Sections

Top 400: Digitising the renaissance

Data is transforming asset management. Are humans a weak link or an indispensable component?

KEY POINTS

  • Creating or finding new data sets will be essential, but data must still play a clear role in an investment strategy in order to add value.
  • Digital tools present a new breed of active equity skill to replace investment skill based on insights drawn from generic market data and financial accounting, which have been commoditised as smart beta.
  • Digitisation levels the playing field. Winners will be firms, large or small, that build organisations and cultures able to integrate data directly into investment and trading teams.

Institutional asset management is undergoing a wave of digitisation, led by active managers looking for data sets they can massage into alpha-generating insights – and the computational scientists and programmers who can use machine learning, artificial intelligence and other techniques to enhance existing investment strategies or develop new approaches to idea generation, portfolio construction and transaction execution.

Several large firms have adopted data-driven investment approaches in the past several years, often aimed at creating a new active equity discipline to supplant the long-standing model of fundamental active research that has faced performance and fee challenges in the past decade. 

But pioneers in the application of digital techniques to asset management say that far from hastening the demise of active equity management, data sciences are fuelling a renaissance in active equity thinking. The revival, they say will favour those firms – large or small – where digital expertise is ‘built in’ rather than ‘bolted on’. The winners will harness technology, primarily software, to augment the inherent talent of portfolio managers and traders, automating repetitive tasks to free up time to improve investment ideas, generate new concepts, and better serve institutional clients.

Digital pioneers believe data sciences will expand the frontier of active equity management, in large measure due to convergence between quant and fundamental approaches. What’s happening, they say, is not so much a blurring of boundaries between established categories as the emergence of new possibilities for identifying sources of value in securities markets and methods to capture that value on behalf of asset owners.

While access to data and data sets will play a big role in determining which managers will take home the competitive spoils over the next several years, the new leaders of the asset management industry will be those firms that get digital culture right. That means not just occupying cool-tech office space, but resolving critical ‘man versus machine’ questions in order to forge teams that meld financial expertise with the ability to automate tasks involving large data sets. In fact, the latest research from scientifically-controlled physiological studies puts hard science around the ‘gut feel’ of traders to demonstrate that humans remain – by a wide margin – the ultimate risk-taking machines.

There will be challenges along the way. Organisational and talent management practices will need to adapt, and firms will need to learn how to develop and control budgets to develop data sources, often in partnership with external parties. 

The stakes could not be higher. The industry has failed to keep pace with rapid technological advances that other industries have already adopted and which clients are coming to expect, according to a recent study from Morgan Stanley and Oliver Wyman. To catch up, asset managers must increase tech budgets by as much as $25bn (€23bn) over the next five years, yet regulatory costs continue to rise and fees remain under pressure. Asset management industry revenue contracted about 5% in 2016, as AUM growth was insufficient to overcome margin contraction, the study found, and barring bearish events, revenue will drop another 3% by 2019. Digitally-based models and execution tools can help managers cut costs, meet compliance requirements, and help develop new products. The ability to harness big data and artificial intelligence, Oliver Wyman says, “is likely to distinguish winners”.

“The rise of new data is requiring a new mindset,” says Gerben de Zwart, head of quantitative equity at APG Asset Management, where he is responsible for €80bn in systematically-managed listed equity portfolios across multiple global strategies both internally and externally managed. De Zwart, who holds a PhD in empirical finance from Erasmus University Rotterdam, continues: “In the past there were a limited number of data sets available and all quant managers were using more or less the same data sets and trying smart algorithms to extract as much value as they could from these individual data sets.”

Digitisation means data and techniques will become more diverse, de Zwart explains. “The question that each quant manager has to ask is, which data sets do I need to spend my time and resources on to improve my investment process.” With different data sets the likelihood increases that active returns will start to vary more than in the past. “The rise of big data and artificial intelligence, I believe, will lead to a wider spread in active returns for active fund equity managers, both quant and fundamental,” de Zwart says. 

