The performance of social responsible investment in U.S.
stock market under different market states
Keyword: SRI, Portfolio, Crisis, US, Screening
Abstract:
Social responsible investment as a growing segment of market has attracted more and more attention from both scholars in academia field and investors in practice. This research will investigate the performance of social responsible investment based on data from stock market in U.S. In addition, this thesis will compare the SRI performance during the financial crisis with non-crisis period to evaluate the relationship between SRI performance and financial crisis.
Student name: Chenxi Huang Student number: 2842823 Study program: MSc Finance Supervisor: Prof
November 8, 2017
1. Introduction
Individual awareness of environmental, social and ethical issue is strongly influencing buying behavior. This development motivates private and institutional investment decisions toward social responsible investment (SRI) and ethical or sustainable investing (Renneborg, Terhorst, & Zhang, 2008). Social responsible investment as a growing segment of market has attracted more and more attention from both scholars in academia field and investors in practice. Social responsible investors add social criteria when evaluating investment strategy to ensure their personal value and belief are consistent with their invested securities (Sauer, 1997). As the proportion of assets that invested in social responsible investment keeps growing, the methods that investors use to screens social performance into their investment process need to be studied. With the fast development of SRI, investors also concern the financial performance of SRI. This study tests several portfolios that constructed based on social-responsible rating in different time periods to evaluate their financial performance. Using stocks traded on NYSE between 2004 and 2015, I build high-rated and low-rated portfolios based on their ASSET4 ESG rating1. A simple investment strategy of buying high-rated stocks and selling low-rated stocks provides evidence on the performance of portfolios constructed by social-responsible screening. Recent studies on SRI funds (e.g., Areal et al., 2013; Nofsinger & Varma, 2014) suggest that the types of screens has impaction on fund performance differently across crisis and non-crisis periods. To further test the impaction of different time periods, I distinguish non-crisis periods with pre-crisis and post-crisis periods and test the performance of these portfolios separately.
The main finding of this thesis is that SRI provides insignificant abnormal normal if crisis period and non-crisis periods are not distinguished. However, investors can obtain significant positive abnormal return during crisis period based on the trading strategy described above. More importantly, SRI significantly underperform conventional investment during post-crisis periods. The opposite performance of SRI during crisis periods and non-crisis periods drives the financial performance of SRI not significantly different with conventional investment for the entire sample period.
My contribution of this research is to provide more evidence on the performance of SRI on US stock market based on different social-responsible ratings and different time periods than previous researches. Furthermore, this research provides investors and researchers evidence on SRI performance in different time periods. Formal researches that study the performance of SRI under different market states focused on the performance during crisis periods and non-crisis periods, I continue their study and separate non-crisis periods into pre-crisis and post-crisis periods. By doing this, I highlight the performance of SRI in different market states and find stronger evidence in this field.
Next chapter provides literature review of this research. The data and methodology is introduced in chapter 3. Chapter 4 shows descriptive data of sample. Chapter 5 and 6 present estimation result as well as robustness test result. The conclusion of this thesis is in chapter 7.
2. Literature review
2.1 Theoretically Foundation
with better management skills and governance method, eventually improving portfolio performance (Bollen, 2007).
Traditionalists see a negative relation between corporate social profit and corporate financial profit because the rising cost of CSR activities declines corporate profit (Oh and Park, 2015). In contrast, supporters of SRI believe as more and more investors conduct SRI, they can place pressure on the companies which are not responsive to social concerns. Apart from social influence, many investors believe social-responsible firms are stronger than their competitors in financial performance (Sauer, 1997). To illustrate this point, they suggest environmental-responsible firms are less likely to be fined against environmental issues. Similarly, social-responsible firms are more exceptional for law suits against product liability and social issues.
2.2 Empirical Evidence
Empirical evidence regarding the performance of SRI is mixed and inconclusive. Older studies (Luther, Matatko, and Corner 1992) apply a simple market model in their analyses, which uses excess returns of an adequate market portfolio as the only explaining factor for excess returns of SRI funds. The study by Bauer, Koedijk, and Otten (2005) uses multi-factor models in combination with a matching approach to test the performance of SRI funds in the UK, Germany, and the USA. Recent researches often apply multi-factor models than a simple market model. Jo and Statman (1993) examine the performance of SRI funds in a very early stage. They find that the average SRI fund underperforms a market portfolio, but does not significantly underperform conventional mutual funds from 1981 to 1990. Statman (2000) confirms his result with data from 1990 through 1998. Statman (2006) finds that the most important SRI indexes for US stocks do not exhibit a statistically significant outperformance. Nevertheless the indexes exhibit clear differences to the S&P 500 index: the average CSR score is higher and the tracking error seems to be relatively strong. Mueller (1991) examines 10 social responsible mutual funds and finds that the risk-adjust return of these 10 mutual funds is 1.03% on average less than comparable, unrestricted investment. Guerard (1997) finds that socially screened portfolios do not have significant different performance with unscreened portfolios.
