The impact of fossil fuel divestment on portfolio performance

Universidade do Minho Escola de Economia e Gestão Nuno Afonso Cruz Loureiro The impact of fossil fuel divestment on portfolio performance Nuno Afonso Cruz Loureiro The impact of fossil fuel divestment on portfolio performance. Uminho | 2020 ii Nuno Afonso Cruz Loureiro The impact of fossil fuel divestment on portfolio performance Master in Finance This work was realized under the supervision of: Professor Maria do Céu Cortez iii DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Licença concedida aos utilizadores deste trabalho Atribuição-NãoComercial-SemDerivações CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/ iv Acknowledgements This dissertation was the last challenge in this big journey that I had in Universidade do Minho. A special thanks to my supervisor Prof Maria do Céu Cortez for the unconditional support and for the guidance provided during this process. I would also like to thank my supervisor for always making time to address my doubts and questions when I was doing my dissertation. Moreover, I want to thank my parents, my brother and my girlfriend for the motivation and strength that they gave me during this time. Last, a special thanks to the whole Universidade do Minho, especially to EEG (Escola de Economia e Gestão) and the teachers for providing me with their knowledge and support throughout these years. This institution is a big part of my educational life. This is where I completed my Bachelor in Economics and Master in Finance and I could not be more thankful for everyone that I met and everything that I learn in this school. v STATEMENT OF INTEGRITY I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho. vi Abstract The fossil fuel divestment movement and its importance on climate change is a recent topic that has been attracting the interest of society. The purpose of this dissertation is to evaluate and compare the financial performance between a portfolio of fossil fuel stocks and a portfolio without fossil fuel stocks from February 2009 to January 2019. To evaluate the performance of the portfolios we use the four-factor model of Carhart (1997). With the objective to overcoming certain limitations of this model, we further evaluate portfolio performance with the conditional model developed by Christopherson, Ferson and Glassman (1998), which allows for time-varying risk and performance. In this model, we use two public information variables that represent the state of the economy. Our results indicate that the investing in a fossil-free portfolio does not penalize investors relative to the benchmark. However, the fossil fuel portfolio outperforms the green portfolio regardless of the model used. Furthermore, we analyze the exposure of each portfolio to the risk factors and observe that the fossil-free portfolio is exposed to small caps. Keywords: divestment; green energies; fossil fuels; portfolio performance; climate change. vii Resumo O desinvestimento em combustíveis fosseis e sua importância nas mudanças climáticas é um dos temas da atualidade. O objetivo desta dissertação é comparar o desempenho financeiro de uma carteira constituída apenas com ações de combustíveis fósseis e uma carteira sem ações de combustíveis fosseis para o período de Fevereiro de 2009 a Janeiro de 2019. Para avaliar o desempenho das carteiras, usamos o modelo de quatro fatores desenvolvido por Carhart (1997). Com o objetivo de ultrapassar algumas limitações do modelo anterior, avaliamos ainda o desempenho das carteiras com um modelo condicional desenvolvido por Christopherson, Ferson e Glassman (1998) que assume a variação temporal do risco e desempenho. Neste modelo, utilizamos duas variáveis de informação pública que representam o estado da economia. Os nossos resultados indicam que investir numa carteira sem empresas de combustíveis fósseis não penaliza o desempenho relativamente ao mercado, embora a carteira de combustíveis fósseis tenha um desempenho superior ao da carteira verde em todos os modelos que implementamos. Adicionalmente, observamos a exposição de cada carteira aos fatores de risco, observando-se que a carteira sem empresas de combustíveis fosseis está exposta a empresas de pequena capitalização. Palavras-chave: desinvestimento; energias verdes, combustíveis fosseis; desempenho de carteiras; mudanças climáticas viii Index 1. Introduction .............................................................................................................. 1 2. Literature Review ..................................................................................................... 5 2.1. Ethical divestment ................................................................................................. 5 2.2.1 Arguments in favor of fossil divestment ......................................................... 6 2.2.2 Arguments against fossil divestment ............................................................... 7 2.3 Performance of fossil fuel firms vs non-fossil fuel firms ....................................... 