The Stock Price Exposure of European Oil Companies

(1)

(2)

The Stock Price Exposure of European

Oil Companies

Abstract

(3)

Table of contents

1. Introduction ... 4

2. Literature ... 6

2.1 Stock price determinants in the oil industry... 6

2.2 Stock price determinants in the gold industry... 9

2.3 Oil price effect on the macro economy and international stock markets ... 10

2.4 Summary ... 11

3. Methodology... 13

3.1 Regression model... 13

3.2 Multifactor models... 14

3.2.1 A two-factor model for testing the oil price exposure ... 14

3.2.2 Four-factor model ... 14 3.2.3 Eight-factor model ... 14 4. Data... 16 4.1 Data sources ... 16 4.2 Variables description... 16 5. Results ... 22

5.1 Analysis two-factor model ... 22

5.2 Analysis four-factor model ... 22

5.3 Period analysis ... 24

5.4 Sector analysis... 28

5.5 The impact of operational cash flow and financial leverage... 29

5.6 High vs. Low levered ... 30

5.6 Analysis common and company specific factors ... 33

6. Conclusions ... 34

References ... 36

(4)

1. Introduction

Oil plays a major role in the world economy today. The worldwide demand for oil has increased about 33% in the last 18 years (1987-2005) from 62 million barrels per day to 83 million barrels per day. The increase is now caused for a large part by emerging markets in the Far East. Increase in the demand of oil has also a large impact on the oil price. From 1987 till 1999 the oil price was rather stable, but since 2002 the price has on average more than tripled (figure 1)1.

Figure 1. Development West Texas crude oil price 1984-2005

There are several papers that studied the impact of the oil price on the stock price of companies. These papers report evidence that there is a significant relationship between the oil price and the stock price. Jones and Kaul (1996), for example, found that oil price shocks have a significant impact on stock return in the United States, Canada, Japan and the United Kingdom. Boyer and Filion (2004) studied the impact of the oil price on the stock price of Canadian oil and gas companies. They found that the oil price has a significant impact on stock prices of oil and gas companies. These findings indicate an oil price effect.

Despite many papers found proofs for an oil price effect, it is striking that the relationship between the oil price and the stock prices of European oil companies has received little attention in the literature. In

(5)

this paper will therefore be investigated to which extent stock price returns of European oil companies are influenced by oil price changes and related stock price return determinants. The exposure of stock price returns of European oil companies will be measured with several common and company specific factors (Boyer and Filion, 2004). The common factors are the oil price, market return, exchange rate and interest rate. Operational cash flow, financial leverage, production and reserves are further applied as the company specific factors. This paper tries to find proof for an oil price effect by European oil companies. Furthermore a comparison with the results of previous studies will be made. The following research question will be used in this paper:

To which extent have oil price changes and related stock price determinants influenced the stock price returns of European oil companies in the period 2000-2004?

This paper follows earlier papers that studied the stock price exposure of oil companies to answer the research question. The paper of Boyer and Filion (2004) who investigated which common and company specific factors determine the stock return of Canadian oil and gas companies is used as guiding principle for this paper. The methodology used in this paper is further based on other studies of for instance Tufano (1998), Faff and Brailsford (1999) and Sadorsky (2001).

The results of this paper present evidence that the common variables market return and oil price have a significant impact on the stock price returns of European oil companies. For the common variables interest rate and exchange rate only significant impacts are found for yearly data. Furthermore the company specific variables operational cash flow and financial leverage have, based on the whole dataset, a significant impact. This paper finds no evidence that the variables oil production and reserves have an impact on stock price returns.The results of this paper offer advantages for corporate managers and investors, because they care about the exposure of firms (Tufano, 1998). A pension fund or private investor can use the results for his investment choices.

(6)

2. Literature

The literature on the stock return of companies distinguishes a considerable number of variables. In sub-section 2.1 literature concerning the determinants of stock price exposure of oil companies will be discussed. In addition the determinants in the gold industry will be discussed in sub-section 2.2. This literature is presented because gold mining companies are also active in the extraction industry and because these companies received quite some attention in the literature. Sub-section 2.3 discusses other literature that uses the oil price variable. Finally, sub-section 2.4 discusses why the described variables are important for this research. The literature is discussed from specific to general because this paper focuses on the crude oil industry. In the appendix an overview is given with the result of all the used papers.

2.1 Stock price determinants in the oil industry

This section describes literature concerning the stock price exposure of oil companies. The literature offers several common and company specific variables that are also important for this paper. Table 1 contains an overview of the results of these papers.

Many studies such as the papers of Jorion (1990), Ferson and Harvey (1994) and Faff and Brailsford use a multifactor model to investigate the stock price exposure. Like these studies Sadorsky (2001) used a multifactor model to investigate the risk factors in monthly stock returns of the Canadian oil and gas industry in the period 1983-1999. The determinants market return, exchange rate, oil price and interest rate are used in his study. He found that the market beta has a significant impact on the stock returns of oil and gas companies and that the oil and gas industry is less risky than the market. The oil price factor is significant and this is evidence that movements in the oil price impact oil stock returns. An increase in the oil price increases the stock returns of oil companies. A negative significant impact is found for the interest rate and exchange rate. Sadorsky (2001) uses a short-term interest rate in his paper and the negative significant impact corresponds with the results found by Bartram (2002). Higher short-term interest rates increase the costs of borrowing and reduce stock returns. The negative significant impact of the exchange rate was unexpected, because a depreciation of the Canadian dollar helps the energy exports. One explanation for the significant exchange rate exposure could be that oil companies do not hedge their exchange rate risk completely and this ensures that exchange rate exposure arises. This explanation is in contrast to the results of Chow, Lee and Solt (1997)2. They only found for horizons of 6 months and longer exchange rate exposures. In this paper, however, no distinction is made between short-term and long-term hedging horizons.

2

(7)

Table 1 Determinants of stock price return in the oil and gold industry

In this table the results of the literature are reported. Only the variables that are also used in this paper are reported. Tables with all the variables of the literature are reported in the appendix. + = positive significant impact on stock returns. – =negative significant impact on stock returns.

Oil/gold price Market return Interest rate Exchange rate Operational cash flow

Financial leverage

Production Reserves

Sadorsky (2001)* + + – –

Faff and Brailsford (1999)* + +

Boyer and Filion (2004)* + + – – + – +

Khoo (1994)* + –

Faff and Chan (1998)** + +

Tufano (1998)** – – +

* Investigated the determinants of the stock price in the oil industry

(8)

Another paper that studies the stock price exposure of oil companies is the study of Faff and Brailsford (1999). The variables market return, oil price and the exchange rate are used in their paper. They found that the market return has a positive impact on the stock returns of Australian oil companies. The oil price variable Faff and Brailsford (1999) used is measured in Australian dollars and in US dollars. For both oil price variables they found a positive significant impact. They found no evidence that the exchange rate factor had a significant impact on the oil and gas industry. Further they concluded, after the studied period has been subdivided in 2 sub-periods, that the oil price sensitivity is a long-term phenomenon.

