Returns such as those reported in the study reflect choices about time frames and trading rules that in any researcher’s hands could be tweaked to maximize output, noted Vincent Deluard, a strategist with INTL FCStone Financial. But Shah’s study looks like “a great paper” with little evidence of data snooping, he said.
“The fact that the results are not that amazing actually increases my confidence in their findings, it suggests little over-optimization,” Deluard said. “From what I can tell without replicating the work myself, it looks like they were quite diligent.”
A separate paper published this month by S&P Global’s market intelligence unit found evidence that stocks responded to price pressures created by ETF flows, mainly those created by buying and selling of ETFs in the open market. The effect “is transient and of only modest magnitude," though might also enhance returns if included in a risk model, it said.
Unlike active funds where managers spend months researching fundamentals in the quest to pick winners, an ETF typically tracks an index and only buys and sells when the members change. Rebalancings, which happen on a regular basis, can trigger money flows with only the thinnest connection to a company’s fundamentals.
The source of inflows isn’t always intuitive. Often a stock’s demotion from a large-cap index can spur passive buying. It takes a bigger share of the small-cap pool than its old weight in large caps, and the downgrade forces ETFs to purchase it.
Such blind buying tends to push share prices beyond what’s justified by fundamentals, Deutsche Bank says. But the gains and losses later unwind. To exploit the pattern, the firm developed a quantitative model that tracks ETF flows for Russell 3000 companies, and each month it ranks stocks by their performance over the previous 12 months.
Signals from ETF flows have shown low correlation with traditional quantitative factors such as momentum or value, making them “an enhancement” to existing models, according to Deutsche Bank.
Keeping other factors static in a multi-factor model, the strategists found that incorporating ETF flows led to proportionate improvement in performance. The more ETF flows are considered, the better the returns.
“Whether or not it’s a factor is debatable,” Shah said. “But this is an anomaly that we found and it’s worth exploring further.”
This article was provided by Bloomberg News.