• No results found

How the Price of Crude Oil Affects the Swedish Stock Market

N/A
N/A
Protected

Academic year: 2021

Share "How the Price of Crude Oil Affects the Swedish Stock Market"

Copied!
33
0
0

Loading.... (view fulltext now)

Full text

(1)

J

Ö N K Ö P I N G

I

N T E R N A T I O N A L

B

U S I N E S S

S

C H O O L

JÖNKÖPING UNIVERSITY

H o w t h e P r i c e o f C r u d e O i l A f

-f e c ts t h e S w e d i s h St o c k M a r k e t

Paper within Economics

Author: Sean Winstanley (810730-4696) Gustaf Hamilton (801206-0052) Tutors: Per-Olof Bjuggren

(2)

Bachelor Thesis within Economics

Title:

How the Price of Crude Oil Affects the Swedish Stock Market

Authors:

Gustaf Hamilton Sean Winstanley

Tutors:

Per-Olof Bjuggren Johan Eklund

Date:

2007-01-23

Keywords:

Oil Price, Arbitrage Pricing Theory, Macroeconomic Factors, Sweden

Abstract

In late summer 2006 we experienced historically high oil prices, and due to this event we found it appropriate to investigate what influence oil price changes has on the Swedish stock market. The purpose with our research was to see the affect that oil price changes has on the Swedish economy, and if the influence of the oil price is still as strong as it used to be. To help us draw conclusions we have applied the Arbitrage Pricing Theory. With use of statistical analysis we have been able to examine the relation between oil prices and other macroeconomic variables, and how these affect the Affärsvärlden Generalindex. Our re-sults show that oil has a significant influence, our regression analysis show that a 1 unit in-crease in the oil price results in a 0.08 unit dein-crease in Affärsvärldens Generalindex. Our study has also given us indications that the oil price effect on the Swedish economy has de-creased since the mid 1980´s. We can also draw conclusions that since the 1970´s, society has moved from heavy oil dependency towards a more diversified usage of energy sources. The results for Sweden are in line with the influence of oil has on other world economies.

(3)

Kandidatuppsats inom Nationalekonomi

Titel: Hur Råoljepriset påverkar den Svenska Aktiemarknaden Författare: Gustaf Hamilton

Sean Winstanley Handledare: Per-Olof Bjuggren

Johan Eklund

Datum: 2007-01-23

Nyckelord: Oljepris, Arbitrage Pricing Theory, Makroekonomiska Faktorer, Sverige.

Sammanfattning

Under sensommaren 2006 erfarde vi historiskt höga oljepriser. Med denna händelse som grund fann vi det relevant att undersöka oljans påverkan på den svenska ekonomin. Syftet med denna uppsats var att se hur skillnader i oljepriset påverkar Sveriges ekonomi och om oljan fortfarande har en lika stark påverkan som tidigare. Som verktyg för att påvisa detta har vi använt oss av ”Arbitrage Pricing Theory”. Med hjälp av statistisk analys har vi kunnat se påverkan av oljeprisfluktuationer och andra makroekonomiska variablers påverkan på ekonomin. Affärsvärldens Generalindex har använts som definition av ekonomin. Våra resultat visar att oljan har en signifikant påverkan på svensk ekonomi, en 1 enheters uppgång av oljepriset resulterar i en minskning med 0,08 enheter på Affärsvärldens Generalindex. Vår studie ger även indikationer att oljeprisets påverkan har minskat sedan mitten av 1980-talet. Vi kan också utläsa att samhället har skiftat från ett tungt oljeberoende i energiförbrukning mot mer diversifierade typer av energikällor, detta sedan 1970-talet. Resultaten visar även att Sveriges relation till olja är i linje med andra världsekonomier.

(4)

Table of Contents

1 Introduction... 1

1.1 Problem ...2 1.2 Purpose ...2 1.3 Method ...2

2 Background... 3

2.1 The History of Oil Price Shocks ...3

2.2 Oil’s Importance to Sweden...4

2.3 Limitations ...5

2.4 Previous Research ...5

2.5 Outline ...6

3 Theory, Macroeconomic Variables and Alternative

Multifactor Models Theory ... 7

3.1 Capital Asset Pricing Model...7

3.2 Arbitrage Pricing Theory...8

3.3 Macroeconomic Variables ...9

3.3.1 Affärsvärldens Generalindex (AFGX) ...9

3.3.2 Gross Domestic Product...9

3.3.3 Consumer Price Index ...9

3.3.4 Morgan Stanley Capital Pricing Index...10

3.3.4 OPEC Reference Basket...10

3.4 Lags...10

3.5 Alternative Multifactor Models ...10

3.5.1 Fama-French Three Factor Model...10

3.5.2 Multibeta CAPM ...10

4 Results and Analysis ... 11

4.1 Result Correlation Matrix ...11

4.2 Regression Models...12

4.2.1 Regression I ...12

4.2.2 Regression II ...13

4.2.3 Regression III ...14

4.3 Oil Price Change in Relation to AFGX...16

5 Conclusion ... 17

6 Suggested Further Research ... 18

References ... 19

Appendix 1 ... 22

Appendix 2 ... 26

(5)

Graphs

Graph 2.1………3 Graph 4.1………..16 Graph 4.2………..16

Tables

Table 2.1……….5 Table 4.1………..………...…...11 Table 4.2………...12 Table 4.3………...12 Table 4.4………...13 Table 4.5………...14 Table 4.6………...14

Diagrams

Diagram 2.1………... 4 Diagram 3.1………7

Equations

Equation 3.1………... 7 Equation 3.2………... 8 Equation 3.3………... 9

Appendixes

Appendix 1………...22 Appendix 2………...26 Appendix 3………...27

(6)

1 Introduction

Our choice of subject for our bachelor thesis in economics has fallen to the price of crude oil, and how changes of the price affect the Stockholm Stock Exchange. We find oil a very interesting and up to date topic because it is discussed almost daily in the financial press. Crude oil is one of the most important commodities in our society, and today’s economies are very dependent on oil.

What makes oil such a competitive energy source amongst other aspects is that it is easy to transport and store, it also has a very high concentration of energy which other sources can not match. Although we have increased our barrel per day consumption over time, the fact is that we have a more energy efficient society today. Oil related products have increased in number, but they are more efficient when it comes to oil consumption today that earlier. The manufacturing industry and transport are two industries that rely heavily on oil. To-day’s largest consumer of oil is the United States, they stand for about 25% of the world’s consumption (The Swedish Petroleum Institute (SPI)).

Oil is found in various places in the world, mainly in areas that earlier existed as ocean. Ar-eas where oil is extracted include the Middle East, USA, Canada, Mexico and North West Siberia. The Organization of Petroleum Exporting Countries (OPEC) has naturally had a major influence on the price of oil since their foundation in 1960 (OPEC). They control a large number of the output and 40% of the world’s oil is produced by the OPEC-countries (SPI). Today, there are 11 member states in OPEC and they together extract and export six different crude oils. In Western Europe the price of oil is related to the Brent oil, which is extracted in the North Sea. The Brent is considered a “marker” crude, crude with better quality will have a higher price than the Brent, and vice versa (SPI).

