Price Clustering of Chinese IPOs: The Impact of Regulation, Cultural Factors, and Negotiation

ABSTRACT

During June 2009–May 2012, the China Securities Regulatory Commission (CSRC) suspended window guidance that limits issue prices. Using this regime change as a natural experiment, we test the combined effects of regulation, culture, and negotiation on price clustering of Chinese IPOs. The proportion of IPOs priced on round number 0 increases from 42.58% during sample periods with window guidance to 79.81% during sample period without window guidance, a level similar to that reported in developed markets . Moreover, we document a connection between whole CNY pricing of Chinese IPOs and several uncertainty measures including a unique uncertainty proxy defined as the time gap between the IPO date and the listing date. Second to the round number 0, issuing firms favour number 8 that associates with fortune, particularly during sample periods with window guidance. Our findings that price restrictions limit the power of negotiations but not the influence of cultural factors contributing to the understanding of price formation process.

KEYWORDS:

• Price clustering
• IPO
• negotiation
• culture
• regulation

JEL CLASSIFICATION:

• G11
• G12
• G15

Acknowledgments

This work was supported by the Faculty Research Fund, Arkansas State University. We thank Jay R. Ritter and participants at the 2017 FMA conference for valuable comments.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Clustering appears not only in stock prices (for examples, see Osborne 1962; Davis, Van Ness, and Van Ness 2014) but also in many other markets such as gold (Ball, Torous, and Tschoegl 1985), foreign exchange (Sopranzetti and Datar 2002), and derivatives (Schwartz, Van Ness, and Van Ness 2004).

² Clustering appears not only in stock prices but also in many other markets. Osborne (1962) examines price clustering in the US stock market. Ikenberry and Weston (2008) find significant price clustering in the US stock market both before and after the decimalization in tick size. Aitken et al. (1996) document price clustering in the Australian Stock markets. Asçioglu, Comerton¯Forde, and McInish (2007) find price clustering behaviour in the Japanese stock markets. Price clustering also occurs in markets such as gold (Ball, Torous, and Tschoegl 1985), foreign exchange (Sopranzetti and Datar 2002), and derivatives (Schwartz, Van Ness, and Van Ness 2004). Clustering is not limited to transaction prices. Analyst forecasts cluster at round numbers (Dechow and You 2012), and so does dividends per share (Aerts, Van Campenhout, and Van Caneghem 2008).

³ An alternative hypothesis is the attraction hypothesis that derives from psychological factors. It argues that people are naturally attracted to salience numbers. Both psychological experiments and evidence from financial markets show that people tend to use round numbers that are easier to recall (for examples, see Schwartz, Van Ness, and Van Ness 2004; Ikenberry and Weston 2008).

⁴ There are two key elements in the initial offer price formula: Offer price = EPS * P/E, where P/E refers to the price-to-earnings ratio, and EPS is earnings per share. On 26 December 1996, the CSRC changed EPS formula from forecasted EPS to realized average EPS in the past three years leading to the IPO. On 10 September 1997, the CSRC changed the EPS formula to weighted average of forecasted EPS during the IPO year (70% weight) and realized EPS in the year before the IPO (30% weight). On 17 March 1998, the CSRC changed the EPS formula back to forecasted EPS with time-weighted average number of shares before IPO and number of IPO shares.

⁵ GTA and RESSET databases have been used widely in academic research on the Chinese financial markets.

⁶ We test other proxies for uncertainty, such as volatility and firm size, using a multivariate analysis framework in the following section.

⁷ The average Gap is 80 days for zero-ending IPOs and 26 days for non-zero-ending IPOs. The difference in Gap is statistically significant at 1% level. We repeat the analysis for samples of whole CNY and fractional IPOs. The results are similar. The average Gap is 82 days for integer IPOs, 2 days longer relative to zero-ending IPOs, but 36 days longer than the Gap for fractional IPOs. These results provide further support for the negotiation hypothesis.