ECML: Costly AI-Driven ETF Producing Inconsistent Results

2025-05-13 17:06:35 ET

Summary

• Euclidean Fundamental Value ETF is designed based on the output of an AI model that uses machine learning to predict future earnings per share for a universe of roughly 1,600 U.S. companies.
• The ECML goal is to select the most deeply discounted 60-70 non-financial stocks that aren't "value traps." Fortunately, I found its value features excellent, and ECML has a 10.92x forward P/E.
• However, its growth and momentum features are poor, and I expect the strategy to produce inconsistent results. Lately, it's been one of the few value ETFs to deliver a loss.
• The current output also suggests its AI model is highly reliant on historical results to select stocks, possibly adopting a contrarian mindset that doesn't consider changing market environments.
• ECML's high 0.95% expense ratio is also a drawback and will put long-term shareholders at a significant disadvantage. Therefore, while interesting, I don't recommend readers buy ECML.

Investment Thesis

While 80% of value-themed ETFs managed positive returns over the last year, the Euclidean Fundamental Value ETF ( ECML ) struggled with a 7.70% loss through April 2025, leading to big questions about its AI model's capabilities. Designed to rotate into 60-70 deeply discounted U.S. securities each quarter while keeping a long-term outlook, ECML currently has an attractive 10.92x forward P/E but overweights risky homebuilding stocks, which is a strange choice, given declining consumer sentiment....

Read the full article on Seeking Alpha

For further details see: