The impact of energy price fluctuations on stock returns in the renewable
energy industry
Patrick Klink (s2016680) University of Groningen
Abstract
The renewable energy industry is a fast growing industry, which makes this industry important for the energy sector now and in the future. This paper examines the impact of fluctuations in electricity, oil, natural gas and coal prices on the stock returns of renewable energy companies. A fixed effects panel least squares regression test is used to estimate this effect. The results of the test are inconclusive. Oil, natural gas and coal prices do have impact on the stock returns of renewable energy companies when the whole sample is studied. However, when the sample in split up and tested for the individual years the test results do not provide a strong relationship. Electricity prices does not have significant impact on stock returns. The results found in this study are not a characteristic of the renewable energy industry, since a paired sample of conventional companies shows comparable results.
JEL classification: G12, Q42,
Keywords: Renewable energy industry; Stock returns; Energy stocks, Asset pricing
Master thesis International Financial Management December 20, 2014
Student: Patrick Klink
1. Introduction
Energy is becoming more and more important in our current societies. The world energy demand is expected to increase with 56% in the next 30 years (International energy outlook 2013, p. 9). Extractible sources of fossil fuel will be exhausted in 2050 (BP statistical review of world energy, 2007). This increase in scarcity of this type of energy and raises the need for renewable energy. There are a number of possible renewable energy sources. The most commonly used are: solar power, wind energy, biomass, hydroelectricity and geothermic energy. Although the renewable energy sector is relatively young, the commercialization of renewable energy sources has accelerated dramatically since 2000 (Kammen, 2006). The renewable energy industry is the fastest growing industry worldwide over the past 20 years. In recent years renewable energy is also becoming important for another reason. Excessive use of fossil energy sources caused a rise in CO2 levels in the atmosphere. This is believed to be a reason for the acceleration of global warming (Cox et al., 2000). Policymaker try to control the CO2 emission by making international agreements and international policies, like the Kyoto treaty. In this manner, countries are forced to reduce their CO2 emission and incorporate CO2 reducing policies or search for alternative energy sources (Mirasgedis et al., 2002).
Energy prices are also an important factor for the increasing interest in renewable energy. With high economic growth in counties like China, India and Brazil, the world energy needs rise. With increasing scarcity and higher production costs, energy prices have risen in the past and are expected rise in the future. These higher prices create an opportunity to substitute fossil energy for renewable energy. The substitution of fossil fuel for renewable energy sources is expected to happen. Renewable energy sources are expected to account for half of the growth in total energy consumption worldwide (IEA: World energy outlook, 2013 p. 5). Also international CO2 treaties support the trend of substituting fossil fuel use for renewable energy. In order to bring CO2 emission down, individual countries make legislation to achieve their target CO2 emission. Looking at individual industries per country, most reduction is made in the power industry (IEA: World energy special outlook, 2013, p. 36). This implies that governments support the use of low CO2 emitting fuels, like renewable energy and reduce high CO2 emitting fuels. Over the past decade energy prices, like for example the crude oil price, have fluctuated heavily. Until the end of 2007 the oil price experienced a sharp increase. With the start of the financial crisis the oil price sharply decreased, after which it steadily increased again. Fluctuations in energy prices can have direct and indirect impact on the global economy. An unexpected sizeable downshift in the oil supply trend of 1%, which results in higher oil prices, slows annual global economic growth with approximately ¼% (IMF: World economic outlook analysis, 2011).
fluctuations in four energy sources, on the stock returns of renewable energy companies. The following research question is used:
What is the effect of fluctuations in electricity, oil, natural gas and coal prices on the stock returns of companies in the renewable energy industry?
There are multiple definitions for renewable energy. In this paper the following definition is used:
"Renewable energy: Any energy resource that is naturally regenerated over a short time scale and derived directly from the sun (such as thermal, photochemical, and photoelectric), indirectly from the sun (such as wind, hydropower, and photosynthetic energy stored in biomass), or from other natural movements and mechanisms of the environment (such as geothermal and tidal energy). Renewable energy does not include energy resources derived from fossil fuels, waste products from fossil sources, or waste products from inorganic sources." (www.treia.org/renewable-energy-defined)
Note that in this definition there is no room for nuclear power as a renewable energy. Policy makers do not agree with each other on the issue if nuclear power can be considered as a renewable energy source. Proponents of this issue argue that nuclear power produces no or little greenhouse gasses, which is the main characteristic of renewable energy sources. Therefore, they consider nuclear power as a renewable energy source. Whereas, opponents argue that nuclear power produces harmful byproducts and relies on extractive industries to produce fuel like uranium (Kanter, 2009). In line with the last argument, in this paper nuclear power is not considered as a renewable energy source.
The renewable energy industry is a relatively young sector. Therefore, fewer research is done on this industry as for example the oil industry. Especially literature about drivers of stock price fluctuations is not much researched. Yet, the importance of this industry is growing rapidly. In this paper the effect of price fluctuations in four energy sources, on the stock returns of renewable energy companies in studied. The energy sources are electricity, oil, natural gas and coal. The effect of oil price fluctuation on oil producing companies is already research (Boyer and Filion, 2007; Nandha and Faff, 2008). Results show a positive relationship between oil price movements and stock prices of oil producing companies. Also the effect of oil price movements on renewable energy companies is researched (Henriques and Sadorsky, 2008). In this study a positive relation is found between oil price movements and the stock returns of renewable energy companies. Similar results are expected for fluctuations in electricity, natural gas and coal prices on stock returns of companies in the renewable energy industry.
the first who studies the effect of price fluctuations in four energy sources simultaneously on stock returns of companies in the renewable energy industry. Further, it examines differences over time, differences in company size and compares US with non-US companies. An adapted model from Sadorsky (2001) is used in this study. In his model the common risk exposure variables interest rate and exchange rate are included. This study is the first to includes four energy prices in one model in order to investigate their impact on stock returns. It further adds to the literature by specifically looking at the renewable energy industry. Finally, The results of the study are compared with the results of a sample of conventional companies in order to see if these two samples react significantly different to price fluctuations in the four energy sources.
The structure of the remainder of this paper is as follows. In section two the literature on the four energy sources is provided. In section three the hypothesis are discussed. Section four provides the methodology and introduces the model, hereafter the data for each variable are discussed in section five. The results are reported and discussed in section six. Section 7 concludes the paper.
2. Literature review
This section provides a review of the literature on the interaction between renewable energy companies and electricity, oil, natural gas and coal prices. This section is organized as follows. First, the renewable energy industry is defined and a few assumptions about this industry are made. Second, literature about electricity prices is reviewed. Third, literature about nonrenewable energy sources (i.e. oil, natural gas and coal) is provided. Finally, literature about other company risk factors (size, exchange rate, interest rate) is reviewed.