positive screening could increase portfolio returns. Jagadeesh and Titman (1993) study the strategy of buying stocks that have good performance in the past and selling stocks that have poor performance in the past. They find this strategy generates significantly positive returns over 3 to 12 months holding periods. That means if CSP (corporate social performance) and CFP (corporate financial performance) are positively or negatively related, then SRI strategies that buy good CSP (or sell poor CSP) stocks should generate significantly positive (or negative) return as well. Derwall, Bauer, Guenster and Koedijk (2004) use Innovest’s corporate eco-efficiency scores to compare returns of SRI portfolios constructed by them. They find that the high-ranked portfolio outperforms the low ranked portfolio over the period from 1995 to 2003. To solve their main research question of whether potfolio performance can be increased by applying social responsible investment screening strategies into investment process, Kempf and Osthoff (2007) use a simple long-short strategy, based on KLD rating, to compare different screening criteria’s affection on financial performance. KLD rating evaluate the social responsible performance of companies based on multiple criteria since 1991. Their result suggest that investors can earn remarkable high abnormal returns by following their simple long-short strategy with positive or best-in-class screening approach. Similarly, Borgers et al. (2015) test SRI portfolios on the basis of KLD data as well and find that SRI portfolios have higher stock performance from 1992 to 2004, but the abnormal return become insignificant from 2004 to 2009. Guenster (2012) also reports the phenomenon of diminishing outperformance of SRI in recent years and concludes this is because the positive alphas in SRI favored firms has been disappearing recently. This explains why the abnormal return for the portfolios of Borgers et al. (2015) become insignificant from 2004 to 2009. Diltz (1995) finds that employing environmental and military screens leads to a significantly positive abnormal return, while all other screens do not have a significant impact.
fund is significantly 5% increased by events that negatively influenced market while negative effect on conventional funds was estimated. Nofsinger and Varma (2014) compare the performance of US SRI funds and conventional mutual funds during periods of crisis in the internet bubble and 2008 global financial crisis, in addition to non-crisis periods during 2000–2011. Their estimation results show that SRI funds significantly outperform conventional funds during crisis periods. However, conventional funds significantly outperformed SRI fund during non-crisis periods. More recently, Leite and Cortez (2015) compare the performance of SRI funds and conventional funds during different market states in France: the period until the technology bubble burst (January 2001–March 2003), the global financial crisis (June 2007–February 2009), and the euro sovereign debt crisis (May 2011–May 2012).Their result shows that SRI funds underperform conventional funds during non-crisis period, but outperform conventional funds during crisis periods, which is consistent with NOfsinger and Varma (2014).
2.3 Social Responsible Rating
2.4 Hypothesis Development
This study focuses on two research questions. The first one is: Can SRI in stock market lead to a positive abnormal return? This question has been studied by many researchers and their results are mixed. In US market, some formal researches find significant positive abnormal return for SRI funds and portfolios (Kempf and Osthoff, 2007). Supporters of SRI claim that the use of social screens could identify companies with better management skills and governance method, leading to a better portfolio performance. Besides with a better corporate governance, social responsible investor believe social-responsible firms are more exceptional for law suits against product liability and environmental issues that would influence their financial performance. For this research, I use ASSET4 data as a reference for screening and assume the performance of SRI would be similar as formal researches. To be able to answer this research question, I construct the first hypothesis:
Hypothesis 1: Social responsible Investment in US stock market, compare to conventional investment, provide positive abnormal return.
The second research question is: What is the performance of SRI under different market states? It should be noted that, better performance in normal period doesn’t mean the performance should be better as well during crisis period.
People are paying more attention to corporate behavior during poor economic states (Hirshleifer, 2008) because they care more about losses than gains based on prospect theory. A common belief described by Nofsinger and Varma (2014) is people believe when companies have good governance practice, they may be exposed to lower risk and thereby achieve better financial performance. A good social-responsible rating is considered as an expression of good governance practice and thus lead to better financial performance especially in poor economic states. Moreover, previous studies argued that social responsible screening helps avoid firms that are more likely to cause high-impact negative news regarding social issues. Thus, investors could use SRI portfolio to secure their return during crisis periods. This lead to the second hypothesis:
3. Data and Methodology
There are formal studies that use KLD rating to screen stocks in US, thus, simply replicate formal studies with KLD rating will obtain similar result. To avoid this problem, I choose to use ASSET4 ESG Data to build SRI portfolios. ASSET4 is a Thomson Reuters database that provides objective, relevant and systematic environmental, social and governance (ESG) information based on 250+ key performance indicators (KPIs) and 750+ individual data points along with their original data sources (Thomson Reuters, 2016). ASSET4 database aggregates the scores of almost all qualitative ESG criteria: resource reduction, emission reduction, product innovation, employment quality, health& safety, training& development, diversity, human rights, community, product responsibility, board structure, compensation policy, board functions, shareholder rights, vision and strategy (Thomson Reuters, 2016).
