Chinese IPO activity, pricing, and market cycles

There are two types of stocks in China. The A shares are denoted in Chinese currency for domestic investors and the B shares are denoted in U.S. or Hong Kong dollars for foreign investors. Since February 2001, Chinese investors can invest in B shares with foreign currencies. This paper focuses on the A share IPO market since it is much larger and more active than the B share IPO market. Chinese A share IPO activity started in the second half of 1991 and it was temperately suspended in the second half of 2005. The activity was resumed in May 2006.

Gross proceeds and money left on the table are not adjusted for inflation.

Ritter and Welch (2002) exclude bank IPOs and IPOs with an offer price less than $5.00. In this study, we include all IPOs for several reasons. First, we would like to examine the IPO activity for the entire sample and compare the results with those from earlier studies that use smaller samples. Second, there are 9 bank IPOs and 102 IPOs with an offer price of one Yuan (the lowest offer price), which is about 8% of the entire sample. Third, the offer price in China is generally low. Nevertheless, we redo the analysis after excluding bank IPOs and IPOs with an offer price of one Yuan and find that the average initial return drops to 151.8% from 238.0%. Other results remain similar.

It is an updated version of the information contained in Loughran et al. (1994), obtained from Ritter’s website.

Majority of Chinese A share IPOs are partial privatisation IPOs over the sample period. In general, there are three major groups of shareholders for each A share IPO: the government and its agency, other legal entities that include institutional investors, and individual investors, each with about one-third of the holdings. Only shares issued to individual investors are floating in the open market for trading. To increase liquidity in the Chinese stock market, China suspended its IPO activity for 1 year to reform the IPO procedure by revising its Securities Law and Corporate Law. Under the revised laws, all public shares issued domestically will be tradable after a certain period set by the CSRC, starting from May 2006 when China resumed its IPO activity.

The Shanghai Stock Exchange was established in late 1990 and initiated 829 A share IPOs while the Shenzhen Stock Exchange was founded in mid 1991 and initiated 551 A share IPOs, over the sample period.

We extend the data to June 2006 when determining stock market cycles in late 2005. The 30% rise or drop to define stock market cycles is arbitrary. We also tried 20%, a typical definition for a bear market, and found similar results. The detailed decomposition of stock market cycles is available from the authors upon request.

There is a double counting issue in calculating the market adjusted average initial return since an IPO sometimes can be a component of the market index. This bias is more pronounced when the overall market capitalization is small relative to the size of the IPO. To reduce this bias, China has announced in 2007 that returns on the first ten trading days from IPOs will not be included in calculating any market index returns.

Traditional factor models use excess returns along with a market factor. Since the risk-free rate and daily index returns are very low during the sample period relative to IPO initial returns, it doesn’t make a meaningful difference if excess returns are used, along with a market factor. Nevertheless, we redo the analysis using excess returns along with a market factor and find that the main results don’t change.

We use the percentage change in firm age to normalize the unit. For example, if A _i is 6 years, \( \bar{A} \) is 5 years, and β_Age is 1.20, then IPO i should earn an additional 24% initial return, compared to an IPO with firm age of 5 in the same sample period, keeping other things constant. Similar arguments apply to the percentage changes in board size, offer price, and offer size.

We treat the system of Eqs. 6 and 7 as if they were “seemingly unrelated equations”, which allows us to estimate one equation at a time to justify the optimal lags of m and p first and then test for significance of the lead-lag relationship. Another way to estimate the system is to estimate both equations simultaneously. Using the same optimal lags of m and q, we perform both tests and find that the results are very similar. Therefore, we only report the results, treating the system as “seemingly unrelated equations”. The detailed procedure is available upon request.

In 1992, the Shanghai Stock Exchange started initiating IPOs with a book value of 10 Yuan (compared to 1 Yuan for all other IPOs in the sample) along with higher offer prices, resulting higher average offer prices. It began to use a book value of 1 Yuan along with lower offer prices from the second half of 1993.

We also report the market adjusted returns in Table 1. Since the IPO underpricing is so severe it doesn’t make a real difference in initial returns with and without the market adjustment. Therefore, we focus the discussion on EWIR₁.

We also estimate the correlation matrix over sub-periods and find that the correlation between firm age and offer size becomes significant over the sub-period 1991–1998. Over the sub-period 1999–2005, we find that the correlation between offer price and offer size is no longer significant. All other correlations remain similar over the two sub-periods.