X hits on this document

PDF document

Searching for Google’s Value: - page 29 / 44





29 / 44

post-IPO values of stocks. The forecasts were quite accurate for Google even before many aspects of the

issue (e.g., the number of shares, initial price range indications, etc.) were revealed.

What can explain the accuracy of these markets? At one level, given the pervasive underpricing,

one might argue that prediction markets may perform well by simply forecasting a market capitalization

higher than that indicated using preliminary price ranges from the prospectus. However, two pieces of

evidence counter this. First, IEM prices predicted well even before preliminary price ranges and share

quantities were available. Second, shortly after the initial ranges were announced, the IEM prices

predicted a market capitalization near the average of the price range, not above the range, and the

prediction fell long before the price range was revised down. Thus, prediction market traders appear to

do more than simply “mark up” preliminary price ranges from the prospectus. Why might this be

possible? Recent evidence suggests that the degree of underpricing may be predicted from publicly

available information that underwriters and/or companies do not build into prices (e.g., Bradley and

Jordan, 2002, Loughran and Ritter, 2002, and Lowry and Schwert, 2004). Participants in prediction

markets may be able to incorporate this information without the biases and conflicts frequently

hypothesized to affect firms, investment bankers and investors.

One might argue that the results here are weakened because Google is a single IPO. We believe

this is not the case for three reasons. First, we argue that unique features of the Google IPO strengthen

the results. Both the unique features of the Google IPO alone and the outcomes of the prediction market

alone provide interesting evidence on theories of IPO underpricing. When combined, the evidence is

particularly compelling. For example, the ability to estimate the demand curve through information about

the auction allows us to determine whether both Google (the issuer) and outsiders (our market

participates) were both informed about the degree of underpricing, allowing for a more complete analysis

of asymmetric information models. Second, we argue that the evidence here is interesting in spite being

generated by a single IPO because the evidence comes from the evolution of prediction market prices

through the entire IPO process. For example, the evolution of ex ante uncertainty through time and the

correlation of uncertainty with ex ante forecasts of underpricing can shed light on theory even from a


Document info
Document views149
Page views149
Page last viewedThu Jan 19 18:04:12 UTC 2017