Corporate social responsibility and stock price performance in the
United States, Europe, and the Asia-Pacific
Bauke ten Brinke
11012552 University of Amsterdam Faculty of Economics and Business – Finance Bachelor thesis Final draft, January 2018 AbstractIn this thesis the relationship between corporate social responsibility (CSR) and stock price performance in the US, Europe and the Asia-Pacific is empirically tested. Synthetic sustainability portfolios are created using Asset4 ESG-scores and portfolio performance is measured using the Carhart four-factor model. It contributes to the current literature by expanding the use of Asset4 data to Europe and the Asia-Pacific for the first time and concurrently increasing the timeframe by 50%. Also, it is the first to measure the performance of sustainability portfolios in the Asia-Pacific using the Carhart four-factor model. The main result of this thesis is that high scoring companies do not significantly out- or underperform low scoring companies in any of the three regions.
Statement of Originality
This document is written by Bauke ten Brinke, who declares to take full responsibility for
the contents of this document.
I declare that the text and the work presented in this document are original and that no
sources other than those mentioned in the text and its references have been used in creating
it.
The Faculty of Economics and Business is responsible solely for the supervision of
completion of the work, not for the contents.
1. Introduction In his essay ‘The Social Responsibility of Business is to Increase its Profits’ Milton Friedman (1970) states that the main responsibility of corporate executives is to maximize shareholder value. If an executive spends corporate money on socially responsible projects that are not increasing shareholder value, that executive is basically spending somebody else’s money. Friedman’s point of view here is that if a shareholder wishes to participate in social responsibility he can do so himself and does not need an executive to do it for him.
This shareholder theory in which the management is solely serving its shareowners is contradictory to the stakeholder theory as described by Donaldson and Preston (1995). They even go as far as to state that this shareholder theory is morally untenable. In their article, they quote The Economist (1992) : “In America, for instance, shareholders have a comparatively big say in the running of the enterprises they own; workers . . . have much less influence. In many European countries, shareholders have less say and workers more . . . [I]n Japan . . . managers have been left alone to run their companies as they see fit – namely for the benefit of employees and of allied companies, as much as for shareholders.” These two different views on corporate social responsibility (CSR) and these differences in corporate culture between the regions raises the question “Is CSR valued differently by investors around the world?”. More specifically, this thesis will try to answer the following research question:
Does corporate social responsibility influence stock price performance and does this differ between the United States, Europe and the Asia-Pacific?
The notion of corporate social responsibility or ethical investing finds its origin in Jewish, Christian, and Islamic traditions (Renneboog, Ter Horst, & Zhang, 2008). It has however only reappeared in modern western society in the 1960s when a series of social campaigns in America made investors aware of certain social consequences of their investments. This lead to the founding of the first social responsible investment (SRI) mutual fund in the US in 1971; the Pax World Fund. SRI came to Europe in the 1980s when the racist system of apartheid in South Africa became the topic of protest by social investors. Environmental disasters such as at the Chernobyl nuclear power plant in 1986 and the spilling of 11 million gallons of crude oil from the Exxon Valdez supertanker in 1989
expended the focus to environmental topics. Since the 1990s SRI has shown a strong growth not only in the US and Europe but all over the world (Renneboog, Ter Horst, & Zhang, 2008). Because of a series of corporate scandals in the 2000s and the rising awareness of climate change, SRI screens have expanded to include the current environmental, social responsibility, and corporate governance (ESG) pillars.
Based around these pillars rating agencies started keeping track of data on corporate performance in these matters. One of the first was Kinder, Lydenberg, and Domini Research & Analytics (KLD) in 1990 with binary scores for strengths and concerns (Halbritter & Dorfleitner, 2015). In 2002 Asset4 was the first rating agency to offer numerical ESG scores. They were soon followed by Sustainalytics (2004) and Bloomberg (2005). Because of the limited amount of data available prior to 2002, and even in the years following because of short time-series, most research from before 2007 has been done using data on SRI mutual funds and comparing their performance with conventional mutual funds (Hamilton, Jo, & Statman, 1993) (Bauer, Koedijk, & Otten, 2005) (Dorfleitner, Utz, & Wimmer, 2014). Utz and Wimmer (2014) compared ESG scores of companies held by SRI mutual funds and conventional mutual funds. They find that SRI mutual funds are useful for investors wanting to avoid the least ethical companies but that the label ‘SR mutual fund (p. 11)’ does not in any way guarantee that there are no unethical firms in the fund. Concluding that ‘SRI appears to have become more of a sales pitch than a reliable path to accomplish ethical preferences (p. 11)’.