2.1 The renewable energy industry
Renewable energy can be generated in many ways. The most important sources for renewable energy are, biomass, geothermal power, hydroelectricity, wind energy, solar power, wave power and tidal power. Different types of energy can be generated from these sources. Biomass produces biogas which can be used for the heating of buildings or as fuel for a car. Furthermore, geothermal power is also used as a heating source for buildings. The other renewable energy sources produce electricity with the use of turbines. Nowadays, all of the above mentioned sources produce mainly electricity. Biomass for example accounts for 0,5% of the electricity production in the USA nowadays. For this reason, it is assumed in this paper that all producers of renewable energy produce electricity and no other type of energy.
As mentioned earlier, the renewable energy industry is a relative young industry. This limits the amount of literature on this industry. There is however literature available about increases in renewable energy consumption (Sadorsky, 2009), and about the effect of an increase in the share of renewable energy among the total energy supply in a country (Chien and Hu, 2007). These studied find that the major drivers behind adopting a higher share of renewable energy in the energy mix are a long term increase in GDP and an decrease in CO2 emission per capita. After the increase in renewable energy use countries experience a rise in technical efficiency. These studies do not relate energy prices with stock returns of renewable energy companies. In a study, which does relate the stock returns of renewable energy companies to oil prices (Henriques and Sadorsky, 2008), a positive relation between oil prices and renewable energy stock returns is found. Though, a stronger effect is found for shocks in technological stock prices.
2.2 Electricity prices
Electricity is generated in multiple ways. The resources for the worldwide production of electricity are: coal (40%), natural gas (22%), petroleum and other liquids (5%), nuclear (12%) and renewable sources (21%) in 2010 (International energy outlook 2013, p. 94). In the future, the total consumption of electricity is expected to rise with 2.2% annually. Further, the share of each resource in the electricity mix is expected to change. Renewable sources and natural gas become more important, with an expected rise of 2.8% and 2.5% annually until 2040. On the other hand, petroleum and other liquids use is expected to decrease with 1.1% annually.
correlation between the electricity price and the S&P 500. Oil is the commodity which is the most related to the market. This result is in line with other research. Kristoufek and Vosvrdra (2014) researched the efficiency of commodity markets. Their result show that oil as an individual commodity and energy as a group are the most efficient commodity markets. All publicly and privately available information is already contained in the market prices.
For the research in this paper the constraint of high volatility does not seem to be a problem since the correlation between electricity prices and the stock market tend to stay relatively stable. Further, electricity prices reflect all available information in the commodity price. This leaves no constraints to incorporate electricity prices into the model of this paper.
2.3 Oil prices
Financial literature about nonrenewable energy price volatility (especially oil) is widely known. The effect of oil price volatility on the stock market is researched in many papers. These papers find empirical evidence that an increase in the oil price has a negative effect on the stock market (Sadorksy, 1999, 2001; Park and Ratti, 2008; Cunado and Gracia, 2014; Asteriou et al. 2013). An explanation they give is that higher oil prices decrease the future cash flows of a company. Since firm value is determined by its discounted future cash flows, an increase in oil price reduces future cash flows and therefore also firm value.
Other research on oil price volatility finds that an increase in oil prices has a positive effect on stock prices of companies in the energy industry. Boyer and Filion (2007) studied the effect of an appreciation of crude oil and natural gas prices on Canadian oil and gas companies. They find that an appreciation of the oil and natural gas prices has a positive effect on stock prices of companies in this sector, with growth of internal cash flows and proven reserves. Nandha and Faff (2008) show that a rise in the oil price has a negative impact on equity returns for all sectors, except for the mining, oil and gas industries. For these sectors a positive impact has been found.
In the paper of Henriques and Sadorsky (2008) an empirical analysis of the relationship between oil prices and the stock returns of renewable energy companies is studied. Their research finds that movements in oil prices, technology stock prices and interest rates all have some power in explaining the movements of the stock prices of renewable energy companies. The theory behind movements in oil prices is that due to a growing concern about global warming and local air quality, a rising oil price should help to spur greater demand and supply of renewable energy. However, investors see renewable energy companies as potential disruptive technology provides. While potential returns from investing in the renewable energy industry are high, so are the associated risks.
2.4 Natural gas prices
Natural gas has compared to oil and coal a low CO2 emission. This makes it popular for countries that are implementing policies to reduce their CO2 emission. Partly for this reason is natural gas the fastest growing fossil fuel worldwide. Natural gas consumption worldwide is expected to grow with an average of 1.5% per year until 2040. In the electricity generation sector the consumption of natural gas is expected to grow with 2.5% annually. In the past years, cross border trade of natural gas increased which made the market for natural gas a more global market (International energy outlook 2013, p. 41).
The effect of natural gas price fluctuations on stock returns is not widely studied. As mentioned earlier, Boyer and Filion (2007) studied the effect of an appreciation of oil and gas prices on the stock returns of Canadian oil and gas companies. They find a positive relationship between these variables. Oberndorfer (2009) focuses on the linkage between energy prices and energy stock prices. Their findings suggest that natural gas prices do not seem to have a relationship with energy companies’ stock prices. The study of Acaravci et al. (2012) do report a relationship. They focus on 15 European countries and finds for most countries a long-term indirect relationship among natural gas prices and stock prices. Although there is no strong evidence of a relation between natural gas prices and the stock market, earlier research at least raises the possibility for a relationship. Considering the fact that natural gas is the fastest growing fossil energy source and the possibility of a relationship with the stock market, natural gas prices are included in the model of this paper.
2.5 Coal prices
Coal is the second largest energy source in the world. Despite many policies and protocols to reduce greenhouse gas emission, the world coal consumption is expected to rise with an average rate of 1.3% per year between 2020 and 2040. Especially China, India and other non-OECD countries will increase their coal consumption significantly. Worldwide coal consumption is concentrated in China (47%), the United Stated (14%) and India (9%) which account for 70% of the worldwide coal consumption (International energy outlook 2013, p. 67). 60% of the coal consumption is used to generate electricity. This 60% accounts for 40% of the total electricity production. Other than oil, the countries that consume a substantial amount of coal also have large coal reserves. This keeps cross border trade low and makes the world market for coal smaller than the world market for oil (International energy outlook 2013, p. 78).
prices of coal companies show a positive relationship between the price of coal and the return of Australian coal companies (Hasan and Ratti, 2014). They study a number of risk factors of coal companies in Australia. The main result of their study is that a 1% increase of the coal price leads to an increase in return between 0.15% and 0.17% for a coal company.
Earlier research found a direct relationship between the coal price and coal producing companies. This might imply that the coal price has an influence on the stock returns of other energy companies, like oil has. Coal account for 40% of the total electricity production and the coal market can be considered as a global market. This justifies the inclusion of coal price fluctuations into model.
2.6 Common risk factors
Exposure to energy price fluctuations is not the only exposure companies in the renewable energy sector face. The renewable energy sector is a young and fast growing industry. This makes it a very risky industry, with a high exposure to macroeconomic variables. Therefore, common exposures factors need to be taken into account. One of these factors is company size. According to the study of Fama and French (1993), small companies experience a slightly lower return on assets than large companies. This negative relation is also found by Keim (1983) who studied the effect of company size throughout the year. He found that size is always negatively related to abnormal returns. Chan et al. (1983) found that the negative relation between company size and stock returns hold between different subperiods.
