I’d like to develop a new investment fund that has two main selling points: a superior strategy and a particularly memorable ticker symbol.
I call it the FAIL ETF.
Now, to be clear, I haven’t yet developed the specific methodology for this fund just yet. But I am in the process of reinventing how new investment funds are introduced.
First, some review. Here’s how new funds are typically launched.
Step 1: Crunch investment results and compare thousands of factors that may be driving performance. In a world of big data, this data mining process is easy and can take mere seconds.
Step 2: Use the output to identify factors that worked in the past.
Step 3: Craft a compelling storyline around these factors to create excitement around your fund’s brilliant new methodology. Although the fund doesn’t exist yet, it can paint a hypothetical scenario, showing people how much more money they would have earned if only they had bought into the fund being launched.
Step 4: Launch the fund for real, market the brilliant strategy globally and reap the financial rewards.
Last year alone, ETF providers launched 247 new products, including both exchange-traded funds and exchange-traded notes, according to SeekingAlpha. Unfortunately, 2016 was also a record year for ETF closures, with 128 ETFs and ETNs going defunct.
THERE’S A PROBLEM WITH HOW MOST ETFS ARE LAUNCHED
I suspect every one of those funds went through the same process of back-tested research, and each created a compelling storyline as to why it would beat the boring strategies such as old fashioned cap-weighted index funds.
So why do so many new funds quickly close down, then? It all starts with the data. For each 512 random factors reviewed, roughly one will have a 99.9% probability of correlation (either positive or negative).
Correlation is not causation, however, and strong past performance doesn’t usually persist. We know that over time outcomes will tend to even out — a phenomenon known in statistics as regression to the mean.
ETF prices have been driven higher in the short run as investors pour cash into these hot new products. But when regression to the mean occurs, and the outperformance turns into underperformance, investors will flee and cause the strategy to do even worse. The end result? Many of these funds are bound for the ETF graveyard.
According to Ben Johnson, director of global ETF research for Morningstar, the majority of strategic-beta ETFs have failed to deliver over the past one to three years. Strategic beta funds use methodologies other than market-cap weighting. Rob Arnott, chief executive of Research Affiliates, warned early last year in a paper that smart beta can go “horribly wrong.” In the paper, Arnott and his colleagues point out that we now have a “factor zoo” where many funds are being launched purely based on past performance and stand little chance to outperform in the future.
At one point, some quants found that butter production in Bangladesh had the highest correlation with U.S. stocks. I feel confident a fund based on this strategy would have been huge had anyone been able to develop a storyline as to why it should work going forward.
Needless to say, the current process behind new fund launches is badly flawed.
THERE’S A BETTER WAY
For my new FAIL ETF, the process would be far superior to the above. Here’s how it would work.
Step 1: Rather than reach for an explanation to falsely explain why random outperformance occurred, I’ll start by theorizing why a strategy should have outperformed.
Step 2: Next, I’ll back test to assure the strategy failed to outperform in the past.
QuoteAt one point, some quants found that butter production in Bangladesh had the highest correlation with U.S. stocks.
Step 3: I’ll assess the likelihood of regression to the mean and future outperformance.
Step 4: If it looks good, I’ll launch the fund and keep it as quiet as I can. No doubt, the FAIL symbol will minimize inflows to the ETF, so no worries about hot money raining on my parade.
A SERIOUS MESSAGE
FAIL is obviously a fictitious fund that I won’t actually launch. Still, I have a serious message. The current flawed process is yet another way we all chase performance. As silly as it sounds, FAIL would be a much better ETF than the abundance of ETFs being launched based on past performance and phony storylines.
Forget back tested logical storylines and compelling sales pitches. Instead, remember how illogical the current process is for launching new funds and investment strategies. Factors that work include some old standbys: fees, tax efficiency, diversification and rebalancing. Though not as exciting as those hot new ETFs, they have a far greater likelihood of actually working.
Someday, if you see me getting out of my private jet, it’s probably not because I launched FAIL — but rather because I devised my own investment strategy from this process.