1
The effects of value and growth investing strategies on the financial
and ESG performance of a best-in-class ESG stock portfolio in
Europe.
Master thesis MSc Finance Rijksuniversiteit Groningen Joppe Verbokkem S2022702Supervisor: prof. dr. L.J.R. (Bert) Scholtens Wordcount: 12.906
Abstract:
2
Content
1. Introduction 3
2. Background 6
2.1. Socially responsible investments 6
2.2. Value and growth investment strategies 7
3. Methodology 14 3.1. Portfolio analysis 14 3.2. Performance measures 16 3.2.1. Financial performance 16 3.2.2. ESG performance 18 4. Data 19 4.1. Data collection 19 4.2. Descriptive statistics 20 5. Results 24 5.1. Results 24
5.2. Robustness and sensitivity 30
6. Conclusion and discussion 33
6.1. Conclusion 33
6.2. Discussion 33
7. References 36
3
1. Introduction
This study investigates the effects of value and growth strategies on the financial and ESG performance of a best-in-class ESG stock portfolio. It does so by constructing portfolios based on value and growth strategies. The existing literature finds a significant positive effect of the value strategy on financial portfolio performance (Lakonishok, Shleifer and Vishny, 1994; Fama and French; 1998, Asness, Moskowitz and Pedersen, 2013). As Socially Responsible Investing (“SRI”) indices tend to be growth stock oriented1
, we investigate whether ESG portfolios constructed based on value and growth characteristics can lead to a difference in financial and ESG performance, as compared to a conventional ESG strategy.
There is a vast amount of literature on the relationship between social performance and financial performance (Revelli and Viviani, 2015). One of the most used measures of a company’s social performance is the ESG rating. The ESG rating is a reflection of a company’s performance in three the three pillars, environmental, social, and corporte governance, which are populated by scores on different criteria in each pillar. Studies on the relationship between ESG performance and financial performance find varying results. In the U.S., there is mostly a positive relationship or no effect of ESG criteria on stock returns (Derwall, Geunster, Bauer and Koedijk, 2005; Hallbritter, 2015). In Europe, most studies find a negative or insignificant relationship (Van de Velde and Corten, 2005; Mollet and Ziegler, 2014; Auer and Schuhmacher, 2015). Most of the studies that focus the relationship between ESG performance and financial performance either select a top percentage of ESG rated companies and control for industries or select the top percentage of ESG rated companies within each industry to create a best-in-class ESG portfolio (Auer and Schuhmacher, 2015). Although the research on ESG performance in relation to financial performance is extensive, not much research has been conducted on incorporating alternative or existing investment strategies in ESG investing.
Abramson and Chung (2000) studied such an alternative investment strategy within SRI as they investigated value premium in the context of SRI indices. By comparing average annual returns and Sharpe ratios, they found that value portfolios, constructed from a social index, can attain competitive performance compared to value indices. The value premium found in the literature refers to the abnormal returns earned by a value portfolio as compared to a growth portfolio. Although sustainability and ESG strategies are relatively new topics in
4 academic research, a rather old debate is that of the value and growth strategies. By characterising equity stocks by their financial ratios, yields, or growth rates, many studies found that value stocks significantly outperform growth stocks in different parts of the world and in different time periods (Bauman, Conover and Miller, 1998; Fama and French, 1998). The debate on this topic can be characterised by Fama and French (1992) on the one hand, who argue that the additional risk of the value strategy explains the value premium and Lakonishok et al. (1994) on the other hand, who argue that the value premium is a result of the strategy outperforming naïve investment strategies. Recent literature on value and growth investment strategies also indicates that a significant value premium exists globally as well as in Europe (Asness et al., 2013).
As Abramson and Chung (2000) construct portfolios of stocks of the MSCI KLD 400 SOCIAL INDEX and find that they can achieve competitive value and growth performance, this study extends this approach by investigating the effect of value and growth investment strategies on a best-in-class ESG environment and comparing it to an ESG benchmark. As to our knowledge, there is no publicly available research on the effects of value and growth strategies on the financial and ESG performance of an ESG investment portfolio. Hence, this study investigates these effects by constructing portfolios from a best-in-class ESG stock universe and adds to the current literature by examining if these strategies can outperform a regular or ‘passive’ ESG investment. By doing so, it tries to answer the following research question:
Does investing in value stocks or growths improve the financial and ESG performance of an above average ESG portfolio?
5 European market. Hence, we test if we can attain similar results to Abramson and Chung (2000) in the European market, and consequently improve the financial performance of an ESG portfolio. As investors are increasingly interested in social performance, we also study the effects on ESG performance of the value and growth strategies.
This study takes the approach of best-in-class ESG investing for the individual investor. Hence, this study takes selects an above average ESG universe and subsequently test the two different investment strategies. By doing so, it aims to find a significant difference in performance of value and growth ESG strategies compared to a ‘passive’ ESG investment. The findings of our study can be summarised as follows. Contrary to existing literature on value and growth, this study finds no significant outperformance of value strategies when compared to growth strategy. Consequently, there is also no difference in the average returns of the value and growth portfolio and the ESG portfolio. However, it does find a higher Sharpe ratio as well as a higher Sortino ratio for the growth strategy indicating a better risk-adjusted return for this strategy as opposed to a ‘passive’ ESG investment. The differences in Sharpe and Sortino ratios of the three portfolios are, however, relatively small when compared to the literature on value and growth strategies.
6
2. Background
This chapter provides a theoretical background for this study. We start with discussing the relevant theory on socially responsible investments (SRI). We continue by discussing the current theories on value and growth investing and analyse the findings of the most relevant literature on value and growth investing. Subsequently, we discuss how value and growth relates to socially responsible investment.
2.1 Socially responsible investments
Socially responsible investments (SRI) have grown rapidly over the past years, increasing globally from $13.3 billion in 2012 to $21.4 in 2014 (Global Sustainable Investment Alliance, 2014). In Europe, total SRI assets increased from $8.8 billion to $13.6 billion during that same period, which made SRI assets account for more than half of the total managed assets in this region. As the interest in SRI increases, so does the search for a balance in financial reward and sustainability, which is often referred to as ‘doing well by doing good’ (Hamilton, Jo and Statman, 1993). Through the years, the research on the relationship between financial performance and SRI has grown into a large number of studies (Revelli and Viviani, 2015). SRI can be described as applying a set of screens to the investment universe as to select or exclude stocks or assets based on ESG criteria, to ensure investors that their investments are consistent with their personal beliefs and values (Auer, 2016). Consistent with this practice, many studies that focus on the performance of SRI construct artificial portfolios based on ESG performance measures (Auer and Schuhmacher, 2015; Halbritter and Dorfleitner, 2015; Auer, 2016). ESG ratings provide insight in a company’s environmental, social, and corporate governance performance. Auer and Schuhmacher (2015) analyse the risk-adjusted financial performance of portfolios formed on ESG rankings. They analysed three different regions, Asia Pacific, U.S., and Europe, and find that for the Asia Pacific region and the U.S., ESG ratings do not have a significant impact on the financial performance of portfolios. However, in Europe they find a significant negative influence of ESG ratings on portfolio performance. On a more general basis, a meta-analysis on 190 experiments by Revelli and Viviani (2015) concludes that globally, there is no real cost or benefit to investing in SRI.
