Until recently, Chinese companies publicly listed in domestic stock exchanges had two classes of stock: tradable and non-tradable shares. These two classes of stock had the same voting, cash flow, and all other legal rights except that non-tradable shares cannot be transferred at the open markets. From 2005 to mid-2007, Chinese government completed the ownership reform, so-called the Split Share Structure Reform (SSSR), to convert all non-tradable shares into tradable shares. Under this reform process, the holders of non-tradable shares had to negotiate with those of tradable shares to determine how much liquidity premium, or the compensation ratio, non-tradable shareholders have to pay to tradable shareholders in order to obtain the liquidity right. This paper starts with a theoretical model to identify the fundamental factors, including price discount before and after the SSSR reform, the percentage of non-tradable shares in total shares, the volatility of tradable share price, and the lockup period, that should determine the compensation ratio. We show that those factors except price discount before the reform are statistically significant in determining the compensation ratio proposed by non-tradable shareholders. We further show that the agency problems also reveal themselves in the compensation ratios. Specifically, when a firm is controlled by a governmental agency, the compensation is higher. However, the compensation is lower when more concentrated in the top ten holders, especially when shares are held by mutual funds. Thus, the evidence is consistent with the notion that the agency problem exists in China’s fund managers. Finally, we show that the existence of agency problems also reduce the importance of fundamental factors in determining the compensation ratios.

In this paper we study how a firm’s capital structure choice affects informed trading of its securities in the secondary markets and consequently, the information efficiency of its security prices. We identify two new factors as the potential determinants of the firm’s optimal capital structure policy: the liquidity premium caused by informed trading and, perhaps more importantly, the improved operating efficiency due to information revelation from its security prices. We show that, from these two perspectives, the optimal debt level is achieved at the point where there is no informed trading in the bond market and the informed traders are just about to trade in the bond market. Thus, the cost of debt financing differs in nature from that of the existing models. This has very different implications for the significance of the cost of debt financing and for financial system design. Our model can also explain the relative trading volumes in debt and equity markets.