What is Sustainable Investing?

Sustainable investing, also known as ESG investing or socially responsible investing, is an approach that asks investors to consider a company’s ESG profile alongside its financials when making an investment decision. Such additional factors include everything from a company’s energy use, waste and pollution, to its working conditions, participation in its community and diversity in its board of directors. Because of these considerations, it is not unusual for sustainability-minded investors to set maximum thresholds or even shy away altogether from less “ethical” sectors such as coal, defence, gaming or tobacco.

“There’s a lot of ESG data out there on firms and these can both be overwhelming and perplexing.”

Prof. Cheng Si

Differences in ESG Rating Methodologies and Rating Systems May Cause Investor Confusion

Perhaps due to its relatively recent arrival in the finance world – the term ESG investing itself was first coined by the U.N. Global Compact as part of a landmark 2004 study titled Who Cares Wins, there is no universal standard nor a commonly accepted methodology for calculating ESG ratings among different agencies. According to KPMG, there are around 30 major ESG data providers worldwide in 2020. These ESG rating agencies usually adopt different methodology and measurements when constructing their ESG scores. It is not uncommon for them to provide different ESG ratings for the same company. For example, Tesla Inc. is rated average by MSCI ESG ratings but categorised as high risk by Sustainalytics.

The new study Sustainable Investing with ESG Rating Uncertainty was co-conducted by Cheng Si, Assistant Professor in the Department of Finance at The Chinese University of Hong Kong (CUHK) Business School, Prof. Doron Avramov at IDC Herzliya, Prof. Abraham Lioui at EDHEC Business School and Prof. Andrea Tarelli at the Catholic University of Milan.

In their study, Prof. Cheng and her co-authors tested their hypothesis using U.S. stocks from 2002 to 2019 and examined the ratings from six major ESG rating providers – Asset4 (Refinitiv), MSCI KLD, MSCI IVA, Bloomberg, Sustainalytics and RobecoSAM. In line with existing studies on ESG ratings, the research team also found considerable disparity across different ESG rating providers. They found that the confusion in the different ratings provided by the ESG rating agencies made sustainable investing riskier and decreased investor demand for stocks.

ESG Rating Disagreement and Problems That May Rise

“Generally, because there’s a lack of consensus in reporting, measuring and interpreting ESG information, there’s a lot of ESG data out there on firms and these can both be overwhelming and perplexing. That’s why it can be difficult for investors to ferret out the ‘true colour’ of a firm, whether that be, green, brown, or something in between,” Prof. Cheng says. “That in turn feeds back into investor appetite in sustainable investment. If an investor is looking for ESG plays and they’re not clear about the sustainability of the stock they’re about to sink money into, then they obviously are going to think twice before proceeding.”

Using data from the six ESG rating providers, the researchers generated an ESG score for each stock, as well as a score to measure the difference in the ESG scores between the six agencies in order to calculate the level of uncertainty in ESG ratings. According to the results, the average rating correlation is only 0.48, and the average ESG rating uncertainty is 0.18. For perspective, this means that a company could be ranked in the bottom third by one data provider but be ranked in the 59th percentile by another.

Using these scores, the researchers looked at how inconsistency in ESG ratings affected whether an institutional owner would invest in a particular stock, and the impact on the stock’s actual performance on the market. The study considered three distinct types of investors. The first type is organisations such as pension funds as well as university and foundation endowments, which constrain their investments to socially acceptable norms (such as by engaging in socially responsible investing) when compared with other institutional investors which are more interested in generating financial returns, such as hedge funds.

The study found that institutions that were more constrained by investment norms were indeed in favour of greener firms, but were less likely to hold green stocks when there is a high level of inconsistency over ESG ratings. For companies with the highest ESG scores, norm-constrained institutions on average hold 22.8 percent of their shares, but only when the ratings put out by the different ESG agencies were in high agreement. When the correlation in ESG ratings between the different ESG ratings agencies was low, the institutional ownership level dropped to 18.1 percent.

In contrast, hedge funds invest more in brown stocks on average, and rating uncertainty mostly affects their holdings for brown stocks. For companies with the lowest ESG scores in the study, the researchers found that hedge funds owned an average of 15.7 percent of shares when there was high agreement between the ESG scores from different ratings agencies. This again dropped to 13 percent when the correlation in the ratings from different agencies diverged. The authors conclude that rating uncertainty matters the most for investors in their preferred investment universe.

And while companies that focus on improving their ESG performance are expected to deliver lower investment returns because they provide nonpecuniary benefits to investors, the study found that this was not always the case. Specifically, it found that brown stocks always outperform green stocks only when ESG ratings ambivalence is low. When there is a high level of agreement between the ratings of different ESG rating agencies, brown stocks surpass green stocks by 0.59 percent per month in absolute returns and 0.40 percent per month in risk-adjusted returns. But when inconsistency between ESG ratings rises, there is no clear relationship between a company’s ESG leanings and their stock performance.

