Constructing Social Portfolios: A Quantitative versus Screening Approach

By Alina Hofer, Lea Katharina Kasper & Dr. Kristjan Jespersen 

◦ 5 min read 

When we talk about ESG, one could argue that there is a strong bias focused on climate investing, reaching net zero targets as well as good corporate governance and diversity themes. But there is much more to ESG. The “Social” dimension of ESG is hugely under explored and developed and covers under studied issues such as how companies treat their employees and care for the responsibility of their products. Still further, assessments linked to human rights codes and social impacts is only now receiving the attention it truly deserves. Although the importance of these topics is undisputed, we see that attention to particularly address the social dimension has been lacking, whereas awareness of other ESG risks has been rising immensely during the past years. 

Not only is the general knowledge and focus on the social dimension of ESG limited, its overall  implementation in portfolio management has not been sufficiently experimented with and addressed.

The delay to properly implement the “S” in ESG is often explained because of the challenges to quantify, assess, and integrate social factors generally.

However, this argument should not be a sufficient justification for neglecting the “S” in ESG and for investigating a possible relationship between a good social rating and superior financial performance. To tackle this lack of awareness, we constructed two portfolios which integrate Refinitiv’s Social ratings based on different integration strategies and test their performance towards the market between 2012-2021.

When integrating social – or other ESG – ratings into the investment process, we find there is often disagreement on how to best consider these factors in portfolio construction. Currently, it is most common to apply screening or best-in-class strategies. These approaches aim to remove assets that do not fulfill certain criteria from a defined investment universe. Negative screening would mean to remove those companies that perform worst from the pool of assets. Inversely, an investor could also only continue with those firms who at least have a certain minimum rating. For both approaches, the portfolio weights are then allocated to the assets that remain. This is done using conventional indicators such as value, size or expected risk-adjusted returns. In our study, we, however observe a clear shortcoming of this approach: After screening out the worst 10% “social performers” and allocating weights based on a risk-return trade-off, the portfolio does not necessarily promise a higher overall ESG score than a portfolio would reach which does not consider the ratings at all. Although the portfolio yields a solid financial performance, this raises the question whether any ESG-related impact has been made with this integration approach.

To make sure an investor can improve his exposure to assets that score well in the social dimension, we integrate the rating scores directly into the optimization problem of our second portfolio. This leads to a very different outcome on the social rating:

Looking closer at the mechanics of this approach, we extend the traditional Sharpe Ratio with the ESG factor, meaning to add by how much it a company “outscores” the market average. This results in the following “Social Sharpe Ratio”:

We add a fifty percent weight split, which can be flexibly adjusted towards investor preferences. And we now balance a risk-return-social trade-off. This explains why the second approach over 9 years constantly beats the market average in respect to the integrated Social factor without sacrificing any performance on the financial side. In fact, we find that in 5 out of 9 years, the second strategy would have also led to higher risk-adjusted returns measured by the Sharpe ratio. Moreover, returns were consistently higher compared to the market benchmark. This result is quite remarkable, given that it is often questioned whether investors need to sacrifice returns in order to make their investments more socially responsible. 

Lastly, our study resulted in one more unforeseen twist when it comes to integrating ESG ratings. That is, the question whether we can actually trust the rating scores. To answer this, we must first understand how scores are created. Rating providers look at an immense amount of publicly disclosed information, reports and policies. And based on what company’s report, rating scores are aggregated. However, it is clear that a firm would only report on things they do well. In fact, we observe that with increased reporting, ESG scores also improve. But what about the real-life actions and impacts? Some rating providers offer a combined score, which also considers media reports on the involvement in controversial actions. As these scores are only available at an aggregate level, we calculate them on a single-pillar level using Refinitiv’s methodology, which adjusts for firm size and industry. Looking at specific examples in our portfolios, we found that the impact of such controversy involvement on the overall score could still be larger. Nevertheless, we stress that in order to have a complete picture of a firm’s ESG behavior, the impact of these controversies needs to be reflected in investment decisions. 

To sum up, given the results of our research, there are three things we aim to highlight:

  • It is crucial to increase investors’ awareness of “Social” matters and provide a better landscape for impact investments in this specific dimension.
  • Integrating ESG ratings does not always promise a better ESG performance for the whole portfolio. Therefore, it is necessary to focus on strategies that lead to actual impact.
  • Third, looking beyond the information that is disclosed by companies themselves, more attention should also be addressed to “real life actions” when making investment decisions. 

About the Authors

Lea Kasper has recently graduated with a MSc. in Finance and Investments (cand.merc.) from Copenhagen Business School. Her interest and enthusiasms about sustainability and how to more efficiently integrate non-financial factors in investment decision-making contributed to her choice to further investigate this topic throughout the master thesis. 

