Quantitative strategies represent the best of active and passive approaches. They’re able to generate additional alpha but at low cost and market levels of risk.
David Wright, Co-Head of Quest, Quantitative Business Strategy Pictet Asset Management.
The world of equity investing has long been divided into active and passive strategies. Each has advantages, but the drawbacks have been hard to ignore. A third approach, however, combines some of the best attributes of the old guard: quantitative investing.
Active investing typically involves building concentrated portfolios invested in deeply-researched companies with the aim of generating above-market performance, otherwise known as alpha. The downside is cost and consistency. Management fees as well as the risk associated with poorly-diversified stocks, make this approach expensive on several levels. Active portfolios tend to be more volatile and suffer bigger drawdowns than the market.
A passive approach, by contrast, mechanically mimics a market index. This is both an inexpensive and relatively low risk way of investing in equities. But, in exchange, investors sacrifice any prospect of outperforming the markets.
Quantitative strategies, particularly those with an enhanced index structure, marry the best of both worlds (see Fig. 1). Built around broad portfolios with extensive – if not quite perfect – overlap with market indices, their breadth reduces risk. Those small differences in portfolios – the subtle but ultimately consequential deviations from reference indices – are where quantitative strategies generate alpha.
Harnessing vast reams of data which are processed by a machine learning-based investment model, these strategies are able to identify which stocks will most likely outperform – or underperform – over the short and medium term. These strategies work because although stocks track economic cycles and megatrends over the longer term, over the short term they are subject to any number of transitory influences that drive their prices away from underlying fundamentals. The vagaries of investor sentiment play a big part here. The market’s reaction to corporate news or changes in analyst forecasts or any number of factors can lead to bouts of volatility that cause anomalies in share prices. And it is those anomalies that create opportunities for quantitative strategies – which can look through these mispricings – to generate alpha.
To pinpoint mispricing, quantitative strategies use algorithms to draw out relationships within the data they are fed. These algorithms act as a map, with in-built rules of the road, to guide data on a journey towards a set of outputs. The more different sets of data, the more the rules need to officiate their interactions and the better the model can predict what will happen in future based on past relationships.
The human touch
These models run automatically. But they need to be built, tested, refined and maintained by human experts. Not only do humans specify the algorithms, but they train them with data and retrain them regularly so that the models keep pace with evolving market dynamics. Experts have to decide which machine learning technique to use, which types of data to use – traditional data sets like corporate accounts can be supplemented with alternative sources such as card transactions or social and traditional media. They are needed to ensure the accuracy and stability of the results and to understand the output. They determine the frequency and extent of training required. And they need to rigorously train and test the models with many years of data. The plethora of possible inputs means that an almost infinite variety of models is possible.
Humans are heavily involved to this point, but once a model is selected, it then dictates the purchase and sale of stocks. This automation not only helps to eliminate the emotional biases to which human investors are subject, but it also cuts costs.
AI boosting quantitative investing…
Artificial intelligence (AI) is not a new field, but recent advances in computing processing, cloud computing and open source tools have lowered the barrier to using machine learning (ML), a subset of AI.
AI is now turbocharging quantitative investing, creating the opportunity to develop Quant 2.0 (see Fig. 2). AI’s unprecedented processing capacity enables the investment models to trace ever more complex relationships among ever greater numbers of data series.
The traditional quantitative approach tended to be restricted to analysing a relatively small number of market effects that result in temporary mispricings. This then provides exposure to broader market drivers, known as factors, such as value or momentum. AI offers the possibility of hundreds of potential signals/features at higher frequency. These are generated from data that include company accounts, share prices, analyst notes, press reports, investor responses to new information over the short and longer term. And that’s just scratching the surface.
Unlike traditional machine learning, which identifies linear relationships within the datasets, AI can capture much more complex associations within the data pool, which allows it to have much greater insights into what’s driving stock prices. Being able to identify more complex, non-linear relationships boosts its ability many-fold to find associations between data series.
Traditional quantitative models specify a linear relationship between stock returns and the signals. For example, in a traditional model, an analyst upgrade of a company would suggest its stock would outperform. However, there are many reasons why such a relationship might not hold on a given occasion, or not be captured in time to generate alpha.
A non-linear machine learning model that is trained with historical data can identify the relationships that tell us when the analyst upgrade would most effectively forecast future outperformance. This could be because there’s a broad range of analyst forecasts, or the particular analyst is an outlier or because of timing – say if the company is due to report results soon. If the stock is widely shorted by hedge funds, the model could identify that it is likely to be subject to a short squeeze resulting in a far more dramatic jump in share prices than the analyst upgrade might otherwise warrant. There are potentially tens of thousands of these “conditioning” non-linear relationship across traditional financial data sets that can generate additional alpha.
…to help deliver factor-neutral returns
This much more intricate framework opens the way for the portfolio managers behind the strategy to isolate the stock-specific effects that influence the stock price. To do so, they strip out a multitude of common factors (market, sector, region, industry, country, styles, economic exposures) from each stock’s performance. In doing so, they can identify and extract pure company-related alpha.
Over time, the algorithms evolve, understanding changing economic and market dynamics and incorporating new data series.
To be effective, AI-driven models need humans to set investment parameters. But once those are specified, the trained algorithm makes the buy and sell calls on individual stocks.
Because these parameters limit risk, they will also limit the amount of alpha the strategy generates. However, the compounding effect even of incremental alpha is powerful over time. And given that expected returns on equities are forecast to fall to mid-single digits amid high valuations following several banner years, even one to two percentage points of alpha will make a difference. For example, if we assume a hypothetical market return over the next 10 years of 5% and an assumed outperformance of 1.5 percentage points per year net of fees, the wonder of compounding will lead to an additional 24.8% of return for the client over the decade.
A flexible approach
One of the biggest strengths of quantitative strategies is their flexibility.
By varying its overlap with underlying indices, a quantitative strategy can be tweaked to adjust the amount of tracking error and risk. It can be customised to an investor’s needs by excluding certain stocks or sectors from its investment universe, while still offering the benefit of a quantitative approach. Given the breadth of quantitative strategies’ portfolios, with holdings across all countries and sectors, these customised adjustments don’t hurt the ability to generate alpha. And although there would be higher tracking error for a sustainable quantitative strategy that, say, excludes oil and gas companies or mining firms, than for one without any exclusions, the core focus on risk management in portfolio construction minimises this impact.
A standard enhanced index offering, with a risk profile very similar to its benchmark index offers clients a core holding that can replace a traditional passive index-tracking approach. At Pictet Asset Management, such an approach would entail a core weighting in our Quest AI-driven strategy, potentially complemented by an investment into one of Quest’s higher tracking error, lower beta sustainable equity funds. And because quantitative approaches give a full range of buy and sell signals, they lend themselves to long/short strategies, such as Quest AI’s.
The key to quantitative strategies is constructing the right model and training it with a sufficient amount of the right data. Then, appropriate parameters need to be set to match an investor’s risk appetite and other requirements. Once all that is achieved, the quantitative strategy can generate incremental stock-specific alpha, removing investors’ exposure to general market factors that muddy performance. And it does so at low cost and at market risk. In other words, it offers the best of both active and passive investment approaches.
