'I'd rather turn them into robo cops': Execs from Man Group, Bridgewater, and Schonfeld explain how they're trying to blend humans and machines
Dan DeFrancesco and Bradley Saacks
Aug 6, 2020, 1
Bringing the two disciplines together is a feat in itself — hedge funds have tried and failed many times to get the culture and responsibilities just right.
But both sides need each other, as fundamental analysts struggle to sort through massive amounts of data and quant models malfunction during never-before-seen market conditions brought on by the coronavirus.
"I don't see my job as automating away humans," Man GLG's Paul Chambers told Business Insider. "I'd rather turn them into robo cops or Terminators, a machine-augmented human."
With a degree in physics and experience working as a scientist on airspace-weapon systems, Paul Chambers has the type of background many believe is the future of Wall Street.
And while traditional portfolio managers might fear for their future at the sight of someone like Chambers, the reality couldn't be farther from the truth.
For Chambers, who heads up quantitative investment and research at $26.7 billion hedge fund Man GLG, it's all about bringing the two sides — man and machine — together.
"I don't see my job as automating away humans," Chambers told Business Insider. "I'd rather turn them into robocops or Terminators, a machine-augmented human."
Chambers and Man GLG aren't alone in their efforts.
Across Wall Street, from highly quantitative shops like D.E. Shaw to well-known stockpickers like Steve Cohen, hedge funds are working to find the sweet spot between human and machine.
The push comes following a recognition in recent years that both sides could benefit from learning from each other. Commonly referred to as "quantamental," the strategy has picked up steam in recent years with funds on both sides of the aisle trying their hand at it.
Traditional managers have felt a growing pressure to use cutting-edge technology such as artificial intelligence and machine learning to digest the seemingly endless amount of data made available to Wall Street. Meanwhile, recent events, like March's market downturn driven by the coronavirus epidemic, have been a lesson for quants about the need to diversify their systematic models.
"The two sides are becoming closer to one." said Chambers, who re-joined Man last year to "supercharge" the work already happening at the hedge fund leveraging quant data in the discretionary world.
"Some quant performance has faded as easy trades have been arbed away, so it benefits them to learn what drives different types of companies to get the most from the data, and discretionary managers want to access the data that quants can unlock in order to remain competitive."
No simple solution to bridging the gap
Quantamental might be a topic du jour among hedge funds, but the increased buzz has hardly led to a standardized approach.
Take mutual-fund manager Strategic Global Advisors. Founded in 2005 as the brainchild of Gary Baierl, a former quant at Causeway Capital, and Cynthia Tusan, a portfolio manager at Wells Fargo, the firm attempted a quantamental approach from the get-go.
Brett Gallagher, SGA's president, told Business Insider the firm's strategy starts with a quantitative model built based on 15 factors. The suggested portfolio is then analyzed by fundamental analysts who have the option to kick stocks out — they typically reject 10-20% — but can't add any in.
That's followed by portfolio construction, which is then given a final look by the portfolio managers to make the last tweaks. Adjustments made by humans to the model are tracked and analyzed to recognize commonalities that could be plugged back into the model to improve it.
"I think having the fundamental overlay — we call them the gatekeepers — is really important to what we do," Gallagher said. "Even moreso in this time of craziness where there are a lot of things that may make the quantitative model look a little bit wonky on certain aspects. We have the analysts there who can scoot out the issues."
PanAgora Asset Management, meanwhile, takes the opposite approach. George Mussalli, PanAgora's chief investment officer, told Business Insider the firm was unhappy taking a purely quantitative approach, finding it difficult to get differentiated results in what was quickly becoming a crowded field.
As a result, the fund adjusted its focus to discretionary idea generation that could lean on the quantitative techniques they'd already built up. At it's core, Mussalli said, it's about understanding the fundamental framework of companies and industries and then finding the necessary data to emulate that.
"What you end up with are very differentiated signals, very differentiated approach," he added. "You're using the power of quant and the techniques of breadth, portfolio construction, and risk management to build the best portfolio. But what you get is a portfolio that acts very differently than a lot of other portfolios."
At Ray Dalio's Bridgewater, the largest hedge fund in the world with $160 billion in assets, the process is led by "human understanding," the firm's co-CIO Greg Jensen told Business Insider in an email.
"Nothing in our process is purely 'quantitative' in the sense that it is based on statistical relationships or historical patterns that machines are discovering. Everything is grounded in human understanding, and every part of our process is inspectable to us, so that we can have the computer read our logic back to us in more or less plain English and ensure that our decisions are consistent with that understanding," Jensen wrote.
Still, there are times that the machines can outweigh the people — Dalio himself was overruled in 2018 about a trade involving the US dollar, according to The Wall Street Journal.
