The Impact of Policies on Investor Perception of Renewable Energy Companies

(1)

The Impact of Policies on Investor

Perception of Renewable Energy

Companies

Michal Lietava

June 2008

UNIVERSITY OF GRONINGEN

Faculty of Economics and Business

MSc BA Finance

Supervisor:

(2)

The Impact of Policies on Investor

Perception of Renewable Energy

Companies

ABSTRACT

The problematic depletion of conventional sources of energy and other environmental strains brings up the issue of exploiting renewable energies. These have many advantages, such as sustainability and low environmental impacts, however their deployment is rather problematic and slow. Many governments, being aware of these issues, support usage and development of renewable energies by various policies. Some studies assess how effective support through such policies is, but so far there is no study that examines how they are perceived from an investor's point of view. This is crucial as renewable energies will be used much more widely when there is sufficient funding from the financial markets.

Therefore, this thesis examines how policies supporting renewable energy companies influence their investors. This is measured in two ways. Firstly, we analyze the influence of supportive policies on sensitivity of renewable energy stocks to exogenous factors. For this, daily stock price returns of 248 companies in EU, USA, Canada, China, and Australia are used. As exogenous factors we use price returns of crude oil, interest rates, country indices, and agricultural commodities. Secondly, we examine the impact of policies on the rate of IPOs in the sector of renewable energies. For this we gathered data on 79 IPO deals.

We can conclude that governmental policies do have an impact on stock sensitivity, but not generally. Only some of the analyzed policies proved to have this impact, mainly those that have a market wide influence and can be applied with only small bureaucracy. Examination of IPO deals does not confirm this conclusion because our hypothesis was not proved in any case.

JEL Codes: G12, Q28, Q42

(3)

PREFACE

In the academic year of 2007/2008 I have been studying an MSc BA program in Corporate Finance at the University of Groningen. During this master program I wrote this thesis in which I tried to merge my financial specialization with my passion and former experience in the field of alternative energies. The aim of my work was to examine if the policies put forward by governments to support deployment of renewable resources have an effect that is measurable on financial markets.

While writing my thesis, I have enjoyed the help of my thesis supervisor Professor Bert Scholtens. We had several discussions on the final content of the thesis which contributed to increasing its informational value. Insights of Professor Scholtens helped me very much in designing a methodology through which I was be able to perform intended measurements for which I would hereby like to thank him. I would also like to thank my parents who made my studies in Groningen possible and supported me along the way.

Michal Lietava Groningen, June 2008

Author’s student number and contact:

(4)

TABLE OF CONTENTS

1 Introduction... 5

2 Literature review ... 6

2.1 The relation between renewable energy demand and policy... 9

2.2 Reasons for stock price sensitivity ... 9

2.3 Characteristics of stock sensitivity and methods of investigating it... 10

2.4 Investment rate and policy... 13

2.5 Company behavior when deciding for IPO ... 15

2.6 Our contribution to results of studied literature... 17

3 Data description ... 18

3.1 Renewable energy stocks... 19

3.1.1 ZEPHYR ... 19

3.1.2 Other internet resources ... 21

3.2 Legal frameworks incentivizing use of renewable energies ... 26

3.2.1 European Union... 27

3.2.2 United States of America ... 29

3.2.3 Canada ... 31 3.2.4 China ... 33 3.2.5 Australia ... 34 3.3 Interest rates... 35 3.4 Indices... 37 3.5 Commodities... 38

3.6 Initial Public offerings ... 41

4 Methodology ... 43

4.1 Measuring sensitivity of stocks ... 43

4.1.1 Justification for used variables and their description ... 49

4.1.2 Comparing coefficients from portfolio and index regressions... 49

4.1.3 Why we analyze only one market companies ... 50

4.1.4 Final form of the regression model ... 51

4.1.5 Comprehensive outline of order of steps in our analysis ... 55

4.2 Measuring changes in IPO volumes ... 57

5 Results ... 59

5.1 Changes in sensitivity of value of renewable energy stocks ... 59

5.1.1 European Union... 59

5.1.2 USA ... 62

5.1.3 Canada ... 66

5.1.4 China ... 71

5.1.5 Australia ... 73

5.1.6 Comparison of changes in sensitivity between countries... 75

5.1.7 Discussion on results of sections 5.1.1 to 5.1.5... 75

5.2 Changes in IPO activity of renewable energy stocks ... 77

6 Conclusion ... 79

7 References ... 81

8 List of tables... 84

9 List of Charts... 85

(5)

1 Introduction

The world is dependent on energy. Availability of energy has always been a crucial factor for development of the economy and we can see this relationship today as well. Fossil fuels are currently the main source for energy but their availability and environmental impacts induce mankind to think about alternatives that are sustainable and clean. Renewable energies, such as solar, wind, hydropower, biomass and geothermal electricity generation, together with biofuels for transport and other green resources give us the possibility to have a source of energy that is not depletable and environmentally friendly. However, we can see that the deployment of renewable energies is not easy and despite their benefits, they have to struggle hard to get interest from investors since they are still more expensive as conventional energy sources. For example in the case of European Union, we have seen, however, that the emergence of supportive policy such as the directives 2001/77/EC and 2003/30/EC have sparked a turnaround and renewable energies such as wind power or biofuels became more widely used. This is documented by Holzer (2005) and European Commission (2007) which prove that these directives have effect. This implies that policy can have large effects on the investors into renewables and thus the characteristics of these effects are worth studying.

(6)

conditions partially determined by a law makes them less volatile. And because of this lower volatility, given the same return, it makes the investment more profitable, and we expect investors to be generally more enthusiastic to fund this industry. One of the ways in which this enthusiasm can be visible is in a higher IPO level and a lower sensitivity to other factors influencing profits of the renewable energy companies.

This research is conducted for five regions: EU, USA, Canada, China, and Australia. The longest time span that we are investigating is for the USA from June 1986 until March 2008. We perform 14 analyses of different laws or other governmental policies in the countries listed above.

The rest of this paper is organized as follows. In chapter 2 we present an overview of literature that addresses our research topic. Chapter 3 describes the data that we use for quantitative analysis and chapter 4 brings a detailed explanation of our methodology. Afterwards, analysis of results follows in chapter 5 and the paper is concluded in the following chapter. After the list of references, in chapter 10 there are appendices with additional charts and tables that will help the reader to gain a better overview about our research.

2 Literature review

Here, we point out pieces of literature that touch on various parts of reasoning that underlies the way in which we build up our research. According to our knowledge, articles that cover all our research questions do not exist, but there are articles that can be used to support our reasoning step by step. Broadly, the literature can be divided into papers that we use by studying impact of regulatory acts on renewable energy stock sensitivity to exogenous factors, and then literature that concerns the sensitivity of investments to regulations supporting a particular type of business.