7 to growth stocks than value stocks (Junkus, 2015; Schröder, 2007; Guerard, 1997; Ferruz, Muñoz and Vargas, 2010). As value and growth stocks tend to show different performance characteristics, this suggests that an ESG strategy based on value and growth factors could lead to a difference in financial and ESG performance as compared to a conventional ESG strategy.
2.2 Value and growth investment strategies
In the current literature, value stocks are commonly defined by having low price to book2 ratios, low price earnings3 ratios (Fama and French, 1992), and low price to cash flow4 ratios (Lakonishok et al.1994). As opposed to value stocks, growth stocks are defined by having ratios in the opposite direction as they are usually defined by high price to book ratios, high price earnings ratios, and high price to cash flow ratios (Bauman et al. 1998). Bauman et al. (1998) and Fama and French (1998) find evidence for a value premium in international markets. A more recent paper on the value-growth effect in Australia by Gharghori, Stryjkowski and Veeraraghavan (2013) similarly argues that there is no question about the existence of the value-growth effect. This value and growth effect is often referred to as ‘the value premium’ and is defined as the difference in returns between the value and growth stocks (Athanassakos, 2009). The debate in the current literature, however, is on the drivers of this effect, namely risk or behavioural factors. This debate is, on one side, led by Fama and French (1992), who argue that value stocks are fundamentally riskier, and, on the other side, by the theory of Lakonishok et al. (1994), who argue that the value strategy outperforms naïve investment strategies that extrapolate earnings growth too far into the future. As Fama and French (1992) indicate that value stocks are fundamentally riskier, they argue that the price to book ratio serves as a proxy for risk. Lakonishok et al. (1994), however, find that size and price to book ratio are not significant in explaining the cross-sectional variation in stock returns when price earnings and price to cash flow ratios are included in the model. Hence, they dismiss the risk argument of Fama and French (1992). Lakonishok et al. (1994) find evidence for their behavioural argument as they find a pattern of expectational errors by investors that explains the superior returns of value stocks.
2
The price to book ratio is also found in the literature as the book to market ratio, which is the inverse of the price to book ratio. For consistency, we will refer to the price to book ratio.
3
The price earnings ratio is also found in the literature as the earnings to price ratio, which is the inverse of the price earnings ratio. For consistency, we will refer to the price earnings ratio.
4
8 To clearly grasp the effect of value and growth strategies we look at different studies performed on different time periods and regions. For a clear overview, we compare the findings of different studies, which show value premiums and significantly higher Sharpe ratios for value portfolios, as compared to growth portfolios. Table 1 shows an overview of various comparable studies on value and growth strategies.
9
Table 1. Literature overview of Value and Growth investment studies.
This table presents an overview of studies conducted on the performance of value and growth investment strategies. The author(s) of the study are in the first column. The second column displays the year that the study was published. The third column is the period over which the study was performed and the fourth is the geographic market of the study. The Ratio column shows the financial ratio that was used to sort stocks into value and growth portfolios. The value and growth columns show the annual returns, monthly returns and Sharpe ratio for each of the two portfolios. The last column shows the spread between the value and growth portfolios. Panel A shows studies which report an average annual return, whereas panel B shows monthly returns. Panel C shows Sharpe ratios from various studies for the value and growth portfolios. All returns are in percentages. The Sharpe ratio is the average rate of excess return divided by the standard deviation of return.
Authors Year Period Geography Ratio Value Growth Spread
Panel A: Annual Return
Lakonishok et al. 1994 1963 - 1990 US P/E 19.00 11.40 7.60
Fama and French 1998 1975 - 1995 International P/E 13.66 6.84 6.82
Bauman et al. 1998 1985 - 1996 International P/E 15.30 12.40 2.90
Abramson and Chung 2000 1990 - 2000 US DY and P/E 17.45
Bauman et al. 2001 1986 - 1996 Pacific RIM P/B 32.00 12.20 19.80
Anderson and Brooks 2006 1975 - 2003 UK P/E 24.26 18.28 5.98
Athanassakos 2009 1985 - 2005 Canada P/E 11.73 5.43 6.30
Asness et al. 2013 1972 - 2011 Global B/M 14.60 8.10 6.20
Asness et al. 2013 1972 - 2011 US B/M 13.20 9.50 3.70
Asness et al. 2013 1972 - 2011 Europe B/M 16.70 11.80 4.80
Hammar 2014 1992 - 2012 South Africa P/E 38.80 27.90 10.80
Panel B: Monthly Return
Fama and French 1992 1963 - 1990 US B/M 1.72 1.04 0.68
Gharghori et al. 2013 1993 - 2009 Australia P/E5 1.20 0.85 0.36
Panel C: Sharpe ratio
Bauman et al. 1998 1985 - 1996 International P/E 0.86 0.68
Abramson 2000 1990 - 2000 US DY and P/E 0.87
Bauman et al. 2001 1986 - 1996 Pacific RIM P/B 0.89 0.65
Asness et al. 2013 1972 - 2011 Global B/M 0.93 0.50
Asness et al. 2013 1972 - 2011 US B/M 0.83 0.53
Asness et al. 2013 1972 - 2011 Europe B/M 0.84 0.64
Hammar 2014 1992 - 2012 South Africa P/E 1.04 0.84
The study of Abramson and Chung (2000), reported in panel A and C, investigates whether a value-oriented portfolio can be created from SRI indices that earns returns comparable to their value-oriented benchmark strategies. They argue that the performance of U.S. SRI indices in the 1990’s might not be due to social responsibility factors but due to the fact that growth-oriented investing and SRI both favour large-cap growth stocks in the technology sector. As
10 these particular stocks have performed well during the 1990s, this might be the cause of the outperformance. The authors construct two value-oriented portfolios from the MSCI KLD 400 SOCIAL INDEX, based on the dividend yield and the price earnings ratio. The first portfolio is a buy-and-hold portfolio for the whole period, which yields an annualized return of 16.45% and a Sharpe ratio of 0.76. The second portfolio is a rebalanced portfolio, which yields an annualized return of 17.45 and a Sharpe ratio of 0.87, as reported in table 1. The benchmark indices produced an annualized return of 15.10 with a Sharpe ratio of 0.80, indicating that both value-oriented portfolios outperform their benchmark on annualized return, and the rebalanced portfolio also outperforms on the Sharpe ratio. Moreover, the percentage annual return and the Sharpe ratio reported by Abramson and Chung (2000) seem to be consistent with the findings of other studies in similar regions and time periods.
Abramson and Chung (2000) are the first to study a value-oriented strategy in SRI. As they find that this strategy can yield returns comparable to other value-oriented strategies, this confirms our suggestion that an ESG strategy based on value and growth factors could lead to a difference in financial performance as compared to a conventional ESG strategy. While many studies6 show that ESG portfolios are tilted towards growth stocks, table 1 shows a consistent outperformance of value stocks over growth stocks. Hence, increasing the exposure to value stocks in an ESG portfolio should yield a higher return as compared to a conventional ESG portfolio, as the conventional portfolio tends to be growth oriented. This leads to the following hypothesis:
H1: The return of an ESG value portfolio is significantly higher than that an ESG portfolio.