New algorithms, artificial intelligence and machine learning will add to the dispersion. “These techniques have not been used in the past or to limited extent,” de Zwart adds, “so their methodologies will be more divergent, which could lead to different return patterns from active managers.” 

APG’s internal quantitative teams explore new data sets to come up with better signals, but also to distinguish their approaches from traditional factor investing, de Zwart says. The mix of styles APG manages is closely held – publicly known is a low-volatility quant strategy. “The other styles,” he adds, “we keep that information to ourselves.”

A step change
The effort to differentiate from a growing menagerie of commoditised factor strategies highlights the potential that managers see for embracing new data sources and advances in machine learning approaches to help build new investment franchises. The industry is undergoing a step change, a mathematical term for a sudden, discontinuous shift, says Vladimir Zdorovtsov, PhD, managing director and director of active quantitative equity research at State Street Global Advisors. The group interacts frequently with fundamental growth and value equity teams and SSGA’s smart beta team. The hallmark of the shift is how data science is creating a new generation of investment skill to take over as the leading edge from the skills based on financial accounting and insights drawn from widely available market data. Those skills have been turned into a commodity by smart beta strategies, he says, in much the same way that patent-protected pharmaceuticals lose their unique status and revenue potential when generic equivalents emerge. “Smart beta is skill that became commoditised,” Zdorovtsov says. “It’s not necessarily skill in creating some factor, it’s skill in delivering a given solution.” 

wesley chan

Concurrently, Zdorovtsov says, active managers “have to push the frontier to find new areas of proprietary skill that deliver additional value, whether in terms of tapping into new data to refine existing model components, finding novel sources of excess returns, managing risk, creating portfolios, or trading execution and implementation.” These new venues to add value “more than offset the potential downside from smart beta,” he adds. Can large investment firms structure active equity capabilities to add significant value? “I would say definitely so,” Zdorovtsov says. “I see it as actually a very beneficial evolution, not just for the industry but for the clientele as well. You can think of active management as life on Earth, and although this looks like a big disruption, life will evolve and it will be better.” 

The additional value active managers can create will derive from “the avalanche of data and the increasing computing power to turn that data into information and ultimately investment relevant insights,” Zdorovtsov says. He expects a massive increase in unstructured data to spark a convergence of quantitative and fundamental equity disciplines. Fundamental information such as site visits and management meetings can be quantified, and increased depth of new data sets about less liquid, lightly-covered companies can help quants develop insights about such securities.

gerben de zwart

Data quest
Success in the digital realm will hinge largely on a manager’s ability to source new data sets that can be utilized to sharpen the investment and trading decisions entailed by a given style, says Wesley Chan, senior vice-president and director of stock selection research at Acadian Asset Management, and holder of a PhD in financial economics from MIT Sloan School of Management. To do that efficiently, he says, managers must “go back to their knitting and figure out what they really have an edge in, and then go look for the data that would answer those questions or help that edge get sharper.”

As a global quant manager overseeing portfolios of hundreds of stocks, Acadian wants data that has wide applicability, rather than being company specific. “We like to push into less well known or less studied factors,” says Chan. “These markets tend to be less liquid, and we think alpha dwells a bit more in those markets.” Finding the right data can be a challenge. “We like a lot of data sources, but oftentimes, we don’t find the broad-based data we need to be readily available so we have to make it ourselves, and that involves partnering with other providers.”

To get ahead of the digital curve, Acadian has developed expertise in managing data partnerships and a digital IP portfolio. The firm has between three and six data engagements active at a given time, carefully defined to help create digital value for specific parts of its investment style. “Running several of these engagements is costly, so you have to sketch out want you want to achieve,” says Chan. “It’s a useful exercise.”

gerben de zwart

In March, Acadian became the first investment firm to collaborate with Microsoft on a project to explore the use of new predictive signals of economic activity derived from the software giant’s Bing Predicts. The prediction tool uses machine learning to aggregate data from trending social media topics, sentiment towards those topics, and trending searches on Bing.