For the first part of our research, studying stock sensitivity, we use literature that shows how regulatory policies can change demand for renewable energy. Since the usage of renewable energies directly influences renewable energy companies’ performance, we can expect it to influence the sentiment of investors as well. This is observable in how the investors react to changes in levels of exogenous factors that influence profitability of their investments.

(7)

to exogenous factors. What is even more important to us is having insight into how this sensitivity is measured and what its characteristics are that should be reflected in our model in order to be able to see significant results. In addition, we also include literature that is valuable because of interesting methodology used to study these issues.

For the part of our research concerning changes in IPO levels of renewable energy companies because of new policies, we look for literature that shows that the investment level may be influenced by such measures. In addition, we also survey literature that generally describes the behavior of firms going public.

(8)

TABLE 1 - Literature Overview

Author (year) Study domain

Study

period Studied subjects Analysis method

Implication for our

methodology Independent variables Dependent variables Results Section Trombork et al.

(2007)

Influence of policy on sensitivity of demand for RE to energy prices 2003 - 2015 (forecast ) Norwegian forest and heat sector

Norwegian Trade Model II

-Wood (prices, production, capacities); households (heating systems, consumption)

Wood based bioenergy production

More subsidies support bigger

production of bioenergy 2.1 Sakkelaris

(1997)

Explanation why stocks react to real market company performance 1959 - 1985 Energy intensive sectors Multivariate structural models

-Prices of labor, energies, materials and capital

Stock value of company

Oil price shocks have an impact on

company value 2.2

Huang et al. (1996)

Studies of reactions of US stocks to oil prices

1979 -

1990 US stocks VAR

-Heating oil and crude oil futures, interest rates

S&P 500, sector indices

Only individual oil companies are influenced by oil prices 2.3 Nandha and

Faff (2007)

Stock reaction to oil prices by sector

1983 -

2005 Various sectors Linear regression

Log returns, reaction asymmetry

World market returns and oil returns

Global industry equity indices

Oil prices have different impacts on

different industries 2.3

Henriques and Sadorsky (2005)

RE company stock reaction to oil prices and interest rates

2001 - 2007 Renewable energies VAR Lagged explanatory variables

Arca Tech 100 Index, crude oil futures, yields of US T-bills

WilderHill Clean Energy Index

Lagged independent variables

cause movements in RE stocks 2.3 Mork et al.

(1994)

Study of oil price shocks on GDP 1967 - 1992 Country economies Multivariate

correlation Reaction asymmetry Oil prices Real GDP growth

Reactions of economy are very

asymetrical or unclear 2.3 Hayo and

Kutan (2005)

Impacts of various factors for Russian equity markets

1995 - 2001

Russian financial

markets GARCH (1,1)

Lagged explanatory variables, ARCH effects

US stock and bond prices, oil prices, news, months, autoregressive terms

Returns on Russian equity and bond markets

Oil prices and energy have a significant effect, while political

news do not. 2.3

Gallagher and Taylor (2002)

Temporary and permanent components on stock prices

1949 -

1997 US stocks VAR

Mean reversion (auto regression)

US macro-variables( interest rates, output, wage rate, consumption)

Real stock prices (S&P 500 adjusted by CPI)

Stock prices are not pure random walks but they revert to the mean 2.3 Nandha and

Hammoudeh (2007)

Sensitivity of Asian stock indices to oil price

1994 - 2004 Asia - Pacific equity indices Linear regression Integration of variables, autoregressive variables, asymmetric

Oil prices, interest rates,

autoregressive terms Index returns

Examined countries are not

sensitive to oil prices 2.3 Constantini and Crespi (2007) Impact of policy on competitve advantage of environmental companies 1994 - 2005 International trade Linearized gravitational model

-Country size, distance between countries, level of environmental regulation*

Level of Bilateral trade between countries

Environmental regulation gives

companies competitive advantage 2.4 Holzer (2005)

Impact of policy on usage of RE 2001 - 2004 German alternative energy market Using results from other vendors - - CO2 emissions

Environmental regulation fosters

utilization of RE 2.4

Edelstein and Kilian (2007)

Sensitivity of investment levels to energy prices

1970 -

2006 Various sectors VAR

between sensitivity of the market and portfolios

Energy prices (Ppi of Fuels and Lubricants)

Non-residential investment levels

Investment level is responsive to

energy prices 2.4

Ritter and Wesch (2002)**

Reasons for companies going public

1982 - 1992

Italian

non-financial firms Probit

-Capex, company size, M/B, Growth, ROA, Herfindhal Index, RCC, Year

Probability of going public

Market timing has an important role

when going public 2.5

Lowry and Schwert (2002)

Cyclicality of IPOs of similar companies

1960 -

1997 US companies

Correlations, F-

tests - Initial returns of IPOs IPO volume

High initial returns predetermine

subsequent IPO volume 2.5 Nelson (2002)

IPO herding, approached

theoretically - - -

-Noisy signals about value of companies going public, make

(9)

2.1 The relation between renewable energy demand and policy

Tromborg et al. (2007) engages in studies of how different policies affect the use of forest based bioenergy in Norway. He uses the Norwegian Trade Model II which is a version of the Network Trade Model (hereafter NTM) that predicts demand for bioenergy based heating up to 2015 based on different types of subsidies and conventional energy prices. In Norway, the heat market is heavily dependent on electricity, solid fuels, and kerosene, as a very large proportion of houses has its own heating and are not connected to central or district heating systems. However, hydro power plants are not built anymore on environmental concerns, and therefore, Norway has become a net importer of electricity. This puts pressure on utilizing alternative energy sources based on wood. Results are formulated in terms of on what level the energy prices must be so as not to need a supportive policy for usage of bioenergy. They clearly show that under any given energy price, the more subsidies available, the more energy generated from renewable resources. This is because the subsidies reduce the costs of using renewable energies and this in turn increases public demand for them which is very sensitive to energy prices. Renewable energy is preferred only when its price is comparable with the one of fossil resources, which means that in times of cheap fossil fuels demand for renewables declines. Tromborg, however, shows that with more subsidies, fossil energies can be cheaper and still need not harm the demand for renewable energies. Such an effect of policy should change investor sentiment towards producers of green energy and thus should also be observable on diminishing sensitivity to substitutes and other investment decision related factors that we are studying.

