X hits on this document

PDF document

Searching for Google’s Value: - page 5 / 44





5 / 44

IPO prediction markets can have a significant practical application as well. If they are accurate

and obvious strategic manipulation problems can be addressed, companies can use the forecasts to set IPO

prices that either (1) avoid underpricing when it is optimal to do so or (2) know in advance and set

optimally the degree of underpricing when it is optimal to underprice (say, in exchange for future

benefits). The overall impact could be substantial. For example, Google’s underpricing left more than

$300 million on the table.3 Prediction markets can help eliminate the underpricing or determine whether

it is the optimal amount.

Our results complement recent research on when-issued trading in German IPOs. Löffler,

Panther and Theissen (2002) and Aussenegg, Pichler and Stomper (2004) both show that when-issued

trading on German IPOs helps forecast post-IPO trading prices.4 From one perspective, our prediction

market evidence is similar: through a market mechanism traders reveal information in advance of the IPO

that forecasts post IPO prices. However, our research differs from the when-issued research in a variety

of important ways. First, our markets provided valuation forecasts before the initial price ranges were set.

German when-issued market runs only after initial price ranges are set, typically one week before the

issue. Löffler, Panther and Theissen (2002, p. 10) report that “pre-IPO prices appear to be unbiased

estimates of the first price established on the exchange only in the second half of the subscription period”

(that is, the last few days). This is important because it severely limits the usefulness of the when-issued

market in setting IPO prices. German IPO price ranges are never adjusted; IPO prices never exceed the

top of the range, prices seldom fall below the bottom and more than half of the IPOs are set at the top of

the range (Aussenegg, Pichler and Stomper, 2004). Thus, the relevant pricing information needs to be

gathered before when-issued trading commences. In contrast, our markets ran for six weeks before the

IPO and two and a half weeks before any initial price ranges or issue quantities were announced. We find

that our market prices were surprisingly accurate even before initial price ranges were set. Second, our

3 4

This statistic comes from Table II, discussed later in the text. Löffler, Panther and Theissen (2002) find that when-issued prices are “highly informative” and Aussenegg, Pichler

and Stomper (2004) find that when-issued prices are indicative but do not “fully supplant information gathering through book building.”


Document info
Document views154
Page views154
Page last viewedFri Jan 20 22:47:53 UTC 2017