The second exposure factor is exchange rate. The return of an international company that has a large percentage of sales abroad can experience significant impact from exchange rate fluctuations. It can experience impact on his accounts receivable from foreign customers, on his account payable to foreign supplier and on intercompany sales. Multiple studies show a negative relation between the stock returns and the exchange rate (Sadorsky, 2001; Louden, 1993; Khoo, 1994). Jorion (1990) find a positive relation between the degree of exposure a US company face and the percentage of foreign sales.
The third control variable is interest rate. The stock return a company has, depends on the opportunities it has to invest and grow. The costs of financing these investments depends on the interest rate at which a firm can borrow money. High interest rates can reduce profit and reduce return on assets. Many researcher prove this negative relation between interest rates and stock returns (Marin and Keown, 1977; Faff and Chan, 1998; Stone, 1974). Sadorsky (2001) emphasizes the importance of the interest rate in time of capital improvements. When companies are upgrading their operations or planning to expand, the impact of the interest rate on stock returns is higher.
exposure to exchange rates and in interest rates. This study controls for these risk exposures in order to filter out possible biases from the main research. This study also controls for exposure within the sample. Following the three factor model of Fama and French (1993) this study controls for differences in company size. Small companies are expected to have a lower return on their assets. This might result in differences in exposure to energy prices.
3. Hypothesis
In this paper the relationship between energy prices and the stock returns of companies in the renewable energy sector is researched. As described before in section 2, changes in oil prices have significant effect on stock returns (Sadorksy, 1999, 2001; Park and Ratti, 2008; Cunado and Gracia, 2014; Asteriou et al. 2013). Further, an increase in oil prices has a positive effect on oil and gas producing companies (Nanda and Faff, 2008), but also a positive effect on companies in the renewable energy sector (Henriques and Sadorsky, 2008). This last relation can be explain by considering renewable energy as a substitute for oil. When oil prices increase, renewable energy becomes relatively cheaper. Cheaper renewable energy increases the demand and could increases stock returns for companies in this industry. This relationship can also be applied to other energy resources. Natural gas and coal are also resources for electricity production. An increase in natural gas and coal prices can also cause an substitution effect to renewable energy. On the other hand, the end product of the renewable energy industry is electricity. A rise in electricity prices results in more revenue for electricity producers. Higher revenue results in a higher stock price. Because of this complementing relation, and the fact that the energy market is an efficient market (Kristoufek and Vosvrdra, 2014), a relationship between electricity prices and returns of renewable energy companies is expected. In order to empirically test if there are relations between energy prices and stock returns of renewable energy companies, this study tries to answer the research question mentioned in the introduction. To empirically test the research question two hypotheses are used. The first hypothesis deals with fluctuations in electricity prices. Based on basic financial knowledge on asset pricing, I expect higher stock returns for renewable energy companies when electricity prices rise. Investors value a firms assets by the cash flows these assets generate in the future. A rise in electricity prices increases this cash flow and thus increases the asset value of renewable energy companies. Furthermore, in the oil industry a relationship between oil prices and returns of oil companies is already proven. Boyer and Filion (2007) found in their study, a rise in oil and gas prices increased the stock returns of Canadian oil and gas companies. I expect the same for the renewable energy industry when the electricity price rises.
The second hypothesis tests whether the nonrenewable energy sources have an effect on the stock return of renewable energy companies. The nonrenewable energy sources which will be tested are oil, natural gas and coal. The effect of oil has already been tested by Henriques and Sadorsky (2008). They find a positive relation between renewable energy stock returns and the oil price. Renewable energy stock returns are explained by past movements in oil prices Therefore, I also expect a positive relation between these variables.
Coal is the second largest energy source in the world. Coal is important in the total energy production worldwide. A rise in the price of coal increases the production cost of coal fired power stations. Electricity generated with coal is the same as electricity generated with renewable sources. Because of this substitution effect, renewable electricity is becoming relatively cheaper when coal prices rise, which should lead to an increase in demand and stock return of renewable energy companies.
The third nonrenewable energy source I am testing is natural gas. Natural gas is in many countries the fuel of choice for producing electricity (International Energy Outlook, 2013). The expected relationship between natural gas and the renewable energy industry is the same as for coal. A rise in gas produced electricity makes electricity generated with a renewable source relatively cheaper. This leads to more demand in renewable energy and higher stock returns of renewable energy companies. Putting it all together, I expect a positive relationship between oil, coal and natural prices and the stock returns of companies in the renewable energy industry.
: Fluctuations in oil price, natural gas price and coal price relate positively with the stock returns of companies operating in the renewable energy industry.
4. Methodology
The effect of energy prices on the stock return of renewable energy companies is estimated with a multifactor model. The model is an extended CAPM model and is also used by Sadorsky (2001). In his study, only one energy source is included in the model to determine stock returns. The study in my paper tests the effect of four energy sources on stock returns. Therefore, the model of Sadorsky is adjusted to fit my research. In order to test for the effect of energy prices on renewable energy companies’ stock return, the imputed energy variables are the returns on energy prices. The returns are the percentage change at time t compared to time . This is calculated as:
the intercept to differ cross-sectionally but not over time, while all of the slope estimates are fixed both cross-sectionally and over time. In this paper the following model is used:
A detailed description of each variable can be found in Appendix 1. In this model , are the energy betas. A significant beta indicates that price fluctuations in the corresponding energy source have significant impact on the stock returns . The value of the beta indicates the percentage change in when the corresponding energy price increases with 1%. All four energy sources are expected to have a positive relationship with the dependent variable. This is based on earlier research of, among others, Henriques and Sadorsky (2008), Boyer and Filion (2007) and Hasan and Ratti (2014). Following Sadorsky (2001) the exchange rate and interest rate are included in the model. According to Sadorsky, petroleum production companies are price takes. This means that they do not pursue risk management. Renewable energy companies are expected to behave the same which leads to exposure to exchange rates and interest rates.
independent variables are characteristics of the renewable energy industry or also present in other industries. To test this question a second sample of conventional companies is selected. Companies are selected on their SIC code, size and country of origin.
