7/27/2019 Investing in Canada's Climate
1/25
INVESTING IN CANADAS CLIMATE: AN
EMPIRICAL ANALYSIS OF RISK AND RETURNS TO PORTFOLIOS OF
CANADIAN ASSETS BASED ON ENVIRONMENTAL PERFORMANCE
Shamai Cohen
ECON 490
April 26, 2013
1
7/27/2019 Investing in Canada's Climate
2/25
Abstract: Environmental factors are increasingly being considered by both investors and corporations,
in terms of socially responsible investing (SRI). The theoretical framework regarding SRI and corporate
environmental initiatives suggests that the effect of environmental performance on returns is ambiguous
due to factors working in opposite direction. We take an empirical approach to analyze the effect of
environmental performance on returns in Canada. Using third party environmental and emissions data
we conduct a portfolio analysis using the CAPM and multi-factor models to estimate susceptibility
to market volatility. In the Oil and Gas sector, portfolios of strong environmental performers have
significantly lower betas, suggesting that the potential costs imposed by regulation are taken into
consideration when investment decisions are made.
CONTENTS
Section 1: Theoretical Background 3
Introduction to SRI: the moral use of money 3
The purpose of SRI: does it have an impact? 5
The returns to SRI: a disputed area 8
Section 2: Question Formation 10
The environment, public policy and risk 10
Four questions and hypotheses 12
Section 3: Empirical Analysis 14
The Model: CAPM and Fama and French 14
Data sources and portfolio formation 15
Table 3.1 Emission Reduction Plan Portfolios 16
Table 3.2 Sample Environmental Score Summary 16
Regression Results 17
Table 3.3 CAPM results 17
Table 3.4 Fama French 3 Factor results 17
Section 4: Discussion 19
Methodological concerns and a look at Sadorsky 20
SRI, returns and risk in light of empirical results 21
Section 5: Conclusion 22
Bibliography 23
2
7/27/2019 Investing in Canada's Climate
3/25
SECTION 1: THEORETICAL BACKGROUND
There is considerable academic interest in the use of non-financial characteristics in decisions in the
finance industry. In a cornerstone paper by Sally Hamilton on the performance of mutual funds that look
at non-financial qualities related to corporate social responsibility (CSR), she determines the possible
causes of different rates of return on these alternative funds. The term socially responsible investment
(SRI) is a term that denotes investment that includes factors related to CSR in the decision making
process. The three following possibilities are explored:
1. Corporate social responsibility is not priced. Therefore SRI does not add or destroy value in
terms of risk-adjusted returns.
2. SRI involves assets with inherently less risk, and drives down the capital and returns. Therefore
SRI portfolios deliver lower returns compared to regular portfolios.
3. SRI involves firms with inherently better fundamental performance. Therefore SRI portfolios
deliver higher returns compared to regular portfolios.
The goal of this paper is to contribute to the body of work that uses empirical analysis to find out which
effect is strongest. Testing returns to portfolios of assets using SRI criteria can help decide which of
the above hypotheses is correct. This paper is structured as follows: in section 1 the theory pertaining
to hypotheses 2 and 3 is discussed (1 is taken as the null hypothesis), in section 2 the question for this
study is formed, in section 3 the methodology and results of this empirical study are given, section 4 is
a discussion of the results and expounds on what conclusions may be drawn from them, lastly section 5
contextualizes the conclusions by making suggestions about policy implications.
INTRODUCTION TO SRI: THE MORAL USE OF MONEY
3
7/27/2019 Investing in Canada's Climate
4/25
Money and its use have always been subject to moral considerations. Often, but not necessarily,
this occurred through religion Islam, for example, prohibits the lending of money for profit. A
Methodist preacher named John Wesley gave a sermon in 1744 titled The Use of Money, in which he
urged his congregation to Gain all you can but to avoid causing harm to the mind or health of oneself
or ones neighbors. Fast forward to present day and shed the tone of religiosity, and there is still a case
to be made for ethical concerns regarding the use of money, and more people are beginning to take
note.
The common saying goes put your money where your mouth is, meaning that to really
support a cause you think is valuable, you must show this with investment rather than just words.
One global cause that has become mainstream in the arena of politics and business is that of the
environment. This cause generally encompasses things like the loss of biodiversity, destruction of
natural habitats, as well as degradation of water sources, release of pollutants and waste into the
oceans and atmosphere, and in the case of atmospheric gasses, the possibility of anthropogenic
climate change. As the environmental movement has grown and matured over the last half century,
awareness of these issues has, too. There is a growing breed of sustainable consumers that desire food
and consumer products to be of a higher standard to reduce the negative environmental impact (think
organic produce and green cleaning fluids). These consumers also want financial products that fit with
their values (Thorgerson 2002).
Step into the world of socially responsible investing (SRI). Though ethical or green mutual
funds have existed for a while, the rate at which large chunks of capital are being directed towards them
is impressively high. In 2006 the United Nations created the Principles for Responsible Investing (PRI)
as a way for investors to publicly commit to a set of goals related to SRI, and as of 2012 the group of
signatories represents over $30 trillion in assets under management. In the United States in 2011, 12.2%
of all assets tracked by Thomson Reuters Nelson fell into the category of SRI (US SIF).