Asset management firms may soon look – and act – more like Microsoft and other software companies. “I’m surrounded by technologists,” says Kristian West, global head of quant equity trading for JPMorgan Asset Management. From London, West leads a team of 46 traders and 49 technology experts across five cities – 48 of the 49 joined the firm during the last two years. The set-up is no accident. “I wanted technology with me,” he says. “It’s quite a profound thing to see if you come to the trading desk, the vast majority of the people you see are technologists.”

The team aims to function like a software firm as well. Its machine learning environment emulates the Google search engine, and JPMAM has adopted an agile digital development process in which technologists build or upgrade code to automate trading tasks throughout the day, instead having traders – and potentially valuable transactions – await periodic releases of software from centralised IT groups typically found in large financial firms. “It’s quite unique from an institutional perspective,” says West. The team mitigates the risk associated with new code by testing upgrades in an alternate platform, and by releasing much smaller pieces of code that carry much less risk. “It was a big mindset change, but through continued success and refinement we achieved buy-in.”

The result is a real-time technology platform. The team’ mantra is click reduction: technologists consolidate tasks to create one-click operations for traders, allowing them to concentrate on higher-value activities such as selecting execution strategies for large or complex orders, or communicating directly with portfolio managers. Even many large orders can now be routed most efficiently by the automated platform, which can select the best venue for execution of trade based on the security, the portfolio requirements and other key data. Overall, automating orders has reduced trade execution costs by about 20% over the past years, West says.

Automated order systems also help managers comply with regulations such as MiFID 2 rules on best execution. “You can evidence why decisions were made,” says West. “From a best execution perspective, we need to show we have a process, a protocol and a methodology to define the parameters and select the order strategy.” Regulators are also more often requiring asset managers to comply with model validation standards created for banks, according to Oliver Wyman. Model validation requirements “are a positive, they force more rigour,” says SSGA’s Zdorovtsov, and digital tools can provide regulatory benefits for active managers that need to meet such standards.

 top 400 2017 big data maturity

Cry freedom
The most important effect of technology may be to set the humans free. While digital methods tap large data sets to generate investment signals, the best signals come from an old source – the ‘gut feel’ of traders. That’s the conclusion of pioneering research into the physiology of traders conducted by Dr John Coates of Cambridge University.

During a two-week experiment on a trading floor, Coates monitored physical and chemical changes as traders received new information. He found that the signals we send to our bodies appear to be immune to the biases expected in behavioural finance. Subjects’ risk aversion declined 44%, “a huge move in a curve that’s not supposed to move”, according to behavioural theory, Coates says. His conclusion: “The dialogue between our brain and our body is the most candid dialogue you can have. The more we can tap into it, the more we’re accessing the best trading signals on the planet.”

Innovations in big data might determine the winners and losers in asset management over the next several years, but at the centre of the technology maelstrom will sit humans. It seems prudent to hedge that bet with a team of good people.

Have your say

You must sign in to make a comment

IPE QUEST

Your first step in manager selection...

IPE Quest is a manager search facility that connects institutional investors and asset managers.

  • DS-2382

    Closing date: 2017-12-14.

  • QN-2383

    Asset class: Residential Property.
    Asset region: Ireland.
    Size: EUR 10m.
    Closing date: 2017-12-18.

  • QN-2384

    Asset class: Equities Switzerland (Large Caps).
    Asset region: Switzerland.
    Size: CHF 550 – 600 mn.
    Closing date: 2017-12-15.

  • QN-2385

    Asset class: Liability Driven Investment.
    Asset region: Europe.
    Size: Size: EUR 1 Billion, Liability size: EUR 3 Billion.
    Closing date: 2018-01-08.

  • QN-2386

    Asset class: Fixed income.
    Asset region: Global developed markets.
    Size: CHF 500 -1000m.
    Closing date: 2018-01-15.

  • DS-2392

    Closing date: 2017-12-21.

  • QN-2393

    Asset class: All/Large Cap Equities.
    Asset region: Europe.
    Size: EUR 200m.
    Closing date: 2017-12-21.

  • QN-2394

    Asset class: Real Estate Industrial.
    Asset region: Europe.
    Size: EUR 10m.
    Closing date: 2018-01-04.

Begin Your Search Here