2.2 Reasons for stock price sensitivity

(10)

firm level financial and accounting data and industry level price data for a period of 1959 to 1985 with main focus on the impact of energy prices on the stock market value of a company. Sakellaris (1997) deliberately concentrates on energy intensive companies according to two or three digit SIC codes, in particular on Paper and Allied Products (SIC 26), Stone, Clay, Glass, and Concrete Products (32), Primary Metal Industries (SIC 33), and Electronic Components Industry (SIC 367) which serves as a control, energy non-intensive industry. The results are congruent with the intuition that oil price shocks do have impact. What is crucial to our research is the link to flexibility of change in used technology that this paper provides. The model that Sakellaris uses reflects the notion of putty-clay capital, that is, the assumption that once production factors are installed it is virtually impossible to change their mix and structure to reflect changes in prices of inputs and outputs. This actually explains, on a structural level, why there is sensitivity of stock to input factors. This assumption is also present in our research as we compare sensitivity of renewable energy stocks to input variables under different legislative support of their existence. If we assume that renewable energy companies also do have putty-clay capital, they are vulnerable to changes in profitability stemming from changes in their inputs and outputs. We expect to see that as more support is given to these companies insuring their profitability, their sensitivity to these factors should diminish. The reason for such an expectation comes from the fact that the legislation alters the impact of market movements on the company insulating it from part of the volatility in demand, output prices, etc. which should be reflected also in the capital markets.

2.3 Characteristics of stock sensitivity and methods of

investigating it

(11)

are competitors of oil companies, the impact of oil prices will also be more pronounced than for the general market.

Results of Nandha and Faff (2007) point out that oil price changes have different impacts on different industries. Their research that has examined this relation on various segments proves that oil and gas, mining, transportation, electrical equipment, and infrastructure technology hardware stocks are positively influenced by unexpected oil price increases and other segments are negatively influenced. Unfortunately, in their research, renewable energy industries are examined separately. For their research, they have used DataStream global equity industry indices from 1983 until 2005. Of particular interest is the methodology used by these authors. They were using log returns of the industry indices, but the explanatory variables were represented by unexpected returns of world market and unexpected return of oil price. The expected return of the world market is calculated through a simple linear regression with expected oil price as the only explanatory variable, where the expected oil price is a product of autoregressive model. The differences between the real and expected returns in world market and oil are used as unexpected returns. Nandha and Faff (2007) also test for asymmetry of response to oil shocks by having two different coefficients for upward and downward movements. By performing the Wald test of equality of these two coefficients, they point out that the impact of oil price increase and decrease is symmetric.

(12)

significant influence. This is, for us, a very challenging result seeing as the study has been done quite recently.

Symmetry of oil shock response is in contrast with the research on impact of oil price shocks on country GDP from Mork and Olsen (1994). They show that with the multivariate correlation method on USA, Canada, Japan, France, West Germany, Norway, and the UK that the impact is asymmetric for the period from 1967 to 1992. Noteworthy in their research is that the effects are rather similar regardless of whether a country is oil exporter or importer. Results of this study are counterintuitive to our expectations of individual stock reactions, because here the economy reacts asymmetrically not only in magnitude but also in direction. This is because Mork and Olsen (1994) found that for US, Canada and West Germany, both, oil price increases and decreases produced a negative reaction while for the rest of the countries the results are not clear.

Asymmetry of oil price changes’ impact is proven also by Hayo and Kutan (2005) who study the impacts of various factors on Russian equity and bond markets. They examine the period from 1995 to 2001 and the main explanatory variables are US stock and bond prices, oil prices, and political and economical news for Russia. They perform a time series analysis where the news is in form of a dummy variable. The model is autoregressive in type because all the explanatory variables, including also the dependent variable, are lagged from one to six days. Taking account of ARCH effects, the model is a GARCH (1,1) type with variance entering the main equation. Their result is intuitive in that the oil prices influence the financial markets, what one expects in the case of Russia. This also applies to energy news, but political news does not have this impact.

(13)

University of Chicago for the time from 1949 to 1997. The whole methodology is rather complex, so we will use only the idea of lagged data from it.

The paper of Nandha and Hammoudeh (2007) is also, among others, studying sensitivities of stock indices of 15 Asia-Pacific countries to oil prices on a weekly basis for one decade from 1994 to 2004. In terms of conclusions, this paper does not provide us with any new knowledge, but it is interesting for its methodology, even when it is quite simple. First of all, the paper is testing unit roots of the explanatory variables to determine what the degree of integration of each of them is. They are using the augmented Dickey-Fuller and Perron-Phillips tests to arrive at the conclusion that the variables of equity prices, oil prices, and also exchange rates are integrated on the first level. This means that shocks to these variables in level are permanent, but in the first differences are transitory. Further on in their regressions, they allow for asymmetric sensitivity of stock returns to oil price changes using dummy variables and they also add to the equation an autoregressive term of first degree AR(1) since it proved to be significant in several regressions.

The literature mentioned here makes us aware of various types of stock price sensitivities to exogenous factors. Particularly in the case of oil, we may see small sensitivity, symmetric, or asymmetric sensitivity. Furthermore, several studies use lagged data as explanatory variables, which we will consider in our analysis too. A special case of lagged data is the use of autoregressive terms. Of further consideration are the ARCH effects that can be present in the data. And lastly, usage of log returns may be necessary too.

2.4 Investment rate and policy

(14)

renewable energies. Results of Constantini’s and Crespi’s empirical work suggest that strict environmental regulation does indeed foster investments into renewable technologies. The way in which they arrive at these results is by using gravitational models that are derived from the Newton’s law of gravity. The basic form of this model gives the amount of bilateral trade between countries as a dependent variable to a ratio of product of sizes of both countries and a gravitational constant to distance between these two countries. By logarithmical transformation of this ratio a linear equation is obtained that can be enriched with other explanatory variables and the regressed with Ordinary Least Squares estimators. The empirical model that is actually used by the authors represents the sizes of the countries as a result of calculation considering GDP and population, and the distance as a result of calculations considering distance, area, and past colonial relationships. Additional variables included are the levels of environmental regulation determined as levels of CO2 emissions, levels of environmental taxes, and

others, then the levels of innovation described by number of patents filed, and auxiliary variable encompassing other control variables such as levels of foreign direct investment. In the analysis 20 developed countries are used as exporting countries and 148 other countries in the role of importers, and the analysis period goes from 1994 to 2005. The results support the Porter and van der Linde hypothesis.

A clearly positive effect of sustainability promoting legislation is documented in Germany by Holzer (2005), who states that the EEG act (Erneuerbare-Energien-Gesetz), as amended to comply with the directive EC 2001/77/EC of the EU, has clearly pushed the utilization of alternative electricity resources beyond expectations. Holzer however, does not perform her own research but bases this conclusion on other statistics such as those from the German Federal Ministry for the Environment that proclaims that application of the EEG Act reduced greenhouse gas emissions by 23 million of tons of CO2 equivalent. A drawback is that this act highly alters the market mechanisms and

virtually puts alternative energy producers away from competition. The effect on investment levels is clear.