5. Data
For this study, a sample of publicly traded companies, who operate in the renewable energy sector is needed. Therefore, this study draws its sample from the Wilderhill Clean Energy index. This sample was also used in other recent studies (Sadorsky, 2012; Henriques and Sadorsky, 2008). The Wilderhill Clean Energy index consist of the 53 largest renewable energy companies worldwide. Appendix 2 provides an overview with the company names per industry. The Wilderhill Clean energy index was chosen out of a list of other indices who represent the renewable energy industry. Ortas and Moneva (2013) provide an overview of these indices. Although a large number of indices, only a few indices
where useful for this study. Most of the indices do not represent the global market or just focus on a segment (like for example clean energy technology companies). Out of the few useful indices the Wilderhill clean energy index was selected, because it is the oldest and largest index. For a company to be included in the index, it needs to have significant exposure to clean energy, contribute to the advancement of clean energy, or be important in developing clean energy. The majority of companies in the index has a market capitalization of more than 200 million and no individual company may exceed 4% of the total index weight. The Wilderhill Clean Energy index was created in March 2005 and it initially started with 24 companies. Due to the rapid growth of the industry more companies matched the criteria to enter the index. These companies where added one by one. None of the companies were removed during the sample period. The low number of companies at the beginning of the sample and the inclusion of 29 companies during the sample period, is a problem in this study. There is no complete data availability throughout the whole time period. This could lead to different relations between the sample in different year. In order to solve this problem, the main research is done for each individual year. The sample period in this study will be from March 2005 until March
Country Number of companies
USA 39 Canada 4 China 6 Chile 1 Germany 1 Brazil 1 Israel 1 Total 53
Table 1: Countries of origin able 2
2014, weekly data is used. This results in nine years that are separately studied. I use weekly data since they provide a good compromise on the issue of daily data and relatively short span monthly data. Daily data can be noisy because of normality issues, cause an increase in variance around events and daily data can exhibit serial dependence (Brown and Warner, 1985). Monthly data is also not optimal, because of the relatively short time period the number of observations would be low, this reduces significance levels. This research has a sample period of 474 weeks. During this time period companies where added to the index. Therefore, not every company has 474 observations. This led to a total of 19.070 individual stock price observations.
One of the criteria of the Wilderhill index is that a company must be listed in the US. However, this does not imply that the whole sample consist of ÙS companies. Table 1 shows the countries of origin for the companies in the Wilderhill Index. Initially the country of origin was determined by looking at the country where the company is officially registered. However, a number of companies is registered in countries which are considered to be tax heavens, for example Bermuda and the British Virgin Islands. Therefore, the country of origin is determined by looking at the establishment of the headquarter. As table 1 shows, the majority of companies has its headquarters in the US. Other companies are mainly from Canada and China.
The most important data source in this study is Thomson Reuters DataStream. Company specific stock returns are calculated from stock prices downloaded from DataStream. Also all the energy prices are downloaded from this source. The sources of the other variables can be found in appendix 1. Stock returns are measured on a weekly basis by the stock returns of company i at time t. The market return ( is measured by the continuously compounded return on the S&P 500 (source: yahoo.fincance.com). Following Henriques and Sadorsky (2008) the risk free rate ( ) is measured by the yield on a 3 month US T-Bill , (source: quandl.com). These returns are converted into weekly returns using the simple interest method. The following formula is used (www.math.hawaii.edu):
of natural gas abroad has the world market for natural gas physically linked to the North American market. Therefore, Asian companies are linked to the Henry Hub price. Coal is not frequently traded across countries because of high transportation cost. This results in differences in coal returns around the world. Large differences and opposite price fluctuations could exist across continents. The correlation matrix (table 4) confirms that the coal price in Europe (HWWI) and the coal price in the US (NYMEX coal index) are not correlated. A problem that could occur when relating Asian
companies to European or US coal prices are possible wrong beta coefficients and wrong probabilities. The sample in this paper includes multiple Asian companies, therefore a Chinese coal index is added. Within DataStream there are not many Asian coal indices available who covers the whole sample period. Therefore the China coal a-ds index selected. This index matches the criteria. The same problem coal has is also applicable to natural gas. The European price (UK NBP) and US price (Henry Hub price) are not correlated with each other. Therefore, an Asian index is added. A detailed description of each energy source is provided in appendix 1.
Table 3 presents the descriptive statistics of all variables in the model. Note that the descriptive statistics in this table are calculated as continues compounded returns. Continues compounded returns reduce the average values (compared to simple returns), since continues compounding assumes a evenly distributed return over the period and it calculates the returns according to the effective interest method.
On average there is a negative weekly return on stocks (-0,0013%). This result is consistent with the information provides by www.invesco.com. This website shows that the Wilderhill clean energy index has on average a negative return over the last 5 year. Only since the last year the index had a positive return. Henirques and Sadorsky (2008) find in their research a positive weekly return for the renewable energy companies (0,020%). An explanation for this different result might be the different time periods these studies were conducted. Compared to the market the return of companies in the Wilderhill index is substantially lower (0,090%). Also the distribution is much wider with a maximum return of 161% and a minimum return of -65% for the sample and 11% and -20% for the market. Among the energy sources there is not a large distribution of average returns. The Asian gas return has
The price indices used per energy source for each continent
Energy source
Continent Oil Natural gas Electricity Coal
Europe Brent oil index UK NBP (National Balancing Point) EEX (European Energy Exchange) Phelix Base HWWI Coal Eurozone EUR US West Texas intermediate
Henry Hub price
(NYMEX) PJM
NYMEX coal index
Asia Brent oil index
Trading Asia Pacific Exchange JP Utility
Natural GAS
PJM China coal a-ds
index
the largest weekly return (0,24%) and the European electricity return and US coal return have both the lowest average weekly return (-0,09%). The average return of natural gas in the US has also a negative weekly return (-0,07%). This result can also be found in the paper of Creti et al. (2013). They find a daily average natural gas return of -0,0004%. In their study, they also find a negative daily average return for electricity (-0,0002%), in my paper the average weekly return in Europe is even larger and negative (-0,09%). For oil returns Creti et al. (2013) find smaller returns (daily return of 0,0004%) as my study (weekly return US: 0,14% and Europe: 0,16%). Also the average oil return Sadorsky (2001) found is smaller as my study shows. He found an average monthly return of 0,33%. However compared to Henriques and Sadorsky (2008) the return on oil prices is low. They found an average weekly return of 0,25%. The variables Oil Return (US), Gas return (Asia) and Coal Return (US) have very high skweness and kurtosis. Their observations are probably not normally distributed.