Quantamental isn't without its own challenges
There is no denying interest in pursuing quantamental strategies is rising, but the trend hasn't been adopted with open arms by everyone in the space.
Josh Pantony, CEO of Boosted.ai, a fintech that helps fundamental managers use quantitative tools, recounted a particularly difficult sales call.
"The most brutal feedback I ever got from a prospective client was, 'Using your platform would be an admission of failure to my investors,'" he said.
Pantony's experience isn't an outlier. In many cases, fundamental managers view the usage of quantitative techniques as a threat to their job.
"I think it's always something that a true MBA from Harvard who is great at Excel and finance but doesn't know anything about coding worries about in the back of their mind," Daniel Goldberg, founder of advisory and consulting firm Alternative Data Analytics, told Business Insider.
Culture needs to be a key consideration of bringing both sides together. Speaking on a webinar in April, Carson Boneck, the chief data officer at Balyasny Asset Management, said transparency between the two groups was critical. Having "translators," or employees that can work with both quants and portfolio managers, also helped the process.
Balyasny ran into its own trouble with the strategy — the firm cut a 10-person team known as Synthesis after only a year because of poor performance in 2019. Coatue, the Tiger Cub manager known for fundamental stock-picking, rolled out a quant strategy a year ago — but has already returned outside capital after the pandemic overwhelmed its data-driven processes.
Industry sources echoed a similar sentiment, adding that quant teams brought into discretionary firms need to have clear responsibilities that show they are having a direct impact on the firm, as opposed to side projects that won't affect actual returns.
"Nobody likes to be put in a room and told to collaborate," Man GLG's Chambers said. "What you need to do is find a way for both sides to get something out of the process. ... Quants can't be seen as service providers."
To be sure, quant-based shops have their own issues attempting quantamental strategies.
For one, quants tend to take a build-over-buy approach, believing proprietary tech to be a key differentiator. But when it comes to the tools needed for quantamental strategies, they often don't have the same expertise and struggle with the process, Emmanuel Vallod, cofounder and CEO of SumUp Analytics, a natural-language-processing platform that analyzes text, told Business Insider.
The issue of explaining models also tends to be a hangup for quant-focused firms, as they aren't accustomed to having to break down models to the level required by fundamental managers.
"There is a need for explicability of the algorithm that is much, much, much higher than what systematic quants are usually requesting," Sylvain Forté, CEO of SESAMm, a provider of alternative-data analytic tools, told Business Insider "Reconciling the two worlds is not easy."
Whichever side you approach the process from, nearly all agreed it takes time. It's not something that can be expected to work overnight. Constant tweaking of the system is required to find the perfect blend.
Most firms can't resist pursuing quantamental strategies
There will always be those who remain either purely fundamental or quantitative. Quants like Renaissance Technology or stock-pickers like Warren Buffet have been successful for too long to drastically change their ways now.
But for the rest of the industry, a shift towards the middle seems more likely.
"I don't think anyone can sit still," Andrew Fishman, president at Schonfeld Strategic Advisors, told Business Insider.
The true definition of quantamental is probably lost forever as it has "one of those wildly generic Wall Street terms that is so amorphous that it doesn't mean much," according to Ryan Caldwell, chief investment officer and co-founder of Chiron Investment Management, a $1.8 billion manager that was bought by FS Investments at the end of last year.
But, Caldwell says, the evolution is necessary with alpha — returns generated beyond what the market churns out for everyone — becoming harder and harder to find.
Fundamental analysts have been pushed even further away from the research that was once the backbone of MBA programs across the country thanks to alternative data that can track supply chains more efficiently. No longer is it about "shaking info out faster than the guy sitting next to you," Caldwell says.
"The fundamental part of this is understanding the macro environment with which we operate in," he added.
And, during a global pandemic that has introduced never-before-seen volatility and record-low interest rates into the equation, quants can't rely on analyzing past situations to react to today because today is unprecedented.
"We would worry about any approach reliant on historical patterns or quantitative models based on historical data, whether combined with fundamental measures or not, as so much of the current environment and where we're likely headed isn't in the data, and instead requires a logical understanding of the economic machine," Bridgewater's Jensen wrote.
Former Goldman Sachs PM Mike Ho put an even finer point on the future. Quantamental, Ho wrote in a 2017 Medium post, will eventually be like "dating in the 7th grade": Everyone says they're doing it, even if they're not. Ho, who ran a firm called Quantavista for a time, is now at Steve Cohen's Point72.
"We suspect quantamental may become an official word in the coming years, too. But, by then we'll probably all drop the 'quantamental' term and it will just be the way that most investment analysis is done," Ho wrote.
"After all, why would you not want to combine quantitative and fundamental analysis?"
