Board structure, corporate governance and firm value: evidence from Hong Kong

Abstract

This article investigates the effects of board structure and internal Corporate-Governance (CG) mechanisms on firm value in an emerging market with concentrated ownership and family involvement. Using a unique Hong Kong (HK) panel dataset from 2001 to 2009, we create a board-structure index that captures board independence, balance of power and conflicts of interest. We also construct other major CG mechanisms to correctly specify our model. We combine the 13 CG attributes, which consist of binary and continuous variables, with four CG mechanisms, using Principal Component Analysis (PCA). In contrast with prior evidence from developed markets, our results indicate that firms with independent board structure are associated with higher firm value and are both statistically and economically significant. The results also suggest that board structure is the most important among the major internal CG mechanisms.

Keywords:

• corporate governance
• board structure
• firm value
• family ownership
• internal corporate governance mechanisms

JEL Classification::

• G34
• G32

Ackowledgements

We thank the HKU Foundation and University of Macau for financial support. We thank Bai Chong-en, Joseph Fan, Zhang Junxi, Kim Woochan, Keith Lam, Lewis Tam and William Cheung, and two anonymous reviewers for their constructive comments on this article. We also thank Ms. Cheung Pui Ki for her excellent research assistance.

Notes

¹ This tends to happen in countries with poor legal enforcement, such as the Czech Republic (Johnson et al., 2000), India (Bertrand et al., 2002), and Korea (Bae et al., 2002; Sung, 2003). Bae et al. (2002), for example, find that Samsung SDS of Korea sold 3.21 million shares of its bonds with warrants to the family of the firm controller at a price per share of 7150 won, while Over-The-Counter (OTC) market shares were trading at 55 000 won each.

² The typical director's fee for independent nonexecutive directors in Hong Kong is about HKD100k per annum.

³ For example, Hung and Kuo (2011) show that in family controlled firms, compared with the firms that are not family controlled, investment is more sensitive to cash flow, which is related to asymmetric information problems.

⁴ Similarly, Roosenboom and Van Der Goot (2006) support that option grants are an increasing function of the employees’ benefits for the firm, and cash constrained firms appear to use employee stock option grants in place of cash compensation.

⁵ For other nonlinear relationships related to ownership structure, Del Brio et al. (2011) find evidence of a quadratic relationship between ownership concentration and diversification.

⁶ Red-chip companies are Mainland-controlled companies incorporated outside of Mainland China, wherein the largest shareholder holds at least 35% of companies’ shares either directly or indirectly through companies controlling these entities, while H-Shares are companies incorporated in Mainland China and are approved by the China Securities Regulatory Commission to be listed in HK.

⁷ The aim of PCA is to find unit-length linear combinations of variables with the largest variance. PCA yields q common factors that reconstruct the p original variables.

where y_ij is the value of the ith observation on the jth variable, z_ik is the ith observation on the kth common factor, b_kj is the set of linear coefficients known as eigenvectors, and e_ij is the residual matrix normally distributed.

⁸ These two questions are ‘whether there has been any controversy over whether the board or senior management has made decisions that favoured them over shareholders’ and ‘whether any decisions by senior management have been perceived to favour majority shareholders over minorities’.

⁹ Board structure would have a range of 0–4, executive compensation and accounting standards 0–3, and ownership compensation 0–2, before normalization to 0–100 for all these indices.

¹⁰ The correlations between all the variables are low, except for that between executive compensation and ownership structure, which has a medium correlation. This suggests that the problem of multicollinearity is not significant for these models.

¹¹ Equals 1 if CEO acts as the Chairman.

¹² Results are similar when substituting with MTBV.