The sample is at the fund-CUSIP level and includes CP (including ABCP) and CDs held by prime MMFs at the end of February with maturity beyond March 31 (i.e., 1 week after the launch date of the MMLF). The dependent variable is the percentage of a security holding by a fund that is pledged at the MMLF during its first 2 weeks of operations, ranging between 0 and 100.

$$log(Time\ to\ Maturity)$$

is the logarithm of the residual days to maturity of the security as of the end of February.

$$Crisis\ \Delta WLA$$

is the net change in the fund’s WLA during the crisis period.

$$Crisis\ fund\ flow$$

is the fund’s cumulative percentage flow during the crisis period from March 9 to March 20.

$$Institutional$$

is a dummy variable equal to one if the prime MMF is an institutional fund. Security-level controls include security yield and share of the security in the fund’s AUM at the end of February. All securities are categorized into four types: nonfinancial CP, financial CP, ABCP, and CDs, and the security type fixed effects are included in all specifications. Fund-level controls include abnormal yield (in excess of cross-sectional average), safe holdings (Treasury and agency debt as share of fund AUM), risky holdings (CP and CDs), logarithms of fund size, expense ratio, bank affiliation dummy, and fund age, as of the most recent Tuesday before the launch of the MMLF. Standard errors (in parentheses) are two-way clustered at the fund and CUSIP levels.