Oil is just as any asset bought and sold on the financial markets, in fact oil is the most traded asset in the world. Important meeting places for trading oil is the International Pe-troleum Exchange in London and the New York Mercantile Exchange. In Sweden the price of oil is determined by factors such as the $ US exchange rate, the international price in dollars, market conditions and taxes (SPI).

On the Stockholm Stock Exchange which is an automated exchange is today currently trading almost 300 company stocks. There are oil related stocks such as Lundin Petroleum and Vostok Nafta Ltd represented on the list.These stocks are naturally more sensitive to variations in the oil price than other stocks that are not dependent on oil. In our study, we are looking at the Swedish market as a whole using Affärsvärldens Generalindex (AFGX) as our measurement for the Swedish stock market. The AFGX includes all stocks traded in Stockholm.

Lately we have seen fluctuations in the oil price. The price of the Brent reached almost $80 per barrel in the summer of 2006 and some oil analysts predicted that a price of almost $100 per barrel was in reach. These events and other recent oil price fluctuations do not seem to have as great impact on the Swedish stock market as events such as the oil crisis in the 1970’s. To get an up to date picture of the oil price just as this thesis is about to be completed in January 2007 we can see that the price of oil is currently trading at around $53-54 per barrel (Dagens Industri). The sudden decrease in the price has erupted due to an unusual mild winter and the high oil reserves in the USA.

(7)

1.1

Problem

The debate regarding that society is facing an oil shortage in the future has opened up the discussion for alternative energy sources once again, although the question arose already af-ter the first oil crisis in 1972-73. Recently we have experienced very high prices of oil due to political unrest and natural disasters. The environmental threat that recently has been a hot topic in the Swedish press is another factor showing that maybe our consumption of oil ought to be decreased and substituted for more environmentally friendly energy sources. In this study the writers’ want to show the reader how oil price changes affect the Stock-holm Stock Exchange. When the writers refer to The StockStock-holm Stock Exchange or the Swedish economy/stock market the definition is always Affärsvärldens Generalindex (AFGX), which represents all stocks traded in Stockholm.

The writers present the following research questions:

 How large is the impact of oil price changes on the AFGX?

 Has the AFGX become less sensitive to oil price changes since the beginning of the 1980’s?

1.2

Purpose

The purpose with this paper is to examine the affect that oil price changes has on the Stockholm Stock Exchange, where the stock exchange is represented by AFGX.

1.3

Method

In our study we have restricted ourselves to only the Swedish economy. To help us draw conclusions we have used the Arbitrage Pricing Theory (APT) to describe the importance of oil to the Swedish economy. To give our model relevance we have included further mac-roeconomic factors that will be explained in depth in a later part of this thesis. We have collected secondary data that reaches from January 1985 to September 2006. Previous re-search and scientific articles in the area has been a way of gaining knowledge about our subject. Following the daily discussion of oil in the financial press has also been a way to gain information and knowledge about our topic.

(8)

2 Background

2.1

The History of Oil Price Shocks

Since World War II we have experienced three major oil price shocks that affected the fi-nancial markets around the world strongly. In 1973-74, what became to be known as the first oil crisis erupted. The OPEC countries reduced their export of oil to countries sup-porting Israel in the Arabic-Israeli War. They also decided to drastically increase the price. Over a four month period the price of crude oil had increased by more than 250% (Na-tionalencyklopedin). The result of this was a move from a booming economy into a reces-sion for most countries in the Western World. The second oil crisis hit the Western economies in 1979. This was a product of unsettlement in the Middle East due to the Ira-nian Revolution and later the Iran-Iraq war. After these two oil crises plans for alternative fuels was a question that rose in people’s mind.

In 1990 the Gulf War erupted. The USA led invasion of Iraq caused the price of oil again to increase drastically as can be seen in graph 2.1. The price of the OPEC Reference Basket increases drastically, from $15-$34 per barrel, this implies an increase of the price by over 120%.

In recent years there have been historically high oil prices. In 2004 the oil price in nominal terms was very high, though as they in real terms have not been as high as earlier oil price shocks. Events such as the hurricane Katrina in 2005, and again unsettlement in the Middle East with the US led war on terror in Iraq has had an influence on the price of oil. The price for the OPEC-basket was fluctuating around $ 70 (US) per barrel in the summer of 2006. Since 1999 we have experienced a steady increase in the price. The below graph indi-cates the fluctuations of the oil price during the time period that we examine in our paper. What one can see is a peak when the Gulf crisis occurred and that there is a sharp drop in the price in 1985-86. This decrease occurred because of OPEC put forward a new pricing scheme that resulted in a decline of the oil price.

Graph 2.1 – Historical Oil Prices (1980-2006)

OPEC Reference Basket

0 10 20 30 40 50 60 70 80 ja n -8 0 ja n -8 2 ja n -8 4 ja n -8 6 ja n -8 8 ja n -9 0 ja n -9 2 ja n -9 4 ja n -9 6 ja n -9 8 ja n -0 0 ja n -0 2 ja n -0 4 ja n -0 6 1980-2006 P ri c e /b a rr e l $ Oil Price Source: Ecowin 2006

(9)

2.2

Oil’s Importance to Sweden

Since the 1970’s, the import of oil products into Sweden have decreased by approximately 50%. Like other Western countries, Sweden is a big consumer of energy per capita. Today, one third off all energy consumption in Sweden is based on oil (SPI). To get a perspective, it is interesting to know that in 1973 the share of world energy consumption that came from fossil fuels was represented by 97 % (Radetzki, 1995). In 2001 that figure was 80%, this implies a reduction of almost 20% (SPI).

Important to stress is the difference between oil price shocks from external events like war, and from cyclical movements in the oil price. These two events have different effects on the market. A moderate increase of the oil price is something Sweden can cope with, but a more drastic oil price shock hits the economy harder (see Graph 4.1 and Graph 4.2). The consumption of oil products can be split into four main categories; Industry, Transport and Industrial Machinery, Houses and Service, Production of district heating and electricity. In the industry, oil is mainly used in the production and heating of facilities. In transport, oil is mainly used as fuel for vehicles, and represents 22 % of total energy consumption, but oil is also used for air and sea transports. Since the 1970-80’s the oil consumption in this cate-gory has decreased much, due to increased taxes. But on the contradictory, machines in the agricultural industry still are dependent on oil related products. To further understand the oil “situation” in Sweden diagram 2.1 is presented to describe the shares of energy sources in Sweden 2001 (SPI).

Diagram 2.1 – Energy types in Sweden (2001)

Energy Types in Sweden

34% 16% 31% 4% 2% 13% Nueclear Power Biological Fuel Oil Coal Gas

Water and Wind Power

Source: SPI 2006

Oil related products only represent 37% of energy production in Sweden (2001) as can be seen in the above diagram. Comparing figures from the 1970´s and 2001 can give a fair de-scription of the development on the Swedish energy market. Since the 1970´s our energy consumption has increased over all. The share of oil consumption has decreased with al-most half, contradictory, both coal and gas has increased but only with a small margin. Nu-clear power has increased with over 200 % and as can be seen above nuNu-clear power has a large share of our energy consumption today. We also have an approximate increase of 100

(10)

% of the biological fuels, this increase goes just as well for the development of energy from water and wind (Radetzki, 2004).