Financial performance is acquired from DataStream Dataset. All the companies which are trading on NYSE will be selected into the sample if there is an associated ASSET4 score. My sample covers time period from 2004 to 2015 and the sample firms are divided into three sub-samples. The first subsample covers time period from 2004 to 2007 and is labeled as “before crisis”. The second subsample consists of companies from 2008 to 2011, this subsample is used to measure the SRI performance during the financial crisis. The crisis period of my research includes the global financial crisis of 2008-2009 and European debt crisis of 2009-2011 so I combine these two crises into one crisis period. The third subsample covers stocks between 2012 and 2015 and is labeled as “after crisis”. The three periods are set to an equal length of 4 years to eliminate potential time bias.
3.1 Portfolio formation
:selected into portfolio. The portfolios are only adjusted within a year when a company vanishes from the ASSET4 database. The bottom 10% of stocks that ranks in ASSET4 rating are selected into the second type of portfolios which is called low-rated portfolio. Based on these two types of portfolios, another portfolio which is called high-minus-low portfolio is constructed by going long in high-rated portfolio and short in the corresponding low-rated portfolio.
3.2 Measurement of performance
To measure the performance of the SRI portfolios, I employ Carhart (1997) model. This model comprises market return, size, value, and momentum factors. Kempf and Osthoff (2007) use a similar model to estimate the difference between their high-rated and low-rated portfolio with respect to the loading of these four factors. I estimate the following regression:
𝑅𝑖,𝑡− 𝑅𝑓,𝑡 = 𝛼𝑖 + 𝛽1𝑖(𝑅𝑚,𝑡 − 𝑅𝑓,𝑡) + 𝛽2𝑖𝑆𝑀𝐵𝑡+ 𝛽3𝑖𝐻𝑀𝐿𝑡+ 𝛽4𝑖𝑀𝑂𝑀𝑡+ 𝜀𝑖,𝑡 The dependent variable is the monthly return of portfolio i in month t in excess of the risk-free rate. The monthly excess return will be explained by the four-factors. A significantly positive alpha would represent a positive abnormal return generated by SRI portfolio. Alpha would be significantly negative if SRI portfolio underperform market portfolio. The risk factors and the risk free interest rates for the US stock markets are downloaded from the website of Kenneth French Data website (http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html).
These risk factors are necessary to estimate risk-adjusted returns that are more reliable than estimates from a restrictive one-factor model based on the CAPM. For robustness test, I also employ CAPM model which controls the factor of market return. The following regression is tested:
𝑅𝑖,𝑡− 𝑅𝑓,𝑡 = 𝛼𝑖 + 𝛽𝑖(𝑅𝑚,𝑡− 𝑅𝑓,𝑡) + 𝜀𝑖,𝑡
3.3 Sample selection
My research use monthly data to test portfolio performance. Monthly return of stocks are used and 144 months are observed in the time period from January 2004 to December 2015. The sample consist of all the companies listed in NYSE which has corresponding ASSET4 ESG rating during 2004-2015. There are 1136 companies that are active in NYSE in 2015, however, all the companies that related to tobacco, alcohol, defense and gambling are eliminated because they are considered as ‘sin stocks’. The nature of these companies ensure that they have a steady stream of consumers and have a relatively inelastic demand. Investing into ‘sin stocks’ clearly disobey the rule of socially responsible investment because their business harm society in some degree. More importantly, researchers suggest that ‘sin stocks’ generate high risk-adjusted returns in the form of positive and significant Jensen’s alpha (Durand, Koh, & Limkriangkrai, 2013; Sauer, Schneider, & Sheikh, 2013; Hong & Kacperczyk, 2009; Fabozzi, Ma, & Oliphant,2008; Salaber, 2007). This attribute of sin stocks would strongly influence my result of this study so I eliminated all sin stocks in my sample. Guenster (2012) describes that the main reason for normal performance of SRI portfolios is because while SRI portfolios include some positive alpha firms, they also exclude some positive alpha firms like sin stocks. To test purely the performance of SRI portfolios, I eliminate sin stocks from my investment universe.
Composite Index is also collected from DataStream database.