Variable Obs. Mean Std Dev. Min Max Skewness Kurtosis
Stock return 19070 -0,0013% 9,93% -65% 161% 0,35 9,91 Market return 474 0,090% 2,628% -20% 11% -0,97 8,99 Risk free rate 474 0,053% 0,059% 0,00% 0,159% 0,61 -1,45 Elect. return (US) 474 0,14% 34,12% -148% 190% 0,13 3,12 Elect. Return (Europe) 474 -0,09% 24,04% -177% 107% -0,77 8,17 Oil return (US) 474 0,14% 5,28% -36% 24% -1,00 7,77 Oil Return (Europe) 474 0,16% 4,34% -21% 12% -0,77 2,06 Gas return (US) 474 -0,07% 9,29% -45% 52% 0,35 7,15 Gas return (Europe) 474 0,14% 6,50% -41% 36% 0,85 12,29 Gas return (Asia) 474 0,24% 2,81% -21% 11% -1,00 8,77 Coal return (US) 474 -0,09% 6,15% -65% 76% 1,35 78,04 Coal return (Europe) 474 0,08% 3,21% -14% 17% 0,15 3,44 Coal return (Asia) 474 -0,02% 5,45% -32% 18% -0,13 2,94
Summary statistics for the sample between 03/03/2005 and 03/04/2014. All variables are weekly returns computed as continues compounded returns: = ln( . The ‘’Stock returns’’ are the returns on the stocks of
renewable energy companies. The ‘’Market return’’ is the return on the S&P 500. The statistics for the risk free rate are
In Table 4 the correlations of the different variables are presented. The returns that are used to compose the correlations are calculated with the simple interest method. Correlations composed with continues compounded returns were also made (not presented in this paper), however this did not result in differences in significant levels. A surprising result is the correlation between the average stock return and the market return. The mean stock return of renewable energy companies is on average negative and the mean of the market return is positive, tough the correlation is high and significant (0.818). Sadorsky (2012) finds in his paper also a positive and significant correlation between these variables (0,262), however not as high as table 4 shows. Furthermore, the correlations between the cross continent indices for oil (0.738) and electricity (0.144) are significant. As mentioned before the coal and natural gas indices of the US and Europe have a low and insignificant correlation (0.062) and (-0,017), therefore the additional coal and natural gas indices are added. Oil prices correlate to many other energy sources. Both US and EU oil prices together correlate significantly to 10 out of 12 other energy sources. Only European natural gas is not correlated to oil prices. A surprising result is that the US natural gas variable correlates significantly with the oil variables of the US (0,128) and Europe (0,194), but not with the natural gas variable in Europe (-0,017).
Table 4 shows that many variables are correlated with each other. There is also a possibility that the variables have some relationship in the long run, they could be cointegrated. In order to test if these variables are cointegrated a Johansen Cointegration Test is done. In table 5 a summary of five different models of cointegration tests is presented. There are 12 variables tested. The results of the different
Table 4: Correlation matrix
Average stock returns Market return Risk free rate Elect. us Elect.
eu Oil us Oil eu Gas us Gas eu Gas as
Coal us Coal eu Coal as Average stock returns 1 Market return 0,818 ** 1 Risk free rate 0,045 -0,014 1 Elect. Us -0,018 -0,038 -0,005 1 Elect. Eu 0,043 -0,041 0,026 0,144 ** 1 Oil us 0,347** 0,343** 0,020 -0,048 0,04 1 Oil eu 0,384** 0,353** 0,029 -0,002 0,084 ,0738** 1 Gas us 0,100* 0,098* 0,012 0,154** 0,076 0,128** 0,194** 1 Gas eu 0,002 0,002 0,029 -0,039 0,112* 0,013 0,049 -0,017 1 Gas as 0,505** 0,474** 0,053 0,011 0,042 0,248** 0,285** 0,099* 0,043 1 Coal us 0,139** 0,133** 0,007 0,006 0,056 0,185** 0,150** 0,096* 0,021 0,082 1 Coal eu 0,084 0,068 0,080 -0,04 0,031 0,091* 0,103* 0,035 0,089 0,081 0.062 1 Coal as 0,127** 0,110* 0,112* 0,003 0,003 0,116* 0,163** 0,059 -0,039 0,303** 0,047 -0,035 1
models are very consistent in the number of cointegrations. The test suggests that there are 11 cointegrating vectors. This result implies that the variables share a common stochastic drift. Note that within this high number, energy sources of different continents are captured. Oil prices in the US are very likely to be cointegrated with oil prices in the EU, furthermore oil prices and gas prices are linked to each other (Stern, 2007). Another reason for the conintegrating variables could be the recent financial crisis. Before the financial crisis energy prices were relatively high. With the start of the crisis energy prices decreased and rose again after the crisis. The result of the Johansen cointegration test implies that in the rest of this study an error correction model should be used to estimate the model. The error correction model used in this study is: Fully-Modified Ordinary Least Squares. The trend specification options in this test will be set on linear trend and quadratic trend.
6. Results
In this section the model is empirically tested and hypothesis are accepted or rejected. Further, robustness checks are done on the basis of company size, comparison between US and non-US companies and between first year and last year results statistically compared. Finally the results are compared with a sample of conventional companies.
6.1 Panel data analysis
To test if energy price fluctuations have significant effect the stock returns of renewable energy companies a fixed effects panel least squares regression test is done. The test captures the cross-sectional variation. As described before, over time the sample became larger. This puts a constraint on the test results when all observations are tested at ones. Since, the end of the period has a larger number of observations the results could be biased, because the relationships might not hold throughout time. In order to get better insight in the relationship throughout time the test is performed
Tabel 5: Johansen System Cointegration Test
Selected (0.05 level*) Number of Cointegrating Relations by Model
Data Trend: None None Linear Linear Quadratic
Test Type No Intercept No Trend Intercept No Trend Intercept No Trend Intercept Trend Intercept Trend Trace 11 11 11 11 11 Max-Eigenvalue 11 11 11 11 11
on each individual year in the sample, starting in March 2005. The results of this test are shown in table 6. The last column test for the whole sample period. The results show large difference in significance between years. The only variable that is significant in almost every year is the excess market return ( . Only in the last year this variable is not significant. For electricity returns only two of the nine years are statistically significant. The betas of these two years have opposite signs. Furthermore, the test for the whole sample period does not result in a significant relationship. This leads to the conclusion that H1 is rejected. Fluctuations in energy prices have no significant effect on the stock returns of companies in the renewable energy industry. This is inconsistent with the work of Boyer and Filion (2007), Nandha and Faff (2008) and Hasan and Ratti (2014) who find that producers of energy sources, like oil and coal, experience a rise in stock returns when the price of their end product rises.
variables in times of crisis. In every year the Durbin-Whatson test result has a value around 2. This indicates that there is no autocorrelation between the variables.