7 3. Methodology ............................................................................................................. 9 4. Data ......................................................................................................................... 11 5. Empirical results ..................................................................................................... 15 5.1 Unconditional model ............................................................................................ 15 5.2 Conditional model ................................................................................................ 19 6. Conclusions ............................................................................................................ 25 References ...................................................................................................................... 27 Appendix ........................................................................................................................ 31 ix List of figures Figure 1 Types of divesting institutions 3 6 divest from fossil fuels (Braungardt, Van den Bergh, and Dunlop, 2019). In 2018, Ireland was the first country that committed to divest from fossil fuels. The latest report by 350.org (2019) indicates that in September of 2019, $11 trillions of assets were divested from fossil fuels. More than 1100 institutions, from various types as shown in Figure 1, committed to divesting from fossil fuels. This represents a huge increase from 2014, when the assets committed were of $52 billion from 181 institutions. The great increase of institutions adopting divestment strategies shows how much this movement has grown around the world. And 70% of the institutions embraced with this cause are from outside of the United States. Although most of the institutions that embargo in the divestment are non-profitable, it is also reported that banks, such as Crédit Agricole and insurance companies, like AXA, have divested from fossil fuel investments. 2.2.1 Arguments in favor of fossil divestment The most common argument for people to divest from fossil fuel stocks is the fact that this sector contributes to climate change. Since investing in fossil fuels harms the planet, some people argue that it puts people’s life in danger (Moss, 2016). Therefore, investing in these stocks means an indirect contribution from investors to climate change. Divestment campaigns put pressure on banks and institutions to divest, which will affect the financing of fossil fuel companies. This will affect their exploration capacity and consequently reduce the supply of fossil fuels (Braungardt, Van den Bergh, and Dunlop, 2019). From a financial perspective, fossil fuel divestment can be beneficial as it excludes assets that can be stranded due to high climate and financial risks (Rezec and Scholtens, 2017; Henriques and Sadorsky, 2018; Hunt and Weber, 2019). The divestment on fossil fuel gives an opportunity to investors to re-invest in low-carbon companies. According to Arabella Advisors (2016), from 2010 to 2015 there was an increase of 20% on clean energy investments and a 4% increase from 2014 to 2015. This increase results from a number of institutions deciding to divest from fossil fuel and channeling their investments towards clean energy projects. Divestment can also have an educational impact. Universities can use divestment as a way to teach younger generations how to invest responsibly (Cleveland and Reibstein, 2015). This can be accomplished by urging students not only about the risks of climate change but also the negative externalities that fossil fuels have on the society. 7 Fossil fuel divestment fails on having a strong direct impact on public-traded companies and government entities. However, it has a several indirect impacts on different parts of the society. For instance, the cultural impact will change people’s point of view about divestment. People take action by making divestment campaigns in order to raise the awareness about the fossil fuel consequences. This will have political impact, which will eventually lead to indirect financial impact (Bergman, 2018). 2.2.2 Arguments against fossil divestment In contrast, some researchers argue that fossil fuel divestment does not have a climate impact and it is only being used as a way of creating political impact with no regards for the environment consequences. Moreover, Ritchie and Dowlatabadi (2015) analyzed some aspects of the divestment movement. They reported that divestment does not reduce the exposure to fossil fuel for institutional investors, because many green energies suppliers still rely on fossil fuels. There are fewer investment opportunities for green energies companies in comparison with fossil fuel companies, so there is a possible loss of diversification in stake. In fact, those that are not in favor of divestment argue that this kind of strategy will result in diversification costs. In general, investors will have fewer stocks to invest in, which may result in a more inefficient performance of the portfolio (Le Maux and Le Saout, 2004). It is also important to understand the lack of direct impact that divestment has on the economy. According to Ansar, Caldecott and Tilbury (2013), the direct impact of divestment of fossil fuel companies on the economy is small. They argue that the amount of capital divested is not enough to affect shares prices and companies like ExxonMobil will not notice any reaction in their stocks. 