(15)

from 1970 to 2006 in the form of Fuels and Lubricants PPI, and data from Bureau of Economic Analysis on National Product and Account. The main methodology used is Vector Auto regression. The reasoning behind the negative impact of energy price shocks on investment is that a rise in energy prices increases marginal costs and lowers demand for firm’s output due to reduction in consumer expenditures. These authors, as many others, studied asymmetry of response to energy price shocks. Their conclusions are valuable in that sense that they found rather perverse asymmetries, often being at odds with economical models and classical asymmetry. In a case of classical asymmetry there is a negative response of investment to energy price increase and a positive response to energy price decrease which is bounded above by the absolute value of the response to price increase. A perverse asymmetry, found by these authors, is anything not fulfilling the definition of classical asymmetry. This can be caused by too large a level of aggregation when doing research. If investments into mining industry (coal, natural gas, oil) are aggregated with other industries this may have quantitative impacts on the analysis. Nevertheless, impact of this study on our research of investment levels is marginal since this paper does not analyze a single industry and changes in policy affecting the industry. What makes this paper still interesting is that it examines how sensitivity of investment levels to energy prices changes in time. The authors found no evidence of change in responsiveness in investment in structures (with the exception of transportation) and some evidence of declining responsiveness of investments into equipment.

The above mentioned paper creates a challenge for us when examining the changes in investment level into renewable energy companies because there needs to be a filter imposed that will distinguish between the impact of supportive policies and a general market trend. Furthermore, the theoretical article of Constantini and Crespi (2007) and the findings of Holzer (2005) support our hypothesis that regulatory policies may increase the level of investment into renewable energies.

2.5 Company behavior when deciding for IPO

(16)

cycle theories that suggest that since a public company is more easily spotted and assessed by a potential acquirer it can be more readily sold. In addition, an entrepreneur that lists his company enables the company to be taken over for a higher price than he would get by an outright sale since acquirers cannot pressure outside investors into such pricing concessions as they can push the owning entrepreneurs. This does not always have to be the case, because by some IPOs the entrepreneurs are the ones who regain control, and the exiting party may be some venture capitalist. If the company is to be sold, it is more advantageous to go public as business angels or venture capitalists do not hold as diversified portfolios, which makes them willing to pay a smaller price for the company as diversified public market investors. Other justification is offered by the market timing theories which in principle argue that IPOs are made in times of bullish market to benefit from firm’s overvaluation in the stock market. In their research, Ritter and Welch (2002) come to the conclusion that the market plays a more important role than the life cycle of the company. They use results of other authors' research, rather than their own. One of them is done by Pagano et al. (1998) where they investigate Italian companies that are public, as well as those who are not, in the 1982 to 1992 decade. Their results point out that companies with higher market-to-book ratio and larger companies are more likely to go public and that this step reduced their costs of credit. Interestingly, they find that IPO activity follows high investment and growth and not vice versa.

As Lowry and Schwert (2002) point out, IPOs occur in cycles. More importantly, similar types of firms choose to go public at around the same time. Here is employed an effect of learning, where the positive information learned from the registration period of one company resulting in positive initial return subsequently encourages other companies to go public at higher valuation than previously expected. The authors study this phenomenon with the help of cross correlations of monthly volumes of IPOs and lagged initial returns on previous IPOs. They confirm their results by using Granger F-tests on initial IPO returns and numbers of IPOs in the US from 1960 to 1997 upon which they can conclude that high initial returns predetermine the consequent IPO volume.

(17)

supportive legal frameworks. This effect can be more observable if magnified by herding of IPO issuers. The herding theory of Nelson (2002) describes this behavior. Nelson uses the model of Bikhchandini, Hirshleifer, and Welch (1992) that has a constant underlying value (the value of the company relative to the stock price) of a company going public, which other agents are trying to guess. Each agent gets a noisy signal about the underlying value and then decides on an observable action. If one agent chooses to ignore this information and herds, this creates an informational cascade and other agents herd too.

According to the articles in this section, there are several reasons to believe that IPOs of renewable energy companies will occur after favorable legislative acts and due to herding and learning effect this will be magnified and thus observable more easily.

2.6 Our contribution to results of studied literature

(18)

shows us that such sensitivity, as mentioned above, exists, is measurable, and with certain characteristics. Thus, our first research hypothesis may be formulated as follows:

Ha1: |βt| > |βt+1|

Ha0: βt = βt+1

Where βt is the stock price sensitivity to exogenous factors in time t.

Furthermore, section 2.4 gives us hints, that apart from existing companies, additional renewable energy supporting policies can influence emergence of new companies and enlargement of the existing ones on the stock markets. Constantini and Crespi (2007) deal with this topic from an international trade point of view and they prove that environmental policies stimulate development of this segment. Edelstein and Kilian (2007) study sensitivity of investment rate to energy prices, but they do not look for influences of policies, which creates opportunity for us to do so and thus contribute to this field of research. This is indeed another way to test the opinion of investors about the expected real effects of legislation on the renewables business leading us to our second hypothesis:

Hb1: IPOlevelt < IPOlevelt+1

Hb0: IPOlevelt = IPOlevelt+1

Where IPOlevel stands for the level of IPOs in renewable energies sector.

If we can reject both H0 hypotheses, this means that we can measure the

effectiveness of legislative actions ex ante through financial market, before the product market is fully influenced by them.

3 Data description

(19)

3.1 Renewable energy stocks

Since we are studying impact of exogenous factors on renewable energy companies the core part of preparation for our research is gathering data on renewable energy stock prices.

3.1.1 ZEPHYR

(20)

deals if target is quoted and nearly 3000 deals for quoted acquirer. Only here, the companies in the deals were repeating themselves and were once in the role of target and once in the role of acquirer. Hence instead, we look only for companies with IPO deals because every listed company has to undergo an IPO. Secondly, we also do not restrict ourselves here only to IPOs that are already completed, since our other criterion is to have known value of the IPO, and so, if fulfilling this criterion, we can also analyze IPOs that are not yet completed. Restricting to IPOs produced a huge reduction in the dataset since out of the companies dealing with renewables, just a few are listed. After this step, the number of results decreased to 347. Restricting to known values of deals further contracted the dataset to 227. Our last criterion is that the deal status is “announced”, “completed” or “pending” whereas the reason for this was not to include deals that were aborted. This produced the final size of our preliminary ZEPHYR dataset, which are 183 deals. These deals represent companies in the role of Target, since that is how ZEPHYR classifies a company doing an IPO. From this point we are manipulating the dataset manually, excluding companies that do not deal with renewables and are included into the results by a coincidence.

For all the companies left, we attempt to download their daily history of stock prices from the DataStream. We download the prices in the form of Total Return Index (data type RI), that comprises capital as well as dividend returns of the particular stocks. We download all the prices in US dollars, so that when analyzing the impacts of crude oil and other commodity prices, we do analyze these impacts clean of exchange rates fluctuations. If the companies have their price history not available through DataStream we also try to download their price history through finance.yahoo.com, since this internet site also offers prices adjusted for dividend returns and splits. For prices quoted in other currencies, we transform them into US dollars with the help of history of exchange rates also downloaded from DataStream.

(21)

traded and its value reaches zero. From the point when the stock reaches zero for the first time we drop the stock out of portfolios that we construct for our analysis.

If it is not directly written in ZEPHYR, we also examine each company’s internet site to determine at what markets it performs its business the most. We do this, since our aim is to investigate and compare the impact of legal acts on the company value sensitivity to external factors. Therefore, when analyzing legal acts of different countries, we have to determine in which of these countries the company performs its business. This part of the data gathering is very time consuming since it involves searching for evidence in the company’s website on which markets it does significant parts of its business. However, this leads to creation of a comprehensive dataset out of which companies doing significant part of business outside European Union, USA, Canada, China, or Australia are excluded. In addition we also search for information on what type of renewable energy the company engages in, since we treat biofuel companies in a way differently as will be obvious later. Since we have identified the companies based on text search with words referring to some type of renewable energy this information is readily available.

As a result of previous steps, we arrive to a dataset of 114 companies. Out of these companies there is 79 companies which we identified as doing their business predominantly only in one of the investigated regions. These results can be compared to those of Pagano et al. (1998) who are also studying IPO behavior of individual companies and work with 69 non-financial corporations. In one respect, their sample is smaller although it has lower restrictions on segment type, but on the other hand they are studying only one country, so the sample does not have to be divided further, which gives them a big advantage in their research. Lowry and Schwert (2002) manage to analyze 586 IPOs on the US markets but also here, they do not restrict themselves to a particular segment and their data are gathered in a time span of 1960 to 1997 what is much longer a time period than ZEPHYR keeps track of. Nevertheless, we believe that our sample, even when it is smaller and for shorter time period, because it is focusing only on renewable energies will contribute to the knowledge about this sector.

3.1.2 Other internet resources

Apart from using ZEPHYR, we also use investor oriented internet sites to identify

(22)

http://www.renewableenergystocks.com/Companies/RenewableEnergy/Stock_List.asp

and http://altenergyinvestor.advfn.com. Advantage of these sites is that they also provide a description of the business of most of the companies. Since the companies are directly categorized according to the type of renewable energy they deal with, this part of information gathering is done very easily. Regarding the major market of these companies, if it is not possible to see it from the description of the company, there is always a link to the internet site of the company where this information can be learned more easily.

Since it is highly probable that companies mentioned in ZEPHYR will partially be the same as those on the internet, we have to deal with this issue as well. This part of data mining unfortunately could not be done automatically through MS Excel functions, such as the lookups, since the full names of the companies were sometimes misspelled or their forms of business such as “Ltd.” or “Corp” were used incorrectly. When a duplicate company is found we give priority to ZEPHYR since it also mentions its IPO value and delete the company from the internet list.

(23)

corporations for which it can be complicated to see how they are influenced by laws of one country. This is for two reasons. Firstly, more countries create new legislation at around the same time, so it is complicated to distinguish, which law is the one influencing stock sensitivity. Secondly, the multiple market companies may be too big and diversified to be influenced by exogenous factors to such extent that changes in their sensitivity can be measurable.

(24)

TABLE 2 - Summary of the results of renewable energy companies' identification process

ZEPHYR Internet TOTAL

1 company 5 companies 10 companies total companies

operating on the market

72 48 120 1. Nov. 1988 17. Jun. 1992 5. Jan. 1998

one market companies 47 35 82 6. Dec. 1988 3. Dec. 1996 14. Dec. 1998

28 106 134 3. Jan. 1973 30. May 1986 10. Jan. 1989

one market companies 9 86 95 3. Jan. 1973 30. May 1986 8. Feb. 1989

7 38 45 1. Sep. 1988 24. Jan. 1990 27. Mar. 1996

one market companies 5 22 27 4. Aug. 1989 2. May 1990 3. Mar. 2000

19 10 29 1. Dec. 1999 26. Mar. 2002 26. Jul. 2005

one market companies 10 9 19 1. Dec. 1999 9. Nov. 2004 11. Aug. 2006

14 19 33 1. Dec. 1999 26. Mar. 2002 26. Jul. 2005

one market companies 8 17 25 1. Dec. 1999 9. Nov. 2004 11. Aug. 2006 Note: "one market companies" are companies that do their business significantly only in one of the analyzed countries. "first occurrence" is specification of the first day when we have data available on a particular number of companies that we can include into the portfolio.

In this table, we present the numbers of companies that we identify through both methods of identification as well as how far back goes their stock price history. This information is used by determining the analysis starting date for every

European Union

first occurrence first occurrence

United States of America

Canada

China

Australia

(25)

TABLE 3 - Descriptive Statistics for Portfolios of Renewable Energy Stocks

ALL One Market ALL One Market ALL One Market ALL One Market ALL One Market

Observations 5095 5027 9181 9181 5095 5060 2392 2392 4854 4854

Start Date 1/9/1988 6/12/1988 3/1/1973 3/1/1973 1/9/1988 20/10/1988 12/1/1999 12/1/1999 4/8/1989 4/8/1989 End Date 12/3/2008 12/3/2008 12/3/2008 12/3/2008 12/3/2008 12/3/2008 12/3/2008 12/3/2008 12/3/2008 12/3/2008

Average Return 0.004 0.002 0.021 0.033 0.006 0.007 0.007 0.007 0.002 0.002 Standard Deviation 0.848 0.034 0.183 0.244 0.848 1.474 0.165 0.165 0.035 0.037 Median 0.001 0.000 0.002 0.002 0.001 0.001 0.000 0.000 0.000 0.000 Minimum -0.199 -0.249 -0.192 -0.192 -0.199 -0.331 -0.583 -0.583 -0.273 -0.273 Maximum 57.64 1.378 15.97 21.51 57.64 99.93 7.192 7.192 1.136 1.136 Skewness 4229 650.8 6399 6622 4229 4215 1505 1505 273.9 253.3 Kurtosis 63.49 18.84 76.13 77.80 63.49 63.47 35.01 35.01 10.03 10.04

Average Return 0.001 0.001 0.001 0.002 0.003 0.003 0.001 0.001 0.001 0.001 Standard Deviation 0.019 0.012 0.020 0.022 0.037 0.043 0.035 0.035 0.014 0.014 Median 0.0003 0.0001 0.0003 0.0001 0.0005 0.0004 0.000 0.000 0.0002 0.0001 Minimum -0.100 -0.124 -0.093 -0.093 -0.097 -0.175 -0.380 -0.380 -0.138 -0.138 Maximum 0.743 0.376 1.230 1.352 1.768 2.004 0.913 0.913 0.330 0.330 Skewness 506.5 261.0 1991 2010 1333 1222 210.6 210.6 101.9 100.1 Kurtosis 14.87 10.26 38.17 38.81 31.42 29.53 7.907 7.907 4.785 5.004

Note: "ALL" stands for all companies engaging in the particular market, "One Market" stands for companies engaging only in one particular market

In this table we present the standard descriptive statistics for portfolios constructed from renewable energy companies relevant for a particular region or country. Construction of these portfolios is described in section 4.1 and the methodology for calculating returns of these portfolios is explained in apperndix 3.

