CEO pay as a reflection of power or performance: an empirical test for The Netherlands, 2002‐2006

Purpose – Previous empirical research interprets results from pay‐performance studies in the light of either agency theory or managerial power theory. This paper aims to directly estimate the relationship between CEO power, and compensation structure, level, and performance‐sensitivity. In doing so, it seeks to test the crucial assumption in managerial power theory according to which more powerful CEOs are able to enjoy higher and less performance‐sensitive compensation. Design/methodology/approach – The hypotheses are tested on a detailed dataset, covering compensation for CEOs in virtually all Dutch stock‐listed companies, for the period 2002‐2006. The paper tests whether the findings are robust against different lag structures and firm size classes. Findings – In general, most of the multi‐dimensional measures of power do not appear to have a strong effect on compensation, with one exception: non‐Dutch CEOs receive more variable compensation, and receive higher and less performance‐sensitive pay than their Dutch colleagues. Originality/value – This paper contributes to the extant CEO compensation literature, which to date relies on interpretations of findings in pay‐for‐performance studies to argue for either agency or managerial power theory. The direct test of the relationship between power and compensation emphasises the importance of one dimension of a multidimensional power construct. As strong effects of performance of compensation are not found either, the paper suggests that the bipolar debate be extended to include other explanations of compensation arrangements.