A final paper, in which the determinants of stock return of Canadian oil and gas companies were investigated, is the paper of Boyer and Fillion (2004). They made a distinction between common and company specific factors that determine stock returns. The common factors they used are market return, interest rates, the Canadian dollar/US dollar exchange rate, oil prices and gas prices. Their results for these common factors are in accordance with the results of Sadorsky (2001). They found that the variable market return, the oil and gas price have a significant positive impact on stock returns of oil and gas companies. Again it appears that oil and gas companies are less risky than the market. The interest rate and exchange rate have a significant negative impact. Boyer and Fillion (2004) use a medium-term interest rate in their paper. Reasons they gave for the use of this interest variable are that the costs of business and equipment maintenance are large for oil companies. Furthermore, oil companies have to invest much in renewals and in finding new reserves. The costs are large and for this reason external financing is necessary. In this way, oil companies meet their growth and cash flow objectives. The negative impact of the exchange rate is consistent with the study of Sadorsky (2001) who found a negative impact for the Canadian oil and gas industry, but inconsistent with the result of Faff and Brailsford (1999) and Khoo (1994)3. The five company specific factors that Boyer and Fillion (2004) used are proven reserves, production volume, debt level, operational cash flows, and drilling success. They assumed that proven reserves, production volume, operational cash flows, and drilling success are used by investors as a criterion of a company’s operational and financial health. The number of production-years left and the growth potential determines the value of a company. Proven reserves and operational cash flows have a positive significant impact on stock returns and production volume a negative significant impact. This negative sign for the volume of production is striking. Boyer and Fillion (2004) gave as a possible explanation that the returns related to the production of crude oil and gas is concave. This means that energy firms experience decreasing returns of scale. For the variables drilling success and debt level they did not find a significant relationship.

3

(9)

2.2 Stock price determinants in the gold industry

The literature addressed above presented the determinants of the stock returns of oil companies. The following part of this section presents literature that discusses the determinants of stock price exposure of gold companies. Table 1 contains an overview of these papers.

Faff and Chan (1998) studied the determinants of the Australian gold stock index. They used the variables market return, gold price return, exchange rate and the interest rate. The market return and the gold price have a positive significant impact on the gold stock index. The risk of gold mining companies appears to be higher than that of the stock market. No significant relationship is found between the exchange rate and the gold stock index returns. For the interest rate they used a short-term, medium-term and long-term interest rates. All these interest rate variables did not have a significant impact.

Another important paper that investigates the stock price exposure in the gold industry is the paper of Tufano (1998). He studied the determinants of the stock price exposure of North American gold mining firms. Tufano (1998) found that gold betas are not stationary in the period 1990-1994. There is a lot of variation from quarter to quarter in the gold betas. The ten determinants Tufano (1998) used in his paper are the gold price, gold return volatility, interest rates, production quantity, financial leverage, cash costs, reserves, percentage hedged, forward prices and percent of assets in mining. He found that the determinants financial leverage and percentage of assets in mining are significantly positive related to the level of exposure. The positive significant impact for financial leverage is consistent with the result of Fama and French (1992)4. A reason for the positive significant impact of financial leverage is that equity-holders run larger risk and they must be rewarded for this larger risk. Furthermore, Tufano (1998) found that the change in the gold price, the gold return volatility, the percentage hedged and the long-term interest rate have a negative significant impact on the gold beta. In the paper a long-term and a short-term interest rate were used. The negative relationship is found for the long-term interest rate and was predicted. For growing companies with growing production levels the relationship between the long-term interest rate and the stock price exposure is positive. Tufano (1998) found no relationship between the gold beta and the short-term interest rate. Striking is the insignificant relationship for the variables production quantity and reserves. Forward prices are also insignificant.

(10)

The discussed literature in this and sub-section 2.1 uses common and company specific factors to explain the stock price exposure of oil and gold companies. The common factors are the oil or gold price, market return, exchange rate and the interest rate. All these variables are able to explain a part of the stock price exposure of oil companies, but commodity price exposure is the most important determinant. The literature found that the market return is also an important determinant. Boyer and Filion (2004) and Sadorsky (2001) found that the exchange rate has a significant impact on the stock return of oil companies. Faff and Brailsford (1999) on the other hand found no proof for exchange rate exposure in the oil industry and Faff and Chan (1998) found no proof in the gold industry. Boyer and Filion (2004) and Tufano (1998) furthermore found proof for interest rate exposure. The most important company specific factors discussed are financial leverage, production, operational cash flow and reserves. Boyer and Filion (2004) found proof that production, operational cash flow and reserves have a significant impact on stock returns of oil companies. Tufano (1998) found that financial leverage has a significant impact on the stock returns of companies in the gold industry.

2.3 Oil price effect on the macro economy and international stock markets

This section discusses the impact of the oil price on the macro economy and international stock markets. Firstly the macro economical impact will be discussed and secondly the impact on international stock markets.

The impact of the oil prices also plays an important role in the world economy. The demand for oil still increases and as well as the oil price. This higher oil price leads to an income transfer from importing to exporting countries. Further consequences of an increase in oil price are higher inflation, a rise in production costs of goods and an increase of the price level. The International Energy Agency (IEA) and the economic department of the Organization for Economic Co-operation and development (OECD) investigated the impact of an oil price increase on GDP in 20045. They estimate that a $10 increase in oil price per barrel from $25 to $35 leads to a decrease in GDP in the OECD6 of 0.4% in the first and 0.4% in the second year of high oil prices.

Jones and Kaul (1996) studied the impact of the oil price changes in the United States, Canada, Japan, and the United Kingdom. They found that the changes in oil price have a significant negative impact on a country’s market index. Furthermore, they concluded that the impact of the oil price changes differed enormously between the countries. A possible explanation for this is that markets differing in concentrations of particularly natural resources and industrial sectors may experience different oil price effects (Faff and Brailsford, 1999). Canada, for example, is a country with a relatively high

5

“Analysis of the Impact of High Oil Prices on the Global Economy” an investigation of the International Energy Agency and the Organization for Economic Co-operation and development (2004),

http://www.iea.org/textbase/papers/2004/high_oil_prices.pdf.

6

(11)

proportion of natural resources, whereas in Japan the proportion of natural resources is limited. For this reason Japan is very dependent on the import of oil and this explains the high relationship between oil price changes and real stock returns in Japan.

In another paper Driesprong, Jacobsen and Maat (2005) studied whether changes in oil prices have an impact on stock market returns worldwide. They found that changes in oil price can predict stock returns in 12 out of 18 developed countries in their sample significantly. Another important finding is that, if there is an increase in oil prices, future stock returns are lower.