2.3 Performance of fossil fuel firms vs non-fossil fuel firms The increasing engagement to fossil fuel divestment strategies casts doubts on whether or not socially responsible investors would be losing returns in comparison with conventional investors. There is no consensus regarding the performance of portfolios when considering fossil fuels stocks divestment. On the one hand, some studies argue 8 that a fossil fuel-free portfolio would outperform the market. For instance, Halcoussis and Lowenberg (2019) compare the rate of return of a fossil fuel-free portfolio with the S&P 500 index. They also compare a portfolio composed of fossil fuels stocks with the S&P 500. Their results show that the portfolio free of fossil fuels has a slightly higher rate of return than the overall market. Moreover, Trinks, Scholtens, Mulder and Dam (2018) evaluate the performance of US portfolios with and without fossil fuel firms and find that fossil-free portfolios do not underperform the market, which could be explained by the “limited diversification benefits” that fossil fuel company stocks provide. Furthermore, during the first years of divestment campaigns, they noted that fossil fuel portfolios would underperform coalfree portfolios. Henriques and Sadorsky (2018) perform a comparison between three portfolios: one with fossil fuel companies and utilities, another with clean energy companies and another portfolio without fossil fuel companies, utilities and also without clean energy companies. The authors show that replacing fossil fuels and utilities with clean energy will result in a higher risk-adjusted return on the U.S market. Hunt and Weber (2019) also find that higher risk-adjusted performance of a portfolio reflecting divestment strategy. This type of evidence was also found by Hunt and Weber (2019), who analyze the effect of divestment strategies in the Canadian market and find that divestment leads to higher risk-adjusted returns. Specifically focusing on the energy sector, Ng and Zheng (2018) compare the performance of US green and non-green energy portfolios and find that that green energy firms perform at least as well as non-green energy firms. 9 3. Methodology To fulfill the objectives of this dissertation, two types of portfolios will be constructed. The first one is a fossil-free portfolio and the second one is a portfolio with fossil fuel stocks. The goal is to investigate the impact that divesting in fossil fuels will have on the financial performance of a portfolio. To evaluate portfolio performance, we use a multi-factor model that captures recognized sources of systematic risk. The four-factor model developed by Carhart (1997) is given by: 𝑟𝑝,𝑡 = 𝛼𝑝+ 𝑏𝑝1(𝑟𝑚,𝑡) + 𝑏𝑝2(𝑆𝑀𝐵𝑡)+ 𝑏𝑝3(𝐻𝑀𝐿𝑡)+ 𝑏𝑝4(𝑀𝑂𝑀𝑝,𝑡) + 𝜀𝑝,𝑡 (1) Where: 𝑟𝑝,𝑡: the excess return of fund p (over the risk-free rate); 𝑟𝑓,𝑡: risk-free rate; 𝑟𝑚,𝑡: the market excess return; 𝑆𝑀𝐵𝑡 (small minus big): difference in returns between a portfolio of small stocks and a portfolio of large stocks; 𝐻𝑀𝐿𝑡 (high minus low): difference in returns between a portfolio of high book-to-market stocks and a portfolio of a low book-to-market stocks; 𝑀𝑂𝑀𝑝,𝑡(momentum): difference in the returns of a portfolio of a past winners and a portfolio of past losers; 𝑏𝑝1,𝑏𝑝2, 𝑏𝑝3, 𝑏𝑝4: factor coefficients; 𝜀𝑝,𝑡: Error term; 𝛼𝑝: Alpha of the portfolio. Considering the limitations of the previous model, namely the fact that it assumes constant risk exposures over time, Ferson and Schadt (1996) develop a conditional of the model for evaluating performance. This model allows beta to vary over time according to public information variables that represent the state of the economy. In general, the model assumes that a set of predetermined variables (𝑍𝑡−1) will influence risk. Christopherson, 10 Ferson and Glassman (1998) extended this model to allow both time-varying risk and performance, as follows: 𝑟𝑝,𝑡 = 𝛼0𝑝 + 𝛼𝑝 ′𝑍𝑡−1 + 𝛽𝑜𝑝𝑟𝑚,𝑡 + 𝛽𝑝 ′𝑟𝑚,𝑡𝑍𝑡−1 + 𝑆𝑜𝑝𝑆𝑀𝐵𝑡+ 𝑆𝑝 ′𝑆𝑀𝐵𝑡𝑍𝑡−1 + ℎ𝑜𝑝𝐻𝑀𝐿𝑡 + ℎ𝑝 ′𝐻𝑀𝐿𝑡𝑍𝑡−1 + 𝑝𝑜𝑝𝑀𝑂𝑀𝑡+ 𝑝𝑝 ′𝑀𝑂𝑀𝑡𝑍𝑡−1 + 𝜀𝑝,𝑡 (2) Where 𝛼0𝑝 : the fund average conditional performance measure; 𝑍𝑡−1 : the vector of lagged information variables measured as deviations; from their averages (zt−1 = Zt−1 − E(Z)); 𝛽𝑝 ′: The vector that measures the conditional beta in relation with the public information variables; 𝛽𝑜𝑝: Average beta; 𝛼0𝑝 : Average alpha; 𝛼𝑝 ′ : Vector that measures the reaction of alpha to the public information variables; 11 4. Data We start by identifying companies in that are fossil-fuel related and