6.2 Robustness
In order to test the robustness of the sample and the robustness throughout the time period a number of tests are done. Table 7 shows the results of the three tests. Tests are done in two stages. First, for each individual company a Fully-Modified Ordinary Least Squares test is done. The trend specifications are
Table 6: Results of the cross-sectional fixed effects panel least squares regression test
Cross-sectional fixed effects panel least squares regression
Variables Time period
3/2005 -3/2006 3/2006 -3/2007 3/2007 -3/2008 3/2008 -3/2009 3/2009 -3/2010 3/2010 -3/2011 3/2011 -3/2012 3/2012 -3/2013 3/2013 -3/2014 Whole sample period Number of companies 24 27 32 39 40 41 46 49 53 53 Constant 1.0864 0.2959 -0.6454 0.3443 -0.1262 -0.1242 0.1208 0.2359 0.1022 -0.0813 (0.0000) (0.0003) (0.0000) (0.0000) (0.0089) (0.0028) (0.0816) (0.0009) (0.1286) (0.0000) Excess market return -0.0304 1.6874 1.4344 1.5918 1.6713 1.4440 1.4303 1.5763 1.4923 0.2924 (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) 0.0000 Elect. 0.1242 0.0070 0.0108 -0.0224 -0.0061 0.0026 0.0022 0.0041 -0.0347 -0.0016 (0.9842) (0.1964) (0.0454) (0.0043) (0.4561) (0.6397) (0.6385) (0.3628) (0.6015) (0.2818) Oil 2.5005 0.1865 0.0406 0.0228 0.0378 0.0749 0.0047 0.0242 0.0044 0.2967 (0.0005) (0.0002) (0.4911) (0.4135) (0.4554) (0.0346) (0.9334) (0.6885) (0.0230) (0.0000) Gas 0.4767 -0.0255 -0.0454 0.2679 0.0185 -0.0021 0.1019 -0.1818 -0.1274 0.0286 (0.2071) (0.1406) (0.1584) (0.0000) (0.1054) (0.9364) (0.0018) (0.0000) (0.0000) (0.0002) Coal -0.2717 0.0884 -0.2430 0.0107 0.2356 -0.0557 -0.1713 0.0659 0.0767 0.0628 (0.0786) (0.3702) (0.0001) (0.4275) (0.0000) (0.2613) (0.1445) (0.4504) (0.3120) (0.0000) Exchange rate 0.0001 -0.0493 1.1107 -0.4034 0.2035 0.1818 -0.0756 -0.2433 -0.1203 0.1240 (0.0000) (0.6203) (0.0000) (0.0000) (0.0027) (0.0005) (0.3332) (0.0036) (0.1895) (0.0000) Interest rate 0.1565 -1.5021 5.6556 3.5318 -1.8030 0.2945 -8.4665 -8.5517 0.6662 0.7655 (0.1458) (0.0839) (0.0057) (0.0001) (0.0001) (0.8131) (0.0003) (0.0004) (0.7967) (0.0006) R-squared 0.0936 0.1366 0.1542 0.3607 0.3007 0.2388 0.2318 0.1349 0.0987 0.0658 Adjusted R-squared 0.0715 0.1162 0.1352 0.3465 0.2852 0.2220 0.2151 0.1158 0.0807 0.0629 Log likelihood 1504.5 1794.9 1913.5 1632.5 2141.6 2910.1 2617.1 2920.9 3167.4 17556.7 F-statistic 4.2 6.7 8.1 25.3 19.4 14.2 13.8 7.1 5.5 22.7 Prob(F-statistic) 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Durbin-Watson stat 2.0632 2.2196 1.9958 2.1374 2.1134 2.1308 2.1945 2.1345 1.9799 1.982349
set on linear trend and quadratic trend to control for cointegration between the variables. From the results of each test the betas of the independent variables are used to make lists of betas per variable (for example the oil betas of all US companies make the list ‘’US companies: oil’’) Each beta is saved in a list per variable. This resulted in two list of betas per variable (for example a list of ‘’US companies oil’’ and ‘’Non-US companies oil’’). Second, the two lists are compared using a test of equality of means between series. The t-values are shown in table 7. The first test is a mean comparison between the betas of the independent variables in the first year of the sample the betas of the independent variables in last year of the sample. In the first year the sample consisted of 24 companies and in the last year of 53 companies. The test shows statistically significant results for the variable natural gas and the common risk exposure variables exchange rate and interest rate. The effect of natural gas on stock return of renewable energy companies is not constant within the sample period. In the first year of the sample period the effect is larger. Also the exposure to the exchange rate and interest rate was significantly larger in the beginning of the sample period. A logical explanation for this are the high values of these variables in the beginning of the sample period. Countries had high interest rates before the financial crisis.
The second test is a comparison between US and non-US companies. The sample consist of 39 US companies and 14 non-US companies. The energy variables oil and natural gas are both significant in the test. This indicates that price fluctuations in both energy sources have a larger impact on Non-US
companies. The last test shows the result of the mean comparison between the largest 30% (16 companies) and the smallest 30% of the companies in the sample. The size of a company is measured by the total assets of the company. The total assets of the company fluctuates over time. For this test the size is measured at 1 July 2013. At this time the smallest company had a size of 2,1 million US dollar and the largest company 33.407,6 million. The test results indicate a significant result for excess market return and natural gas. Larger companies are more exposed to the market and to fluctuations in
Table 7: Mean comparison robustness tests
Table 7 provides a summary of three mean comparison tests. These tests compare betas of each group with the betas of another group. In column two the results are provided for the mean comparison between companies in the first year (N=24) of the sample with the last year of the sample (N=53). Column three shows the results of the mean comparison between US (N=39) and non-US companies in the sample (N=14). In column four results of the mean comparison between the smallest 30% (N=16) of the companies in the sample and largest 30% is shown (N=16). The reported values are t-statistics. ** indicate a significant value at a 5% significance level.
Variable First – Last year US – Non US Small - Large
Excess market return -0.306 -1.382 2.411**
natural gas prices. The excess return on the market is based on the return of the S&P 500. These are the largest companies listed in the US. Therefore, it makes sense that larger companies in the sample are more exposed to excess return on the market, since they all have this feature.
Interesting is the significance result of natural gas over the whole sample period in table 6 and in all three the models of table 7. The effect of natural gas price fluctuations is not constant throughout the whole sample and sample period. To empirically test if these inconsistent results also appear in other samples a second sample is selected.
6.3 Paired sample
In order to test if the expected relationships between the variables are specific characteristics of renewable energy companies, a second sample is selected. Furthermore, the issues on the inconsistency of the natural gas variable, as described in section 6.2 is also tested. The second sample consist of conventional companies which are not active in the renewable energy sector. For each individual company of the Wilderhill index another company is selected. Appendix 3 provides an overview with these companies. The selection criteria are as follows: (1) the first 2 or 3 numbers of the SIC code of the selected company are identical to the SIC code numbers of the renewable company. (2) The total assets of the selected company are approximately equal to the total assets of the renewable company. (3) Both companies have the same country of origin. Table 8 provides the descriptive statistics for the sample with conventional companies. The returns are calculated as continues compounded, like in section 5.
Tabel 8: Paired sample descriptive statistics
Variable Obs. Mean Std Dev. Min Max Skewness Kurtosis
Stock return 22961 -0,022% 9,26% -190% 147% -0,52 25,45
The mean stock return of the paired sample is, like the return on renewable energy companies, also negative (-0,022%). However, the mean stock return of the paired sample is even more negative. An obvious explanations for these negative returns is the recent financial crisis. The standard deviation of the stock return is approximately equal in both sample (9,93% and 9,26%), which implies that both samples are equally distributed. In order to statistically test if the stock returns between the renewable energy companies and the paired sample differ, a number of tests are presented in table 9. Stock returns of the renewable energy companies and stock returns of the paired sample are statistically compared with each other. None of the tests are significant. The results of the table show that the stock returns to not statistically differ between the two samples. This is shown by a t-test. The distributions
of both samples have high kurtosis 9,91 and 25,45, therefore two non-parametrical tests are done. The results of these two tests do not differ from the t-test.
