In an instant, a lively conversation is at full tilt, with people talking about the virtues of data sets, peer reviews and academic papers. The diverse group is connected by a common thread: They are part of State Street Global Advisors’ Active Quantitative Equity (AQE) team, which uses a data-driven approach to scour the market for alpha. This call isn’t simply a weekly check-in. The team is in the midst of “Idea Season,” for which members generated 140 new ideas — culled from academic journals, client conversations and their own market and portfolio observations — to evolve the group’s investment process, which is responsible for managing US$32.7 billion in assets.1
Understanding the importance of Idea Season requires some insight into how AQE differs from other forms of investing. “Many people think investing is all about figuring out which companies or indexes offer the best returns,” says Olivia, head of the AQE team. “Active quant is focused on what is the best way to put together a portfolio of stocks to deliver the best risk-adjusted returns.”
While AQE employs a similar risk focus and the same attention to detail that State Street Global Advisors honed with its index funds, it requires a different set of skills, she says. “Fundamental managers are often experts on individual companies,” Olivia continues. “They know about their leadership, their market and business practices. My team’s expertise is in applying investment insights across as many companies as possible in a highly objective way.”
Doing so requires an assortment of tools, technology and, especially, data. “Reams of it,” says Olivia. She’s not kidding. Every day, State Street Global Advisors’ engine crunches a dizzying amount of data and thousands of lines of code to generate expected returns for around 17,000 stocks — allowing the team to assess companies and markets more quickly and with more granularity than ever before. But getting an edge in quantitative investing also demands that the team be constantly testing, evolving and improving the model it uses to invest client funds. This requires intuition, Olivia says, but not without a strong investment and economic rationale, backed by data and tested through scientific experiment. This is where Idea Season comes into play.
Searching for a Signal
On today’s call, Anna Lester is updating the team on one of the most promising ideas for new signals to be put into the overall model: whether the way a company addresses environmental, social and governance (ESG) issues impacts its long-term financial sustainability. The issue was the top vote-getter at this year’s annual AQE summit, where ideas are pitched, debated, graded and voted on.
The need for such a signal is clear. As information has become more widely available, investment firms have fewer and fewer quality metrics that can differentiate winners from losers. “Investors have mined financial statements to death,” says Anna, a portfolio manager responsible for the team’s US investment strategies. “So, you need to start looking elsewhere for signals on how stocks might perform.”
The emerging field of ESG investing has been a hot topic for a while. “For several years it’s been one of our ideas,” says Anna. “But the data wasn’t great.” And that was a problem, Anna says. “Quants are very picky about data. But with ESG, due to the nature of the things we’re trying to measure — climate impact, social policies — it can be hard to get the level of rigor needed to provide sufficient evidence that the idea has value.”
What changed? While many have noted the rise of socially responsible investing, Anna suggests something more is going on. To illustrate, she points to scandals such as one involving a pork producer that mistreated its animals:2
“It was a big deal in the press,” she says. “And in response, the company completely changed their processes and became a model of good animal welfare policies.” She goes on to describe how companies that are mindful of such issues by their nature are less likely to have a PR fiasco like that one — and hence yield lower risk. The heightened scrutiny has created a trickle-down effect on emerging markets. “Smaller companies find themselves needing to prove their environmental and social bona fides to even supply large companies,” Anna says.
The rapid maturation of the ESG space has led to investors asking more and more questions, resulting in a competitive industry of data providers, with agencies like MSCI, Thomson Reuters and many other smaller firms providing specific and targeted information by industry. The vast improvement in data, Anna says, made it possible for the AQE team to take on the topic of ESG this year. “For us, it’s about identifying intangibles about a company that can be hard to get out of financial metrics. That’s why I really believe ESG is going to be a source of returns going forward.”
From Hunch to Hypothesis
But in quant investing, having a hunch isn’t enough. Despite there being a strong investment and economic rationale that ESG might be related to financial performance, the theorem had yet to be proven. For the AQE team, this meant designing a research experiment involving rigorous testing and experimentation — almost like a drug approval process.