2.4 Summary

In the literature review several determinants are described that determine the stock price return. Many of these determinants will also be used in this paper. This section will give a summary.

Sadorsky (2001), Faff and Brailsford (1999) and Boyer and Filion (2004) found proof for an oil price effect by Australian and Canadian oil and gas companies. Further the studies found evidence that the variables market return and interest rate have a significant impact on the stock returns of Canadian oil companies. For the variable exchange rate only Sadorsky (2001) and Boyer and Filion (2004) found a significant impact. Boyer and Filion (2004), moreover, studied the impact of the company specific variables proven reserves, production volume, debt level, operational cash flows, and drilling success. They found that proven reserves, production volume and operational cash flow have a significant impact on the stock returns of Canadian oil companies.

Faff and Chan (1998) and Tufano (1998) studied the determinants of the stock return in the gold industry. The studies found that the gold price had a significant impact on the stock return of gold companies. Striking results are that Faff and Chan (1998) found no significant impact for the variables exchange rate and interest rate. The study of Tufano (1998) on the other hand found a significant impact for the variable interest rate. Further important results of Tufano (1998) are the positive significant impact of financial leverage and the insignificant impact of production quantity and reserves.

The impact of the oil price on the macro economy and international stock markets is described in section 2.3. A study of the IEA and the OECD estimate that an oil price increase has a negative impact on the GDP of the OECD as a whole. Jones and Kaul (1996) and Driesprong, Jacobsen and Maat (2005) found that the oil price had a significant negative impact on international stock markets.

(12)

(13)

3. Methodology

This section describes the methodology used for this paper. Sub-section 3.1 describes the regression model that is used and the statistics which test whether the regression model has the correct assumptions. Sub-section 3.2 describes the different multifactor models that will be tested.

3.1 Regression model

In this research a generalized least square (GLS) panel regression with cross section weights is used to analyze the relationship between the stock price returns of oil companies and the determinants of these stock price returns. The use of this GLS panel regression technique is based on the paper of Boyer and Filion (2004). This GLS panel regression technique is a combination between a panel data analysis and the GLS method. Panel data contains observations about multiple cases (people, firms) that are observed over multiple time periods. In other words, it contains both cross-sectional and time-series dimensions. Panel data offers several advantages over cross-sectional data or time-series data. An advantage of panel data is that it offers more degrees of freedom and reduces collinearity between the explanatory variables7. It improves the efficiency of the estimates. Another advantage is that it is possible to control for unobserved differences (heterogeneity) between cases. The GLS method is a technique that uses the ordinary least square (OLS) assumptions. This GLS method has several advantages over the OLS method that are useful for cross-sectional and time-series data. The classical linear regression model (CLRM) assumes that the variance of the errors is constant, this is known as homoscedasticity. If the variance of the errors is not constant, they are said to be heteroscedastic8. If the errors are heteroscedastic, but assumed homoscedastic, an implication would be that standard errors estimates could be wrong. The GLS method with the option cross section weights makes corrections for the presence of heteroscedasticity. The second advantage is that the GLS method makes corrections for autocorrelation possible. Time-series data often contain autocorrelation. The CLRM assumes that the error terms are uncorrelated over time. Error terms that are not uncorrelated over time contain autocorrelation or are serial correlated9. To test if the data after the use of the GLS method contains autocorrelation a Durbin-Watson statistic is calculated. The paper of Bhargaza, Franzini and Narendranathan (1982) will be used to check if the Durbin-Watson statistic has a correct value. To test if there is multicollinearity between the explanatory variables a maximum value of 0.7 is used as rule of thumb. This means that if the correlation between explanatory variables is higher than 0.7 there might be multicollinearity. Normality of the data is an important assumption for the GLS method. A Jarque-Bera test is used to test if the data are normally distributed. The data are normally distributed if the probability of this statistic is larger than 0.05. An F-test is used to test whether the results are statistically significant. The equation is significant if the probability of the F-test is below

7_{Hsiao, C., 2003, Analysis of Panel Data, page 3}

8_{Brooks, C., 2002, Introductory econometrics for finance, page 144-148} 9

(14)

0.05. The total relationship between the dependent and explanatory variables in the regression is measured with the adjusted R-squared and lies between 0 and 1. A high adjusted R-squared means that the dependent variable is for a large part explained by the explanatory variables.

3.2 Multifactor models

3.2.1 A two-factor model for testing the oil price exposure

The multifactor model will be used for investigating the oil price exposure of European oil companies. Using the multifactor model is based on the studies of Jorion (1990), Faff and Brailsford (1999), Sadorsky (2001) and Boyer and Filion (2004). A generalized least square (GLS) panel regression with cross-section weights will be used to test the model. The model can be expressed as follows:

Rit = α + β1RMt + β2ROILPt + έt (1)

whereRit denotes the stock price return of oil company i in quarter t, RMt is the quarterly market return,

ROILPt is the quarterly change in the oil price and έt is the error term.

3.2.2 Four-factor model

The second multifactor model that will be used in this study is the model with the four common factors market return, oil price, exchange rate and interest rate. Again a GLS panel regression with cross-section weights will be used. Model 2 that add more macroeconomic factors to model 1 may probable give more information about the stock price exposure. The four-factor model is defined as follows:

Rit = α + β1RMt + β2ROILPt + β3RERt + β4RIRt + έt (2)

whereRit denotes the stock price return of oil company i in quarter t, RMt is the quarterly market return,

ROILPt is the quarterly change in the oil price, RERt is the quarterly change in the exchange rate, RIRt is

the quarterly change in the interest rate and έt is the error term.

3.2.3 Eight-factor model

(15)

is on annual basis as company characteristics were only available on annual basis. The following model is defined:

Rit = α + β1RMt + β2ROILPt + β3RERt + β4RIRt + β5OCFit + β6FLit + β7Prodit + β8Resit + έt (3)

whereRit denotes stock price return of oil company i in quarter t, RMt is the quarterly market return,

ROILPt is the quarterly change in the oil price, RERt is the quarterly change in the exchange rate, RIRt is

the quarterly change in the interest rate, OCFit is the value of the operational cash flow of company i in

year t, FLit is the value of the financial leverage of company i in year t, Prodit is the value of the

number of produced oil barrels of company i in year t, Resit is the value of the quantity oil reserves in

(16)

4. Data

In this section the data used in this paper are described. Firstly the data sources and the period will be presented. Secondly, the variables and their expected impact on the stock return of European oil companies are described. Finally an overview and a description are given of the descriptive statistics and the mutual correlation between the variables.

