Thematic fund survival rates improve as providers look for ways to mine nuggets from social media firehose
While thematic ETFs are by their nature more prone to having short lives – aside from facing heightened concentration risk, their targeted focus makes them more likely to be a miss than a hit among investors – the expanded use of big data and artificial intelligence could help those products broaden their reach.
Citing data from TrackInsight, a report from the Financial Times highlighted how the rates of survival of thematic ETFs have improved recently. While they are still elevated relative to the broader industry – they’ve registered an 8% closure rate this year compared to just over 5% for all global ETFs – that figure still improves over the thematic death toll of just over 10% last year, and over 14% in 2017.
“Size does matter,” ETFGI founder Deborah Fuhr told the Times. She noted that even thematic ETFs built on the best big ideas would find it hard to stay afloat without having an AUM of US$100 million. Anything below that, and the fund will be more vulnerable to reputational risk a news of a constituent security not being aligned with its broadly stated focus could trigger a fatal rush of divestment.
Reaching that asset threshold appears to have gotten easier for thematic funds as they reach a growing number of investors with an appetite for new ideas. Supporting this trend, she said, is the availability of alternative big data, which is transforming the research underpinning those niche funds as well as providers’ ability to market those products.
With more managers casting their eyes at alternative information, non-traditional data providers are becoming more sought-after than ever. That includes Truvalue Labs, which furnishes fund managers and asset managers with insights informed by a variety of sources including legal documents, media and government announcements, and post from trusted individuals on social media.
“You have to really mine the nuggets,” said Susan Lundquist, chief marketing officer at Truvalue, who said the company has found there to be “too much noise” in uncurated data from social media.
Echoing that view was Claire Smith, founder of Beyond Investing, whose VEGN ETF tracks an index of large-cap U.S. stocks screened based on vegan and environmental principles. She told the Times that human researchers at her firm monitor the “Twitter fire house” for clues and “data to support what we’re hearing.”
Last year, her firm attempted to use social media marketing to promote their ETF, but that effort left a bad taste in the mouth as algorithms selected prospects based on the word “vegan” but failed to detect the negative views held by some people it had locked onto.
On that front, hedge fund Periscope Capital might have an edge; it uses machine learning and AI to scour social media and other online sources for large-cap U.S. stocks, then finds which ones have the most bullish sentiment behind them. “We know that in some environments sentiment is predictive,” said Periscope Capital CEO Jamie Wise.
At the moment, much of the attention centres on the potential of AI and machine learning to scrape data from online sources. But many see social media targeting to find buyers through social media as the next logical step.
“I would believe with 100 per cent certainty that it’s something that can be done,” Wise said.