Table 2.1 - Oil Dependence in 1980 and 2004

1980 2004

OECD Sweden OECD Sweden

Oil price, $/bl 36.8 36.8 38.3 38.3

Oil imp/tot exp, % 24.0 24.7 6.9 3.5

Oil imp/GDP, % 4.0 6.0 1.3 1.2

Source :Marian Radetzki 2006

In table 2.1 it is interesting to see that oil import as a percentage of total expenditure has become less, both in Sweden and in the OECD-countries. This gives us an indication that oil dependency in Sweden has decreased since the 1980’s. Just as in total expenditure, the proportion of oil in GDP has become less. This is true for both the OECD-countries and Sweden. This information was supplied to us through e-mail correspondences with one of Sweden’s leading energy experts, namely Marian Radetszki.

2.3

Limitations

Due to the fact that historical data has been limited in certain areas, time series have not been available to the extent that we originally have hoped for. This has resulted in fewer observations in the writers’ statistical analysis than what would have been ideal. We have limited ourselves to the affect of oil price changes on only the Swedish stock market and we have not compared Sweden with other national stock markets. Though, the writers’ have included the MSCI World Index in the APT to get some perspective of the Swedish stock markets sensitivity to the oil price in comparison with the rest of the developed world. In an extended research, the Arbitrage Pricing Theory can be used to its full capac-ity, data can then be applied for different countries and markets, thus giving a more de-tailed comparison.

2.4

Previous Research

Previous research done in our area of study is rather scarce, though some studies are made in related areas with examples of other economies. With this study we hope to contribute to the field of oil’s impact on the Swedish economy. In Sweden, one of the leading profes-sors in the area of energy studies is Professor Marian Radetzki at Luleå University, he has written a great deal about Sweden’s relation to oil. He did supply us with some evidence that Sweden’s dependency of oil has become less. But his material does not give us any support that oil prices have less impact on the Stockholm Stock Exchange today than ear-lier periods.

As a matter of fact we have not been able to find any studies written on the impact of oil on only the Stockholm Stock Exchange, though there is an interesting master thesis from Lund University by Bondesson and Hagströmmer (2005). They investigate the oil price sensitivity on the Swedish and European Union’s stock markets. Their conclusions were that there are some co-integration relationship between oil price change and sub-indices on the AFGX such as the industry, chemical and energy indices.

(11)

Roll and Ross discussed in 1980 that oil price risk is not separately rewarded in the stock market. They test if oil can be seen as a systematic risk factor which it often is in reality. Their conclusion was that oil price change had no overall change on asset pricing.

Hamilton (1983) discusses the US economy’s poor performance since World War II and how oil affects the economy. His findings show that the US economy is affected by oil price shocks and that prior to 1972 oil has contributed to almost all of the US recessions. He also found a strong correlation between change in oil price and GNP growth. Hamil-ton‘s macroeconomic included in the APT were: GNP growth, money supply, unemploy-ment rate, wage inflation and import price inflation. Mörk (1989) extended Hamilton’s work to include not only the affect of the increases in the oil price, but also decreases which was included when the time period was extended up to 1988. Mörk came to the conclusion with this extended research, that oil price shocks only have marginal effects on GNP.

Chen, Roll and Ross (1986) used the APT in an investigation on the US stock market in their paper Economic Forces and the Stock Market. They came to the conclusion that oil price risk was not separately rewarded in the stock market. The macroeconomic variables that were identified in this research paper were: the spread between long and short term in-terest rate, expected and unexpected inflation, industrial production and the spread be-tween high and low grade bonds, they also included the oil price.

Beenstock and Chan (1988) took help from the APT on the London Stock Exchange, their choice of variables to include in the APT fell to inflation, money supply, price of raw mate-rials and interest rate. They did test other macroeconomic variables but it was only the named ones that had a significant effect on the return on the stocks on the British market. The time span they tested ranged from 1977-1983.

More recently Yoon (2004) argues that if oil price has an impact on the macro economy, then it should also affect the stock markets. His work gives us a good review of the link be-tween oil price and the macro economy, and once again it shows that oil has an affect on different economies around the world.

2.5

Outline

In the opening phase of this paper the writers’ have given the reader an introduction to our topic and presented our problem, purpose and method. In this section the writers’ also presented the reader with a brief history and the importance of oil to the Swedish econ-omy. To end the opening part of our thesis previous research in the area is presented. The second part of the paper gives an explanation to our theory. The APT model will be pre-sented and our statistical approach will be put forward and explained. Here, all macroeco-nomic factors affecting the outcome will be presented. In the empirical part all our results will be thoroughly presented which will give the reader a clear view of our work. Finally, we will present to the reader our results and conclusions, we will also name potential areas for suggested further research.

(12)

3 Theory, Macroeconomic Variables and Alternative

Multifactor Models Theory

Analysing the relationship between a macroeconomic variable and the payoff of an asset can be conducted in various ways. One of the most recognized is with the Capital Asset Pricing Model (CAPM) (Bailey, 2005).

3.1

Capital Asset Pricing Model

The CAPM describes the relationship between risk and payoff of an asset. It is a liner rela-tionship which indicates that you can not increase payoff without increasing the risk. The risk of an asset in the CAPM is defined by the beta (β). The beta risk shows how the asset correlates with the market, this risk is systematic, and thus it can not be diversified away. The non systematic risk is approximately zero since it has been diversified according to the theory of the CAPM. The model is based on three main assumptions:

 Asset markets are in equilibrium

 Mean-variance portfolio selection (perfect competition)  Homogeneous beliefs (Bailey, 2005)

Diagram 3.1 The CAPM

Source: Bailey, 2005

The market portfolio has e defined beta equal to one. This indicates that correlation be-tween market and asset are exact. With a higher beta the asset changes more than the mar-ket and with a lower beta the payoff of the asset will change less than the marmar-ket. An asset that lies above the security market line is according to the theory underpriced and when an asset is below the security market line it is overpriced (Bailey, 2005). The connection be-tween expected payoff and the systematic risk is shown in the CAPM formula:

Equation 3.1 rirf (rmrf)j

ri = Expected payoff for the asset

rf= Risk free rate

β = Systematic risk

rm = Expected payoff of the market portfolio

(13)

The expected payoff is then expected payoff minus the risk free rate, multiplied with the beta. A common way of measuring the beta is through a regression analysis but it can also be done analytically. In a regression analysis, historical data is used to estimate the expected beta of the variable (Bailey, 2005).