4. Descriptive data:
My final sample could be divided into different industries types based on their ICB2 industry code. Table 1 reports the number of sample firms in different industries and their average market value in 2016.Their market value is changing during the sample period and their average market value in 2016 only reflects their current market size. The highest number of firms is from the financial sector. In contrast, there are only 4 companies from telecommunications sector that have been selected into sample. However, the average market value of telecommunications sector in 2016 is the highest among other industries. That is because three major companies which are Verizon Communications, Level3 Communications and AT&T, are the top 3 Ethernet providers in U.S that increase the average market value of companies in telecommunication industry significantly.
Table 1: Average market value in different industries.
Industry Number of firms Average Market Value
2016(in million dollars)
Basic materials 24 9,748.63 Consumer Goods 43 30,263.07 Consumer services 37 29,460.54 Financials 89 24,139.93 Health Care 23 57,650.89 Industrials 46 23,265.40 Oil&gas 31 28,779.73 Technology 9 38,438.65 Telecommunications 4 105,436.14 Utilities 21 18825.79 Overall 327 28846.96
This table presents the average market value of the sample firms in different industries in 2016.
The tested portfolios are divided into three groups. The first group consists of firms that have ‘high’ rating and the second group consist of firms that have ‘low’ rating. High-rated means the companies rank top 10% among all firms, and low-High-rated means the companies rank in bottom 10%. Another cut-off of 5% is presented in robustness test.
The third group consist of high-minus-low portfolios which is buying high-rated firms and selling corresponding low-rated firms.
Table 2: Average market value of companies that in portfolios overtime (in billion dollars)
Year ESG
Aggregate
Environmental Social Governance
High-rated Low- rated High-rated Low- rated High-rated Low- rated High-rated Low- rated 2004 57.84 4.48 46.66 8.27 58.89 6.52 29.71 13.34 2005 65.42 5.22 54.58 8.95 62.70 6.96 67.54 6.52 2006 61.24 5.55 67.83 8.63 61.10 7.32 59.12 11.38 2007 68.94 9.35 77.83 8.63 69.37 9.32 60.57 13.29 2008 51.03 4.54 69.66 9.04 58.00 8.18 48.08 9.19 2009 47.43 4.87 55.29 10.12 38.76 5.22 38.89 5.92 2010 40.42 5.98 52.43 5.88 41.00 6.31 31.00 5.14 2011 31.61 7.38 61.74 7.05 35.61 7.77 36.71 9.69 2012 42.37 9.23 54.96 8.47 34.90 8.68 36.37 9.84 2013 38.14 9.23 52.91 8.84 43.86 9.86 49.37 10.71 2014 42.25 6.03 65.79 10.49 50.10 11.06 50.45 7.20 2015 62.83 12.53 72.28 12.18 55.80 13.01 52.08 17.56
This table presents the average market value of the sample firms that are constructed into different portfolios over the entire sample period.
Table 2 reports the average market values from 2004 to 2015 for the companies in the eight high-rated and low-rated portfolios. The average market values for high-rated portfolios are increasing from 2004 to 2007, then heavily dropped down from 2008 to 2011, and slightly increased from 2012 to 2015. This time trend is highly consistent with financial crisis period, showing a preference of large firms are more sustainable before and after financial crisis, while medium-sized firms are more sustainable than large or small cap firms during financial crisis.
doing well in CSR) have higher average market value than others, indicates that firm size has significant positive influence on the practice CSR. They conclude that small company cares less about CSR than large company.
Table 3 descriptive data of the monthly returns of eight portfolios.
Mean Median Max Min Standard
Deviation ESG High-rated 0.43% 1.00% 11.31% -15.45% 4.05% ESG Low-rated 0.42% 1.20% 15.65% -28.45% 5.56% Environmental High-rated 0.51% 1.17% 15.80% -15.25% 4.21% Environmental Low-rated 0.57% 1.11% 14.30% -20.85% 4.88% Governance High-rated 0.59% 0.97% 11.57% -12.29% 3.38% Governance Low-rated 0.54% 1.09% 16.58% -24.19% 5.33% Social High-rated 0.50% 1.24% 12.34% -15.1% 4.08% Social Low-rated 0.73% 1.35% 15.64% -21.27% 4.96%
NYSE Composite Index 0.37% 0.65% 16.61% -19.51% 5.16%
The descriptive data of table 3 shows that the monthly returns of the portfolios studied are similar in terms of their mean, median, and max monthly return. The low-rated portfolios are more volatile than high-rated portfolios during entire period, and that is mainly because their lowest monthly return are evidently smaller than high-rated portfolios. The lowest returns of each portfolio are all acquired during financial crisis period, indicates that crises have negative influence over each portfolio regardless of screening method.
5. Estimation result:
Table 4 presents the performance and risk estimates for the value-weighted portfolios of different screening methods over the entire sample period. It provides results for the high-rated portfolios and the low-rated portfolios from 2004 to 2015. To further enhance the performance comparability I also estimate the performance of the high-rated minus high-rated portfolios (long in the high-high-rated portfolio and short in the low-rated portfolio).