Market Implications

Lastly, the study implies that ambiguity in ESG ratings has an overall impact on the entire stock market. In particular, a higher level of rating confusion is linked with higher market premium, as well as lower stock market participation and lower economic welfare for ESG-sensitive investors.

Green stocks are harmed the most when ESG rating confusion is high. Firms which take a more responsible path in their operations are disproportionately penalised if rating agencies disagree with their ESG profile. This in turn would further limit their ability to make capital investment and generate a real social impact.

“In the face of uncertainty over a company’s ESG profile, ESG-sensitive investors are just as likely to stop making ESG investments or engage in corporate ESG matters,” Prof. Cheng adds.

Overall, the study results have significant implications for asset allocation, investor welfare, and asset pricing. In order to minimise the problems brought to ESG investing by rating inconsistency, Prof. Cheng and her co-authors suggest companies disclose more candid reports on their ESG performance. For ESG rating providers, the researchers advise them to further release and explain their measurement practices and methodologies. Furthermore, they argue that more public discussion on how to measure the ESG performance of companies should help to elevate the quality of ESG ratings.

“Sustainable investing is on the rise. Therefore, the overall impact of ESG rating inconsistency will become even more prominent,” Prof. Cheng says.

Morningstar’s Ratings

Since 1985, Morningstar has offered investors a free backward-looking Star Rating service, ranked from 1 to 5 – based on mathematically derived and adjusted past-performance indicators compared with other funds in the same category.

However, its limitations in predicting future returns led to the 2011 introduction of the company’s Analyst Rating, generated from forward-looking analysis of funds on a rising five-tier scale: Negative, Neutral, and three positive ratings, i.e., Bronze, Silver and Gold.

A top-three rating shows analysts think highly of a fund. The differences between them correspond to the level of analyst conviction in a fund’s ability to outperform its benchmark and peers over time, despite its risks.

Analysts covering funds examine and rank them based on five important pillars – people, process, parent, performance and price – to predict its success in different market environments and highlight key developments in performance and portfolio holdings. In arriving at an analyst rating, they also produce an analyst report through interviewing key parent company executives, risk managers and traders.

Yet as the company’s analyst coverage is limited by the size of its team, it also developed a machine-learning model in 2017 to create its quantitative rating, which is analogous to the rating an analyst might assign to a fund if it were covered. This model also assesses the funds based on the five key pillars. Investors pay US$199 a year to use the two predictive ratings.

The study used monthly analyst ratings, quantitative ratings, and star ratings found on the Morningstar mutual fund database and manually downloaded analyst reports from Morningstar’s website. A final sample featured 3,256 actively managed U.S. equity funds, including 1,056 funds that have been covered by Morningstar analysts at least once.

Human or AI – Which Rating is Better?

Prof. Cheng says the study highlights the importance of analyst reports in providing unique and additional information and insights for retail investors. She also notes that when an analyst adopts a positive tone, it can improve a fund’s annual return.

The study shows analyst reports are even more informative at predicting a fund’s future returns when the tone is at odds with the analyst rating, she says. For example, Gold-rated funds with a more negative tone display a lower future performance, while Negative-rated funds with a more positive tone tend to rebound.

The study also investigates the reaction of mutual fund investors to Morningstar ratings. “We find that investors do not react to analyst ratings, but instead rely on backward-looking past performance and star ratings as well as quantitative ratings,” Prof. Cheng says.

Although analysts’ recommendations are largely ignored by retail mutual fund investors, institutional investors do take advantage of the valuable information provided by the analyst rating and report, for example by withdrawing from Gold-rated funds that have received an assessment with a more negative tone from analysts.

The study also analyses the summary section and the title of analyst reports instead of the full report and finds that only the tone in the full analyst report predicts returns that exceed those of similar funds. This suggests investors need to carefully read the whole report to obtain useful information.

However, investors tend to react strongly to the tone in the summary section and the title, but not in the full analyst report, Prof. Cheng says. This suggests mutual fund investors are not sophisticated in considering the information before them and making investment decisions and are likely to be influenced by the information that attracts their attention.

She believes the study is the first to reveal the informational value of the analyst rating and analyst reports, and to highlight the importance of soft information, expressed as ideas and opinions, in mutual fund investment. The study’s findings suggest mutual fund analysts play an important role in acquiring and processing information as well as facilitating more efficient capital allocation across mutual funds.

“In future, an improved information environment could reduce the search cost in the mutual fund industry and, as a result, lead to a more efficient asset management market and financial market,” Prof. Cheng says.

The study also shows that the analyst rating is easy to access and follow in real time, so it should be easy for investors who rely on the star rating to switch to the analyst rating and improve their performance, she says.

Over-reliance on Fintech?

The findings also touch on the increasing adoption of financial technology (fintech) in the financial industry through the use of statistical methods and machine-learning techniques, such as in credit rating, financial advising and asset management.