Starting in 2005 researchers started creating synthetic portfolios based on ESG scores provided by the various rating agencies (Derwall, Guenster, Bauer, & Koedijk, 2005) (Kempf & Osthoff, 2007) (Statman & Glushkov, 2009) (Mollet & Ziegler, 2014) (Halbritter & Dorfleitner, 2015) (Auer & Schuhmacher, 2016). This is now the most widely used approach. The remainder of this thesis is organized as follows. Section 2 will discuss the recent literature on the relation between CSR and stock price performance (SSP) and its global implications. In section 3 the hypotheses will be formulated, section 4 will outline the methodology, section 5 will introduce the data and section 6 will present the results. Finally, section 7 will conclude and section 8 will discuss. 2. Related literature One of the first to construct their own asset portfolios based on company level sustainability performance were Derwall, Guenster, Bauer and Koedijk (2005). They used Innovest Strategic Value Advisors’ corporate eco-efficiency scores to construct a high-ranking and a low-ranking portfolio. Finding that the high-ranked portfolio provided substantially higher average returns than the low-
ranked portfolio. Many papers followed using different datasets such as KLD (Kempf & Osthoff, 2007) (Statman & Glushkov, 2009), Asset4 (Auer & Schuhmacher, 2016), data from the Zurich Cantonal Bank (ZKB) (Mollet & Ziegler, 2014), Bloomberg, or a combination hereof (Halbritter & Dorfleitner, 2015). Most of them focussed on the US, with only Mollet and Ziegler (2014) expanding this to include Europe, and Auer and Schuhmacher (2016) to include both Europe and the Asia-Pacific. Aside from Auer and Schuhmacher (2016), who used the Sharpe-ratio (1963), the Carhart (1997) four-factor model was used to measure portfolio performance. Combining the findings of about 2200 studies on the relation between ESG criteria and stock price performance, Friede, Busch and Bassen (2015) showed that 90% of these studies find a nonnegative relation between ESG criteria and SSP. Almost 50% of these studies even found a positive relation. However, based on the current empirical literature there seem to be no (more) excess returns on ESG based portfolios in the US, Europe and the Asia-Pacific (Hamilton, Jo, & Statman, 1993) (Bauer, Koedijk, & Otten, 2005) (Mollet & Ziegler, 2014) (Halbritter & Dorfleitner, 2015) (Auer & Schuhmacher, 2016). This is also confirmed by Bebchuk, Cohen and Wang (2013) who found that there have been positive excess returns in the past but that the mispricing has disappeared over time. The use of CSR, however, differs significantly across the world. US firms started first with explicitly reporting their CSR practices, to be followed in recent years by European firms (Matten & Moon, 2008). In Asia, CSR and SRI have gotten off to a slow start but are starting to attract more attention (Cheung, Tan, Ahn, & Zhang, 2010).
Differences herein can be explained by historically grown institutional frameworks (Matten & Moon, 2008). These frameworks can be identified by four key features: the political system, the financial system, the education and labour system, and the cultural system (Whitley, 1999). There are many differences in these systems among the US, Europe and the Asia-Pacific. Some notable differences between the US and Europe are that European governments are more involved in economic and social activity (Heidenheimer, Heclo, & Adams, 1990). In the US, the stock market is the central source for corporate funding while in Europe banks are the main source (Becht & Roëll, 1999). From a cultural point of view Americans are regarded to have a higher relative capacity for participation than Europeans (De Tocqueville, 2002). Asian markets are again quite different from the US and European markets. In Asia, many companies are family or state owned which makes the equity market less liquid. Because of this ownership structure, market discipline mechanisms do not function the same way as in western countries (Cheung, Tan, Ahn, & Zhang, 2010). Also, there are some significant differences between countries within the Asia-Pacific (Matten & Moon, 2008). Take the three largest economies: China, India, and Japan. China is a government-dominated transitional country while in India the government does collect taxes but returns are absent in many regions
(Sharma, 2011). Japan, on the other hand, already has a much more developed economy similar to the US and Europe. Even though CSR is not new to this region, it has only, much like in Europe, mostly manifested itself implicitly rather than explicitly (Baughn, Bodie, & McIntosh, 2007) (Matten & Moon, 2008). Many of the concepts and tools that are now used to describe CSR are derived from western ideas and practices (Chapple & Moon, 2007). The measurable part of CSR is the explicit part, as the implicit part is normally not articulated by individual corporations (Matten & Moon, 2008). Matten and Moon (2008) find that Europe is moving towards more explicit CSR and expect this to expand to Asian countries like Japan, India and South-Korea. Countries which are historically characterized by strong implicit CSR. They expect this transition to take more time in government-dominated transitional countries such as China. And while there is extensive empirical research on the relation between CSR and stock price performance in the US and some on this relation in Europe there is hardly any on this subject for the Asia-Pacific, making it an especially interesting region to incorporate in this research.
Closest to this research are Bauer et al. (2005), Mollet and Ziegler (2014), Halbritter and Dorfleitner (2015), and Auer and Schuhmacher (2016). Bauer et al. (2005) looked at the performance of mutual funds instead of creating high-low portfolios based on company data. Mollet and Ziegler (2014) explored the US and European market using data from the Zurich Cantonal Bank, Halbritter and Dorfleitner (2015) only examined the US stock market. And Auer and Schuhmacher (2016) conducted their research using previously unavailable ESG rating data from Sustainalytics, and a different performance measure from the Carhart (1997) four-factor model. This thesis will contribute to the existing literature by expanding the use of Asset4 data to Europe and the Asia-Pacific for the first time. It will use Asset4 data from 2002-2016, increasing the length of the time-series by almost 50%. It will also be the first to measure the performance of sustainability portfolios in the Asia-Pacific using the Carhart four-factor model. 3. Hypotheses
When looking at stock price performance, a stock (or portfolio), can either be overpriced, underpriced, or be correctly priced. When a stock is overpriced this will lead to diminished returns while the market corrects for this. In case of an underpriced stock there will be positive abnormal returns during this correction period. At the end of such a period, or if there has been no mispricing to begin with, the stock will show no abnormal returns relative to the market. Previous studies on the relationship between CSR and SSP have discussed different hypotheses to explain each of the possible scenarios (Hamilton, Jo, & Statman, 1993) (Bauer, Koedijk, & Otten, 2005).