4
7/27/2019 Investing in Canada's Climate
5/25
7/27/2019 Investing in Canada's Climate
6/25
The overall effect of shareholder activism is not the subject of this paper. In a literature review
of academic papers on the topic (Sjostrom 2009) the effects of shareholder activism are ambiguous,
though most papers take a skeptical stance on its effectiveness. Shareholder activism may or may not
impact CSR, which includes potential environmental projects. If shareholder activism effects returns
exogenously to CSR outcomes, that may be a concern for this paper because those firms with better
CSR are also more likely to be influenced by shareholder activism. If this effect was significant then
the exclusion of shareholder activism from this model would make it subject to omitted variable bias.
However the empirical part of this paper looks at environmental performance as a given value, and
seeing that the effect of shareholder activism is ambiguous, the author does not think that this is an
issue of concern.
Another interesting possibility to consider is that ethical investors, through the process of
excluding companies with poor CSR, increase the cost of capital for them (Renneboog 2011). When a
large enough group of investors decide to divest from a company, the remaining investors incur higher
risks (as the same risk is shared by a smaller number). This has a downward effect on stock prices. If
the total increase in cost of capital becomes greater than the costs associated with making the reforms,
then it is now in the companys interest to do so in order to be considered by the ethical investors
again. This process is the topic of a paper by Robert Heinkel and in their (purely theoretical) model they
do observe the incentivization of reforms, but only at a threshold where at least 20% of investors are
excluding companies for environmental performance. Ultimately, this mechanism may cause investment
in poor environmental performers to give a higher rate of return, but at a greater risk. This mechanism
relates to hypothesis 2 from Hamilton, where SRI receives lower return to investment because of the
lower inherent risk.
The total assets under management in Canada under SRI in 2010 were reported to be $530.9
billion or 19.1% (Social Investment Organization 2011). This number lies too close to the theoretical
6
7/27/2019 Investing in Canada's Climate
7/25
threshold point to claim that the effect of ethical investment on costs of capital is negligible. As a
result, we must consider the possibility that this effect will impact returns endogenously through
environmental performance. If returns to good environmental performers are lower, then this effect is
the dominant one.
Working in the opposite direction is what we will term the fundamental effect. Theorists claim
that companies who are able to make the investments to improve their CSR are also more likely to be
leading performers in other fundamental aspects of business (Cormier 2007). By this logic, investing in
companies with top-notch performance as measured by ESG indicators should yield better returns due
to better company performance. This effect is the basis for hypothesis 3 as proposed by Hamilton. If
returns to portfolios of good environmental performers are higher, then this effect is the dominant one.
The final confounding factor related to SRI is the collection and use of data by investors. The
effect we call reporting bias is that companies are far more inclined to disclose information that is of
a positive nature (Cormier 1999). Put differently, no news isnt bad news, but any news at all is likely
good. Of course this effect occurs to varying degrees, and is dependent on the local regulations and
culture surrounding disclosure. In Canada, there is less of a reliance on mandated disclosure than
in the US (Buhr 2001). Companies are required to meet many standards related to environmental
performance, so voluntary disclosure of these performance indicators plays an important role in the
companys disclosure strategy (Cormier 1999). For our concerns, this causes a problem in sample
selection because we will only include in our analysis companies for which we have all the required data.
There has been a steady increase in environmental disclosure among Canadian firms, causing
us to ask the question whether the increase is due to do better environmental outcomes by firms, or
whether there is some increase in external pressure to disclose. If it is the former then our problem of
sample bias is likely magnified. However, if it is the latter, then it is likely that poorer performers are
also increasing disclosure in the face of such pressure. Unfortunately, enlightening the root cause of
7
7/27/2019 Investing in Canada's Climate
8/25
disclosure for each company used in this study would be too difficult a task to complete. In light of
our firm specific statistics, and our methods of controlling for other variables, we believe our sample
contains sufficient observations from both good and poor environmental performers.
THE RETURNS TO SRI: A DISPUTED AREA
Numerous studies exist looking at returns to ethical investment. There is some variability in the
methodology and sample selection, so here is presented a brief review of some of the literature on the
topic. First a brief overview of methodologies will be given. Then the results for each study are reported
with some context. Lastly, the implication of these results will be discussed and related to form the null
hypothesis of this study.
Returns to SRI can be measured in a few different ways. Some studies look exclusively at mutual
funds and compare ethical ones to conventional ones (Bauer 2005) while others compare ethical indices
to traditional ones (Managi 2012), and still others create their own portfolios based on specific ESG
criteria to perform comparative analysis (Ziegler 2011). The differences here are largely due to the
specific question being answered by the empirical analysis. The rationale for sample selection in this
study will be reviewed in section 2.
Most studies use a single factor model that estimates an assets abnormal return and
susceptibility to market risk. This model is called CAPM. Various studies have implied that CAPM is not
the best tool for conducting this sort of analysis (Chretien 2008), and more useful models have been
created. Most studies that use the CAPM also employ a more robust model: based on Eugene Fama and
Kenneth French (1993) or Mark Carhart (1997) they use a three or four factor model, respectively. The
basis for using both CAPM and a multi-factor model seems to be widespread, so this study considers
that the standard.