Both samples should not have significant different stock returns. This would imply that one sample outperformed the other while both operating in comparable industries. In this study, there are no statistical differences between the two samples, thus the results of the panel fixed effects least squares test can be compared with each other. In table 10 the results of the test are presented. The results again indicate large differences in significance across years. A similarity between both samples is the significant excess market return in almost every year. Also exposure to the exchange rate and interest rate is shown in both samples. Concerning the energy sources, there is no clear pattern. The energy sources in the renewable energy sample have slightly more significant years compared to the paired sample. In the paired sample the natural gas variable shows the best result with a significant result in four of the nine year, followed by electricity (three of the nine year), oil (two of the nine years) and coal (no significant results). When the test is performed for the whole sample period an identical pattern develops as in the renewable energy sample. The variables excess return, oil price, gas price, coal price and exchange rate all have a significant effect on stock returns. Only interest rate and electricity prices are not significant in the paired sample. This results implies that the exposure of renewable energy sources to oil, natural gas and coal price fluctuations is not a specific characteristic of the renewable energy. It is also a characteristic of companies outside of the renewable energy industry. The values of the R-squared and adjusted R-squared are slightly lower compared to the renewable energy sample.
Table 9: stock return comparison between the renewable energy sample and the paired sample
Stock return comparison
Test Correlation Test value Significance
T - test 0,850 -0,31 0,782
Wilcoxon/Mann-Withney test 0,850 0,403 0,687
Kruskal-Wallis test 0,850 0,163 0,687
To statistically compare if both samples are differently exposed to the independent variables a number of tests is done. First, for each company in the paired sample an individual Fully-Modified Ordinary Least Squares test is done, with trend specifications on linear trend and quadratic trend. From the results of these tests the betas per variable are put in a list. These lists are compared with the betas list of the renewable energy sample by three mean comparing tests. The results of these tests are shown in
Table 10: Results of the fixed effects panel least squares regression test for the paired sample
Fixed effects panel least squares regression
Variable Years 3/2005 -3/2006 3/2006 -3/2007 3/2007 -3/2008 3/2008 -3/2009 3/2009 -3/2010 3/2010 -3/2011 3/2011 -3/2012 3/2012 -3/2013 3/2013 -3/2014 Whole sample period Number of companies 44 44 45 46 46 46 50 52 53 53 Constant -0.2367 0.1925 -0.4391 0.2545 -0.1071 0.1185 0.0447 0.0349 0.0843 -0.039 (0.0000) (0.0012) (0.0000) (0.0000) (0.1539) (0.0147) (0.4713) (0.5797) (0.1546) (0.0000) Excess market return 0.7962 1.4196 1.1249 1.1453 1.1194 1.1109 1.1855 0.0780 1.1415 0.2105 (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.3660) (0.0000) (0.0000) Elect. 0.0017 0.0107 0.0100 -0.0105 0.0098 -0.0005 0.0053 -0.0193 0.0004 0.0003 (0.6955) (0.0063) (0.0370) (0.1914) (0.4407) (0.9358) (0.1990) (0.0000) (0.8296) (0.8551) Oil 0.0354 0.1269 0.0632 -0.0090 0.1492 0.0520 0.1142 0.0638 0.0336 0.2218 (0.2842) (0.0004) (0.2288) (0.7531) (0.0588) (0.2089) (0.0259) (0.1843) (0.5651) (0.0000) Gas 0.0079 -0.0368 -0.0621 0.1735 0.0379 0.0518 -0.0399 -0.0136 -0.0331 0.0280 (0.6524) (0.0031) (0.0305) (0.0001) (0.0334) (0.0871) (0.1845) (0.5633) (0.1389) (0.0000) Coal 0.0155 -0.0709 -0.0002 0.0096 0.1260 -0.0583 0.0451 0.0870 -0.1086 0.0424 (0.7637) (0.3188) (0.9970) (0.4894) (0.1112) (0.3138) (0.6742) (0.2327) (0.1043) (0.0000) Exchange rate 0.3798 0.0401 0.7956 -0.3081 0.1845 -0.1267 0.0052 0.0014 -0.1202 0.0718 (0.0000) (0.5760) (0.0000) (0.0000) (0.0806) (0.0377) (0.9407) (0.9853) (0.1365) (0.0000) Interest rate 3.2286 -2.5196 1.7269 2.2087 -0.9860 -1.5587 -5.6966 -7.1966 3.5301 -0.1370 (0.0104) (0.0001) (0.3433) (0.0180) (0.1736) (0.2842) (0.0099) (0.0000) (0.1211) (0.4900) R-squared 0.0580 0.1210 0.1025 0.1907 0.0917 0.1291 0.2343 0.0354 0.0610 0.0419 Adjusted R-squared 0.0373 0.1017 0.0833 0.1731 0.0719 0.1101 0.2164 0.0146 0.0422 0.0394 Log likelihood 2844.3 3117.3 2555.8 1658.8 1209.1 2745.5 2851.5 3478.0 3578.3 21660.1 F-statistic 2.8073 6.2805 5.3261 10.8098 4.6330 6.7986 13.0744 1.7034 3.2510 17.0 Prob(F-statistic) 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0009 0.0000 0.0000 Durbin-Watson stat 1.968473 2.085118 2.099134 2.176519 2.181478 2.205529 2.084122 2.104224 2.182666 2.059854
table 11. The table shows no significant results for the energy sources. Only the variable excess market return is significant. This indicates that none of the energy variables in the model have a significantly different impact on the stock returns of renewable energy companies and conventional companies. Note that the variable natural gas is also not significant. The results in table 10 and 11 indicate that natural gas price fluctuations have the same impact renewable energy companies as on the paired sample.
