production increases logarithmically until about 1970 when the aggregate growth rate becomes notably
linear. Figure III shows the percentage of gas wells drilled in the U.S. since 1945. The sharp increase around
1970 indicates a significant change in energy production and consumption patterns.
World Cumulative Crude Oil Production IBbbls)
900 - I
i
«T ; £ 600 - CO
'
] •■■■■■•World Cumulative Crude Oil Production
; (Bbbls) o 0. I E 300 -
u i I
j 1945 1955 1965 1975 1985 1995
Year
Figure IE: Source: Twentieth Century Petroleum Statistics, 1998
Percentaae Gas Wells Drilled To Total Wells
45.0% - - j
- i
30.0% - (A
I
2 25.0%
»I
O 20.0%
15.0%
|^™"Percentage Gas Wells Drilled To Total Wells |
10.0%
1945 1955 1965 1975 1985 1995
Year
Figure 111: Source: Twentieth Century Petroleum Statistics, 1998
23
Figure IV shows the percentage of U.S. crude oil imports since 1945. It also indicates a sharp break with the
historical trend around 1970. Figure V depicts the average cost per foot for drilling new wells in the United
States. A 700% price increase beginning in 1971 signifies a dramatic change. A reasonable hypothesis is that
a flood of new environmental protection laws such as the National Environmental Policy Act (NEPA 1970), the
Clean Air Act Amendments (1970), the Clean Water Act (1972), the Safe Drinking Water Act (SDWA 1974),
and the creation of the U.S. Environmental Protection Agency (1970) all created vast additional legal and envi-
ronmental costs for companies pursuing oil exploration and development. It is also reasonable to assume that
most of the easy to reach oil reserves had been found and that the preponderance of new oil reserves would ne-
cessitate exploring in more inhospitable and expensive locations.
Percentage of U.S. Crude OiHroports To Total Supply
Ö. 40.0
■Percentage of U.S. Crude Oil Imports To Total Supply
Figure IV: Source: Twentieth Century Petroleum Statistics, 1998
Average Cost to Drill New Wells in U.S.
$90.00
$80.00
$70.00
$60.00
$50.00
$40.00
$30.00
$20.00
$10.00
$-
■Average Cost to Drill New Wells in U.S.
Figure V: Source: Twentieth Century Petroleum Statistics, 1998
24
Figure VI shows total U.S. energy demand over time. For about 25 years immediately following WWII, U.S
energy consumption ramped up steadily and steeply but then nosed over sharply and remained relatively vola-
tile after 1972. Figure VII indicates U.S. economic growth since 1960. This data suggest a period of slow lin-
ear growth ending about 1972 followed by a period of much faster linear growth starting about 1974 and con-
tinuing until today.
Total U.S. Energy Demand
20
5 1? I
c i c
3 8
a
i
0
Total U.S. Energy Demand
Figure VI: Source: Twentieth Century Petroleum Statistics, 1998
10000
9000
U.S. GD P
. „ „
s-C
urre
i
Blll
lor
Q. g 3000
1000
19 60 1965 1970 1975 1980 1985 1990 1995 20
Year
00
Figure VII: Source: www.worldhank.org
25
Figure vill reveals that the U.S. population growth rate declined sharply until about 1969 and then transi-
tioned to a substantially more modest growth rate in the three decades that followed. It also reveals the world
population growth rate during the same period and shows a decade long stretch where growth rate peaked out
followed by a sharp break in 1971 and a steadily declining growth rate since.
World/U.S. Population Growth Rate
2.50% -
2.00%
| 1.50%
/*" *"-%•* "^S-
-- - " I
£ ^"N ^—. —»■"%«, *—■— v
j
rS XM_ —"N » i •■■■■U.S. Population Growth Rate (%)
x:
| 1.00% Ö
0.50%
So/ 'V ^_^"\ J ̂ "^ ^"^"World Population Growth Rate (%)
V^ **^^ ^ ■"•-■w *~~mS ^* •—■» *
0.00%
19 60 19 70 19 SO 19 90 2000
Year
Figure VflT: Source: Twentieth Centurj' Petroleum Statistics, 1998
Conclusion: An examination of each of the eight preceding figures cannot be considered exhaustive or imply
specific causality as to why the economic environment changed around 1970. The results and not the cause
are what is important. The decision to focus on data analysis for the period 1970-1999 and exclude pe-
troleum data prior to 1970 is monumental. Such a decision required that we devote a substantial effort to
convince ourselves that the petroleum industry can be characterized in a markedly different way after 1970.
The data support a strong conviction to this effect. For the purposes of this analysis, the current petroleum in-
dustry will be taken as beginning in 1970.
Stable Or Not?: The period from 1970-1999 has not been entirely stable, however. Oil had become a matter
of state policy and national security and represented power which could be used as a weapon. The Carter Doc-
trine, issued in 1980, clearly established that attempts to illicitly control oil would be viewed as a direct na-
tional threat to the U.S. requiring strong political and military response. 2<i The 1973 Yom Kippur War which
precipitated the first major "oil Crisis" is illustrative of how oil power was able to achieve political aims. 27
The International Energy Agency was formed in 1974 to help understand coordinate and moderate the effects
of global oil pressures. The concept of an "Energy Crisis" was conveyed to citizens the world over to establish
a direct link between oil and prosperity. Oil was no longer a commodity it was^ a vital necessity to sustain-
ing a way of life.
26
Changing Technology & Markets:. Particular to the
modem petroleum era, off-shore drilling and horizontal
drilling were possible and profitable and 3D seismic imag-
ing techniques made the search for and evaluation- of new
oil reservoirs more exact. Oil discoveries in Alaska, Mex-
ico and the North Sea added huge production volumes to
non-OPEC producers. By 1980, non-OPEC producers had
overtaken OPEC in crude oil production, market share
(Figure IX). By 1995, a market share equilibrium had
evolved. This modem petroleum era was marked by sev-
eral large shifts in crude oil production market share and
corresponded to shifts in relative economic and political
power. This doesn't appear particularly stable.
Vf$
HNKx-
World Crude Oil Market Share
80.0% i " ■■ " ' ~
70.0% - I i I
^ £ 50.0%
3= 20.0%
! . % !; 0.0% |
OPEC Market Share
Non-OPEC Market Share
Figure IX: Source: Twentieth Century Petroleum Statistics, 1998
A Different Petroleum Market: Concessions had begun to fall apart in 1950 when a 50:50 profit sharing
deal was struck between ARAMCO and Saudi Arabia. In 1970, the world was at a 99% utilization rate for oil;
the surplus was gone and a sellers market prevailed 1S The initiative had passed to the exporters. In 1970, the
Shah of Iran was able to obtain 55% profits which paved the way for the Tehran Agreement in 1971 whereby
all oil exporters were to realize 55% profits. By 1977 the last of the original oil concessions was gone, sover-
eign ownership was permanently established and the spot market emerged to fill the void left by dissolved
long-term agreements. By 1980, deregulation had successfully lifted oil and gas price controls inthe U.S.
27
Figure X indicates how U.S. and global oil prices tracked from 1970-2000. Notice the large price offset be-
tween Arabian Light (34) (taken as representative of world oil prices) and U.S. domestic oil prior to 1980. Ini-
tially, U.S. prices were off-set higher but from 1973-1979 U.S. price controls held the price off-set lower. Af-
ter 1980, U.S. and world oil prices track much more closely indicating that open market forces were at work.
The oil production and pricing mechanism evolved from a rigid, exploitive system which gave advantage to the
buyers (1950) to an environment of relative shortage which gave advantage to the sellers (OPEC 1974-1978) to
an open market system in which buyers and sellers completed oil transactions on a relatively equal basis (1985-
2000). From this analysis of petroleum markets it appears that the period from 1980-2000 is most representa-
tive of the free market forces and that the oil markets underwent a transition between 1950-1979. However, the
period from 1970-1979 marked the most pronounced change towards competitive markets and it is reasonable
to include this period as being characteristic of the modern petroleum era also.
70 -
60 -
— 50
2. 40 - «i
£ 30 - GB a ? 20,
10
u. S./WORLD C RUDE OIL P RICES
"■""■"U.S. Domestic First Purchase Price
■"■"Arabian Light (34)
f /■"J ——^J I \ #J Jf \ \^ > A f^ *——S V/CV^ \A , jf\ / V^JCr ^^y »*NA
■"■"7 \ *—J 00 19 70 1975 1980 1985 1990 1995 20
Year
Figure X: Source: Twentieth Century Petroleum Statistics, 1998
Conservation And Environmental impacts: Once price controls were lifted in 1980, true energy costs were
felt by industrial and residential consumers. By 1985, conservation efforts had reduced total energy consump-
tion, a significant break from a century long trend of increasing energy demand M As previously mentioned
environmental concerns exerted ever increasing pressure on the petroleum industry, both in areas of exploration
and development and in spill control and liability. The Exxon Valdez disaster (1990) in particular resulted in
numerous regulatory burdens on the oil industry.
Characterisation: Although changes have occurred within the petroleum industry between 1970 and2000,
this period can be distinctly characterized by:
Oil being synonymous with power & an oil-power equilibrium being established
A move towards open markets, sovereign ownership and priee/quota deregulation
New exploration and production technology
Environmental awareness and conservation
28
These characteristics are particular to the 1970-2000 period and thereby constitute the modern petroleum era
we seek to study. In addition to the prominent shift in characteristics and trends which focused our attention on
the 1970-2000 period, there are other data concerns we must address.
Missing & Incomplete Data: Some data simply doesn't exist while other data is incomplete. We alluded to
more exotic regional data such as import/export taxes, rates of economic growth and energy use and exact
population estimates as being difficult or impossible to obtain for developing regions. Where data for basic
economic factors such as crude oil production or imports does exist it often hasn't be recorded credibly for
long periods of time. Older data tends to be more suspect and global data prior to 1945 is largely incomplete or
composed of rough estimates. Lack of credible and consistent data for basic petroleum industry variables tends
to constrain any meaningful analysis to periods after 1945.
Data With Different Units: Throughout the numerous published resources pertaining to thehistory of petro-
leum one can quickly find many different unit bases in use. For example, crude oil production volume is typi-
cally expressed as;
1,000 Tons of Coal Equivalent Metric Tons (MT) Million Barrels (MMbbls) British Thermal Units (BTU's) Cubic Feet (CUFT)
While data can be converted into common units, rounding- errors alone can account for significant incongruity
when attempting to compile a data sequence from multiple sources. For example, the American Petroleum
Institute (API) reports U.S. Proved Reserves in 1950 as 25.3 billion barrels while the Energy Information AxJ-
ministration reports a value of 25,268,398,000 barrels for the same year. The results are close but rounding
error and a differing number of significant digits alone can throw off an analysis. Significant variation will
result simply by changing data sources in mid-stream analysis and results will become more volatile simply
because different data sources were combined To avoid or minimize this effect, a single source of data for a
given time period, 1970-2000 for example, should be used. Lack of consistence data sources prior to 1945 was
a strong consideration in excluding the period from this analysis.
Reporting Source: Different data sources, such as the United Nations, World Bank, American Petroleum In-
stitute, U.S. Department of Energy, U.S. Department of Commerce, Central Intelligence Agency, eta have re-
corded and published petroleum data for different periods of time. In some cases there is overlap, when multi-
ple sources report for the same period of time while in some cases there is no credible data available for certain
time periods. For example, despite an exhaustive search, no global crude oil production data seems to be avail-
able for the years 1936-1937. Even when data is available it does not easily compare with reports from another
source. Production is not always production! Some sources report total supply or refining capacity under
29
Üie heading of production and include byproducts from
thermal and catalytic cracking and reforming operations^
petroleum gas liquids, gas oil. shale oil. inventory draw
downs and eastern block and soviet coniributionsT in add- u\i j #_J|
tion to gross crude distillation. It is often difficult or im-
possible to know which components were included or ex-
cluded from a particular reporting source. Some published
sources clearly indicate what components were included *""1 rfe.
while others fail to explicitly mention what the data is .If
meant to include. Another problem is data sources which. ]: v.: 'r_ j.
report consumption in the form of a major finished product 1^'
such as gasoline while ignoring the volumes of secondary
products. In some cases data is reported for major produc-
ers only, and secondary producers are excluded. fj
Discontinuity: Another danger when attempting to use multiple sources to complete a-lengthy data timeline is
discontinuities created simply by changing the data source. Unless resolved, such discontinuities may provoke
the analysis to conclude that something unusual happened at a certain point in time when in fact, all that hap-
pened is that the data source shifted. For example, examine the data for U. S. Proved Reserves from two report-
ing sources shown below; the API data for 1949-1979 and EIA data for 1977-1998. The three years of overlap-
ping data are shown below:
Year API EIA Change
1977 29.49 31.78 -2.29
1978 27.80 31.36 -3.56
1979 27.05 29.81 -2.76
Volumes Arc in Bhbls (Billions of Barrels)
The problem is that the overlapping years don't match up well. If your data stream had to transition from older
API data to EIA data, then there would be a discontinuous jump in volume by 2.29 to 3.56 billion barrels de-
pending upon the year in which the transition was made. This is Hhistrative of how mach of the data in the
petroleum industry' is recorded and of the danger of carelessly linking different data sources together
indiscriminately.
30
Data Gathering Pitfalls: How the data was gathered in the first place can be another cause for concern. A
prime example leads us to look at published proved reserves data again. This data is derived from surveys sent
though out the world. While respondents are given guidelines, there is latitude for wide interpretation as to
what to include. Some countries fail to update their reserve figures meticulously or at all and their reports are
useless. Others fail to respond at all and estimates must be used or that particular country's contribution left
out. Other countries substantially under report reserve levels in hope of driving prices higher. Often new dis-
coveries take years before being added into the total reserve picture. The bottom line is, even the source, and
methodology for obtaining data may be suspect for a particular published report.
Data Approach: Since we cannot reasonably go out and gather our own data we are constrained to use data
that is already published To minimize the possible negative effects discussed above we must choose both
our data sources and our period of study to ensure maximum consistency and continuity. This means re-
stricting the number of different sources employed and ensuring that continuous data is available for any one
data stream. This approach generally precludes using data prior to 1960 and taken with the earlier discussion
concerning the scope of the modem petroleum era we seek to study, leads us back to the 1970-2000 period.
You Get What You Pay For: It should also be noted that this is an unfunded research process and that all data
used was data that was freely available. There are agencies, such as API, and private companies, such as the
Oil and Gas Journal and WRTG Economics, which possess proprietary petroleum data which can only be ac-
cessed for a substantial fee. The U.S. Department of Commerce also publishes many important economic sta-
tistics which may correlate well with crude oil prices but which can only be obtained for a fee. One example is
time series data for the Total Direct and Indirect Costs of Federal Government Regulation. Such data pertains
to the costs of externalities imposed on industry and consumers by regulators such as auto emission standards,
packaging and labeling requirements, worker health, safety and pollution laws. We already noted in Figure V
a tremendous drilling cost increase which coincided with stricter environmental regulations in the early 1970s.
This leads to the conclusion that there may be significant Department of Commerce, API or other proprietary
data which could be used to establish much stronger correlations to crude oil price than the data which is freely
available. The quality of this data is presumed to be superior to the data used in thi» analysis.
Reminder: It must also be remembered that we are not trying to capture the exact causalities and inter-
relationships of all variables involved. We have taken the practical approach which seeks to use reasonable
macro-economic proxies which have a meaningful correlation to crude oil price. If the trends for the proxy
factors can be related to the trends in crude oil prices, then exact data is not needed. The Part-II analysis
can account for factors we failed to include or didn't recognize and inter-relationships or variables too complex
to model effectively by using properly adjusted correlation coefficients (fudge factors) in our empirical model,
if only we can establish that some relationship exists. To that end we press forward.
31
Raw Data: Exhibit I (5 U.S. Data Tables) and Exhibit II (4 World Data Tables) contain the raw and calcu-
lated data in tabular form for the variables and derivatives we seek to correlate to crude oil price. A 29 year
interval (1970-1998) is taken as the modem petroleum era with the unique characteristics we seek to mode}.
Although 29 data points in a data series (and in some cases as few as 21 data points) may lead one to question
the statistical significance of the results, we will have to accept this uncertainty in our model. A substantial
case has already been made which demonstrates that data prior to 1970 is expected to pertain to a different pe-
troleum era than that which we seek to model and is therefore useless to us. As time goes on, additional data
will become available which can be included in the modeling effort and which should reduce or eliminate the
statistical limitations associated with a limited number of data points.
Data Sources: The table below indicates the raw data source used for each basic factor found in Exhibits I
and II. U.S. First Domestic Purchase Price ($/bbl) provided by the Energy Information Administration, was
taken as a good proxy measure of composite U.S. price. Arabian Light (34) is often considered to be a repre-
sentative global crude oil price and was used as a proxy for a composite global petroleum price. Nominal
prices have already been converted into current dollars (1999) as shown in both exhibits.
Factor U.S. Data Source World Data Source j
Price EIA 20th Century Petroleum Statistics, 1998
Population www.worldbank.org www.worldbank.org
GDP www.bea.doc.gov www.worldbank.org
# Exploratory Wells drilled
Oil & Gas Journal; 31JanOO, p-64 No Data
Production 20th Century Petroleum Statistics. 1998 20th Century Petroleum Statistics, 1998
Oil/Total Energy Use
Oil & Gas Journal; 25Jan99, p-58 Dept of Energy
r Proved
Reserves 20th Century Petroleum Statistics, 1998 Oil & Gas Journal, variou!r(1970-1999)
Refining Capacity
Oil & Gas Journal, various (1970-1999) Oil & Gas JournaL editions (1970- 1999)
OPEC Market Share
Not Applicable 20th Century Petroleum Statistics, 1998
When examining Exhibit I and Exhibit II, note that each key factor and its corresponding data series is indi-
cated with yellow high-lighted columns. Non-htgh4ighted columns contain input data or calculated derivatives
for each key factor (the logarithm or square function). The Price/Price ratio was calculated using the given
price data where the ratio for the current year was calculated using the following formula; P/P Ratio = PNI PN-
i ; where PN represents the current years price and PN.i represents the previous years price. Population and
32
GDP growth rates were calculated from raw population and GDP annual figures. The reciprocal of the number
of exploratory wells drilled of production levels, of proved reserve levels and of reserve-to-production ratios
were used as the prime factors due to their inverse relationship to price. The purpose of using the reciprocal of
the factor and its non-linear derivatives was to generate positive, linear regression results during the statistical
analysis phase. No global exploratory well data could be found and this factor will not be analyzed in the
global model. The Reserve/Production ratio was calculated from the proved reserve and production data given.
When it came to locating refining capacity data, most sources only included the crude distillation component of
total refining capacity. Since this only represents about 60-75% of total refining capacity it was not representa-
tive of the data needed. The Oil & Gas Journal (various issues) did report comprehensive refining capacity
data which included thermal and catalytic cracking and reforming capacity in addition to crude distillation ca-
pacity for both the U.S. and the world. Unfortunately, this data was only reported for the period 1970-1990 and
no other comparable comprehensive data could be located. We are therefore limited to this 21 year period in
calculating and analyzing the Production/Refining Capacity factor. The Oil Energy Use Ratio was calculated
by dividing the magnitude of oil usage by the magnitude of total energy usage. World energy use data was
only available for the period (1973-1996). OPEC market share was calculated as the ratio of OPEC crude oil
production to world crude oil production. Also note that no logarithmic derivative for the World GDP Growth
Rate factor was used because the logarithm of a negative growth rate is mathematically meaningless. For tins
factor, we relied on the square-function derivative alone to capture any non-linear effects.
33
%J#ij
Stage 111: Qualitative Analysis
World Oil Market: Upon review of the petroleum mar-
ket from 1970-1998, hundreds of individual events can
be identified which have potentially impacted petroleum
prices. The vast majority of these events, such as pipe-
line interruptions, refinery fires, new oil field discover-
ies, foreign nationalization of private oil company assets
or unilateral price/production changes, have had rela-
tively negligible immediate impact on crude oil price.
The aggregate effect of these events undoubtedly is important to the overall balance of forces which shape
price, but the individual effect is often impossible to model. For example, throughout the second half of 1999
and into the first half of 2000, a series of events occurred which drove petroleum prices significantly higher.
The sequence began with several U.S. oil refinery fires in the summer of 1999 followed closely by a pipeline
explosion near Seattle, WA and the destruction of a major Turkish refinery by an earthquake These refining
capacity setbacks coincided with the revival of the Asian economies after an 18 month recession. As 1999
wore on, the increased Asian energy demand and an exceptionally long and cold winter in the U.S. caused
prices to increase. Finally, sensing an opportunity to achieve some short term price gouging, OPEC initiated a
modest production cutback to push prices still higher. The point is that any single event would prove difficult
to correlate to price change since there are numerous events occurring simultaneously and which may affect
price in contrary ways and over differing time intervals. A new oil field discover}' in one part of the world may
be offset by a catastrophe elsewhere, with the net result indicating no change. If we are going to incorporate
this qualitative type of data into our model, we must confine ourselves to include only those events in which
crude oil price demonstrates a discretely measurable and singularly significant response to the event. We now
turn to those events.
Major Events: After carefully studying the events of the last three decades, I conclude that there are three
types of events to consider; Extraordinary Events, Significant Events and Aggregate Events. For the purpose of
this analysis, Extraordinary Events are defined as causing a near 100% price increase. Significant Events are
defined as causing a 25%-60% price increase. All other events are considered to be Aggregate Events for
which a cumulative net impact may have an influence on price, but for which an individual contribution is not
easily measured. For this analysis, Aggregate events will be ignored.
Extraordinary Events: Between 1970-1998, five Extraordinary events occurred which must be considered.
• The Arab-Israeli Yom Kippur War and subsequent oil embargo of 1973/1974. This was the first oil crisis and caused nominal prices to rise 400% in four months.
34
%***&*
• The Iranian Revolution and deposition of the Shah in 1979 & the start of the Iran-Iraq War in 1980. Al- though two separate events, it can be considered one continuous crisis within Iran for our purposes. It is con- sidered to be the second oil crisis and caused nominal prices to rise 260% over one year.
• OPEC institutes a series of pricing schemes in an attempt to regain significant price influence on the market in 1986-1987 resulting in a 175% price increase. The effect was short-lived as non-OPEC producers quickly increased production to return to open-market conditions.
• The Exxon Valdez oil spill in March 1989. Although the effect was short-lived, nominal prices spiked up 160% as a result of uncertainty about the flow of Alaskan oil.
• Iraq's invasion of Kuwait in 1990. Although the conflict was settled within six months, nominal prices spiked up 220% as a result of uncertainty about global oil flow, the result of the third oil crisis.
Throughout the modem petroleum era, an Exceptional Event occurs about every 6 years and caused prices to
spike up about 240% on average.
Significant Events: Between 1970-1988, five Significant Events must be considered.
• Supply glut from 1980-1982. Nominal prices fell 25%.
• Open market forces take effect and spot prices are accepted globally in 1985. Nominal prices fall 60%.
• Dissolution of the Soviet Union in 1991. Nominal prices rise 25%.
• Supply glut from 1996-1997. Iraqi exports add to OPEC production increases to fuel global economy. Nominal prices fall 50%.
• Asian economic crisis of 1998 reduces worldwide energy demand. Nominal prices fall 33%
Throughout the modem petroleum era, an Significant Event occurs about every 6 years and causes prices to
change about 40% on average.
The Model: In order to attempt to quantitatively capture the effects of these events, a erode correlation method
will first be employed. Events will be assigned a dimensionless measure corresponding to the magnitude of
price change typical for that event category. A plus (+) sign indicates that an event will increase crude oil
prices and a minus (-) sign indicates that the event is expected to depress prices. An Exceptional Event will be
assigned a magnitude effect of+/-5.0 (on a scale of-5 to +5, a Significant Effect will be assigned a magnitude
effect of+/-1.0 and all Aggregate Effects will be assigned a magnitude effect of 0.0. The magnitude value for
each event will be matched with the year of occurrence or the year in which the majority of the event occurred.
These crude proxy data will be correlated to crude oil price over time. In this rough way we will attempt to de-
termine the proportion of price which can be explained by Exceptional and Significant Events. If a correlation
can be established then an iterative process attempting to match probability distributions and magnitudes can be
used to refine the analysis in Part-II for both Exceptional Events and Significant Events.
35
Crude Quantitative Assigned Values for Qualitative "Special Events"
Year Magnitude Effect
1970 0
1971 0
1972 0
1973 +5
1974 +5
1975 0
1976 0
1977 0
1978 0
1979 +5
1980 +5
1981 -1
1982 -1
1983 -1
1984 0
1985 0
1986 +5
1987 +5
1988 0
1989 +5
1990 +5
1991 +1
1992 0
1993 0
1994 0
1995 0
1996 -1
1997 -1
1998 -1
36
Stage IV: Analytical Results
Data Randomness: The first step used to analyze U.S. and World petroleum market data was to perform a
statistical Runs Tests and generate a Control Chart for the U.S. and World price data series. The results of both
tests revealed that the data was non-random. This confirmed our qualitative observations indicating that the
period from 1970-1998 was not a neatly packaged event comprised of conveniently homogeneous forces and
corresponding data. It also indicated that the extraordinary and significant events previously discussed may
prove to be more influential on price than the other key quantitative factors. The Part-I analysis seeks to deter-
mine if a relationship exists between price and the key factors or their non-linear derivatives and was not con-
cerned with capturing the exact magnitude of particular model coefficients. Therefore, there was no serious
concern about the non-randomness of the data at this point in the analysis.
Individual Regression Analysis:
Each key factor and its associated
non-linear derivative was individu-
ally correlated with price using a
statistical regression analysis soft-
ware package. A summary of the
results is shown in Figure XI at
right. For detailed statistical re-
sults for each individual factor, see
Exhibit III (U.S. Market Results)
and Exhibit IV (World Market Re-
sults). In Figure XI, the R-Squared
statistic indicates the amount of
price that was explained by the fac-
tor on the left-hand side and the P-
Value indicates how confident we
are that the factor's contribution is
significant. For our purposes, the
higher the R-Squared statistic and
the smaller the P-Value (preferably
zero) then the stronger can be our
belief that the particular factor has a
significant effect on price.