8
7/27/2019 Investing in Canada's Climate
9/25
MANAGI, SHUNSUKE ET AL. DO SOCIALLYRESPONSIBLEINVESTMENTINDEXES OUTPERFORM CONVENTIONAL
INDEXES? (2012)
This study compares ethical indexes with traditional ones in an international comparison that spans the
US, Japan and the UK over the years 2001-2008. They find that both indexes exhibit cyclical regimes
(bull and bear), and that the cycles coincide across the SRI and traditional stock indexes. Moreover, the
means and variances on return were not statistically significant, suggesting that there is little difference
between SRI and non-SRI activities.
BAUER, ROB ET AL. INTERNATIONAL EVIDENCE ON ETHICAL MUTUAL FUND PERFORMANCE AND INVESTMENT STYLE
(2004)
Using a database of ethical mutual funds and regular mutual funds in the US, UK and Germany, Bauer et
al. conduct a comparative analysis using the Carhart 4 factor model. They find no significant difference
in risk adjusted returns over the period 1990-2001. Their analysis also casts doubts over the ability of
ethical indexes to incrementally explain returns to ethical funds.
ZIEGLER,ANDREAS ET AL. DISCLOSED CORPORATE RESPONSES TO CLIMATE CHANGE AND STOCK PERFORMANCE: AN
INTERNATIONAL EMPIRICAL ANALYSIS (2011)
Ziegler et al. are interested to find out if disclosing a response to climate change (in the form of a climate
impact statement or an emissions reduction plan) has an impact on the returns of a corporation. Their
study is limited to the USA and EU. They use CAP-M and Carhart 4-factor model to chart returns of
portfolios created using two binary variables related to specific disclosure questions. In the time-series
analysis over the years 2001-2006 they find the excess returns for disclosure are higher in the EU than in
the USA, and that within the USA the excess returns for disclosure are higher in the energy sector than
in other sectors (where returns to disclosure are negligible).
9
7/27/2019 Investing in Canada's Climate
10/25
This sample from the literature is in no way exhaustive, but highlights a few interesting results.
There is little indication of significant differences when assessing SRI at the index or fund level. At the
asset level it appears to be possible to create portfolios that have significantly different results based
on CSR indicators. This study borrows greatly from the methods of Ziegler and focusses on the most
granular level by creating unique portfolios bases on ESG data, rather than ethical funds or indexes.
Also it should be noted that the Canadian sector has had less academic interest, probably because it is a
smaller market.
SECTION 2: QUESTION FORMATION
In this section the theory that pertains particularly to this study is outlined. Factors related specifically to
the environment and regulatory situation in Canada are discussed. Methodological issues are raised as
they pertain to the questions, leaving the models themselves to be explained in the next section.
THE ENVIRONMENT, PUBLIC POLICY AND RISK
Environmental concerns are embodied in a wide variety of corporate activity. So much so,
that narrowing down the list to something comparable across industries poses a problem. One type
of pollutant has been in the public eye extensively over the last decade and this is the category of
greenhouse gasses (GHGs). Carbon dioxide is a prominent GHG and though others contribute, the
industrial measure for emissions is called carbon emissions and is standardized such that other GHGs
get converted to their carbon equivalent. The overall phenomenon of climate change is taken to be the
increased anthropogenic emissions of GHGs that has led to an upward trajectory of global temperatures.
Climate change is a global issue and 194 countries are signed to the UN Framework Convention on
Climate Change (UN FCCC 2013). Because of the ubiquity of climate change, and the fact that the ethical
consumer drives demand for ethical financial products, we believe that a companys stance on climate
issues is likely to be considered a high priority for investment managers in SRI.
10
7/27/2019 Investing in Canada's Climate
11/25
Leaving aside the impact of SRI, what are the costs and benefits of corporate projects designed
to improve environmental performance? It is clear that for a company to engage in a project the cost
must be offset by benefits that come privately in the form of efficiency savings, or publicly if there are
incentives given by government regulation. Some environmental projects are undertaken as part of that
first category, but in the case of GHG emissions there is a clear case of tragedy of the commons in the
sense that private benefits to curbing emissions are negligible if any. Because we are still dealing with a
negative externality, the government ought to play a role and provide public incentives to curb
emissions. There are many forms of doing this, but most involve an emission ceiling, where emissions
beyond a certain point are charged a penalty (Jones 2007).
It is theorized that overcompliance with such regulations may be incentivized through financial
markets due to uncertainty around legislation of regulations (Mallory 2012). The mechanism for this is
two-fold, first as a signal of company value, and second as a method of pre-emption. As a signal,
enacting costly projects shows that the company has healthy finances and the capability to complete a
capital-intensive project, implying to investors the company is doing well in other areas. As a method of
pre-emption, improving a companys environmental performance reduces risk in the face of an
uncertain legislative environment. Regulations may be made stricter and pre-empting them allows
investors to attribute less risk to those assets. Also, by way of pre-emption, overcompliance may signal
to the government that the current regulations are not sufficient or need to be improved and then an
increase in regulation that follows will impose costs on the companys competitors, making their
position in the market stronger.