4.1 Data sources

The sample used for this study consists of 50 European oil companies and are selected from the program Amadeus. This program is used for selecting companies in the oil sector, because it gives access to financial data of more than 250.000 European companies. The oil companies for this study exist of petroleum extraction companies, companies active in the manufacturing of refined petroleum products and oil service companies. Even though Amadeus lists 89 oil companies in these sub-sectors, only 50 oil companies have the required stock data. Table 2 gives an overview of the sample. All the financial data used come from Amadeus, Datastream, Energy Information Administration (EIA) and the annual reports of the European oil companies (see section 4.2). The data are both quarterly and yearly and covers the period 2000-2004. The reason for using both quarterly and yearly data is that for certain variables only yearly data are available. When data with a higher frequency is used, for example monthly, results between monthly and yearly data will probably differ too much to be able to draw a correct conclusion.

Table 2 Overview sub-sectors

Sub-sector (NACE revision 1.110₎ _{Number of Companies}

1110 Extraction of crude petroleum and natural gas 23

1120 Service activities incidental to oil and gas extraction excluding surveying 9

2320 Manufacture of refined petroleum products 14

Active in all sub-sectors 4

Total companies 50

4.2 Variables description

Stock price return. Stock prices of European oil companies are used for measuring the stock price

returns. The stock prices of the oil companies are collected from Datastream. The stock price returns are measured as the logarithm of quarterly or yearly stock price return.

Market return. The Dow Jones STOXX® 60011 index is used as the market index. The data of this index is collected from Datastream. This index consists of large, middle and small capitalisation

10_{Industry activity classification in the EU}

(17)

companies in the European region. Market return is measured as the quarterly or yearly index return. The sensitivity of this variable will show whether the oil companies are more or less risky than the stock market. The expectations are, based on Faff and Brailsford (1999), Sadorsky (2001) and Boyer and Filion (2004), that the market return has a significant positive impact on the stock returns of oil companies.

Oil price. The European Brent Spot Price FOB in US dollars per barrel is used for the quarterly or

yearly change of the oil price variable. The oil spot prices are obtained from the Energy Information Agency (EIA). Most studies use oil future prices as the oil price return variable, because spot prices are more affected by short-run price fluctuations (Sadorsky, 2001; Boyer and Filion, 2004). Temporary oil shortages or surpluses are the main cause of short-run price fluctuations (Sadorsky, 2001). This study uses spot prices, because the available databases do not contain future prices. Faff and Brailsford (1999), Sadorsky (2001) and Boyer and Filion (2004) found that a change in the oil price has a positive significant impact on oil companies. In conformity with these studies it is expected that the oil price has a positive significant impact on stock price returns of oil companies.

Exchange rate. The exchange rate is the value of one US dollar expressed in the local currency

(USD/Local currency)12. Exchange rate data are collected from Datastream. The variable exchange rate is measured as the quarterly or yearly return. The reason for the use of the USD/Local currency exchange rates is that oil prices are in US dollars and therefore oil companies will mainly be paid in US dollars. In this study it is expected that the exchange rate will have a positive impact on the stock price return of European oil companies, meaning that a depreciation of the local currency against the US dollar will lead to a higher stock price return.

Interest rate. The 3-month LIBOR rate is used as the short-term interest rate and is collected from

Datastream. The quarterly and yearly interest rate of the 3-month LIBOR rate is used because LIBOR is the most used short-term interest rate benchmark worldwide13. The investment needs of oil companies make that the interest rate is an important variable in this paper. Oil companies need, for example, to invest to find new oil reserves and often external financing is necessary (Boyer and Filion, 2004). Therefore an interest rate increase or decrease will have a large impact on an oil company. The expectation is that the interest rate has a negative impact on stock price returns of oil companies. This expectation is in conformity to the results of Sadorsky (2001) and Boyer and Filion (2004).

Operational cash flow. The end of the year operational cash flow is used and the data are obtained

from Amadeus. Operational cash flow measures the cash generated from operations and is here defined as earnings before interest and depreciation minus taxes in million of euros14. Investors use operational cash flow as a proxy for a company’s operational health (Boyer and Filion, 2004) and it is expected that operational cash flow will have a positive impact on stock price return.

12_{The value of 1 US dollar for the Euro, British pound, Norwegian krone, Russian rouble, New Romanian leu,} Lithuanian lita, Hungarian forint, Polish zloty, Slovak koruna and Czech koruna are used.

13_{http://www.bba.org.uk/bba/jsp/polopoly.jsp?d=141} 14

(18)

Financial leverage. The end of the year book value of debt-to-market value of equity ratio is used as

financial leverage. Total debt is obtained from Amadeus and the amount market equity from Datastream. Studies of Fama and French (1992) and Tufano (1992) showed that financial leverage is an important determinant of the stock price return. The expectation is that financial leverage will have a positive impact on stock returns (Tufano, 1998) because investors who invest in levered oil companies run more risk and they must be rewarded for this.

Production. Production is measured as the yearly production in million barrels. The quantity produced

is obtained from the annual reports of the oil companies. Boyer and Filion (2004) found that production has a negative impact on stock returns and Tufano (1998) found no impact. The expectation in this study is that production will have a positive impact on stock returns. Higher production indicates a higher demand and therefore profit will rise.The alternative expectations, based on the papers of Boyer and Filion (2004) and Tufano (1998), is that production has a negative significant impact or no impact on stock price returns.

Reserves. Reserves are measured as the end of the year reserves in million barrels and are collected

from the annual reports of the oil companies. Proven15 and probable16 reserves are used because many companies have not published the proven reserves separately. Investors can use oil reserves as a proxy for the operational and financial health (Boyer and Filion, 2004). The expectation in this study is that the quantity of reserves should have a positive impact on stock returns of oil companies. An alternative expectation, based on the paper of Tufano (1998), is that oil reserves have no significant impact.

The four common variables market return, oil price, exchange rate and interest rate are on a quarterly and a yearly basis and the company specific variables operational cash flow, financial leverage, production and reserves on a yearly basis. Table 3 presents a summary of the common and company specific variables.

There are a couple of reasons why for the company specific variables operational cash flow, financial leverage, production and reserves the real values are taken and not the yearly returns. The first major reason is the low data availability in the period before 2000 for the company specific variables and thus makes it impossible to calculate returns. A second reason is the way investors use the company specific variables as an estimation of a company’s operational and financial health. The quantity of

15

Proven Reserves - defined as oil and gas "Reasonably Certain" to be producible using current technology at current prices, with current commercial terms and government consent, also known in the industry as 1P. Some Industry specialists refer to this as P90, i.e., having a 90% certainty of being produced.

http://en.wikipedia.org/wiki/Oil_reserves#Definition_of_Oil_Reserves 16

Probable Reserves - defined as oil and gas "Reasonaby Probable" of being produced using current or likely technology at current prices, with current commercial terms and government concent. Some Industry specialists refer to this as P50, i.e., having a 50 % certainty of being produced. This is also known in the industry as 2P or Proven plus probable.