The first of these hypotheses is that SRI increases demand for high ESG rated stocks, increasing their price and consequently diminishing future returns. This can be explained through the price effect of tastes (Fama & French, 2007). They describe socially responsible investing as ‘an extreme form of tastes for assets as consumption goods that are unrelated to returns (p. 675).’ Investing primarily in socially responsible investments is a violation of the assumption that investors are only concerned with their portfolio return. This deviation leads to a disagreement in the pricing of SRIs. Regardless of this disagreement, the market must clear, meaning that rational investors will overweight non-SRIs and underweight SRIs in their portfolios to compensate for the biased investor. This should, in theory, offset the price effects of the disagreement. The prices will, however, only converge to rational values completely if in the end there is no more disagreement about the valuation of SRIs. But as taste is considered an exogenous variable and there is no economic logic or empirical evidence to suggest that the disagreement will eventually disappear, stocks might remain overpriced, showing no diminished returns (Fama & French, 2007) (Friedman & Heinle, 2016). This is supported by previous studies finding that high rated ESG stocks mostly either outperform the market or perform conform the market. The second hypothesis is that companies invest in CSR, but that the actual value of this investment is not adequately recognized by investors, leading to underpriced stocks. Because also in this situation the market must clear, an outperformance of the market is expected for these stocks. The literature shows that this has been the case in the past for the US market (Derwall, Guenster, Bauer, & Koedijk, 2005) (Statman & Glushkov, 2009) (Edmans, 2011) (Eccles, Ioannou, & Serafeim, 2014). However, Bebchuk et al. (2013) and Borgers et al. (2013) have shown that this is no longer the case. Which leads to the third hypothesis where SRI stocks are not mispriced and hence do not show abnormal returns. This hypothesis is also the one supported by the traditional view of finance; that publicly available information cannot lead to abnormal returns, as its value will be reflected in the stock price instantly. As shown by Bebchuk et al. (2013) and Borgers et al. (2013) this is the current state of the US market. In Europe, CSR seems not to have provided excess returns – historically or current. This is most likely because of the long tradition of implicit CSR and social responsibility through governments (Matten & Moon, 2008). For the Asia-Pacific region, after years of focusing solely on economic growth there is now a transition to sustainability (Berkhout, et al., 2010). Together with a rising purchasing power in these countries and following the developments seen in early SRI in the US, this could be expected to lead to the development of a taste for CSR from investors. Leading to positive abnormal returns for high ESG rated companies. However, with a long tradition of implicit CSR, it is more likely that the Asia-Pacific will follow the course of Europe and show no mispricing of CSR. This is also supported by Auer and Schuhmacher (2016) finding no abnormal returns in this region.
The statistical hypothesis that will be tested for all portfolios in all three regions is then whether or not the estimated alfas differ from zero: H0: αi = 0 H1: αi ≠ 0
For the individual portfolios, there will also be a test for outliers – to find whether any individual portfolio outperforms another. The statistical hypothesis that will be tested for this is whether or not the estimated alphas differ from their respective means: H0: αi = µα H1: αi ≠ µα 4. Methodology and analysis 4.1 Methodology The main models used for measuring portfolio performance are the one-factor market model (Sharpe, 1963) and the CAPM (Lintner, 1965). After discussions about the validity of the CAPM, two control variables have been added to improve its accuracy (Fama & French, 1993). The size variable, Small-Minus-Big (SMB), has been added to correct for the often-higher return of small companies compared to big companies. And the value factor, High-Minus-Low (HML), has been added to correct for the often-different returns of high- and low book-to-market ratio companies. This improved 3-factor model has since been proved to be more accurate than the single variable CAPM (Fama & French, 1996). However, almost instantly the validity of this model was also questioned by Jegadeesh and Titman (1993) leading to the Carhart (1997) four-factor model, which includes the Win-Minus-Lose (WML) control variable. With this fourth control variable – the difference in return of stocks that were rising (winning) in the previous year with those that were falling (losing) – it is now the most widely used model to measure portfolio performance and will also be used here. To measure the influence of CSR on firm performance with an asset pricing model such as Carhart’s four-factor model, the panel-data provided by Asset4 must be converted into time-series data. This is done, in line with current empirical literature, by creating synthetic ESG portfolios (Bauer, Koedijk, & Otten, 2005) (Derwall, Guenster, Bauer, & Koedijk, 2005) (Mollet & Ziegler, 2014) (Halbritter & Dorfleitner, 2015). For each region, a portfolio will be constructed on a yearly basis by buying the top 10% of high-rated firms and selling short the bottom 10%. Also for each region eight 12,5%
portfolios will be constructed representing the entire range of ESG-scores. The performance of these portfolios will be measured using the aforementioned Carhart (1997) four-factor model as specified below: 𝑟",$− 𝑟&,$= 𝛼"+ 𝛽" 𝑟+,$− 𝑟&,$ + 𝑠"𝑆𝑀𝐵",$+ ℎ"𝐻𝑀𝐿",$+ 𝑤"𝑊𝑀𝐿",$+ 𝜀",$, (1) where ri,t – rf,t is the excess return of portfolio i over the risk-free rate, rm,t – rf,t is the excess return of the market proxy over the risk-free rate, and SMBi,t, HMLi,t, and WMLi,t are the size, value, and momentum factors respectively. The coefficients αi, βi, si, hi, and wi are estimated performing a linear regression and 𝜀",$ represents the residual. The main coefficient of interest in this regression is the α, which depicts the monthly abnormal returns of the, in this case, ESG based portfolios. This will then show whether the portfolios have under- or outperformed the market proxy. 