• Lawrence Cummins

Machine Learning for Asset Managers

Updated: Jan 4

BLGI Chronicle’s


Some industry experts argue that Machine Learning will reverse an increasing trend toward passive investment funds. But although Machine Learning offers new tools that could help active investors outperform the indexes, it is unclear whether it will deliver a sustainable business model for active asset managers.

A form of artificial intelligence, Machine Learning enables powerful algorithms to analyze large data sets to make predictions against defined goals. Instead of precisely following instructions coded by humans, these algorithms self-adjust through trial and error to produce increasingly more accurate prescriptions as more data comes in.

Machine Learning is particularly adaptable to securities investing because the insights it garners will act quickly and efficiently. By contrast, when Machine Learning generates new ideas in other sectors, firms must overcome substantial constraints before putting those insights into action. For example, when Google develops a self-driving car powered by Machine Learning, it must gain approval from an array of stakeholders before it can hit the road. These stakeholders include federal regulators, auto insurers, and local governments where these self-driving cars would operate. Portfolio managers do not need regulatory approval to translate Machine Learning insights into investment decisions.

In the context of investment management, Machine Learning augments the quantitative work already done by security analysts in three ways:

Machine Learning can identify potentially outperforming equities by finding new patterns in existing data sets.

For example, Machine Learning can sift through the substance and style of all CEOs' responses in quarterly earnings calls of the S&P 500 companies during the past 20 years.

Machine Learning may generate insights applicable to current CEOs' statements by analyzing the history of these calls relative to good or bad stock performance. These insights range from estimating the trustworthiness of forecasts from specific company leaders to correlations in firms' production in the same sector or operating in similar geographies.

Some of these new techniques produce significant improvements over traditional ones. In estimating the likelihood of bond defaults, for example, analysts have usually applied sophisticated statistical models developed in the 1960s and 1980s, respectively, by Professors Edward Altman and James Ohlson (notably the Z and O scores). Researchers have found that Machine Learning techniques are approximately 10% more accurate than those prior models at predicting bond defaults.

MACHINE LEARNING can make new forms of data analyzable.

In the past, many formats for information such as images and sounds were understood by humans, and such compositions were inherently challenging to utilize as computer inputs for investment managers. Trained Machine Learning algorithms can now identify elements within images faster and better than humans can. For example, by examining millions of satellite photographs in almost real-time, Machine Learning algorithms can predict Chinese agricultural crop yields while still in the fields or the number of cars in the parking lots of U.S. malls on holiday weekends.

A flourishing market has emerged for new forms of these alternative datasets. Analysts may use GPS locations from mobile phones to understand foot traffic at specific retail stores or point of sale data to predict same-store revenues versus previous periods. Computer programs can collect sales receipts sent to customers as a by-product of consumers' various apps as add-ons to their email system. When analysts interrogate these data sets at scale, they can detect useful trends in predicting company performance.

Machine Learning can reduce the adverse effects of human biases on investment decisions.

Behavioral economists and cognitive psychologists have shed light on most humans' extensive range of irrational decisions. Investors exhibit many of these biases, such as loss aversion (the preference for avoiding losses relative to generating equivalent gains) or confirmation bias (the tendency to interpret new evidence to affirm pre-existing beliefs).

Machine Learning can be employed to interrogate portfolio managers and analyst teams' historical trading records to search for patterns manifesting these biases. Individuals can then double-check investment decisions fitting into these unhelpful patterns. To be most effective, individuals should use Machine Learning to check for bias at every level of the investment process – including security selection, portfolio construction, and trading executions.

Despite these substantial enhancements to investment decisions, Machine Learning has its own very significant limitations, which seriously undercut its apparent promise.

Machine Learning algorithms may exhibit significant biases derived from the data sources used in the training process or from the algorithms' deficiencies. However, Machine Learning will reduce human biases in investing. Firms will need to have data scientists select the right sources of alternative data, manipulate the data, and integrate it with existing knowledge within the firm to prevent new biases from creeping in. This ongoing process requires competencies many traditional asset managers do not currently have.

