This is almost completely backwards. Here's an example from the Facebook IPO: http://blogs.reuters.com/felix-salmon/2012/05/21/morgan-stanleys-2-4-billion-faceb

This is almost completely backwards. Here's an example from the Facebook IPO: http://blogs.reuters.com/felix-salmon/2012/05/21/morgan-stanleys-2-4-billion-facebook-short/ http://blogs.reuters.com/felix-salmon/2012/05/21/morgan-stan.... This is how an IPO works: 1. The underwriters talk to potential buyers to gauge how much interest there is and how much they're willing to pay. 2. The underwriters set the IPO price in consultation with the company. Then they buy X shares from the company at the IPO price minus a fixed-percentage underwriter's fee, and sell Y shares at the IPO price to investors. Since the percent fee is negotiated ahead of time, the underwriters have an incentive to price the IPO as high as possible so that their fee is higher. 3. In step 2, Y is typically greater than X, by up to 15%. The underwriters sell more stock than they bought, so they make money if the price falls after the IPO. The reason that the company allows (or encourages) this is so the underwriters can buy back the stock after it opens, at or slightly below the IPO price, and prevent the stock price from falling. So: the underwriters make larger fees if the IPO prices higher, and they make larger trading profits if the price drops following the IPO.