Anna began by assembling a 10-person working group with a diverse array of experience and expertise. “The idea wasn’t to put a bunch of ESG experts in the room,” she says, “but to use our skills as quantitative investors to build a solid ESG signal that is backed up by rigorous testing.” Rather than diving right into number crunching, the team began with a series of longer discussions about ESG and its impact on the markets. These discussions led to an investment hypothesis: ESG attributes are linked to future financial performance because firms that lag on ESG metrics are more likely to be involved in scandals or other improprieties. The hypothesis acts as an anchor for the team’s research — will the research support or refute this claim?
The team got a head start on research because they were able to leverage the extensive ESG experience of State Street Global Advisors’ Asset Stewardship group, which regularly monitors and engages with thousands of companies and provided valuable insights. The team then fanned out to learn more, with one portfolio manager attending ESG conferences and reporting back to the group, while another combed through ESG academic papers. One of the researchers ran all of the back tests looking at efficacy of a factor in predicting future returns, while another was charged with overseeing simulations — using the factor to build a historical portfolio, and evaluate its characteristics and performance. Even though the team was spread across Boston, London and Sydney, they met twice a week by phone. “Three a.m. phone calls were pretty rough on our researcher in Australia,” Anna admits. “But they helped maintain momentum.”
One of the first challenges they tackled was identifying the right data sets. They quickly homed in on a few well-established and respected data vendors. Before choosing a provider, the team analysed the quality, coverage, history and methodology, asking candidates detailed questions over a series of calls. From there, the team considered a few different ways of testing their idea. Some approaches seemed promising at first but turned out to be a bust when team members dug into the data. Others were deemed too challenging to execute. Eventually, the team landed on measuring how material ESG issues are according to their industry. The ability to tailor ESG evaluations by industry, Anna says — for instance, a utility’s impact on climate — is an area where academics had interesting results, and the team believed it was an important area to explore further.
Tracking a series of metrics that data vendors, sell-side firms, nonprofits and other entities have identified as relevant to each industry, Anna created an elaborate directory from which she saved “materiality maps” to guide the work ahead. Charging each member of the working group with becoming an expert in two or three industries, Anna assigned each with the job of surveying the common themes for their industries across the different materiality maps. “Our goal,” she says, “was to identify common themes across the different ESG experts and not just use our personal judgment.”
How does this work in practice? “I was our team’s ‘real estate expert,’ ” she says. “So, it was my job to identify which ESG issues were relevant for that industry. I didn’t make these decisions based on my knowledge of the industry. Rather, it was led by the materiality maps we had developed, which had put a lot of emphasis on the environmental issues.” Out of this work, the team assembled its own proprietary “mega” materiality map to create a signal they could test. With that, the model was ready to be built using data from their data provider.
“Ideally,” Anna says, “you want a long history of data that uses a consistent framework to judge.” But the best ESG data goes back to only 2009, so the team had to settle. The team ran numerous analyses. “We would build it,” Anna says, “then we’d look at the results to make sure they were consistent with the hypothesis and make adjustments as needed.” Once the team was satisfied, they tested it again, this time using the out-of-sample data. Why split data like this? “Data mining,” Anna says. “The chance of a false signal working on out-of-sample data is pretty small. If the model still works, we have reduced the chances that our own biases have skewed the results.”
After many months and numerous check-ins over the phone with the broader AQE team, Anna’s working group was ready to present its findings to the Technical Committee — the Idea Season equivalent of defending a Ph.D. dissertation. There, the team demonstrated how the signal they developed functions both as a stand-alone metric and when combined with the team’s overall model.
Satisfied that the signal would help the AQE team avoid riskier stocks and improve fund performance when put in the context of overall stock selection, the committee gave Anna’s signal the seal of approval. It’s one of only a handful of signals this year that will be written into the code and become part of the team’s daily investment process, helping forecast returns the larger AQE team generates for every stock, portfolio rebalance and trade.
Sitting in her office, Anna reflects on the experience. “Managing money is very intuitive. But this process forces you to take all of your ideas that intuitively ‘feel right’ and put them down on paper. For every factor we use in the model, we can tell you exactly why it’s in there.” In this sense, she says, the team’s ever-evolving investment model is so much more than an algorithm. “It’s really a living organism — an expression of our investment philosophy, converted into a rigorous process.”