Three Essays in Executive Compensation.

This thesis investigates three theoretical problems in executive compensation literature. They involve extension of a standard principal-agent model, incorporating taxation into the valuation of executive stock options, and the pricing of executive stock options in the presence of managerial effort. Empirical literature has long addressed the endogeneity of capital structure and executive compensation. Yet few models, which optimally determine executive compensation, explicitly introduce capital structure choice. Chapter 2 proposes a principal-agent model in which the capital structure, compensation and managerial actions are simultaneously determined. Based on our numerical results leverage has two effects on managerial actions. One is to discipline the manager and the other is to replace the incentive effect of compensation. Two such effects exist because volatility is chosen by the manager. The basic model is also extended to include debt-like compensation. Our results show that for a given leverage level rewarding the manager with debt makes her work harder but take less risk. But. debt, compensation cannot limit risk neutral shareholders' risk appetite; we hence conclude that only a combination of capital and pay regulation, which restricts both risk-taking of shareholders and incentives of the manager, can significantly reduce the firm's risk. Taxation is an important consideration in the design of executive (and employee) compensation. It directly affects the firm's revenue as well as the executive's after tax income. Once the compensation is granted, taxes also affect the early exercise strategy of the components of the compensation. Chapter 3 explores the executive (and employees) compensation with tax. Specifically, we build a tax-inclusive valuation model,. The new feature of the model is an addition of a tax decision, which allows the executive (and employees) to optimally sell stock to maximize after-tax terminal utility. The stock selling decision is very similar to an option exercise decision. The valuation model essentially has two embedded options: one option is when to exercise the stock option and the other option is when to sell the stock. This new feature allows different exercise policies for executive stock options under different tax schemes. We apply the model to the US and the UK tax system. The findings suggest that restricted stock is the preferred form of compensation in the US. In the UK, restricted stock is only preferred when the executive has low wealth. We also investigate incentives of a special tax scheme - section 83b election - which gives employees a choice to pay income tax at grant date. This voluntary election allows the executive to accelerate tax on restricted stock. Our results suggest that 83b election is not optimal for the manager, who would get double-taxed. And it is not optimal for the issuing firm either, as restricted stock without the election can provide higher incentives at lower cost. The value of executive stock options (ESOs) should depend on the manager's ability to influence firm value. ESOs are granted under the assumption that the executive could make the firm value increase. However, ESOs are always valued with no managerial influence. Chapter 4 examines valuation of ESOs, with the assumption that the manager can influence the firm value via her effort choice. The manager influences stock prices by exerting effort, which increases the firm's stock expected return. Effort leads to a disutility (which can be regarded as effort cost) to the risk-averse, utility-maximizing manager. In addition to the effort choice, the market asset is also introduced to the manager's investment set. Effort increases the manager's subjective valuation as well as the cost of ESO. The standard principal-agent model is not strictly speaking consistent with general equilibrium models like CAPM. Managerial effort is generally not priced under these equilibrium models, because all managers are price-takers. For this reason, we assume that CAPM does not strictly hold when effort is introduced. Our results show that the manager's value and the cost increase with the correlation, because the manager delays a value destroying early exercise. We also show that the manager's subjective value of the ESO is higher than the cost only when the manager has low wealth, low risk-aversion, and the stock has a low volatility. Under these scenarios, the manager's marginal utility is high and effort has a large impact on the manager's valuation. (Abstract shortened by ProQuest.)