Dan DeFrancesco and Bradley Saacks
Aug 6, 2020, 1
Quant and fundamental firms are looking to bridge the gap between the two sides in an effort to get the best out of human and machine. Samantha Lee/Business Insider
"Quantamental" is gaining hype on Wall Street. It's a type of investing strategy that tries to blend the best of quant funds and fundamental analysis.
"Quantamental" is gaining hype on Wall Street. It's a type of investing strategy that tries to blend the best of quant funds and fundamental analysis.
Bringing the two disciplines together is a feat in itself — hedge funds have tried and failed many times to get the culture and responsibilities just right.
But both sides need each other, as fundamental analysts struggle to sort through massive amounts of data and quant models malfunction during never-before-seen market conditions brought on by the coronavirus.
"I don't see my job as automating away humans," Man GLG's Paul Chambers told Business Insider. "I'd rather turn them into robo cops or Terminators, a machine-augmented human."
With a degree in physics and experience working as a scientist on airspace-weapon systems, Paul Chambers has the type of background many believe is the future of Wall Street.
And while traditional portfolio managers might fear for their future at the sight of someone like Chambers, the reality couldn't be farther from the truth.
For Chambers, who heads up quantitative investment and research at $26.7 billion hedge fund Man GLG, it's all about bringing the two sides — man and machine — together.
"I don't see my job as automating away humans," Chambers told Business Insider. "I'd rather turn them into robocops or Terminators, a machine-augmented human."
Chambers and Man GLG aren't alone in their efforts.
Across Wall Street, from highly quantitative shops like D.E. Shaw to well-known stockpickers like Steve Cohen, hedge funds are working to find the sweet spot between human and machine.
The push comes following a recognition in recent years that both sides could benefit from learning from each other. Commonly referred to as "quantamental," the strategy has picked up steam in recent years with funds on both sides of the aisle trying their hand at it.
Traditional managers have felt a growing pressure to use cutting-edge technology such as artificial intelligence and machine learning to digest the seemingly endless amount of data made available to Wall Street. Meanwhile, recent events, like March's market downturn driven by the coronavirus epidemic, have been a lesson for quants about the need to diversify their systematic models.
"The two sides are becoming closer to one." said Chambers, who re-joined Man last year to "supercharge" the work already happening at the hedge fund leveraging quant data in the discretionary world.
"Some quant performance has faded as easy trades have been arbed away, so it benefits them to learn what drives different types of companies to get the most from the data, and discretionary managers want to access the data that quants can unlock in order to remain competitive."
No simple solution to bridging the gap
Quantamental might be a topic du jour among hedge funds, but the increased buzz has hardly led to a standardized approach.
Take mutual-fund manager Strategic Global Advisors. Founded in 2005 as the brainchild of Gary Baierl, a former quant at Causeway Capital, and Cynthia Tusan, a portfolio manager at Wells Fargo, the firm attempted a quantamental approach from the get-go.
Brett Gallagher, SGA's president, told Business Insider the firm's strategy starts with a quantitative model built based on 15 factors. The suggested portfolio is then analyzed by fundamental analysts who have the option to kick stocks out — they typically reject 10-20% — but can't add any in.
That's followed by portfolio construction, which is then given a final look by the portfolio managers to make the last tweaks. Adjustments made by humans to the model are tracked and analyzed to recognize commonalities that could be plugged back into the model to improve it.
"I think having the fundamental overlay — we call them the gatekeepers — is really important to what we do," Gallagher said. "Even moreso in this time of craziness where there are a lot of things that may make the quantitative model look a little bit wonky on certain aspects. We have the analysts there who can scoot out the issues."
PanAgora Asset Management, meanwhile, takes the opposite approach. George Mussalli, PanAgora's chief investment officer, told Business Insider the firm was unhappy taking a purely quantitative approach, finding it difficult to get differentiated results in what was quickly becoming a crowded field.
As a result, the fund adjusted its focus to discretionary idea generation that could lean on the quantitative techniques they'd already built up. At it's core, Mussalli said, it's about understanding the fundamental framework of companies and industries and then finding the necessary data to emulate that.
"What you end up with are very differentiated signals, very differentiated approach," he added. "You're using the power of quant and the techniques of breadth, portfolio construction, and risk management to build the best portfolio. But what you get is a portfolio that acts very differently than a lot of other portfolios."
At Ray Dalio's Bridgewater, the largest hedge fund in the world with $160 billion in assets, the process is led by "human understanding," the firm's co-CIO Greg Jensen told Business Insider in an email.
"Nothing in our process is purely 'quantitative' in the sense that it is based on statistical relationships or historical patterns that machines are discovering. Everything is grounded in human understanding, and every part of our process is inspectable to us, so that we can have the computer read our logic back to us in more or less plain English and ensure that our decisions are consistent with that understanding," Jensen wrote.
Still, there are times that the machines can outweigh the people — Dalio himself was overruled in 2018 about a trade involving the US dollar, according to The Wall Street Journal.
