Oil sensitivity and systematic risk in oil-sensitive stock indices

Abstract

This paper has two primary objectives. First, is to examine and compare the oil sensitivity of equity returns of non-Gulf, oil-based countries (Mexico and Norway) with that of two major oil-sensitive industries (US oil and transportation industries). Second, is to examine and compare the oil sensitivity of those returns with their sensitivity to systematic risk with respect to the world capital market. The findings suggest that the oil price growth leads the stock returns of the oil-exporting countries and the US oil-sensitive industries, with the US oil industry showing the greatest sensitivity. The results also indicate that investors view the systematic risk more importantly than the oil sensitivity in pricing those oil-sensitive returns, regardless of the direction of the world capital market.

Introduction

The empirical research in International Finance has focused on the sensitivity of the stock markets of the major oil-consuming countries to changes in the oil price. The sensitivities to oil price growth¹ of stock index returns of the world's major oil-sensitive industries and the oil-exporting countries are not well researched in this financial literature. In particular, the jury is still out on the extent of oil price sensitivity of the stock markets of the oil-exporting countries whose governments own the oil industry, which thus has no stocks traded on stock exchanges. If sensitivity to the oil price changes is found to be significant for the individual country/industry and for all of them as a group, then it will be interesting to compare this sensitivity to the inherent systematic risk defined with respect to the world capital market. In this case, we seek answers to the following questions: (1) how sensitive to oil price growth are the returns of the composite stock indices of the oil-exporting countries, which have zero oil industry equity capitalization? (2) how does the oil sensitivity of those returns compare with that of the returns of oil-related and oil-sensitive industries such as the oil and the transportation industries whose stocks are traded on the world's major stock markets? (3) does oil price growth matter more than the world capital market return in influencing the oil-sensitive stock returns? Or does oil sensitivity matter more than systematic risk?

Hammoudeh and Eleisa (2004) investigated the oil sensitivity of the stock markets (without including the systematic risk) for five members of the oil-rich Gulf Cooperation Council (GCC). These members are Bahrain, Kuwait, Oman, Saudi Arabia and UAE which, all except Bahrain, are major oil exporters and have oil-anchored economies. They found that on a daily basis only the Saudi stock market has a bi-directional causal or mutual predictive relationship with oil price growth. They, on the other hand, found that the stock returns of the smaller oil exporters, Kuwait and Oman, have no causal relationships with oil price changes. These results are puzzling because the GCC countries are strongly oil exports-anchored and are similar in their economic structures. Do the results have something to do with the fact that the GCC economies are very small and their stock markets are relatively unknown?

The above study has motivated us to extend this type of research to examine other oil-exporting countries that have larger economies and more well-established stock markets than the GCC countries. We also included two major oil-sensitive industries. We hope to find more meaningful directional relations or sensitivity between the oil price growth and the oil-sensitive stock returns. In this extension we are limited by several factors including daily historical data availability, relatively well-established equity markets or industries and larger, more diversified economies. The possible country candidates that fit these criteria include Indonesia, Mexico, Norway, Russia and Venezuela.² The two major oil-sensitive industries include the US oil industry and the transportation industry. The daily sample period considered in this paper is the period 1986–2003. This sample factor excludes Russia's market because its daily data starts from 1995 only.³ Venezuela cannot also be included because we could not find consistent daily data that fits our lengthy sample, and Indonesia will also be excluded because its stock index is stationary in (log of) level. We hope that examining the oil/stock link for the oil exporters, Mexico and Norway, and the oil-sensitive oil and transportation industries would balance the results found by Hammoudeh and Eleisa (2004), and would also further our understanding of the behavior of other oil exporters’ stock returns such as those of Indonesia, Russia and Venezuela. These results may also have general information on whether the oil-sensitive stock markets have a forecasting power of daily oil prices, and on whether the systemic risk matters more than the oil price sensitivity.