Figure XI: Individual Regres- sion Results Comparison
ua US. World VUxld RSquared p-vaue
4.3%
p-vaue Price/Price (P/P) Ratio a6% 0.13
LN of P/P Ratio ao% 0.14 Square of P/P Ratio a8% 013
Population LN of Population
Square of Population Population Growth Rate
LN of Population Growth Rate &3% O.20 Square of Population Growth Rate &5% ai9
GDP 49% 0.25 LNofGDP
■'' :i
NA
Square of GDP GDP Growth Rate ao% a2i
TüM LNofGDP Growth Rate 48% 026 Squareof GDP Growth Rate ■
1/ExrJoratDryV\äls(Bty .■fiEii:^|||||;=;.§ NA NA LN1/EW NA NA
Square 1/EW .... ,; ~-±-&- IW II NA NA 1/Production(PROD) ^'ftl'lVf;;'
LNaM/PROD Square of 1/PRDD
Oil Energy Use Ratio (OEUR) &7% 0.26 LNofOEUR a8% 0.22
Square of OEUR gp'I'K/li 1/Proved Reserves <PR)
LNofWR
in ,
^ßäHEI
pSSp
SquareoffflR 1/Reserve-PrDductJon Ratio(FPR)
LN1/RPR Square 1/RPR
Production/Refining Capacity Ratio(PRCR) eßmlim^Sß^SmmS/iß^Mmn- LNofPRCR .< , iir'l
■:■'■■
-. Ill' U".
Square of PRCR OPEC Market Share NA NA
LN of OPEC Market Share NA NA Square of OPEC Market Share NA NA '" '?:>
Special Events ■■•.-.i;:;,:.-:,':
37
Results high-lighted in green indicate factors which had a strong influence on price while those high-lighted in
yellow indicate only a weak influence. Results high-lighted in red indicate no meaningful influence on price at
all. In addition to the factors included from Exhibit I and Exhibit II, price was also correlated with our table
of magnitude effect levels for the extraordinary and significant qualitative events discussed earlier. These re-
sults are listed at the bottom of Figure XI under the factor heading "Special Events".
Interpretation of Individual Analysis Results: The initial step was to observe the similarities and differ-
ences between the U.S. and World market results. The previous price, population, population growth rate,
GDP growth rate and ratio of oil energy use to total energy use had essentially no influence on price for either
model. GDP had only a weak correlation to price but the square of GDP had a remarkably strong and signifi-
cant influence on price for each model. We conclude that price is affected by GDP in a non-linear way.
It was somewhat surprising that previous price levels appeared to have little effect on subsequent price levels.
Perhaps this indicates that the petroleum market is more sensitive to supply-demand pressures and less sensi-
tive to momentum and perception titan other commodities. The fact that population levels appear not to influ-
ence price while GDP does is suggestive that it is the relative wealth level per capita, and thus the energy con-
suming potential per capita, that is important in driving energy demand Also surprising was the lack of influ-
ence of the oil energy use ratio. Increased dependence on oil should correspond to relative price insensitivity
and thus higher prices but this is not borne out by the results. Perhaps this ratio is too simplistic and simply
fails to adequately capture the effects of petroleum substitutes and their influence on price. This aspect will
require further investigation.
It was interesting to note that production levels had a remarkably strong influence on U.S. prices but negligible
influence on World prices. Conversely, U.S. proved reserves had little impact on price while World reserves
made a significant contribution. No clear explanation presents itself as to why this could occur. It may be that
our assumption that the U.S. market could be treated as an independent and isolated market is flawed. The U.S.
market may be so integrated into the world market that separate analysis is not possible. If this is the case, then
our final conclusions should be drawn largely from the world market analysis since U.S. market results are
likely to be distorted.
The number of exploratory wells in the U.S. had a profound influence on price. Although no World explora-
tory well data could be located, the strong U.S. correlation should convince us to redouble our efforts to locate
data for this potentially important global factor.
38
As expected, OPEC market share had a strong
correlation to crude oil price. Also, the Reserve/
Production ratio and the Production/Refining Ca-
pacity ratio had the most influence on petroleum
prices for both models. Our logic concerning the
effects of a relative scarcity of reserves compared
to production levels and the relative scarcity of
refining capacity compared to production levels
seems to be borne out.
Finally, the "special events" factor appeared to
have no correlation to crude oil price for either
model. Our crude magnitude effect modeling at-
tempt was ineffective at capturing the effects of
extraordinary and significant events. A better
model must be devised and tested since "special
events" most assuredly impact prices.
Multiple Regression Analysis: After eliminating those factors which had negligible effects on price (red
high-lights in Figure XI), a multiple regression analysis was performed on all remaining-factors and their de-
rivatives. One problem encountered was the differing time periods covered by different factor data series. For
example, production data ran from 1970 to 1998 while exploratory well data ran from 1972 to 1998. In order
to combine different factor data series into a single regression analysis, an identical number of data points had
to be used for each factor. Thus, the factor data series with the fewest number of data points became the limit-
ing element determining which data was ultimately used. IiLthe example above, if both production and ex-
ploratory well data were included in a multiple regression analysis, data would be constrained to the inclusive
period from 1972 through 1998. If refining capacity data were also included then the overall analysis wouldbe
limited to the period from 1972 through 1990. Utilizing a mere 19 data points begins to generate difficulties
when used as a basis for generating significant statistical results. This also undermines our intention of using^a
long time interval to base our modeling efforts upon. The only viable solution is to obtain more complete and
higher quality data which completely covers all time periods of interest. Such data was not available for this
analysis and any conclusion drawn from continued analysis must keep this limitation in mind.
Multiple Regression Results: Several different regression runs were made using combinations of factors giv-
ing ever smaller inclusive periods of data coverage (1970-1998, 1972-1998, 1972-1990,). Surprisingly, the re-
sulting factors which had a significant price influence remained largely unchanged as the model period was
truncated to include more factors. The best results came from regression runs which used all of the factors
39
and thus had the shortest inclusive data period (1972-1990). The U.S. and World multiple regression results
are shown below:
U.S. Model
Overall R-S(marcd = 98%
Factor P-Valuc
SQGDP 0.002
LN1USEW 0.000
1RPR 0.004
PRCR 0.006
SQPRCR 0.009
GDPGR 0.017
World Model
Overall R-Squarcd = 94.7%
Factor P-Valuc
SQGDP 0.004
1PR 0.002
PRCR 0.077
OPEC 0.000
LNOPEC 0.000
GDP 0.003
U.S. PRICE - 109 4 0.668 (SQGDP) - 46.3 (LN1TISEW) - 471 (1RPR) - 715 (PRCR) I 910 (SQPRCR) - 73 (CI)POR) 1 WORLD PRICE = -1684 + 0.365 (SQGI)P) + 79.2 (1PR) - 69.6 (PRCR) -18.6 (OPEC) + 670 (LNOPEC) -10.7. (GDP)
The actual coefficients resulting from the regression analysis were not important. What was important was that
six U.S. factors and six World factors were identified as having a significant influence on price (low P-values)
as well as being able to explain 94.7%-98% of price when their effects were combined. The amount of price
variation which could be explained by these factors (9+.7%-9&%) was surprising given that this was our first
modeling attempt.
40
V: Conclusion
Assumptions and Limitations: Before recounting and consolidating the results of this Part-I analysis, it is
prudent to review the assumptions made and associated limitations placed upon those results.
• The Part-I analysis was an attempt to identify factors which had a measurable impact on crude oil price.
Once identified, these factors could then be examined more carefully (in the Part-II analysis) to create a model
which could be used to predict general price levels. It was assumed that the petroleum market could be con-
veniently captured through the concept of aggregate macro-economic factors and that such factors could be eas-
ily combined with a global or U.S. wide scope. The great simplification was that the market was entirely ho-
mogeneous and that differences in geography, national borders, languageT culture, currency, politics, regula-
tion, economic goals and collusion did not exist. This is obviously not the case. The incredible difficulty of
trying to account for the myriad complexities and interactions left no choice but to paint with a broad brush.
This modeling effort was never intended to predict daily, weekly or even monthly price variations. Such pre-
dictions are simply impossible given the complexities involved and the random and unpredictable nature of
daily events.
• We assumed that there was one single representative market price for one single type of representative crude
oil. In fact there are many variations in the quality and composition of crude oils and even wider variations in
the price of such oils throughout the world. The assumption more practically asserts, that although no single
price actually exists, all of the various prices for various crude oils will respond in similar and proportional
ways to the forces shaping the supply and demand for petroleum energy.
• Inventory levels and inventory build-ups and draw-downs were ignored as being of small influence when
compared to aggregate annual production and consumption volumes.
• Taxes, tariffs, quotas, currency exchange rates, price controls and individual price off-sets were ignored; es-
pecially as they pertained to regional differences.
• The effect of periodic weather changes was ignored.
• Gross crude oil production was assumeda good proxy for total petroleum supply. Secondary sources were
considered negligible.
• The geophysical characteristics of individual reservoirs was ignored.
• The U.S. Strategic Petroleum Reserve was ignored.
41
• The current petroleum era was taken to be 1970-1998. It was assumed to be homogeneous in terms of the
forces shaping the petroleum industry. In fact, the petroleum industry has never experienced a period of homo-
geneity. It has been in constant flux and continues to experience change in technology, regulation, and politi-
cal and economic influences and may never be able to be modeled as a simple, consistent and homogeneous
event for any substantial time frame.
• Tire statistical analysis capabilities and experience of the author are limited and should not be construed as
being comprehensive or exhaustive.
• The data is severely limited. Numerous problems with the quality, quantity and availability of aggregate
data make results suspect.
• The time-lag associated with any key factor's influence on price was ignored.
• The use of factor derivatives such as logarithms and squares is by no means exhaustive It was simply an
attempt to see if any type of non-linear relationship might exist. There are an infinite number of mathematical
functions available (square root, cubes, ARCTAN, etc) which might prove to better represent the influence of a
factor on price.
• The use of aggregate factors or proxy factors using data that was freely available was employed to construct
an empirical model. There is no grand theoretical model on which to-base our efforts although much of our
reasoning was loosely tied to economic supply-demand theory.
Results: Using individual and multiple regression analysis for U.S. and global petroleum market data indicated
that the following factors had a strong influence on crude oil price (strongest to weakest):
Production/Refining Capacity Ratio
Non-linear derivative of Production/Refining Capacity Ratio
Reserve/Production Ratio
Non-linear derivative of the Number of Exploratory Wells
OPEC Market Share
Non-linear derivative of OPEC Market Share
Non-linear derivative of GDP
GDP
GDP Growth Rate
Crude oil production levels
Proved Reserve levels
42
Not only were important factors which influence crude oil prices identified, this process also eliminated several
factors which were initially under consideration as having potential influence.
Recommendations: After attacking this thesis project for many months it is now apparent that considerable
more time and effort will be required to complete the job adequately.
• I suggest revisiting the numerous modeling efforts made by other academic and industrial scientists in
greater detail. Each of these books/studies/reports may be able to provide new insights into the past failures
and point the project in a more productive direction. With over 2,000 references at the University of Kansas
alone, there simply hasn't been sufficient time to thoroughly review these works and develop a comprehensive
knowledge base on the petroleum market modeling subject. I am confident that a good start has been made and
that the factors identified above can be useful in constructing an empirical model in the Part-II analysis. There
simply hasn't been time to pursue this project further and tins project was probably better suited to a 3-4 PhD
program vice a 3 credit hour thesis project.
• A better quantitative model must be developed to capture the effects of extraordinary and significant events
which most assuredly have an impact on price. The crude model used in this analysis was inadequate. Perhaps
assigning event magnitudes which are proportional to the maximum price percentage increase for the event
would prove more effective.
© The Oil Energy Use Ratio did not have significant impact on oil prices and a better factor or combination of
factors must be developed to capture the effects of alternative energy sources on oil prices. Correlating key
factors pertaining to natural gas, coal, hydroelectric power, nuclear power etcetera may provide better results
and more fully capture the effects of alternative energy sources. The time-lag effect of changes in alternative
energy sources and pricing impacts must also be explored.
• The breadth and quality of data must be vastly improved. I feel that this can only be accomplished through
the purchase of highly reliable data from an authoritative source (API). This will also improve the sheer num-
ber of data points used and minimize the errors caused by using relatively small data sets. Department of Com-
merce data consisting of numerous leading economic indicators is another place to locate additional factors
which may correlate to price.
• Instead of choosing a single crude oil price (Saudi Light 34, for example) as our representative pric^ per-
haps we should consider calculating a weighted market average price, aggregating the price and volume infor-
mation for each crude oil product produced in a given year. Although such a price would be substantially more
representative of an overall global price, I fear that data availability may limit the usefulness of this approach.
43
• A method of quantifying environmental impacts should be attempted.
• Annual aggregate data may be too course of a measuring stick to adequately capture the qualitative forces
(extraordinary and significant events) which influence price. Such events are of a short duration and perhaps
daily or monthly data should be used in lieu of aggregate annual data. I am not sure if such data is even avail-
able but it may be the only way to gain sufficient sensitivity to account for the qualitative forces impacting
price.
• Inventory levels including the Strategic Petroleum Reserve (SPR) should be correlated with price to see if
any influence on price is evident.
• The time-lag effects of exploratory well drilling should be explored and global well drilling data should be
obtained and incorporated into the model.
Bring additional petroleum and mathematical modeling expertise into the project. More minds brainstorming
possible approaches to developing correlations will produce better results. Additionally, different combinations
of non-linear factor derivatives should be tried to identify an optimal non-linear relationship for each factor.
Final Thoughts: The Part-I analysis must continue in an effort to acquire better data, identify additional im-
portant factors of influence and to develop more sophisticated factor-models before a refined model attempt
(Part-II analysis) can be made. Tins analysis was able to make substantial progress into understanding the
forces influencing crude oil price, understanding the assumptions and limitations associated with a macro-
economic approach and in identifying several key factors which demonstrated a significant correlation to price.
44
(•- ai OK-' 0. a» 9: o a: feJÜ o V) § CO ^ 3 o- 3 x
tf> *" """ a CO
Di ^ <* a
1* o: w
z (0 3^ -J 3 ^—
z- _j
Q£ Di Q. 0-1 JQ£ cd (0 w 3 3 ■^ ■""j
c * .2 0£
siil 0. Di JIis 3
Q.
° Di o 0. » Q. O (0 « W O 3 3 _j T- T~
(0 C0]
<* M- oi 0- O O. co
5§ 3
z _1
Di Di a o-1 CO CO 3 3 "*■
.»••* w <fl
cd Q-
= 2Ii co 3
DC S
i_ LU n o >
CO
cd
eo CM
eo
o q ©"
IO to o ©
CO 5 CM
©
<o
d CO q CO
-* CO o» d
IO <o o
o> CO CO-
CM
o CO
CO IO o CO
«o
CM
CO
<o en ■*
oi
IO IO
CM
to CO
CM
«o eo q
© © IO
d © q CM
©
CM
CM
«o
©
■«*
ST-
to ©
©
o ©
d
o o o d
CD ©
r O O ■s*
CM
CO to CO
CM ■
CM © CO CM
i
o *•- CO
CM
© to CO
CM
CO
CM
CM 4» CM
<M i
o> CO
CM i
JO, CM
00 ■sl- CM CM
i
CM
CM i
00
CM i
0> CO
o> CO
CM 1
©
CM i
©
CM
© © «r-
CM i
© o> T-
CM i
© © CM CM
CM
© © CM
CM j
O CM
<M i
©
CM
CM i
CM CM CM CM
i
e» CM CM
CM 1
© CM CM CM
i
© © CM
CM i
© © ° ©
o o d
CO o © O) o d
CM © O) o d
io CO O) o d
© IO CO O) o d
CO
CD o> O
d
CO o CM O
d
CO
d
O
d
-1^ ■O m o
d
<o CO CO o
d
co CO
d
CO
d
«5 CM ■*
d
CO CM IO
d
«0
©
<o o CM
d
CM
d
o © o
d
■«t © CM O
d
© » © O
d
© © O
d
CM
cr> o
d
C0 » o
d
© ©
o
d
© ©
o
d
W ■<* ^» o
d
CO 00 q
o CM q ■A
d
to o W
d CO
d
o> CO d
o> CO
d
CO o>
d
CO
(6
o IO ^0
d
CM
d
TO
d
to CM CO
d
CO
d
CO m
»
© N. ©
d
» »
© CM w «
CM © ©
d
o CM ©
d
©
d
s T-
©
© o d
© © T-
©
© CM CM
d
© CM
©
© © CM
©
©
©
o o o d
© IO CO
ö
1»-
(0 d
CO CO m
d
§ CO ©
CO CM © CO
d
CO CO 0)
d
CO CO ■«* q
CM o CD o> CM
CM CM CM r-
o 05 CM
o CO in T-'
00 CO CO CM
«M CO a CO
© o> CO CM
o 0> © CM
a © © ©
©
CO
CO T- O) CO
CO
CM
CO
IO ■<*
©
©
■sl- CM
o o ©
© o 00
T- « w ©
CM
© o CM
©
CM © © to
r
CM CO CO CO
<?'
CO CO o> ©
©
to m
CO CO CO IO CO i
CO CO 00
«?
CM CO <er CO
1
CO
00 CO
<?"
CM -<» CM CO
CM CO o> CO
«?
CM CO O) CO CO
1
CO CO
9
o CM CO CO
IO *!" CM CO
*?"
© ■<* CO CO
1
©
©
•>*
CM
d ■
o © o CO
© 1
© CO © CM
« CM
<7
©
© CM
© © o CM
9
CO
«7
CM « ■r-
CO CM
©
o
«7
CM CM o> O
©
© © CM |o ö
CO CO CM o d
©
CM o d
CO CO CM O
d
CM © CM O
d
CD O CO a d
CO TM CO o d
CO CO o d
O CD CO q d
in CO CO o d
CO CO CO o d
o ^>* CO o d
o> «o CO o d
CO CO o d
CM TO CO o d
CM © CO o d
CM T^ CO O
d
© © o d
©
© o d
© q d
CO CO o d
© o o d
T- CM
O
d
CD CO
O
d
©
o d
-s*
O
d
©
o d
o d
i© © CO
CO CO o cej CO
O) CO CO
d CO
© o CO
© -CO
o IO CM
«O
-CM CO ©
CM ©
<M
d CO
CO CO
d CM
o CO
CM
o «0
d <M
IO o CO
d CM
CO CM
d •CM
CO IO 00
CM
IO CO
CM
©
d CM
d CM
© © CO
©
© w CM
r>-' CM
IO CM ©
© CM
r- © ©
© CM
© CM
© CM
CM © ©
CM
©
d CM
© ©
CM CM
IO ^,
CM CM
w « CM CM
O CM CM
©
©
CM CM
O Ir». © © ©
eo
© © ©
co en
CO
o>
o co co
o>
CM CO en
«o CO o>
CO o>
IO
o>
©
©
N. co e>
eo © ©
© © ©
O © ©
© ©
CM © ©
© ©
© ©
IO © ©
© © ©
1». © ©
© © ©
UJ — _l "S DO
X . UJ CO
3
Sq
uar
e o
f O
il E
ner
gy
Use
Rat
io
3 UJ O (0 3 a to
to CO ©
©
co © ©
d
ID
V CM
d
o
CM
d
io CM ^- CM
d
CO to
CM
d
co CM CM
d
CM N. CO CM
d
CM CO CO CM
d
co
CM CM
d
to CM o CM
d
co to co
d
CO CM eo
d
to
oo
d
to to
© ©
to
m
©
co CM to
d
© CM » ©
©
T"
d
© O f«. T"
©
© •<* ©
©
© © ©
©
CM ©
d
©
©
©
©
©
©
M ©
©
« © ©
d
M o ©
©
LN
ofO
il
En
erg
y U
se
Rat
io
Oi 3 Ol O to 3 z
CM 00 d
to
3 9 d
1
V) r«. d
i
xt tx
d
CO «
9
en
I*- d
en
o
CM CM
d
co IO
d
en en r«. d
en « co o
to CO
d i
CO to co d
i
o
CO
d
CM
CO
9
co co d
IO CO ©
1
© co ©
■
© © ©
9*
w 00 d
i d ■
© © ©
d
© o ©
9
O ©
1 d
i
CM ©
d i
© ©
© d
i
Oil
En
erg
y U
se R
atio
3 UJ O (0 3
o
d d
o
d
en co
d CO
d co
d
CO
d CO
d
to CO
d d
o to
d
CM co
d CM ■*
d
CO CM
d d
eo T-
d
CO CO
d CM
d CM
d
© CM
d
CO
d
© o
d
eo O ■<*
d o ■«t d
© o
d © CO
d
w © CO
d ©
d o
d
U.S
. T
ota
l E
ner
gy
Use
(101
2B
TU
s)
©
r- to
eo <* CO eo ©
CO
(O
CM co CM
r».
CO ■«t IO CM r-.
co ■* to ©
CM to co ^t
eo to «M to r-
en CO o «0
CO en co «0
to to en to
o en en co
CO
oo ©
CM IO ©
co a> to
© CM
r--
©
©
w
CM o w
w © CO V" ©
n © CM
©
» © T-
©
©
CM W
w © ©
© © © © w
© © CM o
o ©
© © ©
o © CM
©
CO w
.5« CO* o 3 ^-
© © CM
©
© © CO
to CO en CM CO
o
co ■<*■ CO
IO in
CO
T"- co
CM ©
to r».
CO
<M CM
1» ©
to CO en
co
«0 CM
to
<M o CM
CO
co en
«0
CO CM © CO
IO
CO
IO o
CO
<M CM en © CO
©
CO
to © © CM <o
CM CM CM *± to
T-
Of
co to © © ©
©
© CM ©
CM © © © ©
© «
©
© © ©
©
© w © ©
©
©
O
© © ©
£ £ (0 A. 3-OT
to ^
D O DC a. to 3
a to
CO -co o CO o d
CO CO CO CO o d
r- CO CO o d
CO •A CO CO o d
o r- co o d
CM ©
o
d
o CM
d
CO
o
d
o © CO CO o d
CM ©
d
o ©
o
d
CO
CM o
d
CO CO o o
©
«o 0> » o ©
CO ©
CO o ©
CD <M CO © O
©
CM © CO CO o ©
© ^M
O
©
r-- co CM
©
CO o CO CM
d
CO
r» CO
d
CM O to CO
d
CO CO ■<* ■<*
d
«0 IO
d
m © CO CO
d
© CM
o"
o r~
d
CO o
©
in o CM CO
d
LN
of
1US
PR
OD
a O Q: Q. to 3 r z
1- m CM
■
© CO CM
CO CM
i
CM CM
CM
i
CO
L 1
CO o
a CM
1
co CO in
CM ^o CO
1
co to o **
IO o> II- IO
I
CO to
i
CM © CO
i
CM CO IO
CO ©
o CO O) o
K CO CM o
CM
CO CO
9
lO t>. CO CO
9
CO CO CO
d
T" CM CD
9"
o> CO CO
9
CO 1-
co
9"
co IO CO CO
9
iO co © CO
9
o m CM CO
9
Q o a. <o 3
D O tu a. (0 3
© CO
CO CM
d
CO IO CO CO CM
d
o "* en CO CM
d
v> CO CM
d
to CM CM T" CO
d
->* f- CM CO
d
o © CO CO
d
to CO CO CM CO
d
-«t CO
CO d
xo CO © CM CO
d
v- 00
CO
d
CO CO a> ^» CO d
CO
^™ CO
d
xo CO tn
CO
d
*M
O CO
d
«0 CO Ui o CO
d
CO CO m CO d
©
C- CM CO
d
CM to CO CO d
m CM O) m CO d
CM T- V r-- co d
CO CO CO CO
d
■O o CO CO d
o ©
CO
d
o
d
o © ■t d
o CO o CM
d
CO CD ■<!■ CM
d
CO CM CO CO
d
U.S
. C
rude
O
il P
rod
uct
ion
(M
MM
bb
Is)
a O
a. <o 3
w
*^ CO
o> CO © ■*
CO
CO CO CO m IO
co"
CO o CD o CO CO CO
m CO m CM o CM
CO
CO
CO m © CO*
00
CO !•» CO CM
m to CM CO © o CO
CO
CM CO
co'
CO
*~ CO
CO r- It) CO -«T T-
CO
'S- CM CO CO CM
co"
in
r- to to
co"
CO © CO o
co'
CO CO CO O) "* CM
CO
CO
IO ■<t t- CM
CO'
CM m CM CO CO T~
CO
CO
© © CO
CM !■»■
CO CO CO
CM
CO © to eo
CM
m o> CO CM
IO
r- CM
m to
CM CO CM
O» w o> T- T- lO CM
CO
CO CO CO
CM
© to o
CM
CO
CO CM
©
© CO
CM"
O) T- o> co CM CM
CO m >
UJ Q O O
O
en 6»
CM t». en
CO
en en IO
en
CO
en en CO
en
en
en
o CO en
co en
CM 00 en
co eo en
eo ©
IO CO ©
© co © ©
eo eo ©
© © ©
o © ©
© ©
CM © ©
© ©
"It © ©
© © ©
© © ©
1». © ©
w © ©
co UJ
— _l
X . LU CO
<4- °5 U] £ ui co « CO 3 = 3 T-" 0* ^ o
CO <0
% ? UJ til (0 to 3 3 ^" T~ ■
z Z _l _l:
5 ^1 UJ UJ iß tOH 3 31 T-
B C? (0
£ 3 » ■2 .5 2 5 111
xplo
el
ls
(10 CO
3
wg <U o:
O Q. co —. Q-1 „DKO
Co . +J _ 3(0 J^ CO ff4 oi 3
a\ CO
.c <* »1 o
Q. 3 o a> Q
LN
of
GD
PG
R
a O CO 3 Z
S.
GD
P
irow
th
ate
(%)
O Q. D O
^üQ: CO 3-j
*ft a\ CO
£ o tt-l
= W O 57 -; CO w => 3j
CO 3 O. O * Q CO O (5 ^ z 2 _J _JJ
J-1,
GD
P
llion
s a. a] O
">£ CO 3
^ w
UJ CQ Q 2 ü
co IO is.
p d
tD IO CO
o d
CO
CM
p d
P d
■o
o d
xn o o d
<o CN 00 o p d
9) to «0 o o d
o to o o d
cn tN CO o o d
t». cn «o o p d
T- tN to O p d
CO
o p d
Is- to O p d
<M IO cn
p d
xs o
© d
eo
CM p d
CN
co o d
to cn rs. co p d
Is. to o to p d
Is. T- fO CO o d
o cn is, is. o d
to Is. o N. O
d
o CM Is. 00 p d
cn CO
cn o d
cn o d
IO cn
d
CM CM p CN
o IO p «si c4 «4
CN co -c* CN
i
*0
«N i
o o •* CN
i
*•> r>. *o
i
IO IO p CN
1
eo IO to
to to
c4
TO CN to
i CN
1
<M O p oi
i
CO to p
tM co p
eo
CO
1
to Is. p
to CO p en
<* i
co ■<*
CM
1
to Is. «M
CM p i
o CM CM
i
to cn
1
Is. o to p d
i
CN co
d
r- co CN
d
■o
d
IO o
d
CO
o
d
to eo en p d
o o> p d
CN co cn p d
p d
IO p d
CO to o d
CN CN r» p d
to to p d
cn
eo p d
cn CO
d
cn
■r-
d
IO Is. IO T-
d
Is. eo
d
eo ■*■
cn T-
d
lO CM CM
d
eo 00 CM
d
o>
CM
d
to to CM
d
CO IO cn tN
d
CO CM o p d
cn o CO
d
Is. co 00 p d
IO p is.'
ts. »>. isl
cn CO p CO
o> IO
o> p d d
CO p IO *o N.
d p CN
CO tN eo 00
d
to
p d
00 CO
d
00 o tN
to to
h~'
co o cn d
to p d
to co CO
d
CO CO
d
CO Is. to
d
CO 00 IO
d
to Is.
d
to eo
d
CO o p d
to CO CM
d
to 00 p CM
o IO tO
Is!
o o CN O
d
O O Is. p d
o is. Is. p d
o
o> p
o o to 1^
cö
o o to p CO
o o tN
d
o o o p d
o IO eo p N.'
o o CD oo «*"
o CN IO p
CO
to
d
o r». cn
O IO N.
o eo
p CM
o cn o
d
o
cn cn d
o CO
d
o CM O
d
CM IO cn Is.
d
o o eo p d
'T- IS.