Alina Hofer has recently graduated with a MSc. in Finance and Investments (cand.merc.) from Copenhagen Business School. Being passionate about creating impact through ESG-aligned investments, she was excited to further focus on her interest in this field throughout the master thesis.

Kristjan Jespersen is an Associate Professor at the Copenhagen Business School. He studies on the growing development and management of Ecosystem Services in developing countries. Within the field, Kristjan focuses his attention on the institutional legitimacy of such initiatives and the overall compensation tools used to ensure compliance.


Image source: SustainIt

Why we should not call for experts instead of experiments

By Jan Michael Bauer

◦ 3 min read 

The recent elections in Germany turned out as the historic loss for conservatives that pollsters have predicted a few weeks before. Responding to the declining poll numbers, the conservative party presented a “team for the future” consisting of several field experts that should help the candidate addressing the big challenges of the country. Their slogan was „Experts instead of experiments“. The message was clear: we know how to solve these issues and voting for another party would be an experiment and therefore risky. While this might appeal to a conservative base, I think this slogan sends a wrong if not hypocritical message.  

Framing an experiment as something uncertain and dangerous that should be avoided taints one of sciences most successful methods and obscures the undeniable level of uncertainty associated with policy decisions.

Even after acknowledging the turbulent times during Merkel’s legacy, remarkably little was done to address the challenges of the future. While some inaction might be attributed to a lack of courage or lobbying by special interests, it certainly constitutes the lack of obvious and simple solutions for the many problems the country is facing. 

The challenges we face are new and unpredicted in magnitude. While few would disagree with the need for action, there is disagreement about what needs to be done. Experts argue with each other, often struggle to persuade their colleagues, and remain unconvinced by the evidence that substantiates the oppositions claims. Fierce debates about the right course of action often overshadow a sad truth that in many cases no one is and really shouldn’t be sure that the proposed path will be the most successful one. Even more disheartening might be that even after the implementation of a policy, we often have a hard time qualifying if the intervention did more good than harm or quantifying these benefits. 

A fundamental problem, particularly economists try to resolve through (quasi-) experimental research methods to understand how a specific policy intervention works. Broadly speaking, they become experts because they do experiments. Pioneers in the study of causal relationships were recently awarded the Nobel Price. Among them David Card who is famous for a study on the effects of a higher minimum wage on employment exploiting a so-called natural experiment.  

Experimentation can help us to find out if our ideas and theories work in practice. They should increase our confidence in the people applying them rather than creating a fear of uncertainty.

Our knowledge that COVID vaccines are effective mostly relies on the result of randomized experimental trials. An approach increasingly used to answer questions in the social sciences. For instance, we don’t know how people will respond to universal basic income, which is why a three-year experimental study is currently on its way in Germany.

Pharmaceutical trials are also designed to show that potential drugs have no sever side-effects. While the necessity to ensure the safety of a drug is quite intuitive, the unintended consequences of non-medical products and services are less straight forward. For instance, social media has been suspected to inadvertently contribute to political polarization and erode democratic processes. While many of these claims are based on anecdotes, recent experimental studies from researchers outside Facebook have added hard evidence to the debate and conclude: 

Our results leave little doubt that Facebook provides large benefits for its users [but also] make clear that the downsides are real. We find that four weeks without Facebook improves subjective well-being and substantially reduces post-experiment demand, suggesting that forces such as addiction and projection bias may cause people to use Facebook [..] it also makes them less polarized by at least some measures, consistent with the concern that social media have played some role in the recent rise of polarization in the United States.

– Allcott, Hunt, Luca Braghieri, Sarah Eichmeyer, and Matthew Gentzkow. 2020. “The Welfare Effects of Social Media.” American Economic Review, 110 (3): 629-76.

There are obvious differences between vaccines and a social media platform, and probably nobody would suggest that Facebook should have undergone a randomized safety study in the mid-2000s before going public. Such products and services develop over time and can be used in very different ways. However, despite these differences there is an open question about the potential side-effects and the burden of proof. To ensure a healthy society, it might be worth considering that at least with a reasonable initial suspicion of harm, also non-medical companies should be obligated to proof their products’ safety using a suitable experimental design.

There will remain many problems where experiments are unfeasible, and, as seen with the development of Facebook, even the results of the best experiment today might not be a valid description of tomorrow in an increasingly complex and dynamic world. Such a world, however, should also humble us and our experts but foster an acknowledgement that there are many questions for which we don’t know the answer. Hence, we should make use of the best scientific methods available to reduce this uncertainty, which ultimately means that we need more experiments and not less.  


About the Author

Jan Michael Bauer is Associate Professor at Copenhagen Business School and part of the Consumer & Behavioural Insights Group at CBS Sustainability. His research interests are in the fields of sustainability, consumer behavior and decision-making.


Photo by Scott Webb on Unsplash