CEP Discussion Paper No 1135 March 2012 CEO Bonding : Who Posts Performance Bonds and Why ?

Despite their theoretical value in tackling principal-agent problems at low cost to firms there is almost no empirical literature on the prevalence and correlates of performance bonds posted by corporate executives. Using data for China we examine their incidence and test propositions from principal-agent theory regarding their correlates. Around one-tenth of corporations deploy performance bonds. They are sizeable relative to CEO cash compensation. Ceteris paribus, CEO's posting performance bonds are more likely than other CEO's to have their compensation linked to firm performance in other ways and the elasticity of their pay with respect to firm performance is greater. They are also more likely to hold company stock. Thus bonds appear to be complements to rather than substitutes for other forms of corporate incentive. The negative association between bonds and sales volatility is consistent with principal-agent theory. Positive associations between performance bonds and firm age, the CEOs ranking in the Communist Party, and city-level clustering in the use of bonds are all consistent with "legacy" effects dating back to the use of performance bonds in the early reform period. The only corporate governance measure that is strongly and robustly associated with an increased use of bonds is employee representation on the board of directors. JEL Classifications: G34; J31; J33; M12; M52; O16; P31