CM
o CM
P d
o
co T"
CM
o 00 p CM
o CM CM
d
o o IO «* CM
eo co
■«t
CN
<o
o co CN
i
IO co
CN i
CN
P CN
eo co <o ■*
CN f
CN t
cn oo to
CO IO
o oi
■-
eN ■o
5 CN
CN 1
to
o tN
i
o ■»t CO CM
d
o CN to c4
CM T— CM
1
CM «N I«, to
CM CO
p CM
i
IO oo cn f.
CM i
oo 00 IO p CM
CM Is.
i
eo Is. oo p CM
eo to to
•?
cn eo CM
1
to
IO p CM
i
Is.
P CM"
1
CO
Is. p d
cn cn N. p CM
N. eo co eo CM
I
eo ■o o p d
oS
©
©
s?,
^- p eo
oS
p d
CN p d
1»- N.
d
CO CO CO p c4
CN p T-
0s- 00 p d
cn *™
tN cn d
p d
p d cn
d CO
d cn p d
Is.
Is!
CO p fs."
S5 r- IO
d
S5 CM p CM
oS IO IO
d CM
d
OS
P d
s5 OS CM to
öS 00 o d
©s to p d
os to cn
cn Is.
o CM
O
IO
en is.
oo
CO
CM
CM
co ST». IO
<M
CO
d
o> to o> p d
oo IO en •>*
o o>
^O*
00 CO p
CO r«. co «N d
to eo I-- cn d
o to
p
o fs. CO ro
o TOO 00 to
d T"
o is.
tM
d
o ^* o p d CM
O co o p d CM
o fs. Is. to
CM
o CO
p d CO
O
to
CM CO
o co CO p d co
o |s- T- «M
o Is. en
d
O tM Is.
d
o CM to
IO
o o «M
CO
o
|sl CO
cn oo © o ©
eo T" o> p ©
|s. 00
d
00 eo o» CN
d
oo co is. p ©
ov o to
d
co
p d
o o eo p d
co
o p d
IO
P d
CO CO cn p d
to
T1»
CN o IO
co Is. CN tN
CO p
co oo cn p
CN
IO
co o
p
cn
00 p
CO o to p
IO
Is.
CM t Is.
co cn Is.
Is. <* p
IO to o p
Is. o IO p
CM O O p CM
N. to O
«N
to o T-
CN
en oo © p
co en p
00 to o CN
CO 00
p
co r«. to <*
IO IO co p 1»;
0» co
p co CM CN
o eo eo
o cn o r>-
CO cn CO p d
eo IO
d
cn CN
d
to eo N.
d
CM oo ■»* o
eo co CM
eo CM IO
eo eo Is. eo
cn o to CM
d IO IO
d
cn o |s.
d
Is. Is. CM p d
Is.
CO
d
cn CM CM Is.
d
to CO CO o is!
to o cn p is!
CO CO CM p Is!
CO o CM
d
© Is cn
*s. cn
CN Is. en
CO
cn cn
IO
o>
CO
o> o>
00
cn
o>
cn
o eo cn
eo cn
CN eo cn
CO eo cn
eo cn
IO co cn
CO eo cn
Is. 00 cn
00 CO cn
cn eo cn
o cn cn
cn cn
CM cn cn
co cn cn
<* cn cn
IO cn cn
to cn cn
|s. cn cn
to cn cn
CN III ~~ ...J
.*ti m u < JC H X LU CO
— © a t- c "S it) o o re —»
■=? 1* o 0£
CO 3. U
.S.
opul
at
ow
th I
(x100
O a. O O" a. w °-o to 3
a a:
of
U.S
. pu
latio
n w
th R
at o
a. O 0. co
z o o 3! -«Q-Ö Z,
c oil
U.S
. o
pu
lati
o
Gro
wth
R
ate
(%)
o Q. <o
0. ' 3
o o Of w
quar
e U
.S.
spul
at
o 0)
w a. 3
w o o —' «0 Q- *- ■= <0 £ ä 3j z o Z j-£ _l
c _ © (0 a.
(0 re o 3 a. E
o S (01 3
a. -'
uare
of
ce/P
ri c
e R
atio
a co a: Q-
CT-r- (0 w £ 3
ct\
LN
of
ce/P
ri
Rat
io
a. 3 Z]
0. _i|
a>
U.S
. ce
/Pri
c R
atio
a a. co ac 3]
Q.
u <u
a> ■£ XJ t» 0.1 O 3 —' O) (01 Q 0. a> en 3
. ^ U T-
U.S
Fi
rs
Pri
(
HI CO Q *
co to
o
to (O
c!
<M CM en o d
to
«o p d
o co o d
<M CM <J> O
d
o CM o
d
CM T-
d
«M CO CM
d
CM CM 0> O
d
o «0 o> o d
co o eo p d
to «* «0 o d
to
p d
«M en o d
IO to co p d
o CO o d
CO CM CO p d
co o en o d
CM
d
to
d
o CO CM •^ d
to to
d
o 00 en p d
co o en p d
to CO p d
IO to co o d
CM en r«. p d
CM CO to CO
CM CO CM IO
o to *t to i
<o CO co <o
«M IO •T- to
i
o <o to
i
«M IO
IO
o> to
IO
«o o o IO
o to
to to
i
IO to IO to
i
to CO CO to
<M
i
CO N.
to
o w
en en to
i
IO (O IO CO
i
en to
IO
i
IO
co IO
i 5 CM eo
CM IO
i
CM to
CO
i
IO co u> to
i
to CO co to
CO
to
i CM N.
i
p o
CM CO p
2? co p d
CM p d
en p d
to o> d p
to p
to p d
o> p d
IO p d
CM p d
25
p d
en p d
S5 CO p d
o en d
T- en d
IO p d
to p
S5 p
25 25 CO p
en en d
25 IO en d
25 CM p d
eo p d
S5 en p d
io
o CM
co CM
CO ^-
CO IO o
to © en •*
CM co I-. IO -*
co «o to
<* IO
to o co
CM •«* IO o>
CM ■O to O IO
CM CO
IO
to IO to CM to
to to to eo IO
co IO to
to
to IO to
to o CO to IO
CM to CO
to
o CO to
o CO
en IO
o CM en o to
o CM CM CM (O
CO to
IO CM o IO to
IO CO
to to
to
to
CO
s en to
o CM CO o r-
CO to V-
to
en CM
co «*> CM CO
d
cr> w CO CO
IO
to IO ^- CO
IO
CM co IO co IO
CO IO to CO
IO
IO r«. CO
U5
CO CO
IO
en CO IO
co IO o
IO
to
IO
o> IO CM «t
IO
co IO co "# id
CO IO
id
to "* ■d
eo to
id
en
d
CM V co
d
o en
d
CM en en ■>»
d
co o IO
d
CM en IO
d
en en CM IO
d
CO
IO
d
CM IO IO
d
co to IO
d
CM
r- IO
d
o CO IO
d
to en 00 to d
IO co en IO
d
IO o IO © CM
to p
o CM
o °i d © CM
p
CM
IO p CO ^" CM
i^ o> IO
CM
^1
o CO T- CM
CM
d CM CM
CO p CM CM CM
to o IO CM CM
CM CM
CM CM
I1«,
d CM CM
to p
CO CM
1»;
d CO CM
CM CO
d CO CM
CM p
CO CM
CO
d CM
en CM
CM ■>*
CM
o IO
CM
CM p d ■<*
CM
3 d CM
CM "^ CM IO CM
o o d to CM
IO
IO CM
en CM
d to CM
to
CM to CM
CO
d to CM
p
to CM
«o o d CM
©
p
en o p ©
s co
CO CM v> evi
o p o CM p o CO p d
o CO p ^0
to CO p o co CM
CM
«o co is.
CO IO o
d
T"" CM CO
d
IO CM o en d
o o p d
o to CM
d
to CO o p d
o o o o •*
CM p d
to CO eo p d
en to IO
d
o o p d
it to to
en CM T™ p
00 CM p d
CO CO
d
CO en o d
IO o CO p d
i
o
p d
co to «« d
CO
o d
o o p d
CM o
CM o CM O
d i
CO CM
d
CM
d
CM to «o CM
d
N.
d 1
IO to V"
d 1
CO
IO p d
o. CO CO !•- p d
o
d
to 00
CM
d 1
co CM co
d
co CM CO
d
IO co CM
9
en to p d
i
co en CO
d
IO o
d i
o r«. N p d
o r«. o CM
d
CO
en p d
i
o CO
d 1
CM p p d
CO ©
an p
CM p p
CO p d
ea p d
to CM p
CO p
^1
p d
CO p d
IO p d
o p d
IO
d
en co
d
o CM o CM
en
d p d p d
o en d
co p
co CM p d
CM p d
CO
.«fl-
IO
<o «O
(O p CO
CM co «*. to
o> CM
CO CM
CM p cö Ol
CM o
CM t»
p d CM
co p CO CM
to p d CM
o>
to
o
d IO 1»
p d p d
T- p
co
p d «o
IO to
CM CM
r«. IO p
CM
co to
d CM
CM
d CM 4»
p d d
oo p d p d
CO p IO
ea en en
CM
o> eo O» 0>
IO
o> to
o> o> co o>
o> N. o> o CO CO
eo CM co e»
co co er>
co en
IO CO en
to CO en
CO en
CO CO en
en CO en o en en
en en
CM en en
CO en en
en en
IO en en
CO en en
en en
co en en
HI — _l
3
(0 c o o 3
O a:
D
O 0)
(0
K to
o a fo
re ■D O 3 a 3 ^ a.
a- a. re (0 V) Ü
o *-» a: to 3
o ?£ 0£
o o 3
■c £* « 2
a. to
Z o is re a. at
Ü _l
«*^ c o D> fr oi c
0£ 3
■a a» tn\ o 0£ Ü 3 Q.
*—< Ol >« c .*: XJ
(0 3
E 0) re S.
DC ° 2 UJ
(0 Q 2 a
|o o T™ en oo eo CM 00 «* o oo IO o IO o CM to * eo co lio ^ O eo to co T- *— o o o CO o CM CM CM T- o cn r» N. 1*™ CM CM ■r- T- T- T" T- T- T* ^ o ▼■ V- ^ T" T- T" o o o 1° O O O o O o o O a o o o o o o o o o a o
|o ^ CO M CM CM o ■* CM o cn ■* r>~
- oo co l>«. ▼" T- h».
IO eo o co en en V en ^ co <* T~ CM to CO ■o cn CM eo I-. cn |o> i^ 00 00 00 en © o *— T" T- CM ^ © o o o T" T" CM CM lo © o o ■ o © 1 i i 1 1 i i 1 I t ■ 1 1 i ■
k- eo eo IO o CO to cn CO N. N. ■* to ■* r>~ ^, to r«. cn IO leo u> ^i eo *- CO CO CM CM ■^™ en CM It to ->» CO CM o r«. I-- *? Tt ^f Tf Tf CO «o CO CO CO CO CM CO co co CO co CO CO CM CM o o o O o o o o o o O o o o o o o o o o o
[eo N. CM IO ^ en r». CM en
r» >* CM to CO r- i^ CM 00 cn CM IO to f** ■o >* ^p—- eo CM to h» 00 cn
CM cn
eo «O •* to r* CO CO cn T* co oo eo ■* 00 eo CM IO co o * Tt to CM K ■* IO ^ CO IO ^ -* © C+ o CO oo (O «o IO IO t- oo CM cn tSr T- (O o © IO h>. K. oo CM en to to o r- ^ ■* ^t CO r». cn eo °* N. K. r^ r». *0 00 CO «> ■Of 9> en en cn en cn -cn cn cn cn
© 1- CM CO *■ IO (O h- oo cn o CM co ■* IO to »- 00 cn o T— CM co <* IO to i^ co !■*■ N. N. t>~ N. 1». r* i». i». N. oo eo eo eo oo CO eo CO 00 eo cn cn cn cn cn cn cn en cn O). en cn 0> en en en en cn en cn cn cn cn cn en cn cn cn cn en cn en en cn cn cn cn cn
in UJ
— _i
X . UJ CO
•- * D£ *- o c 'S o re ^ O t3 ~ XC o — « o
te — -c o a O
(U 3
w °-c5
a.
1 V)
2cJ tore CD O ™ * a.
. § JS x: O
LN
of
Pop
u G
row
t a.
z! -J|
c 73 ',C *> o» TT « 5 a. O 3 O S O 5 to^
0.
CL
S
o o °-i ® 2 ■•*"■ i_ i_ re 0-1
Sq
ua
Wo
Pop
ul
5 a eo
-C o 0-1
•*r T3 ^S o o — re 0- Z O S 5 z . -J§ Q- o XL
_t
c -» O (/)
'O ^ c Ü. ■c re o o
Wo
op
ul
Tri
lli
0-1
OL *-
o o &*] C O- -p
0- 0-1
|S£ 5 a «£ to
0) o o^J 's ,= o 0- °a~ 0- z ® £ ^ J o D; ^
a. _l
0) o "O "C o a:] tD.5 0-1 ° 5> ffl 0-1 5 1 * §
k. a
<•-■*.
m to - <" ■2 — JS <n 0-H
Ara
b L
igh
t ($
/b
(199
5
UJ i_ ol re 0) o >- o
1«- CO
xt d
co o CM xt o
o CO «n CO
Ö
co to CO
d
tM
m co d
o <M CO d
xo en o co d
CO <o -o CO
d
«o en o co d
oo CM o CO
d
co in
CM
d
CO co -o CO
d
CM O co d
CM CM oo CM
d
CD «n oo CM
o
co CO o co d
co co
CO
d
eo CM o CO
d
co m en CM
d
co in en CM
d
■n to CM
d
x— xt CO CM
d
m xt O CM
d
o CO en
d
eo oo <n XT"
d
o CO «n
d
to
o CM
d
co CM CO
d
o o> xt co CO*
1
co 1- co co CO
1
o
X— cn CO*
1
CM o> CO
■
<o CO to a> CO
i
o co CM o
i
en cn eo o
i
co in o
en en co o
CO
in o
i
co CM CO o
1
co in xt o
co
in o
i
xt co co o
■f
xt o eo o
■t
co m xt o
co m xt o
co x— in o
00 CM CO o
co CM CO o
CD CO
x—
i
en en
x—
■t
in
xt CM
i
r» 09 CO CN
CD x— co CM
i
N. co co CM
■f
in
s CM
i m o co
i
Us |o 1° lo
-s co X—
CM
o CM
cn <n
as o» CO
o> N.
to i».
in co xt N.
-s CM m xt co
CO
£5 cn CO
m a? m
S5 xt CM CM CO
CO
S5 CO in
CO xt
a5 o xt xt
5? o xt
CM xt
m CO
loo CM
lx—
1"? CO*
CM xt o> o
X—
o CO
co xt
to
to CM
in
CO I*. 00 in
o> oo r«. ■>*
tö
m
o
CM CO eo (O r->" -T-
o> o CO 09 -T-
o> m en CO
CO CM CO
d
co xt o CO
d CM
cn o CM O
CM
O en ■n
<M
xt xt en xt CM* CM
X—
CO CM
CO CM
X— O co o xt CM
in
CO en xt <M
CO 00 o 00 iri CM
o CD O I»; tö CM
en
co p
CM
to en CO in tö CM
cn oo x— xt d CM
in CD CD CM
d co
CM
CO
in xt O O CM CO
CO oo o en CM* CO
P CO co
en 00 CO r- xt CO
CO
CO
en CM CM co
CM CO xt CO
o CM to CO
eo CM oo co
o o
09 eo *^ •<*
CO <o CO
CO in x*
CM X" 1«.
x* 09 09 xj-
xt in o m
CO CM CM in
o o
s co CO in in
CO
in
o en m
co o co
CO m CM CO
CM xt co
in en in CO
CO in
CD
eo o cn CO
o m a
o cn x—
cn CM co
en CO xt N.
o X— co
xt xt 1«. N.
cn lO Is- *°, CO
CO CM xt in
CO
CO X— X—
co co CO*
CO CO
o p cö
CO
00 an eo
CM
m o
00 CM CM CO
m o CM
en CO
CM
CM x* in CO
xj
N. 09 en CM x*
xt*
oo o CO o m xt
■n oo xt oo m xt
co xt co CO
xt
CO co CM xt r- xt
co en CM CM CO
xt
m
r>» o en xt
m X— co cn cn xt
en X™ O oo o in
eo
co
iri
CO CO m CM
iri
m CM CM xt CO
iri
CM cn CO CM xt iri
in
o m iri
00 CO 00 f~ m iri
in CM
m CO
iri
CO co co N. iri
CM CM 00 X" OO
m*
CM m CO cn p iri
N. CO v> x—
(O CO CM X—
to
CO ©
X1-
o> CO
o o o o v-
CO
CO o
O eo en d
■«t o CO cn d
to CO 00 00 d
xt 00 m en CM
to in x»- CO X"
o o o p
en o xt
d
xt CM
to d
en o xt
d
in CM o en d
CD CO x— co d
CO T" 00 p
xt 00
in d
cn co CO
x—
m CM CD in x—
co en o xt d
cn o CO p X—
xt 00
d
o o cn p X-
CO CO CM
co
CO in X—
CO m CO xt d
CM
o
CO in o o
CO o o
<o o o o d
cn co p d
o ^—
o d
i
o CM O
d
CM co o d
i
CM xt in d
co
d
o o p d
o co p d ■
CO en xr-
d i
o co o p
w p d
o co in d
i
en CO p d
en CM CO
d i
m CO CM
d
CO CM CM
d
CO xt xt d
i
o CO p d
cn CM p d
i
CM to CM
d
CO m p d
in
CM
d
CO x— xt d
i
xt CM
X—
to o X—
o <M o
o p en en d
CO en d
o> d
CM co X— p
f».
d
CM 00 d d
m p d
CO in d
xt p x—
CM r». d
co lO CM
X"
«t p d
co p x—
CM
d
o to x—
CO p x—
-xt CM
CO p d
CM CO
id
o CM
r-"
o p I*-"
09 in CM CO
co
CM «
m »»- CO CO
co CO
CO in CM co
co d co
xt
CM in
CM 00 tö Of
CM eo
«n
in p i»" w
CM *°. d ■xf
CO o in
CO p CM 'Xt
in
co* CM
in r- xt CM
o
x— t»
o m CO* CM «>
xt d CM
CO
eo
CM xt d
eo p co"
CM
d
CO p d x—
X—'
p co" CM
00 CO
id x—
Is» en
X— CM co en
iß
cn
to r-. 09
CO
en
o> N. o>
o CO en
■x— oo o>
CM eo en
co oo en
-xt CO en
in CO en
to CO en
00 en
oo 00 en
en CO en
o en en
en en
CM en en
CO cn cn
xt cn cn
m en cn
to en cn
1». en cn
oo cn cn
UJ _I
= m — < S H j= Q X _i
UJ 01 o
Sq
uar
e o
f 1W
PR
OD
(x
100
0) Q
o 0-
r o w
LN
of
1WP
RO
D D
o DL
z Q O OS 0.
o
Wo
rld
C
rude
Oil
Pro
du
ctio
n
(MM
Mb
bIs
)
Q-
Sq
uar
e o
f W
orl
d G
DP
G
row
th R
ate
(x 1
000)
o
o
Wo
rld
G
DP
G
row
th
Rat
e (%
)
ü Q-
O
Sq
uar
e o
f W
orl
d
GD
P
O
a
LN
of
Wo
rld
G
DP
Q.
o
_l|
Wo
rld G
DP
($
Tri
llio
ns)
0. a
UJ
I10. Icö
u>
cö
o cn co oi
co co o "* CM
r-
CO
CM
O co r>- co oi
co co co CM
CM
o <o
CM
o CM
o CM p
IO
CM
CM
IO
c4
IO
CM p CM
CM
CM
o
IO
CM
co «o r- p CM
CO ■*
p CM
eo IO co p CM
CM co co
CM
o
o> p oi
cn co >* p CM
co o p CM
00 to p CM
to co p oi
o co
p oi
eo <M CM p
co
P T-
cn o cn p
IO co CM p T-
|oi
io o r- eq oi
CN CO CM p CM
1
co ■«t
o cö
1
CO co
p cö
1
CO p CM
i
co
p co"
1
p cö I
o IO CO p cö
1
s CM
cö 1
eo
p cö
CM
p cö
1
p CM
CO CO «0 o> CM
i
CM
p CM
co o C0 cn CM
IO
CM p CÖ
i
CM p CÖ
i
co co p
■
co
00 o cö
1
co cn p cö
o o cn p cö
■
cn CM cn p cö
i
cn p «ö •
co co o
cö
o r«. CM
cö 1
CM IO IO
CÖ
co
cn
cö
C0
p ei
Im Icn o> u> Ip [d
r- co co IO p d
CO r~ co IO o d
00 o o>
o d
co cn co
o ö
CO
■o p o
co co 1*«. ■*• p d
o o co
p d
co r«. IO
p d
CM
co
p d
co o co ■»!■ p d
IO
p d
CO CM
IO p d
r- ■^-
u> p d
o IO p d
IO
O ■O o d
CO !•- 00
p d
CO co 00 ^> p d
IO co
o d
CO
lO
o d
V- CM IO
p d
o IO ■o
p d
r- co IO
p d
IO
IO ^> p d
o cn
o d
IO co co
p d
CO co CM
p d
CM
p d
co co co C0 p d
ho CM
CO (O
CM IO
1^ <» CO
1^
IO co o
o CD
cd *—-
IO ~OJ CM co
CO-
© CM
o>
00 T* CN t
Ö CM
CM
o to CO co o>
o>
O)
ö CM
I--
CO CM ■o
CM
CM CD o> <o ax
CM
co CM o f>« <o
oi CM
T- CM m o> o
CM-
IO ■^~
0> +«•
co d
-c* CM 00
io
co to CO o CO
0)
co co IO ■o •er CD O)
CM o> >ef CO O
d
O) o co CM CM IO
d CM
r» IO IO o> C0 CO
d CM
co <o o o IO
CM
o>
IO IO p
CM
(O
co co
CM" CM
o> CO O) O)
p
CM
co co co o ■"fr
o CM' CM
CM >« co o o. CM CM
o> co o
CM CM CM
IO o CO CM CM
CO o> co o> IO Tt
CÖ CM
C0 IO o
CO
■Cf" CM
IO co
CM
o co o d
o IO co cö CM
o »-. eo cd
O 00 CO
IÖ CM
o IO o* cö
co CM p d
o
o cd
o o»
C0
o 5 IO CM
O CO CM
CM
co IO r«.
CM o> CM
d 1
o ■<* r«. d
o
CM
<M
■T"
«M
O CO p CM
o p d CM
o co
■*"
IO
p o CM p d
o o co IO
CM CO ■»t
o CM
d
IO CO CO
IÖ
o co
d T"
CM cn CM
d
co <n o o
IO 00 CM
oi
O
d O d
eo CM
IO
IO co
CM
SS co cn IO
P d
CM
cö
CO p IO cn IO
00 IO
0s" 0> r--
1
CM I»;
CM
CO CO
o p d
to
CM p CO p d
cn
CM
eo p d
0s" IO
1^
CO p d r«.
CO p d
co
i
io CO co
d
co CM
o cn CO co d
CO
o co
oo CM p d
CM f- co IO
•>* CM
CM o> p cd CM
IO o p <ß co
co o
p 5
IO
p eö co
o co o>
d o
CO
lO
O»
s eo
d
co CM o p d
CO co 00
d CM
p cö co
CM
co co
CM
co
co p d co CM
o
d V co
co 00 00 p d co co
eo o co p d eo
to
00 p d IO IO
to
cö N. ■O
o IO o p oi CO CO
co
o eo
oo T-
id co CO
CM 0. IO T-
d co
eo co cn p oi
o d
IO
o ©
r- o
co
IO 00 CO
T- o o p
co eo N. p
CM
o p CO CO p
co
CM
o io CM
CM
o CM
CM N. CM
CM
IO
O p CM
co o IO p CM
eo IO cn p CM
o>
p CM"
IO
eo CO
CM
cn cn
CM
eo co eo p CM
co 00 r- p CM
IO co p cö
CM T- CO
cö
CO IO 1^
CÖ
CM eo
p cö
cn co
p cö
eo co co p cö
s p cö
IO o p cö
CO co
CM
CM IO CM r»; CM
IO
cö
co *~ CM 00
cö
o co
IO co IO
IO co IO co to'
^1
IO
o d
co
1^
co co eo CM
cn
d
io ■«t
p d
IO CM
d
CO o> CO
d
CM 00 eo-
d
to co
o> d
o> r-.
CM
IO o IO co
to o co
d
r«. CM oo 00
CO p d
C0 o CM T-
CM CM
^1
oo cn p cö CM
o ^,
cn CÖ CM
IO
id CM
co
d CM
IO cn cn co d CM
co o CM CO
d CM
to CM IO OI
1^ OI
cn
CM
cn
co *-- o»
<* IO cn
CO
o»
CO
0)
o oo
CM <o
CO co C7»
eo o»
IO 09- 0»
CO 00 co
a>
co 00 cn
cn oo o cn cn
cn <n
CM cn cn
co cn cn
cn
IO cn cn
co o> cn
0» cn
eo cn cn
CM UJ I CQ
a _J Qi o
Sq
uar
e o
f 1W
PR
Q.
a CO
■ to ■'s B*°.
CO IO o IO
IO eo
c«i
<o CO IO
CM
o co eo p CM
co Tt o eo c4
en is. CO r>- CM
cn h» en eo es
CM cn CM
CM
CM cn CM
CM
CO Is. Is. co CM
eo CM CM
CM
CO CM CM
CM
eo CM eo CM
<4
eo co
o
eo o
o
CO IO o CM
o Is. CM
T" CM
CM
CO p
co
O P
CO
o p
o to o p
o co o p
CO T- o p
o o o p
eo to co p d
<o eo CO cn d
CM CM co p d
LN
of
1WP
R Di
a.
