Causes of changes in ultimate share ownership

Part of the Contributions to Economics book series (CE)


The purpose of the section is to provide first-hand evidence on the determinants of changes in ultimate share ownership from an unbalanced panel of cross-sectional data of Chinese listed firms over an eleven-year period (1996 to 2006). The aim is to test hypotheses one to four, as outlined in section 2.3.


Firm Performance Total Asset State Transfer Firm Characteristic Large Shareholder 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 117.
    For instance, Wei et al. (2005, p. 93) follow this approach.Google Scholar
  2. 118.
    Bai et al. (2004, pp. 608 and 611) define two alternative measures of Tobins’q by taking a 70 per cent and 80 per cent discount for non-tradable shares compared to the price of tradable shares. Their results are, however, not affected by the different measures of Tobins’q.Google Scholar
  3. 120.
    Huang and Xu (2005, p. 11) were among the first to suggest this variable.Google Scholar
  4. 121.
    Other potential proxies for the default risk on bank credit include the relative position of bank credit to total debt and the interest coverage ratio. Unfortunately, data on net interest and bank credit was not available for the universe of publicly listed firms in China. The corporate bond market in China is, however, very small. Therefore the debt positions of firms are mostly related to bank financing. See OECD (2005a, pp. 158–160) for a description of the corporate bond market in China and Guo and Yao (2005, p. 223) for alternative measures of the default risk on bank credit used in the privatization literature.Google Scholar
  5. 122.
    See e.g. Barclay and Holderness (1989, pp. 385–386).Google Scholar
  6. 123.
    The Wilcoxon rank-sum test is a non-parametric alternative to the two-sample t-test. See Greene (2003, p. 106) for details of the t-test; Sun and Tong (2003, p. 195) apply the Wilcoxon rank-sum test. The STATA code for the t-test is “ttest” and for the Wilkoxon rank-sum test “ranksum”.Google Scholar
  7. 124.
    For a discussion of this literature see Jefferson and Su (2006, p. 152).Google Scholar
  8. 126.
    The following section is based on Wooldridge (2002, pp. 482–492), Cameron and Trivedi (2005, pp. 795–797), and Greene (2003, pp. 695–701).Google Scholar
  9. 127.
    See Katz (2001, pp. 379–380).Google Scholar
  10. 128.
    See Greene (2003, pp. 695–697) for a detailed technical report on the unconditional FE model with a binary dependent variable and a description of the incidental parameters problem.Google Scholar
  11. 129.
    Wooldridge (2002, p. 491) points out that the fact that this conditional distribution does not depend on ai is a feature of the logit functional form. The same argument does not work for the probit case.Google Scholar
  12. 130.
    See Chamberlain (1980, pp. 228–230) and Greene (2003, pp. 696–698); for a detailed deviation of this estimator see Baltagi (1995, pp. 178–180).Google Scholar
  13. 131.
    See Cameron and Trivedi (2005, pp. 796–797) for a more technical illustration of this point.Google Scholar
  14. 132.
    See Greene (2003, p. 699) and Wooldridge (2002, p. 492).Google Scholar
  15. 134.
    Wooldridge (2002, p. 57) contains a general description of the Huber/White estimator and Greene (2003, pp. 519-521) details the Huber/White estimator in ML estimation. Table 4.3 contains the goodness-of-fit measure pseudo Rsquared as suggested by McFadden and the p-value of the model measured by the Wald-statistic. See Wooldridge (2002, p. 465) for the definition of McFadden’s likelihood ratio index and Greene (2003, pp. 676–678) for a discussion of the Wald-test.Google Scholar
  16. 137.
    See Liu et al. (2006, p. 2024) and section 2.1, pp. 9–10 for details and further references.Google Scholar
  17. 138.
    See Sun and Tong (2003, p. 215).Google Scholar
  18. 139.
    See Wang et al. (2004, p. 485).Google Scholar
  19. 140.
    An alternative interpretation concerns wealth constraints of investors. Shleifer and Vishny (1992, pp. 1362–1364) show that the market for corporate control is less liquid as firm size increases.Google Scholar
  20. 142.
    See Cai and Chen (2004, p. 69).Google Scholar
  21. 143.
    See Köke (2004, p. 75).Google Scholar
  22. 144.
    Wei, Xie, and Zhang (2005, p. 106) report that when government ownership decreases and other blockholder become dominant, firm value increases. Bai et al. (2004, p. 610) support this result.Google Scholar
  23. 146.
    Alternatively one could argue that potential private benefits from control are higher in more concentrated firms increasing the probability of a change in control. In this respect, Dyck and Zingales (2004, p. 572) show in a cross-country study that ownership is more concentrated in countries in which private benefits of control are larger.Google Scholar
  24. 147.
    See Green and Liu (2005, p. 130).Google Scholar

Copyright information

© Physica-Verlag Heidelberg 2008

Personalised recommendations