Conversely, as growth oriented portfolios show to predominantly underperform value stocks, allocating the ESG exposure to growth stocks is expected to yield a lower return. This leads to the following hypothesis:
H2: The return of an ESG Growth portfolio is significantly lower than that of an ESG portfolio.
Lakonishok et al. (1994) indicate that value stocks are underpriced and hence earn abnormal returns. They argue that this holds, as they report that the excess returns of value stocks can be explained by investor based errors. This suggest that a value strategy is not significantly
11 riskier than a growth strategy. Other studies reporting a value premium take the standard deviation of the returns as risk indicator and find, in line with Lakonishok et al. (1994), no significant difference in the standard deviation of a value strategy and a growth strategy (Bauman et al., 1998; Asness et al., 2013). Similarly, Chan and Lakonishok (2004) find in their analysis on previous research and an updated value investment strategy, evidence that value stocks are not fundamentally riskier than growth stocks. Opposing evidence is found by Fama and French (1992) and Fama and French (1998) who support the metaphysical explanation of risk, in which a higher return by definition bears more risk. Gharghori et al. (2013) find supporting evidence for this metaphysical explanation in a more recent study on the value premium in Australia. As Gharghori et al. (2013) indicate that, although there is no question about the value-growth effect, there is not yet a consensus about whether it is driven by risk or behavioral factors. Regarding the lack of consensus on the drivers of the value-growth effect, we investigate in this study whether an ESG value portfolio bears additional risk, as compared to an ESG portfolio. This leads to the following hypothesis:
H3: The standard deviation of an ESG value portfolio is not significantly higher than that of an ESG portfolio.
Conversely, as we investigate an ESG portfolio constructed on growth stock characteristics bears less risk than a portfolio constructed on ESG ratings alone. Hence, we test the following hypothesis:
H4: The standard deviation of an ESG growth portfolio is significantly lower than that of an ESG portfolio.
12 H5: The Sharpe ratio of an ESG value portfolio is higher than that of an ESG portfolio.
Conversely, as growth stocks tend to show lower Sharpe ratios7, we investigate whether increasing the exposure to growth stocks of an ESG portfolio leads to a lower Sharpe ratio. Hence, this leads to the following hypothesis:
H6: The Sharpe ratio of an ESG growth portfolio is lower than that of an ESG portfolio.
A potential downside of the Sharpe ratio, however, appears when returns deviate from the normal distribution, as it does not account for the proportion of downside risk relative to upside risk. An approach to overcome this deficiency is the Sortino ratio, as applied by Charles, Darné and Fouilloux (2016) in their research on the impact of screening strategies on the performance of ESG indices. The literature on value and growth investment strategies does not provide insight on the direction (up- or downside) of risk associated with these strategies. Hence, we add to the existing literature by investigating whether value and growth strategies applied to an ESG portfolio lead to a different Sortino ratio. This leads to the following two hypotheses:
H7: The Sortino ratio of an ESG value portfolio is different than that of an ESG portfolio.
H8: The Sortino ratio of an ESG growth portfolio is different than that of an ESG portfolio.
In their research on a value investment strategy applied to SRI indices, Abramson and Chung (2000) indicate that companies characterised as value stocks tend to be companies that may have difficulty meeting SRI criteria. Additionally, Auer and Schuhmacher (2015) report a small yet negative correlation between the book-to-market ratio and the disaggregate ESG scores. They find correlations of -0.04 (Environmental), -0.03 (Social), and -0.11 (Governance). As a high or low book-to-market ratio can be used to identify a value or growth stock, respectively, this infers that a value portfolio yields a lower average ESG score and a growth portfolio yields a higher average ESG score. This leads to the following hypothesis: H9: The mean ESG-score of an ESG value portfolio is significantly lower than that of an
ESG portfolio.
H10: The mean ESG-score of an ESG growth portfolio is significantly lower than that of an ESG portfolio.
13 Auer and Schuhmacher (2015) additionally show in their research that the average ESG ratings of portfolios, which are selected on ESG ratings, fluctuate over time. As we investigate the influence of value and growth strategies on ESG performance, this fluctuation of the average ESG rating of a portfolio can be considered as a risk associated with the ESG performance of the portfolio. When applying value and growth strategies to an ESG portfolio, the association of value and growth characteristics with ESG ratings might lead to additional fluctuations in the average ESG rating of a portfolio. Although the literature on ESG investment strategies, or value and growth strategies, does not indicate any existing theories or empirical findings on the risk of an average ESG rating, we investigate the following two exploratory hypotheses:
H11: The standard deviation of the ESG-score in an ESG value portfolio is significantly different from that of an ESG portfolio.
14
3. Methodology
This section describes the methodology this paper uses to answer the research question. We examine value and growth portfolios of individual stocks in the European equity market. To do so, first, we discuss the general approach to answering the research question, namely the portfolio analysis. Second, we will discuss the portfolio construction process and provide a graphic overview of this process. Subsequently, this section describes the various performance measures and the statistical tests this study uses in answering the research question.
3.1 Portfolio analysis
To investigate the effect of value and growth investment strategies, this study uses the portfolio analysis method in which we form portfolios based on rankings of the ESG score and value and growth variables. A portfolio analysis allows identification of any relationship between the sorting variable and the subsequent financial performance (Gharghori et al., 2013) as well as ESG performance (Auer and Schuhmacher, 2015).
We construct three main portfolios and subsequently compare their performance. The first portfolio is the ‘passive’ ESG portfolio, which represents the best-in-class stocks for each industry, based on their ESG score. We rank the stocks within each industry according to their ESG score and subsequently select the highest rated stocks to form equally weighted portfolios, as in Auer and Schuhmacher (2015). A cut-off rate indicates the top percentage that is included in the portfolio. Different studies apply different cut-off rates to select the best ESG stocks8, as there is no general rule for which cut-off rate to apply. Halbritter and Dorfleitner (2015) find no significant difference in applying cut-off rates of 1, 5, 10, 25 and 50%. As, Auer (2016) argues that as portfolio sizes become smaller, diversification benefits are lost, we apply a positive selection of the top 50% stocks based on ESG ratings within each industry. In line with Auer and Schuhmacher (2015), we select a top percentage within each industry to attain a best-in-class portfolio and not exclude industries. As ESG scores and most earnings data are updated annually, the portfolios in this study will be rebalanced on an annual basis, in line with studies on value and growth strategies of Dimson, Nagel and Quigley (2003), Anderson and Brooks (2006), Athanassakos (2009) and the study on SRI of Galema, Plantinga and Scholtens (2008).
15 Subsequently, we select the value and growth portfolio, based on the lowest and the highest 30% of price earnings (P/E) ratios, respectively, consistent with Lakonishok et al. (1994), Fama and French (1998), Bauman et al. (1998), Anderson and Brooks (2006), Athanassakos (2009), and Hammar (2014).
The selection process yields the following three portfolios:
ESG portfolio: The ESG portfolio represents the equally weighted portfolio containing the top 50% of ESG rated stocks within each industry.