_l
Ho ■ CM ■ cn
■ o
CM o> IO
ö
IO o
o
IO to
ö
o IO co
d
IO is.
d
CO CM T- IO
d
eo
to
d
CO eo T*
d
CO eo
d
co CO
d
en o> eo d
CM O O
d
o
d
to co to eo d
co CO to eo d
o CO co d
CO en
■r- d
Is. en p d
o en eo o d
o> o o p d
o> o o o d
o co o p d
o eo o p d
cn o o p d
o o o o d
Tt Is. o p d
Is. CO
P d
co cn
p d
i
1/W
PR
(x
100
0)
or a.
l(£> ■ lO ■ co ■ <£>
eo CM eo IO
IO tn en
(O CM
IO
CM
CM to
co
IO
is. «n CO CO
IO co-
IO
CO CO to to
CO CO to to
o CM
to
o
en CM o>
5 en
CM
eo CO
en eo CO
o Is. CM
o CM O
co cn CO o
en o o o
o> o o o
o co o o
o to o o
cn o o o
o o o o
CO CM cn en o
IO
00 cn o
cn o 00 cn o
Wo
rld
P
rove
d R
eser
ves
(MM
Mb
bIs
)
Cd a.
B** ■ (£> eo
(0
cn d CO CO
isl CM «0
Is.
io"
CO
is.
cö to CO
p d o> to
CO to ■>* CO
co
Tt co
to
CO
to d CO
Is.
d ts. CO
CM
d CO
eo d to CO
is.
d cn CO
d o Is.
cn CO
eo |sl CO CO
Tt
Is." o CJ>
Is.
to d o> o>
d en cn
p is! o> cn
o is! cn cn
d cn o>
p d © p
to is! o
co d p "T-
p d
p
Sq
uar
e o
f O
il E
ner
gy
Use
Rat
io
Di
UJ
a CO
eo O «0 CM CM
O
Is. co CM CM CM
d
Is. Is. (O O CM
d
cn CM e»
CM
d
CO CO T" CM CM
d
CO CO IO
CM
d
CM CO o CM
d
eo
CM O CM
d
CM O to CO
d
o o CM Is.
d
CM o <o CO
d
CO
CO to
d
IO co
d
to Is.
d
eo en CM
d
to CM CO
d
eo
CM
d
CO CO 00 CO
d
Is.
cn CO
d
Is.
CO CO
d
CO IO CO co
d
CM CO CO CO
d
to co co eo
o
lO
eo
d
LN o
f O
il E
nerg
y U
se
Rat
io a:
UJ
2 _1
IO
en eo is.
© i
co IO o IO
d 1
co o co co Is.
d i
CO CO 00 to
d i
CM co CM IO Is.
d 1
cn CO Is. CO Is.
d
CO co CO is. Is.
d
CO CO CO cn Is.
d i
CO eo ■* CO
d i
o CO eo o
CM co
en CO
o
cn Ti- ls. CM en d
CM o CO CO o» o
to CO CO to en d
CM IO CM
o> d
1
CO
Is. en o
Is. CM
Is o» d
CO CO is. 00 en p
CO en d
CO to CO CO a> d
■
o to IO 0>
°i d
00 cn Is. o> p d
i
Tt CM CO o p
1
o
Tt o p T-
1
Wo
rld O
il E
ner
gy
Use
Rat
io
a; UJ
5
CM IO Is- is. Tt
d
o> o CM Is.
d
CM Is. Tt lO Tt
d
CO CM CO CO
d
o 1^
d
o> o
CO <* d
eo en CO IO
d
CM cn cn
d
O CO
d
eo Is.
d
co «* Is. o
d
to IO IO o> CO
d
cn to o co CO
d
co
co eo d
eo
eo Is. CO
d
co «* co is. eo d
Is. ti- ls. IS. CO
d
Is. eo CM Is. p d
CO ■<* eo Is. m d
CN Is. n. d
o IO cn co p d
CM co cn co eo d
to to co p d
is. CM CO CO p d
O 3 2 •- >. fc — Ü 00 -2£ 1_ CM
o g © s £ z.
o CO » to 1*
o Is. IO
d CM
o o ls^
eo"
CM
o cn
CM CO CM
o V- CM
CM
o
cn CO Is. CM
o o *s_
en CM
a CO Is.
■*" CO CM
o IO eo en Is. CM
o
co is. CM
o 'T- IS.
of Is. CM
o
IO cn CM
o CM_
eo o eo
o co cn CM
eo
o CM
d CM to
o co |s^
CM eo eo
o IO p cn eo eo
o Is. |s^
isT "* eo
o o co
eo
o CM p isT
eo
o
Is.
d eo
o
p d to eo
o Is. CO
co co
o
IO Is. co
0) w «r
o m 2 2 w O O T-
5 ~
o CO eo isT
o CM co IsT
o en © eo
o CM cn CM CM
o IO |s^
isT CM
o IO
CO" CM
o |s. co eo CO
o CM
CO CM
o CO
d CM
O
lO
o |s. cn CO
o CO co d
o o
to
o
CM
d CM
o CO
CM
o eo cn to" CM
o CO cn isT CM
o o IO
d CM
o Is. |s-_
to CN
o eo
d CM
o CO oo_ d CM
o CO
d eo
o CM p d eo
o en p isT eo
CO CD
>
UJ Q O ü
Ho ts. en
cn
en
eo
en en IO
09
CO |s- en
is. Is. cn
CO Is. cn
cn Is. cn
o CO cn
CO en
CN CO en
eo CO cn
co cn
IO co en
CO CO cn
Is. CO en
co CO cn
cn co cn
o o cn
cn cn
CM cn cn
co cn en
Ti- en en
to cn en
CO en cn
1» en cn
CO cn cn
CO UJ _J
= m *- < '■*= Q X _l
UJ (£ O
K—
qu
äre
o O
PE
C
Mar
ket
Sh
are CO
o a (0
OT
p s£ s _j O 5 to
o Zj
^•^
ü^ OT
O OP
E
Mar
k ha
re
(0
o
Sq
uar
e o
f W
orl
d
rod
uct
ion
/ R
efin
ing
p
acif
y R
ati 02
<->
0-
§ a. re OT O
5 c a:] J? o o> J*? § ■■= -£ o o 5 = = « "-S
02 0-
o "a oJ re 02 ^ Z 2 02 O z _l 0-
— o "?,& Oi] *= .S Ö P O
rod
ut
Ref
in
Cap
a R
at
021 Q.I
Q. -—v
a&2 13 C ~ SI
Wor
t ef
ini
apac
M
Mb
a: o 5 "-—'
*^ 02] © 02 o 0-1 0) 0. O 02 J
- -3 > «" er? 3 T—1
c (0 w| (A
«- a: 0-1 o Q- arl 7 o- 55
T- zl _ll
02 02] a. a.1 02 oq •5 T- r_l
c
Wo
rld
R
eser
ve/
rod
uct
io
Rat
io 021
0-1 021 5
a. uil
i» ol re o| >- o|
Ico IM
IO CM t- CM 27
98
3014
«0 <ct o CM
O CO IO CM
co
CM
o to IO CM CM
O co CM
CO
o CM
-* CM CO
CM CO CM
CM CO o
^t CM O
o o o
to IO o
to o
r>» o
CM co CO
r». CO
«0 o
o to
O co t».
r-. Is»
Is»
r».
r». Is»
to
«0
IO CM co
IN Io
Ico"
IO IO o co"
co CO q CO
to o o
IO o o co
it CM
co co q CO
CM CM q
O q co
O
q CO
o q CO
co 0» co co
co to q co"
CO o «t cö
CO co
co"
o
co"
co Tt co"
Is» co <* co"
q co"
CO o co CO
IO
q co"
CO IO CO
co
o h» co"
co Is- co"
co co N. co"
co CO N.
CÖ
co co r». co"
CM IO r-. <o"
IO o q co"
1 V llO
CM
CM m
0) CM m
O Tt TO
CO
-t <o d CM
TO d TO
O)
o> CO
to d id
CO
o>
CO
o CM CO
o d CO
IO
oi CO
cM* CO
CO
<o
O) to CO
q CO CO
1».
d CO
q d
Is»
^1-
q q q q CM d
CO
IN Iff) llO l"«t Io
o •4 o Is- IO o
T- co o q d
IO CO CO Is» co d
CO r». CO CO q d
*o co ■* T" lO d
o co
o
d
o co CM
d
CO co
q d
450
co O
d
co o IO q o"
Is» co co o CM
d
CO io o CM
d
CO co r«. o CO
d
o o CM q d
"«t IO CM q d
to
CO q d
■<* O «* q d
IO co co q d
co U) co co q d
IO o co q d
I-— loo 1« Io Ito |d 1 1
co CM Is- CM
O i
CO 00 CM IO CM
d 1
o co to o
d
o co CM CM
d i
o CO •CO
q
CO to CO to
d i
CO <M
d
to CO O o -* d
co to
d
o o ■* CM q d
to •^ co
q d
i
Is- o to o q d
IO CO o oo q d i
to o 00 CO q d i
CO CO T" CO to o
o o o o q d i
Is» CO to CM to d i
o o
d
co o o o
d i
I». h» O
d
Ico
Io
■*t CM
CO r». o
o co CO Is- I-.
d
o IO CM CM CO
d
l>- O co o Is» d
CO IO
d
CO ■* co CO q d
CO CM ^t q d
CO CM CM
q d
to CO q d
to
CM O
d
co
q d
o co
q d
o co
IO q d
CM
to CO q d
to
o 1^ q d
CO o IO o q d
co
o q d
IO o Is- o q o"
o o Is- o q d
O o r«- o q d
Io CM r».
"* CM
CM co
CO CM
CM IO q CO CM
*™ Is» t». -*t CM
r». u> to •o CM
o CO CM
co •*
to" CO
io CO
CO CO
CM
*o CO
o CM co to CO
CM CO q to CO
CO q Is-" CO
IO co co to" CO
o CO o
CO
CM CO CM
io" CO
o s •«t CO
o Is» q co" CO
co en
CO
00
to" CO
o o q d CO
CM oo q d co
CO to
o
co Is- Is» o
CO r». !•» d
CO co IO q
CM CM o
«* co CO
d
o CM CO CM
«O CO
CO CM CO
CO o co CM
CO CM CM
«o o> d
00 CO ■<* 4»
d
o
co d
CO CM
BO
d
CO o CO Is- d
Tt
co co d
o CM
to d
CO q d
co to
IO
d
oo co
d
o
oo
d
oo co
d
co co -* d
CO
d
co
q d
o CM xt q d
CM
Is» lO d
o en CM Is- o*
CM
q CO
co Is» IO CO*
o t» IO co"
GO CM
CO
CO o
CO 1
co CM to CO
to q CO*
o q co'
o Is- q co"
o CO q co*
■
o q CO*
i
«o o
co
CO CO q co" i
o <* q co"
i
CO to q co"
o q CO
o CM q CO
o to Is» co
CM
r». co"
■
CO t». co
o
q co"
Is-
q CO
CM
oo
-«t
q CO
■«t o q CO
oo r». CO
i
o co Is»
CO
■o co 1».
CO
co
CO
■
to Is» CM q o
o r». CM q o"
o r». CM q d
to CM co o d
CM CO *0 q d
o CM q d
lO CO q d
CO CO q o*
<0 q d
to IO CO q d
to co CO q d
co o « q d
o CM q d
o o CM q d
to co CM q o*
co CM o o*
o <M o o"
CO CO CM q d
Is» CO CM q d
Is» CM CM o d
CM CM q d
o CM CM q o"
CM CM q d
CM CM q d
co CM CM q d
CO CM CM q d
co CO CM q o"
o CO CM q o"
o i». CM q d
co **. CO CO
o CO
to CO
to CO
IO CO
q d CO
d CO
o tt co
CM V)
CO CM
OY <o d CM
q d CM
co
CO CM
q d CM
CO o CM CO
CO
CO
to
■«t CO
CO o id co
co to id co
CO q CO* co
o q CM •>*
o CM
CM
o o
3 id
CO
id
CO CM
id
•— CO
id
co CO
Tt
IO co CO
to q CM ^t
o CO
o q
CO
o t~ o
Is» o
CM r». o
CO Is- o O
to
a»
co
o
r- o
CO
o o o
o CO o
CO o
CM CO o
CO CO «n
<•* 00 o
«o co o
CO CO o
co CO CO o
o CO o
o o o
o o
CM O O
CO o o
Tt o o
•o o o
CO o o
Is» o o
CO o o
111 _l
= DO
lc Q X _l
UJ a: O
r— ■* CM co CO IS) o 1^ CM CM 00 T— 1^ CM in ^— CO ■*
O O) in in T— Ul i^ x—
CO CO oo o in a> o in t^- CN o o ,_:
.10 o o T— *c 01
H- 00 =)
Q)
H (A CO 3 CO 3 L_
o {? u (0 o CO
> en I—
111 o DC <
=> W
cc
Mul
tiple R
R S
quar
e ■a CO
w 3
<
"E CO
■D
CO
GO
TO
CD CO jQ
o
2 o
to
in o f-- ■«1- Oi o o CO
co co o
CM co CO ■<* ■«*
CM
CM m co m
oo CO O) ■<t o i^ oo co co CO T-
CM CO CD co oo oo O CO CO ON« T- CO CM co m o tciod r^ o) CM co in o co co co
T- (D S CM CN
c o _
tO 3 £ -g a> to CD CD cc cc
■* CM O CM co r- CM CO co m CO CD oo ■* ■* ■*
o Tt co co
00 T- CD CO O T- T- CD m oo o> co ■* CD
"if CN O CM <o r*- CN CD co in co co oo
d ■* co co
00 T- co co o -I- T- co in oo Öl co Tj- CD Tf oo
CD
CO ID oo o co r>- r- ■* CO O) co o co o CM co co d d O) O) o> r-- CM m T— oo ■«t o o CM CO co o CO o m •"- T_
•t r-- 05 05 in ^* in T— r^ "S- co a> m co O) T—
o> in at
oo m CD m i~~ o o at O) o CO o T— oo o 00 o ■*
CD
n .Q m CO n if_ i— co CD > _c X
Exhibit Hi Individual Regression Results U.S. Market Data
PRICE vs PRICE/PRICE RATIO
■* CM ■* m CO 00 CM oo CM 00 ^ CJ> O) CO CO m CO oo CO i^* CM T— ^— 00 oo
.0 H3
CO o ■* CO oo oo •* h~ CM o o ,_:
,<0 o o o T- HS en
1- to 3 0. c p 1- f/> CO D CO 3
o ? u
CO
"O CD W
o C/>
> <
O) Oj ft: a:
CD Q.
"5
P CO 3 CT (0
LJJ ■g CO XJ c CD
cz o
CD in n
CO :> Q: < co O
2 o
CO
CO m CM CM
oo oo oo CM o oo co CO CM CM
CM h~ 00 00
O) lO
O CM
^ CO
CO CO CO
CM CO CD CO CO CO ~ lO CO
h- 00 CM CM o o
o o o irl CO
CO CO T- CM CO O) CO CO
T-(ON CM CM
c o </> m w -i CD T3 O) </> CD CD a: a: \-
CM CO CD CO CD O co m
O) co CM CO
oo <* ■* o o co lO CM CD W m co T- O
d T CM IT)
i
CM CO CD CO CD O oo m oo ■* oo o r- oo oo ■* a» co CM co
Ul CM co w m co T- o
d T CM lO
T- CD
ll'l ">
X w CO <d-
■<- CD
at ■* co Tf CM lO m o m O) c» m ■«a- in co o _' *o
h~ oo co CD ■c- o CM CM CM CO at co
CM O
if) T- m m oo co co w CM o h- lO o> lO CM T-; in' •* CM T-
_CD
S.-S CD .15 i- CO
3 > = X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN PRICE/PRICE RATIO
=> Q. H Z> o > a: <
CO
CD
to c .0 55
& CD
0> OO 1- CM CO CO ■>*■ Oi CO CM O (D T- CD CO O O O) "* O 1^- h- -<t CO 00 CN o> o m "d9 o o
CO
CO
o- o CO ■
?u till! # g. I -o £ 2 a: < co o
o z <
CO
CO CN
o o co
CM
CO O CN CM <J> m co t^ co
a> co CM 00 r~- co o> 00 CO CN CO o ■*' 00 00 CN lO O) co co
■<- CD r- CM CN
c o CO w £ -o CD CD a: on
Co
t/i CD
in o> O) o CD CN r» CD m CD h- o co co ■<* o cb co CN T-
r>- CD CN T- o co r» O) CO CM r-- T- T- CM m o CO T- CD CM
lO 05 01 o CD CN h- CO m co h- o co co "* °. 00 co' CM
r~- co CM T- O CO
co CM r» T- y- CN m o CD T- cd CM
•* CO co 00 CO T- CO T- in •<* r~- •* CM CD o CM o ■«- Ö Ö
CM m ■* o> 00 in r- CD CD CO 00 o> r-- 00 o h- co in CO T^
CO 1- CM •* in ■<- ■* CM 00 co
r~ o
CO Tf in Tt
a> 00 co co in co co co o> CM co ■*■ m co
a> co
_CD
S.-S CD 55 u 'b "- CD £ > £ X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs SQ PRICE/PRICE RATIO
T— CD CO m 05 o O v~ i^ CM co CO co O) o CO o> o CM CO T— 1^ o CO ■* O) CO m o o
£0 .O d
CM q T— o CM
y> d d ■G i
CO H^,
H CO D c CL .o S. 1- 55 CO Z> to 3 l_
o or
CO 2 t/>
co a: cu
o: UJ
"5 03
T3
o
<
2 CO
Q. as zs CT
CO
T3 CD W 3
03 £ CD
z> "5 H? CO X) CO S Qi < CO O
o
i- 00 Q. Q. rj
CO a. O
8 c 8
h-
CO
CO >c o 1 5 ■*
d o co
co
co CO m
cb 3
u.
7223
0
.711
2 76
79
CO Tf OJ 00 CO 42
S
104.
0 14
6.3 CO
CO CO w K» CM CO m o CM r~- ■«*• Ö f- CO CO ■* f» m
to CO
00 i- o ^J
CM
iri 1 Tf in m ■o o en o c T- CO ■t -2
to T- h~ 00
CM CM 42 c
.CD H-. G T3
Reg
ress
ion
Res
idua
l
"co o H
i CD O O
CM ■* ■* CJ> co co oo oo ■* CM co o> ■** o
■* OJ CM CM oo O) CD CM CM CO T- |v- CO CO c- co cd d
CM ■* •* O) CD CD co oo ■* CM eo c» —• o
Tl- O) CM CM co o> CD CM CM CD T- r- co CO h- co oo d
i
r- co o o CM i^- oo t>» o) co oo ■* CO CD o> o o ■* d o
oo oo co
co r-- co co in co co co co co <<t h~ oo
a> co CM o
CM 1^
CD
CM
00 T- 00 o CD CO oo o co m 00 CO m 00 O 05 cb °.
CD
n .O ni CO fi 'il L_ co CD >
_C X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs POP
D Q. h- => o >- a: <
CO
-2 (O c .o 55 <0
& CD CD
1^ oo CO 00 OS o> r»- o ■«*• CM o> Oi t^- t— 00 O) 05 h~ a> m en CO ■* co r- CM m oo T~ ■«*• co ^ o T— o o CM o o 1
CD L_ CD 3
o co (/J
a: a> Q.
"5
8> HI
■p
o
CO 3 o-
a> "55 3 TT
CO ■o c CO
CD CO J3
J> on < CO u
2 o z <
CO
CM
O O) O
CO oo ■«a-
d
co co o o>
o m
CM ■<- ■* m co oo
co
'. c-
CM o> m M- o> m (D O) •* m co oo TJ- o m •<*■ co CM
o m oo in o> o co ■<*
r- co CM CM
<0 3
O) to CD CD a: cr
* CM ■<*■ ■* -if o CO o>
<o" O ■>* oo •>*
O) O) T- l~ CO T- CD
# CM "-.
■* co
co o) e? ■* CD o oo ■<*•
r^- r- id m a> O) r-- r- >». a> co
i h-' "* h- CD o T- h-
-J 1 1
CM •*
^ CO O) in O Tf C3> oo ■* »^ O) T- CD CO T-
1 CM r~-
■* co
co o) •<*■ CD
ON m
00 •*
in o>
>~ o> CO I o
^ "*. h- CD
1 1
CM CM r-- T- r- o
CD .3
o o> co o
SP in <<* > O) CO
CL oo oo co ■**■ ö d ■* T- CM T- CM CM
•*- CO
in r-- m CM
•JS ■<* CD CO r»- o H-j CO 1-
d ^ o
1
^ CD f» o O CM
£ CO CM m o> h~ CM
■o T- O c. ■ |»-
« 5 » c T- r-- co CM
Sw
to
oo co CO O) CO
■c CO CM c ■<- CD .15 T- CM G T- CO
JE CD
CM CD
CM ®. O CO (35 o T— T—
1
CD
Q.-S CD ™ o c <- co s > = X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN POP
oo T- SS o in o CO CO
■* co vi CO CD o> ■■- CM i- a> o CD h- O
1
8
66.
2
0.0
^ ■<*■ o o CM O
m o) id in co O) CO r-~ i» T- o 3? h- o ä m o o o> Ö -j
00 T-
^ o m CO CD
in Tf CO O) co co i-. T- CM CD <J> o Q. r-- o § CD °
CO o
CO 00 CM U. OJ ■* o
8 8
CO 00 O CO o> 65
52
7390
a> a! T- O IC CO i r» o
.1 CO m o Ö o ai o
CO
o> O) r
CO 5
I-- CO co en ■* eo
oo O) oo w o> o
u. CM co
f- CO O OJ T- CO
o> o co d d d CD ■* oo r» ■<t ■* y- CO h- r- CM O O T- m oo
to T- O) 3s oo in CO o CM r-
§
137.
14
5.1 CO
2.73
8 -0
.972
CO O) in L_ CD T- ■* o m 5 O CO f~- h- ■* ä O CO o r- 00 00 Tf t- ■<*■ m m ■*
CO to
CO o> CM ■p CO CM
CO T- in
3.94
.0
00
T- O) o c CO ■* -2
co T- O
m m T- m CJ T- r>- 00 CD in CO CM •* O) CM CM CM CO i- t-- (0 0)0(0 •vs o> CO O O CO CM c o> h- t— m CM T- CD ■* co 0 00 01(0 fc G i- CM •<t CO T- xfr Tj
O
oo o
.8 00 CO o o T- O -°.°N cd °.
.^ O O O T- O CO o »G 1 CO H^
h- to => C a. .o CD
i- to CO D 00 => ,_ o >- CD
lulti
ple
R
Squ
are
djus
ted
R S
q ta
ndar
d E
rroi
'b
serv
atio
ns
c T-
a: or o CD
< 2 O 2
00 -j CD -S O) 00 CD CD
re "o te
rcep
t V
aria
bi
CO s a < CO O < a: a: 1- £ x
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs SQ POP
ü. H => o >- 01 <
D
.8
■S co c .o S5
& CD
Ul NID o> o> r-
CN 00 CD co co in 00 CM •* o o o ci b
CD O) Ol O CM O 00 (^ o
CO CM °. Cxi
CD CD
ti g a: tu o
£ 1 I « I — cn .S, £ tn
a: CD
<U
01 < w O
io O)
CM CO r«- oo co r~- T- O CD CM CD T- Oi ■ CD £1 ■* CO ■* CM
■«fr T- O CO CO r- a> ■*■ co oo
in
14. in co
CD Ci
CM CM
c m -fr Cl T~
ic- o cr oo C7) o to
CD
CM u.
0945
0.
0645
69
72
00 co co CD co t^ oo CD ^
o r-
m TJ- m •* Mm O) 00 -«fr o o oo co o m
co co
.667
6 04
5.5
055.
2
<» "tf "*
T- r- oo CN CM
»»-. ■o
o
o z eg
ress
es
idua
ot
al
< er oi i-
CM CO r- oo co h- T- o CD CM CO T- o> ■ CD ^ "fr CD ■<fr CM
o>
CJ co o> -fr co oo co o> m
m co o
o CM i
■tj- w r- co o CM oo CM o o> -fr m <p-
in o o oo o o
CM T- h- r- co o r- o co x— r^ o i— -fr o> in o CM
"- o
co co CM 0> ^- O) o ■"fr- co ei r- CO co q CM
CO r— r— r~ T_
-fr CM co CO o CO to r- o> o> o> CO "fr co o e- rM oo CM CM
CD
"ri .Q CD tu ci L. i_ CD CD >
_c X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs POP GR
00 T— Is- CO 00 T— ■* Tf ^- CM T— m O) in ^ CD Is- CM o> O) in OS o> o> o> r- o CM CO c»
.8 in CO CM CO CM O o
<o o o o ^— •C m
1- co D r Q_ ,o (Li
1- to CO Z> (0 3
o co o v> > a> Q: III n on ir tu (11 "E
CO Tl
< 5
CD
Q. CO 3
0> CO
:> CO M
ulti
RS
q
3
<
c CO
CO
</)
o
CO a. r- 0) 00
00 r- CO m co ei Is-
«er a> ST
,—
O) o CO
at o 05 OJ CM
u.
2202
1
.748
0 62
25
CO m o •d- CD
ü
5220
2
24
7.
7618
4
141
.5
2838
6
CO ■* m o CO Is- O 00
TJ- 00 CM CM CO O)
co co
T- CD Is- CM CM
«»-. ■o
c o
2 o
egre
ss
esid
ua
otal
< a: tr H
3? m T- Tf CO
o Is- in o) r- in" m r- O) co ■*
I» Is- CO Hi
§: CM °>. T- CO
3^ *— ^
oo co ^ m Is- o m o)
co o) id ■* Is- O) co co
I Pi 8 co Is- o 1 co
~j ■
m T- ^ ■* co
Is- m in O) Is- o> m Is- >- co •* Q) Is- co
1 CM °> 1- co
co co m r-
lO o>
m es co en ■* Is- co co
1 o -J
1^8 co Is- 1 co
co co O Is-
5 co co CD CO T- CO
5P 00 (O s» CD r- d CM T-
Ö o
O) T- Is- CM Is- co
co T- O T- CO
■c CO T- CO O CM **«- T- CM
T c0. 1 1
k» CO Tf o o co
ul T- CM 00 o CM CO
t) CO Is- c CO T-
■8 ai °°. c T- in co T- CM ti CO
CO T- CO ■>- CM
*M CT> CM c CM T- 05 CD m ,G m TI-
»E m co Cu CM ^ o co ■* O T- CO
1 1
05
Q--S a) 55 O '- t- CO Ä > £ X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN POP GR
D D. H D o > a: <
3 CO
-2 to c .o to to
& O) CD
I^ co T-~ co oo 0> CM ■* CM OJ CM CD CM in CM OJ CN OMNN -* CM CM CO m in co oo m CD CM CO CM O O o o d
0) 1_
CO
tig a: in o T3 "2 00 S ro > w -a a, .5, E to - =5* 2 -a a < (o o
Q. = •j= er ■3 W
T_ u. T—
n CO m
C- o m CD ei co
CO
Ol o CO
CO
co u.