Table 4 shows that not only the market risk has a significant impact on the excess returns of the portfolios, but also the size, the book-to-market, and the momentum factor. The respective coefficients are significant in most cases. Therefore, all these effects should be controlled when comparing the high-rated and the low-rated portfolios. For robustness test, Jensen’s alpha which use market return as the only factor, is also tested. The result is similar and is presented in chapter 6.
Carhart’s Alpha estimates the abnormal return of tested portfolios. For the portfolios constructed on ESG aggregated score, the alphas for high-rated and low-rated portfolios are both not significantly different from zero. Furthermore, the alpha of conventional investment which is investing in NYSE Composite Index does not significantly different from zero as well. I cannot conclude any difference of SRI and conventional investment in terms of abnormal return based on the performance of ESG aggregated screening portfolios during the entire sample period.
market index because only this alpha is significantly above zero, and the evidence is weak. This result is consistent with table 3, which shows that low-rated stocks based on social screening have the highest average monthly return and NYSE Composite Index has the lowest average monthly return during the entire period, showing that low-rated stocks based on social screening achieve the best performance among all other portfolios.
Table 4 Value-weighted portfolio performance during 2004-2015
ESG Aggregated Alpha Market SMB HML MOM 𝑹𝟐
High-rated 1.53 0.90∗∗∗ -0.32*** 0.16** 0.09** 0.89
Low-rated 0.11 0.91∗∗∗ 0.07*** 0.44*** -0.16*** 0.86
High-rated minus low-rated 0.43 -0.01 -0.27 -0.25 0.23 0.26
Environmental Alpha Market SMB HML MOM 𝑹𝟐
High-rated 0.47 0.822*** -0.19** 0.162** 0.06** 0.92
Low-rated 2.46 0.73*** 0.15* 0.32*** -0.08* 0.83
High-rated minus low-rated -2.01 0.068* -0.32*** -0.151* 0.14 0.23
Social Alpha Market SMB HML MOM 𝑹𝟐
High-rated 0.9 0.809*** -0.19*** -0.01 0.06** 0.91
Low-rated 3.53* 0.752*** 0.147 0.289*** -0.88* 0.81
High-rated minus low-rated -2.7 0.057 -0.031 -0.27 0.13 0.25
Governance Alpha Market SMB HML MOM 𝑹𝟐
High-rated 1.55 0.730*** -0.18*** 0.043 0.15*** 0.86
Low-rated 1.25 0.92*** 0.157** 0.122* -0.01 0.89
High-rated minus low-rated -0.35 -0.17*** -0.34*** -0.08 0.15** 0.41
Conventional Alpha Market SMB HML MOM 𝑹𝟐
NYSE Composite Index -2.91 1.099*** -0.139 0.161* -0.05 0.82
Table 5: Portfolios performance in divided time periods.
This table presents statistics on the performance of SRI portfolios during different time periods. This table summarizes for each screen the annualized abnormal return, factor loadings, and the adjusted R2 of a portfolio strategy using the Carhart four-factor model. SRI stocks refer to stocks trade in NYSE with social responsibility screens. The high-rated portfolios consist of the top 10% of all companies with the highest rating. The low-rated portfolios consist of the bottom 10% of all companies with the lowest rating. Time period from 2004 to 2007 is defined as before-crisis, data from 2008 to 2011 is defined as during-crisis and data from 2012 to 2015 is defined as post-crisis. Significance at 10%, 5% and 1% is represented by *, **, *** respectively.