3.2

Arbitrage Pricing Theory

The APT was developed by the economist Stephen Ross in 1976. It is an alternative, or rather an extension of the CAPM. Instead of showing the relationship between risk and expected payoff of an asset with one factor, it shows the relationship of expected payoff of an asset from the influence of several factors (Bailey, 2005) (Clare, Priestley & Thomas, 1997).

The APT is a two step model. The first step as a descriptive model, an extension of the CAPM, and describes what the returns should be in efficient market. In the second step, is the theory of arbitrage, here only the betas should be priced. This step is not for any use to the writers’ since we are not interested in the arbitrage opportunity, thus we are satisfied at the first step of the model in our analysis. The first step of the APT can be defined as a liner regression equation:

Equation 3.2 Ra11 22  nn 

R = the rate of return on the security over a given time interval. a = a constant term, the risk free rate.

βn= sensitivity coefficient the security to each factor.

δn= return of a factor (percentage variation).

n= number of factors.

ε = random term to the specific security.

In the APT every factor has its own beta to show the unique sensitivity to changes in that particular factor and the return of the security. In a financial perspective it opens up to a more detailed and custom made approach to portfolio risk management than the CAPM. This can be explained through the increasing usage of more widespread derivatives instru-ments. (Shanken, 1982)

To state the main differences between the CAPM and APT gives a clearer view why the writers’ choose the latter one. In the APT, portfolios are diversified and the unsystematic risk is approximately zero, this also goes for the CAPM. In the latter one, there is only one factor influence and no arbitrage assumptions are made. In equilibrium the CAPM shows the market portfolio that every investor wants to hold and shows the value of the asset in relation to its systematic risk. The APT instead relies upon the absence of free arbitrage opportunities since the theory says that there are no arbitrage assumptions at the market during a long time, since all investors would seek this opportunity (Bailey, 2005)

(14)

3.3

Macroeconomic Variables

Below we present the factors that will be included in our model. We have chosen the fac-tors mainly in the presence of availability and the coherence of data. We have worked with several other macroeconomic factors than those presented in the next section, including money supply, exchange rates, interest rates and industrial production index. These factors have not been included because they have not had a significant influence in our model and there has been a lack of historical data. When studying which factors other researchers have chosen the results have been very different. The APT does not exactly tell which factors to use, it is up to the researchers to include the ones that they believe are relevant. As the APT model considers changes in macroeconomic variables we have chosen to take the percent-age change in each factors according to the below formula:

Equation 3.3 (Pt+1-Pt)/Pt

3.3.1 Affärsvärldens Generalindex (AFGX)

The AFGX is the most commonly used index for financial investigations in Sweden and was often used in the Swedish reports that we have read to gain knowledge for our re-search. We have chosen this index since this is the only index that reflects the Swedish Stock Market back to 1985. It gives us a wide appropriate description for the Swedish economy since it represents all the stocks traded on the Stockholm Stock Exchange (Af-färsvärlden).

3.3.2 Gross Domestic Product

GDP is a measure of how much that is produced in a country. Or, more exactly, the value of all end products and services produced in a country during a certain time period, usually one year. GDP is calculated on a quarterly basis and is based on a large range of primary statistical resources. The quality of GDP is directly connected with these sources. The cal-culations are made according to EU standards and have to be correct, include as much as possible and be comparable. GDP is an important factor to use in comparison with the oil price since many products are directly dependent on oil products (SCB).

3.3.3 Consumer Price Index

The CPI is used to measure the average development of prices of the entire private domes-tic consumption, the actual price the consumer pays. CPI is used to calculate compensation and inflation calculations in Sweden. CPI has been calculated since July 1954 on a monthly basis. CPI is used by the department of finance, the Swedish central bank, different organi-sations and companies to calculate indexes and future levels of rates etc. CPI is calculated of price changes of the private consumption on a monthly basis (Statistics Sweden (SCB)). Information of prices are collected by Statistics Sweden through interviews to businesses of several different commodities and services, called “represented commodities”. The sys-tem of gathering prices is divided in three different ways. CPI is constructed as a chain in-dex with links on a yearly basis. Every link measure how much the average price level is during the actual year has changed from the average price level from recent year. Weighted figures represent the geometric average value of the two affected year’s consumption vol-ume. Since companies are affected by the oil price when producing goods, the price of goods can be interesting comparing with oil price fluctuations (Ecowin).

(15)

3.3.4 Morgan Stanley Capital Pricing Index

The MSCI-World is an index measuring the global developed market equity performance for 23 countries, or 1400 stocks listed on exchanges in the US, Europe, Canada, Australia, New Zealand and the Far East. We have chosen this index because it will be interesting to make a comparison between the AFGX and the world markets sensitivity for oil price fluc-tuations (Ecowin).

3.3.4 OPEC Reference Basket

The OPEC Reference Basket average prices are based upon seven crude oils expressed in US dollars. The different crude oil’s are: Algeria Saharan Blend, Indonesia Minas, Nigeria Bonny Light, Saudi Arabia Arab Light, UAE Dubai, Venezuela Tijuana Light and (non-OPEC) Mexico Isthmus. We chose this index since it would give the most general price of oil. The time span has been from January 1985- September 2006 (Ecowin). We preferred the OPEC basket in front of the Brent and other oil’s because of the availability of data. Since the APT relies on changes in factors, it has of little importance which oil to use, dif-ferent oil’s move together when it comes to changes in price.

3.4

Lags

To get the most appropriate fit of the regression model lags are implemented. Because it can take time for a dependent variable (AFGX) to adjust to changes in the independent variables you may lag the variables. In our case GDP was lagged two periods. By this we mean that it takes two periods for the AFGX to adapt to changes in GDP. The other three independent variables in our model we have chosen to leave without lags.

3.5

Alternative Multifactor Models

The APT is naturally not the only multifactor model that can be used in a similar investiga-tion. In the below section two alternative models that can be used in place of the APT will be explained briefly.

3.5.1 Fama-French Three Factor Model

This model was developed to describe market behaviour. The model states that size and value of stocks is the most significant factor outside of systematic risk that explains realised returns on stocks. These three factors describe the total risk on an asset. Fama and French also claimed that small cap stocks and stocks with high BV/PV outperformed the market as a whole. Since we include different variables and more than three, we decided that the Fama-French Model was not appropriate for our study (Index Fund Advisors).

3.5.2 Multibeta CAPM

Since the traditional CAPM only considers systematic risk the multibeta CAPM was devel-oped. In this model extra risk factors are added so it is no longer just the systematic risk de-fining the level of stock returns. What these extra risk factors are exactly is not explicitly defined, and it is up to the researchers to include which ones they find appropriate. Since we want to examine how the oil price and other macroeconomic factors affect the Swedish stock market the APT is chosen in front of the multibeta CAPM.

(16)

4 Results and Analysis

4.1

Result Correlation Matrix

The outcome from our correlation matrix can be seen in table 4.1. We included all our macro economic variables that we used in the APT to get a clear view of the correlation amongst the different variables, and thereby helping us to draw comparisons among the different variables.