4.2 Analysis First, all data retrieved from Datastream and Bloomberg Professional is downloaded into excel. In excel all columns (date range, companies, indices and risk-free rates) are given unique identifiers in the first row so that they can be imported into STATA. Datasets downloaded from the Andrea Frazzini data-library and the central bank data-libraries are already in excel format and have unique identifiers. All data is imported into separate STATA files. ESG company data and company return data is then first transposed to create panel data that can be merged. Twelve USA companies constantly returned an error when downloaded from Datastream; these were removed from the dataset. All companies were then sorted by date and merged with the control variables. In the now complete dataset the percentile portfolios were created and heteroscedasticity-robust regressions were run on each of the (in total) 30 ESG-portfolios. On all data, the preparation of the datasets to make them ready for regressions and the regressions themselves were performed with the same STATA-code to prevent human discrepancies between the 3 regions and different portfolios. Where necessary yearly data was transformed to monthly data and/or was multiplied by 100% for all data to be in the same order of magnitude. 5. Data and descriptive statistics As a measure for corporate social responsibility the ESG data from Asset4 is used. Asset4 started collecting data on more than 1000 companies in 2002 and currently collects data on more than 6000 companies worldwide. They collect this data from publicly available sources such as annual reports and CSR reports. After collecting this data on environmental, social, and corporate governance
performance it is processed to an indexed score between 0 and 100 to create a useful dataset for quantitative analysis. These scores are then discounted if companies are subject to controversy that year to get the final ESG score (ThomsonReuters, 2017). The data is retrieved for three distinct regions: The United States, Europe and the Asia-Pacific. Summary statistics are as follows: Table 1: Summary Statistics
Mean Std. Dev. Min Median Max #Companies Months
USA 51,15 29,35 2,96 50,17 98,60 805 180 Europe 62,26 29,46 2,50 74,90 98,33 750 180 Asia-Pacific 41,22 29,63 2,41 33,43 98,08 953 168 The number of companies is the average number of companies that were in the Asset4 database per year. The data for the Asia-Pacific region starts 1 year later on 01/01/2003 as opposed to 01/01/2002. Full summary statistics can be found in appendix 1. All three regions have roughly the same amount of observations. It is however noteworthy that data on the Asia-Pacific region only starts in 2003, one year later than the data on the US and Europe. Europe clearly has the highest mean and median and the Asia-Pacific the lowest. The US is in between for most years, only to fall back starkly in 2015 and 2016, when the number of observations also almost double. This might have to do with legislation such as the California Transparency in Supply Chains Act or the new European law which requires publicly traded companies that have more than 500 employees to report on sustainability factors, coming into effect in 2017.
To calculate the returns of the ESG-portfolios monthly company-return data is used from ThomsonReuters’ datastream. ESG scores and company returns are retrieved from the same database to avoid any errors that might arise when merging datasets. Monthly data is used and trades are made once a year (31 December) based on the then available ESG-scores – which are only provided on a yearly basis. The market proxy used for the US will be the S&P500, since this is the most widely accepted and used proxy. The use of this proxy will make the US results most comparable with other research. For Europe and the Asia-Pacific the market proxy will be the regional subsections of the STOXX Global 1800: the STOXX Europe 600 and the STOXX Asia-Pacific 600. This index is used because it provides the most complete proxies of the entire regions, rather than national market indices. Data for the S&P500 and the STOXX Europe 600 are retrieved from Datastream, data for the STOXX Asia-Pacific 600 is retrieved from Bloomberg Professional. The risk-free rate is defined by the 1-year treasury bill rate of the Federal Reserve, European Central Bank, and the average of the People’s Bank of China (PBC), Reserve Bank of India (RBI), and Bank of Japan (BOJ) respectively. The data for the US and Europe are retrieved from their respective online (federal) databases, while the data for the PBC, RBI
For the control variables SMB, HML and WML for the Carhart (1997) four-factor model, data is retrieved from Andrea Frazzini’s data library. This library provides separate data for the US, Europe and the Pacific. An important addition to Kenneth French’s US-only data as global data is a pour indicator for local returns (Fama & French, 2012).
6. Results
6.1. High-low portfolios
Table 2 shows the estimated results for the high, low, and high-low 10%-portfolios. The estimated parameters are those from the Carhart four-factor model. It shows the results for the US, Europe and the Asia-Pacific. Standard errors are reported in parenthesis. The high-rated US portfolio shows a weakly significant negative loading of the WML control variable. For the low-rated US portfolio the SMB variable is significantly different from zero. None of the European or Asia-Pacific portfolios show any significant correlation with the control variables. For all three regions both the high- and low portfolios show alphas significantly larger than zero. Leading to the main results of this table which are the insignificant negative alphas for the US and Asia-Pacific high-low portfolios and the insignificant positive alpha for European high-low portfolio. This is in line with the third hypothesis as introduced in section 3, that SRI stocks are currently not mispriced. Table 2: 10% high-low portfolios ⍺ MKT SMB HML WML R2 USA High 0,60*** 1,13*** -0,07 0,01 -0,05* 0,91 (0,14) (0,03) (0,06) (0,06) (0,03) Low 0,82*** 1,14*** -0,19** -0,04 -0,01 0,86 (0,18) (0,04) (0,08) (0,08) (0,04) High-Low -0,21 -0,01 0,12 0,05 -0,04 0,05 (0,22) (0,04) (0,10) (0,10) (0,05) Europe High 0,47*** 1,06** 0,05 -0,02 0,00 0,91 (0,14) (0,03) (0,06) (0,08) (0,03) Low 0,45** 1,08** -0,01 0,09 -0,06 0,83 (0,21) (0,04) (0,10) (0,12) (0,05) High-Low 0,02 -0,02 0,05 -0,10 0,06 0,08 (0,25) (0,05) (0,11) (0,14) (0,06) Asia-Pacific High 0,51** 0,86*** -0,04 -0,00 0,01 0,71 (0,21) (0,04) (0,10) (0,12) (0,06) Low 0,94*** 0,87*** 0,01 -0,01 -0,04 0,66 (0,24) (0,05) (0,11) (0,13) (0,07) High-Low -0,44 -0,02 -0,06 0,00 0,04 0,05 (0,32) (0,07) (0,15) (0,17) (0,09)
High portfolios are comprised of the 10% companies that have the highest ESG-score. Low portfolios of companies that have the lowest scores. ⍺ represents the ⍺ in equation 1. MKT represents the β, SMB the s, HML the h and WML the w. All estimated results are monthly values. Standard error is denoted in parentheses. *, **, and *** represent a 10%, 5%, and 1% significance level respectively.
