|dc.description.abstract||The apparent anomaly that initial public offerings (IPOs) are mispriced across time and markets has been a focus of academic research for over four decades. Previous studies on the subject have focused on the underpricing of IPOs. Using a sample of 6171 IPOs that were issued from 1995 to 2013 in four markets – the United States (US - 2458 IPOs), Australia (1095 IPOs), China (2199 IPOs) and Malaysia (419 IPOs) – we also find that IPOs are on average underpriced but that a concentration on the average overstates the extent of this underpricing with a significant proportion of IPOs actually being overpriced. In the US, the mean mispricing is 34.90%, and the median is 2.40%, with 35.50% of IPOs being overpriced. In Australia, the mean mispricing is 25.51%, the median is 10%, and 37.70% of IPOs are overpriced. In China, the mean mispricing is 112.10%, the median is 71.40%, and 6.30% of IPOs are overpriced. In Malaysia, the mean mispricing is 1.80%, the median is -17.53%, and 59.42% of the sample IPOs are overpriced.
A wide dispersion in mispricing exists across the four markets. The IPOs range from highly overpriced to extremely underpriced IPOs. We assemble a large number of firm-level and country-level variables to explain the mispricing, and we show how their impacts vary across the range of mispricing and across the four markets. The firm-level factors examined include company characteristics, offer characteristics, issue certification, prospectus disclosure, market sentiment, and institutional characteristics. These factors are found to have a varying impacts across different levels of mispricing.
Country-level variables include institutional quality and economic strength. Our findings suggest that poor institutional quality adds to the uncertainty about the value of the firm and leads to more mispricing.
Our cross-country examination of mispricing is one of the first studies to examine the relationship between a country’s economic strength and IPO mispricing. We find that the largest mispricing occurs in developing countries experiencing high economic growth, and that larger economic size reduces mispricing. We further find that, while the country-level characteristics are differentiating factors across our sample markets, they have their greatest explanatory power for moderate levels of mispricing, and that extreme levels of mispricing are better explained by firm-level factors.
We use the traditional ordinary least squares (OLS) and the more appropriate quantile regression (QR) methods as our methods of examination. The OLS approach focuses on the average impact that the independent variables have on mispricing. In this approach, the latent characteristics of the mispricing distribution, given that it does not follow a normal distribution, remain unexamined. On the other hand, the QR approach allows us to examine the varying effects that the independent variables have at different levels of mispricing due to the asymmetric distribution of returns. The QR approach enables us to identify the impacts of each variable on IPOs at particular levels of mispricing. The QR approach is a robust method which is able to deal with potential heterogeneity in the distribution, as was the case with our sample. We are able to compare the results derived from the median QR with those derived from the traditional OLS to enrich the literature in terms of the analysis of a skewed distribution. The QR method also enables us to examine different segments of the distribution of mispricing, including the tail regions. By doing this we are able to compare the impacts of the explanatory factors on mispriced IPOs that range from extremely overpriced to extremely underpriced.||