Lars G. Hassel The Economic Value of ESG/SRI EAPSPI Conference on Sustainable Investments October 24, 2014 Hotel Kristina Outline for an Academic Approach • • • • Sustainable and Responsible Investment strategies SRI value creation in financial markets ESG metrics in research papers Errors in investors’ expectations and learning – The alpha puzzle and market efficiency – Learning and Disappearing Association between Governance and Returns (Bebchuk et al., JFE 2013) – Stakeholder Relations and Stock Returns: On Errors in Investors’ Expectations and Learning (Borgers et al., JEF 2013) • Evidence from academic studies on portfolio analysis, event studies around earnings announcements, analysis of errors in analysts’ forecasts, and firm value studies • Conclusions with implications Sustainable and Responsible Investment Strategies European SRI Study 2014 SRI in Europe - Eurosif Key Features of the European Market 35 CHARACTERISTICS OF INVESTORS ASSET ALLOCATION The impressive growth of SRI in Europe can, just as in previous years’ reports, be a ributed to institutional investors. The overall split between retail assets and institutional assets has tilted further in favour of the la er, and now stands at 96.6% institutional. Part of this growth can be a ributed to the addition of Finland, Norway and Sweden to this data for 2013, but even correcting for this shows minimal growth for retail assets on aggregate. Figure 13 provides the evolution of the European market breakdown of retail vs. institutional since 2009, while Figure 14 illustrates the differences across the European countries. Figure 15 details SRI asset allocations by country and provides an asset-weighted European average. At European level, equities represented about 50% of the SRI assets in December 2013. This compares with 33% for both 2011and 2009, thus representing a signiﬁ cant increase. By contrast, the allocation to bonds fell from 53% and 51% in 2009 and 2011respectively to 40.0% in 2013. The allocation to money markets also f ell sharply from 6.9% in 2011to 1% last y ear. FIGURE 13: Institutional vs. Retail Europe 2009 – 2013 (%) This year, for the ﬁ rst time, Eurosif broke down the allocation to bonds. In 2013, 21.3% of SRI bond assets were invest ed in corporate bonds, 16.6% in sovereign bonds and 1.4% in supra-national bonds (Figure 16). Allocation to so-called alternative assets remained relatively ﬂ at between 2009 (9.3%) and 2013 (9.1%), but fell compared to 2009 (12%). However, real estate and commodities recorded signiﬁ cant growth between 2011 and 2013, while the allocation to hedge funds decreased. FIGURE 15: SRI Asset Allocation by Country (2013) Source : Eurosif 31 SRI – by Country ESG Value Creation in Financial Markets • When benefits to ESG materialize, SRI investor enjoys positive earnings “surprise”? • ESG at firm level associated with productivity, efficiency, and hence, cash flow Alpha ESG drives future cash flows Errors in Impact on consensus company expectations valuation • The market overlooks the benefits of ESG? • How does market price ESG ? Extrafinancial value ESG Metrics - dimensions Thomson Reuters (ASSET4) MSCI ESG Research (KLD Research and Analytics) Global Engagement Services Semenova and Hassel: On the Validity of Environmental Performance Metrics, JBE 2014 ESG and Stock Return – Alpha Puzzle Institutions often justify their responsible practices using the argument that ESG factors generate positive investment returns (alpha). Investors (and companies) that exploit this market inefficiency will benefit from an early mover advantage that can last decades before risk-return equilibrium is established (2012) ESG and Stock Return – Alpha Puzzle Abnormal return! Derwall et al (FAJ, 2005): The high-ranked portfolio based on Innovest ecoefficiency ratings provided substantially higher average returns than its lowranked counterpart over the 1995-2003 period. ESG and Stock Return – Positive Alpha Factors: market, small-large, valuegrowth, momentum Abnormal Return - Mispricing? Stock market anomalies • Classical EMT suggests that trading strategy based on public information cannot produce abnormal returns (Fama, 1970). • Adaptive EMT suggests that the ability of trading strategy to generate abnormal returns will disappear over time (Daniel and Titman, 1999). • Errors in expectations and Learning – evidence from SRI studies (Bebchuck and Borgers studies, 2013) Errors in Investors’ Expectations and Learning Intangible nature of ESG Mispriced information eventually disappears when investors learn about the anomaly Increased integration of ESG factors Media and research attention More shareholder proposals on ESG issues Positive risk-adjusted returns on portfolios of high-ranked ESG stocks will disappear ESG earnings surprises will disappear Positive association with firm value (Tobin’s Q) will persist Governance and stock returns • Learning and disappearing association between governance