The team successfully concluded a great amount of research in 2020, including into the predictive power of text, directors’ transactions in their own companies’ stocks, ESG, factor rotations, value as an investment style, and stocks’ trading volumes. The fruits of this research quickly started making a meaningfully positive contribution to returns in the portfolios managed by the team.
The Jupiter systematic Team

The Jupiter Systematic team, headed by Amadeo Alentorn, and including Tarun Inani, James Murray, Yuangao Liu, Sean Storey and Matus Mrazik (from left to right)

Central to our philosophy as a team is a continuous and disciplined research effort to ensure that our investment process improves over time, and that we try to learn lessons from new market environments and more challenging performance periods. Over the past 15 years, this philosophy has resulted in a regular stream of evolutionary changes to our investment process, leading to improvements in our expected risk adjusted returns over time. In addition to the research activity undertaken by our in-house research team, we have also continued to interact very productively with our team of academic advisors, with external researchers in our Virtual Laboratory, and, since the Merian acquisition, also with Jupiter’s in-house Data Science Team.

The aim of this note is to give an overview of research activity undertaken by the systematic equities team over the course of 2020. In particular, we outline a number of projects that were successfully concluded, and which led to the introduction of three new stock selection components in our alpha model during the year. More detailed notes covering each of these three new components are available on request.
Harnessing the predictive power of text
The last few years has seen an explosion in so called “alternative” and “big” data, typically offering very granular insights into a small subset of stocks. While we remain somewhat sceptical about the use cases for many of these datasets, one area in which we have had tangible success is textual analysis.

We undertook a long project to systematically utilise information that company managements transfer to investors when engaging with analysts. It uses textual analysis to accomplish this. This not only reduces the information advantage fundamental investors have due to information obtained through this channel, but also creates our own information advantage via the breadth of companies we are able to analyse through systematising the process. Furthermore, systematising the process helps us to avoid behavioural biases to which fundamental investors may be prone when evaluating company management, such as anchoring current opinions to past outcomes.

When examining management communications, we explicitly focus on sentiment, believing it offers insight into management thinking, helping us forecast future company fundamentals and asset prices. It is important to consider that company management are incentivised to portray their company in the best light. To incorporate this feature of our text, we include measures of quality, designed to distinguish between realistic and embellishing sentiment. v The successful completion of this project led to the introduction of a new stock selection component within our analyst sentiment stock selection criterion. The incorporation in a systematic manner of soft, non-factual information for the first time within our investment process was a significant milestone in the evolution of our process, and an augmentation of our existing information mosaic to enhance the systematic analysis of companies and to help us make better investment decisions.
Directors’ Transactions
Another area of alternative data that we have been evaluating and monitoring for some time is that focusing on directors’ transactions. After revisiting the results from a project that we undertook a few years ago on this topic, we were encouraged by the consistency of results with an updated dataset, and its performance during the out- of-sample period.

There is strong academic support for the usage of information based on directors’ transactions in their company’s stock, with one of the main arguments being information asymmetry. This argues that company directors enjoy information advantages over external investors about the firm’s profitability and prospects. The fact that they have disproportionally more information than outsiders can result in information about their transactions in their company’s stock, when noise is removed, being very powerful.

Many features of directors’ transactions can be important when determining the strength of the information of those trades. For example, directors’ trades in their companies’ stocks can be triggered for various reasons, and can involve different transaction types (buy, sell, awards, etc), resulting in them having different impact. Size of trade in both dollars and percentage of the director position being traded can also be important, as is the seniority level of the individual making the trade. It is also important to incorporate company characteristics, such as current valuation of the company, and recent price movements.

All these considerations have helped us take the seemingly random transactions made by company directors and distil them into a useful, proprietary stock selection signal. This new signal was introduced in the model in April 2020 as part of the company management stock selection criterion. It has been designed to complement and diversify the existing signals in this component, which mostly focus on evaluating the historic corporate decisions made by a management team. With this new signal, we now also take into account decisions made by individual managers in their personal portfolio, as they trade in and out of the stock of their company.
A Systematic Approach to ESG
Investors have a responsibility to integrate environmental, social and governance (ESG) considerations into their investment decision- making process. All parts of the investment chain, from asset owner, investor and the investee company themselves, face societal demands to focus on ESG. We fully understand and support our stewardship responsibilities to our clients and to wider stakeholders.

