Should Firms Look to an Insider or an Outsider When Hiring a New CEO? Evidence from China

ABSTRACT

Using survey data from China, we demonstrate that the performance consequences of managerial turnover depend on firm characteristics and managerial ties outside of the firm. An insider CEO has better knowledge of the organization but may be bound by existing social ties within the firm. An outsider CEO may turn around the firm more efficiently and possess valuable business and political ties. The author finds outside CEOs are associated with higher efficiency gains in state-owned firms with a large labor force and dependent on government support. The finding is robust after controlling for the selection of CEOs.

KEYWORDS:

• CEO turnover
• China
• corporate governance
• executive succession
• managerial ties
• productivity

Notes

1. One can argue that wages above market rate can be justified by an efficient wage rationale as demonstrated by Yellen (1984). But this rationale can be ruled out in the context of this study—the wage level in most SOEs in China is set to a large extent by the government, so it is not at the discretion of the manager. Bonuses, however, do reflect the incentive mechanism the manager can use to motivate the workers. Groves et al. (1994) illustrate the performance consequences of using bonus as incentives in the context of China’s reform.

2. For example, despite the declining return to capital in the state sector from 25% in the early 1980s to less than 10% in 1992, wages and bonuses per worker increased from 4% per year to 7% during the period (Qian, 1995).

3. An enormous body of theories suggest that managers’ incentives are related to their compensation and career concerns (Fama, 1980; Holmstrom 1979; Murphy, 1999), measured by their payment, education, and tenure in the firm. We will control for these factors in the empirical tests.

4. See, for example, Maskin, Qian, and Xu (2000) and Roland and Sekkat (2000).

5. The earlier part of the same survey data from 1980 to 1989 was used in the empirical works on China’s reform of state-owned enterprises. The description of the data and the history of the survey can be found in Li (1997).

6. The three surveys were conducted in such a way as to record retrospective data that could introduce survival bias and memorial errors.

7. We estimated the selection equation with data one year prior to CEO turnover, and the result is qualitatively similar.

8. All the firms in the data set experienced at most one turnover during the period 1994 to 1999.

9. The result is qualitatively similar when we estimate a Cobb-Douglas production function.

10. The estimation using employment as labor input generates qualitatively similar results.

11. A more accurate measure of the scale of customer credit is the ratio of accounts receivables over total revenue. But total revenue on the income sheet is highly correlated with industrial output, our dependent variable.

12. The estimation using Levinsohn-Petrin proxies gives similar results so the results are not reported.

13. We tested an alternative hypothesis that outsiders and insiders are different in their recruiting behavior. In the fixed-effect estimation we found no evidence that outsiders and insiders are different in hiring or firing workers, nor in recruiting more college graduates. This is probably due to the relatively tight government regulation on labor during the period.

14. The estimation is restricted to the sample after matching and thus has controlled for the selection effects of CEO successors.

15. The negative coefficient on the outsider CEO in the NSOEs also holds to this test, suggesting these enterprises may be different from SOEs in productivity improvement and managerial turnover.