PORTFOLIOS PERFORMANCE IN DIFFERENT TIME PERIOD
Alpha Market SMB HML MOM R2
Before -1.6 0.928*** -0.299*** -0.05 -0.09 0.78 ESG HIGH-RATED Crisis 3.2* 0.79*** -0.06 0.068 0.165*** 0.94
After 0.4 0.812*** -0.207 0.16 0.325*** 0.80 Before -0.53 0.91*** 0.26*** -0.12 -0.09 0.71 ESG LOW-RATED Crisis -1.2 0.79*** 0.182 0.445 -0.163*** 0.84
After 2.3 0.911*** 0.04 0.147 0.129 0.87 Before -0.4 0.90*** 0.22* -0.04 -0.26** 0.72 ENV HIGH-RATED Crisis 3.32* 0.8*** -0.183** 0.249** 0.10*** 0.96 After 0.5 0.87*** -0.223* 0.20 0.273*** 0.86 Before -2.16 0.912*** -0.33*** -0.09 -0.09 0.79 ENV LOW-RATED Crisis 0.16 0.71*** 0.184 0.462*** -0.05 0.86 After 6.4** 0.93*** -0.06 0.14 0.075 0.86 Before -1.1 0.93*** -0.308*** -0.09 -0.093 0.81 SOC HIGH-RATED Crisis 2.42 0.79*** -0.14* 0.01 0.082 0.95
After -0.7 0.912*** -0.23** 0.068 0.211** 0.87 Before 2.4 0.84*** 0.25** -0.169 -0.195** 0.78 SOC LOW-RATED Crisis 2.11 0.72*** 0.152* 0.46*** -0.06 0.85
After 6.3** 0.88*** -0.15 0.05 0.08 0.82 Before -2.1 0.903*** -0.295*** -0.09 -0.07 0.73 GOV HIGH-RATED Crisis 4.8** 0.813*** -0.13 0.10* 0.193*** 0.94
After 0.17 0.90*** -0.31** 0.06 0.328*** 0.82 Before 1.0 0.86*** 0.05 -0.26*** -0.11 0.78 GOV LLOW-RATED Crisis 0.4 0.935*** 0.32** 0.16 0.02 0.93
Table 5 presents the estimation result computed from Carhart 4-factor models in different time periods. The alphas are annualized for better presentation. I divide the full time period from 2004 to 2015 into three sub-periods based on the event of financial crisis. From 2008 to the end of 2011, there are two main financial crises happened which is the global financial crisis and the Euro Debt Crisis. I combine these two crises into a financial crisis period and set the time periods before 2008 as pre-crisis and time periods after 2011 as post-crisis to estimate the same regression again. Each period contains the same length of 4 years.
The four top-rated portfolios generate negative but not statistically significant abnormal return during 2004 to 2007. Low-rated portfolios based on social and environmental screening generate positive but not significant abnormal return in the same period. However, the evidence is not strong enough to draw conclusion on the performance of SRI during the period of pre-financial crisis.
The high-rated portfolios perform much better than low-rated portfolios during crisis periods. The high-rated portfolio based on ESG aggregate score generates a significantly 3.2% abnormal return during crisis period, outperforms the low-rated one which generates -1.2% abnormal return. Separating environmental, social and governance screening categories, I find that all three high-rated portfolios generate positive abnormal returns during crisis period. The portfolio focusing on governance issue shows a particular strong crisis alpha of 4.8%. Except the portfolio focusing on social issue, all the high-rated portfolios show significantly positive abnormal returns during crisis period. Looking into low-rated portfolios, none of them generate significantly positive abnormal returns during crisis period.
The low-rated portfolios are the winners in the period of post financial crises. All the low-rated portfolios generate positive abnormal returns during 2012-2015. The separating portfolios focusing on environmental, social and governance issues show 6.4%, 6.3% and 4.8% significantly positive abnormal returns. The low-rated portfolio based on ESG aggregate score also generate 2.3% positive abnormal return, however not significantly. Looking back to high-rated portfolios, none of them generate significantly positive abnormal returns after crises.
for the entire period. It because that the abnormal returns in the three periods are adverse and they are equaling out the abnormal returns for the entire period. This result is consistent with Nofsinger and Varma (2013), which they show SRI portfolios are performing better in crisis period than non-crisis period. However, they did not distinguish non-crisis period into pre-crisis and post-crisis periods. My result shows that high-rated portfolios are performing better than low-rated portfolios during crisis periods, and low-rated portfolios are performing better than high-rated portfolios after crisis periods. The evidence on the performance of SRI portfolios during pre-crisis period is very weak.
Table 6 High-minus-low portfolio performance in divided time periods.
This table presents statistics on SRI portfolio performance during different time periods. This table summarises for each screen the annualised abnormal return, factor loadings, and the adjusted R2 of a portfolio strategy using the Carhart four-factor model. SRI stocks refer to stocks trade in NYSE with social responsibility screens. The high-rated portfolios consist of the top 10% of all companies with the highest rating. The low-rated portfolios consist of the bottom 10% of all companies with the lowest rating. Time period from 2004 to 2007 is defined as before-crisis, data from 2008 to 2011 is defined as during-crisis and data from 2012 to 2015 is defined as post-crisis. Significance at 10%, 5% and 1% is represented by *, **, *** respectively.
Table 6 presents the performance of the high-minus-low portfolios during different time periods. All the alphas are annualized for better presentation. Comparing the result of high-rated portfolios and low-rated portfolios with high-minus-low portfolios, result of high-minus-low portfolios show similar performance at a higher significance level. For the high-minus-low portfolio based on ESG aggregate rating, it generates an alpha of 4.1% during financial crisis period at 5% significance level. This portfolio generate negative abnormal returns before and after crisis periods, however, not significantly. This result is consistent with Nofsinger and Varma (2013), which they acquire a
significantly 1.66% alpha for SRI-minus-Conventional (long in SRI and short in conventional investment) fund during financial crisis period.
