Why Low Vol Funds Are Bleeding

Sometimes the simplest questions are the most important to ask. Such is the case when I looked at the flows into low-volatility funds last week as part of the regular checkup on ETF fund flows.

While the rest of the ETF market is pulling in money by the fistfuls, low-vol funds, as a group, are actually having a pretty awful year. Through the end of February, they were down $578 million in assets.

The bulk of the outflows happened in two funds that anchor the space, the PowerShares S&P 500 Low Volatility Portfolio (SPLV), which lost $239 million through February, and the iShares Edge MSCI Min Vol USA ETF (USMV), which lost a whopping $702 million.

So what’s going on here? The performance so far hasn’t been awful—with the S&P up 5.94%, SPLV and USMV returned 5.14% and 5.86%, respectively. Not beating, but hardly going backward—which led me to ask, when exactly are these things supposed to ‘work’ versus the market?

It turns out, there’s a lot of research on when certain factors work out there, including solid academic work from folks like Cliff Asness at AQR and the quants at BlackRock. Most of that work talks about theories of why low-vol stocks occasional outperform, with a lot of handwringing about whether low-vol stocks are “expensive.”

Volatility For Low Vol?

But I was more curious about when they might outperform, and whether there’s anything to be learned about investor psychology here. I wanted to just keep it simple (because I’m not an academic), so I just looked at SPLV over the last five years, and how it’s done relative to the actual volatility of the market, proxied by VIX:

Here we’re looking at the monthly returns of SPLV versus SPY. I think that’s a pretty good measurement of “did it work?” The correlation here isn’t perfect, but I think it’s pretty illustrative. Almost every spike in VIX shows a concomitant month where the fund beat the market. I don’t think I even need to include the line for market returns—we all know when VIX tends to spike: when markets go down.

That suggests that low-volatility investment is essentially serving as a kind of long-term bet on volatility itself—you need volatility to go up for the strategy to deliver on the promise.

The problem, of course, is that we’ve been at extremely low VIX levels for many of the last few years, with small, notable exceptions. And that has the effect of “taking the bloom off the rose” of these strategies. So it’s not that they’re underperforming badly, they’re just not delivering on the low-vol promise.

To put the above chart in further context, on the average month where SPLV beat the SPDR S&P 500 ETF Trust (SPY), VIX averaged 18.42. On an average month where it failed, VIX averaged 15.4.

Investors Do Even Worse

The actual news for real investors is much, much worse. Let’s move VIX to the side and just look at investor behavior:

The yellow bars here represent fund flows in and out of SPLV for each month. I wish I could say I was surprised that, almost universally, investors are plowing money in when it works, and pulling it out when it doesn’t.

Those few examples going the wrong way—like in late 2011—are just as painful, because you can see the money chasing last month’s performance. But I’m not surprised, because we’ve been to this rodeo before (and I wrote about it a few months ago).

To put this in context further: On an average month where SPLV “worked,” it pulled in $10.6 million. On an average bad month, investors pulled out $1.3 million.

Conclusion

Maybe it’s a giant “duh” that low-volatility strategies need high-volatility markets to do well, but I’ve honestly never really thought about it.

The moral of the story is that when you’re considering any factor investment—whether that’s a value fund or a momentum fund, or anything in between—ask yourself when and why you expect it to work.

No factor or strategy always works. But if you don’t know when and why, maybe it’s time to have second thoughts.

At the time of writing, the author held none of the securities mentioned. You can contact Dave Nadig at [email protected].