those that are not. To select fossil fuel companies, we identified in Datastream 372 US stocks belonging to the oil and gas sectors. Of those 372, we excluded 167 of them because of limited data on Datastream. Regarding non-fossil fuel firms, we used The Carbon Clean 200 (CC200) list of 2019 to identify clean energy companies. This list reports the 200 largest public companies ranked by green energy revenues. It uses negative proxies in order to exclude oil and gas companies. The companies who are on this list have 100% of the energy that they consume coming from renewable sources. From this list, we selected U.S. companies, resulting in 34 stocks. However, from those 34 we were not able to extract data from 2 of them so we ended up with 32 stocks. From the Datastream database provided by Thomson Reuters, we extracted companies’ monthly total return series from January 2009 to January 2019, and calculated returns in a discrete way. Next, two portfolios were formed, the first one composed of fossil fuel-related companies and the second one is a portfolio formed with companies from various sectors, that use clean energy, which we call the green portfolio. In table 1 we present the companies that form the fossil-free portfolio and the sector that it belongs to. 1 Table 1 – Companies belonging to the fossil-free portfolio Company Sector Alphabet Communication Services Cisco Systems Inc Information Technology HP Inc Information Technology Tesla Inc Consumer Discretionary CSX Corp Industrials Ecolab Inc Materials Ball Corp Materials Air Products and Chemicals Inc Materials 1 The list of the fossil-fuel companies is presented in Appendix 1. 12 Green Plains Inc Energy Acuity Brands Inc Industrials Emerson Electric Co Industrials McCormick & Company Inc Consumer Staples First Solar Inc Information Technology Workday Inc Information Technology Prologis Inc Real Estate Autodesk Inc Information Technology Republic Services Inc Industrials BorgWarner Inc Consumer Discretionary Renewable Energy Group Inc Energy Quanta Services Inc Industrials SunPower Corp Information Technology EMCOR Group Inc Industrials Pacific Ethanol Inc Energy Cree Inc Information Technology Owens Corning Industrials Avangrid Inc Utilities Itron Inc Information Technology Clearway Energy Inc Utilities Andersons Inc Consumer Staples Hubbell Inc Industrials Regal Beloit Corp Industrials Timken Co Industrials We also computed the difference portfolio, corresponding to the difference between the returns of the green portfolio and the fossil fuel portfolio. Table 2 describes the portfolios formed. 13 Table 2 –Summary of the portfolios used As the market benchmark, we used two indexes: a general market index (Standard & Poor’s composite 500) and a sector index (Standard & Poor’s 500 Energy). Data on these indexes were collected from Datastream. The remaining risk factors, SMB, HML and MOM, as well as the risk-free rate, were collected from Kenneth’s French website 2 . Finally, for the conditional models, two public information variables were used: the dividend yield and the short-term rate. These variables were also used by Ferson and Warther (1996) for the US market. The dividend yield is based on the Dow Jones Industrials Average Index and retrieved from Datastream. The short-term rate is the monthly interest rate of the United States and was retrieved from the OECD database. Both variables are lagged 1-month. Because these variables tend to have a higher level of correlation, they were stochastically detrended, as suggested by Ferson, Sarkissian and Simin (2003). The public information variables are also used in terms of their deviation from the mean. 2 http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/index.html Portfolio Portfolio’s description Number of stocks Fossil fuel Portfolios It is an equally weighted portfolio is composed by NYSE and NASDAQ oil and gas stocks retrieved and identified in DataStream. 205 Green Companies Portfolio The stocks were retrieved from DataStream and identified using the Carbon Clean 200 list. This list uses negative screens to exclude oil and gas companies. 32 Difference portfolio This portfolio is formed by performing the subtraction between the green portfolio and the fossil fuel portfolio. 14 Table 3 presents the descriptive statistics of portfolio excess returns and of the variables involved in these regressions. VARIABLES N mean p50 sd min max kurtosis skewness jbera p-value Green 120 0.015 0.018 0.059 -0.168 0.188 3.908 -0.183 4.793 0.091 S&P Energy 120 0.005 0.012 0.057 -0.129 0.170 3.125 -0.180 0.729 0.69 Fossil Fuel 120 0.051 0.031 0.118 -0.171 0.511 5.091 1.120 46.94 6.4 S&P 500 120 0.012 0.015 0.039 -0.107 0.109 3.615 -0.370 4.637 6.4 SMB 120 0.001 0.003 0.024 -0.05 0.061 2.591 0.172 1.428 0.489 HML 120 -0.001 -0.003 0.025 -0.073 0.083 4.249 0.614 15.34 4.7 Mom DY STR 120 120 120 -0.268 0.00 0.00 0.002 -0.012 0.04 0.047 0.281 0.439 -0.344 -0.764 -1.680 0.103 1.483 0.752 24.18 10.43 8.001 -3.361 1.237 -2.133 2470 306.8 216 0 2.4 1.2 By comparing