Table 4.1 – Correlations (1985-2006)

AFGX MSCI OIL CPI GDP

AFGX Pearson Correlation 1 ,607 -,215 -,106 ,078

Sig. (2-tailed) ,000 ,001 ,089 ,212

MSCI Pearson Correlation ,607 1 -,154 -,028 -,044

Sig. (2-tailed) ,000 ,013 ,658 ,483

OIL Pearson Correlation -,215 -,154 1 ,065 -,040

Sig. (2-tailed) ,001 ,013 ,297 ,523

CPI Pearson Correlation -,106 -,028 ,065 1 -,028

Sig. (2-tailed) ,089 ,658 ,297 ,650

GDP Pearson Correlation ,078 -,044 -,040 -,028 1 Sig. (2-tailed) ,212 ,483 ,523 ,650

To begin with, we can see that the MSCI has a strong correlation with the AFGX. This can be explained by the Swedish stock markets strong relationship with other developed econo-mies. The outcome supports the statement that economies do tend to move in the same di-rection when it comes to cycles in the economy. We can also state that we are experiencing an increase in globalization.

Beforehand, we did believe that oil prices would have an impact on the Swedish stock mar-ket, and by looking at the correlation we can see that this is supported, i.e., as the price of oil increases the effect on the Swedish financial market is negative. We can also see that consumer prices do have a negative correlation with the AFGX. This is in line with what is generally explained in economic theory. As inflation increases the market reacts negatively (Varian, 2003).

The GDP has a very low correlation to the AFGX, which to the writers is rather surprising. In our correlation matrix GDP does not seem to give any significance or correlation. This can partly be explained that GDP-data is only available on quarterly basis and thus differs in the coherence compared to the other variables.

The correlation between oil and MSCI is weaker than the correlation between oil and AFGX. This means that the Swedish stock market, to more extent is negatively correlated, or less correlated with oil than the World market. One of the explanations to this can be the lack of large oil related companies represented on the AFGX and thereby it can be as-sumed that the influence of oil is not as great on the AFGX, as it is on the World market as a whole.

The relationship between oil and CPI is often closely monitored in economic studies. This is due to the fact that production and transport costs are heavily dependent on the price of oil. An increased price of oil makes it more expensive to produce the goods and services represented in the CPI. We therefore find it appropriate to further investigate the

(17)

correla-tion between oil and CPI and we have run a correlacorrela-tion test with just oil and CPI repre-sented.

Table 4.2 – Correlation between oil and CPI (Jan 1985- Sep 2006) OIL CPI CPI Lag1 CPI Lag2 CPI Lag3 OIL Pearson Correlation 1 ,065 -,058 -,143 -,111

Sig. (2-tailed) ,296 ,351 ,022 ,076

Since it takes time for the consumer prices to adjust to increases in the price of oil, we tested different lags alternatives. In the above table we can see that it takes approximately two months to see the affect from an increased oil price, i.e. if the price of oil goes up in for example January the effect from CPI is shown first in March. Period with two lags gives us the strongest correlation and a significance that qualifies on the 5% level.

Since oil has such an affect on CPI, multicollinearity must be taken into consideration. This can explain why our significance level for CPI in regression model II (table4.4) does not qualify on the 5% level.

4.2

Regression Models

In this part we will present our regression analysis. Altogether we have run three different regressions to help us see what influence oil has on the Swedish stock market and whether the importance of the oil price has changed over time.

4.2.1 Regression I

In this section we will continue our analysis between oil and CPI since we think it is important to see the relationship between these two variables and how they affect the Swedish economy. To do this, four regressions have been run with different lags. We tested different lag alternatives just as we did in the correlation analysis, these were t0 to t -3. This

resulted in t-2 giving the highest significance, in fact it was the only lag that qualified on the

5 % level.

Table 4.3 – Regression analysis Oil and CPI (Jan 1985-Sep 2006)

Variable Coefficient t-statistic Sig.

OIL (Dependent) 0.013 2.044 0.042

CPI lag2 – 2.440 – 2.298 0.022

R-squared 0.020 F-statistic 5.281

Adjusted R-squared 0.016 Prob(F-statistc) 0.022

To give the reader an economic perspective we will present an example explaining the re-sults above. Once again we can see the strong impact of oil, this time it shows on CPI. A 1 unit increase of the price of oil would result in 2.44 units increase in CPI. This gives us further support that the oil price has a significant influence CPI. Though, one should not

(18)

value is very high suggesting that the model does not explain a great deal. But as we men-tioned earlier it still gives us indication that the oil price is a strong factor explaining the level of CPI, i.e. an increased price of oil yields a higher CPI.

4.2.2 Regression II

This section presents the regression with all independent variables included in our chosen time span (January1985-September 2006), and this is in line with the full APT model.

Table 4.4 – Regression analysis AFGX, MSCI, OIL, CPI and GDP (Jan 1985-Sep 2006)

Variable Coefficient t-statistic Sig.

AFGX (Dependent) 0.007 2.066 0.040 MSCI 0.907 11.976 0.000 OIL -0.080 -2.331 0.021 CPI -0.951 -1.630 0.104 GDP 0.051 1.988 0.048 R-squared 0.400 F-statistic 42.110

Adjusted R-squared 0.390 Prob(F-statistc) 0.000

By looking at the beta parameters (coefficients) we can see that an increase in the price of oil gives a negative payoff on the AFGX even though its effect is rather weak, beforehand we actually did think that the impact would be greater. If this relation is explained in nu-merical terms it would imply that a 1 unit increase of the oil price would give a negative 0.08 unit payoff on the AFGX. What can explain this rather low influence is that the changes in oil price is partly accounted for in the CPI, we found signs of this relationship in our first regression (regression I) when looking at oil’s influence on CPI. But still, we can see that oil has a significant influence on the Swedish stock market by examining the beta coefficient for oil.

It is also of interest to look at the other macroeconomic variables respective coefficients, we can see that they show what one can anticipate, namely negative coefficients on oil and CPI and positive coefficients for MSCI and GDP. This gives support to the strength in the model, thus the accuracy of the beta coefficient for oil. The model also explains a great deal because the variables are significant. Although CPI is not significant on the 5% level it still gives a contribution to our model since multicollinearity can be assumed. The impact of CPI on AFGX is very strong, this is also interesting when looking at our earlier regression analysis between oil and CPI. This gives strength to our hypothesis that oil prices have an impact on the Swedish stock market.

(19)

4.2.3 Regression III

To get a broader perspective of the influence of oil prices on the Swedish stock market over time, we divided the data into periods. Due to the lack of observations we limited ourselves to just two periods. In a scenario with more periods and a larger amount of ob-servations one would probably get a clearer and more reliable picture that the Swedish stock markets oil price dependency has decreased over time. By comparing the two periods below (table 4.5 and 4.6) one can see that the Swedish stock market has developed a less dependency to the price of oil over time.

Table 4.5 – Regression analysis period 1 (1985-1995)

Variable Coefficient t-statistic Sig.

