
July 2002. IPOs issued during January 1991 to December 1998 are included in the
sample. However, we use IPO aftermarket returns three and a half years after
issuance to measure IPO long-run performance. To avoid possible noise in the right
aftermarket, we measure the IPOs’ aftermarket performance from the second trading
month, where one trading month is defined as 25 trading days.
Following Bossaerts and Hillion (2001), our IPO sample is classified into a sub-
sample of winners and a subsample of losers on the basis of a firm’s status on the ref-
erence point to measure IPO performance under an ELM. Winners are those who
survive for at least 43 trading ...