As a team, we have been conducting research for some time on the potential usefulness of ESG within our investment process. For us, the incorporation of ESG into our strategy is not about greenwashing the portfolios. It is about systematically incorporating ESG information alongside all our other stock selection criteria, to fully integrate ESG considerations in making stock return forecasts, while maintaining the large breadth of our investible universe.

Our research has identified a number of aspects that can help turn raw ESG scores from data providers into a useful alpha strategy. One of the first considerations is to understand the accidental style exposures of these raw ESG scores. Some of these style exposures may be due to structural biases in the data collection process or inherited from the human biases of ESG analysts determining the scores. Other accidental style biases in ESG data may be more time-varying, driven by market forces, like the recent elevated valuations of some highly scoring ESG stocks.

Another interesting finding that our research has uncovered is the benefit of considering not only the levels of ESG scores, but how they have been evolving through time. There is an argument for favouring companies with good ESG scores, where these have been improving, helping us to capture potential turnaround situations. In particular, we are interested in capturing the impact on the future stock price of companies whose management teams have been making efforts to improve their lower ESG scores.

As a result of this body of research, a new ESG based stock selection component was introduced in June 2020, as part of our company management stock selection criterion. Since then, ESG has been fully integrated in our investment process, whereby the daily updated return forecasts for all the stocks in our universe are incorporating ESG considerations in a systematic manner.
Higher Conviction Factor Rotatings
Another area where we have continued to dedicate research efforts is in our dynamic rotation model. The implementation of our first dynamic rotation model more than a decade ago has been key in generating strong and uncorrelated performance over time, by systematically rotating in and out of investment styles according to changes in the type of market environment. A number of key enhancements were made to that model in September 2019, incorporating information from downside risk measures as additional inputs in determining optimal factor weights, in order to improve both the strategic positioning across factors, and reduce exposure to factors that exhibited higher downside risk in a given type of market.

Our research focus during 2020 built on top of that work, shifting to identify and strengthen significant relationships between our market environment indicators and subsequent factor payoffs.

Where our methodology identifies a higher degree of certainty about a relationship, our model will rotate with higher conviction; and conversely, weaker relationships result in less conviction. This methodology has been also translated into our dynamic valuation stock selection criterion to improve the rotation between value and quality, and to unify the various factor rotation aspects of our model in a more generalised framework. Throughout 2021 we shall continue to work on how to further improve the efficacy of our dynamic rotation model, by looking to expand the number of conditioning variables, and exploring more advanced econometric models. Part of this work will continue to be done through our collaboration with academics from Imperial College London.
Re-Examining Value
Value as an investment style has experienced poor performance for some time now, and particularly over the last couple of years. Despite this, our value-based signals have still been a significant source of positive returns over the long run. This has been thanks to our proprietary approach to constructing valuation models, as well as our dynamic approach to tactically deploying this style of investing, increasing its weight in periods of high risk appetite, while reducing it in periods when risk aversion dominates investor behaviour.

In 2019 we enhanced the dynamic valuation stock selection criterion, by making the factor forecasts for value and quality more robust to short term noise, and by taking into account the time varying nature of investor demand for each individual style. This improves the model’s ability to both mitigate downside risk and participate in positive performance in periods where both value and quality are either in or out of favour simultaneously.

In 2020, we continued our research efforts in this area, looking to learn further lessons from the behaviour of this style in recent years. Specifically, we initiated a new research project to re-examine the payoff profile of our valuation framework and how it relates to the style’s theoretical underpinnings. We hope this will enable a more precise decomposition of the value premium into its different underlying drivers, each with different payoff horizons and risk profiles. This will hopefully help mitigate future drawdowns if value as a style were to underperform again, while continuing to benefit from the upside which the strategy offers.
Trading Volume
In academic literature, a lot of attention has been given to price- based strategies, whilst strategies utilising trading volume have not attracted as much attention. However, for many equity market participants, signals combining price and volume are widely used, particularly for forecasting stock returns over shorter horizons. While our investment horizon is more medium term, we have continued to explore how volume indicators can further help us refine our price driven signals. We have been exploring how an increase or decrease in the trading volume of a stock can help determine the strength of a trend, and the impact of volume spikes in the future behaviour of a stock. It also offers opportunities to improve signals to capture positive price trends, like momentum, but also for signals that look to capture reversals in prices. This is an area of research that we shall continue to work on during 2021.