Table 3:
12,5% complete range portfolios
⍺ MKT SMB HML WML R2
USA Best performing 12,5% 0,68*** 1,15*** -0,08 0,05 -0,06* 0,91
(0,14) (0,03) (0,07) (0,06) (0,03) 12,5% - 25% 0,73*** 1,17*** -0,13* 0,07 -0,04 0,90 (0,15) (0,03) (0,07) (0,07) (0,03) 25% - 37,5% 0,71*** 1,11*** -0,10 0,09 -0,04 0,92 (0,13) (0,03) (0,06) (0,06) (0,03) 37,5% - 50% 0,75*** 1,12*** -0,05 0,08 -0,05* 0,90 (0,14) (0,03) (0,07) (0,07) (0,03) 50% - 62,5% 0,63*** 1,14*** -0,15** 0,04 -0,04 0,91 (0,14) (0,03) (0,06) (0,06) (0,03) 62,5% - 75% 0,69*** 1,18*** -0,17** 0,10 -0,03 0,86 (0,18) (0,04) (0,08) (0,08) (0,04) 75% - 87,5% 0,47*** 1,05 -0,04 0,08 -0,00 0,87 (0,15) (0,03) (0,07) (0,07) (0,03) Worst performing 12,5% 0,82*** 1,12*** -0,15* -0,04 -0,01 0,86 (0,17) (0,03) (0,08) (0,08) (0,04)
Europe Best performing 12,5% 0,49*** 1,07*** 0,08* 0,05 0,01 0,91
(0,10) (0,02) (0,04) (0,05) (0,02) 12,5% - 25% 0,63*** 1,07*** 0,07 0,05 -0,02 0,87 (0,12) (0,02) (0,05) (0,07) (0,03) 25% - 37,5% 0,62*** 1,05** 0,01 0,01 -0,02 0,85 (0,13) (0,02) (0,06) (0,07) (0,03) 37,5% - 50% 0,66*** 1,04* -0,02 0,07 0,00 0,86 (0,12) (0,02) (0,06) (0,07) (0,03) 50% - 62,5% 0,62*** 1,04* 0,02 0,04 -0,05* 0,87 (0,12) (0,02) (0,06) (0,07) (0,03) 62,5% - 75% 0,48*** 1,05** 0,00 0,06 -0,05* 0,86 (0,12) (0,02) (0,06) (0,07) (0,03) 75% - 87,5% 0,60*** 1,05* 0,00 -0,02 -0,06** 0,84 (0,13) (0,03) (0,06) (0,08) (0,03) Worst performing 12,5% 0,59*** 1,06** -0,00 -0,00 -0,07** 0,83 (0,14) (0,03) (0,07) (0,08) (0,03)
Asia-Pacific Best performing 12,5% 0,50** 0,87*** -0,03 -0,03 -0,01 0,70
(0,22) (0,04) (0,10) (0,12) (0,06) 12,5% - 25% 0,71*** 0,85*** -0,01 0,01 -0,02 0,70 (0,22) (0,04) (0,10) (0,12) (0,06) 25% - 37,5% 0,72*** 0,89** -0,03 -0,03 -0,05 0,68 (0,24) (0,05) (0,11) (0,13) (0,07) 37,5% - 50% 0,84*** 0,80*** -0,03 -0,00 -0,02 0,69 (0,21) (0,04) (0,10) (0,11) (0,06) 50% - 62,5% 0,77*** 0,90** -0,08 -0,01 -0,05 0,73 (0,21) (0,04) (0,10) (0,12) (0,06) 62,5% - 75% 0,84*** 0,91* -0,01 -0,06 -0,05 0,69 (0,23) (0,05) (0,11) (0,13) (0,06) 75% - 87,5% 0,76*** 0,90** -0,05 -0,02 -0,11* 0,70 (0,23) (0,05) (0,11) (0,13) (0,06) Worst performing 12,5% 0,93*** 0,86*** -0,01 -0,02 -0,04 0,68 (0,23) (0,05) (0,11) (0,12) (0,06) Percentage ranges indicate the relative ESG-scores of the companies that make up the designated portfolio. ⍺ represents the ⍺ in equation 1. MKT represents the β, SMB the s, HML the h and WML the w. All estimated results are monthly values. Standard error is denoted in
Graph 1: USA Graph 2: Europe Graph 3: Asia-Pacific
Alphas of all 12,5% portfolios depicted against their relative regional ESG-performance
6.2. 12,5% complete range portfolios The results of the 12,5% complete range portfolios are depicted in tables 3 and 4, and Graphs 1 through 3. Table 3 shows the estimated values for the 12,5%-portfolios for the entire spectrum of ESG scores. First the results for the US are shown, listing the portfolio comprised of the highest ESG scoring companies on top and the lowest ESG scoring companies at the bottom. In the middle, the same is shown for Europe and on the bottom for the Asia-Pacific. Regression factors are again from the Carhart four-factor model and standard errors are reported in parenthesis. The slightly different composition of the portfolios (12,5% against 10%) in itself does not show any significant differences. All alphas for the high- and low portfolios are still significantly larger than zero and all high-low portfolio alphas are still not significantly different from zero. The best performing US portfolio still shows a significant negative correlation with the WML portfolio and the worst performing with the SMB portfolio. All other estimated alphas are also significantly different from zero. For Europe now all four worst performing portfolios show a significant negative loading for the WML factor. While in the Asia-Pacific, except for one weakly significant negative correlation of the 75-87,5% portfolio with the WML factor there are still no significant correlations with the control variables. Table 4:
Portfolio alphas and corresponding p-values Best performing 12,5% 12,5-25% 25-37,5% 37,5-50% 50-62,5% 62,5-75% 75-87,5% Worst performing 12,5% µ USA ⍺ 0,68 0,73 0,71 0,75 0,63 0,69 0,47 0,82 0,69 p-value (0,98) (0,76) (0,85) (0,64) (0,69) (0,98) (0,15) (0,43) Europe ⍺ 0,49 0,63 0,62 0,66 0,62 0,48 0,60 0,59 0,59 p-value (0,32) (0,70) (0,77) (0,55) (0,80) (0,40) (0,95) (0,97) Asia-Pacific ⍺ 0,50 0,71 0,72 0,84 0,77 0,84 0,76 0,93 0,76 p-value (0,24) (0,82) (0,86) (0,70) (0,97) (0,73) (0,98) (0,45) Percentage ranges indicate the relative ESG-scores of the companies that make up the designated portfolio. ⍺ represents the ⍺ in equation 1. p-values are relative to their respective means as denoted in the last column. µ represents the mean of the respective alphas.
