An IQ Test For The "Smart Money"

This report explores four classes of stock selection signals associated with institutional ownership (‘IO’): Ownership Level, Ownership Breadth, Change in Ownership Level and Ownership Dynamics. It then segments these signals by classes of institutions: Hedge Funds, Mutual Funds, Pension Funds, Banks and Insurance Companies. The study confirms many of the findings from earlier work – not only in the U.S., but also in a much broader geographic scope – that Institutional Ownership may have an impact on stock prices. The analysis then builds upon existing literature by further exploring the benefit of blending ‘IO’ signals with traditional fundamental based stock selection signals.

Key Findings:

• Among the four classes of ‘IO’ signals, Ownership Dynamics showed the greatest efficacy, both on the long and the short side, after controlling for value, market, size and momentum, and across all geographies (except Japan).
• In the U.S, ‘IO’ signals showed varying strength depending on the class of the institution. Factors constructed based on Hedge Fund activity yielded the strongest results, while those based on Bank and Insurance Companies holdings yielded the weakest.
• Results generated by larger institutional investors were stronger than those generated by smaller institutional investors, independent of institution type.
• In examining institutional ownership signals among non-U.S. countries/regions, we observed statistically significant return spreads – although the factors are generally less effective compared to those in the U.S.
• Strategies constructed using ownership data generally show low correlation with signals constructed from fundamental data sets in the U.S; blending ‘IO’ signals with fundamental signals improved the annualized information ratio (by 34%), long-only return (by 23%), and long-short return (by 32%) compared to a standalone fundamental strategy among Russell 3000 companies.