7. Conclusion
At one point in the future fossil energy sources will be exhausted. In order to fulfill the worlds’ energy needs other energy sources are required. An alternative for fossil fuels is renewable energy. The renewable energy industry is the fastest growing industry worldwide and is becoming more important. In this study the effect of electricity, oil, natural gas and coal price fluctuations on stock returns of companies in the renewable energy sector is researched. This is tested using two hypothesis. The first hypothesis states that the fluctuations in electricity prices are positively related to stock returns of renewable energy companies. The second hypothesis states that oil, natural gas and coal price fluctuations are also positively related to stock returns of renewable energy companies. This study used the model of Sadorsky (2001). It is a multifactor model where exchange rates and interest rates are added to account for risk exposure. The sample consist of companies from the Wilderhill Clean Energy index. The sample period in this study is March 2005 until March 2014, weekly data is used. The results of the fixed effects panel least squares regression provide mixed results. Over the whole sample period, oil, natural gas and coal prices are significantly positively related to renewable energy stock returns. No significant results were found for electricity price. Based on these results H1 would be rejected and H2 would be accepted. However, when looking at the significance levels in the individual years the observed relations are not so strong anymore. A mix of significant and insignificant years is found for each variable. Based on these result H1 and H2 could neither be
Tabel 11: A comparison between the betas of the renewable energy sample with the paired sample
Comparing betas sample – paired sample Variables
Test type Excess
market return
Electricity Oil Natural gas Coal Exchange rate Interest rate t-test 0.0124 0.3373 0.9998 0.2691 0.6381 0.2895 0.2834 Satterthwaite-Welch t-test 0.0132 0.3395 0.9998 0.2703 0.6382 0.2913 0.2852 Welch F-test 0.0132 0.3395 0.9998 0.2703 0.6382 0.2913 0.2852
accepted or rejected. The sample seems reasonably robust. Only the exposure to natural gas prices is not constant over the whole sample and sample period. The results found in this study correspond to the results of Henriques and Sadorsky (2008) who also find a positive relationship between oil prices and stock returns of renewable energy companies. However, this relationship is not corresponding with the general theory that higher oil prices are bad for stock returns and the economy as a whole (Sadorksy, 1999, 2001; Park and Ratti, 2008; Cunado and Gracia, 2014; Asteriou et al. 2013). The results for natural gas are not consistent with Oberndorfer (2009) who did not find a relationship between energy stock and natural gas prices. In order to test if the relations are characteristics of the renewable energy industry or just common relations, a paired sample was drawn. The same fixed effects panel least squares regression test was performed on this sample. The results show a similar pattern as for the renewable energy companies. Over the whole sample period significant results are found for the energy variables oil, natural gas and coal price and an insignificant result for electricity price. The significance levels over the individual years also show a similar pattern, a mix of significant and insignificant results. Tough, the paired sample shows a slightly lower number of significant year results as the renewable energy companies. To empirically test if the exposure to the price fluctuations in the four energy sources is significantly different for the two samples mean comparison tests between the betas of both samples is done. The results show no significant difference in exposure between the two samples.
On a whole, the findings in this study show that the oil, natural gas and coal do have at least some explanatory power in determining stock return of companies in the renewable energy industry. This result suggest that renewable energy could be a substitute for other energy source. A higher price of energy produced with oil, natural gas and coal makes renewable energy relatively cheaper, which increases demand for renewable energy and raises stock returns of renewable energy companies. This study adds to the literature by identifying energy sources that drive stock prices of companies in the renewable energy industry.
References:
Acaravci, A., Ozturk, I., & Kandir, S. Y. (2012). Natural gas prices and stock prices: Evidence from EU-15 countries. Economic Modelling, 29(5), 1646-1654.
Asteriou, D., Dimitras, A., & Lendewig, A. (2013). The Influence of Oil Prices on Stock Market Returns: Empirical Evidence from Oil Exporting and Oil Importing Countries. International Journal of Business and Management, 8(18), p101.
Boyer, M.M., Filion, D., (2007). Common and fundamental factors in stock returns of Canadian oil and gas companies. Energy Economics 29 (3), 428–453.
Brown, S. J., & Warner, J. B. (1985). Using daily stock returns: The case of event studies. Journal of financial economics, 14(1), 3-31.
Chan, K., Chen, N. F., & Hsieh, D. A. (1985). An exploratory investigation of the firm size effect. Journal of Financial Economics, 14(3), 451-471.
Chan, H., & Faff, R. (1998). The sensitivity of Australian industry equity returns to a gold price factor. Accounting & Finance, 38(2), 223-244.
Chien, T., & Hu, J. L. (2007). Renewable energy and macroeconomic efficiency of OECD and non-OECD economies. Energy Policy, 35(7), 3606-3615.
Cox, P. M., Betts, R. A., Jones, C. D., Spall, S. A., & Totterdell, I. J. (2000). Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model. Nature, 408(6809), 184-187.
Creti, A., Joëts, M., Mignon,. V. (2013) ‘’On the links between stock and commodity market’ volatility’’, Energy economics 37: 16-28
Cuñado, J., & Pérez de Gracia, F. (2014). Oil price shocks and stock market returns: Evidence for some European countries. Energy Economics, article in press
Ellerman, A. D. (1995). The world price of coal. Energy Policy, 23(6), 499-506.
Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of financial economics, 33(1), 3-56.
Geman, H., & Roncoroni, A. (2006). Understanding the Fine Structure of Electricity Prices. The Journal of Business, 79(3), 1225-1261.
Hasan, M. Z., & Ratti, R. A. Australian coal company risk factors: coal and oil prices. Business and Finance ESEARCH, 57.
Henriques, I., & Sadorsky, P. (2008). Oil prices and the stock prices of alternative energy companies. Energy Economics, 30(3), 998-1010.
Jorion, P. (1990). The exchange-rate exposure of US multinationals. The Journal of Business, 63(3), 331-45.
Keim, D. B. (1983). Size-related anomalies and stock return seasonality: Further empirical evidence. Journal of Financial Economics, 12(1), 13-32.
Khoo, A. (1994). Estimation of foreign exchange exposure: an application to mining companies in Australia. Journal of International Money and Finance, 13(3), 342-363.
Kristoufek, L., & Vosvrda, M. (2013). Measuring capital market efficiency: Global and local correlations structure. Physica A: Statistical Mechanics and its Applications, 392(1), 184-193. Loudon, G. (1993). The foreign exchange operating exposure of Australian stocks. Accounting &
Finance, 33(1), 19-32.
Martin, J. D., & Keown, A. J. (1977). Interest rate sensitivity and portfolio risk. Journal of Financial and Quantitative Analysis, 12(02), 181-195.
Mirasgedis, S., Sarafidis, Y., Georgopoulou, E., & Lalas, D. P. (2002). The role of renewable energy sources within the framework of the Kyoto Protocol: the case of Greece. Renewable and Sustainable Energy Reviews, 6(3), 247-269.
Medlock III, K. B. (2014). Natural Gas Price in Asia: What to Expect and What It Means. James A. Baker III institute for public policy
Nandha, M., and R. Faff. (2008). Does oil move equity prices? A global view, Energy economics 30: 986-997.
Oberndorfer, U. (2009). Energy prices, volatility, and the stock market: evidence from the Eurozone. Energy Policy, 37(12), 5787-5795.
Ortas, E., & Moneva, J. M. (2013). The Clean Techs equity indexes at stake: Risk and return dynamics analysis. Energy, 57, 259-269.
Park, J., & Ratti, R. A. (2008). Oil price shocks and stock markets in the US and 13 European countries. Energy Economics, 30(5), 2587-2608.
Sadorsky, P. (1999) “Oil Price Shocks and Stock Market Activity,” Energy Economics 21: 449-469. Sadorsky, P. (2001). Risk factors in stock returns of Canadian oil and gas companies. Energy
Economics, 23(1), 17-28.
Sadorsky, P. (2009). Renewable energy consumption, CO2 emissions and oil prices in the G7 countries. Energy Economics, 31(3), 456-462.
Sadorsky, P. (2012). Modeling renewable energy company risk. Energy Policy, 40, 39-48.
Stern, J. P. (2007). Is there a rationale for the continuing link to oil product prices in continental european long-term gas contracts?. International Journal of Energy Sector Management, 1(3), 221-239.
Stone, B. K. (1974). Systematic interest-rate risk in a two-index model of returns. Journal of Financial and Quantitative Analysis, 9(05), 709-721.
