Say Hello to FAIL, My New ETF

The stan­dard process for launch­ing these funds is flawed. Here’s a mod­est pro­posal about a bet­ter way to do it.

Financial Planning - - CONTENT - By Al­lan S. Roth

The stan­dard process for launch­ing ex­change-traded funds is flawed. Here’s a mod­est pro­posal about a bet­ter way to do it.

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’ll 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 ETFS 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 have worked some­time in the past.

Step 3: De­velop 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.

In 2016 alone, ETF providers in­tro­duced 247 new prod­ucts, in­clud­ing both ex­change-traded funds and ex­change­traded notes, ac­cord­ing to Seekingal­pha. Un­for­tu­nately, 2016 was also a record year for ETF clos­ings, with 128 ETFS and ETNS go­ing de­funct.

What hap­pened?


I sus­pect that 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, a 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.

Smart beta can go “hor­ri­bly wrong,” warned Rob Arnott, CEO of Re­search Af­fil­i­ates, in a pa­per in early 2016.

Arnott and his col­leagues wrote that we 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 that 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.


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.

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 there will be no wor­ries about hot money rain­ing on my pa­rade.


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.

At one point, some quants found that but­ter pro­duc­tion in Bangladesh had the high­est cor­re­la­tion with U.S. stocks.

For­get back-tested log­i­cal sto­ry­lines and com­pelling sales pitches. In­stead, con­sider how il­log­i­cal the cur­rent process is for launch­ing new funds and shap­ing 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.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.