However, because on the wider range of portfolios analysed, there is the possibility to look for outliers in the results of table 3. This is done by graphically showing the alphas against their relative regional ESG performance in Graph 1 to 3 for the US, Europe and the Asia-Pacific respectively and also numerically in Table 4. This table shows the same estimated alphas as in table 3, however, now they are tested against their means, which are shown at the far right of the table. In parenthesis are the p-values of the alphas, corresponding to their means. Graph 1 (USA) shows a lower performance for the second to last rated portfolio only to bounce back up for the lowest rated ESG-portfolio. This drop in stock price performance might be caused by negative media attention for the worst performers, leading to a lower demand and lower returns. One assumption, however, has to hold for this explanation to be plausible: the worst performing portfolio has to be comprised mainly of so called sin-stocks (like alcohol and tobacco companies), pushing the non-sin, low ESG rated companies into the second to last portfolio. This large presence of sin-stocks then explains the rebound of the performance for the lowest rated portfolio, as these stocks generally get little media coverage and often outperform the market (Fabozzi, Ma, & Oliphant, 2008) (Hong & Kacperczyk, 2009). Table 4, however, shows that these differences are not significant which is again in support of hypothesis 3. Further research might provide a definite answer. Graph 2 (Europe) shows hardly any variation in the portfolio returns. This is also supported by the insignificant alphas as shown in table 4. Graph 3 (Asia-Pacific) shows an outperformance of high-rated ESG companies by low-rated ESG companies, suggesting a correction from an overvaluation of SRI stocks in the concerning time-period, in support of hypothesis 1. Only, here too, the alphas are insignificantly different from zero which also supports hypothesis 3. 7. Conclusion
This thesis empirically tested how corporate social responsibility influences stock price performance in the US, Europe and the Asia-Pacific. This was done in line with other research by creating synthetic sustainability portfolios. These portfolios were created with Asset4 ESG data. It has contributed to the current literature by expanding the use of the Asset4 ESG data to Europe and the Asia-Pacific and by being the first to use the Carhart (1997) four-factor model for sustainability performance analysis in the Asia-Pacific.
The main results of this thesis are the insignificant abnormal returns for the high-low portfolios in the US, Europe and the Asia-Pacific, as well as that none of the 12,5%-portfolios had significantly divergent returns. The research question: “Does corporate social responsibility influence stock price
therewith be answered in the negative. No, corporate social responsibility does not influence stock price performance and no, this does not differ between the United States, Europe and the Asia-Pacific. These results are in support of the third hypothesis “SRI stocks are currently correctly priced and hence do not show any abnormal returns”. Based on this research, however, it cannot be concluded that mispricing has not existed before 2002 or does not exist in a region outside of the US, Europe and the Asia-Pacific Non-significant variations in the returns of the 12,5%-portfolios have been found for the bottom two US portfolios and all portfolios in the Asia-Pacific. The US showed relatively low returns for the second to last portfolio and then relatively high returns for the lowest ranked portfolio. This can be in support of the inverse of the second hypothesis, that investors are moving away from low rated stocks, but only if the assumption holds that the lowest ranked portfolio is comprised mostly of sin-stocks. The Asia-Pacific portfolios showed an outperformance of high ranked portfolios by low ranked portfolios. This could be in support of the first hypothesis. Because of the statistical insignificance of these particular results, however, nothing definite can be concluded from these variations.
8. Discussion
All high-low portfolios constructed for this thesis showed no significant abnormal returns. In contrast, however, all individual portfolios did show significant (positive) abnormal returns. This led to the overall conclusion that ESG scores do not influence stock price performance. It does, however, not say anything about the effect of reporting on ESG criteria towards stock price performance. Judging solely on the significant positive alphas there are two possible hypotheses. Either reporting on ESG criteria improves stock price performance, or, companies that outperform the market are more likely to report on these criteria. It would require further research to find an answer to this.
Another possible influence on the results are the business cycles. During the researched time period (2002-2016) there was a severe regression (financial crisis, ~2007-2011). If the valuation of ESG scores differs between such periods, any positive (or negative) results from one period could be offset by the other. Dividing the 15-year time period into three 5-year periods should show any possible susceptibility to these cycles. Then lastly, there is the difference between the United States’ and European betas (larger than 1), and the Asia-Pacific’s betas (smaller than one). This might be due to different types of companies reporting in these regions. A possible hypothesis is that in the Asia-Pacific it is mostly larger (multinational) companies that report on ESG criteria. This would be in line with current literature that in the Asia-Pacific CSR is still mostly implicit rather than explicit. However, also here, further research would be required to be able to support this hypothesis.