the results of the fossil fuel portfolio and the green portfolio, it is possible to see that the average excess return of the fossil fuel portfolio is 5.1% while the green portfolio has an average excess return of 1.5%. If we turn now to the values of the standard deviation, they show that the fossil fuel portfolio has a higher risk than the green portfolio, 11.8% and 5.9%, respectively. This means that the fossil fuel portfolio exhibits higher returns, but it also has a higher risk. Furthermore, continuing the analysis of the descriptive statistics reported in table 3, the excess returns from both benchmarks are 1.2% and 0.53%, with the S&P 500 S&P Energy indexes, respectively. This is evidence that both the fossil fuel portfolio and green portfolio exceed both market indexes in terms of the mean average excess returns. The fossil fuel portfolio is the one with higher average excess returns among the portfolios and the market indexes. Table 3 also shows the descriptive statistics for the rest of the risk factors of the four-factor model of Carhart (1997) - SMB, HML and MOM - and for the public information variables (short-term interest rate and the dividend yield). Table 3 This table shows the descriptive statistics of monthly excess returns of the equally weighted green and fossil fuel portfolios, the benchmarks used as the market risk factor, S&P Energy and S&P 500, the remaining Carhart four-factor model risk factors, SMB, HML and Mom and the public information variables used in the conditional models, dividend yield (DY) and short term interest rate (STR) both lagged 1-month. 15 5. Empirical results In this section, we will analyze the performance of the fossil fuel portfolio and the green portfolio from February 2009 to January 2019. We start by presenting the results of the unconditional models followed by those of the conditional ones. Two benchmarks are used as the market, so the regressions are performed using a market index and a sector index. 5.1 Unconditional model To evaluate the performance of the portfolios we use the four-factor model developed by Carhart (1997). As a proxy for the market portfolio, we are going to use two benchmarks. The first one is the S&P 500 index. This allow us to have a better understanding of how the performance of the overall market affects the portfolios returns and the level of correlation between the variation of the overall market and the variation of the portfolios. The second benchmark used is the S&P Energy index, which is a more specific market index formed by companies from the energy sector. Firstly, we are going to analyze the unconditional model using the S&P 500 excess return as our market risk factor. Table 4 presents the results. 22 Similar to the results reported in table 4, the performance of the green portfolio is neutral and the fossil fuel portfolio has a positive and statistically significant alpha at a level of 1%. The regression results using the S&P 500 are presented in table 6. As expected, the conditional model showed a slightly greater coefficient of determination in comparison with the non-conditional model using the S&P 500, in table 3. The additional variables improved the explanatory power of the model. The SMB risk factor is positive statistically significant for the green portfolio, at a level of 5%, which indicates exposure to small-cap stocks. The HML factor continues to be statistically insignificant for both portfolios. In contrast to what we observed in table 4, the momentum risk factor has a statistically significant impact on the fossil fuel portfolio instead on having on the green one. Since the coefficient is negative, it means that these portfolios are more exposed to stocks with poor performance in the recent past. There is only one public information variable that has a significant impact on returns - the dividend yield. This variable has a negative impact on the returns of the green portfolio, at a level of significance of 5%. This means that a higher dividend yield causes lower excess returns for the green portfolio. The short-term interest rate presents little evidence of explaining the variation in portfolio performance. According to the results given by the Wald test, for both portfolios we cannot reject the null hypothesis for Wald 1 meaning that there is no evidence of time-varying alphas. The same is observed with Wald 2, suggesting that there is no evidence of timevarying betas. The results of using the conditional model with the S&P Energy index are presented in table 7. Similar to what we observed when we used the S&P 500 Energy in the unconditional model, in table 7 we see that both portfolio alphas are positive and statistically significant at a level of 1%. The 𝑅2 is higher when compared with the nonconditional using S&P 500 Energy but is smaller if we compare with the conditional regression using the S&P 500. 