AFGX (Dependent) 0.009 1.406 0.162 MSCI 0.707 5.857 0.000 OIL -0.103 -2.007 0.047 CPI -1.020 -1.225 0.223 GDP 0.086 2.407 0.018 R-squared 0.288 F-statistic 12.552

Adjusted R-squared 0.265 Prob(F-statistc) 0.000(a)

In the above table (4.5) we can see that oil is significant in the model on the 5%-level. The beta coefficient for oil tells us that a 1 unit increase in the price of oil would give a 0.103 unit negative effect on the AFGX.

Table 4.6 – Regression analysis Period 2 (1995-2006)

Variable Coefficient t-statistic Sig.

AFGX (Dependent) ,007 1,831 ,070 MSCI 1,115 12,981 ,000 OIL -,050 -1,151 ,252 CPI -1,805 -1,899 ,060 GDP -,034 -,914 ,362 R-squared ,597 F-statistic 46,594

Adjusted R-squared ,584 Prob(F-statistc) ,000(a

In the above table (4.6), the beta coefficient for oil does not show significance on the 5%-level, this the writers find interesting since also gives a result in itself. Since the oil coeffi-cient does not qualify on the 5% significance level the writers’ can not reject the null

(20)

hy-pothesis that oil does not have an impact on the Swedish Stock Market ( Ho: Beta=0). This gives the writers’ evidence that he oil price does not have as much of an impact in pe-riod 2 compared to pepe-riod 1.

The beta coefficient for oil does indicate that a 1 unit increase in the price oil would lead to a 0.05 decrease in the AFGX, but since this is significant only on the 25,2% level too strong conclusions should not be made.

We can see that in the first period the beta parameter for oil is significant and that it is not significant in period 2. When looking at the beta parameter of oil in period 2, we can see that the beta value is reduced by half when comparing it to period 1. The reduced beta value tells us that the influence of oil on the AFGX has decreased, and by this our hy-pothesis that the influence of oil on the Swedish economy has decreases over time is fur-ther supported.

We can also see that there is a big difference in the t-statistic and R-square between the two periods. The R-square in period two is much higher implying that the model explains more, the independent coefficients stand for 59,7% of the factors affecting AFGX. This also gives a suggestion that the effect of oil prices on the Swedish stock market has decreased. An explanation of Sweden’s lowered dependency of oil can be linked Table 1.1. As we know, the consumption of oil has decreased since the 1970´s and these results further en-hance that Sweden’s oil dependency has decreased. What we can bring with us from this analysis is not only that the consumption of oil has been lowered, but also that the overall affect of oil on the Swedish economy has moved in line with the lowered consumption. This is clearly seen when comparing the results from our regressions and Table 1.1.

(21)

4.3

Oil Price Change in Relation to AFGX

To further support the findings from regression II the writers’ will show graphically how oil price changes have fluctuated, and how the AFGX has responded to these changes since 1980. As one can see in the below graphs 4.1 and 4.2, the AFGX seem to follow the development of the oil price. This gives us a clear picture that oil has an impact on the Swedish stock market.

Graph 4.1 – Percentage change in Oil and AFGX (1980-1993)

Percent change in Oil and AFGX

-0,3 -0,2 -0,1 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 fe b-80 fe b-81 fe b-82 fe b-83 fe b-84 fe b-85 fe b-86 fe b-87 fe b-88 fe b-89 fe b-90 fe b-91 fe b-92 fe b-93 1980-1993 % C ha ng e

AFGX% Oil price%

Source: Ecowin 2006

Graph 4.2– Percentage change in Oil and AFGX (1993-2006)

Percent change in Oil and AFGX

-0,3 -0,2 -0,1 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 ju l-9 3 ju l-9 4 ju l-9 5 ju l-9 6 ju l-9 7 ju l-9 8 ju l-9 9 ju l-0 0 ju l-0 1 ju l-0 2 ju l-0 3 ju l-0 4 ju l-0 5 ju l-0 6 1993-2006 % C h an g e

AFGX% Oil price%

Source: Ecowin 2006

Here, one can see that in general the oil price and the AFGX move in opposite direction, i.e. an increased oil price affects the Swedish economy negatively. It is interesting to men-tion an oil price shock, as one can see from graph 4.1 it is clear that the oil price went up drastically during the Gulf Crisis in 1990-91 and that the AFGX responded negatively to this price increase. Graph 4.1 and 4.2 also indicates that a progressive change of the oil price does not seem to affect the AFGX as much, i.e. the changes in the price of oil do not

(22)

5 Conclusion

The Oil price still has a significant influence on the Swedish stock market. We have shown in our research that a 1 unit increase in the price of oil leads to a 0.08 decrease of the AFGX (which is the writers’ definition of the Swedish stock market). Also, the correlation between oil and the AFGX index is strong, this clearly gives us further evidence that oil has an impact on the Swedish stock market. But on the contrary, we have learnt during our research that oil does not have as much of an impact today that it had in the 1980’s, i.e. the impact of oil on the Swedish stock market has decreased over time.

The relation between the oil price and CPI is of strong importance, since approximately 30% of the production of goods and services are dependent of oil it is obvious that fluctuations in the price of oil will affect CPI. From our outcome from the regressions we can see that CPI has an affect on AFGX. Here, oil price change is partly included for in the CPI.

The impact of oil on the Swedish stock market is clear, but we did beforehand think it would have had a greater impact. Since we use Affärsvärlden Generalindex to describe the Swedish stock market one of the reasons for the weak impact can be explained by the lack of large oil related stocks traded on the Stockholm Stock Exchange. Since the usage of oil has decreased today compared to the 1980’s it gives us suggestion that this decrease shows that the economy is not relying as heavily on oil today as it did earlier.

What further supports that the Swedish stock market is less dependent on oil today than in the 1980’s is by examining our models and tables presented throughout this paper. Swedish economy’s weakened oil dependency is also supported from our analysis of the Arbitrage Pricing Theory in relation to the results from both period 1 and period 2 in regression model III. Our findings are also further supported from the literature we have read with evidence of other researchers.

The strong correlation between MSCI and AFGX is of interest, from this fact we can draw valuable conclusions, one of them is that it gives support that we are experiencing an in-crease in globalization.

From graphs 4.1 and 4.2 we can see that changes in the price of oil does not seem to have as much of an impact on the AFGX as in the period 1. When comparing the results from regression II with graphs 4.1 and 4.2 it is obvious that oil price changes does not affect the Swedish stock market as much today as it did in the 1980’s. The Swedish stock market has learnt to cope with fluctuations in the price of oil and the market has created a more controlled relation to oil.

To sum this thesis up it is appropriate to answer the research questions presented in the problem part of the this paper .It can be said that the oil price still has an influence on the Swedish stock market, but the importance has decreased over time.

(23)

6 Suggested Further Research

What can be done in a more in depth study is to examine different sub-indices such as industry, medicals, finance and commodities that lie under the AFGX. It would be interesting to see the impact of oil price changes and how these different indices are affected. By doing this the APT can be developed to its full capacity and we could probably clearer see oil’s impact on the economy, and that oil probably has less impact today than before.

