That kind of sounds to good to be true... This [0] seems to suggest that the low-beta anomaly is essentially an artifact of how risk is measured. Toughts?

[0] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2503174