This table reports regression results of U.S. Treasury daily returns on bond fund ownership around the COVID-19 crisis and the Fed announcement. The dependent variable is a Treasury’s daily return, expressed in basis points. In the first two columns, the sample period spans from February 24th to March 20th in 2020. In the last two columns, the sample period spans from March 9th to April 3rd in 2020. We decompose bond fund ownership into two components: Ownership (High LMI) is the proportion of total market value of a Treasury that is held by high-LMI funds; Ownership (Low LMI) is the proportion of total market value of a Treasury that is held by the rest of the funds. Liquidity management intensity (LMI) is defined as the coefficient estimate from a univariate regression of NetBuy (US Treasuries) on Fund Flow using data from the past 12 quarters, where NetBuy (US Treasuries) is the percentage change of a fund’s total holdings in U.S. Treasuries relative to its beginning of the quarter holdings, and Fund Flow is the quarterly fund flows. We require at least five quarterly observations to estimate this coefficient. The top one-third of funds in each quarter are identified as funds with high-LMI based on this estimate. High Ownership (High LMI) is a dummy variable that equals one for Treasuries with Ownership (High LMI) in the top tercile. High Ownership (Low LMI) is a dummy variable that equals one for Treasuries with Ownership (Low LMI) in the top tercile. After is a dummy variable that equals one for the 2 weeks of crisis period in the first two columns, and the 2 weeks following the Fed intervention in the last two columns. We divide all Treasuries into six maturity buckets (less than 2 years, 2–4 years, 4–6 years, 6–8 years, 8–10 years, 10+ years) and control for maturity-day fixed effects, as well as bond fixed effects. Standard errors are clustered by Treasuries. t-statistics are reported in parentheses.

p < .1;

p < .05;

p < .01.