Value portfolio: The value portfolio represents the equally weighted portfolio with the 30% lowest P/E ratios, derived from the ESG portfolio selection. Growth portfolio: The growth portfolio represents the equally weighted portfolio with the
30% highest P/E ratios, derived from the ESG portfolio selection. Figure 1 below shows a graphic overview of the portfolio construction process. The most left block represents all the stocks in Europe. Secondly, the figure shows the selection of the ASSET4 rated stocks. Subsequently, the figure shows that from the ASSET4 rated stocks only the stocks with a score within the top 50% of their respective industry are selected, which will form the ESG portfolio. Finally, the figure shows the selection of the 30% highest and 30% lowest P/E, P/B, and P/C ratings, resulting in the value and growth portfolios. In the main results, we will confine ourselves to the value and growth portfolios constructed based on the P/E ratio.
16
3.2 Performance measures
To test the performance of the different portfolios, we use Microsoft Excel to program a model to construct the different portfolios, calculate the returns and perform the statistical tests. To assess the difference in the financial performance of the portfolios and test the corresponding hypotheses, this study measures the financial performance by the average monthly return, the standard deviation and the Sharpe ratio, in line with Asness et al. (2013). Additionally, we use the Sortino ratio as a measure of downside risk. We measure the ESG performance of the different portfolios as the average ESG rating of the portfolio and the risk of the average ESG score as the standard deviation.
3.2.1 Financial performance
We calculate the average monthly return in excess of the risk-free rate of each portfolio, in line with Asness et al. (2013). The monthly return of all assets in a portfolio is divided by the total number of assets in the portfolio, as in equation (1).
𝐴𝑀𝑅𝑡 = ( 1 𝑁 ∑ 𝑅𝑖 𝑁 𝑅𝑖,𝑡 ) − 𝑅𝑓 (1)
Where, AMR is the arithmetic monthly return, 𝑅𝑖 is the return on asset i, and 𝑅𝑓 is the risk-free rate, which is the return on a 10-year German government bond.
The return is subsequently used to determine a spread between the mean returns of different portfolios. A t-test for equality of means is used to determine if this difference is significant. Prior to the t-test, an F-test is conducted to determine equal or unequal variance in the different portfolios and hence to decide which t-test to perform.
The standard deviation will be used as a measure of portfolio risk. It is the standard deviation of the average monthly portfolio return for each portfolio as described in equation 2. A higher standard deviation implies a riskier and more volatile strategy.
𝜎𝑅𝑖,𝑡 = √∑(𝑥 − 𝑥̅)2
17 Sharpe Ratio
The widely-used Sharpe ratio represents the average return per unit of total risk and was first used by Sharpe (1966). For each of the constructed portfolios, the Sharpe ratio (3) is calculated as a measure of risk-adjusted return.
𝑆ℎ𝑎𝑟𝑝𝑒 𝑅𝑎𝑡𝑖𝑜𝑖 = 𝑅𝑖,𝑡− 𝑅𝑓,𝑡
𝜎𝑅𝑖,𝑡 (3)
where 𝑅𝑖,𝑡 refers to the stock return of portfolio i in year t, 𝜎𝑅𝑖,𝑡 refers to the standard deviations of the equally weighted portfolio returns from equation 2, and 𝑅𝑓,𝑡 refers to the target return which in this study is equal to return on a 10 year German government bond. Sortino Ratio
The Sortino ratio (Sortino and Price, 1994) will be used to determine the downside risk of the specific portfolios and is defined in line with Charles et al. (2016) as follows:
𝑆𝑜𝑟𝑡𝑖𝑛𝑜 𝑅𝑎𝑡𝑖𝑜𝑖 =
𝑅𝑖,𝑡 − 𝑅𝑓,𝑡
𝐷𝑜𝑤𝑛𝑠𝑖𝑑𝑒 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 (𝑅𝑖,𝑡) (4)
Where 𝐷𝑜𝑤𝑛𝑠𝑖𝑑𝑒 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 (𝑅𝑖,𝑡) is measured as:
𝐷𝑜𝑤𝑛𝑠𝑖𝑑𝑒 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 (𝑅𝑖,𝑡) = √1 𝑁∑(𝑀𝑖𝑛(0, 𝑅𝑖 − 𝑅𝑓,𝑡)) 2 𝑁 𝑅𝑖 (5)
18 3.2.2 ESG performance
To measure the effect of the value and growth strategies on the ESG performance of a portfolio, we use the average annual ESG rating, calculated as in equation 6.
𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝐸𝑆𝐺 𝑠𝑐𝑜𝑟𝑒 = ∑ 𝑥𝑖 𝑛 𝑖=1
𝑛 (6)
19
4. Data
This section describes the data that we use in this study. First, we describe how we collect the data to construct the various portfolios and measure their performance. Second, we provide an overview of the description of the different variables. Third, we discuss the characteristics and the descriptive statistics of the sample.
4.1 Data collection
To measure ESG performance, this paper uses the ASSET4 Database from Thomson Reuters, as it is, as part of Thomson Reuters, one of the largest ESG rating agencies (Halbritter and Dorfleitner, 2015). From the ASSET4 Database, we download the ‘ASSET4 Europe’ constituents list with, at that moment, 1056 constituents. We take the ten year period from 2005 to 2014, as ASSET4 Europe data is limited before 2005 and after 2014. In line with Dimson et al. (2003), investment trusts are deleted, leaving a sample of 1023 stocks. The ESG-scores9 were collected from the ASSET4 Database from Thomson Reuters. To isolate the effect of the investment strategies, a balanced sample was created from the 1023 stocks, meaning that only the stocks for which an ESG rating is provided during all consecutive years are included in the sample. This leaves a total sample in this study of 653 stocks. In sorting the stocks on ESG rating within each industry, this study uses the Industry Classification Benchmark from Worldscope to distinguish between ten different main industry classifications.
The ESG ratings from ASSET4 are updated annually at the beginning of the year. Similar to other studies10, the rebalancing will take place every year at the end of June, to make sure that the ESG data and financial information are known. For the earnings data, the rebalancing uses the data of the previous fiscal year-end to ensure data would have been made available. Hence, a buy-and-hold strategy of one year is applied, running from July until June each year. We obtain the Total Return Index from DataStream for each month in the ten year period. We take monthly returns as opposed to annual returns, in line with Fama and French (1992) and Gharghori et al. (2013), to create more data points and thus assure more accurate measurement, as opposed to annual returns. The total return index was then adjusted for any fluctuations in currencies as these the returns are measured monthly. From the Total Return
9
See appendix 1 for a detailed description and methodology of the ESG ratings from Thomson Reuters.
10
20 Index, the monthly return can be calculated for each stock in the sample. This index accounts for dividends, stocks splits and repurchases. Annual price earnings ratios for all companies are obtained from DataStream.
Additionally, other commonly used indicators for value and growth strategies, namely price-to-book (P/B), price-to-cash flow (P/C), were obtained from DataStream to use for robustness of the study. Similarly, disaggregate ESG scores were obtained for the same purpose. Table 2 below gives an overview of the descriptions of the variables that this study uses.