(19)

debt, for example, is for an investor a correct manner to assess the way a company is financed and not the yearly change in it. A final reason is that for the operational cash flow it is impossible to calculate the yearly change because there are negative values presented.

Table 3 Measure of common and company specific variables and there source

Variable Source Measure17 Effect on stock price

Market return

Datastream, Stoxx

Quarterly return Dow Jones STOXX® 600 oil and gas index - 1 month LIBOR rate

+

Oil price Energy information administration

((Europe Brent Spot Price FOB in $/Barrel)t/(Europe Brent Spot Price FOB in $/Barrel)t-1)-1

+

Exchange rate

Datastream ((Exchange rate US$/Local currency)t/(Exchange rate US$/Local currencyt-1)-1

+

Interest rate Datastream (3 month LIBOR rate)t/(3 month LIBOR rate)t-1)-1 –

Operational Cash flow

Amadeus and Datastream

Operational cash flow in EURt +

Financial leverage

Amadeus and Datastream

Year-end long term debt/Year-end market value of companyt +

Production Annual report Production in million of barrelst +

Reserves _{Annual report} _{Reserves in million of barrels}

t +

Table 4 presents the descriptive statistics of the common variables. The numbers of quarterly observations included in this study are 993, while the original sample contains 1000 observations. The probability of the Jarque-Bera statistic in figure 2 shows that with 1000 observations the sample is not normally distributed. The stock price returns in figure 3 show that extreme outliers are the reason that the data are non-normally distributed. In total there are 7 outliers excluded with quarterly changes above 500% and below -80%.

Table 4 Descriptive statistics of common variables on quarterly basis

Stock price return Market return Oil price Exchange rate Interest rate

Mean 0.0112 -0.0148 0.0367 -0.0075 -0.0174 Median 0.0108 0.0181 0.0296 0.0001 -0.0286 Maximum 0.4435 0.1504 0.4377 0.1316 0.1900 Minimum -0.5626 -0.2334 -0.2055 -0.1774 -0.1841 Std. Dev. 0.1007 0.1005 0.1638 0.0461 0.0957 Observations 993 993 993 993 993

The correlation matrix in table 5indicates that all common variables are positively correlated with the stock price return of oil companies. The only two variables which correlate negatively with each other are the market return and the exchange rate. Table 5 suggest that there is no multicollinearity presented among the explanatory variables.

(20)

Figure 2 Distribution of stock price returns Figure 3 Quarterly stock price returns for thousand observations

Table 5 Correlation matrix common variables on quarterly basis

Stock price return Market return Oil price Exchange rate Interest rate Stock price return 1.0000 0.2739 0.1120 0.0103 0.0350

Market return 1.0000 0.0023 -0.0039 0.0987

Oil price 1.0000 0.1741 0.3432

Exchange rate 1.0000 0.3174

Interest rate 1.0000

In table 6 the descriptive statistics of all the variables are presented. Only 11 companies with total 55 observations are included. The reason for the low number of observations is the lack of data. This is mainly caused by the variables production and reserves. Table 7 shows that there exists multicollinearity between the variables operational cash flow, production and reserves. This multicollinearity may also be caused by the small number of observations. When the regressions are tested the presence of multicollinearity will be taken into account.

0 40 80 120 160 200 -0.4 -0.2 -0.0 0.2 0.4

(21)

Table 6 Descriptive statistics of common and company specific variables on a yearly basis Stock price return Market return

Oil price Exchange rate

Interest rate Operational cash flow Financial leverage Production Reserves Mean 0.0283 -0.0629 0.1316 -0.0424 -0.0479 7766.4040 0.6796 324.6430 3652.5130 Median 0.0151 -0.0519 0.0060 -0.0695 -0.1266 5634.0000 0.5578 216.9580 2018.6960 Maximum 0.5441 0.1368 0.5566 0.1005 0.4626 20370.1100 1.6238 899.0000 10730.0000 Minimum -0.2375 -0.3247 -0.1430 -0.2268 -0.3295 -0.9129 0.0842 0.0200 8.2000 Std. Dev. 0.1254 0.1718 0.2720 0.0909 0.2838 7088.0220 0.4125 343.4347 3763.0400 Observations 55 55 55 55 55 55 55 55 55

Table 7 Correlation matrix common and company specific variables on a yearly basis

Stock price return

Market return

Oil price Exchange rate

Interest rate Operational cash flow

Financial leverage

Production Reserves Stock price return 1.0000 0.4952 -0.1353 0.0138 0.0085 -0.3049 -0.4286 -0.3144 -0.3240 Market return 1.0000 -0.3388 -0.0639 0.0990 0.0332 -0.1654 0.0167 -0.0678 Oil price 1.0000 -0.6910 -0.0846 -0.0431 0.1461 0.0244 0.0005 Exchange rate 1.0000 0.4734 -0.0098 -0.2157 -0.0916 -0.0346 Interest rate 1.0000 0.0443 -0.1049 -0.0095 -0.0040 Operational cash flow 1.0000 0.1775 0.9620 0.9368

Financial leverage 1.0000 0.1603 0.1080 Production 1.0000 0.9579

(22)

5. Results

5.1 Analysis of the two-factor model

Table 8 presents the results of the two-factor model. The probability of the Jarque-Bera statistic shows that the regression model is normally distributed. The probability of the F-test shows that the results for this regression are statistically significant. Table 8 shows that all coefficients are significant at the 5% level. Market return and the oil price have significant positive impact on the stock price return of European oil companies. An interesting observation is the market beta of 0.2902. This observation shows that European oil companies are less risky than the overall market. It also shows that oil stocks are not interesting hedging tools in the period 2000-2004. The impact of the oil price, with a coefficient of 0.0825, appears to be lower than the impact of the market return. This is in contrast with the results of Boyer and Filion (2004). They found a higher impact for the oil price coefficient. The adjusted R-squared shows that the two variables can explain 27.28% of the stock price return. The positive impact of market return and oil price corresponds with the hypotheses that movements in the stock market and in the oil price have significant positive impact on stock price returns. The results correspond with the findings of Faff and Brailsford (1999) and Sadorsky (2001).

Table 8 Regression two-factor model

Variable Coefficient Std. Error t-Statistic Prob.

Constant 0.0125 0.0018 7.0480* 0.0000

Market return 0.2902 0.0170 17.0418* 0.0000

Oil price 0.0825 0.0105 7.8957* 0.0000

Panels included: 50 F-statistic 175.4413 Total panel (unbalanced) observations: 993 Prob. F-statistic 0.0000 Probability Jarque-Bera statistic 0.3117 Adjusted R-squared 0.2728 The table reports the following regression:

Rit = α + β1RMt + β2ROILPt + έt

whereRit denotes the stock return of oil company i in quarter t, RMt is the quarterly market return, ROILPt is the quarterly change in the oil price and έt is the error term. Timeframe is 2000-2004. Econometric model is a GLS panel regression with cross sectional weights.