Machine Learning can be highly effective at examining vast amounts of past data from one specific domain and finding new patterns relative to an express objective. It does not adapt well to rare situations such as political coups or natural disasters. Nor can Machine Learning predict future events if they are not closely related to past trends, such as the 2008 financial crisis. In these cases, investment professionals must judge future trends based partly on their intuition and general knowledge.

Finally, many of the patterns Machine Learning identifies in large data sets are often only correlations that cast no light on their underlying drivers, which means that investment firms will need to employ skilled professionals to decide if these correlations are signal or noise. According to a Machine Learning expert at a large U.S. investment manager, his team spends days evaluating whether any pattern detected by Machine Learning meets all four tests: sensible, predictive, consistent, and additive.

Even when Machine Learning finds patterns that meet all four tests, these are not always easily convertible into profitable investment decisions, which will require a professional's judgment. By sifting through social media reams, Machine Learning might have been able to predict — contrary to most polls that Donald Trump would be elected president in 2016. However, making an investment decision based on that prediction would present a difficult question. Would Trump's election lead the stock market to go up, down, or sideways?

The bottom line is that while Machine Learning can significantly improve data analysis quality, it cannot replace human judgment. To utilize these new tools effectively, asset management firms will need computers and humans to play complementary roles.

As a result, firms will have to make substantial investments in technology and people, although some of these costs will offset by cutting back on traditional analysts' numbers.

Unfortunately, most other asset managers have not gone far down the path to implementing Machine Learning. According to a 2019 survey by the CFA Institute, few investment professionals are currently using computer programs typically associated with Machine Learning. Instead, most portfolio managers continued to rely on Excel spreadsheets and desktop data tools. Moreover, only 10% of portfolio managers responding to the CFA survey had used Machine Learning techniques during the prior 12 months.

Perhaps predictably, it is the largest asset managers, like Schroders, BlackRock, and Fidelity, leading the way, nurturing relationships with information suppliers, technology providers, and academic experts. But they are unlikely to open a significant gap over competitors as the scale is not necessarily an advantage in active investment.

For instance, trading in large volumes can carry high costs, and firms are constrained in the overall exposure they can bring in stock.

Mid-size asset managers should also benefit because they are likely to attract and retain high-quality data scientists. The latter may see more opportunities for advancement there than in large firms. Besides, mid-size firms can afford access to alternative data through third-party vendors, high-quality algorithms from open source libraries.

And sophisticated tools from the technology companies (e.g., Amazon and Google) are already offering cloud-based services to many industries.

The losers are likely to be small firms (with less than $1 billion in assets under management). They are likely to have trouble attracting enough talent and absorbing the cost of developing the technology, given the intense downward pressures on active managers' fees. Management fees for active equity managers are roughly 20% lower in 2018 than in 2008, partly because passive funds have become so cheap. Asset managers are also under regulatory pressure to pay their cash for outside securities research instead of paying with "soft dollars" by allocating brokerage commissions to good research firms. Therefore, investments required by Machine Learning come at a difficult time generally for the asset management industry, and this will be particularly challenging for small firms.

What's more, it is unclear whether substantial investments in Machine Learning will lead to a long-term sustainable business model for active asset managers. If Machine Learning generates unique alpha for an investment firm, it cannot sit on its laurels for long because other firms are likely to simulate its investment methods. And suppose other asset managers derive similar insights from similar Machine Learning techniques. In that case, they will be buying or selling the same securities simultaneously, which may have the effect of wiping out any gains the understanding can generate.

For example, over three days in 2007, several significant hedge funds, using quantitative models based on the same factors, liquidated their positions simultaneously and suffered substantial losses as a result.

To sum up, Machine Learning may be seen initially as the savior of active investing. It indeed can allow early adopters to find new sources of alpha and outperform the indexes. Yet if other managers copy Machine Learning insights as they develop Machine Learning capabilities, it may become even more difficult to find publicly traded stocks and bonds that outperform their benchmarks. Over time, will actively invest augmented by Machine Learning increase security pricing efficiency and reinforce the current shift to passive investing.

©BLGI 2021
2027 W. Division St.
Suite 137
Chicago, IL 60622
  • Linkedin
  • Facebook
  • Twitter
  • Instagram