South Korea takes on the turtle traders
Regular columnist Allan Lane of Algo-Chain runs the rule over the new Qraft AI Enhanced US Large Cap Momentum ETF from Exchange Traded Concepts…
One has to go back to the 1920s to trace the foundations for momentum-based investment strategies as exemplified by the legendary ‘Boy Plunger’, Jesse Livermore. During a period of 25 years he made and lost several fortunes as he analysed the ticker tape that relayed the prices of commodity futures between Wall Street and the brokerage houses in the likes of Kansas City.
Here we are in 2019, and news reaches us of the latest iteration in the never-ending quest to beat the markets by focusing on the momentum factor. In partnership with the US firm Exchange Traded Concepts, based in Oklahoma City, a specialist firm in AI algo design from South Korea has launched the Qraft AI Enhanced US Large Cap Momentum ETF, with ticker AMOM.
As the number of ETFs placing themselves in the AI camp continues to grow, one is reminded of the late 1990s when a plethora of new black box hedge funds joined the party. Over time many ‘Masters of The Universe’ came to realise that selling a product, where you were not prepared to tell your investors what you were actually doing, wasn't the easiest conversation to have. However, when it comes to momentum investing, there is far less of a mystery about the technical details that are employed behind the scenes.
Over the last 12 months or so, some of the better-known firms in the systematic space such as GLG and Winton Capital, have had to address the issue of overcrowding in the momentum space. This ETF designed by Qraft looks to hold 50 large cap US stocks, selected with their own in-house AI driven model, and comes with an annual management fee of 0.75 per cent. Whether Qraft’s quant team are completely aware of this, I would hope so, but it wouldn't be the first time that on launch a systematic strategy just happens to go through one of its rough patches.
Looking at the Qraft web site, one is in awe at the sheer depth of talent as they proudly display their 50 data scientists ready to do battle with the West. To be fair, I could probably name 100 hardcore quants who have studied this very problem for the last 20 years, so it will be of interest to see if the machine learning techniques from South Korea can improve on some of the design flaws that often comes with momentum investing. Why do you think Jesse Livermore went bust more than once?
As I study the factsheet it is with particular interest that I read this fund will invest at least 80 per cent of its capital, but why stop at 20 per cent of the investment potentially sitting in cash? Why not be prepared to go 100 per cent cash? I'm reminded of that fateful day on Friday afternoon on Sept 12th 2008. I was sitting at my Bloomberg terminal with Axel Lomholt, who is now the Head of ETFs, International at Vanguard, and both of us looked in astonishment when it was clear Lehman was going bust and Merrill Lynch might go under as well. Qraft's FAQ document talks of the model being thoroughly tested over the last 10 years. Sorry guys that won't do, unless you back test all the way until 2000, you might skip what could happen should we repeat 2008 or indeed the popping of the Tech Bubble in 2001. Does this sound familiar to anyone?