*significant at the 5% level

5.2 Analysis of the four-factor model

(23)

Table 9 Regression four-factor model

Constant 0.0127 0.0018 6.9360* 0.0000 Market return 0.2874 0.0171 16.7924* 0.0000 Oil price 0.0814 0.0112 7.2872* 0.0000 Exchange rate 0.0317 0.0378 0.8378 0.4024 Interest rate 0.0015 0.0200 0.0768 0.9388

Rit = α + β1RMt + β2ROILPt + β3RERt + β4RIRt + έt

whereRit denotes the stock return of oil company i in quarter t, RMt is the quarterly market return, ROILPt is the quarterly change in the oil price, RERt is the quarterly change in the exchange rate, RIRt is the quarterly change in the interest rate and έt is the error term. Timeframe is 2000-2004. Econometric model is a GLS panel regression with cross sectional weights. *significant at the 5% level

The first interesting observation in this model is the insignificant impact of the exchanger rate on the stock price return of European oil companies. This insignificant impact does not correspond with the hypothesis that the exchange rate has a significant positive impact. A reason for the insignificant impact of exchange rate return could be the degree of foreign involvement of oil companies. The degree of foreign involvement means the ratio of foreign trade to total trade (Jorion, 1990). Jorion (1990) found that there is a relation between the exchange rate and the degree of foreign involvement. European companies possibly have a low degree of foreign involvement so that exchange rate has no significant impact. A second reason could be that European oil companies hedge their exchange rate risk and as a result of this the exchange rate exposure disappears.

The second important observation, and in contrast with the hypothesis, is the insignificant impact of the interest rate. This result suggests that a change in the interest rate has no effect on the stock price return of European oil companies. Hedging the interest rate risk could be an explanation for the insignificance. An oil company could, for example, use interest rate swaps18, which would limit the exposure of interest rate changes.

As mentioned in section 4.1, quarterly and yearly data are used for this paper. To investigate whether the results for quarterly and yearly data correspond with each other a second regression is run with yearly data. Table 10 shows the results for this regression.

(24)

Table 10 Regression four-factor model on annual level

Constant 0.0640 0.0067 9.5058* 0.0000 Market return 0.3564 0.0444 8.0275* 0.0000 Oil price 0.1399 0.0396 3.5301* 0.0005 Exchange rate 0.3229 0.1196 2.7010* 0.0075 Interest rate -0.1018 0.0270 -3.7691* 0.0002

whereRit denotes the stock return of oil company i in quarter t, RMt is the quarterly market return, ROILPt is the quarterly change in the oil price, RERt is the quarterly change in the exchange rate, RIRt is the quarterly change in the interest rate and έt is the error term. Timeframe is 2000-2004. Econometric model is a GLS panel regression with cross sectional weights. *significant at the 5% level

Before the results in table 10 can be analysed it should be mentioned that the regression contains autocorrelation. During the interpretation of the results one should take this into account. The results in table 10 present different results than table 9. The exchange rate has a positive significant impact on yearly basis and it supports the hypothesis that the exchange rate has a significant impact. This means that a depreciation of the local currency against the US dollar leads to higher stock price returns. The second interesting result is the significant negative impact of the interest rate. This corresponds with the hypothesis that the interest rate has a negative impact on stock price returns of oil companies. In contrast to the quarterly results, these results show that European oil companies have not completely hedged their exchange rate and interest rate exposure. Further, the results in table 10 show that the market return and the oil price have a larger impact on yearly basis.

5.3 Period analysis

This section studies the differences between the two time periods 2000:1-2002:6 and 2002:7-2004:12. The first time period is characterised by large fluctuations in the development of the oil price. The lowest oil price in 5 years, less than 20 dollars per barrel, lies in this first period. The second period is characterised by an increasing oil price. In this period the oil price has more than doubled with a top above 50 dollar per barrel. After this top a considerable correction took place. Figure 4 shows the development of the European Brent Spot price for the whole 5-year period 2000-2004.

(25)

shows that in the first period all the variables have a significant impact on the stock price return. In the second period only the exchange rate has no significant impact.

Figure 4 Development European Brent Spot Price FOB 2000-2004

European Brent Spot Price FOB in US dollars per barrel

0 10 20 30 40 50 60 4-0 1-0₀ 4-07-0₀ 4-01-0₁ 4-07-0₁ 4-01-0₂ 4-07-0₂ 4-01-0₃ 4-07-0₃ 4-01-0₄ 4-07-0₄ Source: Energy Information Administration

Table 11 Regression four-factor model for the periods 2000:1-2002:6 and 2002:7-2004:12

2000:1-2002:6

Constant 0.0075 0.0025 2.9627* 0.0032 Market return 0.1999 0.0293 6.8164* 0.0000 Oil price 0.1148 0.0148 7.7719* 0.0000 Exchange rate 0.1594 0.0486 3.2840* 0.0011 Interest rate -0.0422 0.0247 -1.7069* 0.0885

Cross-sections included: 50 F-statistic 49.7450 Total panel (unbalanced) observations: 495 Prob. F-statistic 0.0000 Probability Jarque-Bera statistic 0.7985 Adjusted R-squared 0.2960

2002:7-2004:12

Constant 0.0154 0.0030 5.1247* 0.0000

Market return 0.3118 0.0215 14.4740* 0.0000

Oil price 0.0631 0.0142 4.4356* 0.0000

Exchange rate -0.0601 0.0670 -0.8960 0.3707

Interest rate return 0.0663 0.0288 2.3035* 0.0217

Cross-sections included: 50 F-statistic 64.9221 Total panel (unbalanced) observations: 498 Prob. F-statistic 0.0000 Probability Jarque-Bera statistic 0.0158 Adjusted R-squared 0.3660 The table reports the following regression:

(26)

The results in table 11 show that the impact of the market return is much higher in the second period. Figure 5 shows that the development of the DJ STOXX® Oil & Gas index in the first part of the first period is much better than the development of the DJ STOXX® 600 index. This means that oil companies performed better than the total market in the first period. In the second period the development of both indexes is more similar. The large difference in development of the market return and the stock price return of the European oil companies in the first period is a possible reason of the lower impact of the market return in the first period.

Figure 5 Development DJ STOXX® 600 and DJ STOXX® Oil & Gas

0 50 100 150 200 250 300 350 400 450 4-0 1-0₀ 4-07-0₀ 4-01-0₁ 4-07-0₁ 4-01-0₂ 4-07-0₂ 4-01-0₃ 4-07-0₃ 4-01-0₄ 4-07-0₄ DJ STOXX® 600 DJ STOXX® Oil & Gas

Source: Datastream

An interesting observation is the higher impact of the oil price in the first period. A possible explanation is the large fluctuations in the oil price in the first period. These fluctuations in the oil price possibly have a large impact on the stock price return of oil companies. A strong rising oil price in the second period on the other hand has less impact on the stock price return than large fluctuations in the oil price. An increasing oil price therefore does not mean that the impact on the stock price returns of oil companies becomes higher.