(26)

3.2 Legal frameworks incentivizing use of renewable

energies

One of the most crucial parts of our dataset are the histories of legal acts that support the use of renewable energies in the investigated regions. These we use as the milestones (alternatively called as landmarks) in every country to divide the stock price histories of respective renewable energy companies into periods with different level of legal support. To do this for every region we also had to choose what types of legal acts will we put into consideration. Since in every system there are laws which set the general targets and policies but there are also abundant regulations which give specific instructions how to put the idea of the law into practice we had to choose such a type of act that will be feasible to process. Because regulations often come about quite often throughout the time until the general law is superseded and they may consider only some part of the industry or type of the renewable energy, it would be difficult to find landmarks among regulations that influence our whole portfolios. Using regulations would require us to divide the dataset into too many small pieces and also probably to divide portfolios according to type of renewable energy. This however, would pose one significant problem, and that is that there are companies that engage in more renewable energy types at the same time. Therefore, to avoid dividing the data into small periods, and in every period to perform analysis on every type of renewable energy separately, we decide to regard only laws, or EU directives, as the major events. We consider laws to be sufficient, as they mandate the governmental offices to execute the policies and they already state the main targets and incentives, so that the market participants can already predict the course of action government is going to take. Apart from some exceptions, mainly in the EU, the laws also cover all the renewable energy sources. We hold to the principle of regarding only laws as the main milestones as consistently as possible with only a few exceptions that are underscored and explained.

(27)

3.2.1 European Union

The region of the European Union is processed as one country since legal acts of EU have a heavy influence on the support of renewable energy usage. This is because member countries have to adapt their legislation to the legislation of the EU, as it is mentioned for example by Holzer (2005) who analyzes the impact of German legislation that was amended based on existence of the 2001/77/EC directive. As European Commission (2007) mentions, directives 2001/77/EC and 2003/30/EC set targets for the share of renewable energies and require the member states to set their own targets reflecting the values stated in the two directives. This is the basic mechanism of influence of the European directives on national legislations.1

Another specific trait of the European renewable energy policies is that, despite the fact that EU is now perceived as the leader in green matters and sustainability, it reached this position with only a few legislative acts. This points out the power that just a few but crucial regulations can have on the market. We of course cannot forget that we have this impression also due to the fact that the burden of complex regulations touching every type of renewable energies is still on the individual members of the community.

Events that are important for renewable energies in the EU, but are not included in our analysis due to unavailability of data on interest rates are listed below in numbered points. After these we describe in more detail the landmarks that we used for our analysis. When deciding what landmarks to select, primarily we base this decision on own experience which was connected to European renewable energy support. To back our decision, we consult the document of European Commission (2002) that states which pieces of legislation are regarded to be the most crucial ones. For more recent acts, we support our decision with Hendricks (2008). In addition we also confront our findings with International Energy Agency (2008).

Unused events listed here are not directives, but at that time were the only main community documents promoting renewable energy use:

1. White Paper on Energy policy, COM(95)682 Final from 18th December 1995. This paper has already objectives on environmental

1

(28)

protection but its aim is only inviting community and national authorities to adopt necessary policies.

2. The first effective document is the White paper on Renewable energy Sources & Action plan, COM(97)599, from 26th November 1997. This paper, although not being a directive, sets the first targets for share of renewable energy and also sub-targets for particular sectors. The first general target that it set was to double the share of gross renewable energy share from 6% to 12% by 2010.

The first landmark in the European legislation according to our criteria is the directive 2001/77/EC on the promotion of renewable energy resources for the electricity market that has come into force on 27th October 2001. This directive concerns, however, only electricity production and does not deal with biofuels. The issue of biofuels was tackled only in the year 2003 by the directive 2003/30/EC on the promotion of use of biofuels or other renewable fuels for transport. Since in our portfolios we are considering all renewable energy companies together, we will analyze periods before 2001/77/EC and after 2003/30/EC come into force that is before 27th October 2001 and after 17th May 2003.

The second event is the adoption of EU targets for share of renewable energies on energy consumption of the member countries that has been agreed upon on the EU summit on 9th march 2007. The targets are 20 % reduction of CO2 emissions by 2020

compared to 1990 and 20% share of renewable energies by 2020. This event is not a directive but still the targets are binding, moreover they are a follow up to the to the Renewable Energy Roadmap that was communicated by the European Commission at 10th January 2007, and they lead to the proposal of directive on the promotion of the use of energy from renewable sources communicated on 23rd January 2008. This directive if adopted and published, would serve as another landmark superseding the 2001/77/EC and 2003/30/EC directives. (Hendricks, 2008)

(29)

TABLE 4 - Overview of Milestones Used for EU

Name of law/directive date of coming into force

White Paper on Energy Policy, COM(95)682 Final 13 December 1995 White Paper on RE sources & Action Plan, COM(97)599 26 November 1997

Directive 2001/77/EC on the promotion of RE for electricity 27 October 2001 Directive 2003/30/EC on promotion of use of biofuels for transport 17 May 2003 Adoption of EU targets for share of RE in energy consumption* 9 March 2007

Used milestones Unused milestones

Note: RE stands for renewable energy, * stands for such a milestone that is not a law/direcive but we still use it in our analysis

3.2.2 United States of America

For the USA we are restricted with the availability of data on interest rates as well, since the history of acts that promote renewable energy production goes as far back as to 1978. The description of the effects of all the acts mentioned in this section comes from Energy Information Administration (2008) which we also used as a guide as to what acts to include in our analysis. This is because this agency presents these acts based on their impact on the whole renewable energy market. Those acts that we present and analyze here are those which are officially deemed to have the widest influence. In addition, we also confront our findings with International Energy Agency (2008).

List of those acts not analyzed due to unavailability of other data follows: 1. The Energy Tax Act of 9th November 1978 that provides income tax

credits on equipment expenditures for alternative electricity sources. 2. The Crude Oil Windfall Profits Tax Act which became a public law

on 2nd April 1980 increased the energy tax credits laid out in the Energy Tax Act and extended these credits also to investment into equipment used for converting biomass into fuels. It also allowed for tax-exempt interest on renewable energy industrial development bonds.

