First things first—please check out my interview with derivatives professional Devin Anderson on The Monthly Dirtcast. You don’t get this type of intellectual discussion on TV. Just saying.
Now, to smart beta. My first encounter with a smart beta ETF was a WisdomTree product over ten years ago. As you may know, WisdomTree uses index weighting methods other than market capitalization for its ETFs. This one in particular was weighted by earnings or dividends—I can’t remember.
But wait—isn’t the whole point of an index to track the market? Since when are we constructing indices to outperform the market?
This is what is known as “smart beta,” the idea that you can build an index that will provide superior returns at a lower cost, with perhaps reduced volatility.
So how do people come up with these things?
Building a Smart Beta ETF
You start from some economic premise—say, that companies with lower valuations will outperform over time. Next, you build a hypothetical index, and then you go back in time (using the computer) to see if that is true. If you find it to be true, you build an index around it.
I think the whole concept of “backtesting” is one of the funniest things in finance. Mutual fund companies are quick to tell you that past performance is not indicative of future results, but a lot of money plows into smart beta strategies based on past performance!
Unfortunately, you can pretty much backtest anything to see if it outperforms the market. In one of the greatest pieces of financial journalism, ever, Dani Burger at Bloomberg created a hypothetical smart beta ETF that was based on cats.
It outperformed the market by nearly 850,000%.
Ha—I am already long four cats, maybe I should get more!
Some people would call the cat study silly, because there is no economic basis for believing that cat investments will outperform. But that is kind of where we are today in ETF-land. Researchers spend a lot of time backtesting stuff until they find something that works, then build a marketing plan around it.
Backtesting is not forward-testing. I’m predicting headlines in 2-3 years about smart beta ETFs failing to deliver on their promises.
If there is one thing we should have learned from the passive investing revolution, it is that you cannot systematically outperform the market over time. Even if you figured out how to, other people would discover the anomaly and pile into the strategy, driving down returns.
There might be other reasons to invest in smart beta products—say, you really believe in the economic premise of the ETF—but I believe more in the market’s ability to exhibit nonstationarity, rendering all strategies useless over time.
That’s one of the reasons why trading is an exercise in emotional fitness. You have to be willing to switch things up when they’re not working anymore.
Same trend in my ETF survey*: a large percentage of you are sweating over what happens ETF liquidity in the event of a downturn. “Will it disappear? Are ETFs the portfolio insurance of this decade?” And so on.
So, if you buy an ETF, will you be able to get out?
For most ETFs, the answer is yes.
Let’s look at SPY for instance.
In SPY’s case, lots and lots of robots post liquidity—bidding and offering to buy or sell stock “out loud” in SPY, and ready to hedge with either S&P 500 futures or a basket of S&P 500 stock. These robots are arbitrageurs, and are competing with each other to make tiny profits on each share of SPY bought or sold.
The liquidity in SPY is very, very deep. In 2006, I did a trade in SPY that was about $400 million… and it hardly moved the market at all.
If you own shares of SPY, chances are you are going to be able to sell it—unless there is an electromagnetic pulse. In fact, it’s not likely that SPY would deviate much from its net asset value even if someone dumped $10 billion worth on the market.
The same is not true for all ETFs.
It All Comes Down to Arbitrage
There are a dozen different ways to arbitrage SPY: through S&P 500 futures, to options, to the underlying stock, or other derivatives. There is a very complex web of derivatives that holds the liquidity in the complex together. This isn’t true for a lot of ETFs.
There is hardly any liquidity at all in the example we used last week, the Nashville ETF (NASH). As I write, the market is 27.86 bid, at 27.99, 300 x 700 shares up.
Likely, the issuer of NASH has persuaded one lonely market maker to provide liquidity in it. It is probably unprofitable for him to do so. Is there a chance that he could pull the plug on the computer during a crisis? Yes. And then you would be stuck with NASH, and unable to sell it.
That is an extreme example. There are other funds that can be temporarily overwhelmed by volume, usually around the market open, but those intervals of time are pretty short.
In my old job, I would always tell people that the average daily volume you see on the screen is not a good measure of liquidity, because the liquidity of the ETF really depends on the liquidity of the underlying.
But what you care about as an individual investor is the likelihood that you can get out of the trade at a price that approximates the NAV of the fund.
That will be the case 99% of the time in 99% of the ETFs. You could have a temporary dislocation, like the flash crash. As a general rule, anyone who sells in a panic is usually unhappy with the prices later.
Sell when you can, not when you have to.
*The survey’s offline now—thanks to the thousands of you who took it. If you have any questions or comments about this 10th Man ETF series… send them in here.