For the portfolios based on separated ESG ratings, some statistically significant abnormal returns are acquired for each period. Social and governance screening portfolios generate negative abnormal returns in pre-crisis periods, the alphas are both statistically significant at 10% level. Environmental and governance screening portfolios generate significantly positive abnormal returns during crisis periods. Unlike other portfolios, social screening portfolio doesn’t generate positive alpha during crisis periods. However, Nofsinger and Varma (2013) found significantly positive alpha for all the three ESG separated portfolios during crisis periods.
6. Robustness Test
Nofsinger and Varma (2013) as well as Leite and Cortez (2015) build equally-weighted portfolios instead of value-weighted portfolios to test SRI performance. Kempf and Osthoff (2007) build both equally-weighted and value-weighted portfolio to test SRI performance. To eliminate the potential size effect, I also build equally-weighted portfolios for robustness test.
Table 7 presents the performance of equally-weighted high-minus-low portfolios during different time period. CAPM model is also tested for robustness test. Formal researches use both CAPM and Fama-French factors to test performance, Nofsinger and Varma even add Carhart 4-factors model. For robustness test I use both CAPM and Carhart’s alpha. Moreover, the equally-weighted portfolios are constructed with stocks that not only ranks on top (bottom) 10% among my investment universe, but also another cut-off of 5%.
Table 7: performance for equally-weighted high-minus-low portfolios
Entire-time Before Crisis During Crisis After Crisis
Carhart Alpha CAPM Alpha Carhart’s Alpha CAPM Alpha Carhart’s Alpha CAPM Alpha Carhart’s Alpha CAPM Alpha ESG5% 1.29 1.48 -2.12* -2.21* 2.33* 2.44* -2.53 -2.55 ENV5% -1.71 -1.62 -1.84 -1.69 3.61* 3.56* -5.54*** -5.67*** SOC5% -1.21 -1.02 -4.24** -4.21** 1.68 1.59 -4.91** -4.88** GOV5% -1.91 -1.73 -3.84* -3.92* 4.78** 4.88** -6.94*** -6.85*** ESG10% 0.49 0.55 -1.21 -1.30 1.99* 2.04* -4.11** -4.29** ENV10% -0.58 -0.64 -1.98 -1.99 2.87* 2.69* -3.49* -3.22* SOC10% -1.25 -1.26 -3.66* -3.58* 1.47 1.40 -4.99** -5.10** GOV10% -1.43 -1.35 -2.76* -2.81* 3.39* 3.58* -5.29*** -5.19***
This table presents statistics on SRI portfolio performance during different time periods. This table summarises for each screen the annualised abnormal return using the Carhart four-factor model and CAPM model. SRI stocks refer to stocks trade in NYSE with social responsibility screens. The high-rated portfolios consist of the top 10% of all companies with the highest rating. The low-rated portfolios consist of the bottom 10% of all companies with the lowest rating. Time period from 2004 to 2007 is defined as before-crisis, data from 2008 to 2011 is defined as during-crisis and data from 2012 to 2015 is defined as post-crisis. Significance at 10%, 5% and 1% is represented by *, **, *** respectively.
7. Conclusion:
The number of investors that incorporate SRI screens into their investment decisions have dramatically grown over the past couple of decades. Previous studies that try to answer the question of how SRI strategy influence their portfolio return present very mixed results. Many people believe SRI pays a cost compare to conventional investment, but the ideal of paying a cost for limiting portfolios with attributes is in contract with the popularity of SRI. In this paper, I test the performance of portfolios based on ESG ratings in different time periods to analyze whether investors could increase their performance by implementing social responsible investment.
In contract with Kempf and Osthoff (2007), the result of my study doesn’t support that investors can earn remarkable high abnormal returns by following the SRI strategy described above. More importantly, SRI shows different performance under different time periods. In general, SRI outperforms conventional investment during financial crisis with a cost in non-crisis periods. The result of this paper shows that SRI portfolios perform significantly better during crisis period than during non-crisis periods, which is consistent with Nofsinger and Varma (2014)’s study on SRI funds. SRI portfolios based on aggregated ESG rating shows significant 4.1% Carhart’s alpha in crisis period. SRI portfolios based on environmental and governance rating also shows significant and positive alpha in crisis periods. However, SRI portfolios based on social rating doesn’t generate positive alpha in crisis period. Moreover, unlike previous researches on SRI performance under different market states, I separate non-crisis periods into pre-crisis and post-pre-crisis periods. The result shows that SRI portfolios generate significantly negative alpha in non-crisis periods, and SRI portfolios perform even worse in post-crisis periods than pre-post-crisis periods. SRI significantly underperform conventional investment in any ESG screening method in post-crisis periods. The portfolios constructed on separate environmental, social and governance screening generate -6.9%, -6.8% and -4.3% annualized alpha in post-crisis periods. The result is not significant for aggregated ESG screened portfolio. The insignificant abnormal return of SRI in entire sample period is driven by their adverse performance in crisis and post-crisis periods. Their significantly negative abnormal return in non-post-crisis periods equal-out their significant positive abnormal return in crisis-periods. Robustness test which test equally-weighted portfolios with both Carhart’s and Jensen’s alpha support my main finding. Overall, my results suggest that past SRI ratings are a valuable information for investors. However, investors need to carefully implement SRI strategy because it shows different performance under different market states.
