58 Pages Posted: 22 Mar 2009
Date Written: March 18, 2009
We test the Miller (1977) prediction about IPO overvaluation in a sample of 7,212 U.S. IPOs from 1980 to 2003. According to Miller (1977), owing to the lack of short sales prior to the offer, IPO prices are determined by the most optimistic investors, leading to IPO overvaluation; this initial overvaluation is corrected in the years after the offer as the uncertainty of IPO firms subdues. We hypothesize that investors categorize IPOs by their industries, and we use an industry-level uncertainty measure - industry volatility in excess of the market volatility - as an ex ante proxy for investor heterogeneity of beliefs. We therefore predict that IPOs in volatile industries should have high initial returns, followed by low long-run returns.
Generally, IPOs in industries with high excess volatility have much higher initial returns and lower long-run returns in the following 2-3 years than IPOs in industries with low volatility. In time-series tests, we find that excess industry volatility explains half of the annual time-series variation in aggregate initial returns. In cross-sectional tests, when we sort IPOs into three portfolios by excess industry volatility, we find that over the full sample period, high excess industry volatility IPOs have 18% (7%) higher mean (median) initial returns, and 18% (31%) lower mean (median) style-adjusted returns over a 1.5-year period after the offer (skipping the initial 6-month aftermarket lockup period) than low excess industry volatility IPOs. These results are robust to various risk-adjustment procedures.
The effect of excess industry volatility on initial returns is about four times stronger during the "tech bubble" period (01/1997 - 03/2000) than during other periods, and the extreme bubble-period IPO initial returns in volatile industries are not reversed until about 4 years after the offer, consistent with the late 1990s' market being a market bubble fueled by investor euphoria.
We also find that firm age - our firm-level proxy for uncertainty and investor heterogeneity -affects both short- and long-run IPO performance, in coalition with industry volatility. Our results suggest that investor opinions are more diverse among young firms, as well as among firms in volatile industries. Adding age to industry volatility strengthens the return predictability of investor heterogeneity.
These findings lend strong support to Miller's hypothesis that in markets with restricted short-selling, valuations tend to reflect the most optimistic investor's appraisal in the short-run, and revert to the average appraisal in the long-run.
Keywords: Investor heterogeneity; Divergence of opinion, IPOs, Overvaluation,; Industry volatility, Limited attention, Category investment
JEL Classification: , G12, G14, G32
Suggested Citation: Suggested Citation
Dong, Ming and Michel, Jean-Sebastien, Does Investor Heterogeneity Lead to IPO Overvaluation? (March 18, 2009). Available at SSRN: https://ssrn.com/abstract=1364059 or http://dx.doi.org/10.2139/ssrn.1364059