Table 2. Variable Descriptions
This table summarizes the descriptions of the variables used in the main analysis as well as the robustness checks. The left column indicates the name of the variable while the right column outlines the description.
Variable Description
Return Returns are derived from the Total Return index from DataStream which accounts for This shows a theoretical growth in value of a shareholding over a specified period, assuming that dividends are re-invested to purchase additional units of an equity or unit trust at the closing price applicable on the ex-dividend date.
ESG score The ESG score is an equal-weighted rating of a company’s financial and extra-financial performance based on the ASSET4 economic, environmental, social and corporate governance ratings. The ratings are derived from 226 Key Performance Indicators (“KPI), which are populated by over 500 separate data points.
Environmental score The environmental score is based on a company’s impact on living and non-living natural systems. It measures to what extend a company avoids environmental risks and exploits environmental opportunities. The environmental score is based on 70 KPI’s.
Social score The social score measures how well a company can generate trust and loyalty with its workforce, its customers and with society. It is a measure of the company’s reputation and hence its license to operate. The social score is based on 88 KPI’s.
Governance score The governance score reflects how well a company’s systems and process ensure that its board members and executives act in the interest of its long-term shareholders. It reflects the company’s ability to direct and control its rights and responsibilities through the creation of incentives, as well as checks and balances to generate shareholder value.
Price to earnings (P/E) The price-to-earnings ratio (“P/E”) is the stock price per share divided by the earnings per share at the required date.
Price to book (P/B) The price-to-book ratio (“P/B”) is the stock price per share divided by the book value per share at the required date.
Price to cash flow (P/C) The price-to-cash flow ratio (“P/C”) is the stock price per share divided by the cash flow per share at the required date.
4.2 Descriptive statistics
21 columns indicate the number of companies in the sample and the percentage of the total sample for each industry classification and currency.
Table 3 shows a rather even distribution of the sample over the different countries, with low percentages per country, except for the United Kingdom. The 33.6% indicates that more than one third of our sample is made up by companies from the U.K. This might be since the U.K. knows relatively many listed companies, as compared to other European countries. As this large tilt towards U.K. firms might influence our results, we will we will perform an additional analysis that excludes the U.K. from the sample and include the results in the sensitivity analysis.
Table 3. Sample characteristics
This table shows the data sample composition. The column 'Country' indicates the different countries. The column 'Industry' indicates the different industry classifications from the Industry Classification Benchmark present in the sample. The column 'Currency’ indicates the different currencies. The column ‘Nr.’ represents the number of companies per item and the column ‘%’ shows the percentage of the total sample of the respective item.
Country Nr. % Industry Nr. % Currency Nr. %
Austria 16 1.6% Oil & Gas 58 5.7% British Pound 344 33.6%
Belgium 26 2.5% Basic Materials 76 7.4% Czech Koruna 3 0.3%
Czech Republic 4 0.4% Industrials 220 21.5% Danish Krone 27 2.6%
Denmark 26 2.5% Consumer Goods 112 10.9% Euro 429 41.9%
Finland 25 2.4% Health Care 56 5.5% Hungarian Forint 4 0.4%
France 102 10.0% Consumer Services 151 14.8% Norwegian Krone 25 2.4%
Germany 91 8.9% Telecommunications 32 3.1% Polish Zloty 30 2.9%
Greece 18 1.8% Utilities 40 3.9% Swiss Franc 68 6.6%
Hungary 4 0.4% Financials 227 22.2% Swedish Korona 60 5.9%
Ireland 15 1.5% Technology 44 4.3% Turkish Lira 26 2.5%
Italy 47 4.6% Not available 7 0.7% U.S. Dollar 7 0.7%
Netherlands 36 3.5% Norway 26 2.5% Poland 30 2.9% Portugal 10 1.0% Spain 48 4.7% Sweden 60 5.9% Switzerland 69 6.7% Turkey 26 2.5% United Kingdom 344 33.6%
Total 1023 Total 1023 Total 1023
22 line with what is observed in general market indices, the large distribution towards ‘Financials’ seems unusually large. Since we do not exclude any industries in this study, this might have an influence on our results. As such, we will test for robustness if we exclude the ‘Financials’ industry classification from our analysis and will include the results in chapter 7.
Table 4 below presents the summary statistics of the sample. The data show characteristics of non-normality. To see if this non-normality has an influence on the results, we will test for robustness of the results as we the data will be trimmed at the 5th and the 95th percentile, as this removes outliers and increases the normality characteristics of the data. The results of this will be included in the sensitivity analysis.
Table 4. Summary statistics
This table presents the descriptive statistics for the sample used in this study. The left column indicates the variable. The ‘mean’ column indicates the mean value of the total sample. The ‘median’ column indicates the median value of the total sample. The column ‘Std. Deviation’ indicates the standard deviation for each variable. The columns ‘Maximum’ and ‘Minimum’ indicate the maximum and minimum values for each variable in the sample, respectively. The last two columns provide the Skewness and Kurtosis for each variable.
Variable Mean Median Std. Deviation Maximum Minimum Skewness Kurtosis
ESG-score 59.23 70.51 32.59 98.34 0.00 -0.56 -1.15
Returns 1.09% 1.00% 11.40% 952.26% -79.30% 6.45 453.30
P/E 24.87 15.10 145.97 9,216.20 0.10 46.83 2,597.84 P/B 25.51 1.80 1,622.19 140,608.00 -320.10 77.77 6,337.47 P/C 7.58 8.74 250.13 2,540.87 -20,997.46 -68.18 5,464.90
23
Table 5. Correlation table
This table depicts the correlations between the variables that this study uses. On each axis of the table, the different variables are indicated.
ESG Environmental Social Governance P/E P/B P/C
24
5. Results 5.1 Results
This section describes the main findings of this study as well as the implications of these findings. We will discuss the findings of our analysis in the light of the existing literature and discuss similarities and divergences.
Table 6. Main results
At the end of each period (t), portfolios are formed based on the ESG score and price earnings ratios. Panel A shows the returns and standard deviations for the three portfolios. It shows the arithmetic average monthly returns that refers to the respective portfolio. The spread indicates the differences in return or standard deviation between the respective portfolios. Panel B shows the Sharpe ratios and the Sortino ratios for each portfolio. Panel C shows the mean ESG-score and the corresponding standard deviations.
1 2 3 Value Growth
Time Period Value Growth ESG Spread 1 - 2 Spread 1 - 3 Spread 2 - 3
Panel A
07/2005 - 06/2015
Geometric Mean Monthly Return 0.51% 0.58% 0.57% -0.08% -0.06% 0.01%
Geometric Mean Annual Return 6.26% 7.23% 7.08% -0.96% -0.82% 0.14%
Arithmetic Average Monthly Return 0.70% 0.69% 0.71% 0.00% -0.01% -0.02%
p-value1 (0.4980) (0.4926) (0.4892) Standard deviation 6.21% 4.67% 5.27% 1.54% 0.94% -0.60% p-value2 (0.0021) (0.0733) (0.1916) Panel B 07/2005 - 06/2015 Sharpe Ratio 0.11 0.15 0.13 Sortino Ratio 0.17 0.22 0.20 Panel C 07/2005 - 06/2015 ESG score 89.50 87.86 88.74 1.64 0.77 -0.88 p-value1 (0.2629) (0.5129) (0.5814) Std. Deviation 1.98 3.73 2.82 -1.74 -0.84 0.90 p-value2 (0.0739) (0.3067) (0.4209)
Note: 1: P-value for the t-test for equality of means, 2: P-value for the F-test for equality of variances.