The legislative environment is full of uncertainty and varies country by country. The propensity
of a government to enact environmental regulation cannot be modelled, but we may look at regional
differences to get an idea of comparative likelihood. Canada, for example, signed the Kyoto Protocol
in 1997, but opted out of it in 2011 after more than 5 years of not attempting to achieve the emissions
11
7/27/2019 Investing in Canada's Climate
12/25
standards (IPCC 2012). The United States never signed or ratified the Kyoto Accord, while the European
Union did and remains committed to the reduction plan. Based on these differences, Ziegler expected
higher returns to good environmental performers in Europe, but not in the US (and confirmed that
result). Canada lies in between the two so the results should be interesting and help determine if
the regulatory environment impacts returns. Within this framework a subsector analysis will also be
conducted. The oil and gas sector is subject to more stringency in terms of regulation due to the high
amount of emissions it is responsible for. This makes the effect of environmental performance more
important. Ziegler found higher returns to good environmental performers in this sector in the United
States, and this study hopes to replicate that result in Canada.
So, given that compliance with future environmental regulations is a costly risk carried by firms,
it follows that investors would likely view good environmental performers as less risky. Assets that carry
less risk are going to have a lower rate of return to induce investment. Working in the opposite direction
is the theory that good environmental performance goes with good performance in fundamental
financial areas. So which way does the overall effect go?
FOUR QUESTIONS AND HYPOTHESES
Specifically, our questions are as follows:
(1)Compared to portfolios of poor environmental performers, are portfolios of good
environmental performing assets less risky?
(2)Compared to portfolios of poor environmental performers, do portfolios of good
environmental performing assets provide a higher return?
And two variants to the model:
(3) Does relying on overall environmental performance rather than corporate emissions
disclosure for portfolio formation yield a larger difference between good and poor
environmental performing portfolios?
12
7/27/2019 Investing in Canada's Climate
13/25
(4)Does the Oil and Gas sector exhibit a larger difference between good and poor
environmental performing portfolios?
The method used to test these questions empirically will be capital asset pricing model (CAPM) and a
multi-factor model. From this analysis we can learn about returns and volatility of assets grouped in
portfolios based on environmental performance. The model variation implied by question (3) means
that this study will use two different methods for estimating company environmental performance. The
first is the same as one used by Ziegler and is a measure of corporate disclosure. This result is a binary
variable and as such only reveals environmental performance to an extent. A more direct analysis of
environmental performance may yield different results, which is what the answer to question (3) hopes
to ascertain. The variant in question (4) simply means that the analysis will be done on a subset of assets
that comprise the oil and gas sector, to see if the results here are different than across all sectors.
We put forth a null hypothesis, given that Canadas regulatory environment occupies a grey area
between the United States and Europe.
HYPOTHESIS (1N): portfolios of good and poor environmental performers will exhibit no significant
observable difference in risk.
HYPOTHESIS (2N): portfolios of good and poor environmental performers will exhibit no
significant observable difference in risk-adjusted returns.
In answering the 2 variant questions we put forward positive hypotheses that elaborate on the expected
outcomes, given that these variants were selected for their hopefully explanatory outcomes.
HYPOTHESIS (3P): portfolios based on overall environmental performance will yield a larger
difference between good and poor environmental performing portfolios.
HYPOTHESIS (4P): portfolios formed of assets in the subsector of oil and gas will yield a larger
difference between good and poor environmental performing portfolios
13
7/27/2019 Investing in Canada's Climate
14/25
SECTION 3: EMPIRICAL ANALYSIS
In this section the models of financial analysis are explained and the method by which
environmental performance is integrated into these models is explained. Then the empirical findings are
given. Further discussion of these findings can be found in section 4.
THE MODEL: CAPM AND FAMA AND FRENCH
The Capital asset pricing model (CAPM) is used to estimate an asset or group of assets risk
adjusted return and volatility (Fama and French 2004). Using time series data of the assets returns,
the market rate of return, and the risk free rate of return, those parameters can be estimated using an
ordinary least squares (OLS) regression. The equation for CAPM looks like this:
Jensens alpha is the intercept parameter and accounts for abnormal return also known as risk
adjusted return because the model assumes a linear relationship between risk and return. If an asset
has a positive or negative Jensens alpha, then the asset deviates from that assumption, hence the
term abnormal returns. As mentioned risk is estimated by the regression coefficient (beta) and it is
important to note that this captures only systematic risk, also called market risk or undiversifiable risk. It
is defined as such because beta estimates the movement of the price of the asset in relation to market
returns. This means a beta value of 1 moves exactly as the market does, less than 1 and it moves with
the market but to a lesser degree, and it is possible to have a negative beta meaning it moves counter-
cyclically to the market.
