Western pension funds are growing increasingly aware of the importance of technology and data analysis, both for how they invest and what they invest in. But they are also becoming more wary of the potential pitfalls, like some of their peers in Asia.

Both the Teacher Retirement System of Texas (Texas TRS) and UK defined contribution scheme National Employment Savings Trust (Nest) rely on external managers to invest in fast-evolving areas of artificial intelligence (AI) and so-called big data.

“We’ve tended to let AI and big data be externally driven portfolios,” said Jerry Albright, chief investment officer at Texas TRS, a $151 billion pension fund based in Austin. “We invest with some of the top names in artificial intelligence and big data companies." 

“[Investing in such areas is] very expensive to do internally and very competitive,” he told AsianInvestor. “You have to hire the top-10 graduates from the top-10 schools to be competitive, and we don’t have those capabilities, so we pass that onto our partners to do.”

However, Texas TRS does run quantitative portfolios in-house and has been building those capabilities, Albright said. Indeed, about half of the human resources the institution expects to add as part of a planned big increase in headcount would comprise operations and IT professionals.

Albright has proposed nearly doubling the investment team to around 270 from 150 by 2023.

Certainly, the hunt for technology talent is something AsianInvestor said at the start of the year would be a key hiring trend, with the likes of asset manager BlackRock and Singapore sovereign wealth fund GIC pushing ahead in this area in Asia.

Nest is also eyeing technology-driven investment with interest. John St Hill, the £4 billion ($5.25 billion)-pension fund’s deputy chief investment officer, told AsianInvestor: “We think this is a fascinating area. Lots of people have got involved in the AI and big data space recently.”

NEW RISKS

One key point to be aware of, St Hill noted, is that big data introduces new types of risk into portfolios, such as from legislation.

For instance, he said, if GDPR-type regulation – stricter European rules around how data is used that came into effect in May – were introduced globally, this might have a systemic impact on large data-driven companies in various regions.

John St Hill, Nest

“That could change the way one constructs portfolios at the securities level," he added, describing it as "something that a lot of people haven’t started to think about yet". 

It is an area Nest is looking at with its external managers.

There are also issues around quantitative analysis. While data is now much cheaper to obtain and easier to analyse, St Hill said, that also means it’s much easier to find something by luck that doesn’t actually apply when it is used in the marketplace.

“This means you need have a strong theory base before you starting sifting through data to find an investment idea,” he said.

Certainly, fund houses such as BlackRock and Columbia Threadneedle recognise the mportance of gleaning actionable insight from data, as executives from both firms flagged at AsianInvestor's 2nd Chief Operating Officer Forum in April.

St Hill said the better asset managers will test their investment views using relevant data, whereas “less sophisticated users of data will say ‘we have a lot of data and can search through it to find some interesting patterns’".

“Of course, if you have a billion pieces of data and use them to look for a pattern, you’re bound to find something," he noted. "That doesn’t mean that pattern will continue."

Ultimately, concluded St Hill, fast computers and large data sets are no substitute for smart, well-thought-out and reasoned investing.

Doris Ho, executive director who oversees investments for Hong Kong’s Hospital Authority Provident Fund, takes a similar view, as she told AsianInvestor in May.

“While quant models have improved a lot, including the growing use of machine learning and artificial intelligence, for now I still hold the view that the machine might not always win against the ingenuity of the human brain,” she said.

For example, the turmoil of the 2008 global financial crisis underlined how technology-based models were inadequate to cope with such events.