0249
1
.814
3 22
86
CO "<f CM CO
§
4024
9
256.
2 94
361
141
28
386
CO co
1
256
.23
26
3671
.7
27
3928
.0
<*.. t3
c o
2 o
C0 3 £ -a _ Oi "to jS ^ ID <D O
< K. a:\-
•* CN ^ T- O o CM O)
T- m lO r- co O) m co i> o m CD ^co- s co oo 3 CD
CD ■* ^ T- CO o CM ■«*
O CM in r- m O) CO CO >». CM m
1 CD in T— T— o ■*
-j
•«fr CM M? ON
i- o CM OJ
lO T- m O) h- CD i- m co CD o m
I * co- co oo CD
CD •<*• T- CO
in CM ■* O CM r- m CO CD
1 CM m CD lO
o 1
CO T- h- T-
03 3
co co ■* m ■<t o
5 CN CD > CM CO
oL o oo O T-
ö d 00 Tj- CO CO CO I--
CO CM en ■iS CO CO CO a> co <k< CO ■*
CO CO
CO T^ 1
l^ CD CM o co o
ä ■* m Tf CO o ■*
•ö co co l~ O) T-
■8 ^ CM c CM CM co T~ T~"
Hw
CO
h- CD to 00 CD
CD CM CD CO
CD CO ■«* ,G CO T-
JE CO
co m
o T- CD O
1
CO
co .2 ü •- c CO 3> £X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs SQ POP GR
a. h- Z> o > a: <
D 00
5 oo c .o
to
in CO
CO h-
o> o> CM co (D N m O) CO CM CO
m o
O) co m o d ö
o>
co
&c5 w E o: LU T3 "P
.9- 2. Vi "o
2 a: < w
u. CM
n O
r- oo (Tt oo C1 f^
ic- •«J-
cr CM CD o
00
m o oo CO
u.
2357
1.39
48
7489
00 CO CO O T- s
6235
7 1
99.2
05
222
1
42.8
58
455
CO CM »' ^ ^ m m to o> co m O) no T- TJ-
T- r*- co CM CM
»»-. T3
c o
o z egre
ss
esid
ua
otal
< ir a: H
CD T- ^ r» T- o ■* co
■* o in co o> O) m a) v~ 05 CO 0) o) r-
# oo w. co o
CM O) ^ o ^ 5
05 CO lO r>- m O) CM CO w i- m 9? T- m ä d ^ o CM CO -j
CD ■<-
S5 I1- T- ■* CO
in ■* O o> co at v« m o 0) o> CD Q. O) h-
§ od °°. co o
CM a> ^~" ^~
m O)
«- ■* O) CO h- in CM CO
JR i- in
1 o -J
T- in d <N CM CO
h- CM O h-
<b liS .3 oo eo co 3 °° £ oo i*. _i o> ^* a. 05 CM
CD d
CM T- CD CM CM CD O) CM
CO Tf CO •is CO O oo CM T- ■fci CM CO
"* T CD T^
i
L» r-- r- o CM CO
fi CM r- lO CM
•o CO T- c T- h- ■S o o ^3 CO o C CO "tf T^ ti CO
■* ■* to a> m >*o r- o c CD CO ." O CO ,S -f ■* IE m o)
tf) oo o o> T O CM T-
1
<D
a) co ü •- i- cs 3 > = X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs GDP
Tt- co co in o> - " T- r» CM CM CM T- r» in 00 oo r- <35 CM r- T- CO m co r- in CM r>- r>- 00 ^— m ■*
10 o o CO CM
o o °. Cxi co o O t-
H2 1
m H to D c- a. ,o CD
H co CO D to = ,_ o o- o
>- CD
sted
R
dard E
< It IT
CD Q. 3
^ '■>-i CT CO
3 C D 3 TT W CO :> a: < CO
LL. o ai r» o r- CO m CO fi in
ie- in
c- oo rn o to
CM m CO co CD
u.
1024
0
.033
8 59
02
CO CM CD m o
^
2102
4
5.0
80
7793
4 1
50.0
58
455
CO m T- CM
to
.08C
05
0 05
5
in -«j- ■*
T- 1^ 00 CN CM
>*-. ■o
c o
2 o gr
ess
sidu
a al
.<£ CD CD o < tu K. i-
T- CM es T- O o O CM
r~- oo in" O CM O) CO CM v. •* o CD m m
I in" «P CO CD
1- in <£ co m o at T-
CO CD lO CO CM o> CM r- l« CM o 9? CO CO 5 in °> o T- r^
«J ■
T- CM
«£ T- O O CM
in r>- oo o> O CM i- CO CM CD ■* O Q. m m
§ in » CO CO
T- m co m
m 05 T- CO CO CO CM CM r-
if; CN o
1 o
co co in °>
■
m ■* o o
CD UJ£ ^ co co CO co co S in m
r- m d ■* co ^-' ©
m co CM h- ■* CO
42 CM CO O O
to O ■* <tw r~- oo
CM 1- iri o
^. m co o CO CM
£ ■* m co o oo f-
■o co •* c ■* o ■S m co T3 C CD
oo in •* CO
•C co
T- r- CO r- i^ •*£ ■* a> c o oo CD c- ■* ,G CM CM
CD co m m in
o in <°. o CM O i
CD
"S.-S CD CO O «- >- CO
3 > EX
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN GDP
=> ti- l- ID o > DC <
Z> CO
5 co c .o 53 to
S CJ) Q3
1^ co 00 CO r» co oo o co CM oo CO oo CM 1^- CO O) CO 00 o a> 1^ o co rf CO oo •* co CO T— oo m CO T— o o o o
p
a: <u
CO
a: in o
■= 2 £ to ~ CO ~ # 5-11 a) - /A .5, E to OT =o iS -Q
a < CO O
u. R CM r- ■■a- m 00
£ CM r-
c o CT) o L0
CD oo
o>
U.
7719
3.
4844
98
83
CO 05 CO CM CM
^
9771
9 4
63.5
28
683
1
33.0
58
455
CO CM r-- CM
CO •°. T-: in co o> m co m o Tf CO •*
T- r- oo CM CM
«*-. ■o
o
2 o 2 eg
ress
es
idua
ot
a I
< Dl ÜH H
CO T- c£ o ^ 5)
o o in CD m o> 00 CM i^ r- O CD ■* CN I lO °
co o
^ r- T- CD O)
O T— T—
U5 O)
■* oo T- in CM OO
l^ CO CM I o
CD ■* CM Ö CM '
CO T-
* ir> o o O) co m v» 00 CM OJ h- O Q. ■<* OJ
§ in q co o
r- T- CD O)
m O)
T— T— ■«fr CO T- m CM 00
JR CO CM
1 o -j
CD Tf c\i d CN '
O r- T- OJ
<b 1 ^
UJ CN .3 £ "* TO co oo > m CM
i 00 fC cd °
Ö
r- co CM f» m o> ■* in
CO T- r*- ■»2 ■* CO CO ■<- CD •w oo co
CN CO
oi T^ 1
k«. CO CO o r>- co
fi m co 00 CO
•Q ■**■ o> c: o o -2 CO 1- T3 c T ö co CO
H—
CO
■<*■ in Ü3 T- m
r^ co Q> CO i-
m Tj-
1 in o O CM
03 oi o o O CN '
CD
Q.-2 CD .2 o «- "- CO
B > EX
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs SQ GDP
r-^ lO CM m co ^— T— Tj- CO CM CO m oo m m o CM CD •>* o Tf m r»- a> r- r»
£ -* a> CO T- ■* w CM O)
■C CM o o T^ y> o o O T- *z CR
H 10 3 a. c
o £ i- in CO D CO 5. >_ o ? % 2 co >
<
D
03 a: a:
03
a. "5
03
co 3 D-
CO jjust
ed R
! an
dard
Er
bser
vatio
n
CO 2 Qi < co O
u. CM
* r- o> <n CO
£ o c CM
m Ö to
CM
■<* co
u.
9228
1
.656
83
64
CO O CM a> o
™ CM m -a- CO 1- CM
00 CO CD CM CD 00 CN -'S- CO 0> f-- 00 O CO CM
(0 o> h- o
to °H CM cb in o> CM CO CO 0) CM CO CO
i- CD r» CN CM
>»~ ■C3
o
O z egre
ss
esid
ua
otal
< cr cc \~
CO N- ^ O CN o CM a>
CM CM in" ■«*■ •»- O) r- T- i- O O) 03
1 O CD CO CM
■* r- as co o o m co
T- O) id in r- O) T- CO l^ m co ¥ m m 5 ■* ^
o co co —J i
co h-
<£ O CM CM O)
"O CM CM O) ■* T- I» r~- 1- 03 O OJ
1 O CO CO CM
■«* r- oo o
s? in co
lO T- o> m i*- i- 00
JR in oo
i o
m m
CO CO ■
CM CM OJ T- T- f-
03 3
T- r>- C- O)
5 o co > co a> d T- o
O CM
Ö Ö
■<* CM O CD O i-
H— CM 00 CD ■* o
<o 00 T-
«fc- r- OO in CM
CM T-1
u». h- CM o ■* co
£ T- m m co CO CD
73 CO CD fe CO ■*
■2 in T-
<9 b CO co oo *2 CO
T- 00
.2 oo o> c 03
CD ■* 00 CM CM T- 1 i- CD CO 1-
03 ■^ CO o o r- o
T— T—
T—
03
Q.-S 0) Co O >- <- co B> = X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs GDP GR
a. H
o > a: <
D 00
■2 oo c .o
CO
& CD
00 CO (O to oi s CN O CD co m o ■* CN T- T- O Tf o> 00 T- T— ^" T— c\i o o odd
03
T- oo o CN CO CN O) O)
<u a: Q. _
3 W
frig CO 3
P 'w ■O w - Ä «5 > CO -O CD
Qi < 00 O
u. co
0) ei O r- CO m m
& CN co
c CN
C71 o 00
oo CN CO
u.
9385
1
.311
6 47
87
oo CO CN ■* CN CO CO
00 CO 00 ■* T— *—
W N CD oo ■* co co -d- co at ■* oo CO 00 CN
fo ■* CO o
oo <9 o> oo OO CO CN oo r~- <j) T- CO CO
T- CO s CN CN
»•-. t3
o
2 o z eg
ress
es
idua
ot
al
< EKI-
^ oo in co r-
o ■* h-
U5 CO O) I-- 00
O) 00 ■* »^ ■* "* CD T- o>
§: iri oo
3^ r-- T-
CO CO öS r» T- o o> co
■* r*- lO T- •* o> O) CD ^ CO r- 9? h- oo ä d «? o T- m -J i
oo m c£ co r*- lO CD O) O) r- oo k» oo •* CD ^* •* Q. T- O)
§ in oo
CD CO I*- T-
lO O)
o> 00 ■* r-
OJ CD I- CO 1*- I o
~1
r-~ oo
v- m
O) CO o> o> in T-
CD .3
o> o CM CO
!P O) m :» O CN
CL. T- CD O CM
d d OO h- co ■<- CD ■«* **~ h- O
CD C5 r--
oo O CN ■>3- in
H-w r— •"■!• CN T T_
L in TJ- o CO CN
£ CN CD Oi OO CO o>
■o r- r- fe co oo
42 CD i- ^3 iri °>. CO T- in
H-j
oo h- O)
i2 O CD
c CD
f~- OJ in ■* OJ CN I co co M- 00
CD O) r» O O CN ^
■tf CD
T—
CD
"S--S CD <" Ü «- v- CO
£ > £ x
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN GDP GR
■* CM ■<- Is- 00 T— CM m co CM CO CM T- o> ^~ "* CM CM 05 r- ■* r- T— co CO T- O) o> 00 Is-
CO CO ^— T— T— T- o °. CM
.^ O o O T- *G 1
CD S*j
H to 3 c Q. o £ h- 55 CO D to => u. O »2» >- or
03 QL 0)
jjust
ed R
ta
ndar
d E
bser
vatio
i
< CO Q.
3
es 3 cr (0
CO 2Q:<WO
2 o
CO
CM m oo co a> O)
■* 00 O) lO
CM o> o o T- in cd ^ Is- oo
■* CM co en a) oo
oo co CM in oo "~ CM CM
a> o T- 00 m CM oo a> co co
T- (ON CM CM
tfl 3 £ -o - en w 2 co co o a: £ h-
O)
CO
c» -«j- r» t- CD Is- CM CM co ■* o o> T- 0> o co T^ co CO T^
■* co o> o CM T- O) CO CM O) o) in m T- [*- CD <* <P T- O
o> ■*
co h- CM CM CO ■* O O) T- C» o co
CO
CO T^
co o
CO W
a. CM
Tf CO O) o CM T- O) co CM O) c» m m T- Is- co ,♦: to
•«a- m •* o> CO o 00 CD O T- T- r» O) CD 00 T- r» Is- iri o
co Is- CO T- •«t co co co CM co
co 5 in o) OJ ■* co d
a> oo m o in oo oo m CM m. CM O
_co
CO 2
i- CO
Ä > £ X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs SQ GDP GR
00 w 00 rr f» •<* "<t ■* CM CO f- r~ co CD m m in ■*— ■* ^— T—
co CO r^ 01 10 m •* in
to r^ <* CO 00 co ö ■* 0
y> ö 0 oi ■c co
H~i
\- CO D c Q. .0 CD
1- CO 03 D <o 3 ^ o >- a:
O) (X to
O I— 1—
a: a: a> a: 111
"E cc
■0 c CO
0
< a> a.
co 3
(0
■a a> w
CD
a> CO jQ
CO 5 a < CO O
u. CO
8 O ST ^ S O ic- O cr O O O
to
o> m m
u.
1839
2
0.99
4 57
09
CO 00 00 T- CO
S
8183
9
173
3.
4642
7 8
2.55
28
267
to T- 00 O
to CO CO f» CO CO O) r- 0 r- T- CM CO
T- m co CM CM
«t— T3
0
O -z. eg
ress
es
idua
ot
al
< UL ai\-
CD ■* ^ in co 0 en m
■* CM «i O) r- o> 00 r- »^ CO •* CD CM in
I h-^ t^
■
Tf CO äS T- ■*" 0 r- 00
00 CO in r- T- O) CM O) i~ CM T-
1 w. in CM CM
0 CO 1- -j 1
CD ■<*
^ m co o> m
10 ■* CM o> o> h- V» 00 r- CD co ■* Q. CM in
§ h-' t^
"* T ■* CO t- -«t
3? o>
r» 00 00 CO I-- T- CM OJ
i». CM T- I 0
« in CM CM CO T-
1
T- CO T- T-
CD W 0 5 CO T- CO £ *- s S 0 _■ CO O Q- co 0
CO ö
CM CO CM T- O O T- CM
CD r-- 00 *2 CM OJ </) CO T- ■kw T- CO
■^ m
1
^ 00 CO 0 00 00
£ CD CD 00 in CD 00
■0 CM OJ t: r~ ■*
■8 m 00
c co 00 CO ^~
H«
CO
m es 00 CO O)
H-i co <<!- c CD 0 CD 00 r- ,G O OT
CD CO CD 00 CO
0 ai co 0 CO 00
1
V
<D
"S.-S CD .2 0 c 1- CO
2 > EX
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs 1/USEW
CM ^— O) CO f- O CO CO ■* CM T— a> CO o> m co 05 o CO T™ OJ CN CM w •* f- o ■>* Oi x—
fJ oi CN O in .o r~ CO co in
CO o o o i-- ■A en
H CO Z> c- Q_ ,o UJ
1- co CO Z> <0 3
o >- a:
1% o CO
o L_ CO
CD Dd (11
££ W
"E CO
c CO
o
< CD
Q. '■♦-'
3
c CO 3 cr (0
to 3
Co
CD CO
CO ^ a: < CO Ü
o <
CO
r- o UJ co CO co T— -* o>
co O) CM m eg T—
oo m ■,—
■<*
CD "fr CO CM o o> CD CO ■>- TJ- CO 00
co r-- CM in
CD 1- h- CO CO CD O CN CN CO CN CO 1— X— CM CO h- O
T- in r- r- CM o> CO Tf r» CN i- co
T- in CO CM CM
c o CO CO 'P 3 f> -D O) CO CO a> CD o a. ir H
co ■«* s? •* CM ö co oo
o ■* lO r- T- O» a> oo ^ 00 CM CD r- •* Q. CD ,_:
s- CO r-
■t r- ^ CD O O co o
CO T- Vi r- a> a> o co i» co m
1 T— T—
OO' CM o T- CN -j i i
co ■*
^ ■<* CM co oo
in o ■* O) h- T- ^ o> oo CD OO CM
Q. 1^- ■*
Ä co ,j ^ CO T-
1
■* T- co o
in o> 07
33
3910
1 o
—i
co m r— T—
00 CM T- CM
i i
■* h- o o OJ ^ IM CM UJ 3 o to
"c5 in co S ■* oo _i oo ▼" a. " t
Ö O)
f- CO ■* r- OJ CM
CO o m T- CO
H2 r*- co CO ■>*• ■*
<kl CD ■* m. cd
i
^. o> h- o o ■<*■
£ CO CO i^ in O CM
•G 00 CR »». "<* co ■2 05 O
c co
CM CD
iri CM ■C CO
O) CM CO O) i- <^ •* a> c T- CM CD 00 o ,G 00 i-
IE CD
m a> CM h~
O «N cd O
1 1
T-
0)
CD CO
Ü 1- <- co B> = x
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN 1/USEW
■* m CO m Is- v™ CO co CO CM ■■a- CD co m CN T™ ■* o CD oo 00 CO oo o •* CO
.2 oo m 00 TI- r-- eo o CS m co CO o
.to o o o T—
en H 00 Z) Q. o CD
h- CO CO D CO 3 t o ?
u CO QL T3 0)
1?) 3
<
o CO
> DU <
D CO M
ultip
le R
R S
quar
e
w "E CO
T3 C CO
CO
c o
CO
CD C/> 13
O
u. CO CD
(1) CM CO
r- m co IT) C1
ic- o c- o O) o
CO
CM CD CO O CD
u.
4667
1
2.59
8 85
44
CO ■t- 00 co co s
4667
1272
. 71
36
100.
82
67
co T- _,; CM « S °.
co M in s r^ CM co CM r-- T- CO
T- m CD CM CM
»(-. tj
o
2 o 2 eg
ress
es
idua
ot
al
< DL DC H
oo m feS CM CM o 00 Is-
m co «i CM Is- O) CD T- k^ T- CO OJ CD T-;
1 r-! CD' CO Is-
■
co r-- öS CM o o o m
Is- co in" O) T- O) O CM ^ co r-
1 <P <o co oo o CM CM
-J i
oo in
o^ CM CM oo r-
lO m o) O) CM r- i^ CD T- cB T- CD Q. CD 1-;
§ Is-' cb co r-.
O) Is- CM O
CO
o m Is- CD O) T- O CM w CO h- I
o —1
<9 (D' co co CM CM
1
CM O) T- CD
03 ll'l ™ 5 co
58 •« s CO T- _' CM O 0. CD O
co o
CO "tf T— T— m CM ">*■ T-
co o m co £ S ***. ^ ■*
m CO
L Is- CO 5 o o
ul co Is- CO CM
"O Oi OO c; CD T-
« CD i- CD c
CD CM in ■**.
CO
a> -t •2 CM CM
c ■* Is-
.<u ^— ^— CD ■* 1 CO ■* CM ^J
CD T": oo o CM T- O CO '
T-
CD
"S.-S a) co o <- ■- CO
B > £X
Exhibit Hi Individual Regression Results U.S. Market Data
PRICE vs SQ 1/USEW
D Q. H
o > a: <
c .o CO CO
s, O) CU
•* in
m 01 m NCOO
O) (N
to CD m I-- m o> m a ID
oo o
Tl-
Oi
O O O T-
03
CD ^ CO Q. 3
a: Lu g ■o "E " CD
T3
CO ,- TO C*
■J Ö- 3 ? CD
2 a: < co o
2 o
to
CO o CO
<o co co co
in oo m r- co co co co
co co
co CM ■* CM co co •»3- T-
in CN m m
CO CO CO s M m
». ° iri •* m in co co o
h- oo CN CM
c o 10 on </) -i QJ ■a O) CO CD <D
in CO
o: ct i-
(O co CD •* CN O) r» o CM m CD h- O T- co m a> co co
oo m ■* oo m T-
CM CM co r» f- h- <N oo CN 00 CM T-
CD OO CD ■* CN 0> t-~ O CM m CD r-- O r- co in O) co' oo
oo m TJ- oo m T-
CN CM CO 1^- h- h- <N CO CN 00 CN T-
00 CO •e- 00 CM CO 00 U) O T- CM CO O CO O O O CD
m m co r-- CM oo
m •«*■ CO Tf ei in CD a> m oo TJ- m co r-^
O)
CN m o CO r-
CO CM O T- CD T^ m o)
S--S
£ > S x
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs 1/PR0D
o^ T- CO co oo
O m co
O)
o> ■* O CM r-- o
CM g
§ T- CO
0^ CO CD O O) CO
vi r- r-- O) ■* CO w CM O
I co h-- <6 in o ■t co
-j ■ i
T- CO
8 co co m co
IO OO ■* O) O CM l^ r- o
8° CM CO ^ •*
CM T- CD u. O CO CD
8 O 8 IO
G> CO CM OO f-
2 CO h- r- CM Tf CO CO !R CM o
IC CO 1 o
co t^
s to
O d co m"
■>a- co ■ i
CM m CM CO oo o co
CD 5 o o
in CM CM i-~- CO CD
u. in CO
2 Tl- CM r- co m eo
CM co o CO d ö co co •<- ■*
co oo co m co m CO Tj- CM OO CM oo
to •* CO 4§ c- in T- CM CO CD
§
433.
9 13
4.1 to
-0.4
48
-1.7
98
co o) m ^_ CO OJ CO oo m s ■«t TJ- CD i- ■* £ CD O CM ■* oo CM m ■* ■* m ■** r-
to to
T- CO CM T3 00 CO
CO CM in in
CM CO m o
CO CO o c oo r^ ■* CO •«t -2
to " CM s eo in O) T- h- oo CD Tt m o> co m CM CM CM CO O T- i- co co m ■^ CO CD 0)0(0 00 c ■* h- O O CM T- cb CO 00 i- O CJ> i- fc G CD ■<- i-- r-- co co ^3
o h- oo
.8 NOMO O CD CO t- O ^J M. d
y> d o d T- O oo co «c 1 1
CD H»*
H to =) c Q. .o cu
I— H <0 CO D <o =3 >_
o
Mul
tiple
R
R S
quar
e A
djus
ted
R S
q S
tand
ard
Err
oi
Obs
erva
tions
cr T-
IT Q: o CD
<
D to
2 O z < R
egre
ssi
Res
idua
l
TO O
Inte
rcep
t X
Var
iab
Exhibit ill Individual Regression Results U.S. Market Data
PRICE vs LN 1/PROD
O) *— r— ^ 05 o r- oo CM r-~ (O co T—
CO 00 t^ t^ in 1*- r^ eg CO it) CO t^ T— O co CM
to CO co 05 ■* .O CO ^— o ■G y> o Ö ö ^~ HS (0 H^
1- CO Z> c Q. .o a> 1- 5 CO D CO 3 u- o cr
CO O 1_
CO c >- 0) Q: UJ o {£ o: Qi CD "E
cc
c CO
<
Z)
CD Q.
'■«-«
3
CO 3 cr (0
■D 03
"So
OB
CD CO .a
to 2 tt < CO O
2 o <
to
o
co h- o w CD CM CM w o CM co co m m oi o CM co
co m m o a> m co ■*- •>* CM co co o CM co co r~- CM
o> CM m CM lO O m co TJ-
T- r- co CM CM
o CO en CO -i 0) 1— ■o O) CO a> CD K a: H
CO <r- CM T- m ■* m it r- CM co o> co CO CM O)
CO CM
■<* r~- CD ■* CO «o t- CM co co
as r^ CO CD CM CD
CM
00 T- CM T- lO Tj- CO ■* r-- CM co o> CO CO CM 0>
m CM
■* r~- co ■* co m
CM CD CO
a> ?~ CO CD CM CD
CM
m CO o o> liJ o CM £■ OJ ^
co o
O) 00 at o CO IT) ■<- f- CD CO "* f- o> CO O) T- ^ o
CM
CO O) O i-
oo en CO CM
co in O CD CD *- ■* h- T- o
_CD
CD .JS c-> E >- co £ > EX
Exhibit ill Individual Regression Results U.S. Market Data
PRICE vs SQ 1/PR0D
r> Q. I- D o >- a: <
CO
.8
a co c .o CO to
& O) CD
O) CM CM in CO o> CM CO CM if) CM eg m -* CD O o o Ö O O T-
9> ™ Co Q. 3
•j= cr
5 a
*- CD a> CO CO CM IO CM in o CM CO CD O) CM CM CO CM ° CM o ■>-
1
CD CD
£* co t r- a: w o ■o T» CO 3 co r* CO "O 3 C CO
< to O
2 o
to
CM M- CD r~- CD CO
CD co in co m co r~- oo co in o in in m ™ a>
CD oo m m in in
O 00 o m
i CM 00 co in o m ■* o
r» oo CM CM
£ 3 S s D) CO CD CD
DU CC
"in W O H
oo -* a) oo CO CD co co h- co co h-
m. ^ CD OJ co 1-
CD CO CO -<fr T- CO m o CM CM
CD
co ri- al co CO CD co co
v'. f- CD O) CO T-
CD 00 CO ■* f- oo in o CM CM 00 Tf m !•--
CD
CM m CO o CO Tt CD m r«~ co CO CM CO r>- i>- o d CM co
5) co oo co CD CM O CD in o) co co ,-; co
CD O)
CM CM CO CD co m
co m oo O) r- iri CM co co
co oo
f» CD h- in O T- co T- CM cd
CM
0)
CD 5
<~ co £ > s X
Exhibit ill Individual Regression Results U.S. Market Data
PRICE vs 0EUR
r> D.
o >- <
D CO
to .cj
y>
-2 to c .o 53 to
&
CO o> CM CN O CM r^ CO r- CN CO o r^- o r-- oo CM I-» o CO oo IT) •* oo O) o> CD io o T— i*- o CO CN o o o O O O T-
CD t— CO
cr w - «=
rv o: üj o
■£ 5 £ «5 >
2 K. < to O
CO
2 O
to
co r- CM CM CD CO
co
o> co in in oo in
a> eo o o CM a>
co CO CM
co
■<*■
•* ■* m o> co in o m rf CM CO oo en oo m ■* m CM
CO CO in m
CO o o CM ■* M-
v- r»- oo CM CM
c o to en 10 -i a> ■D __ O) CO CO CD CD r> a: E H
T- CN 1^- CM O) ■* ■<* m OJ O) oo co in oo ■* oq o» cd o> oo
CM oo co co T- CD oo co ■* CD oo in m O) co a> in oi CM m
O) c» oo oo m co •* oo oi cd o> oo
CN oo CD CO
co 00 CD ■* CD oo m m a» co o> in o> CM m
T— a> CM co ■«t r- CD CN CD CN CO CO CO co CM OS O CO
o
a> o> T— CO CO V-
o 00 i*- 00 CM T—
o 00 T— OJ CM CO
•<- o
CN ro CM o> ■* ■*
CM O) o> o 00 OO •* CM O CD CO CO
■* CM o a> ■* CO CO ■<fr CM CO O CO O ■*
o> ■*
co ■*
CO
CD
n n ni co o ■e
co dj > c X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN 0EUR
"«i- T— CO t*~ a> ■<t T— co O CM 0 h- CM CM ■<*• ■* CO m CO O» CO 0 in CM m Oi m ■* CM CM
^ co O CO CM 0 O 0 CM
<o 0 O 0 ^^- ■iS 1
m
h- CO D ST a. .O 01
\- co cc D (0 =5
o cr CO
O
> u> DU III n at u- ir m "E
CO ■D c CO
CO
<
=> CO M
ultip
le
R S
quar
CD
<
CO
CD W
O
u. 00 m
* in 0
r- in m m C1 in
CO
m 0
co
CO
m u.