25 for significance of these spreads. Panel A further presents the standard deviation for the arithmetic returns of each portfolio. Panel B shows the risk-adjusted performance of the portfolios measured by the Sharpe ratios and Sortino ratios. Finally, panel C shows the average ESG ratings of the portfolios and their spreads, as well as the standard deviations of the average ESG scores.
In panel A of table 6 we observe that the three spreads between the arithmetic average monthly returns of the portfolios are negative but close to zero. Moreover, the p-values of the spreads indicate no statistical significant difference between the portfolios. This result means that there is no difference in the average monthly returns of the value and growth portfolios as compared to the ESG portfolio. Similarly, we observe no difference in return between the value and growth portfolio. As we hypothesized in hypothesis 1 that the return of the value portfolio would be significantly higher than that of the ESG portfolio, this result provides no evidence to confirm hypotheses 1. Additionally, we hypothesized that the return of the growth portfolio would be lower than that of the ESG portfolio, for which we find no evidence. Hence, we cannot confirm hypothesis 2. To compare our findings to the spreads between the value and growth portfolios in panel A in the literature overview in table 1, we look at the spread between our value and growth portfolio. We find our results to be inconsistent with the value premiums we find in the literature for similar markets and time periods. We observe no significant difference between either the value and growth, value and ESG, or growth and ESG portfolio. This suggests that, when screening on ESG criteria, sorting stocks based on value and growth characteristics does not provide a difference in average monthly returns. This result is surprising since the research of Asness et al. (2013) shows a value premium in Europe for a period that largely covers our sample period. However, Asness et al. (2013) take the period from 1972 to 2011 whereas our study starts in 2005 and ends in 2014. Hence, this might be a reason for the insignificant difference in average returns between the value and growth portfolio. The fact that we do not find a difference in average returns might also be due to the fact that we incorporate ESG factors in our portfolio construction. This implies that a high rated ESG stock portfolio may not possess the same characteristics as regular stock portfolios that allow investors to realize a value premium.
27 this might yield different results as our sample from which we construct the value and growth portfolios has different characteristics. Moreover, not only are the Sharpe ratios lower than what we observe in table 1, also the differences between the Sharpe ratio for the value and the growth portfolio are considerably lower. This indicates that there is less difference in performance, based on the Sharpe ratio, than what we observe in panel C in table 1. As we observe significant differences in the standard deviation of the portfolios, yet not in the returns, the lower difference in Sharpe ratios is largely attributable to the lack of difference in monthly returns. We also try to overcome the deficiencies of the Sharpe ratio in our analysis, as we are primarily interested in the implications for the individual investor. Hence, we also look at the Sortino ratio as measure of the downside risk. Panel B in table 6 shows Sortino ratios of 0.17, 0.22, and 0.20 in for the value, growth and ESG portfolio respectively. This indicates that, compared to a general ESG portfolio, an ESG portfolio based on growth stocks yields a higher return for every unit of downside risk, while an ESG portfolio based on value stocks yields a lower return for every unit of downside risk. The literature on value and growth strategies does not provide Sortino ratios as results, hence we are not able to compare our results with these studies. However, Auer (2016) finds a Sortino ratio of 0.24 for a European stock portfolio which was screened based on ESG criteria. As we compare our findings for a similar European ESG portfolio, we see that the growth portfolio shows the most resemblance, based on its Sortino ratio. This indicates that when constructing an ESG portfolio based on growth stock characteristics yields similar downside risk performance as other European ESG portfolios. This is in line with the existing literature that shows that SRI funds are more exposed to growth stocks than to value stocks11. Moreover, in line with Aueer and Schuhmacher (2015) we find small but negative correlations between ESG ratings and value and growth stock indicators12. Panel C in table 6 shows, however, that the value portfolio has the highest average ESG rating over the total period of ten years. Although we observe this small difference in average ESG ratings between the portfolios, this difference is not significant. Based on these findings, we cannot confirm hypothesis 9 that hypothesizes a lower average ESG score for the value portfolio. Similarly, we do not find evidence to confirm hypothesis 10 that states that the average ESG score of the growth portfolio is lower than that of the ESG portfolio. Although the average ESG score of the portfolios are not significantly different, panel C in table 6 shows that the value strategy leads to a significantly lower standard deviation in the average ESG score of 1.74, as compared to the growth
11
Junkus, 2015; Schröder, 2007; Guerard, 1997; Ferruz et al. 2010
28 strategy. However, the differences in standard deviation are smaller and therefore insignificant when we compare the value and growth portfolio to the ESG portfolio, as panel C shows differences of -0.84 and 0.90, respectively. As these differences are insignificant, we cannot confirm hypothesis 11 that states there is a difference between the standard deviation of average ESG score of the value and the ESG portfolio, as well as hypothesis 12 that states there is a difference between the standard deviation of the average ESG score of the growth and the ESG portfolio. As these hypotheses were of exploratory nature, there is no literature that suggests a similar or contrary relationship. Nonetheless, we cannot conclude there are differences in the risk characteristics of the ESG ratings when applying a value or growth strategy to construct an ESG portfolio. As we observe no difference in the ESG performance of the three portfolios, this suggests there is no significant difference in the composition of the portfolios, with respect to ESG characteristics. As we select the top 50% of ESG rated companies, this leaves a relatively large range for the ESG performance. Our findings suggest that, although the literature finds that sustainability indices are tilted towards growth stocks, value and growth characteristics cannot be used as indicators for ESG performance.
29 Table 7. Summary of the hypotheses
In the left column are the numbers of the corresponding hypotheses. The middle column shows the hypothesis and in the last column is a ‘Yes’ indicating that the related null hypothesis is rejected and a ‘No’ if the results do not find evidence to reject the related null hypothesis. Panel A exhibits the hypotheses related to the financial performance of the portfolios. Panel B shows the hypotheses related to the ESG performance of the portfolios.