Fintech can greatly reduce information production costs and enhance financial inclusion, but the study highlights one of the drawbacks, she says. The quantitative rating cannot be considered a like-for-like substitute for an analyst rating because of the selection of analysts’ coverage and the information value of analyst reports.

Cheng says the research also has implications for investor education and financial service provision. While individual investors can outsource their day-to-day portfolio management decisions to professional fund managers, the growing market size and variety of financial products mean that fund selection can be complicated.

The study’s findings show there is a need to offer continuous financial education to individual investors and inform them of the up-to-date, valuable financial services and tools, she says.

It’s been called the holy grail of finance. Is it possible to harness the promise of artificial intelligence to make money trading stocks? Many have tried with varying degrees of success. For example, BlackRock, the world’s largest money manager, has said its AI algorithms have consistently beaten portfolios managed by human stock pickers. However, a recent research study by The Chinese University of Hong Kong (CUHK) reveals that the effectiveness of machine learning methods may require a second look.

The study, titled “Machine Learning versus Economic Restrictions: Evidence from Stock Return Predictability”, analysed a large sample of U.S. stocks between 1987 and 2017. Using three well-established deep-learning methods, researchers were able to generate a monthly value-weighted risk-adjusted return of as much as 0.75 percent to 1.88 percent, reflecting the success of machine learning in generating a superior payoff. However, the researchers found that this performance would attentuate if the machine learning algorithms were limited to working with stocks that were relatively easy and cheap to trade.

“We find that the return predictability of deep learning methods weakens considerably in the presence of standard economic restrictions in empirical finance, such as excluding microcaps or distressed firms,” says Si Cheng, Assistant Professor at CUHK Business School’s Department of Finance and one of the study’s authors.

Disappearing Returns

Prof. Cheng, along with her collaborators Prof. Doron Avramov at IDC Herzliya and Lior Metzker, a research student at Hebrew University of Jerusalem, found the portfolio payoff declined by 62 percent when excluding microcaps – stocks which can be difficult to trade because of their small market capitalisations, 68% lower when excluding non-rated firms – stocks which do not receive Standard & Poor’s long-term issuer credit rating, and 80 percent lower excluding distressed firms around credit rating downgrades.

According to the study, machine learning-based trading strategies are more profitable during periods when arbitrage becomes more difficult, such as when there is high investor sentiment, high market volatility, and low market liquidity.

“Machine learning methods require high turnover and taking extreme stock positions. An average investor would struggle to achieve alpha after taking transaction costs into account.”

Prof. Si Cheng

One caveat of the machine-learning based strategies highlighted by the study is high transaction costs. “Machine learning methods require high turnover and taking extreme stock positions. An average investor would struggle to achieve alpha after taking transaction costs into account,” she says, adding, however, that this finding did not imply that machine learning-based strategies are unprofitable for all traders.

“Instead, we show that machine learning methods studied here would struggle to achieve statistically and economically meaningful risk-adjusted performance in the presence of reasonable transaction costs. Investors thus should adjust their expectations of the potential net-of-fee performance,” says Prof. Cheng.

The Future of Machine Learning

“However, our findings should not be taken as evidence against applying machine learning techniques in quantitative investing,” Prof. Cheng explains. “On the contrary, machine learning-based trading strategies hold considerable promise for asset management.” For instance, they have the capability to process and combine multiple weak stock trading signals into meaningful information that could form the basis for a coherent trading strategy.

Machine learning-based strategies display less downside risk and continue to generate positive payoff during crisis periods. The study found that during several major market downturns, such as the 1987 market crash, the Russian default, the burst of the tech bubble, and the recent financial crisis, the best machine-learning investment method generated a monthly value-weighted return of 3.56 percent, excluding microcaps, while the market return came in at a negative 6.91 percent during the same period.

Prof. Cheng says that the profitability of trading strategies based on identifying individual stock market anomalies – stocks whose behaviour run counter to conventional capital market pricing theory predictions – is primarily driven by short positions and is disappearing in recent years. However, machine-learning based strategies are more profitable in long positions and remain viable in the post-2001 period.

“This could be particularly valuable for real-time trading, risk management, and long-only institutions. In addition, machine learning methods are more likely to specialise in stock picking than industry rotation,” Prof. Cheng adds, referring to strategies which seek to capitalise on the next stage of economic cycles by moving funds from one industry to the next.

The study is the first to provide large-scale evidence on the economic importance of machine learning methods, she adds.

“The collective evidence shows that most machine learning techniques face the usual challenge of cross-sectional return predictability, and the anomalous return patterns are concentrated in difficult-to-arbitrage stocks and during episodes of high limits to arbitrage,” Prof. Cheng says. “Therefore, even though machine learning offers unprecedented opportunities to shape our understanding of asset pricing formulations, it is important to consider the common economic restrictions in assessing the success of newly developed methods, and confirm the external validity of machine learning models before applying them to different settings.”