Reference
:
Bauer, R., Koedijk, K., and Otten, R., 2005, International Evidence on Ethical Mutual Fund Performance and Investment Style, Journal of Banking and Finance, 29, 1751-1767
Borgers, A., Derwall, J., Koedijk, K., & ter Host, J. 2015. Do social factors influence investment behavior and performance? Evidence from mutual fund holdings. Journal of banking and finance, 60,112-126
Bollen, N. (2007). Mutual fund attributes and investor behavior. Journal of Financial and Quantitative Analysis, 42(3), 683–708.
Carhart, M. M. 1997. On persistence in mutual fund performance, Journal of Finance, Vol. 52, 57–82.
Chatterji, A., Levine, D,. Toffel, M,. 2009. How Well Do Social Ratings Actually Measure Corporate Social Responsibility? Journal of Economics & Management Strategy, 18, 125–169
Derwall, J., Bauer, R., Guenster, N., and Koedijk, K., 2004, Socially Responsible Investing: The Eco-Efficiency Premium Puzzle, Working Paper, Erasmus University.
Diltz, J. D., 1995. Does social screening affect portfolio performance?, Journal of Investing, Vol. 4,
Edmans, A. 2011. Does the stock market fully value intangibles? Employee satisfaction and equity prices. Journal of Financial Economics, 1–32.
Guerard, J. B., Additional evidence on the cost of being socially responsible in investing, Journal of Investing, Vol. 6, 1997, 31–36.
Guenster, Nadja, 2012. Performance implications of SR investing: past versus future.In: Baker, Kent, Nofsinger, John R. (Eds.), Socially Responsible Finance and Investing. John Wiley & Sons Inc.
Hamilton, S., Jo, H., and Statman, M., 1993, Doing Well While Doing Good? The Investment Performance of Socially Responsible Mutual Funds, Financial Analysts Journal, Nov/Dec, 62-66.
Hirshleifer, David A., 2008. Psychological bias as a driver of financial regulation. European Financial Management 14 (5), 856–874.
Jegadeesh, N., & Titman, S. 1993. Returns to buying winners and selling losers: Implications for stock market efficiency. The Journal of Finance, 48(1), 65.
Kempf, A., & Osthoff, P. 2007. The effect of socially responsible investing on portfolio performance. European Financial Management, 13(5), 908-922.
Leite, P., & Cortez, M. C. 2014. Selectivity and timing abilities of international socially responsible funds. Applied Economics Letters, 21(3), 185–188
Luther, R. G., J. Matatko, and D. C. Corner. 1992. The Investment Performance of UK ‘Ethical’ Unit Trusts. Accounting Auditing & Accountability Journal 5 (1): 57–70. Merton, R, 1987. A Simple Model of Capital Market Equilibrium with Incomplete Information Journal of Finance, 42(3), 483-510
Naika, M., Yamaguchi, K,. &Takeuchi,K. 2016. Can SRI funds better resist global financial crisis? Evidence from Japan. International Review of Financial Analysis 48 12–20
Nofsinger, J., Varma, A., 2014. Socially responsible funds and market crises. Journal of Bank. Finance. 48, 180-193
Oh,W& Park,S. 2015. The Relationship Between Corporate Social Responsibility and Corporate Financial Performance in Korea. Emerging Markets Finance & Trade, 51, 85-94.
Renneboog, L., Terhorst, J., & Zhang, C. 2008. Socially responsible investments: Institutional aspects, performance, and investor behavior. Journal of Banking & Finance, 32(9), 1723–1742.
Sauer D. A. 1997 The Impact of Socially-Responsible Screens on Investment Performance: Evidence from the Domini 400 Social Index and Domini Equity Mutual Fund. Review of Financial Economics 6(2):137–149
Semenova, N,. Hassel, L,. 2015. On the Validity of Environmental Performance Metrics. Journal of Business Ethics, 132, 249-258.
Statman, M., 2000, Socially Responsible Mutual Funds, Financial Analysts Journal, May/June, 30-39.
Statman, M. 2006. Socially Responsible Indexes: Composition, Performance and Tracking Error. Journal of Portfolio Management 32 (3): 100–109.