23 VARIABLES Green Fossil Fuel Difference STRt1 -0.01400 -0.01206 -0.00194 (0.160) (0.660) (0.02905) DYt1 0.00442 -0.05924 0.06366 (0.776) (0.169) (0.04545) S&P Energy 0.69456*** 0.97979*** -0.28523 (0.000) (0.000) (0.21020) S&P EnergySTRt1 -0.39479** 0.12447 -0.51926 (0.033) (0.806) (0.53735) S&P EnergyDYt1 0.18224 0.37718 -0.19493 (0.562) (0.664) (0.91899) SMB 0.58429*** 0.27395 0.31035 (0.00) (0.542) (0.47577) SMB*STRt1 -0.79807** 0.47110 -1.26916 (0.031) (0.641) (1.06869) SMB*DYt1 0.38391 -2.84840* 3.23231* (0.527) (0.091) (1.77243) HML -0.30793* -0.39147 0.08354 (0.070) (0.401) (0.49349) HML*STRt1 -0.86815*** 0.32711 -1.19526 (0.006) (0.702) (0.90664) HML*DYt1 -0.69724 0.03273 -0.72997 (0.321) (0.987) (2.05248) Mom -0.01828 -0.70605** 0.68777* (0.882) (0.039) (0.35920) Mom*STRt1 -0.46588* -0.14382 -0.32206 (0.066) (0.836) (0.73544) Mom*DYt1 -0.39876* -0.40492 0.00615 (0.095) (0.537) (0.69421) Alpha 0.00937*** 0.04251*** -0.03313*** (0.007) (0.000) (0.00996) Observations 120 120 120 R-squared W1 W2 W3 0.68567 0.0004 0.3658 0.0008 0.39571 0.7319 0.3334 0.5822 0.16559 0.2285 0.3780 0.2192 P-values in parentheses *** p<0.01, ** p<0.05, * p<0.1 Table 7 Reports the results of the regression for the equally-weighted portfolio with the S&P 500 Energy as our benchmark and using the extended conditional model by Christopherson, Ferson and Glassman (1998) during the period from February 2009 and January 2019. The public information variables used are the short-term interest rate (STR) and dividend yield (DY). Heteroskedasticidty and autocorrelation were tested by using the Breusch Pagan and the Breusch Godfrey test. If both heteroskedasticity and autocorrelation are detected the Newey and West (1987) procedure is applied to correct the model. If only heteroskedasticity is detected, we corrected it by using White (1980) robust models. W1, W2 and W3 correspond to the p-value of the Wald test for the null hypothesis that the coefficients of the conditional alphas, the conditional betas and the conditional alphas and betas, respectively, are jointly equal to zero. Difference is the portfolio that was constructed by subtracting the green portfolio with the fossil fuel portfolio. 24 Once again, the SMB factor is positive and statistically significant at a level of 1% for the green portfolio, meaning that the portfolio is exposed to small-cap stocks. Moreover, HML factor is negative and statistically significant at a level of 10% for the green portfolio. This is the first time that this factor is statistically significant in any of the models we performed. The negative coefficient of -0.30, suggests that the green portfolio tends to be exposed to growth stocks. The momentum factor, just like we have observed before on table 6, only has a statistically significant impact on the fossil fuel portfolio. However, the negative impact is lower than it was on the regression done with the S&P 500 index, with the coefficient decreasing from -1.02 to -0.706. The Wald test provides us similar results for the fossil fuel portfolio as we have seen on table 6. We are not able to reject the null hypotheses that the alphas and betas are jointly equal to zero. However, for the green portfolio we are able to reject the null hypothesis that the coefficients of beta are jointly equal to zero. 25 6. Conclusions In summary, with this dissertation, we aimed to provide a better understanding about fossil fuel divestment and its consequences. This is a hot topic nowadays and has several recent studies addressing the topic. On the one hand, some researchers prefer to focus more on the financial consequences of divestment. On the other hand, others prefer to emphasize the environmental consequences of divesting raising awareness about divestment and its importance in combating climate change. In our study, we decided to focus on the financial consequences of divesting. We have managed to form two portfolios, one consisting of fossil fuel stocks and another of fossil-free companies, in particular, companies that use clean energy. The period time we used for our analysis was from February 2009 to January 2019. Our aim was to compare both portfolios and figure it out which had the best performance. To help us compare both portfolios we created another portfolio – the difference portfolio - that results from the subtraction between the green portfolio and the fossil fuel portfolio. To evaluate portfolio performance, we used the four-factor model developed by Carhart (1997). However, considering the limitations presented by the previous model we also used the conditional model of Christopherson, Ferson and Glassman (1998) model to allow both time-varying, risk and performance. In our conditional models, we used two public information variables that represent the state of economy. These variables were the short-term interest rate and the dividend yield. We contribute to the literature by using robust models of performance measurement, as previously mentioned. Taken together, the results we gathered from both conditional and non-conditional models show that the alpha of the fossil fuel portfolio was always positive and statistically