The CAPM is widely used and widely criticized as a model. One more robust variation proposed
by Fama and French includes other slope estimating parameters that are supposed to give more
explanatory power to the model. With the addition of the factors for asset capitalization and book-to-
market ratio and the market movements of these factors, additional betas can be estimated. Small cap
14
7/27/2019 Investing in Canada's Climate
15/25
assets tend to have higher returns than big caps, so the variable SMB (small minus big) represents the
returns of a portfolio of small cap stocks minus big cap ones. Similarly, HML (high minus low) is a
variable determined by the difference in return to portfolios of high and low book-to-market ratios. A
fourth term called the momentum factor can be added for even more explanatory power (Carhart
1997). This factor takes into account the movement of the market over the previous months, it is usually
an average of returns from the previous 12 or 6 months. In all now, we have 4 factors and the equation
looks like this:
Using these two formulas, we take a closer look at Canadian markets, but first we need environmental
variables.
DATA SOURCES AND PORTFOLIO FORMATION
The Canadian Fama and French monthly factors as well as monthly market returns and t-bill
rates were generously provided by Claude Francoeur of HEC Montral. This data spanned from 1995 up
until the end of year 2009. I am very grateful for his contributions, without them I would have had to
compute the factors myself. Asset-specific data was all accessed via Datastream, provided by Thomson-
Reuters to the University of British Columbia. Also accessed via Datastream was asset ESG data compiled
by a third party Swiss company called ASSET4. ASSET4 bases its entire ESG database on publicly available
information; all of the environmental parameters that were looked at had been disclosed by the
company, much of it on a voluntary basis.
The first environmental aspect that was looked at is whether the company has an emissions
reduction policy. The variable is simply a yes or no, but we take this variable to be a good indicator
of the companys overall commitment to combatting climate change. The other environmental factor
considered is a score out of 100 that is given by the researchers at ASSET4. They describe this score
15
7/27/2019 Investing in Canada's Climate
16/25
as follows: The environmental pillar measures a company's impact on living and non-living natural
systems, including the air, land and water, as well as complete ecosystems. It reflects how well a
company uses best management practices to avoid environmental risks and capitalize on environmental
opportunities in order to generate long term shareholder value (ASSET4 2013). To include these
environmental factors into our analysis, portfolios were created for a comparison between better and
poorer environmental performing companies. For the emission reduction plan the portfolios were
divided on the yes/no basis, giving two portfolios of different sizes. For the environmental score, the
median was calculated and assets with scores above the median were in the good portfolio and below
the median were the poor portfolio.
TABLE 3.1 EMISSION REDUCTION PLAN PORTFOLIOS
All companies n=89 Oil and Gas n=19
Emission
Reduction
Plan (Y)
No
Plan
(N)
Market Cap
Ratio Y/N
Emission
Reduction
Plan (Y)
No
Plan
(N)
Market
Cap Ratio
Y/N
2005 25 (28%) 64 0.74937754 8 (42%) 11 2.557029
2006 29 60 1.05077807 11 8 4.598691
2007 44 45 2.00314465 14 5 16.14236
2008 56 33 3.54523756 17 2 51.25082
2009 59 (66%) 30 3.76894898 17 (89%) 2 75.20784
TABLE 3.2 SAMPLE ENVIRONMENTAL SCORE SUMMARY
Mean Median
2005 39.03382 24.07
2006 40.26674 32.26
2007 41.94292 29.65
2008 48.86708 44.38
2009 52.58213 59.73
Also note that the environmental parameters are updated on a yearly basis by ASSET4, while all
the other data was monthly. Furthermore, though ASSET4 has data on more than 250 Canadian
companies, the number of companies they have full data for that overlaps with the Fama and French
factors was considerably smaller. We selected the period 2005-2009 to maximize the number of
16
7/27/2019 Investing in Canada's Climate
17/25
observations while keeping the sample size large enough. In the end this left us with 89 companies and
60 monthly return observations for each. To deal with the dynamic nature of environmental parameters
we used a value weighted floating portfolio where each year it is as if the portfolios are liquidated and
then reformed based on updated environmental criteria and using that years asset market
capitalisation divided by portfolio market capitalization to give new asset ratios. An additional 4
portfolios using the same methodology were created using only companies from the oil and gas sector
(yes and no emission reduction plan, and good and poor environmental score). Thus, in total, we have 8
value weighted floating portfolios to conduct analysis on.
REGRESSION RESULTS
Regressions were run using Stata statistical software at the University of British Columbia.
Displayed in tables 3.3 and 3.4 below are the results from CAPM and Fama and French 3 factor models.
Regressions using a 4 factor model including momentum were also done, but the results had less
significance than the three factor model and lower adjusted r-squared values so those results are not
included here.
