This paper examines underwriters' pricing errors and the information conten
t of first-day trading activity in IPOs. We show that first-day winners con
tinue to be winners over the first year, and first-day dogs continue to he
relative dogs. Exceptions are "extra-hot" IPOs, which provide the worst fut
ure performance. We also demonstrate that large, supposedly informed, trade
rs "flip" IPOs that perform the worst in the future. IPOs with low flipping
generate abnormal returns of 1.5 percentage points per month over the firs
t six months beginning on the third day. We show that flipping is predictab
le and conclude that underwriters' pricing errors are intentional.