What also could be suggested as further research is comparing Sweden to other economies. Maybe a study could be made on Scandinavia, this would be interesting since Norway extract and export a lot of oil from the North Sea and the Stock Exchange in Oslo (Oslo Börs) trades oil intensive stocks such as Statoil and Norsk Hydro. Taking this even further one could also include Denmark and Finland and thereby get a broader Scandinavian perspective.

Another area that could be investigated in a further research is to compare oil with other energy alternatives. Today, to give an example, uranium is a hot topic and it would be of in-terest to see the development of uranium and its influence on a specific country or stock exchange. Biological fuels are another interesting and up-to-date energy source that one could use in the APT as a substitute to crude oil.

(24)

References

Literature

Bailey, R. E. (2005). The Economics of Financial Markets. First Edition. Cambridge University Press. Cambridge p.183-222

Bondesson, M., & Hagströmmer, B. (2005). Oljepriskänslighet på Sveriges och EU:s Aktiemarknader. Magisteruppsats Lund Universitet.

Chan, K.F., & Beenstock, M., (1988). Economic Forces in the London Stock Market.

Ox-ford Bulletin of Economics and Statistics, 50(1), 27-39.

Chen, N. F., Roll, R., & Ross, S. A.., (1986). Economic Forces and the Stock Market.

Jour-nal of Business, 59(3), 383-403.

Clare, A., Priestley, R, & Thomas, S. (1997). The robustness to the apt to alternative esti-mators. The Journal of business finance & Accounting, 24 (5), 0306-686X.

Elton, E. J. Gruber, M, J. (1995), Modern Portfolio Theory and Investment Analysis, fifth edition. John Wiley & Sons, Inc. New York.

Gujarati, D.N. (2003). Basic Econometrics. Fourth Edition. McGraw-Hill Higher Education. Hamilton, J.D., (1983). Oil and the Macroeconomy since World War II. Journal of Political

Economy, 91(2), 228-248.

Mörk, K.A., (1989). Oil And the Macroeconomic When Prices Go Up and Down: An Ex-tension of Hamilton’s Work. Journal of Political Economy, 97(3), 740-744.

Radetzki, M. (2004). Svensk Energipolitik Under Tre Decennier. SNS Förlag. Stockholm. p.88. Radetzki, M. (1995). Tjugo År Efter Oljekrisen. SNS Förlag. Stockholm. p. 9-15.

Roll, R., & Ross, S. A.., (1980). An Empirical Investigation of the Arbitrage Pricing Theory.

Journal of Finance, 35(5), 1073-1103.

Shanken, J. (1982). The Arbitrage Pricing Theory: Is it testable?. The Journal of Finance, (5) 1129-1140.

Varian, H. R., (2002), Intermediate Microeconomics, A Modern Approach, sixth edition. W. W. Norton & Company, Inc, New York

Yoon, K; (2004). Oil Price Shocks and the Economy. Retrieved 08-26, Retrived 2006-08.25. From: http://web.missouri.edu/~econprm/ec413f04/kyoon_ls.pdf

(25)

Internet sources

Affärsvärlden, (AFGX). Webpage. Retrieved 2006-09-06 & 2007-01-10 http://www.affarsvarlden.se/

http://bors.affarsvarlden.se/details/details.aspx?id=AFV00

http://bors.affarsvarlden.se/aboutafgx.aspx

Dagens Industri. Webpage. Retrieved 2007-01-20

http://www.di.se/Nyheter/

Graphpad. Multicollinearity in a Multiple Regression. Motulski 2005. Webpage. Retrived 2007-01-12.

http://www.graphpad.com/articles/Muliticollinearity.htm

Index Fund Advisors. Webpage. Retrieved 2006-12-10.

http://www.ifa.com/pdf/FamaFrenchThreeFactor_Tuck2003.pdf

Morgan Stanley Capital Price Index (MSCI). Webpage. Retrieved 2006-09-06 http://www.mscibarra.com/

http://www.mscibarra.com/products/indices/stdindex/performance.jsp http://www.msci.com/stdindex/MSCI_Nov06_STMethod.pdf

Nationalencyklopedin. Webpage. Retrived 2007-03-06

http://www.ne.se.ezproxy.ub.gu.se/jsp/search/article.jsp?i_art_id=275497&i_wor d=oljekrisen

Organization of the Petroleum Exporting Countries (OPEC). Webpage. Retrieved 2006-11-15.

http://www.opec.org/aboutus/

Statistiska Centralbyrån (SCB). Webpage. Retrieved 2006-09-06 http://www.scb.se/

(26)

http://www.scb.se/templates/Product____33769.asp

Svenska Petroleum Institutet (SPI). Webpage. Retrieved 2006-10-07 http://www.spi.se/

http://www.spi.se/produkter.asp?kat=5# (History)

Databases

Ecowin Database. Retrieved 2007-01-06

ew:swe01750 (Sweden, Expenditure Approach, Production Approach, Gross Domestic Product, Total at market prices, Current Prices, SEK)

ew:com2121010 (World, Energy, Oil, OPEC Reference Basket Price, Aver-age, USD)

ew:swe15005 (Sweden, Affärsvärlden, General, Index (AFGX), Price Re-turn, End of Period, SEK)

ew:swe11800 (Sweden, Consumer Prices, By Commodity, All Items, To-tal, Index)

(27)

Appendix 1

Regressions OIL and CPI with lags.

Regression no lag Model Variables En-tered Variables Removed Method 1 Change CPI(a) . Enter a All requested variables entered.

b Dependent Variable: Change in oil

Model Summary

Model R R Square Adjusted R Square Std. Error of the Estimate 1 ,065(a) ,004 ,000 ,092536653828 a Predictors: (Constant), Change CPI

ANOVA(b)

Model SquaresSum of df Mean Square F Sig. Regression ,009 1 ,009 1,108 ,294(a) Residual 2,209 258 ,009

1

Total 2,219 259

a Predictors: (Constant), Change CPI b Dependent Variable: Change in oil

Coefficients(a) Unstandardized

Coeffi-cients

Standardized Co-efficients

Model B Std. Error Beta t Sig.

(Constant) ,004 ,006 ,671 ,503

1

Change CPI 1,125 1,069 ,065 1,053 ,294 a Dependent Variable: Change in oil

(28)

Regression 1 lag Model Variables En-tered

Variables

Removed Method 1 Change CPI

+ 1lag(a) . Enter a All requested variables entered.

b Dependent Variable: Change in Oil

Model Summary

Model R R Square Adjusted R Square Std. Error of the Estimate 1 ,061(a) ,004 ,000 ,092695610401 a Predictors: (Constant), Change CPI + 1lag

ANOVA(b)

Model SquaresSum of df Mean Square F Sig. Regression ,008 1 ,008 ,963 ,327(a) Residual 2,217 258 ,009

1

Total 2,225 259

a Predictors: (Constant), Change CPI + 1lag b Dependent Variable: Change in Oil

Coefficients(a) Unstandardized

Coeffi-cients

Standardized Co-efficients

Model B Std. Error Beta t Sig.