Hypothesis Expectation
Confirm Yes/No Panel A: Financial hypotheses
1 The return of an ESG value portfolio is significantly higher than that an ESG
portfolio. No
2 The return of an ESG Growth portfolio is significantly lower than that of an
ESG portfolio. No
3 The standard deviation of an ESG value portfolio is not significantly higher than
that of an ESG portfolio. Yes
4 The standard deviation of an ESG growth portfolio is significantly lower than
that of an ESG portfolio. No
5 The Sharpe ratio of an ESG value portfolio is higher than that of an ESG
portfolio. No
6 The Sharpe ratio of an ESG growth portfolio is lower than that of an ESG
portfolio. No
7 The Sortino ratio of an ESG value portfolio is different than that of an ESG
portfolio. Yes
8 The Sortino ratio of an ESG growth portfolio is different than that of an ESG
portfolio. Yes
Panel B: ESG hypotheses
9 The mean ESG-score of an ESG value portfolio is significantly lower than that
of an ESG portfolio. No
10 The mean ESG-score of an ESG growth portfolio is significantly lower than that
of an ESG portfolio. No
11 The standard deviation of the ESG-score in an ESG value portfolio is
significantly different from that of an ESG portfolio. No 12 The standard deviation of the ESG-score in an ESG growth portfolio is
30 5.2 Robustness and sensitivity
Ratios
Lakonishok et al. (1994) characterizes growth stocks according to different financial ratios and finds different results. These ratios could therefore have a different influence on the financial and ESG performance of value and growth portfolios. To test if there is a difference the value and growth portfolios are constructed based on the to-book ratio and the price-to-cash flow ratio. The results of these portfolios can be found in appendix 2 and 3.
We first look at the price to book ratio in table A. Although the differences in return are still insignificant, the standard deviation of the average returns of the growth portfolio is with -1.05% significantly lower than that of the ESG portfolio. As this is in line with the expectations based on the current literature, the price to book ratio may be a better indicator of value and growth stocks. The Sharpe and Sortino ratio are robust, with the growth portfolio still showing the highest ratios. As we observe no significant differences in ESG performance, our results are robust.
Table B in appendix two shows the results when sorting the value and growth stocks based on the price to cash flow ratio. Although the differences in average monthly returns are larger than in our initial result, they are still insignificant. Consistent with the results from the price to book ratio, all three differences in standard deviations are significant, indicating a higher standard deviation for the value portfolio and a lower standard deviation for the growth portfolio, as compared to the ESG portfolio. Regarding ESG performance, the results in table 2 show only a significant difference between the standard deviation of the value and the growth portfolio. As the other differences are insignificant, this is consistent with our earlier results.
Disaggregate scores
31 corporate governance rating, as provided by ASSET4. The results are presented in tables C, D and E in appendix 4.
Table C shows the results from applying the environmental score in our analysis. The results show no significant difference between the standard deviations of the value portfolio and the ESG portfolio. This deviates from our initial results, while the rest of the results is in line. Table D shows the results from applying the social score in our analysis. The results are consistent with our initial results, except for the difference in standard deviations between the value and growth portfolio, which, in this result, is not significant. Table E shows the results from applying the corporate governance score in our analysis. The results are also consistent with our initial results, except for the insignificant difference between the standard deviations of the value and growth portfolio.
ESG Cut-off rates
There is no general rule in academic research on the cut-off rate to apply to construct portfolios based on high ESG ratings. As the diversification of a portfolio is influenced by the applied cut-off rate, we test whether a different cut-off rate yield different results in our analysis. As such, we apply a cut-off rate of 25% and check whether this has a significant influence on our results. The results of our analysis are presented in table F in appendix 5. The only difference with our initial results we observe, is a significantly lower standard deviation of the average ESG score of the value portfolio, as compared to the ESG portfolio. This means, that when selecting a smaller range of only the best 25% of ESG rated companies within each industry, a value strategy yields a less volatile ESG score.
Ratio cut-off rate
32 Trimming
In table 4 we indicate that our variables show characteristics of non-normality. To see if this has an influence on our results, we trim the variables on the 5th and 95th percentile which yields a distribution closer to that of normal distribution. The returns, as well as the P/E ratio show non-normality, which is why we apply the trimming to both variables. The results of this analysis are presented in table H in appendix 7. Compared to our initial results, we observe less significant differences as only the difference in standard deviations between the value and growth portfolio is significant.
Sample characteristics
33
6. Conclusion and discussion 6.1 Conclusion
We investigate in this paper whether incorporating alternative investment strategies into ESG investing yields different results. In doing so, we include value and growth investment strategies into an ESG strategy and study the effects on the financial and ESG performance of the portfolios. This way, we try to answer the research question of this paper, which is: ‘Does investing in value stocks or growths improve the financial and ESG performance of an above average ESG portfolio?’. To investigate the effects of value and growth strategies on the financial and ESG performance of a portfolio of high rated ESG stocks, we apply a portfolio analysis, in which we construct three portfolios. The first portfolio is the ESG portfolio which consists of the companies with 50% highest ESG ratings within each industry. The second portfolio we create from the ESG portfolio as we select the companies within the 30% lowest price earnings ratios to form a value portfolio. The third portfolio we create also from the ESG portfolio as we select the companies within the 30% highest price earnings ratios to form a growth portfolio. We collect ESG data from DataStream for the ASSET4 Europe constituents list. Similarly, we collect annual price earnings ratios and monthly returns from DataStream. By investigating the average monthly returns, standard deviations, Sharpe ratios, Sortino ratios, average ESG scores, and standard deviations of the average ESG scores, we compare the portfolio performance of the three portfolios. We find, in line with the theory of Fama and French (1992), that a portfolio of high ESG value stocks is significantly riskier than a portfolio of ESG stocks, although this result is not robust for all robustness checks. Additionally, we find, although the differences are small, that a growth strategy yields the highest return for every unit of risk as well as downside risk. We find no significant differences in the ESG performance of the value, growth an ESG portfolio. These results infer that when investing in high rated ESG stocks, an investor is best off investing in stocks that show characteristics of growth stocks. However, as we find no significant difference in returns between the portfolios and the difference in standard deviations appears not to be robust, we cannot conclude there is a significant difference in financial performance and ESG performance when value and growth investment strategies into an ESG portfolio.
6.2 Discussion
34 our time period and the incorporation of ESG factors in our portfolio construction. Lakonishok et al. (1994) indicate that the abnormal return from value stocks can be explained by behavioural errors from investors. Due to these behavioural errors, growth stocks are overpriced and value stocks are underpriced, which yields the abnormal returns of value stocks. As this study focusses on a recent period in which there is more transparency and information is more quickly and freely available, this might reduce behavioral errors by investors and hence reduce the value and growth effect. We also indicate that the incorporation of ESG ratings in our portfolio construction might lead to different results as this is the most significant difference in our analysis, compared to conventional value and growth studies. Companies with high ESG ratings are known to have a high transparency about their business activities and performance. This high transparency could also lead to a reduction in behavioral errors by investors. High transparency about a company’s information may lead to less underpricing and overpricing as more information is known which leads to more accurate expectations. Hence, a higher transparency for the companies in our sample could reduce the value and growth effect. Furthermore, our robustness checks indicate that the significant positive difference between the standard deviations of the value and the ESG portfolio is not consistent across different methods and metrics. Our main result indicates a significant difference at the 10% level. As we test twelve hypotheses this increases the chance of a type 1 error by which we falsely reject the null hypothesis and confirm a significant difference between the standard deviations of the value and the ESG portfolio. This would suggest no significant difference in risk and return between the value and growth portfolios and the ESG portfolio and would suggest that there is no effect of either strategy on the performance of an ESG portfolio. As we suggest that transparency about company information could be a cause for this lack of difference, further research could indicate the significance of the influence of transparency on the value and growth effect.