6295
0
.116
4 71
45
co T- O) CO ■* T- in * oi h- <* T- T~
in CM m O) 01 10 CM CM ■* CD CM 00 •<- ■* m
CO CD CO CM
CO
1
17.
41
27
4037
28
4055
H~ ■0
c 0
< > O z eg
ress
es
idua
ot
al
< on a: H
CD CM OR m co 0 00 0
■* CM id 0 m O) CM CM
CO
# to 55 . CM in CM in
■«a- m c£ >* m 0 -a- r»
0 co in m m O) ■* T-
1 ~ CD CO T-
0 ^. 1
-j 1
CD CM
^ in co co 0
10 ■* CM O) 0 in ^ CM CM CD
1 m. CM m CM in
■* m ■* m
in ■* r- O CO m m
I 0
~ CD co 1-
-J 1
CD CO 1^- w
5 co m 0 0 in m
SP r~- m > T- m
oL T- CO co i«~ 0 0
CD in co 1-- m •<*■
tw CD CO CD 00 in
co O CM CO T-
Hw 0 <* T^ CO
d
I» 00 m 0 O T-
£ CD CM m CM o> 1^
TD "«l- O) C 0 r- •§ T- CO
C co in co 1- o> Hi CO
co ■<- to 0 in *J CM CO c h~ CO CD r- T- ,5 CO CD
CD CD ** CD m
0 00 CM 0 1- co
CD
CD .2 0 <- <- cc £ > £X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs SQ OEUR
in CO •* r- CO CO 1^ m T— CM ■* 1^ co Ol o ■^ CO CO ■* lO "ä- CM CO r^- CM CO lO m CO in
fJ i^ o o T-
o o o CM to o o ■ x— *G m
1- LO D C CL ,o UJ
H CO CO D <o 3
o Dl
a1
CO o 10
>- a:
(D OH 01
or UJ T3
CD
c CO
o
< CD
Q.
3
li- re 3 tr
CO
■o CD
W _3
T?
CO
CD w
CO :> a: < CO O
u. 8 CD
o c co CM
£ O c h- Ol O
10
T- CO
u.
4416
0
.150
4 95
16
CO O T- CD O
Ü
0441
6
22.2
1 50
742
1
47.7
66
786
CO CD CM ■*
CO
1
22.
21
26
3840
27
3862
1^ t>
c o
2 o gr
ess
sidu
a al
z: CD CD o < CC K. \-
3? CD Tf
o co co CD CO ifi O CO O) O O) ^ oo oo
CD Q. <9^ §
T- 00 CO CO
o> m öS o o O ■* co
CM T- «3 co ■<* o> m ■*
1 §3 m. o o O x- -J 1
CD ■<*■
öS CO CO in co co o> o co ^ O O) CD co co
# <p <* t- 00 co co
o) m o o
es m
■«a- CO CM T- co ■«*• in Tt
i O -1
« o O T-
1
■* r- o r-»
1 co co O O) CM O)
5> 00 CM > r» t-
0L ■* o O f»- ö ö in T- x- r- CO O) CD 00
CO O CM ■iä o co CO I-- r- ■kM h- CO
O CO CM O
1
l^ CM CO o o> o> b T- ■*
lil r^ en ■* o
73 in oo fe CM CD
"§ °>. Ö C •* T-
co 1- ■* •»2 co
CO T- <o ■* O ■c 05 h- c O) <J> .© CM CO G o r>-
0) O CM °. a>
o T- m O CO 1-
03
CD S5 ü >- c- co 3 > £ x
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs 1/USPR
D CL h- D o >• tr <
00
.8 s3
■2 oo c ,o 53 to
& CD
CO
in
oo
o
a:
T- CO a> oo O) CO ■** CM co co o T- CD CO m o> CM co o CO O 00 co O CO T—
o o CM o o 1
T—
CD !_ CO 3 CX 00 L_
(Y 111 o £ T> V re
Squ
a dj
uste
CO
CO
CD CO
QL < 00 U
u. CO oo
a> CM ri r- r~- ttt m £ •«a-
CM
c a> CT) o 00
m CO CO 05 CO
u.
4167
0.
0091
21
35
00 CO <*■ CM O
fe
3416
7 1
.357
09
552
1
48.5
66
786
00 (MT-^f
00
.357
86
1 86
2
T- CO CO
i- CD r~ CM CM
>*~ ■o
o
2 O gr
ess
sidu
a al
^ CD CD o < K. DC H
a> co oo co r» co tv. co r- ■* ■* m o> oi M. T- 00 T- CM
oo o> CO CM ^ r>- CO co in !•»• o h- *— h- co O oo ,—
■ co
OT co 00 CO h~ CO r»- CD r- ■<a- ■<d- m en ^~ OJ co ^^ oo T~ CM
co O) CO CM T— r~- co CO m r- o r- T- h- CD o oo ■*^
CO
co co T— oo co CM CM ▼—
CM f- CM m CM ■*
r- CM co o> o o
O) IT) CD CM CD O "* O co O 0> CD r- in CM o> ■* o o o
1
r*- r- n oo 3) CO
■* r- S) ■T— » h- Ji ■*
CM «tf oo ■*
■<*• T_
■* ■* ^> CO o Tf M o £> m » *— ID T— ID 00 o CO CM
CD
CD <S
*— co 3 > sx\
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN 1/USPR
■* ■* CO x— 00 Ti- OS CN r*~ CN en ^— OJ CO CN CN ■* CN o> oo ■<* o> co CO r- CD CN 00
CO cg V O o ,cj T— o Ö CN ."> Ö o i ^~ •C CD
Hw
H to D c Q. .o £ H 55 CO
o > a:
to
CD a: a: CD
CO a:
o L_
LU
"E CD
"O c CD
CO c o
<
3
0) Q.
"5
CD 3 cr
CO
■D a> to 3 la*
TO
CD co -Q
CO ^ CC < CO O o z <
CO
co o O) CN CO
en co o
co o co co
ö
O) CO
CD CN m CN
fN co co ■* CO T-
o
CD O) co
CN 0> CN a> co co r^ oo co co co
co h- CN CN
c o CO m CO ~i CD TJ O) CO IU CD CD r>
0C a: 1-
CN 00 OJ CN O) O) o r--
CN 00 o r^- CO CN o> co CN CO CO CO co co T in ■* T-
CN 00 O) CN
O h-
CN 00 O h- CO CN o> oo CN CO CO CO CO CO
00 CO w T- r» O ■* o> CO CN co co O T- O CO °. d o
in r-- O CO co a> m o CO «o "* o co oo oo in CO <9
CN 00 <<* CN m OJ r- o OJ ■* oo 1- co oo T- m oo CD r^ in
r- oo O) co CM O) o ■* o ■* oo co a> co
CN o ^. CO CO
CD n .Q » CO c> i_ i_ CD CD > c X
Exhibit ill Individual Regression Results U.S. Market Data
PRICE vs SQ 1/USPR
in oo CD h- OO *— CO fx. CO CM r^. i^ oo o> CO x— CO CO O) OO ■^t co o CO O) O) CM CM m
to O <o T— T— ,o •<* O T- <s> O o T-
•C m
1- oo D c- Q. ,o a> H to cc 3 to =3 L. o we« >- a:
CD CT o: (It
CC 111 o
< CD
Q.
3
co
cr CO dj
uste
d ta
ndar
c bs
erva
CO 2 a: < co o
u. co
CD CN
r* O TT O) (" co
ic- CO c o CD o
CO
O)
in
u.
4605
5
.014
4 75
51
CO 5
O CO o> in
-<t CM CM T- CO
m m CD O CM 00 CD co r- ■* co co O MD
(n Oi Oi •*
CO °*. N CM ■* CO CD CM CM 00 CO CO CO
T- (ON CM CM
H~ ■o
c o
2 O z < R
egre
ss
Res
idua
T
otal
o V) o> l~ CD
co co 00 CM m ■*
^ oo O T- CM T-
CO 00 co o co co CO CM in oo co ■* co O) CM m ai o o) m
co co oo CM
CM oo
o 1- CM T-
co oo co o co co CD CN in oo CD ■* CD 0> CM m ai ri
r- eo o ■* r- CM CM o> co o •<*• o> oo co
co o
CM CM 00 O CM CO CO O) CO CO ,_: CM
■ CM
■«J- CO in ■* o o T- CD CO CO O) T- CO CO
°. in CM CM
Tt in in co CD 1- CD ■* CO o m T—
l>- m m CO 0) r— CO CD
CD
CD « u E >- co 3 > = X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs 1/USRPR
m CD ■* CO co 0 CO ■* o> CM co 0 1^ CN co 10 CO h- 0 CO CO O) CN 00 co
CO 0 0 t*- t— 0 CO CN *-
.JO 0 0 O
en H CO Z> a. 0 01
i- </l Co D <0 3 L_
o £ u CO tr ■a CD *—* CO 3
0 10
>- on <
D
O) CD a:
CD Q.
"5
CD b- co 3 O- 00
1_ uu
CO n c CO
c 0
'•4-1 CO
CD CO
CO ;> 0: < to 0 o
to
o r-- co 00
co o
CO co eg 10 co
00 m
■«a- 00 co co CN co co 00 CM CM co co co r-
CN
■* O CD CD m 00
■* h- ■* co
CN CO CD CO CO •*
CO CN •* CD CN 00
co co co
CD r-- CN CN
c o _ CO 3 £ -o - o) 'tö S CD CD o
en an \-
■<* CN o o CO o m m O O) 06 o Oi CN CN T-
m T- ■tf CO m o) o co CO CN T- CO
O) T- f- ■«J- CN o o CO o m m o o> 00 ö
"SIO) CN -J CN T-
1-- ■<*■ m T- Tf CD m o o co CO CN T- CD T- co co °°. co -«t
■* t- o o CN r» <j> eo o 00 ■* •*
■«t co 1- o ° 0 o
CO CO in 1- h- T- CD CD o r-- co co CN CN CN CN ® CN CN
r- T- in co a> r-- CNI CM CD CM CD ■<*■
r- CN co 00 CD CN
CM h- 00 co
co in
00 co ~ O)
1*» °°. CD CN
CD
CD
8-S
EX
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN 1/USRPR
H => Q. H
o >- <
D CO
CO .O ■c ."> •2 CO c .o 55 to
& O) CD 0£
r~ co CD •* m O)
TJ- T- oo CO 00 CM ■* oo
r-- m co co (NO) O CO NOIO O co CM o in O CD CO T- "* T- T- ^J
odd T-
CD t_ Co
tis Kill o
co O *3
LL. CM
CD CM r» T~"
m co r" co
!*= o c- o CT)
CO
■* o o
IJ-
4352
5
.063
1 32
12
CO CD CM CD •*
Ü
6435
2 6
29.5
00
351
1
24.3
66
786
CO CO
1
629
.56
26
3232
.9
27
3862
.4
>».. ■o
c o
2 o z eg
ress
es
idua
ot
al
< on a: \-
r- oo c£ 00 T- o T- o
CM oo lO m CM O) o> co k- CD o> CD
CM ^ 3^ w
<* m SR m r» o 1*- T-
CM r- vi & oo O) oo o) v.. oo in
i Oi CM oo d
O co ~J ■
r- oo c£ oo i-
T- O in CM CO o> m CM i» c» co CD co o)
1 CM? CM "^
m
■<t m m r-
a? lO O)
r- T- CM r- o> oo oo o)
1~ oo in
1 o
o> CM cd d co
i
0> CM T- O)
CD o> •* O) CM f» 1-
5° I-- CO i. T- CO
i o> o "> d o
en h- r- o V T"~ CO O)
co CD Tf ■is O) T- CO CM O Hw ■* in
in CM d CM ■
^, O) h- o m co
£ CM CM CO O) ■* 00
■p ■* CM I» o in •2 o co
in <*>. CD
V" T_
■*«
CO
OO Tf CO T- 00
■*^ Tj- 00 c ■«a- o CD in o> ,5 oo in !£ CO "* CD ■>a- o O t- O
O CO CO 1
T~
CD
CD .2 o <= <- co Ä> -EX
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs SQ 1/USRPR
CO CO c£ h- O O T- O
co ■* lO* CM CD O) ■«t CO I» T- m CD
105.
-8
.40
I ■*- m
^ ■«* T- o o o
CO CO in" T- co O) co m V» ■* co
1 « d O CM o CO CM ~J i
CO CD
OR h- O
in co -<a- o> CN CO V« •«* CD 03 i- in
Ä S"5 ^ T- 00 1
co ■<j- in u. CO ■* T-
0) CJ c 8
CM o o CO CM CO
CO CO 1- CO co in
m v^ ■* co CO o 1
o -J
« d O CM
to d CO CM
i
CO CD CO co T- CO T—
as 5 in CM
T— "* CO CO CO CM
u. CM co o
CM CO
O CO T— o o in d o m ■«* CM CO ■* co ■* oo 1- r- o oo m i- CO CO
CO co o § CN CO CM m T- O §
700.
4 13
7.2 CO
3.79
1 -2
.259
in co ■* k^. T- oo ■* o CM o r- m T- XT CO £ m o m co CD CM o co in r~ 00 CO
CO CO
CM i*- T— ■Q 00 T-
o o CO o
CD CO 00 CD
O CD CO c h-' d r- CM CO
CO T- m
OXD^NT- T- CO o T- 00 CD CM CD ■<- CM r— CM to ■* CM f~- CO h~ CM s^ CM CO m ■>* CD co c o TJ- CO CM CO CO 0) CO CO T-NNK) =5 G ■* r» O T- O r-
o
h- co to (D T- S S ■* CM T- ^
CD T- y> ö o d T- O •C TO 1M
H CO D c Q. o CD
1- to CO D to 3 ._ O >-
O)
Mul
tiple
R
R S
quar
e A
djus
ted
R S
q S
tand
ard
Err
oi
Obs
erva
tions
c T—
a: or o a> <
CO
2 O < R
egre
ssi
Res
idua
l
ID o
Inte
rcep
t X
Var
iab
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs PRCR
m CO m CM ,— m h~ a> CD CM co t^ CO 00 CM ^- CM ■* CM ■«* ^ in oo •* ■* x— oo 05 O)
.8 -Ü- o in h-- -* CM T_
T—
y> o Ö ö ^~- •C CD H^
H 00 Z) c Q. 1-
.o 55 CO
Z> o > a:
to
ft: a: CD
3
t/) a:
o
HI "2
10 c o
< CD a.
"5
C T3 to CO 3 CT
CO
CD CO ■D c CO
CD CO
CO s a: < 00 O
2 o
co
CD ^r- 00 CO co CM
o
o> O) h- oo C\J 05 f»
CM 00 co m CO CM o ~ in o
co o ■t o
CO O) CO CO CO T-
CO Oi CM co oo o)
CD •* O CD CO CO CO CM CO
T- O O T- CM
c o CO m CO -i a> "o D) CO CD CD
in TO
a: a: \-\
m o> O) o> ■«t m M- i^- CO h- oo OJ co ■* CM r^ CO i^- CM ■
c~- co 00 t—
oo O) r- CM ^ O) 05 r^- o OT CO oo
■
m Ol o> O) ■* m ■* i^ co r- oo o> CD ■*
CM r*- CD i^- CM ■
f- CO 00 oo a> N- CM T— O) 35 h- O O) CD 00 in r--
O) CD
co oo m co o co oo CM r- T- m •* -* o ö d
co ■* r- T- m h~ m ■* o co O) CM r- O) in oo r-- T- d c\i
CM CM r-~ o o o TJ- CM f~- co co ■* o> ^ i- 00
oo CO 00 ■* T— CO CM ^~ r^ h- T- r~- ^— CO OJ o ■* •*
CD *-< n .Q a> CO ci t_ i_ CO £ > ■E X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs LN PRCR
D Q. I-
O >- a: <
CO
.U
42 to c .o
Co
Oi CO CM CM CO T—
■* CM o> co CO 1^ "a- CO ■* CO Ul o CD O co T—
r~- 1^ CO o ■* CM 00 CO r-> CM T-
CM
O O
CO
Co
tig a: w o T3 "E CO £ w > w -o oj
— ■Ö' iS -Q CK < co O
co Q. —
- CO
2 O
to
in r^ co •«J- r-~ o co io LO
co r- m m o> co
o
co m
m co
co a> TJ- m CM CM en CD a) CM r^ CD O) O) s r- o T-
T- in o m in co
CM co
a> o T- CM
c o (/> m W -i cu "O __ en w CO CO CO n tr a: h-
a> ai
T- CM co en co o CD CM co m co o CM CO T- CM cd cd CD r-
CO •* co oo r*- CM a> CM ■* CM CO f- oo m CM ,_: ai o CM CO
t- CM CO O) 00 o CO CM co m co o CM CO
CM CO 00 CO
co ■* oo oo f- CM OJ CM ■* CM co r- co m CM ,_• ai o CM CO
oo CO o> m co CO CM
CO Ö
co t •* oo 00 N- co o> in m ■* co o CM CM CD Tf CO iri CM
CD CM CM 00 co oo o> CO CM T-
oo CO
co co
CD
Q. CO
JQ CO
(■) u. L_ co CO > c X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs SQ PRCR
D Q. H
o > <
D CO
.8 «G
-2 co c .o 55 to
& CO
n « in CM o)
o o T- h- CM oo o CM co co co
m o m CD co m CM CD
CM
CO CM
CM o o o
CD
CO
tig * • S S * ■S ? Ä CO >
CO O *S 2 o
CO
CO CO CD CM CM w CO
oo co CO O) a> co o
O) ■<*• Oi CM o> oo CM
O) CD cn
oimm oi Mm O) OTf CM -t CO ■* co m
CO ■* in CM
CO g oi in moo ■<-■*■*
T- r-. oo CM CM
c o w cn co -i CD ■D O) CO 0) CO
Q; Di O H
^ CO CD r>- w
o CD CO
ui r-- co o> oo co «f. co co (13 ss 3 T—
CO h~ o^ CO CD o co co
T- 00 lO T- 00 o> CM CO »~ TJ- co 9? o> CO ä oi «H o i- CM -J i
co co ^ CD CO lO r*- co O) OJ T- ^ 00 CO CO co co
I 30.
1.64
CO h- C0 CD
2116
88
83
I- ■* CO
§ o
o> CO
T- CM i
O CO T- CO
CD iA8 5 co £S S O) CO
i oo ? *- o'
0> ■* CO •* r- ■* O CM
CO in oo ?s r» oo CO xj- CO Hw 00 CM
00 CO oi o
i
l^ in CM o o ■* ? O) CD
■xa in o CO CO
73 O) O) JR m CM 42 Tj- in t3 C co
m O) CM Ö
si co
CM m to oo o
H~ T- in c ■* r» .1> <y> co ,S m CD !£ CO 00 0) o
i- o in M.
O CM O i
T—
CO
Q.-9 tu .5 o c: >- co 3 > s X
Exhibit III Individual Regression Results U.S. Market Data
PRICE vs SPECIAL EVENTS
OJ eo 00 CN oo w CO OJ CO CM o> 1-^ T— T- CD CO o o co o CO CM o> m o> in in
10 o p ö q£
CO o O ■G i
CTI
1- CO Z> c- Q. ,o ^ h- to CO D to =3 !_ o CO b >- tu
sted
R
da
rdE
IT < 2
It ir CD Q.
8> CO 3
t> '■— CJ CO
3 C 3 3 ■■¥? » CO 5 a: < co
2 o
to CM CO h- 00 m
r-- ■* co CO •* O T— CN
CO co ■<a- oo ■* ,_: CD ■* r- CM
r*- 00 CO CO CN ■* ■* CD h- T— T- It ^— r- •<-
CO CO T- CD
CO M ■*
f- CO 00 CD
CD CM CO t- CD CD
T- CD h- CN CN
c o _
& TO CO 3 cu -f 05 CO CU CU
co m m CD CN co m ^^ CN o>
in O) m o co ro CN co T~ t^ i^ o CO o 1^ 1X1 o o o> ^ CN T— CO 1^ •* •<t o T—
i
CO in w co CM CO m v- CM o> T— m 05 m o co O) CM co x—
h~ 1^ 0> CD o f^ X) O o o> ^ CM ^ oo T— ■*
OJ T- m in co CM r-- co i- t>- o oo o r- o m o d
CO co h- co 00 CO in oo CM ■* 00 CM h- CM r- co co in ■* ö
CD T- co o> co co CO O) CO 05 CD O) f- T- o f- CD Tf
oo CO
co m CD <°. CM CM
CD
CD <5 <-> E ir CO £ >
Exhibit IV Individual Regression Results World Market Data
PRICE vs PRICE/PRICE RATIO
O) co ■* CO oo r^- oo CO h- CM co m CO T—
0> -^ o r»- co CO 1^ 00 m r- CO co CO co o co
io to eg ^ Tf ,_ o o m (0 o o o T-
HS 1
ro H CO 3 c- Q_ ,o ID H CO Co 3 CO 3
o o GO
> a:
Oj a: 01
a: UJ
"S CO ■D c
o
< 5
CD
Q.
c. CD
■a CD
3
CO
CD
D (0
3 "■» 5 £K <
CO J3 o
u. CO co
CD m r- CM m in et m ie- o tr •<* Ot o
10
CO CM O) CO co
u.
4053
0
.714
9 39
76
to o ■<* o o CO CM
o oo r*- co ■r- CM
co oo co w co •» oeos •<* •* Tt O T- T-
CO CO
1
170
.30
26
6193
.3
27
6363
.6
»t- ■C3
o
2 o 2 eg
ress
es
idua
ot
al
< B:D:I-
m o> i- CD
CM oo CM ^ CO o
CO t*- ">* m o CO <* in 1^ co co
CM
co ■*
co T—
oo o ■* co oo O) i*- r*- o> T—
CM o CO a> CM ■
CM CO CM ^ CO o T— CO f~- ■*
m o co ■*
m 1^- CO CO
CM
co •<* co T—
oo o •«a- oo CO o> h- r^- O) ^- CM o CO O) CM 05
o 00 ^ CO
UJ m O) CM m m O) m CM O
CM d
o) m o m o> co r>- r- ■* co oo m r-- in
05 co o> d
CM ■<* ■* o oo co T- O oo in co co co co co oo O T- CM oo
CM in co in o o -I- in co r>- co CM O) CO T-
ai °> CM CD
0)
CD .JS c- co
LEX
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN PRICE/PRICE RATIO
=> a. D o >-
<
CO
to c ,o 55 to
03
CM O) m in 00 oo ^ O) ■* CM (M to 00 00 o in o> o 1^ <s> •* CM o T— o ^~ w CO (0 r- o CO in o o o o o o
CD
a>
CD
tig OC LU O
® re Q. 3
— ^n
V. "B. v re re
■=L £ m a: < w o (0
2 o z <
CO
00 CM O) 00 1^- 00
co o o oo o m CM CO
CM r- m co oo in <o co Tf co r- T- co m
co CO CM
■* CO
oo CM r- CO h- ■<t ■* ■«t T— r~- co CO co CM CO
co CO
EM co CO CO co co
,_ CO r- CM CM
c o to m w -i 0) ■o O) in re CD a> n (T C£ 1-
co in ■* •«- o co CM m CD 1- (M <* o> oo in m. CO CM
co CM 0> T- ■* CM CO CM r» i*- co co o co CM » CM T-
co m •* T- ■* T- o co CM m co ■<- CM ■* OJ 00 m" « CO CM
CO CM OJ T- ■* CM CO CM
co co ~ co
co
co
o I
co oo oo
co 00 OT ■* CO CD 00 T- T- O o ■<* co co oo CO
O) oo CD -I- Oi 00
CM r- co m
co oo co »
OO T- co in o> O) CO T- 0) O) co co a> oo en in
co
co re o «- c- re Ä > -EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ PRICE/PRICE RATIO
■<*• CM ■* ID o> m o> o> ■fl- CM m o> ^ m o CO r- o a> CO x— ■*
CM CM m CO O) CO o ■*
.8 r- o CO O) o o o m
CO o o o ^— •Ä ■ cn
1- CO z> c D. ,o 0)
\- to CO Z> <o 3
o >-
OH (11
O" CO
en
o 1—
UJ "E CD
T3 C CO
CO
c o
<
D CO
(U Q.