TABLE 3.3 CAPM RESULTS
Portfolio Jensen's Alpha Beta R-squared
ENVSCORE-TOP 0.009(0.008) 0.352(0.111)** 0.147
ENVSCORE-BOT 0.007(0.007) 0.36(0.109)** 0.157
O&G-ENVSCORE-TOP 0.011(0.011) 0.307(0.156) 0.060
O&G-ENVSCORE-BOT 0.005(0.01) 0.459(0.151)** 0.136
E REDUC-YES 0.01(0.008) 0.352(0.114)** 0.140
E REDUC-NO 0.005(0.008) 0.345(0.117)** 0.139
O&G-E REDUC-YES 0.011(0.01) 0.317(0.153)* 0.069
O&G-E REDUC-NO 0.01(0.117) 0.496(0.173)** 0.125
**means statistically significant to the 1% level
* statistically significant to the 5% level
TABLE 3.4 FAMA FRENCH 3 FACTOR RESULTS
Portfolio
Jensen's
alpha Beta SMB HML
Adjusted R-
squared
ENVSCORE-TOP 0.009(0.007) 0.216(0.152) 0.19(0.261) 0.464(0.162)** 0.216
ENVSCORE-BOT 0.007(0.007) 0.285(0.152) 0.06(0.262) 0.382(0.162)* 0.195
17
7/27/2019 Investing in Canada's Climate
18/25
O&G-ENVSCORE-TOP 0.012(0.01) 0.182(0.222) 0.179(0.382) 0.411(0.237) 0.063
O&G-ENVSCORE-BOT 0.004(0.009) 0.511(0.208)* -0.316(0.357) 0.507(0.222)* 0.200
E REDUC-YES 0.011(0.007) 0.207(0.158) 0.219(0.271) 0.447(0.168)** 0.196
E REDUC-NO 0.005(0.008) 0.286(0.162) 0.012(0.279) 0.402(0.173)* 0.171
O&G-E REDUC-YES 0.011(0.01) 0.209(0.217) 0.131(0.374) 0.425(0.232) 0.075
O&G-E REDUC-NO 0.009(0.011) 0.551(0.249)* -0.235(0.428) 0.257(0.265) 0.101
First, note that in all the regressions there is not a single significant value for alpha, meaning
that none of the portfolios experienced abnormal returns. From this we may choose to infer that the
null hypothesis 2N is confirmed, because there is no significant difference in returns across portfolios.
Next, observe the extremely low r-squared values in both tables (adjusted r-squared for 3-factor
model), this indicates the extent to which the independent variable influence the dependent variable.
In this instance it seems that these models explain only a small portion of the returns on these assets.
The adjusted r-squared values for the 3 factor model are higher than those using CAPM in 6 out of 8
instances, lower in 1, and roughly the same in another. The low explanatory power poses a problem for
drawing conclusions from this analysis, and this will be elaborated on in the next section. For now, we
will allow inferences from the available data to be made on the basis that low r-squared values do not
make them invalid merely less explanatory; in other words they do not tell us the whole story.
Using CAPM the beta estimates are significant to the 1% level in 6 instances, to the 5% level in
1, and only insignificant in 1. The difference between good and poor environmental performers across
all sectors is negligible; however in the oil and gas sector we get an interesting result. The difference
between oil and gas assets with an emissions reduction plan, and without a plan is 0.18, significant to
the 5% level. This result is 0.15 in the oil and gas environmental score portfolios, but not significant.
What this difference implies is that poor environmental performing oil and gas firms move pro-cyclically
(with the market) at a rate of 0.18 higher than good environmental performers in that sector. One way
to interpret this is to say that oil and gas firms without an emissions reduction plan are 44% (0.496/
18
7/27/2019 Investing in Canada's Climate
19/25
0.345) more volatile. This implies greater risk for these assets, adjusted for returns suggesting that
though the null hypothesis 1N holds across all sector, the hypothesis 4P may be inferred as correct
based on the data.
Also from the CAPM regressions, the difference in estimates across environmental parameters
is negligible with the largest being less than 0.04 and most within 0.02 points. The results given from
portfolios created based on emissions reduction plan data compared to portfolios created based on
environmental score data are virtually the same. This suggests that hypothesis 3P should be rejected
and the two environmental assessment methods can be said to be largely interchangeable for our
purposes. This result is possibly explained by the portfolios having similar composition of assets. One
should note that the trend is for more companies to come out with an emissions reduction plan, so in
the future this variable will likely not be a good indicator as nearly all companies will have one.
Turning now to the 3 factor analysis, the first thing to notice is that we have lost most of the
significance for the beta values. Thus, we cannot infer much about riskiness of assets using this model.
Based on the market capitalization ratios of the portfolios given in table 3.1 it is clear that market
capitalization between good and poor environmental performers could play a role in their rate of return.
For example the average market capitalization ratio of oil and gas firms with emissions reduction
plans to those without emission reduction plans is 30:1. Therefore it makes sense that we have more
significant results for the HML factor than for market returns. However no clear pattern emerges from
this data, confusing the analysis slightly.
SECTION 4: DISCUSSION
In addressing the results, the first issue to discuss is the low explanatory value of the
regressions. While there are some reasons to think that a flaw in methodology is to blame, comparison
with a similar empirical study shows that these results are likely not an anomaly. Then after a look at
19
7/27/2019 Investing in Canada's Climate
20/25
how the results fit with various theories, future areas of study are recommended.
