Seattle-based investment advisor Euclidean Technologies Management has made its ETF debut with the launch of an actively managed US equity fund that uses artificial intelligence to identify value stocks.

The ETF employs machine learning capabilities to identify companies with superior earnings forecasts relative to enterprise value.

The Euclidean Fundamental Value ETF (ECML US) has been listed on NYSE Arca with an expense ratio of 0.95%.

The fund has come to market with approximately $130 million in assets under management.

Security selection is based on an initial universe of US-listed stocks with market capitalizations greater than $1 billion. Non-US-based companies as well as those classified to the financials sector are removed from the selection pool.

Leveraging machine learning capabilities, Euclidean generates a forecast of next year’s earnings for each company within the investment universe. The firm’s model uses a wide range of company data in its forecast including income and cashflow statements, balance sheet data, short interest, historical stock price change data, industry classification, and market size category.

The model also estimates the uncertainty in its calculations, using this uncertainty to discount earnings such that the final output of the model is a discounted earnings forecast.

Each company is then ranked by how inexpensive they are relative to its discounted forecast of earnings. That is, each company’s “earning yield” is calculated by dividing the model’s discounted forecast of earnings by the firm’s total enterprise value.

Next, the model is used to identify and screen out potential “value traps” which are defined as companies that are considered inexpensive based on current valuation multiples but are likely to have bottom decile price performance over the subsequent year. The model re-ranks the eligible universe, adjusting lower those companies with a higher probability of being value traps.

Euclidean will generally select 60 to 70 stocks with favourable rankings to comprise the ETF’s final portfolio. Each constituent is individually vetted by Euclidean to ensure the model has not received erroneous data about that company or that there exists highly relevant negative information about that company that the model does not have access to (such as announced bankruptcy filing, indictments for financial crimes, or a restatement of financials due to material oversight or misrepresentation).

Constituents in the ETF are typically equally weighted and rebalanced on a quarterly basis.