The results show that there only is a positive significant exchange rate exposure in the first period. This positive impact indicates that a depreciation of the local currency against the US dollar leads to higher stock returns for European oil companies. Figure 6 shows that the exchange rates are relatively high in the first period for the US dollar against the Euro and the US dollar against the British pound19. These high exchange rates indicate a strong US dollar against the local currencies and this might explain the positive exchange rate exposure in the first period. In the second period the local

(27)

currencies strongly appreciate against the US dollar. Hedging the negative impact of an appreciation is a possible explanation of the insignificant impact of the exchange rate.

Figure 6 Cumulative development exchange rate

Development exchange rate

-0,3 -0,2 -0,1 0 0,1 0,2 0,3 4-1_-0 0 4-7_-0 0 4-1_-0 1 4-7_-0 1 4-1_-0 2 4-7_-0 2 4-1_-0 3 4-7_-0 3 4-1_-0 4 4-7_-0 4 USD/Euro USD/GBP Source: Datastream

A final interesting result is the different impact of the interest rate return in the first and second period. In the first period the interest rate return has a negative significant impact on the stock price return of European oil companies. The second period on the other hand shows a positive significant impact. The development of the interest rate in figure 7 shows that the interest rate is relatively high in the first period and relatively low in the second period. As earlier mentioned oil companies need often external financing for their investment need. A relatively high interest rate as in the first period means that the costs of debt are relatively high and this possibly explains the negative impact of the interest rate return in the first period. The relatively low interest rate in the second period means lower costs of debt and this possible leads to higher stock price returns.

Figure 7 Development 3-Month Libor rate

Development 3-month Libor rate

(28)

5.4 Sector analysis

This section investigates the differences between the sectors. The results for the different sub-sectors are presented in table 12. The four companies who are active in all three sub-sub-sectors (table 2) are excluded from the regressions. It should be mentioned that the results of the oil service companies contain autocorrelation.

Table 12 Four-factor regression for the three sub-sectors

Extraction of crude oil and natural gas

Constant 0.0148 0.0027 5.5691* 0.0000 Market return 0.3019 0.0247 12.2128* 0.0000 Oil price 0.1029 0.0161 6.3776* 0.0000 Exchange rate 0.0372 0.0570 0.6533 0.5139 Interest rate 0.0122 0.0289 0.4210 0.6740

Service activities incidental to oil and gas extraction excluding surveying

Constant 0.0202 0.0057 3.5729* 0.0005 Market return 0.3293 0.0527 6.2523* 0.0000 Oil price 0.0586 0.0344 1.7049* 0.0901 Exchange rate 0.3024 0.1189 2.5456* 0.0118 Interest rate 0.0036 0.0617 0.0590 0.9530

Manufacturing of refined petroleum products

Constant 0.0039 0.0037 1.0588 0.2907 Market return 0.2139 0.0347 6.1643* 0.0000 Oil price 0.0528 0.0226 2.3311* 0.0205 Exchange rate -0.1136 0.0730 -1.5552 0.1211 Interest rate -0.0354 0.0405 -0.8731 0.3834

(29)

The market return and the oil price are the variables that have a positive significant impact on the stock returns of petroleum extraction companies. The oil price compared to the other sub-sectors appears to have the largest impact on petroleum extraction companies. A possible explanation is that petroleum extraction companies are the more dependent on oil prices than the companies in the other sub-sectors. The only products petroleum extraction companies produce compared to the other companies in the other sub-sector is crude oil and gas. The variables exchange rate and interest rate both have an insignificant impact. Hedging the exchange and interest risk is possible the explanation for this insignificant impact.

Petroleum manufacturing companies show many resemblances with petroleum extraction companies. The difference is that the market return and the oil price have a smaller impact on the stock price returns of petroleum manufacturing companies. A possible explanation for the smaller impact of the oil price variable is that petroleum manufacturing companies are less dependent on the oil price than petroleum extraction companies, because of the semi-finished and finished products they produce.

In comparison with petroleum extraction companies and petroleum manufacturing companies there are three variables that have a significant impact on oil service companies. Market return, oil price and the exchange rate are positively significant. Oil service companies are the only companies in which the exchange rate is significant. A possible explanation for this exchange risk exposure is the global activities of oil service companies. Global activities are a possible reason that service companies are paid in different local currencies.

5.5 The impact of operational cash flow and financial leverage

This sub-section investigates the impact of the inclusion of the company specific variables cash flow and financial leverage in the four-factor model. The mean reason for the use of this model before the eight-factor model will be used is the low data availability for the variables production and reserves. Table 13 shows the regression results for the multifactor model. The model shows that both company specific variables are significant. The results further show that the variables market return and oil price are positively significant. Interesting is (in comparison with table 10) that the exchange rate and the interest rate are no longer significant. The adjusted R-squared is in comparison with table 10 more than doubled to 60.45%.

(30)

Table 13 Impact operational cash flow and financial leverage

Constant 0.0935 0.0257 3.6360* 0.0004

Market return 0.2597 0.0369 7.0337* 0.0000

Oil price 0.0813 0.0335 2.4277* 0.0169

Exchange rate 0.0080 0.1090 0.0733 0.9417

Interest rate -0.0349 0.0228 -1.5303 0.1290

Operational cash flow 1.75E-05 5.20E-06 3.3686* 0.0011

Financial leverage -0.1348 0.0187 -7.2125* 0.0000

whereRit denotes stock return of oil company i in quarter t, RMt is the quarterly market return, ROILPt is the quarterly change in the oil price, RERt is the quarterly change in the exchange rate, RIRt is the quarterly change in the interest rate, OCFit is the value of the operational cash flow of company i in year t, FLit is the value of the financial leverage of company i in year t and έ_t is the error term. Timeframe is 2000-2004. Econometric model is a GLS panel regression with cross sectional weights. *significant at the 5% level

A final, not expected result of the regression is the negative significant impact of the variable financial leverage. The expectation was that financial leverage would have a positive impact on stock returns. A possible explanation for this negative impact is that a part of the companies is relatively highly levered and because of this it is difficult to get new debt.

5.6 High vs. Low levered

The previous sub-section studied the impact of the variables operational cash flow and financial leverage. The results showed that the variable financial leverage had a negative impact. This sub-section will investigate whether this result is the same for all European oil companies. To investigate whether there are differences between oil companies, a distinction is made between high levered and low levered companies. Table 14 shows the regression results.