METHODOLOGICAL CONCERNS AND A LOOK AT SADORSKY
As noted above, the low r-squared values initially were viewed as possibly caused by
methodological concerns. A usual r-squared value for CAPM lies around 0.7-0.8 and for 3 factor models
that value is higher, more like 0.85-0.9 (Fama and French 1993). What the low r-squareds mean is that
there are other factors that are impacting the dependant variable, in this case asset returns minus the
risk-free rate. CAPM only takes into account market risk, and does not account for any ideosynchratic
or asset-specific risk. It is possible that the assets that were used happened to have high levels of
ideosynchratic risk that is not accounted for in either model. Another possible confounding factor in our
analysis is the financial crisis that hit in 2007 and continued to impact global financial markets into 2008
and 2009. It is possible that the portfolios we created were not representative of the overall Canadian
market, in the sense that some sectors were affected differently than others and if our portfolios do
not share similar constituency as the market, then market returns are not a good indicator for portfolio
returns.
On the topic of portfolio composition, a common criticism of these models is that they
over-estimate returns due to an effect known as survivorship bias. Survivorship bias can lead to an
overestimation of 0.14-0.3% in ethical mutual fund returns (Bauer 2005), and is an issue when looking
at any sector. The process that was used to select assets for use in this study was not arbitrary, but was
based on which companies had complete environmental data in the ASSET4 universe over the specified
time period. Companies with incomplete data were not used, meaning that companies that went out
of business during the period were not available for selection and this should put an upward bias on
return results. Though this does not directly affect r-squared values it is a methodological concern worth
noting.
20
7/27/2019 Investing in Canada's Climate
21/25
Given these concerns, let us do a brief comparison with results from a peer-reviewed journal
article by Perry Sadorsky titled Risk factors in stock returns of Canadian oil and gas companies. In
this study CAPM is used, along with multi-factor models using factors specific to oil and gas in Canada,
namely the price of crude futures and the Canadian/US exchange rate. In this analysis the r-squared
using CAPM was 0.04, and in the multivariable model 0.22. This shows that it is likely that there is more
going on than can be explained by these financial models. There are many critics of these models, but
we will leave the solution to this problem to another paper, perhaps.
SRI, RETURNS AND RISK IN LIGHT OF EMPIRICAL RESULTS
The main focus of this paper is to determine a causal relationship between these key factors in
the Canadian market. Based on the results, there does not seem to be a higher return to investing in top
environmental performers. Based on this, it seems the results support the null hypothesis of Hamilton,
that CSR is not priced and there is no difference in returns when investors try to include CSR in
decision making. At the least, it appears that SRI does not suffer from inherently lower returns. Ethical
consumers, who wish to engage in SRI but are concerned about sacrificing profits, should evaluate their
decision in light of this result. If we wanted to extend the model and use a utility function to represent
returns to ethical investment with two inputs of financial returns and stringency of ethical requirements,
we could do that. What the expected result would be is that consumers with preferences for ethical
investment should engage in it because there seems to be no sacrifice in profit. In other words, this
paper predicts the continual growth of the market for SRI so long as the market segment of ethical
consumers continues to grow. It will be interesting to how the empirical results of this paper hold over
time, as the market for SRI grows. We mentioned the 20% theoretical threshold that has not yet been
crossed, but it will be soon. If the market for SRI becomes over-saturated then it is possible that returns
will suffer and profit-driven investors will be able to capitalize by investing in companies with poor CSR.
21
7/27/2019 Investing in Canada's Climate
22/25
The one exception to the above discussion is the interesting result of higher betas in poor
environmental performers in the oil and gas sector. The literature (this paper included) is guilty of
conflating the terms risk and volatility, to some extent. What the higher beta means is not that those
assets are less likely to achieve returns (risk) but that the price fluctuates more (volatility). Still, volatility
is potentially something that should be avoided. The higher volatility for poor environmental performers
in this sector is probably explained by the overcompliance theory. As companies make investments to
curb environmental destruction, those who dont make those investments are subject to an increasing
likelihood of government intervention. Investors respond to this by making bets on whether they think
regulation ill or will not happen (by making investment decisions on the more polluting firms), and this
causes more movements in asset prices. To find out more about this effect, an event study framework
that analyzes prices and how they respond to new regulation could give more insight.
SECTION 5: CONCLUSION
If socially responsible investing wants to grow to become more than a niche market, there
needs to be proof that selecting assets with positive ESG profiles yields as good, if not better, results
as a no-holds barred approach to finance. The financial world is notoriously profit-driven and if ethical
finance is going to have the opportunity to promote corporate social responsibility, it needs to do so
by making money as well as its not-so-ethical competitors. This paper looked specifically at investing
in Canada based on environmental criteria, to see if portfolios of good environmental performers
performed better than poor environmental performers. The results showed that there was no significant
difference, but this may be due to limitations of the CAPM and multi-factor financial models. One
interesting result of this analysis was that poor environmental performing companies in the oil and gas
sector do exhibit higher volatility to market risk than good environmental performers, by an estimated
22
7/27/2019 Investing in Canada's Climate
23/25
factor of about 44% higher.