At the industry level, we examine the oil sensitivity of two US industry stock indices: one for the oil industry and the other for the transportation industry. These two indices are, respectively, represented by the Amex Oil Index and the NYSE Transportation Index. The oil industry stock index includes companies engaged in oil exploration, drilling, production and refining. The transportation industry includes airlines, trucking and railroad companies. In examining the oil price–stock linkages for those two industries, our main objective is to determine whether or not these industries have stronger links with oil prices than do the national composite stock indices of the oil-exporting countries. The national indices considered in this paper have virtually zero-oil equity market capitalization because the oil industry is owned by governments.⁴ Moreover, we are interested in ascertaining whether the oil industry, which depends heavily on crude oil as its major input, has a stronger predictive power or causal relationship with the daily oil prices than the transportation industry, which has stronger links with the over-all economy or vice versa.

The other primary objective of this paper is to compare, through the use of an international APT model, the oil price growth sensitivity with the systematic risk relative to the world market return when this market is aggregated and when it is in up and down modes. Based on the two broad objectives, we can formulate two basic hypotheses that set the stage for our analysis.

Hypothesis 1

In a system of oil-based stock returns, there are causal or predictive relationships running from the oil price changes to the oil-sensitive stock returns on a daily basis.

This hypothesis will be examined using a vector error–correction model. If in this hypothesis we find (which is the case) that for most of the stock returns there are directional relationships running from the oil price to the stock returns, then the oil price is an important factor in determining the oil-sensitive stock returns as a group. We will then formulate a factor model that compares the returns’ oil sensitivity with their systematic risk in both specifications of the world capital market. In this case we will test a second hypothesis.

Hypothesis 2

The oil-based stock markets are on a daily basis more sensitive to changes in the world oil price than to the changes in the world capital market index price regardless whether the world market is in an up or down mode.

This hypothesis will be tested through the use of international APT models.

As a result, this paper will help to fill the gaps in the VAR and APT literatures on the sensitivities of the oil-based stock markets. Since the period under consideration is marred by international events, this study will also examine the impact of three well-known political and economic events on those returns.

The main findings of this study can be summarized as follows. (1) Oil price growth leads the stock returns of all the individual oil-exporting countries and the US oil-sensitive industries. Moreover, the world capital index MSCI was found to simultaneously have a particularly strong predictive power of these stock returns, pinpointing the importance of the systematic risk with respect to the world capital market. These results which show long-run relationships among the variables pave the ground for using the factor model. This model compares the sensitivity of the national and industry stock returns to the oil price changes with their sensitivities to the own market risk for the specifications of the standard aggregate world market and its up and down markets as stated in Hypothesis 2. (2) Based on the extent of returns’ positive dynamic sensitivity to oil price growth, investors interested in the oil-sensitive stocks should invest first in the stocks of the US oil industry and then in the Mexican stocks before they invest in the Norwegian market to take advantage of higher oil prices. (3) Based on the dynamic sensitivity to the world capital market, investors should also invest first in the US oil industry followed by the US transportation industry, Norway and then Mexico. (4) Similar to Saudi Arabia in Hammoudeh and Eleisa (2004), only Norway and the world capital market index have a lead relationship with the oil price, with Norway having a positive 1 day lead and MSCI showing a negative 1 day lead. (5) All the stock returns display sensitivity to major economic and political events, with Norway showing the strongest relative sensitivity. (6) Aggressive investors and aggressive growth portfolio managers interested in oil-sensitive stocks with higher returns may be inclined to buy stocks that belong to industries or countries with higher betas (systematic risks). Among the returns considered, the US transportation and oil industry have the highest betas. However, those investors should be willing to accept abnormal losses in the down market if they invest in those oil-sensitive markets.

The organization of this paper is as follows. After this introduction, Section 2 presents a brief review of the literature that used the VEC and International APT models to study the oil-sensitive stock markets. Section 3 describes the oil and financial series used in carrying out this study and also presents their descriptive statistics. Section 4 discusses the methodology and empirical results for the unit root tests, the cointegration tests and the error–correction models. Section 5 compares the stock returns’ oil sensitivity to the systematic risk by the means of using a factor model. Section 6 concludes the paper and offers recommendations.