2
co
(0
■a 03 V> 3
<
CO
CD (/) -O o
2 o
o O) m co CM co co
CM ■*
r^- CM CM co o
OJ CO CD
CO ■<r co T- r- co co CO CM
CO lO ■* CM
T- co CO CO
iri o> CM
CD CO 00 Oi ■<(■ CD CD
r~- co CM CM
c o w m CO -i CD T3 ^_ O) to CO CD CD n
ÜCL QC h-
io O)
03
51
CO CM r^ oo ■* i- CD CM CO CO O) Tj- O) CM CM CO T-1 ^ 00 1^
CM OJ O CO 00 OS o co O CM CO CO ■* CM oo o> ™ d CO ■«-
CO CM h- co •* -I-
CD CM co co Oi tt O) CM CM CO
CO 1^
CM O) O lO 00 O) O CO O CM CD CO ^* CM 00 O) ™ d CO T-
r~- o m a o> tn
?s O CM O) «0 O CO d o f- CO Oi o i- h~ CM O r- co co co
m ■*
co a>
co CM CO o o co co •<a- CM CO CO
CM -«t
O) lO o> o
C0 CO CO o r- o> O T- lO CO
i^ » CO T-
<-> E i- co B>
Exhibit IV individual Regression Results World Market Data
PRICE vs POP
o> in m oo o> ■<* O) CD h- CM O) CM o ■*
eo o o CO CD O) r— co "* a> O oo T- o CD o>
f? CO o CO tri .o o o o -G T~"
y> d d d H5 ■ CD
H—
H to D c CL 1-
.o 55 CO
3 O > 01
£0
CD Qi 03
3
CO 2 HI "E CO
CO c o
< 2 02 CO
Q. 3 CD
CO
co 2
2
Mul
ti R
Sq
3 c CD CO o
Z> CO <
CO
55 o z <
CO
o> CO LO co CM
00
co w ■* CM m co •<- ■* CM co
co m in
CD CM
a> o> ■* co m T- n oo o ■* r- o in co T- T- CO CM
CM en o o o> O)
co co co
i-- co CM CM
CO 3 £ -o - ra'iö ™ CD CD o a: a: i-
«n
CD*
V- CD 1*- i^
co CN co CD a co ^" m CM CM
co O o> T~ CO
O) ■*
o CM co CD oo CO r^ h- O) ^ o o> o ■*
CO CO co
i 1
,_ co r^- ^ CD CM co CO o co T- m CM CM T— CO o o> T~ co
a> ■*
o CM oo CD oo CO h- h- a> T—
o O) o ■*
co CO co
i i
CM O) m O) m CO ■«• m co CO co CM CO T-
o f». CO 00 o o ,_ T_ r» oo 00 ■<*■
00 00 00 O) m in CM CO ■<* co O T—
T— o 1
CO oo CO r~- T— m CD ■*
m ■* ^ in o 00 r^ O) CM O CO CM
O) CM ■t ■* O) O CO I--- m oo o •>- o co T- co
CO CO
_JD
CD CO CJ != i- CO 3 > EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN POP
D D. I- D O >- <
D
.8
42 co c .o 55 to
& O)
m lO co T— 05 CN CM CM CO CN CO CO O) 00 T~ 00 O CM ■* co 00 CN o m O CO ^* T— T— f- CM o CN 00 T™ o O m o O T—
2 CO
a: • * S £ tsS«l Q. 3 •« -n £. .9- Ew "a
J2 w :u iS -a 5a<wo CD _Q
2 o
to
CN
co co
CN m
CN CO m o CN o> T— CN
■* r- co co oo a> co m o O) co ^j cd m O CN
r» T- t- o
oo co o co o T- _ C» CM
CO
CN O) o o co o> CO CD
r- co CN CN
c o _
(D '55 CO
CO CD _ OH DU H
lO o>
co oo o <tf m I*- O) oo CN 0> o ■* ■* Tf oo <°. m o
r». co Oi h- m o co co h- CM m in o> CN
co oo o ■* m h-
en oo CM a> o ■* ■«*• ■*
oo ». m o
h- CD en r- m o co co r-- CM ID m a> CM o> •* CN ™
00 CM o> ^t- m I-- oo CM CD CO O O CM O»
o
CO CO r» o CM T- co ■* T- m co m co c» CM ■* CM CO co d
oo o) o co co o m co o m in o> CD 05
in
O) CO
co oo
co
cü 2 " b i- co 3 > = x
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ POP
8 c
1 Ä CO
co ti- ro CO
CM
oö
D a. i- Z) o > a: <
CO
CO
y>
to c .o 5; to
& <b
* * O) m oo O) OJ CM CD CM ■* •»- m _
CD CO
o) m m o r>- OJ CM T- CO CM CO CM ■* o n <D p °. q ^ O O O T-
<L CO
^c5 to • CO
2 a: LU o
5 a < CO O
2 O
CO
CO r-- 5 ■«*
h~ m o
CM CM
m oo co o
o CO CO CM °. * T- CM
■* r-~ co S CO M- m h- r- o r- ■<* o oo T- co m co gain " *9 CO Tf CO CO ■<- CO CD
CD r*- CM CM
o) co «2 CD CD o o: a: h-
o in a>
m
co r- h- T- oo a) oo ■*»■ ■* co CO O)
O) co
r-- co
CM CO ^ co T— o O) CM ■<*- ■<*■ x~ oo oo o o oo 1^- CM
i CM
i
CO r^ r^ ̂ oo o> oo ■*
■* co co o> o> co m •<!■
r» CO
CM CO T- CO ^ o 3) CM ^t <* T- oo 00 o
CO T- CM CO CO ■<* CO CO h- CO m ■* «* CM CD T- co co O Ö
a> CD CM o oo O) T- CO -i- oo h- CM O) CM CO G> CM d o CM T- CM O in 1- r>- co CM C7> o> CO 00 K-
oo -<t oo o
CM CO CM **■ m m CO CD CM m
co
CD .J5 >- co 3 > -EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs POP GR
CM i^ T— T— 00 O •* o T— CM ^~ co t— ro m CN ■<* ■*
00 co h- CM o CNJ 00 CM CO o> 00 h-
.§ o> o CN m o o O m
(0 o o O ^— s3 i
cn H LO D C Q. 1-
.O 55
at
CO 3 to 3
o >
CO 2 co CD or Q: «11
DU ill "5 CO
T3 C CD
o
< OJ Q.
"5
CO 3
a CD
w 3
co
to
CO :> a: < CO O
o
co
05 o co r» t^- CM CN CM Ö
in co o m oo in
a>
' . CM oo ■* m CM m m co o ■* Tj- CO (O N o> CO ■*■ o co T- •<- oo co
Tj- CO o co
00 CO CO in co co
T- CD r- CM CM
c o U) m m -i !° •o ^ O) in CO CD (D o a: K 1-
o «5
m m CO CM CN ■>*• O T- m o oo co T- m oo n. oo co CN CD
oo co o> CO
CN o co r-~ co T- m 1- r- a> °°. CN oo T- co
m m co CN CN ■* o T- m o oo co T- m cd « oo co CM CD
co co o> co ■<*- CM O CO I-- CD T- m T- r>-
a> °°. CM 00 T- CO
•»* CO a> CN a> •*
. CD O) O) CO CO O CD ■* CM
CD O O
m oo T- CO O CM T- r«. CO i- m CN CO CM 00 O)
ö o G> ■* CM O) CO CD 1- m r» •* o h- m co ,_: oo o ■<* T- CM
CO CO 00 O) CO o oo o) O) -r-
m CM ai CM
CD «S Ü T^ £- CO
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN POP GR
H D CL H
o >- or <
D CO
.O
.<"
co c ,o s; CO
CD m o
T- ■* CO "<fr 00 CM
.„ ■* ■* IP CO Tf h? CO CO
° iri O T-
co
CD
Cd 0)
.£ CO Q. 3
■i= CT
CD L_
CO
tig Q; lil o T3 ^E CD £ tö > = c <» 3 5 w
' TJ is ■" cr < w o
O)
O)
o CO CO CO
O) m m co ■<* o o o
CM r» O) CM CO 1- CM CO
CO T- O CD in CD r~-
O CM
CM co co O T- ■* co oo r» CM O TJ- T- m T-
CO CO
1
0.1
066
26
6363
. 27
6363
.6
»•-. TD
o
2 o z eg
ress
es
idua
ot
al
< o: a: i-
o id
T- m oo o> m m o>
co oo m oo
O)
3
r» co r- CM o ■* 05 O) o to m oo m co co m °°. d CM a>
T- m oo o> m t-
co oo m oo
h- co r- CM o ■* a> o) o co m co m co co m °°. d CN OJ
a> CD •<j- a> oo r- ■* o co m T- co co oo o O) d d co m co o) -* o
CM oo T- o r» CM CM o CM d
CO CM
m r*-. T- en co o co co
co co co
■* Oi •*t CO CM T- 00 T- h~ in oo co in co
at
a> <H " E i- co
Ä> = X
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ POP GR
Q. H Z> o >- on <
00
■2 oo c ,o 5> CO
& CD
CD CM CO CO 0> O) 00 00 CD 0O 1- O) CM CO St« o o •* r- ,-; r- o ° o CM o
CM
01
Q.
PJ
CD 03
w t a: ai
cu ■o "E OB 2 TO g; w -D a,
■2, £ w — -o iS JQ ce < oo o
■^3 CT - W
U. m 00
CD C1
o
C o CD ^ fi
oo CM
Ol O
L0
co ■t oo r~- oo
u.
7459
1
.209
6 30
62
CO o co 00 CM
Ü
0745
9
296.
2 29
268
2
44.9
10
014
00 CO Ol CM
00 <N CM ai CD T- O O) CD O) CM CD CD
T- h- CO CM CM
«fc. T3
o
2 o 2 eg
ress
es
idua
ot
al
< ttKh
3? o «5 o> cB Id
oo a> CO CM co CM co co m x* r>- co oo m
co
co CM CM ■*
oo o) co CM CO CM co co m ■* r- co co m
CO
co CM CM •>*
co m o oo
Is- T- 2 oo O CM x- d
O) o Ol T- oo oo o> m T- oo o a> m a> CM o CD
oo -t o CD O) o> in CM -a- m co CM m T-
•* CD in in !•*- CM oo
CD
Q. (11
n co
n 11 L_ CO CD > C X
Exhibit IV Individual Regression Results World Market Data
PRICE vs GDP
CM m (D O) O) CM CO m m CM O ^ O) T- ■* r^ O) O) O ■* CO O) 1^ co <o m CO •^ m oo
10 CO o co a> o o °. in
to o o O T- H3 1
en
t- co D ST Q. ,o CD
H CO CB D CO = ,_ o
O) »^ >- a:
0) on (11
djus
ted
R
tand
ard
E bs
erva
tio
<
D
Q.
"5
li- re 3 O" (0
CO 5 a < co o
u. CD CM
CM 1~ r^ m o C1 o
ic- m cr OO o> o to
O) CO o co CM
u.
5555
0
.036
4 19
27
CO oo m
.307
9 25
5.5
o>
Cfl CO if in CM i- m o o m CM o co o T-
(O a> o> CM
CO o °> °> ". co O) ro co co
T- r^ co CM CM
v_ ■o
c o
2 o z eg
ress
es
idua
ot
al
< a: a: H
CO CM öS h- m o oo r-
CO CM id T- r- O) o o V» CM r» CD
1 47.
8.87
CM h~ äS r>- m o CO T-
o) co If) 1^- CM O» CM CM i- ■* CO 9? m co £ r-; co o co' r^ —i *
CO CM
^ r- m oo r-
lO CD CM O) T- i*- l^ o o CD CM r- Q.
47.
8.87
§ CM r- r-- m
in CO T- a> co h- CM CM CM
i- Tt CO
1 m co h- co
o co' r-^ i
T- r>- co h-
1 OJ CM Oi CM CM i--
JP 0O o > o o oL o oo
o' d CM o CM CO r- CM
H^i m co TO C» -<il-
fc O 00 l-~ o
*♦*, CO c» r-- T-
CM CD
U* Tt CO o oo oo
£ r- co co r» CM rj-
■o ■* CM is co oo
■2 m m t3 C cn
00 O) OJ CO
ti CO
o> oo CO m o>
S-S co r~- c 00 •* CD O) CM ,C h~ "*
CD h- in & in
O CD ^ O CM O
CD
S--S OJ 03 Ü «- <- co 2 > sx
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN GDP
r-- CM m r*- co CM O CM CM 00 r» co o T— CD T- f— OS r- T- •<* co r- co m m CM CM
to C\J o r-- T- ■G co *■; °. ui to o o O T- s3 m
1- 10 D c- a. ,o ^ \- to CO Z) to 3 ,_ o > Q>
stedR
d
ard
E
on < 2
u; Q. 3
k *-■ O1 3 C
CO 3 W 5 a =5" a
< CO
en
2 O
co
r-- 0>
co o
co co CM co co CD r--
CO
co co CXI CM CO o 00 CO CO O) CM CZ> CD r» T-1 00 CO CM h- CM
co OJ CO co oo oo CM m CM CD h- CR
r- o CD CD
T- r- co CM CM
(/) en w -i CD "O o> V) CD OJ Q: a:
CO
5
O) a> O 00
£ UJ £ cp co o)
CO o
S» CO ri ^ "" CM
CJ> T- CJ> T- co ■*• O CD CD CM CO CD ■* CM ■* O T^ ° ■* d
i^ oo CD CM CM CD
CM m oo CM oo oo m co
in 9 CM O
a> T- CX> <t- co ■* o CD CD CM CO CD ■* CM Tf O ,-■ °. -* o
h- co CD CM CM CD
CM w 00 CM 00 00 in co
o Ö
m co CM O m 1- co oo co o m CM
CM h- CD CM o co in CM
CO r» en co a>
co o
co en oo o in co CM m m oo in T-
S.-S tD .f5 i- co 2 > -EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ GDP
ü. H D O > a: <
D
.8
a to c .o 5) to
s. O) Cb
DC
CO O) N it oo Tj- 00 h~ •* CM 0*000 00 00 CO o M- T- in in O T- <35 O) o co o m O) eo o co T- O O Ö Ö Ö
lO
CD 1— CO
tig a: w o
I ™ «! öj >
5 a: < w o
(U p
2 O <
CO
m
CM CO CO
m oo o> co CM
r-
co r- co co ■* m T- CM
00 o> ci in CM CO CM CM
co a> in o
co co co'
a> co co CM T- CO CM co co
T- (D S CM CM
c o _ CO 3
05 W CD CD
££ Di
3 ^1
m co
oo ■* O) o o> T- r» o o TI- o T-
co h~ CM co
m co
oo ■* O) o
r» o o ■* q T-^
■* oo
o
CO UJ
25
CO h- CM CO
CO CO CO Ö
co in o r»
00 00 CD OJ CD Ö
r»- oj 1- O) o r-- r- CD 00 r» 00 ■* CD m co o
CM CM
CM in in co
CD <5 <- 03 £ > EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs GDP GR
H Z> 0. H D o
on <
CO
■S CO c .o to co
or
CO CD CD CD CO O CO T- O) CN co m TJ- o CD ID O TJ- CD CD T- Tf CO N O) ■* "«t T— ■* T- O <* O 00 " o q „■ O O O 1-
01 L_ CD
3-* CO t
rv a: w CD W 0) Jo Q. = co T3
3 w ^2 2 a: < co o
to
co CN CO f- co CD O) CN
ID CO CO T—
05 o O y—
o CN co CO r- lO w "* CO CO CN CN
ID CO CO CD ■* "* Ol (ON O^t OOr CO CO CD "^ N-' CO m o> co CD O CO CN CD CD
T-(DN CN CN
CO CO to -I a> TJ _ O) 10 03 a> CD n
OH OH 1-
os i» CO
3
oo •* oo o>
oo m r- o CD CM <* o o) oo CD CO O CD CN CD
CN O
oo ■* oo o> O CD CN
CN co ■* CN Ö
oo m r- o CD CN ■* o o> oo CD CO O CD CN CD -* ^ CN O
oo co
CO
111 CO
CN co r- 00 CD
8§ CN°
co in 00 CO T- m in ■* o> co CO o oo •
CO CD m h- I*- ■* ■T) Oi X) CN T— CD n in » ■* *— CN
■* o
co in co m CO CM oo r- co oo a> ■* O T- oo co CM <N CO o
_CD
S--S e CO £ > E X
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ GDP GR
Z> Q_ I- D o > <
=) CO
■2 co c .o «0
co co co m CM N T" CO r- T- CM O) CO T- O) W (D co co m m S t- M CO CO o o
o o
Oi
Co
i. 1 s 11 5 a < co O
o: ©
CO u. co (11 co CJ c m
CO
£ co tr o CD
CO
CM I-- co o
IL
4117
0.
0376
41
67
CO CO o r-- «a- co m
CO in'
O) CM
r- m ■* T- CM T- 1- T- o ^■0)0 co in T-
CO
.618
7 68
99.
909.
2
co
O) CO
T- r- co CM CM
»»-. T3
c o
2 o z eg
ress
es
idua
ot
al
< cm tr \-
CM CO e? CO T- o CM lO
CM h- ui T- O) O) 1- co
tu £S D • CO 5 CO ■>* 5 00 T-
m co SS co o o in en
CO T- ui CM CO o> CO h-
1 8§ T- CM o ■* T-
~J 1 1
CM CO
e* CO T- CM lO
lO CM r» O) T- O) v. T- CO CD
1 n CO CO ■* CO T-
m co 00 o
es
O) CO T- CM CO co r>-
I o °g ■<- CM
T T
m co O CO
CO I-- CO CO T- m co
SP r- Is- i> T- "*
i r-- oo
o
r- co co m CM m
•+»« f- CM CD CO T-
CO oo o o) «a-
•ki CM o> r- i- Ö Ö
V*. co ■* o h- Oi s c» r~- lis CM 00
O h- ■o CM lO c h- T-
s . t™-
c T- m CO co co
■*i co
■«*■ oo jo CM TT c
.3>
co o n- oo ■t CM 1 T- CO m co
05 o O
^ CO- CM CM CM T-
T™
(D
Q.-S a> .2 o c >- co 3 > £ X
Exhibit IV Individual Regression Results World Market Data
PRICE vs 1/PR0D
=> Q. I- D o > a: <
D CO
co .U *G .«0
to c .o 5) CO
g, a:
at
CD h- oo en CM O ID CO CM T- OJ CO CM lO r- o o o Ö Ö o
CO CO Oi I»- CO CM
r-
m TJ- T- m CO O) """ in
CD CD
OH a, ® CO Q. 3
••3 cr 3 W 2 a:
co t a: LLJ
CO ■o 3 C
=5" iS -Q < to O
o z: <
CO
CO
o CD at co o
Ol CM co co in CO
■* CO T- CO If) CO CO !•» CO IT)
co in
CO Tt co m
CM
■* CM T- CD If) T- 00 CO CO CD ■* •<a- CM ' . CM 0) (OSO co co o)
CD CO
i- r- co CM CM
c o _ CO 3 CO -g
D) CO CD CD
to cc
a: a: H
o
O)
cB
lO O) v. CD
If) r~- If) If) 05 CO CM if) co co if) CD oo •»* oo "* If) If) CM S
r-~ If) co eo CM CM CM Oi co Tj- 00 at T^ CM if)
m r>- lf) If) Oi CO CM if) CO CD If) CD oo ■* oo "* If) If) CM r»
f- if) CO CO CM CM CM at CO o
■<3; CO ai T^ CM in
CO
CO h- . T- O CM CO T- TO O) CO O) o at r- * d o
CD ■* CM CO CD m CM CM oo -<t co at oo at CD C- i-i CO
T- oo CM CD O CD ■»t 00 o co CM in CM CM CO o
at co
co at o m •* CO CD O 00 m s- o oo s s-
CD
CD
CD 5 CJ !=
3 > £ X
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN 1/PR0D
3 Q. H
o >- a: <
D CO
•c
42 co c .o 55 to
& 03
■*<DS N O) co o m co CM ■** IM "* ~" ^ \a "* io oo co
oo o O "* °. CM o
o m
co
CO O)
° in
2 CO
tig ™ ™ ~° 15 CD
■2 W 03 CD > t o- % E «3
' co O 5*3
2 o
to
co CO
o CN m co O)
co
co m co o i- co o o O) co r«- oo
H in o CM
co -<a- M- co CM r- 00 O O CM O O) O) T- T- ^ CM
öS a> o o CD co
T- r- oo CM CM
o w CO CO 1 (!) 1— T3 O) V) (1) OJ a: a:
O) CO O) "* r- O T- CO CM CD m o co o
ö co CO •<-
co m T- O CM m CM m
co
Q> CO Q. O) So
co
O) co O) -«* i- o ■«- co CM CD m o
co m T- o CM m CM m ■* m CM o
co
co to CD 00 CO I-- 00 o CO CM ■* IO m co r- CJ> o o> d d
"* T- O T- co T- co co oo o) a> r- oo co •<*■ o oq o T^ d
Tf co co m
o f- o
o t-- co
O) -r- 00 CO ■* !•» oo ■*
h- ■<*• c3> ui oo o oo d CM ■
OJ ™ cj E i- co 5 > S X
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ 1/PR0D
CO CO CM ■f ■* o> CO O 00 CM in ao o> ^— T— oo CM o r-~ o CM T—
l>- w CD r- CO CO CO CM
to CM m ^— CO ri o o Tf
."> ö Ö ■^~
■G CD
H«
h- CO D c Q. 1-
.o 5) 03
3 to 3 ^ o > on
o- C0
o to
CD ft: OH CD
K. 111
"2 CO
CO
_o <
Z>
CD Q.
"5
co 3 CT
CO
•o
"cio _3
"to
Co to £1
CO 2 a: < W O
2 o <
CO
CO
o CO
CO CO co CM
00 CM T—
m co
r- ■* oo CO r- o oo T- ■* CM T- m o> O) T^ co 00 T- CM CM
r*- M- co oo oo co 1^ CO CM 00 CO
a>
co a> oo co oo o oo
oo r- ei CM TJ- Tt
T- CM CO CM CM
c o to en to i tu T3 C3> to a> CO
CO
a: tr \-
•* 05 ■* CO CD ■* co ><t a> to 01 co
§3 O CM CO
CM 00 m at 00 CM f- a> CM co O) oo CO CO m co co co' co co
■* co CO "tf CO ■* O) in O) co
8.8 O CM co
CM CO m a> oo CM h- O) CM co O) oo CO CO in oo co CD CO CO
CM CO 00 T- co o oo co O "* O) co CM CO m co OJ CM ö c>
co in 00 CD O) T-
co ai r-- oo O) m o
O) CM co m CO CM r» co CM CO ■* T- OJ CM ^ CM
CM
00 CM CO o O CO
co f- o CD CO o> o> r- oo ^CM 1- oo
CD
CJ S i- CO
B> S X
Exhibit IV Individual Regression Results World Market Data
PRICE vs OEUR
D CL I-
o :>- a: <
=> CO
42 00 c ,o 55 to
& 03
<D w CD eo CD CM
m
m o CN
o> o t^ CD
CO co co CM m
o _«:
CM ■* o 00 CD
O O T-
03
CO
Qi eo o> eo Q. 3
o: in X3 P
fag ü> CO
CO
CD
a: < oo o CO
2 o
co
CO r-
CD ^- CN
CN
T— CN m CO m CD
a> oo o> oo CN CN co in
o in
CO CN
■* T- CO CN
O) CO CO G> CO CO CN CO CN CO a> •<!• CO
t- CO m oo
co o)
CM CO CN CN
c o _ & ro If) 3 £ -D O) (/) 03 CD
QL a:
V)
V. 0)
I
CO T- OJ •* ■* o h- CJ3 co oo CN T-
CM *-. CN O 1- o
co m m CM OJ CO co oo 03 CO in m OJ CM co o d ■* T- CM
CO T- O) ■* <tf O r- as co oo CM T-
co m m CM o> CO co oo 03 co in m OS CM co o ö ■* ■<- CM
OS CO o r- m ■* oo r-
CM O) CM CD CM T- O CN O Ö
T- CD
03 T- m h- ■«* CN CM -^
CM OJ T- in O T- 00 00 co m ■* o co co 03 03 CD O) CM CM
CM ■* o CO T- r~ co co co
co o CD CO CD CO
co 2 u != i- co 3 > £ X
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN 0EUR
■* r*- c£ in co o T- ■*
oo o io co o> O) oo co v. T- CD CD
# «N oo oo co ■* CM
CO h- as ■«* CO o oo co
in •«* wj ■<* CM O) o> <* v. O CM
I *— T— ■«t d o T- CD —J 1 1
cs IO co ■r- -* in 00 o O) CO o> i-. 00 CO CD 1- CD
# 48.2
26
8.
■* CO f>- u. m ■* CO CD ü c
1 o 3?
CO O)
oo co co
•<3- CM CO 0> Tf co
i o
O CM
£ co ■* d
CO Ö
■ i
O) r>- ■* ^ CO m 00
CD 3
oo o CM h- 00 1^ T— ^
u. oo CO
co co oo eo CO T-
o CM CO
->- d d
CM lO CO T- m co a> in CM CO 1- h- h- 05 ■*»« CO CO
CO 00 CO CD CO CM CM a> CO o> a> CO CM ■* T-
O CD CO T-
•*w CO CO
CM CM ■<- T~
CM f- CO U» ■* CM m co CO 5 a> o CM lO CM £ CO CO r- CM x— h~ CD CO CO CO O) 00
CO CO
CM m 00 73 co m m. oo o in
oo oo
C S
CM O) o ■*
co r-» O) c iri cd CM -<t ■*
CO T- 00
O) 1^ ■<* |-~ -<fr •>- CM CO CD CO S O) CM 00 CM CM CM i? in o oo vo co co c
CD T- o
CO 00 CM CO T- O 00 CO T- o CD "tf ■* T- oo I-» =o 1 ■* r- .8 •>- CO CM O in m CM ■* o r~ CD O CM
«G ö<=>.°* O h-^ o> .CO O O T- O T- 00
TO Sw
1- CO
CL c .o 1_
1- CO 03 D eo = u- o g.
lulti
ple
R
Squ
are
djus
ted
R S
q ta
ndar
d E
rro
obse
rvat
ions
> a: <
CD a:
2 O z
c o CO 3 CD -E5
CD CD US 'S te
rcep
t V
aria
ble
1
CO 2: 01 < CO O < a: a: H -EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ 0EUR
D CL H Z) o >- DU <
3 CO
-2 co c o to
& Q:
n N N ■* ro m i- co in CM
co
CM CO
CM O
in CM co
CO CM
1^ o
1-^ o
o o o
CD
CO
cr co
ai (U o 1_
CO Q. 3
or 3 (/J
2 CO i- c
tt UJ o
"S "2 '™ S CO >
.5, E w — -o £5 -Q a: < co o
2 o
co
CO I-- CM CO Tl- m I-- o
co m 00
co T- oo
CO
co co CD TT O 1- CO T- CO Tj- CO 1-
CM CM
co ■* r- (D 0)0 o o o CO 1- T- co co CM
CM co co co o
T- CO co co
r- to CM CM
o _ (/> -J
O) </) <D tl! o: a: i—
i-» co <*; CO ■* o o oo
a> ■* uj 1- r>- o» 00 T- v o oo (1) CM r»-
1 CM T- co ■*
oo oo * r- o o O CM
o ■«- u> CM CO o> co m I- T- Tt fl) o r- s Is-' T Ü CM CM ~J ■ i
f» CD OS
co TI- CS CO lO co TJ-
U) T- r~- I. 00 1- tu o oo Q. CM r--
£ CM T- CO Tj-
oo co r» o o CM o T- CM CO co m ■>- ■* o r»
CM CM
Tf CD T- 1^ CO CM O CO 00 •* CO lO co r»- m o oo
oo in «* m TJ- oo co co CO T- r- co co co i*- <* T- oo Ö T^
CO "* CM CO h- CD CM 00 OO Tf CM ■>- CO T- ■* 1^-
■* Ö
r- co CO CD h- 00 Tl" O co co
co co ". CO- CM T-
S--S a> IS ü «- c- co 3 > = X
Exhibit IV Individual Regression Results World Market Data
PRICE vs 1/PR
S? £ 2? 00 f-- O ■<* 0
«3 "* o> co m O) o> 0
i- co ■* 0) O T- §■ T-" Tt ^ co m
^ CO CO O co co
T- CO W5 CM CD o> 0 r~- v. at co 3? r- ■* ä CM q 0 -j ■
00 OJ
^ co r- ■* 0
10 ■* OJ O) co m i- OJ 0 OJ CO ■* Q. O T-
§ co 10
CM T— T— u. co co co CD 0 SS co co Ü 10 10
Oj 1- 00
c O) CM CD CO co 0 h- CJ 00 !R o> CO IC m 1
0
r- ■*
.1 0 0
CM 9
to
10
1 6E-0
5 95
062
u. co CO £; tt> CO S 2 °° c» 05 m 00 OL CD 0 CO co 0
O) CO r«. co <J> O) CM •* co co CM CO T- CM ■*— O CM
to T- CO CO OJ OJ CO CM co cq to CM CO T-' CO •*«* h~ CM 00 CM
O) -<* ■* ^, m co OJ CM ^ O CD 0 CO 00 0
& ■<- CM
T- CO 0 T- CO ■<- CM T— T- CO
to to
00 00 CM tj r-- co
i- CO oi O
f- 0 o> c ■tf CO 00 co co 42
to T—
rooMno) T- h- co O) T- r>- 00 T- CM CM CM CM ¥> m CM O N 00 CO c
CD CM CO CM CO CO T- (Or m 0 a> a> T- T- 1^. ^J- fc § ■* CD
(0 in CD co in fc ■«- TJ- m CM en o) CD a> m *°. T °. ^ 0 0 O T-
O O ■r^ CD
CM CM CD •***
H to D CL c
.0 H s CO D £0 = t. o £ W gg
rv UJ 0 > c: ■"-
a: <
D
Q:
lulti
ple
R S
quar
e dj
uste
d 1
tand
ard
'bse
rvat
i
O 2
0
£ s O) en Q) 01
TO O te
rcep
t V
aria
ble
CO 2 Q: < to 0 < a: a: Hi £X
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN 1/PR
m •* in co T— CM CM o o. CO CM O ■* "* in co ■* oo CD CO in CO CM in
to CM r^ ■* a> T- at CD T-
«c co o °. iri y> o o O T- ig m
h- to D 0. c
o I—
1- (/I co D <0 H ' o > D)
(11
w E a: uj
< 5
Q: U.
