In a crowded world where active equity strategies increasingly look alike, machine learning’s potential to deliver genuine differentiation as well as mass-customization make it an invaluable tool for investment managers.
While traditionally there has been a wide chasm between quantitative and fundamental investing, the next frontier lies in building a strong bridge between machine and human. This goes beyond a ‘quantamental’ approach which blends both, to a process that leverages the best of both worlds. More specifically, the promise of machine learning is to combine the breadth and dynamism of quantitative equity investing with the depth and analytical rigor normally seen in top human managers.
What are the benefits of machine learning technology in investing?
Advancements in machine learning (ML) technology bring several key benefits to fundamental investing, including the ability to:
- Systematically apply the rigor of fundamental research to a wide breadth of stocks
- Identify market inefficiencies and opportunistically take action when a great stock becomes temporarily mispriced
- Create a portfolio that is customizable to meet specific client objectives in a cost-effective manner
As a result, the strategy does not have static style or factor (or even human) biases. This leads to a portfolio that behaves differently than that of other traditional active managers. Moreover, the anti-crowding, contrarian approach can result in a return stream that is uncorrelated to both active and passive equity funds.
Where does machine learning fit into the evolution of investment management?
According to the CFA Institute, "the evolution of investing can be divided into three “waves.” The first and second waves of “fundamental discretionary” and “quantitative” investing, respectively, have been followed by an emerging third wave that involves sophisticated computing techniques and leveraging the power of machines. Strategies in the third wave are a confluence of 5 factors: 1) The exponential growth of data, 2) Data science, 3) Machine learning, 4) Record-low processing and storage costs, 5) Playing close to the information edge."
Can machines assess stocks based on more than fundamental (i.e. structured) data?
Recent academic research suggests that predictions of company performance can be improved by using Natural Language Processing (NLP) techniques on text contained in company filings. Li (Journal of Accounting Research, 2010) showed that companies expressing positive sentiment in their MD&A (management discussion & analysis) section tend to have continuing business momentum that will translate into rising share prices.1 Identifying positive sentiment within hundreds of stocks would be difficult and time consuming for a human analyst, but, as noted above, machines, using NLP technology, are able to analyze thousands of company filings to help improve forecasts.
Will robots replace human investment managers?
A key advantage the machine has in competing against humans is that it can assess the entire stock universe rapidly. But, unlike traditional quant strategies that can have wide breadth but limited depth, not only can machine-learning technology quickly assess information, it can make sense of it, too.
Nevertheless, even the best machine intelligence system is not ‘smarter than humans’ in a pure sense. It (just) makes fewer mistakes. Some humans can apply the discipline needed to pick stocks. Yet, even the most successful stock pickers makes mistakes along the way, for instance “falling in love with a stock” (the case of Valeant pharmaceuticals being a classic example). Successful investing over the long-term, by definition, requires standing apart from the crowd. This is uncomfortable, and not in a metaphysical way. Learning the rules of the investment game is hard, despite the brain having incredible pattern recognition technology.
While machines are indeed learning to emulate the skills of the world’s best stock pickers, investing is not Chess, and we believe the best approach is one of symbiosis – combining the strengths of human and machine.
What is the difference between AI investing and algorithmic trading?
The ‘rise of the machines’ in investment management began with trade execution (that is, the implementation of buy/sell trade decisions) and then was extended to market making, followed by fast-trading strategies like statistical arbitrage. Almost all of this is now algorithmic. Hence, when the media reports that “robots dominate trading” what this means is simply that algorithmic trade execution is now the norm.
Applying AI tools to fundamental investing is the next logical step – using machine intelligence to select stocks over the medium to long-term.
What is Voya Investment Management’s approach to machine learning?
Voya IM’s Equity Machine Intelligence (EMI) team uses machines to learn - and consistently apply - the rules of the game. That is, to findand exploit persistent patterns in company fundamental data.
Through the EMI team, Voya’s Quantitative Equity Platform intends to lead the way in combining the best of traditional quant/human processes with advanced artificial intelligence and machine learning techniques, increasingly seen as the next frontier for asset managers.
Our approach is evidence based, and data-centric. Before the machine can deploy its algorithms, the human team builds rules to integrate and scrub complex data sets consisting of 10,000 unique data points per company (spanning across macro and sentiment, fundamental and ESG data). While perhaps this is the least sexy part of the process, data preprocessing is vital for success. Once complete, the next step is to move from data to features, involving a hybrid human + machine effort. The resulting feature set (think complex financial ratios) help the virtual analysts find predictive patterns based on a deep, holistic view of each company in the universe.
This process, dubbed feature engineering in the jargon, means that the input variables (252 in total) are explicit, rules-based, and fully transparent. Moreover, they are carefully curated – by humans – to ensure that each feature is backed by an economic rationale and is not conflating noise with signal.
We also have the added benefit of working with Voya’s team of fundamental and quantitative analysts, who improve upon our insights and feature construction.
What level of human intervention is involved?
There are tens of thousands of data-points at a company level that could be assessed to decide whether that company will outperform its peers. While this is too much for any human to handle, machines are the perfect partner with their ability to rapidly process vast sums of incoming data in a completely probabilistic manner.