(Constant) ,009 ,006 1,475 ,141

1

Change CPI + 1lag -1,051 1,071 -,061 -,982 ,327 a Dependent Variable: Change in oil

Regression 2 lag

Model Variables En-tered RemovedVariables Method 1 Change CPI

+ 2lag(a) . Enter a All requested variables entered.

(29)

b Dependent Variable: Change in Oil

Model Summary

Model R R Square Adjusted R Square Std. Error of the Estimate 1 ,142(a) ,020 ,016 ,091932248803 a Predictors: (Constant), Change CPI + 2lag

ANOVA(b)

Model SquaresSum of df Mean Square F Sig. Regression ,045 1 ,045 5,281 ,022(a) Residual 2,180 258 ,008

1

Total 2,225 259

a Predictors: (Constant), Change CPI + 2lag b Dependent Variable: Change in Oil

Coefficients(a) Unstandardized

Coeffi-cients Standardized Co-efficients

Model B Std. Error Beta t Sig.

(Constant) ,013 ,006 2,044 ,042

1

Change CPI + 2lag -2,440 1,062 -,142 -2,298 ,022 a Dependent Variable: Change in oil

Regression 3 lag

Model Variables En-tered RemovedVariables Method 1 Change CPI

+ 3 lag(a) . Enter a All requested variables entered.

b Dependent Variable: Change in Oil

Model Summary

Model R R Square Adjusted R Square Std. Error of the Estimate 1 ,108(a) ,012 ,008 ,0924988073 a Predictors: (Constant), Change CPI + 3 lag

(30)

ANOVA(b)

Model SquaresSum of df Mean Square F Sig. Regression ,026 1 ,026 3,026 ,083(a) Residual 2,199 257 ,009

1

Total 2,225 258

a Predictors: (Constant), Change CPI + 3 lag b Dependent Variable: Change in Oil

Coefficients(a) Unstandardized

Coeffi-cients Standardized Co-efficients

Model B Std. Error Beta t Sig.

(Constant) ,011 ,006 1,808 ,072

1

Change CPI + 3 lag -1,860 1,069 -,108 -1,740 ,083 a Dependent Variable: Change in Soil

(31)

Appendix 2

Regression analysis, AFGX, MSCI, OIL, CPI, GDP

Model Variables En-tered RemovedVariables Method 1 GDP, CPI,

MSCI, OIL(a) . Enter a All requested variables entered.

b Dependent Variable: AFGX

Model Summary

Model R R Square Adjusted R Square Std. Error of the Estimate 1 ,632(a) ,400 ,390 ,050159848762 a Predictors: (Constant), GDP, CPI, MSCI, OIL

ANOVA(b)

Model SquaresSum of df Mean Square F Sig. Regression ,424 4 ,106 42,110 ,000(a)

Residual ,637 253 ,003

1

Total 1,060 257

a Predictors: (Constant), GDP, CPI, MSCI, OIL b Dependent Variable: AFGX

Coefficients(a) Unstandardized

Coeffi-cients Standardized Co-efficients

Model B Std. Error Beta t Sig.

(Constant) ,007 ,004 2,066 ,040 MSCI ,907 ,076 ,591 11,976 ,000 OIL -,080 ,034 -,115 -2,331 ,021 CPI -,951 ,584 -,080 -1,630 ,104 1 GDP ,051 ,026 ,097 1,988 ,048

(32)

Appendix 3

Regression analysis in period 1 and period 2

Regression period 1

Model Variables En-tered RemovedVariables Method 1 GDP, OIL,

CPI, MSCI(a) . Enter a All requested variables entered.

b Dependent Variable: AFGX

Model Summary Model R R Square Adjusted R Square

Std. Error of the Estimate 1 ,537(a) ,288 ,265 ,057755326799 a Predictors: (Constant), GDP, OIL, CPI, MSCI

ANOVA(b)

Model SquaresSum of df Mean Square F Sig. Regression ,167 4 ,042 12,552 ,000(a)

Residual ,414 124 ,003

1

Total ,581 128

a Predictors: (Constant), GDP, OIL, CPI, MSCI b Dependent Variable: AFGX

Coefficients(a) Unstandardized

Coeffi-cients Standardized Co-efficients

Model B Std. Error Beta t Sig.

(Constant) ,009 ,006 1,406 ,162 MSCI ,707 ,121 ,451 5,857 ,000 OIL -,103 ,051 -,154 -2,007 ,047 CPI -1,020 ,833 -,093 -1,225 ,223 1 GDP ,086 ,036 ,183 2,407 ,018

(33)

Regression period 2 Model Variables En-tered

Variables

Removed Method 1 GDP, MSCI,

OIL, CPI(a) . Enter a All requested variables entered.

b Dependent Variable: AFGX

Model Summary

Model R R Square Adjusted R Square Std. Error of the Estimate 1 ,772(a) ,597 ,584 ,039239585533 a Predictors: (Constant), GDP, MSCI, OIL, CPI

ANOVA(b)

Model SquaresSum of df Mean Square F Sig. Regression ,287 4 ,072 46,594 ,000(a)

Residual ,194 126 ,002

1

Total ,481 130

a Predictors: (Constant), GDP, MSCI, OIL, CPI b Dependent Variable: AFGX

Coefficients(a) Unstandardized

Coeffi-cients Standardized Co-efficients

Model B Std. Error Beta t Sig.

(Constant) ,007 ,004 1,831 ,070 MSCI 1,115 ,086 ,743 12,981 ,000 OIL -,050 ,043 -,068 -1,151 ,252 CPI -1,805 ,951 -,114 -1,899 ,060 1 GDP -,034 ,037 -,054 -,914 ,362

Figure

Diagram 2.1 – Energy types in Sweden (2001)
Table 2.1 - Oil Dependence in 1980 and 2004
Table 4.1 – Correlations (1985-2006)
Table 4.3 – Regression analysis Oil and CPI (Jan 1985-Sep 2006)
+3

References

Related documents

The recorded music industry still betted on the success of digital retailers like Apple’s iTunes Store, selling music in the form of single-track downloads (IFPI 2007).. But

46 Konkreta exempel skulle kunna vara främjandeinsatser för affärsänglar/affärsängelnätverk, skapa arenor där aktörer från utbuds- och efterfrågesidan kan mötas eller

Från den teoretiska modellen vet vi att när det finns två budgivare på marknaden, och marknadsandelen för månadens vara ökar, så leder detta till lägre

Furthermore, the Trade balance and Non-oil trade balance, comparing the two countries, the oil price seem to explain a larger proportion of the fluctuations for Norway (around 5%

The result from the granger test shows that oil price changes do affect Household consumption, Disposable income, and the short-term interest rate NIBOR. Since

The Random Walk method has been compared with linear regression, average and last known value across four periods and has that the Random Walk method is no better that the

The table shows the average effect of living in a visited household (being treated), the share of the treated who talked to the canvassers, the difference in turnout

This study took a similar approach using a screening process, but instead of looking at the world’s largest markets, our focus was on the Swedish stock market. This study was not