Bibliography
Ariely, D., Bracha, A., & Meier, S. (2009). Doing Good or Doing Well? Image Motivation and Monetary Incentive in Behaving Prosocially. The American Economic Review, 99(1), 544-555.Auer, B. R., & Schuhmacher, F. (2016). Do socially (ir)responsible investments pay? New evidence from international ESG data. The Quarterly Review of Economics and Finance, 59, 51-62. Bauer, R., Koedijk, K., & Otten, R. (2005). International evidence on ethical mutual fund performance and investment style. Journal of Banking and Finance, 29, 1751-1767. Baughn, C. C., Bodie, N. L., & McIntosh, J. C. (2007). Corporate Social and Environmental Responsibility in Asian Countries and Other Geographical Regions. Corporate Social Responsibility and Environmental Management, 14, 189-205. Bebchuk, L. A., Cohen, A., & Wang, C. C. (2013). Learning and the disappearing association between governance and returns. Journal of Financial Economics, 108(2), 323-348. Becht, M., & Roëll, A. (1999). Blockholdings in Europe: An international comparison. European Economic Review, 43(4-6), 1049-1056. Berkhout, F., Verbong, G., Wieczorek, A. J., Raven, R., Lebel, L., & Bai, X. (2010). Sustainability experiments in Asia: innovations shaping alternative development pathways? Environmental Science & Policy, 13, 261-271. Borgers, A., Derwall, J., Koedijk, K., & Ter Horst, J. (2013). Stakeholder relations and stock returns: On errors in investors' expectations and learning. Journal of Empirical Finance, 22, 159-175. Carhart, M. M. (1997). On Persistence in Mutual Fund Performance. The Journal of Finance, 52(1), 57-82. Chapple, W., & Moon, J. (2007). Introduction: CSR Agendas for Asia. Corporate Social Responsibility and Environmental Management, 14, 183-188. Cheung, Y. L., Tan, W., Ahn, H.-J., & Zhang, Z. (2010). Does Corporate Social Responsibility Matter in Asian Emerging Markets? Journal of Business Ethics, 92(3), 401-413. De Tocqueville, A. (2002). Democracy in America. (B. Frohnen, Ed., & H. Reeve, Trans.) Washington: Regnery Publishing, Inc. Derwall, J., Guenster, N., Bauer, R., & Koedijk, K. (2005). The Eco-Efficiency Premium Puzzle. Financial Analysts Journal, 61(2), 51-63. Donaldson, T., & Preston, L. E. (1995). The Stakeholder Theory of the Corporation: Concepts, Evidence, and Implications. The Academy of Management Review, 20(1), 65-91. Dorfleitner, G., Utz, S., & Wimmer, M. (2014, December 4). Patience pays off - financial long-term benefits of sustainable management decisions. Retrieved from SSRN: http://ssrn.com/abstract=2533957
Eccles, R. G., Ioannou, I., & Serafeim, G. (2014). The impact of corporate sustainability on organizational processes and performance. Management Science, 60(11), 2835-2857. Edmans, A. (2011). Does the stock market fully value intangibles? Employee satisfaction and equity prices. Journal of Financial Economics, 101, 621-640. Eichholtz, P., Kok, N., & Quigley, J. M. (2010). Doing Well by Doing Good? Green Office Buildings. American Economic Review, 100, 2492-2509. Fabozzi, F. J., Ma, K. C., & Oliphant, B. J. (2008, Fall). Sin Stock Returns. The Journal of Portfolio Management, 82-94. Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. The Journal of Finance, 33(1), 3-56. Fama, E. F., & French, K. R. (1996). Multifactor explanations of asset pricing anomalies. The Journal of Finance, 51(1), 55-84. Fama, E. F., & French, K. R. (2007). Disagreement, tastes, and asset prices. Journal of Financial Economics, 83, 667-689. Fama, E. F., & French, K. R. (2012). Size, value, and momentum in international stock returns. Journal of Financial Economics, 105, 457-572. Friede, G., Busch, T., & Bassen, A. (2015). ESG and financial performance: aggregated evidence from more than 2000 empirical studies. Journal of Sustainable Finance & Investment, 5(4), 210-233. Friedman, H. L., & Heinle, M. S. (2016). Taste, information, and asset prices: Implications for the valuation of CSR. Review of Accounting Studies, 21(3), 740-767. Friedman, M. (1970, September 13). The Social Responsibility of Business Is to Increase Its Profits. The New York Times Magazine, 33, pp. 122-126. Halbritter, G., & Dorfleitner, G. (2015). The wages of social responsibility - where are they? A critical review of ESG investing. Review of Financial Economics, 26, 25-35. Hamilton, S., Jo, H., & Statman, M. (1993). Doing Well while Doing Good? The Investment Performance of Socially Responsible Mutual Funds. Financial Analysts Journal, 49(6), 62-66. Heidenheimer, A. J., Heclo, H., & Adams, C. T. (1990). Comparative public policy: The politics of social choice in America, Europe and Japan. New York: St. Martin's Press. Hong, H., & Kacperczyk, M. (2009). The price of sin: The effects of social norms on markets. Journal of Financial Economics, 93(1), 15-36. Jegadeesh, N., & Titman, S. (1993). Returns to buying winners and selling losers: Implications for stock market efficiency. The Journal of Finance, 48(1), 65-91.