(31)

The results also show that there is an important difference between the financial leverage effect on high levered and low levered companies. Financial leverage has a negative significant impact on the stock price return of high levered oil companies and an insignificant impact on low levered oil companies. This result corresponds with the possible explanation given in sub-section 5.5 that a high levered oil company can have difficulties in issuing new debt. As a result of this oil companies have no resources to invest in renewals and in finding new reserves.

Table 14 High vs. low levered oil companies

High levered

Constant 0.1313 0.0429 3.0619* 0,0034

Market return 0.2316 0.0447 5.1861* 0,0000

Oil price 0.0413 0.0395 1.0465 0,3000

Exchange rate -0.0638 0.1218 -0.5235 0,6028

Interest rate -0.0242 0.0311 -0.7782 0,4399

Operational cash flow 1.91E-05 1.01E-05 1.9014* 0,0626

Financial leverage -0.1367 0.0198 -6.8896* 0,0000

Cross-sections included: 15 F-statistic 18.6636

Total panel (unbalanced) observations: 75 Prob. F-statistic 0.0000 Probability Jarque-Bera statistic 0.2210 Adjusted R-squared 0.6623

Low levered

Constant 0.0932 0.0739 1.2619 0,2139

Market return 0.3016 0.0812 3.7122* 0,0006

Oil price 0.1436 0.0743 1.9346* 0,0598

Exchange rate 0.1879 0.2991 0.6283 0,5332

Interest rate -0.0728 0.0512 -1.4222 0,1624

Operational cash flow 1.34E-05 7.87E-06 1.7048* 0,0956

Financial leverage -0.2175 0.1372 -1.5855 0,1204

(32)

Table 15 Regression eight-factor model

Constant Market return Oil price Exchange rate Interest rate Operational Cash flow

Financial leverage Production Reserves

0.0855 0.2524 0.0531 0.0798 -0.0252 7.09E-06 -0.1551 3.57E-05 -1.99E-06 (0.9025) (3.2687)* (1.0549) (0.5151) (-0.7555) (0.9101) (-2.9162)* (0.1194) (-0.2284) 0.0688 0.2606 0.0603 0.0999 -0.0381 8.18E-06 -0.1557 3.90E-05 (0.8055) (4.9773)* (1.2679) (0.6474) (-1.1776) (1.1509) (-3.5389)* (0.2255) 0.0471 0.2739 0.0713 0.0882 -0.0473 1.15E-05 -0.1516 1.12E-06 (0.6809) (5.0547)* (1.5456) (0.5756) (-1.5055) (1.7346)* (-3.3697)* (0.2298) 0.1446 0.2595 0.0502 0.1151 -0.0169 -0.1548 3.92E-05 -2.75E-06 (1.8574)* (3.3085)* (0.9855) (0.7497) (-0.5294) (-2.8896)* (0.1198) (-0.2979) 0.1379 0.2732 0.0574 0.1401 -0.0295 -0.1527 2.60E-05 (1.9931)* (5.2134)* (1.1946) (0.9157) (-0.9286) (-3.4583)* (0.1257) 0.1455 0.2834 0.0611 0.1396 -0.0341 -0.1503 -2.53E-07 (3.4936)* (4.5855)* (1.2315) (0.8900) (-1.0446) (-3.2777)* (-0.0413) The table reports the following regression:

Rit = α + β1RMt + β2ROILPt + β3RERt + β4RIRt + β5OCFit + β6FLi t + β7Prodit + β8Resit + έt

whereRit denotes stock return of oil company i in quarter t, RMt is the quarterly market return, ROILPt is the quarterly change in the oil price, RERt is the quarterly change in the exchange rate, RIRt

is the quarterly change in the interest rate, OCFit is the value of the operational cash flow of company i in year t, FLit is the value of the financial leverage of company i in year t, Prodit is the

value of the number of produced oil barrels of company i in year t, Resit is the value of the quantity oil reserves in barrels in year t and έt is the error term. Timeframe is 2000-2004. Econometric

(33)

5.6 Analysis of common and company specific factors

This sub-section investigates the impact of the four common variables and the four company specific variables. Table 7 in section 4 showed that there is multicollinearity between the variables operational cash flow, production and reserves. To limit the impact of the multicollinearity on the regression results several different regressions are run. The results of these regressions are presented in table 15. Furthermore, it should be noticed that all the regression in table 15 contains autocorrelation.

Table 15 shows that the results for the common variables market return, exchange rate and interest rate correspond to the results in table 13. The market return has a positive significant impact and the exchange rate and interest rate have no impact on the stock price return. Interesting is the insignificant impact of the common variable oil price. This result corresponds with the insignificant impact of the oil price on high levered oil companies.

(34)

6. Conclusions

This paper studied the stock price exposure of European oil companies. Common and company specific factors are used to explain the stock price exposure. These variables originate from earlier studies related to stock price exposure and the findings of these studies are used to support this paper. The data of 50 listed European oil companies are collected. Multifactor models are used to investigate the relationship between the explanatory variables and the stock price return in the period 2000-2004.

The paper firstly studied the impact of the common variables on the stock price return. A two-factor model is used to investigate the impact of the common variables market return and oil price. To investigate the impact of all the common factors a four-factor model is used with the variables exchange rate and interest rate included. The period 2000-2004 is divided in two periods to investigate whether there are differences in the impact of the common variables in both periods. Further, the impact of the common variables on the three sub-sectors petroleum extraction companies, oil service companies and petroleum manufacturing companies is investigated.

As expected the variables market return and oil price have a significant positive impact on the stock price return of European oil companies. Interesting is that the market return coefficient of the oil companies is smaller than one and this indicates that European oil companies are less risky than the market. The results for the sub-sectors further show that the oil price effect is the strongest for petroleum extraction companies. The impact of the variables exchange rate and interest rate is not always clear. No impact is found for both variables when quarterly data are used. When yearly data are used a significant positive impact is found for the exchange rate and a negative impact for the interest rate. When the data are divided in two periods, only a significant positive impact is found for the exchange rate in the first period. Hedging activities could be the reason for the insignificant impact in the second period. The variable interest rate shows two different impacts. In the first period a significant negative impact is found and in the second period a significant positive impact. Finally, when the data are divided in three sub-sectors a significant positive impact is found only for the exchange rate on the stock price returns of oil service companies.

(35)

An interesting result is that with the inclusion of company specific variables the common variables exchange rate and interest rate are no longer significant. The variable operational cash flow has a positive significant impact and financial leverage a negative significant impact on stock price return. The negative impact of financial leverage was unexpected. A reason for this effect could be that several companies are highly levered and for these companies it is difficult to issue new debt when it is needed. When the sample is divided in high and low levered companies this appears to be correct. For high levered companies a negative impact is found for the variable financial leverage and for low levered companies no impact is found. Further, no impact is found for the variables production and reserves.