This thesis is a study of the effects of abnormal items and extraordinary items on Australian CEOs' remuneration. Since the study argues that CEO pay is ex ante (implicitly or explicitly) a function of bottomline net income, abnormal items and extraordinary items can potentially affect CEO pay ex post. The underlying research issue is whether boards of directors ex post shield CEO pay from the effects of unusual gains (abnormal plus extraordinary gains) and/or unusual losses (abnormal plus extraordinary losses). The specific objective of this thesis is to explain cross-sectional differences in the extent of shielding of these gains/losses.
The general pattern that emerges from studies in the U.S. is that gains (losses) from unusual transactions, such as change in accounting methods and restructuring events, are on average not shielded (shielded) from affecting CEO pay. In addition, evidence
also exists that the extent of shielding varies with the impact of the transaction on a bank's capital adequacy position, frequency of the (restructuring) transaction and expected horizon of the CEO. However, the role of the board and the CEO in the shielding decision has not being investigated.
The major contribution of this thesis is a more general and fundamental explanation for cross-sectional differences in the extent of ex-post shielding, and it is also an explanation that explicitly takes into account the roles played by the board and the CEO in the shielding decision. The thesis initially proposes that the extent of shielding of unusual gains/losses will depend on the judgements of the board regarding the controllability by the CEO over the gain/loss and the expected benefit to the company from the gain/loss. Ceteris paribus, more controllable gains (losses) will be shielded less, and more beneficial gains (losses) will be shielded
However, this thesis also argues that these judgements are themselves subjective outcomes dependent upon bargaining between the board and the CEO. The CEO has more information on the gains/losses, and he will naturally prefer all losses to be shielded and all gains to be not shielded so as to maximise his compensation. The board will be wary about favourable rationalisations by the CEO regarding the gains/losses, and will resist the shielding of losses and promote the non-shielding of gains lest they be taken to task by shareholders for being too lenient to the CEO. Their reputations and personal shareholdings may also be at stake. The model then proposes that the outcome of the bargaining depends on the power/influence of the board relative to that of the CEO. Ceteris paribus, the more powerful the board is, the more likely it will succeed in arguing that gains (losses) are not controllable (controllable) by the CEO and are
not beneficial to the company, and so it will shield the compensation effects of the gains (losses) to a greater (lesser) extent.
Drawing from the corporate governance literature, this thesis proposes four dimensions of board power. The board is regarded as more powerfiil or influential than the CEO if it has a non-CEO chairman; if it consists of proportionately more independent directors; if its independent directors have more firm-specific knowledge than the CEO and other directors; or if its independent directors hold a greater proportion of the total ordinary shareholdings of the board.
The sample for the study was selected from the top 500 companies Usted on the Australian Stock Exchange. Data was obtained primarily from published accounts for fiscal periods spanning 1986 to 1995. The sample comprised 426 CEO-years from 89 companies. A pooled cross-sectional time-series CEO pay-earnings model was estimated using ordinary least
Results from the base (main) analysis of the data support two hypotheses. First, the extent of shielding of unusual gains (losses) varies directly (inversely) with the presence of a non-CEO chairman of the board. Second, the extent of shielding of unusual losses is also found to vary inversely with the independent directors' ordinary shareholdings. Various sensitivity analyses report that these results (especially the first) are generally robust. Other hypotheses on the relationships between the extent of shielding and the other dimensions of board power are supported only in sensitivity analyses, and may depend on whether the chairman is a CEO or not. In almost all the regressions of this study, serious to severe multicollinearity is present, which may be the reason for the relatively "poor" results.
In conclusion, the results suggest that boards do intervene ex post to shield CEO pay from the
effects of unusual gains and unusual losses, and the extent to which they do so depends on their power/influence over the CEO.