Kempf, A., & Osthoff, P. (2007). The Effect of Socially Responsible Investing on Portfolio Performance. European Financial Management, 13(5), 908-922. Lintner, J. (1965). The valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets. The Review of Economics and Statistics, 47(1), 13-37. Matten, D., & Moon, J. (2008). "Implicit" and "Explicit" CSR: A Conceptual Framework for a Comparative Understanding of Corporate Social Responsibility. The Academy of Management Review, 33(2), 404-424. Maxwell, J. W., Lyon, T. P., & Hackett, S. C. (2000). Self-Regulation and Social Walfare: The Political Economy of Corporate Environmentalism. The Journal of Law & Economics, 43(2), 583-618. Mollet, J. C., & Ziegler, A. (2014). Socially responsible investing and stock performance: New empirical evidence for the US and European stock markets. Review of Financial Economics, 23, 208-216. Renneboog, L., Ter Horst, J., & Zhang, C. (2008). Socially responsible investments: Institutional aspects, performance, and investor behavior. Journal of Banking and Finance, 32, 1723-1742. Sharma, S. (2011). Corporate Social Responsibility in India. Indian Journal of Industrial Relations, 46(4), 637-649. Sharpe, W. F. (1963). A simplified model for portfolio analysis. Management Science, 9(2), 277-293. Statman, M., & Glushkov, D. (2009). The Wages of Social Responsibility. Financial Analysts Journal, 65(4), 33-46. The Economist. (1992, September 11). Corporate governance special section. The Economist, pp. 52-62. ThomsonReuters. (2017, October). ESG scores factsheet. Retrieved from https://financial.thomsonreuters.com/content/dam/openweb/documents/pdf/financial/esg -scores-factsheet.pdf Utz, S., & Wimmer, M. (2014). Are they any good at all? A financial and ethical analysis of socially responsible mutual funds. Journal of Asset Management, 15(1), 72-82. Visaltanachoti, N., Zheng, Q., & Zou, L. (2011). The Performance of Sin Stocks in China. Journal of International Finance Studies, 11(3). Wall, L. D. (1995). Some lessons from basic finance for effective socially responsible investing. Economic Review, 1, 1-12. Whitley, R. (1999). Divergent capitalisms: The social structuring and change of business systems. Oxford: Oxford University Press.
Appendix 1
Summary statistics
Obs. Mean Std. Dev. Min
10th percentile Median 90th percentile Max USA 2002 310 45,91 28,87 4,96 12,12 38,54 91,15 98,60 2003 311 45,88 29,03 4,14 12,59 41,16 91,83 98,60 2004 443 53,79 26,40 3,56 19,55 52,48 93,28 98,31 2005 503 55,05 27,76 4,28 19,55 50,17 95,30 98,44 2006 512 55,25 27,80 4,46 20,55 51,66 95,41 98,22 2007 547 55,08 27,74 4,68 18,48 52,65 92,88 97,25 2008 709 52,33 29,08 4,02 15,13 48,30 93,35 97,42 2009 804 51,80 29,20 2,96 15,82 45,73 93,87 97,11 2010 849 54,05 28,46 4,05 17,96 50,49 93,69 97,08 2011 869 54,83 28,77 4,11 19,11 52,37 93,52 96,46 2012 874 52,50 29,04 4,12 16,06 49,56 93,19 96,75 2013 892 54,52 28,91 4,06 17,03 51,74 93,56 96,93 2014 922 55,73 28,38 3,30 17,61 56,08 93,06 97,04 2015 1579 43,09 32,31 4,66 8,59 35,85 91,69 96,26 2016 1954 37,40 30,61 5,49 10,12 19,47 90,38 96,18 Europe 2002 339 56,67 31,00 3,05 10,23 62,27 93,37 98,06 2003 344 55,50 30,83 2,87 10,50 61,13 91,96 98,14 2004 538 57,34 31,42 3,70 10,43 60,77 94,16 98,06 2005 642 57,16 30,79 3,37 10,83 61,14 94,85 98,33 2006 654 57,15 30,96 3,08 11,68 61,37 94,96 97,94 2007 705 59,44 30,18 2,67 11,89 65,50 93,74 96,93 2008 753 60,64 30,07 2,50 12,55 69,07 93,20 97,45 2009 791 64,17 29,38 2,75 14,94 74,90 94,63 97,09 2010 837 67,28 28,03 3,03 19,09 79,20 94,12 96,63 2011 873 65,99 28,71 3,25 17,73 77,99 94,20 96,47 2012 892 66,74 28,06 2,86 19,86 77,54 93,83 96,72 2013 906 66,24 28,11 2,91 18,05 77,45 93,93 96,77 2014 934 66,07 28,37 3,04 17,35 78,18 93,16 96,61 2015 1052 65,84 30,27 4,67 11,46 79,91 93,39 95,88 2016 997 67,65 29,18 5,29 16,12 82,13 93,82 96,12 Asia-Pacific 2002 0 2003 44 52,51 30,00 4,40 9,41 54,93 91,36 97,65 2004 330 35,77 27,86 3,41 5,79 26,73 76,84 98,01 2005 502 37,00 29,17 3,08 5,60 25,96 82,08 98,08 2006 508 37,81 29,54 2,90 5,00 30,75 82,12 98,02 2007 560 37,06 30,16 2,53 3,60 29,92 81,19 97,02 2008 669 42,10 30,71 2,41 4,47 40,24 85,02 96,96 2009 849 39,88 29,48 2,78 5,07 34,96 82,57 97,44 2010 1219 37,60 29,44 3,15 5,41 29,33 82,89 96,62 2011 1304 38,83 29,70 3,26 5,52 31,54 81,87 96,01 2012 1352 40,00 30,29 2,87 5,13 33,56 82,61 96,49 2013 1452 39,61 29,96 2,69 5,22 33,30 82,43 96,11 2014 1537 40,86 30,01 2,80 5,40 34,63 82,74 96,27 2015 1628 47,31 29,12 4,27 8,48 47,89 85,07 95,69 2016 1389 50,70 28,87 5,05 10,69 52,90 87,23 96,05
