A Robot Tried to Fix Value Investing and Ended Up Buying Amazon

(Bloomberg) -- Artificial intelligence has been touted as a solution for everything from ending lines at the checkout to rooting out systematic bias in Wall Street hiring. It was almost inevitable that someone would suggest using it to fix value investing.

The strategy of buying stocks that appear cheap relative to their fundamentals has been struggling for more than a decade, but a South Korean money manager reckons its AI-augmented exchange-traded fund is the answer.

Qraft Technologies filed on Friday to create the Qraft AI-Enhanced U.S. Next Value ETF, ticker NVQ. It says this strategy can revive the factor by estimating a firm’s intangible assets based on financial statements and patent databases.

NVQ posted a simulated return of 13% in the year through July, compared with minus 3% for the S&P 500 Value Index -- but value traditionalists will be shocked by the composition of the portfolio.

The top three holdings of the machine-guided fund in July were Amazon.com Inc., Alphabet Inc. and Facebook Inc. Those are far from the kind of undervalued stocks typically favored by a value strategy. But to Qraft, it’s just value 2.0.

“Intangible assets have become a more important factor in the actual value of the company due to the development of information technology,” founder Hyungsik Kim wrote in an email. “It is easy to tell which of the following is more important in measuring the value of Amazon: warehouses (tangibles) or automated logistics systems (intangibles).”

It’s the rallying cry for many remaining proponents of value: The factor isn’t dead, it’s simply plagued by outdated accounting rules that treat intangible investments such as research as expenses rather than capital.

As a result, knowledge-intensive firms end up with much lower book values and higher costs, which make them look more expensive than they actually are.

The new ETF’s eye-catching backtests also speak to the variety of methods underlying even the best-known equity factors. One study estimated there are well over 3,000 different ways to define a value strategy.

To some quant traders, that kind of performance-chasing flexibility suggests a lack of rigor that means failure in the long haul.

But Qraft can at least point to its AI system’s track record: Its $6.7 million multi-factor U.S. large-cap ETF (QRFT) has beaten both the S&P 500 and the iShares MSCI USA Multifactor ETF (LRGF) over the past year.

©2020 Bloomberg L.P.