That said, there is a strong need for a ‘human in the loop’ – in both an active monitoring role and in terms of managing the data (inputs) and in risk management of the portfolio (outputs).Machine learning (ML) technology delivers the virtue of patience and the ability to act quickly and decisively when opportunities arise, but the pilot of the plane should always be human
Hence, while ML is at the core of our stock selection process, the approach is by no means machine only, as shown in the diagram below. Our virtual analysts and virtual traders are solely responsible for picking stocks; the human team oversees all processes and controls both the inputs (data) and outputs (risk managed portfolio).
Importantly, we rely on human portfolio managers to oversee risk management and trade execution. Every trade is reviewed by the team and they will override a trade if deemed necessary, such as if there is a corporate action which is not yet captured by the model. This step is key to managing the risk of an algorithmic error, which is sometimes experienced by high frequency traders (and a popular hobgoblin of those critical of machine-only processes!).
What is the AI approach that Voya Investment Management applies?
Instead of a single ‘magic algo’, there are 26 machine learning models (we call them virtual analysts) that compete to select stocks and 45 virtual traders who help cut losses and time trades. It’s an ensembled and layered approach.
Our virtual analysts have the same starting point as everyone else: publicly available information which, in aggregate, comprises 10,000+ unique data points per company. Within this stock selection framework, our virtual analysts become “rejection machines”, saying “no” to more than 96% of stocks - rather than analyzing a few and getting to yes. When the system does finally say yes, our virtual analyst’s decision is unemotional and based on sustainable patterns in 20+ years of fundamental, ESG and sentiment data.
Does AI impact entry/exit timing of buys and sells?
Prior to purchase, all buys and sells must pass through our ‘team’ of 45 virtual traders, specialized models which analyze shorter term top-down data (e.g. price and sentiment) to determine the entry/ exit timing of each stock as well as position sizing. The decision is void of emotion, allowing the virtual traders to take positions in stocks that may be otherwise uncomfortable to portfolio managers and avoid chasing after crowded trades.
Can AI investing be explainable or is it inherently a ‘black-box’?
The EMI team makes use of tools known as interpretable or explainable AI (XAI), which allows the basis of decision making to be make explicit. This is a key benefit of the technology – both for clients and for helping direct work on system enhancements.
More specifically, EMI’s proprietary technology allows human portfolio managers to interrogate the system and gain visibility on the rationale of each buy and sell decision. Indeed, they can go into the same level of detail that a fundamental analyst would in support of a bottom up stock-picking strategy – minus the human bias. A caveat – this does not mean that AI can replace human analysts (who, for instance, may be better placed to forecast long-term trends or structural changes) – but it does mean that AI can be geared to make fewer mistakes than most humans.
What is a recent example of an AI-driven stock pick?
To illustrate how the investment rationale can be made explicit, let’s take the recent purchase of Wells Fargo & Co (ticker: WFC) – a position which was initiated in October 2020 as a top-5 portfolio holding in EMI’s US Opportunistic Alpha mandate.
From October 5, 2020, several of EMI’s virtual analysts found the stock to be attractive because it fell into a broad predictive pattern: a cheap, distressed, fallen angel where sentiment is overly negative, yet fundamentals appear to be slowly turning around and the stock price may be bottoming (i.e. a classic “Turnaround” play).
How do we know this? By using inhouse explainable AI (XAI) tools to make the decision-making logic of the virtual analyst explicit, including features which underpin the pattern. For illustrative purposes, these features include:
- Value: Cheap relative to peers and history according to various adjusted metrics such as EV to free cash-flow, price to earnings and price to book
- Quality: Dozens of important balance sheet and cash-flow statement items are improving, as are various proprietary ESG metrics such as governance, compensation and other issues
- Sentiment: Historically low ratings and price targets from sell-side analysts; various market signals show signs of investor capitulation
- Technicals: Price bottoming, relative strength, and long-term negative momentum (reversal signal)
In addition to the basic investment rationale described, we can clearly identify why a stock was bought or sold at a particular time from the virtual traders. In the case of Wells Fargo, sentiment is so negative that human nature could deter even a skilled analyst from purchasing the stock despite improving fundamentals. Conversely, our virtual analysts and traders remove emotion from investing and do not shy away from opportunistically buying into a story identified as a potential turnaround.
Finally, and perhaps more importantly, the technology allows for performance attribution at the level of the patterns themselves – to determine what is really driving portfolio returns. Coming back to WFC, the original purchase represents an opportunistic turnaround play with a time horizon of 3 – 24 months.
1 Li, F. (2010) “The Information Content of Forward-Looking Statements in Corporate Filings—A Naïve Bayesian Machine Learning Approach,” Journal of Accounting Research, vol. 48(5), pp. 1049-1102.
This commentary has been prepared by Voya Investment Management for informational purposes. Nothing contained herein should be construed as (i) an offer to sell or solicitation of an offer to buy any security or (ii) a recommendation as to the advisability of investing in, purchasing or selling any security. Any opinions expressed herein reflect our judgment and are subject to change. Certain of the statements contained herein are statements of future expectations and other forward-looking statements that are based on management’s current views and assumptions and involve known and unknown risks and uncertainties that could cause actual results, performance or events to differ materially from those expressed or implied in such statements. Actual results, performance or events may differ materially from those in such statements due to, without limitation, (1) general economic conditions, (2) performance of financial markets, (3) changes in laws and regulations and (4) changes in the policies of governments and/or regulatory authorities. The opinions, views and information expressed in this commentary regarding holdings are subject to change without notice. The information provided regarding holdings is not a recommendation to buy or sell any security. Fund holdings are fluid and are subject to daily change based on market conditions and other factors.
Past performance is no guarantee of future results.