Quantamental isn't without its own challenges
There is no denying interest in pursuing quantamental strategies is rising, but the trend hasn't been adopted with open arms by everyone in the space.
Josh Pantony, CEO of Boosted.ai, a fintech that helps fundamental managers use quantitative tools, recounted a particularly difficult sales call.
"The most brutal feedback I ever got from a prospective client was, 'Using your platform would be an admission of failure to my investors,'" he said.
Pantony's experience isn't an outlier. In many cases, fundamental managers view the usage of quantitative techniques as a threat to their job.
"I think it's always something that a true MBA from Harvard who is great at Excel and finance but doesn't know anything about coding worries about in the back of their mind," Daniel Goldberg, founder of advisory and consulting firm Alternative Data Analytics, told Business Insider.
Culture needs to be a key consideration of bringing both sides together. Speaking on a webinar in April, Carson Boneck, the chief data officer at Balyasny Asset Management, said transparency between the two groups was critical. Having "translators," or employees that can work with both quants and portfolio managers, also helped the process.
Balyasny ran into its own trouble with the strategy — the firm cut a 10-person team known as Synthesis after only a year because of poor performance in 2019. Coatue, the Tiger Cub manager known for fundamental stock-picking, rolled out a quant strategy a year ago — but has already returned outside capital after the pandemic overwhelmed its data-driven processes.
Industry sources echoed a similar sentiment, adding that quant teams brought into discretionary firms need to have clear responsibilities that show they are having a direct impact on the firm, as opposed to side projects that won't affect actual returns.
"Nobody likes to be put in a room and told to collaborate," Man GLG's Chambers said. "What you need to do is find a way for both sides to get something out of the process. ... Quants can't be seen as service providers."
To be sure, quant-based shops have their own issues attempting quantamental strategies.
For one, quants tend to take a build-over-buy approach, believing proprietary tech to be a key differentiator. But when it comes to the tools needed for quantamental strategies, they often don't have the same expertise and struggle with the process, Emmanuel Vallod, cofounder and CEO of SumUp Analytics, a natural-language-processing platform that analyzes text, told Business Insider.
The issue of explaining models also tends to be a hangup for quant-focused firms, as they aren't accustomed to having to break down models to the level required by fundamental managers.
"There is a need for explicability of the algorithm that is much, much, much higher than what systematic quants are usually requesting," Sylvain Forté, CEO of SESAMm, a provider of alternative-data analytic tools, told Business Insider "Reconciling the two worlds is not easy."
Whichever side you approach the process from, nearly all agreed it takes time. It's not something that can be expected to work overnight. Constant tweaking of the system is required to find the perfect blend.
Most firms can't resist pursuing quantamental strategies
There will always be those who remain either purely fundamental or quantitative. Quants like Renaissance Technology or stock-pickers like Warren Buffet have been successful for too long to drastically change their ways now.
But for the rest of the industry, a shift towards the middle seems more likely.
"I don't think anyone can sit still," Andrew Fishman, president at Schonfeld Strategic Advisors, told Business Insider.
The true definition of quantamental is probably lost forever as it has "one of those wildly generic Wall Street terms that is so amorphous that it doesn't mean much," according to Ryan Caldwell, chief investment officer and co-founder of Chiron Investment Management, a $1.8 billion manager that was bought by FS Investments at the end of last year.
But, Caldwell says, the evolution is necessary with alpha — returns generated beyond what the market churns out for everyone — becoming harder and harder to find.
Fundamental analysts have been pushed even further away from the research that was once the backbone of MBA programs across the country thanks to alternative data that can track supply chains more efficiently. No longer is it about "shaking info out faster than the guy sitting next to you," Caldwell says.
"The fundamental part of this is understanding the macro environment with which we operate in," he added.
And, during a global pandemic that has introduced never-before-seen volatility and record-low interest rates into the equation, quants can't rely on analyzing past situations to react to today because today is unprecedented.
"We would worry about any approach reliant on historical patterns or quantitative models based on historical data, whether combined with fundamental measures or not, as so much of the current environment and where we're likely headed isn't in the data, and instead requires a logical understanding of the economic machine," Bridgewater's Jensen wrote.
Former Goldman Sachs PM Mike Ho put an even finer point on the future. Quantamental, Ho wrote in a 2017 Medium post, will eventually be like "dating in the 7th grade": Everyone says they're doing it, even if they're not. Ho, who ran a firm called Quantavista for a time, is now at Steve Cohen's Point72.
"We suspect quantamental may become an official word in the coming years, too. But, by then we'll probably all drop the 'quantamental' term and it will just be the way that most investment analysis is done," Ho wrote.
"After all, why would you not want to combine quantitative and fundamental analysis?"
No comments:
Post a Comment