and returns by Bebchuk, Cohen, Wang (JFE, 2013) • Investor Responsibility Research Center (IRRC) Corporate Governance data on S&P 500 – Democracy firms (Good G) and Dictatorship forms (Bad G) based on • G-Index (24 IRRC provisions; such as voting rights, CEO protection, tactics for delaying hostile bidders, other takeover defences) • E-Index (limited to 6 IRRC provisions) – High G-Index and E-Index (protection) = poor Governance – Periods: 1990-2008; 1990-1999 and 2000-2008 Good Governance Effect • The disappearance of the governance-return correlation was associated with an increase in the attention to governance by a wide range of market participants; • Until the beginning of the 2000s, stock market reactions to earning announcements reflected the market’s being more positively surprised by the earning announcements of good-governance firms than by those of poor-governance firms; • Stock analysts were also more positively surprised by the earning announcements of good-governance firms than by those of poorgovernance firms until the beginning of the 2000s but not afterwards; • While the G-Index and E-Index could no longer generate abnormal returns in the 2000s, their negative association with Tobin’s Q and operating performance persisted (good-governance firms higher market value). Attention paid to G by media and institutional investors increased from 2000 Structural break in abnormal G returns occurred in November 2000 Critical learning point occurred in October 2001 Governance indices and abnormal stock returns Abnormal returns fade out over time Governance indices and operating performance and Tobin’s Q Association between governance indices and operating performance and Q persist during both 1990-2001 and 2002-2008 Environmental and Social information and stock returns • Stakeholder relations and stock returns: On errors in investors’ expectations and learning (Borgers et al, JEF 2013) • Stakeholder-relations index, SI – (KLD ES ratings: emissions and pollution prevention, environmental management systems, community involvement, diversity, employee relations, product quality, human rights) • High SI = high ES ratings • U.S large cap 1000-3000 firms (1992-2009; 19922004 and 2004-2009) ES mispricing is eliminated • Provide evidence that the Stakeholder-relations Index (SI) explained errors in investors' expectations about firms' future earnings 1992–2004. – SI was positively associated with long-term risk-adjusted returns, – Earnings announcement returns, and – Errors in analysts' earnings forecasts over the period • When attention for stakeholder issues became more widespread, these relationships diminished considerably. • The results are consistent with the idea that increased investor attention for stakeholder issues eliminates mispricing. Number of shareholder proposals CSR in the news Number of occurrences for CSR in the newspapers, including the Wall Street Journal, the Financial Times (Capelle-Blancard and Petit, 2011) Learning effect: time variation in abnormal returns top-minus-bottom ES ranked portfolios Critical learning point occurred in April 2004 ES and risk-adjusted returns ES and earnings announcement returns Subsample 2: April 2004 – December 2009 1992-2004: higher abnormal returns around earnings announcements 2004-2009: earnings surprises disappear Errors in analysts’ forecasts Error: actual EPS minus consensus analyst longterms forecast from I/B/E/S Positive Errors 1992-2004 Zero Errors 2004-2009 ES and Operating Income Performance of shunned-stocks and high employee relations portfolios 1992-2008 (Derwall et al., 2011) Shunned stocks: KLD Controversials (tobacco, alcohol, gaming, nuclear, firearms) Values-driven v. profit-seeking SRI Performance of controversial stocks Study Hong and Kacperzyk (2009) Kempf and Osthoff (2007) Statman and Glushkov (2009) Salaber (2007) Fabozzi et al. (2009) Visaltanachoti et al. (2009) Region and Period U.S. 1926-2006 U.S. 1991-2004 U.S. 1992-2007 Europe 1975-2006 21 countries 1970-2007 China 1975-2006 Tobac. Alc. Game Weap. X X X X X X X X X X X X X X X X X X X X X X X Controversial stocks outperform! Nuke Biotech Adult Alpha Positive Positive (non-significant) Positive (non-significant) Positive X X X X Positive Positive Conclusions and Implications • Mispriced ESG information has generated in the past superior risk-adjusted returns in the short run • Increased attention for ESG issues has eliminated errors in investors expectations in the long run (learning effect) – Correlation between G and abnormal returns disappears from 2000 – Correlation between ES and abnormal returns disappears from 2004 • Public ESG information used by researchers is not able to provide a basis for profitable trading strategy. • Positive association with operating performance and market value persists in the long run • ESG becomes more significant in the market and to be taken into account by investors Doing well while doing good!