Q. CO 3
£ co :> CO M
ulti
RS
q
Adj
u S
tan
CM
u. m co
R in
£ oo a>
c- o Oi ö to
co oo oo o
u.
5071
2
.923
0 57
43
to T- O) CM OJ o> CO ■* O h~ CO CD CM
T- t- ■* r- o T- o in o m co o 1- 00 T-
to to
1
674
.92
27
6234
.2
28
6909
.2
<4- T3
c o
2 o
egre
ss
esid
ua
otal
< OH IT H
•* CD SS CO T- o co a>
uo o) lO r- 00 O) CM T- V» ■«- m 0J "* ^~: a CM iri s CO 1-
T- ■*
SS o> o o in o
i-- co in r- o) o> o r- k« CD CM 3? CO CO ä oo ■* o Ö T-^ -J ■ i
•<* CO CO T- co O)
lO m o o> r- oo k» CM 1- OJ T- m Q. ■* r--
§ CM in CO T-
T- Tt" O) O
m o>
m o h- CO I-- CO o r»
i- CD CM I CO CO oo TJ-
o O T^ 1 1
co m T- CO
.3 o ■* ■* m m o)
S> CM h- :> CM 00
cL co o> o o o d r» T- CM CO o. en
s— T- o CD co oo
to o co in o>
tw ^- o en r^
■■- '-
^ CD CO ö CD in
ul •<- o CM CO in in
•Q in co C o> r-
c CD
o co
00 Tf ■*;
to
r^ co i2 co r-- c 0)
h~ m co o co m 1 co m h- at
05 t~- m o in T O r- r-
OJ
"S--S Q) w o «- i- CO
3 > S X
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ 1/PR
•<*• r- x— CM en O) T— CM m CM co r^ in en en co o CM ■■a- CO T— CD o> m m r>-
.8 T- T~ o ** o CO CO CD
*c ■* *" ■»- Tt-' .*> Ö d d •c CD
<hj
H- CO
Q_ .o cu H CO CO 3 CO 3 L_
o 8. D" CO a: "0 CD
"</) _3
<
o w >- 01 <
Z> (0
O)
Mul
tiple R
R S
quar
e
LU
"E CO X3 c CO
CO
g
w .a O
u_ ■*
CD Ü
5 o c CD CO CD Q O
5 CO O
.O) d to
oo
o u.
7651
5
.202
8 05
05
CO h- CO CM m
§
7765
1
111
6.
3236
2 2
14.5
10
014
to CM en CM CO
1
11
16
27
5792
28
6909
H~ ■ö
c o
2 o gres
si
sidu
al
tal
z <D CO o < DU K. H
^ CD h- r» T-
O cn m vi Ol
■* cn ■* m co co
#
co «S m co CM CM
CM 1^ ^ m o o m m
CM T- Ui CM CO O) co m SR o ■*
i T d m Tj-
o ■«t T- -J ■
CD 1^
^ 1^- r- cn m
lO ■<*■ cn en ■* m V- co oo 0) co .„•
# co *Q m co CM CM
CM r» m o
lO o>
m m CM T- CM CO co m
v.
1 o
o ■* T d m ■*
■
■* ■* co t-
1 O CO in o m co
SP Tf <o > cn o CL CD CO
m o d d
m T- co r~- o r-
cn CD in cn co
■US oo cn CO m o H^ r- oo
m CM d CM
i
k^ in TJ- O r>- c»
£ T— ^r- CD 00 CO CD
73 ■<*- CD C CD cn ■8 *-. CM c h- T- co T- CD si to
co ■<+-
i2 co co
.1> r- co CO CD co m 1
CD
-* T-
ss o O
Oi CO ' T-
0>
^■1 o> .2 o E fc- co £ > £ x
Exhibit IV Individual Regression Results World Market Data
PRICE vs 1/RPR
<£ ID 00 r- ■*
o ID ID
V) o>
co co O h- o a>
v~ ID CO CD
1 CM r-
£ CM CD ■* o
O co O) «5* O)
ID CM CO ID co O)
i^ ^— ^— 1 O
~1
CO CM
Tf ID
ID CO
c£ ID ID ID CO CO O) O 1^- »». O O) a> ID CO
s 297.
74
.5
O) CM CD u. r*- Tf O CO o> M?
ID O)
CO o> CM ID CM
c CO
ID CO
CO ID CO O) 1 m
CM 1 o
00 CM £ o r-' « .O) CO
ö "<t ID
O) O) o> ID f- r- in
3 ID O)
ID O CM ID CM ID
u. CM CO o
ID CO 00 ID O CM
CD O O ID O O
T- CO TJ- CD CD -st ID h- ID O) (O CO c» r-- •*«* i- a>
CO CM Is- i0 ID T-
187.
4 11
.91 co CD T-
f» r- Hw CO CD
CO CO
T- CM CM CM
t- CO ■* V^ 00 CM CO ID T— o r- T- ID <st o £ o> o O) O o ■*J- CM CM 00 T— ■st ID
CO ■* h- CM ■p 00 O)
CO h-^ T^ oi co §• 00 CM O 73 T- h~ O) c co T- ID co 42
CO
O CM 00 CD 0> T- r- oo ■* O) M * in m (N CM CM i" CM T- CM 00 O CO c
CO
ID Tf CM i- O OJ 00 TJ- CD CD O) CO co oo WOOT-N fe 1 ID ID
.8 ■<*• 1- 1- to CD ID •<- r- TJ- m ^ - -. ^
CD O
CM °> ."5 O O O T- o T- CO ■G CO H-
H CO D CL
c H 55 CO D to 3 !_ o %.
lulti
ple
R
Squ
are
djus
ted
R S
q ta
ndar
d E
rro
bser
vatio
ns
> cm <
CD
2 o z
c o 52 ro
2> -D D) (/> CD CD
15 "o te
rcep
t V
aria
ble
1
CO 5 a: < co O < a: K 1- ■EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN 1/RPR
Z> Q. h- Z> o >- OH <
CO
co .O
y>
CO c .o 55 CO
en
Q:
CM CO CO h-
■* in o) CM in CM
CO o CM m CO ^ CM T— r^- O) h- <o ■* CO CM CO O) T~ O CO oo m CM t^ CO T— T—
o o o
cu
Q 2 °"
Co
tgg cm o ■O "2 CD OJ CO > CO "O m
K < co O CO o
to
CM
CM oo
m CD O) o r--
»• o
CM
CO CO CO
CD O CO OT in CD
r» co CM CM
c o _ & ro CO -j
£ -a _ 05 "55 iS CD CD o a: K H
o vi
0) S: oo CD .£*■ rvi *+
O) CM T- OJ CM CO O ir- in CO OO •* a> CD co TJ-
CM ■*
r«- co co a> m co CO h- co co
CO O) to in CO T^
O) CM
CM co O T- m co co ■<* O) CD co Ti- ed cd CM •*
h- 00 co a» in co co h- oo co o> •*!• m o co ro in in 00 T^
CM eo co co
O) T- r- <o in r- O) co r» co CM o ö o
■<- CO CD CO
r-
CM h- O)
CM
i- in Tj- CO
co o co m
o> CM
0)
cu .«5 t- 03 3 > EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ 1/RPR
CM ■* CM 1— ■* r^ CO CD CD CM r- T- ■* CO a> ■>* CO *— CM T- ■* O) CN m in T— co co
^ T— T— 00 CN .0 r~- m "«I CM co 0 0 O r- *G en
H CO D c CL ,0 CD
1- CO CO D CO 2. ,_ o cr 0
CO g >- a: <
CD cr: CD Q.
CD
cc 3 st
edR
dard
E
5> •t-l cr 3 c D CO
3 «■» 2 a: < CO
% o z <
CO
co CN O CO CO CM O O O
m co 00 in ■f o co r«. CO CM ^ ». o o o m CO -i-
m T- m co T- o) ■* ■* 00 CO O) CM t- co 00 CD m T-
r» in co r» 00 co CM in
1-0)0 r- CM
C O _
CO 3 £ -g CO "co 0) CD a: ai o
I-
m r- 0^ ■* r- 0 r- T-
in •* in r- CM O) 0 co ^ co 0
# 168.
-7
6.6
O) CD öS CO 1- 0 m m
m co in 00 m O) co ■* 1» O) CO 0) 00 ,_: ä O T- 0 CO CM -j ■
m 1--
S5 m m ■* O) T- CM V» 0 CO Q) co 0 Q. CO <°.
§ CO CO T- 1—
o> CD CO T-
öS
o>
m in m co co in co •«*■
I- a> co I 0 -J
00 ^ Ö T- 00 CM ■
CD CO O CM
0> UJ§ 5 CO
co co 2 w s O) 0
-1 CM O Q. CD O
CD Ö
r» co o> CM T- Tf 00 00
^5 T- CD •4Ü T- O co r- ■* Hw co co
05 Tf in ■*
^ co 00 O ■* o> E co ■<*
iS h- co T3 0 r-~ C 00 0
■8 co co
c d c\i co CM CO -ti co
a> r~~ CO TJ- CD •w T- O) C 1^ CM .® h~ •* ,S O) CM s in CM
OJ ■* ■* 0 CM TJ- O
1
v-
<D
Q.-2 CD .2 O = i- co 3 > £ x
Exhibit IV Individual Regression Results World Market Data
PRICE vs PRCR
a. \- Z) o >- Cd <
D CO
42 co c o to CO
& O) CD
Q:
1^- •* c- CO CO T"" m o r-- o CO T—
O) <* CO o T~ if) m CM o CM o CO m O co CM CO ■* T— a>
CM
O O O 1-
8> CO
a) i_ CO 3 cx n
CO (Z UJ
■a ■p 0) CD
.9- 5- OT T3
3 CO ^TO
s a: < co
u. co CO
sr ■*
m T-
fi o «r- o c o Oi co-
CO
in co co CO
u.
1141
2
2.32
7 72
39
co <7> CM O CD
Ü
9114
1
317
4.
9175
4 1
42.1
82
895
CO O O 1-
CO TJ- T- in t^ o f- T- h~ CO co CM in
T-C90 <r- CM
>»-. ■o
tz o
2 o z eg
ress
es
idua
ot
al
< tr on \-
CD CM i^ T- m o CD CO
m m id T- CO o> CD CO ^ m t- 03 •* CO
1 °°. in
O) O) äS o> co o T- r-»
co ■* ui co -a- O) CM OJ i- ■* CM
1 -* co- co ■* o CO T-
~J 1 1
CD CM
c^ T- m CD CO
m m m o> T- CO i^ CD CO £ m 1- Q. ■* co Ä °°. m' ^3 r-- in ■
G> O) a> co
m t- i>~ CO ■* CO Tf CM a> V. Tl- CM I
O -* co- co -«a- CO r-
i i
■* CO CO T-
1 M- CO CO h- O) ■*
CO a> r- >. o o <t CM O
o
■* CO O CD in co CO CO
^5 CO CD ?- m i- CO r~- in -+M o> CM
CM f-
■ i
w^ T- OJ o O CM
fi T- TJ- m o CM m
■D CM CO i~ CO T-
■S co o T3 CO c . *— co <3> CM
■<;
co 0> CM
CO T- CO <t* co 1- C m •>* CD CO CD ,G CO o
CD O) CO ^ oi
O CM O) O T— 1
1
T-
cu
CD .2 o c i- co £ > ■EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN PRCR
D Q. I-
o > en <
CO
.8
CO c .o CO co
& as Q:
CD Tf CO OO -I- ■* o w m CM T-IDCM^t 0(0 t<- 1- 1- CD 00 m eo o> -I- o> to io ■* CO "tf lO CD
"* CN O T-
o o
CO 3 O" (0
a: p a: w •o E
_ .. O o
— /n
.32 cö
(0 3 C
2 on < w o
IL o> CO
r- CD TI ■<t
£ o o
C o CD Ö
co CD
O o> CN
u. CM CD
h-'
-<t 05 in If) oo r- 00 ■*
oo m 0) en £ co oo
co o> co m CM
■"~
■* T- m m "fr o> oo O CO oo ■f CM 00 a> 00
CO CD Tf T-
CO CO CD m CO CO Is- 00 o oo CM co in
T— a> o T- CM
»t-, T3
c o
2 o 2
CO C/5
L-
cu esid
ua
otal
< K en h-
CO 00 s? CO T- o CM CO
O) ■* IT) ■* o O) T- CD ^ CM CO CU <* o>
# a> m i
O» T- öS r- CM o CO CM
CD m lO CM CO O) m T- w oo •*
i °. * co m o m T-
-J i
co oo
es CO T- CM CO
w> o> ■* o> 1- o u» <r- CD 0) CM CO a. "* o*.
§ r^ T^ a> m
■
G> T- h- CM
in o>
CO CM CD m CN CO m 1-
!R 00 ■>*• i o
° <* co in in i-
r- *- o o>
0) tu s 5 CO
1- to
5 CM O _i •«-■ o Q. CO o
O) Ö
t-~ oo CO o O CM CO CO
CO CN CO •J; T- m CO in ■* •*■* O T-
i- CM 1-^ ■*
1
w. ■* CM o O T-
ä m ■* CM T-
T3 T- o C a> oo
■§ in ■* c d ■* co T- CM -o co
CD CD CO O CM •*«* 00 00 c f- Is- ."> CO Tf ,5 CO h-
m T-
™ co o in o O
CO
Q.-S (D .2 Ü c t- CO B> £ X
Exhibit IV individual Regression Results World Market Data
PRICE vs SQ PRCR
h~ CO CO T- OO CD CO CO in CO O) O 00 CM Tf h- CO t^ O CD o> (O T- CM c- •«a- ■tf CO
i/j T— t^ ■* m ■* ■»" T Tt
<o O 0 O T- "G en
H to D cr Q. .0 V t- to CO ZD CO = i_ o >- CD
sted
R
da
rdE
< 2
It It CD
Q. CO 3
2> ♦J CT 3 c
CO 3 «■» 5 K < CO
CO u. CM
8 CM
f- co m 0
£ CM c- O m O
10
IT) CN
a.
0109
5.
7142
85
15
to in o) CO T-
§
5010
9
120
6.
5990
4
21
1.
1001
4
to 0O CO CM
to CO CM O) OOO N S O) 1- m co
r- r- 00 CM CM
S-. T3
c 0
2 O
egre
ss
esid
ua
otal
< a: a: H
r- 05 ä? co m O m m
r» o) «j t- OJ O) Oi 0 i~ co 0 03 CM CM
1 CD T O) 0
■
35 CD CO in r>-
0 00 r*- 10 in m
in m !R 00 h-
1 0
co in co ^ CO '
r~- 0
öS co m m m
m r- O) o> T- O) v*. o> 0 CD co 0 Q. CM CM Ä CO T S O) O
1
CD CO m r-
IO O)
00 r- ■* ^- m m m m
1^ CO 1*-
1 O
co m CO T-^ CO '
■>*■ 00 m CM
1 CM CM o> r- CO CD
SP CM 0 :> 0 ■*
«t O CM 0 0 0 0
CO h- co m CO CD CM CO
CO ■"«i- in ■♦2 m M- to 1- 0 iw CO 05
CM CO
■* CM 1
I*. CO CO Ö m co
£ co m 00 co ■* CD
■Q co co C t- -«i- ■2 co in T3 C m" *? co t- 0 Hi to
CD •** to o> ■* **^ ■<- CO c co co CD CO 0 ,G CM CO
CD 1- r~- 00 ■*
O * °° O CD O
1
CD
"S.-S CD W O i i- co 3 > EX
Exhibit IV Individual Regression Results World Market Data
PRICE vs OPEC
D Q_ h- D o >- <
3
-2 to c .o to to
& O)
Q;
oo m co o> ■* CM o
o> O in oo CM 00 co m CN CO v- in 00 co co ■*
CN oo in •* ■* T— T—
o o o
CD
CO
&c5
OH
CL
a: ai o T3 "2 S CO
3 C
Of < CO O
2 o <
CO
■t •* co CO r- ■* o CM o d
CD O) CO T~
CN T™
co CD O CO
oo m C- CD CN 00 CN CN T- O o r- i^ °> co oo CN O ■>- CN
CO CO ■o- 1*- CO V
CN h- o CN r^- o T- a> TT-
O 1- CN r-^ c\i O) CD ■* O CN CD 05 ■r- in CD
i- i*- 00 CN CN
c o
'to CO H? = £ T3 _ O) t/3 CO to o o
DU 01 H
Is- co c£ CM ■* o CO T-
O) h- io oo r- O) ■* m i~ h- m 0)
I CN CO i
I-- T- ^ ■«t o> o o m
CO TT in co r- O) oo o> v^ m T-
i CO 05 T-1 CD o in co
-J ■
h- CO
^ CM ■* CO T-
lO O) h- o> 00 I*- k^ ■«t m Q) >• m Q. oo S Ä " ° ID CM CO
i
f- i- •<fr O)
OS
o>
o in CO Tj- CD h- oo O)
v» m T- i O
—J
co o> T^ cb in co ■
m ■* •>- ■*
co 5 m co CO CO m r»
$ o ■* > co o
d. O CM o o b d r- in co o> O CO CO o
CO 00 TJ- ■is CD CO CO OO CM ■+W r- co
o> ■* CM c\i
^ CM 00 o in oo
£ r*- ■<* o r-~ CD CO
-Q t- TJ" i- T- CM
■8 Tf 00
c in ■* co in 1-
■♦£ CO
CO CO to t- in
■fci OO i- c h~ co .Q> CO CD ,5 in CM
IE CO
o o iri "> o CD CO
O T- CO i
CD
"S--9 CD CO O «- i- co
B > £ X
Exhibit IV Individual Regression Results World Market Data
PRICE vs LN OPEC
=> CL h- D o > <
co .O ■G .«"
•S CO c .o 55 CO
& Q:
o> co 01 CO lO CM CM CM (NO>fflN MO (O O) if) <D 00 ■"!»• r- CM co co MD m o O CD CO CD "* T *r <* O O O T-
CD CD
cr co
a £
2 o: Lü
3 TO > w -o a, .5, E w
— -6^ 3 -Q a: < co O
Q. 3
3 W
2 o
to
CO oo
oo CM o
CO oo O)
CM
oo co
T- CO CO CM •tf CD co r>- CD OJ
co
CM
T- n ■* co oo T— ■* io o CO CO o CO ■* T- h- •* CM co ö oi Tj- CD O
h- OJ T- in CD
T- r» oo CM CM
0 _
'85 § £s -= o) in 3 CD CD o 01 Cd h-
CM T- C0 v-
vi ■* T- O) co O ^ ■* O CD co d a CD '
o T- O) CM O
If) to CO O) OJ T-
k^ oq o
1 o> Ö CN '
o ^ CM T-
U) CO T- o> "* T-
co o fe ■t °. Q. CO O Q. CD '
CM CO ■<- Tt r» CO o o> CM O vo CO
!R CJ> T-
i o
oo o O) Ö CM '
CO CO o oo & w$ 5 "* T- CO O) T- s ^ 00
it CD O
■* d
m to •* CM co o O h-
s r-- co co ■*
to co o <*»* O CM
r-- co if) CM
^. co oo o h- CO
fi O CM o oo Oi CO
13 CM r~ C o ■<* •S oo o ^3 c co
T- O oo d
si to
O) CM CO CO CM
■*^ •«t lO c CD CM .°J h- O ,S OJ CO IE 0) o
r- en CD O CD °.
CJ ■* o 1
T—
CD
CD .2 o c i- CD 3 > £ X
Exhibit IV Individual Regression Results World Market Data
PRICE vs SQ OPEC
3 0. I-
O > a: <
Z) CO
to c ,o CO CO
03 a:
t*~ co ■<!■ CJ) h- CD O CM T- CD O O T- O i*- r- CM «o co CO CM ■* f- o co CO o o m o o
1 "
CD L_
CD 3 ^ O" n
CO ir l/J c-
rr UJ o CD ■p TO
Squ
a dj
uste
CO
c CO
CD W J3
a: < CO u
u. co CM
fl) ■*
c m
CO co
£ co co
c- r» Oi o
CO
<* oo m CO
u.
8964
0
.069
5 96
76
(0 oo oo C-- co
^
8896
4
17.
75
5212
4
255.
2 10
014
(O !■"- ■<* CM
to
1
17
.75
27
6891
28
6909
H- 13
c o
2 o gres
s si
dua
al
^i CD CD ö < K CC H
CM CM öS 1- r>- o T- CO
t^ T- in" O) CM CO o »n V. co m 03 CO CM
1 iri <N CO CM
m T- <£ r~- co O CO ■*
CM CM id CO CM O) «fr T- >>. ^ CM S> co co ä CM °°. o CM CM -j
CM CM
^ T- co >o h- T- O) 03 CM v» o m 0) o> m Q. CO CM
§ CO CM
IO T- r-- co
O)
CD Tf CM CM CO CM
fc ■* CM 1 o
co oo CM °°. CM CM
i
CO CO O CM
Q> äi? 5 CO CO CO h- CO s S co _■ co o> 0. •<- r»-
c\i o
CO CO CD ■* co o h- oo
CO CM CO ■SJ ■* r-- to T- CO •ki r-. co
r~- CM oo d
^, CD CO 6 CM CM
£ ■* CM CO CO CO IO
■Q CO CD C. IO TJ-
« CM ■* CO CM c
co CO T-^ <w (^
CO ■* CO CO ■<*■
H-i CO IO c ■* o .13 co o ,5 r- co £ co oo 03 o
T- CM co co
O CM O i
CD
CD CO ü ^ t- co B > = X
Exhibit IV individual Regression Results World Market Data
PRICE vs SPECIAL EVENT
Sources: 1.MacAvoy, Paul W. Crude Oil Prices As Determined by OPEC and Market Fundamental». Cambridge: Ballard Publishing Company, 1982 (p 189).
2. Houthakker, Hendrik S. The World Price of Oil: A Medium Term Analysis. Washington, D.C.: Ameri- can Enterprise Institute for Public Policy Research, 1976 (p. 2).
3. Ibid., p. 1.
4. Ibid., p. 8.
5. Ibid., p. 7.
6. Pindyck, Robert S. The structure of World Energy Demand. Cambridge: MIT Press, 1979 (p. 23).
7. MacAvoy, Paul W. Crude Oil Prices As Determined by OPEC and Market Fundamental». Cam- bridge: Ballard Publishing Company, 1982 (p. 26-27).
8. Ibid., p. 59.
9. Houthakker, Hendrik S. The World Price of Oil: A Medium Term Analysis. Washington, D.C.: Ameri- can Enterprise Institute for Public Policy Research, 1976 (addendum p. 1).
10. MacAvoy, Paul W. Crude Oil Prices As Determined by OPEC and Market Fundamentals: Cam- bridge: Ballard Publishing Company, 1982 (p. 33).
11. Ibid, p. 56.
12. Houthakker, Hendrik S. The World Price of Oil: A Medium Term Analysis. Washington, D.C.: American Enterprise Institute for Public Policy Research, 1976 (p. 36-37).
13. MacAvoy, Paul W. Crude Oil Prices As Determined by OPEC and Market Fundamentals. Cam- bridge: Ballard Publishing Company, 1982 (p. 21).
14. Houthakker, Hendrik S. The World Price of Oil: A Medium Term Analysis. Washington, B.C.: Ameri- can Enterprise Institute for Public Policy Research, 1976 (p. 5).
15. Ibid, p. 14.
16. Ibid, p. 6.
17. Ibid., p. 17.
18. Ibid, p. 19.
19. Ibid, p. 3.
20. Ibid, p. 61,67.
21. Yergin, Daniel. The Prize. New York: Simon & Schuster, 1991 (p. 541-542).
22. Ibid, p. 550-553.
23. Ibid, p. 499,538.
24. Twentieth Century Petroleum Statistics: 1998. Dallas, Degolyer and MacNaughton, 1998 (p: 2-3).
25. Yergin, Daniel. The Prize. New York: Simon & Schuster, 1991 (p. 586).
26. Ibid, p. 702.
27. Ibid, p. 588-652.
28. Ibid, p. 718.