How do mutual funds and hedge funds change their stock holdings in response to a sharp market crash? In their July 2020 paper entitled “Where Do Institutional Investors Seek Shelter when Disaster Strikes? Evidence from COVID-19”, Simon Glossner, Pedro Matos, Stefano Ramelli and Alexander Wagner analyze changes in institutional and retail stock holdings during the first quarter of 2020. Using a February-March 2020 snapshot of returns and firm accounting data for non-financial stocks in the Russell 3000 Index, institutional holdings of these stocks as percentages of shares outstanding during the fourth quarter of 2018 through the first quarter of 2020, and number of Robinhood clients (representing retail investors) holding these stocks on December 31, 2019 and March 31, 2020, they find that:
- Stocks in the sample have average 80% institutional ownership at the end of 2019 and average cumulative return -39% during the crash and .
- After controlling for firm characteristics and industry, stock return during the crash relates negatively to institutional ownership. In other words, higher institutional ownership of a stock translates to a more negative return.
- Controlling for overall institutional ownership, stocks held disproportionately by active funds perform relatively worse during the crash than stocks held by passive funds.
- During the crash, mutual funds, investment advisors and pension funds shift toward stocks with relatively low debt and high cash, while hedge funds sell indiscriminately.
- Retail investors are on the other sides of institutional changes in holdings.
In summary, evidence suggests that stocks with high institutional ownership fare worse during a crash, and most institutions shift toward low-debt/high-cash firms during a crash.
Viewed differently, preemptively holding low-debt/high cash stocks may suppress stock portfolio drawdowns. And, mimicking holdings of active funds may expose a portfolio to deep drawdowns.
Cautions regarding findings include:
- The COVID-19 stock market crash may not be representative of future crashes.
- The study does not relate findings to fund-level returns and drawdowns.