3. The Economic Recovery Tax Act of 13th August 1981 which allowed for accelerated depreciation of capital.

(30)

Crude Oil Windfall Profits Tax Act back to the level guaranteed by the Energy Tax Act of 1978.

The first event in our work is the Energy Policy Act, known as EPAct, which has become a public law on 24th October 1992. This law instigates an incentive of 1,5 US cents per kWh of energy produced from renewable resources including biomass from publicly owned utilities and rural cooperatives known as Renewable Energy Production Incentive. Another provision of this act introduces production tax credit for wind and biomass facilities that are privately and investor-owned. Even when this law should not affect private, quoted, producers of renewable energy it affects producers of equipment and that is why we still regard it as an important point supporting the industry.

An extension to the provisions of the EPAct of 1992 is provided in the Tax Relief Extension Act which became a public law on 17th December 1999. The extension concerns the above mentioned production of energy through usage of wind and biomass so that also government owned facilities are eligible for it. Another provision of this law is giving the opportunity for a 20% tax credit for incremental research expenses incurred in a trade or business.

We do not use the Economic Security and Recovery Act of 2001, that was signed by the president of the USA on 9th March 2002 as it has only impact on the availability of production tax credits for wind and biomass facilities which were under the EPAct available only for 10 years. Under this law, the tax credits are available until 31st December 2003. Therefore, this law does not provide any incremental change in the incentives already in place and only concerns production tax credits for wind and biomass facilities. Thus, the provisions of EPAct 1992 that introduce the 1,5 US cents pre kWh incentives for production of energy from all the types of renewable energies remain in force.

(31)

purchase renewable energy equipment for home use and purchase of cars that use alternative fuels.

As it is visible from the text above, we will be using here three. Dates of becoming into force of all the laws mentioned here are taken from The Library of Congress (2008). Table 5 summarizes the acts mentioned above.

TABLE 5 - Overview of Milestones Used for USA

Energy Tax Act of 1978 9 November 1978

Crude Oil Windfall Profits Tax Act of 1980 2 April 1980 Economic Recovery Tax Act of 1981 13 August 1981

Tax Reform Act of 1986 22 October 1986

Economic Security and Recovery Act of 2001 9 March 2001

Energy Policy Act of 1992 24 October 1992 Tax Relief Extension Act of 1999 17 December 1999

Energy Policy Act of 2005 8 August 2005

Unused milestones

Used milestones

3.2.3 Canada

Unlike other regions so far mentioned, in Canada we did not find legal acts at the level of federal laws that we cannot include due to unavailability of other data. This can be partially caused by a relatively large level of independence that provincial governments enjoy and partially also by a perceived lagging of Canadian renewable energy policy compared to EU or USA. (The Pembina Institute, 2008) The decision as to which legal acts to choose for analysis is based on the report by The Conference Board of Canada (2003), the report of Brown (2007) originally prepared for International Energy Agency, and International Energy Agency (2008) itself. These sources give overview of the major renewable energy initiatives and their effects. Descriptions of the acts and dates of their coming into force are taken from Department of Justice of Canada (2008).

(32)

Another point that we will use in our analysis is an amendment to the Income tax act of 1985 made by the Budget Measures Implementation Act adopted on 2nd June 1994. This law amended the Income Tax Act to enable Accelerated Capital Cost Allowance for the investments into renewable energy equipment. Under this provision, taxpayers can make an accelerated write off of up to 30% of certain equipment that is purchased for the purpose of producing energy from alternative energy sources. (Brown, 2007)

Next landmark is the so called Action Plan 2000 on Climate Change that was adopted on 6th October 2000 with the aim of meeting at least 1/3 of the targets lined out in the Kyoto Protocol. Under this action plan, the Market Initiative Program was established which has the aim of creating such renewable electrical energy sources that they are competitive on the market by 2010. This program has funding of 25 millions of Canadian dollars with maximum of 5 millions dollars per recipient. The funding was available until 31st March 2006. (Conference Board of Canada, 2003)

Of particular importance is the establishment of the Canada Foundation for Sustainable Development Technology which was carried out by the Canada Foundation for Sustainable Development Technology Act in force from 22nd March 2002. This foundation has the aim to finance the development and demonstration of clean technologies which provide, among others, solutions to climate change problems. The foundation manages a fund of 550 millions Canadian dollars. On average, the foundation will finance 33 percent of eligible projects. (Brown, 2007)

(33)

we would start the analysis in this case after 1st April 2008. However, since this day is beyond the scope of our analysis, we here make a major exception to our rule, and take into account 17th January 2007 from when the existence of the programs was already officially anticipated. (Brown, 2007; Office of the Prime Minister of Canada, 2007; ecoACTION 2007; ecoACTION, 2008)

As a result of our analysis we use 5 milestones for Canada which are summarized in table 6.

TABLE 6 - Overview of Milestones Used for Canada

Energy Efficiency Act of 1992 22 June 1992 Budget Measures Implementation Act of 1994 2 June 1994 Action Plan 2000 on Climate Change 6 October 2000 Canada Foundation for Sustainable Development Technology Act 31 March 2006

ecoEnergy programs* 17 January 2007

Used milestones

Note: * stands for such a milestone that is not a law/direcive but we still use it in our analysis

3.2.4 China

When making our decision on which regulations to use in our analysis as landmarks for China, we considered the article from Stender and Lestelle (2006) and the database of International Energy Agency (2008). These sources list Chinese laws for the segment of renewable energies with various importance and breadth of impact. As a result of our research we identify three policies that have an impact on all the types of renewables in this country.

The first important legal act in China that falls into our analysis period and meets our criteria is the Renewable Energy law effective from 1st January 2006. This law sets middle and long term targets for the total volume of renewable energy development. Developers that will obtain a permit to proceed with their projects will be selling energy to grid operators at guaranteed prices. In addition there will be a special fund set up for the development of renewable resources. (International Energy Agency - Renewable Energy Law, 2008; Stender and Lestelle, 2006)

(34)

quantified target for reducing greenhouse gas emissions. Nevertheless, it sets an energy efficiency objective to reduce the energy consumption per unit of GDP by 20% by 2010 and to quadruple the GDP by 2020 with only doubling the energy consumption. (International Energy Agency - National Climate Change Program, 2008)

From September 2007 there are preferential tax policies in force that encourage the development of energy conservation and renewable energy. The policies include cut on income tax for the producers and consumers as well as a reduction of the import tax for “green” equipment. We are using these policies as a milestone as well in our analysis. (International Energy Agency – Preferential Tax Policies for Renewable Energy, 2008)

All the above mentioned acts are summarized in table 7 below. TABLE 7 - Overview of Milestones Used for China

Renewable Energy law 1 January 2006

National Climate Change Program* 4 June 2007 Preferential tax policies for development of RE* 1 September 2007

Used milestones

Note: * stands for such a milestone that is not a law/direcive but we still use it in our analysis, RE stands for renewable energy

3.2.5 Australia

(35)

results up to 2003 when their paper was published. They also identify other policies but these are state based, which is the reason why we do not use them, as in the case of EU, USA, and Canada.