significant at a level of 5%. When using the market index as the benchmark, the green portfolio performs neutral and the fossil-free portfolio exhibits positive performance. When using the sector index, both portfolios outperform the sector-. However, the performance of the fossil fuel portfolio is always higher than that of the green portfolio, regardless of the benchmark used. Furthermore, the results from the market factor document positive and statistically significant coefficients at a level of 5%, for both portfolios. This means that a change in the return of the indexes translates into a positive variation for the portfolios. In general, for both portfolios, the S&P 500 Energy reports the highest coefficients. This may be 26 explained by the fact that both portfolios are composed of energy stocks, so the correlation is bigger. The findings of our study indicate that the green portfolio is exposed to small-cap stocks. This is shown by the fact that the SMB factor has a positive and statistically significant impact for the green portfolio whatever model is used. The fossil fuel portfolio is more exposed to the momentum factor. This factor has a negative impact on the variation of the fossil fuel portfolio. The HML factor does not add much explanatory power to the returns of both portfolios, since the only coefficient that is statistically significant is only significant at a level of 10%. There is no evidence of time-varying alphas and betas in for the fossil fuel portfolio This means that this model does not add much explanatory power to the variation of excess returns for the fossil fuel portfolio. However, the results show that for the green portfolio there is evidence of time-varying betas. This means that the public information variables have an impact on the variation of the green portfolio’s excess return. Ultimately, as we said previously, this dissertation investigates the differences in the financial performance between the fossil fuel portfolio and the green portfolio. After analyzing all the models and the descriptive analysis of the variables, we observe that all the alphas from the difference portfolios are statistically significant at a level of 1%. It is clear to conclude that the fossil fuel portfolio outperforms the green portfolio, and, in the end, divestment has a negative financial impact. Our findings differ from recent studies. For instance, Trinks, Scholtens, Mulder and Dam (2018) found an underperformance of fossil fuel portfolio in comparison with fossil-free portfolios. However, this was only visible for a short period of time, when the fossil fuel prices were down. Another example is Henriques and Sadorsky (2018), who concluded that it is possible to have higher-adjusted returns by divesting in fossil fuel and including clean energy. Our work has some limitations, such as the fact that the stocks that were used to create the fossil fuel portfolio are only from oil and gas companies. This is a clear limitation since it does not include coal stocks, which is an important fossil fuel. Another possible limitation is the difference on the number of stocks that each portfolio has. The fossil fuel has 205 stocks, while the green portfolio has 32. Based on the results there are some recommendations we would like to suggest for future research namely extending this research to other regions, for example Europe, and compare the results. 27 References Ansar, A., Caldecott, B. L., & Tilbury, J. (2013). Stranded assets and the fossil fuel divestment campaign: what does divestment mean for the valuation of fossil fuel assets?. Arabella Advisors (2016). The Global Fossil Fuel Divestment and Clean Energy Investment Movement. Retrivered from https://www.arabellaadvisors.com/wpcontent/uploads/2016/ 12/Global_Divestment_Report_2016.pdf. Braungardt, S., Van den Bergh, J., and Dunlop, T. (2019) Fossil fuel divestment and climate change: Reviewing contested arguments. Energy Research & Social Science, 50, 191-200 Bergman, N. (2018). Impacts of the fossil fuel divestment movement: effects on finance, policy and public discourse. Sustainability, 10(7), 2529. Carhart, M. M. (1997). On persistence in mutual fund performance. The Journal of Finance, 52(1), 57-82. Christopherson, J. A., Ferson, W. E., and Glassman, D. A. (1998). Conditioning manager alphas on economic information: Another look at the persistence of performance. The Review of Financial Studies, 11(1), 111-142. Cleveland, C. J., & Reibstein, R. (2015). The path to fossil fuel divestment for universities: climate responsible investment. Working Paper. Available at SSRN 2565941. Dam, L., and Scholtens, B. (2015). Toward a theory of responsible investing: On the economic foundations of corporate social responsibility. Resource and Energy Economics, 41, 103-121. 