If the public decides that it is the governments role to reduce volatility in financial markets,
then making stricter environmental regulations may be a good solution. By increasing the regulatory
standard, there will be fewer companies that are seen by investors as being subject to future regulatory
change. Companies near the bottom will be forced to make the capital investment, but then the playing
field will be more level and speculation will be less. This solution has a double effect of improving
environmental and economic outcomes. It is also important to note that the government can reduce
uncertainty by making its regulatory goals clear and communicating them to the companies, investors
and the public. Also on the topic of information, the collection and dissemination of information related
to ESG factors is still reliant on corporate disclosure and this is likely not an efficient solution.
Government regulations surrounding disclosure could help in this case, as well.
BIBLIOGRAPHY
Bauer, Rob, Jeroen Derwall and Roger Otten. The Ethical Mutaul Fund Performance Debate: NewEvidence from Canada. Journal of Business Ethics 70.1 (2007): 111-124
Bauer, Rob, Kees Koedijk and Roger Otten. International evidence on ethical mutual fund performance
and investment style. Journal of Banking and Finance 29 (2005): 1751-1767.
Buhr, Nola and Martin Freedman. Culture, Institutional Factors and Differences in Environmental
Disclosure Between Canada and the United States. Critical Perspectives on Accounting 12
(2001) 293-322.
Carhart, Mark. On Persistence in Mutual Fund Performance. The Journal of Finance 52.1 (1997): 57-82
Chretien, Stephane and Frank Coggins. Cost of Equity for Energy Utilities: Beyond the CAPM. Draft
February 2008 accessed online.
Cormier, Denis and Michel Magnan. Corporate Environmental Disclosure Strategies: Determinants,
Costs and BEnfits. Journal of Accounting, Auditing and Finance 14 (199): 429-451.
23
7/27/2019 Investing in Canada's Climate
24/25
Cormier, Denis and Michel Magnan. The revisited contribution of environmental reporting to investors
valuation of a firms earnings: An international perspective. Ecological Economics 62 (2007)
613-626
Fama, Eugene , and Kenneth French. Common risk factors in the returns on bonds and stocks. Journalof Financial Economics 33, (1993): 3-53.
Fama, Eugene and Kenneth French. The Capital Asset Pricing Model: Theory and Evidence. The Journal
of Economic Perspectives 18.3 (2004): 25-46.
Hamilton, Sally, Hoje Jo and Meir Statman. Doing Well While Doing Good? The Investment
Performance of Socially Responsible Mutual Funds. Financial Analysts Journal November-
December 1993: 62-66
Heinkel, Robert, Alan Kraus and Josef Zechner. The Effect of Green Investment on Corporate Behavior.
The Journal of Financial and Quantitative Analysis 36.4 (2001): 431-449
IPCC. Fourth Assessment Report: Climate Change 2007. Web accessed on April 26, 2013.
Jones, Charles and David Levy. North American Business Strategies Towards Climate Change.
European Management Journal 25.6 (2007): 428-440.
Kim, Eun-Hee and Thomas Lyon. When Does Institutional Investor Activism Increase Shareholder
Value?: The Carbon Disclosure Project. The B.E. Journal of Economic Analysis and Policy 11.1
(2011): article 50.
Lyon, Thomas and Jay Shimshack. Environmental Disclosure: Evidence from Newsweeks Green
Companies Rankings. Business Society 51:3 (2012)
Mallory, Julie. Is the Financial Market a Mechanism for Environmental Overcompliance? PhD thesis
submitted to Department of Economics at the University of Toronto (2012).
Managi, Shunsuke, Tatsuyoshi Okimoto and Akimi Matsuda. Do Socially Responsible Investment
Indexes Outperform Conventional Indexes? MPRA paper no. 36662 posted 15 February 2012.
PRI Association. About the PRI. Web accessed on April 26, 2013. http://www.unpri.org/about-pri/
about-pri/
Renneboog, Luc, Jenke Ter Horst and Chendi Zhang. Is ethical money financially smart? Nonfinancial
attributes and money flows of socially responsible investment funds. Journal of Financial
Intermediation 20 (2011) 562-588.
24
7/27/2019 Investing in Canada's Climate
25/25
Sadorsky, Perry. Risk factors in stock returns of Canadian oil and gas companies. Energy Economics 23
(2001): 17-28.
Social Investment Organization. Social Investment Trends 2010. Web accessed April 26, 2013.
www.socialinvestment.ca
Sjostrom, Emma. Shareholder Activism for Corporate Social Responsibility: What Do We Know?
Sustainable Development 16 (2008): 141-154.
Thogersen, John and Folke Olander. Human values and the emergence of a sustainable consumption
pattern: A panel study. Journal of Economic Psychology 23 (2002): 605-630.
US SIF. Report on Socially Responsible Investing Trends in the US 2010. Web accessed April 23, 2013.
http://www.ussif.org/files/Publications/10_Trends_Exec_Summary.pdf
Williams, Geoffrey. Some Determinants of the Socially Responsible Investment Decision: A Cross-
Country Study. The Journal of Behavioural Finance 8.1 (2007): 43-57
Ziegler, Andreas, Timo Busch and Volker Hoffman. Disclosed corporate resonses to climate change and
stock performance: an international empirical analysis. Energy Economics 33 (2011): 1283-
1294.