ETF flows could offer clues about investor sentiment

Do exchange-traded product flows provide a signal for investors?

Unlike mutual fund flows, which are reported monthly, ETP flows can be discerned daily and are seen by some as representing the trading views of institutional investors.

While other data points are helpful for spotting anomalies or trends around individual products, flows can be compared in aggregate or individually, by fund sponsor or index provider, by size or sector, by asset class and geography.

“Watching fund flows between comparable products can be instructive,” said David Nadig, president of ETF analytics for San Francisco-based IndexUniverse LLC. He points to the “classic example” of assets moving from an iShares emerging markets ETF to a Vanguard ETF tracking the same index in 2010. (iShares has since introduced a more complete product to replicate the MSCI Emerging Markets Index and Vanguard has switched its offering to track the FTSE emerging markets index.)

But Mr. Nadig balks at the suggestion that fund flow data can be traded on. In order to do that, “you have to be able to separate investor sentiment with non-investment activity in the primary market,” he said.

The primary market, which is overseen by brokers known as authorized participants, is where cash and securities actually flow into or out of the ETP. The vast majority of ETP trading is conducted in the secondary market, with investors swapping existing shares.

New shares of an ETP can be created (or redeemed) for many reasons. Arbitrageurs spot pricing anomalies to the futures or underlying securities. Brokers “create to lend.” Others may use ETFs in hedging.

Research papers over the past several years have attempted to discern if daily flows on an ETP could not only predict returns, but also directly influence daily results. “Can you torture the data enough to extract value?” asked Mr. Nadig. Not including the noise already mentioned, he also pointed out that “it is extremely buggy data.”

“There's an indeterminate lag in reporting and it's unregulated. There's nothing on the line for being wrong,” he said.

But, across the several financial data firms aggregating flow data — and piping data to algorithmic traders — there may yet be enough information in ETP flows to discern investor sentiment and risk appetite.

Market sentiment indexes are numerous, drawn from investor surveys, options activity, recommended asset allocations and even analyst research.

Mutual fund flow data also has been stretched and prodded, but the nature of these flows — largely from defined contribution plans — has been shown to be extremely backward. According to data from the Investment Company Institute, flows for equity mutual funds have been minimal to down since the nadir of the financial crisis, despite the market rally.

“Mutual funds are primarily retail,” said Michael Rawson, a fund analyst at Morningstar Inc. in Chicago.

“Those flows reflect average investor sentiment rather than institutional behavior.”

Attempting to capture that sentiment in institutional behavior, BlackRock Inc., the issuer of iShares ETFs, recently has begun publishing the BlackRock Risk Sentiment Measure, based on monthly flow data from all U.S.-listed ETPs.

The data series, going back to 2006, uses 60-day realized return volatility to divide ETPs into high and low risk. The sentiment figure is determined by subtracting low-risk ETP flows from those of high risk ETPs. The measure is zero-scaled to the one-year moving average. A positive figure represents relative risk seeking behavior.

Using 60-day realized return volatility means that sectors and asset classes can shift from being high risk to low risk across measurement periods, regardless of their perceived risk characteristics.

“We're looking to add more context to raw flows,” said Ananth Madhavan, managing director and global head of research for iShares in San Francisco, “The verdict is still out.”

Moreover, this measure of sentiment runs counter to some of the conclusions drawn from standard flow data.

For example, in July, the sentiment measure indicated risk aversion in equities, even as nearly $39 billion flowed into U.S. and international equity ETFs. (August didn't disappoint, with major U.S. and international equity indexes falling.)

In August, the sentiment measure indicated a significant spike in equity risk tolerance, even as net equity outflows topped $13 billion. (Again, September's start has rewarded the risk takers.) Fixed-income risk tolerance also re-emerged for the first time since January.