28 Derwall, J., Koedijk, K., & Ter Horst, J. (2011). A tale of values-driven and profit-seeking social investors. Journal of Banking & Finance, 35(8), 2137-2147. Ferson, W. E., Sarkissian, S., & Simin, T. T. (2003). Spurious regressions in financial economics?. The Journal of Finance, 58(4), 1393-1413. Ferson, W. E., and Schadt, R. W. (1996). Measuring fund strategy and performance in changing economic conditions. The Journal of Finance, 51(2), 425-461. Ferson, W. E., & Warther, V. A. (1996). Evaluating fund performance in a dynamic market. Financial Analysts Journal, 52(6), 20-28. Halcoussis, D., and Lowenberg, A. D. (2019). The effects of the fossil fuel divestment campaign on stock returns. The North American Journal of Economics and Finance, 47, 669-674. Henriques, I., and Sadorsky, P. (2018). Investor implications of divesting from fossil fuels. Global Finance Journal 38, 30-44. Hiltzik, Michael, “When is it worth it to divest?” Los Angeles Times, January 17, 2016,. C1–C8. Hong, H., & Kacperczyk, M. (2009). The price of sin: The effects of social norms on markets. Journal of Financial Economics, 93(1), 15-36. Hunt, C., & Weber, O. (2019). Fossil fuel divestment strategies: Financial and carbonrelated consequences. Organization & Environment, 32(1), 41-61. Kemp, L. The Fossil Fuel Divestment Game Is Getting Bigger, Thanks to the Smaller Players. Available online: https://theconversation.com/the-fossil-fuel-divestmentgame-is-getting-bigger-thanks-to-the-smaller-players-65109 (accessed on 19 December 2017). 29 Le Maux, J., and Le Saout, E. (2004). The performance of sustainability indexes. Finance India, 18, 737. Linnenluecke, M. K., Han, J., Pan, Z., & Smith, T. (2019). How markets will drive the transition to a low carbon economy. Economic Modelling, 77, 42-54. McKibben, B. (2012, July 19). Global warming’s terrifying new math. Rolling Stone. Retrieved from https://www.rollingstone.com/politics/politics-news/global-warmingsterrifying-new-math-188550/. Moss, J. The Morality of Divestment (2017). Law Policy, 39(4), 412–428. Newey, W. K., & West, K. D. (1987). Hypothesis testing with efficient method of moments estimation. International Economic Review, 777-787. Ng, A., & Zheng, D. (2018). Let's agree to disagree! On payoffs and green tastes in green energy investments. Energy Economics, 69, 155-169 Schueth, S. (2003) Socially Responsible Investing in the United States. Journal of Business Ethics 43(3), 189-194. US SIF. (2018). Report on US Sustainable, Responsible and Impact Investing Trends. Wallace Global Fund. Retrieved from https://www.ussif.org/files/Trends/Trends%202018%20executive%20summary%20FIN AL.pdf Rezec, M., & Scholtens, B. (2017). Financing energy transformation: The role of renewable energy equity indices. International Journal of Green Energy, 14(4), 368378. 30 Ritchie, J., & Dowlatabadi, H. (2015). Divest from the carbon bubble? Reviewing the implications and limitations of fossil fuel divestment for institutional investors. Review of Economics & Finance, 5(2), 59-80. Ripple, W. J., Wolf, C., Newsome, T. M., Barnard, P., & Moomaw, W. R. (2020). World Scientists’ Warning of a Climate Emergency. BioScience, 70(1), 8–12. The Economist. (2015, June 25). No smoking. The Economist. Retrieved from https://www.economist.com/leaders/2015/06/25/no-smoking Trinks, A., Scholtens, B., Mulder, M., and Dam, L. (2018). Fossil fuel divestment and portfolio performance. Ecological Economics, 146, 740-748. White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica: journal of the Econometric Society, 817838. 350.org (2019). $11 Trillion and counting. Retrieved from https://www.ussif.org/files/Trends/Trends%202018%20executive%20summary%20FIN AL.pdf 31 Appendix Appendix 1List of fossil fuel companies Domination Resources Black Warrior Trust Cross Border Resources Gase Energy Infinity Energy Resources Sentry Petroleum Evolution Petroleum Chelsea Oil and Gas Concho Resources HyperDynamics Callon Petroleum Company Mogul Energy International Tianci International Royale Energy Alamo Energy Diamondback Energy Deep Well Oil & Gas Avoca West Texas Resources Aztec Oil & Gas Comstock Resources Freestone Resources Kodiak Energy Approach Resources Western Midstream Partners County Line Energy SRC Energy New Source Energy Partners Axis Energy Daleco Resources Parsley Energy Houston American Energy Panhandle Oil & Gas Montage Resources Victory Oilfield Tech EOS Petroleum Viper Energy Partners Erin Energy BP Prudhoe Bay Royalty Trust Empower Clinics Cygnus Oil & Gas Cabot Oil & Gas Pioneer Natural Resources Mexco Energy Torchlight Energy Resources ConocoPhillips Cimarex Energy Blue Dolphin Energy EQT Treasure Island Royalty Trust Range Resources CHEVRON Delta Oil & Gas Lilis Energy EXXON MOBIL Daybreak Oil & Gas White Label Liquid Occidental Petroleum Devon Energy Marathon Oil HESS New Frontier Energy Abraxas Petroleum Murphy Oil Chancellor Group Solar Integrated Roofing Noble Energy Spindletop Oil & Gas Minerco North European Oil Reserve Petroleum Company Gulport Energy Tellurian Contango Oil & Gas FieldPoint Petroleum Antero Resources Sky Petroleum Striker Oil & Gas California Resources Whiting Petroleum Laredo Oil Barnwell Industries Britannia Mining Octagon 88 Resources Apache Companies