35 Sharpe ratio, indicating a worse financial performance of our value strategy. As their study focusses on the U.S. and the 1990’s, and our study takes Europe between 2005 and 2014, this could be a cause of the different result. However, studies on the value and growth effect in Europe for similar time periods to ours, find Sharpe ratios for value portfolios comparable to what Abramson and Chung find. Hence, the difference in our results may be caused the fact that we construct our portfolios based on best-in-class ESG rated stocks from ASSET4 while they use constituents of the MSCI KLD 400 SOCIAL INDEX. This MSCI index excludes so called ‘sin’ stocks13
and targets high ESG large and mid-cap companies. The difference between the MSCI and our sample is most observable in the fact that more than 25% of the MSCI constituents is made up by the information technology sector. As the study of Abramson and Chung (2000) was performed during the boom of technology firms during the 1990s, their results might be influenced by the time and the market. Further research should investigate the influence of the diversity of the sample on value and growth strategies as well as the influence of different ESG metrics or sustainability indices.
13
36
7. References
Abramson, L., Chung, D., 2000. Socially responsible investing: viable for value investors? Journal of Investing 9, 73 – 80.
Anderson, K., Brooks, C., 2006. The long-term price-earnings ratio. Journal of Business Finance and Accounting 33, 1063 – 1086.
Asness, C., Moskowitz, T., Pedersen, L., 2013. Value and momentum everywhere. Journal of Finance 68, 929 – 985.
Athanassakos, G., 2009. Value versus growth stock returns and the value premium: the Canadian Experience 1985-2005. Canadian Journal of Administrative Sciences 26, 109 – 121. Auer, B., 2016. Do socially responsible investment policies add or destroy European stock portfolio value? Journal of Business Ethics 135, 381 – 397.
Auer, B., Schuhmacher, F., 2015. Do socially (ir)responsible investments pay? New evidence from international ESG data. Quarterly Review of Economics and Finance 59, 51 – 62.
Bauman, W., Conover, C., Miller, R., 1998. Growth versus value and large-cap versus small-cap stocks in international markets. Financial Analysts Journal 54, 75 – 89.
Bauman, W., Conover, C., Miller, R., 2001. The performance of growth stocks and value stocks in the Pacific Basin. Review of Pacific Basin Financial Markets and Policies 4, 95 – 108.
Chan, L., Lakonishok, J., 2004. Value and growth investing: review and update. Financial Analysts Journal 60, 71 – 86.
Charles, A., Darné, O., Fouilloux, J., 2016. The impact of screening strategies on the performance of ESG indices. Unpublished working paper. Université de Nantes, France.
Derwall, J., Geunster, N., Bauer, R., Koedijk, K., 2005. The eco-efficiency premium puzzle. Financial Analysts Journal 61, 51 – 63.
37 Fama, E., French, R., 1992. The cross‐section of expected stock returns. Journal of Finance 47, 427 – 465.
Fama, E., French, R., 1998. Value versus growth: the international evidence. Journal of Finance 53, 1975 – 1999.
Ferruz, L., Muñoz, F., Vargas, M., 2010. Stock picking, market timing and style differences between socially responsible and conventional pension funds: evidence from the United Kingdom. Business Ethics: A European Review 19, 408 – 422.
Galema, R., Plantinga, A., Scholtens, B., 2008. The stocks at stake: return and risk in socially responsible investment. Journal of Banking and Finance 32, 2646 – 2654.
Gharghori, P., Stryjkowski, S., Veeraraghavan, M., 2013. Value versus growth: Australian evidence. Accounting and Finance 53, 393 – 417.
Global Sustainable Investment Alliance, 2014. Global sustainable investment review, report. [Online] Available at: http://www.ussif.org/Files/Publications/GSIA_Review.pdf [Accessed 30 September 2016].
Guerard, J., 1997. Additional evidence on the cost of being socially responsible in investing. Journal of Investing 6, 31 – 36.
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, 62 – 66. Hammar, S., (2014). Value and small firm premiums in the South African market. South African Journal of Business Management 45, 71 – 91.
Junkus, J., Berry, T., 2015. Socially responsible investing: A review of the critical issues. Managerial Finance 41, 1176 – 1201.
38 Mollet, J., 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.
Revelli, C., Viviani, J., 2015. Financial performance of socially responsible investing (SRI): what have we learned? A meta‐analysis. Business Ethics: A European Review 24, 158 – 185.
Schröder, M., 2007. Is there a difference? The performance characteristics of SRI equity indices. Journal of Business Finance and Accounting 34, 331 – 348.
Sharpe, W., 1966. Mutual fund performance. Journal of Business 39, 119 – 138.
Sortino, F., Price, L., 1994. Performance measurement in a downside risk framework. Journal of Investing 3, 59 – 64.
39
8. Appendices
Appendix 1: ESG ratings
From: The Thomson Reuters Corporate Responsibility Ratings (TRCRR): Rating and Ranking Rules and Methodologies
I. General Description
The Thomson Reuters Corporate Responsibility Ratings (“TRCRR”) measure the
environmental, social, governance and composite Environmental/Social/Governance (“ESG”) performance of over 4,600 companies worldwide.
The TRCRR are based on data provided by ASSET4, a leading global provider of ESG data. Ratings are derived by company comparisons for a total of 226 Key Performance Indicators (“KPI”). The 226 KPIs are derived from over 500 separate data points to facilitate accurate and transparent ESG screening.
Environmental Ratings are derived from a total of 70 KPIs; Social Ratings are derived from a total of 88 KPIs; and Governance Ratings are derived from a total of 68 KPIs.
Ratings are designed to offer a “best in breed” measure. Accordingly, ethical exclusions are not part of the process.
II. The Ratings Committee
The TRCRR Ratings Committee (“The Committee”) is comprised of not less than three members. The Committee Chairman will have extensive experience with and expertise in responsible and/or ethical investing. The other members will have experience in corporate investor relations, financial markets, corporate governance, the environment and/or corporate ethics. The number of committee members may be expanded from time to time.
The Committee will be responsible for 1) overseeing the role of the calculation agent
(Thomson Reuters); 2) overseeing the production of the ratings pursuant to the rules contained in this document; and 3) voting on changes to the rules and/or methodologies defined in this document.
40 III. Data Sources
The ASSET4 unit of Thomson Reuters provides the data used to calculate the ratings. This unit employs over 120 analysts who compile ESG data on 4,600 companies worldwide. Data is collected on over 500 separate data points from multiple sources, including a)
company reports, b) company filings, c) company websites, d) NGO websites, e) CSR Reports and f) established and reputable media outlets.
These 500+ data points roll up into 226 KPIs. The KPI values form the basis for the rating process.
The KPI fall into three pillars:
1. The Environmental Pillar. Examines factors including resource usage and reduction; emissions and emissions reductions; environmental activism and initiative and product or process innovation.
2. The Social Pillar. Examines factors including employment quality, health and safety issues, training, diversity, human rights, community involvement and product responsibility.
3. The Corporate Governance Pillar. Examines factors including board structure,
compensation policy, board functions, financial and operational transparency, shareholder rights and vision and strategy.
IV. Timing of Ratings Updates
Ratings are updated monthly on a dynamic basis. Ratings updates occur on the last business day of each month.