The first major law is the Renewable Energy (Electricity) Act of 2000 in force from 18th January 2001 that gave rise to the Mandatory Renewable Energy Target (MRET) valid from 1st April 2001. According to this target it is required to produce 9500 GWh of electricity from renewable resources per year by 2010 and 45 000 GWh by 2020.

Another important change to legislative framework of Australia concerning the renewable energies was the Renewable Energy Amendment Act in force from 11th September 2006. This law sets time limits for creation of renewable energy certificates. Further, it accommodates changes in the renewable energy industry as they were assessed through the process of the review of MRET that commenced in 2003 and resulted in this amendment act. (Office of the Renewable Energy Regulator of Australian Government, 2008).

A summary of these milestones is presented in table 8 below. TABLE 8 - Overview of Milestones Used for Australia

Renewable Energy (Electricity) Act of 2000 18 January 2001 Renewable Energy Amendment Act 11 September 2006

Used milestones

3.3 Interest rates

(36)

TABLE 9 - Descriptive Statistics of Benchmark Interest Rates

EU USA Canada China Australia

Date First Published 31/12/1998 3/1/1986 2/5/1990 10/1/2002 24/9/1986 Vendor LDN:BBA LDN:BBA LDN:BBA People's Bank of China LDN:BBA

Observations 2400 5789 4661 1610 4854 Start Date 31/12/1998 3/1/1986 2/5/1990 10/1/2002 4/8/1989 End Date 12/3/2008 12/3/2008 12/3/2008 12/3/2008 12/3/2008 Average Return 0.0002 -0.0001 -0.0002 0.009 -0.0001 Standard Deviation 0.006 0.009 0.014 0.149 0.008 Median 0.00006 0.000 0.000 0.000 0.000 Minimum -0.086 -0.134 -0.178 -0.725 -0.128 Maximum 0.145 0.103 0.326 2.333 0.137 Skewness 194.5 30.04 100.9 69.69 63.64 Kurtosis 4.034 -1.383 4.085 5.628 0.204 Average Return 0.00006 -0.00008 -0.00012 0.00003 -0.00007 Standard Deviation 0.003 0.004 0.006 0.056 0.003 Median 0.00003 0.000 0.000 0.000 0.000 Minimum -0.039 -0.062 -0.085 -0.560 -0.060 Maximum 0.059 0.043 0.123 0.523 0.056 Skewness 169.8 32.82 75.73 26.22 63.43 Kurtosis 2.548 -1.798 2.528 -0.094 -0.533 Log Returns Normal Returns

Note: LDN:BBA stands for London British Banker's Association. Start dates of descriptive statistics are identical to start dates of descriptive statistics of porfolio returns, as presented in table 3 or if such a long history is not available, date of first publishing is used.

In this table we present the standard descriptive statistics for 3 months interbank offered rates that we use for a particular region or country. Calculation of returns of these rates is explained in apperndix 3.

(37)

European Union as a whole. This adds to our consistency, even though we believe that these rates are correlated to EURIBOR rates. We have downloaded all the interest rates on a daily basis from DataStream.

3.4 Indices

Since we want to study sensitivity of renewable energy companies to exogenous factors, it is hard to produce a model where all the exogenous factors will be explicitly included. To represent those factors that we do not want to analyze in detail one by one, we include into our analysis equity indices. We take country specific indices since these capture local conditions and thus help to capture other country specific factors, including also those legislative ones, that influence all the companies and not only those engaging in renewable resources. This approach is justified also because countries like USA adopt new legislation often in packages which comprise acts influencing different individuals and business types or the whole environment. In the European Union this measure is of course still quite rough since it is hard to estimate how is the influence of national laws of the member countries reflected in a pan European index. Nevertheless, a regional or national index helps us to filter the influences of acts that are general and not touching specifically renewable energy companies. These indices, moreover, are also influenced by global factors which have gross impact on the whole economy. Thus, we are not omitting these factors either. If we selected global indices, such as S&P 500, these track only the 500 largest companies, which may not have much in common with companies listed in Australia or China.

As a vendor for the indices we used MSCI Barra that publishes a comprehensive set of indices for countries, industries, types of stocks, etc. We work with the country indices which are available for every country investigated and Europe as well. The MSCI Europe ex. Switzerland index encompasses only countries of the old EU 15, which suffices because most of the renewable companies reside in these counties and the advantage of this is that there are no European countries that are not influenced by the directives of the European Commission.

(38)

downloaded from DataStream in US dollars on a daily basis. Table 10 shows an overview of indices that we are using.

TABLE 10 - Descriptive Statistics of Equity Index Returns

EU USA Canada China Australia

Date First Published 31/12/1969 31/12/1969 31/12/1969 30/12/1992 31/12/1969 Vendor MSCI Barra MSCI Barra MSCI Barra MSCI Barra MSCI Barra

Observations 5219 9181 5059 2392 4854 Start Date 1/9/1988 3/1/1973 20/10/1988 12/1/1999 4/8/1989 End Date 12/3/2008 12/3/2008 12/3/2008 12/3/2008 12/3/2008 Average Return 0.0004 0.0003 0.0004 0.001 0.0003 Standard Deviation 0.011 0.010 0.010 0.019 0.012 Median 0.000 0.000 0.001 0.000 0.000 Minimum -0.121 -0.204 -0.093 -0.102 -0.099 Maximum 0.115 0.173 0.054 0.117 0.077 Skewness 21.58 41.77 5.651 3.166 3.713 Kurtosis 0.706 -0.466 -0.507 0.102 -0.224 Average Return 0.000 0.000 0.000 0.000 0.000 Standard Deviation 0.005 0.004 0.004 0.008 0.005 Median 0.000 0.000 0.000 0.000 0.000 Minimum -0.056 -0.099 -0.042 -0.047 -0.045 Maximum 0.047 0.069 0.023 0.048 0.032 Skewness 21.477 48.827 6.168 3.138 3.964 Kurtosis 0.332 -1.179 -0.624 -0.044 -0.325 Log Returns Normal Returns

Note: Start dates of descriptive statistics are identical to start dates of descriptive statistics of porfolio returns, as presented in table 3 or if such a long history is not available, date of first publishing is used.

In this table we present the standard descriptive statistics for indices that we use for a particular region or country. Calculation of returns of these indices is explained in apperndix 3