Does One Hand Wash the Other? Testing the Managerial Power and Optimal Contracting Theories of Executive Compensation

Abstract

ABSTRACT Executive compensation forms a litmus test for corporate governance law's effectiveness. The interests of management and shareholders diverge most sharply in this arena, where every dollar taken from shareholders transfers directly to management. Legal theorists have broadly divided into two opposing camps on their views of the executive compensation process: those who believe the chief executive labor market is efficient and should be left alone (Optimal Contracting) and those who believe that managerial power over nominally independent directors creates a market failure that requires regulatory intervention (Managerial Power). This article tests the two hypotheses with an experimental model of the executive compensation process. The results indicate that the Managerial Power camp is correct that chief executives' power over directors results in excessive compensation. The article argues that corporate governance reform should therefore focus on minimizing managerial power over directors through mechanisms such as truly competitive elections for directors. I. INTRODUCTION Jack Welch looms as one of the greatest chief executive officers of the late twentieth century. During his twenty-year plus tenure as CEO of General Electric (G.E.), the corporation's earnings rocketed from $1.5 billion in 1980 to $12.7 billion in 2000, revenue quintupled to $129.9 billion, and G.E.'s stock performed more than 3.5 times better than the Standard and Poor's (S&P) 500 Index, increasing 5,096% (while the S&P 500 rose 1,4330Xo).1 Welch for a time became almost a brand name, as his protegees secured CEO positions at over a dozen corporations.2 In exchange for his benchmark performance, Welch received generational wealth from G.E.'s grateful board of directors, $120 million in his last year as CEO alone.3 Was Welch worth the money? This question is surprisingly difficult to answer, even for someone with the colossal reputation of Jack Welch. The answer depends on a counterfactual: how would G.E. have performed with someone else at the helm? Considering the multifarious variables that produce success in an enormous multinational-from the quality and motivation of employees to the inspiration of inventors to interest rates, consumer sentiment and countless other factors-it appears impossible to determine how much a single individual impacts the whole. Perhaps a better question would ask why we should care. Even if Welch was overpaid, his compensation came out of General Electric's pocket, not ours. Most of us did not own G.E. stock during Welch's tenure and, even for G.E. shareholders, Welch's compensation amounted to a negligible sum on a per share basis. Few shareholders could have felt very sharply the impact of any excessive pay. Other than envy, what motivates the fascination of so many legal and finance scholars with chief executive officer compensation? The answer involves the core issue of corporate governance law: how can we structure legal constraints that will minimize the agency costs inherent in the separation of ownership and control in publicly held corporations? As Berle and Means first pointed out over seventy years ago, those who own the public corporation do not control it and those who govern the public corporation do not own it.4 The corporation's shareholders are often dispersed and seldom own a sufficiently large stake in the company to justify close monitoring of particular corporate decisions.5 Therefore, the executives who exercise day-to-day control over the corporation are potentially capable of diverting corporate assets to themselves without any fear that shareholders will catch them in the act. For example, a chief executive officer might hire his or her brother-in-law to provide legal advice to the corporation in exchange for an inflated fee, or overstate corporate profits (thereby earning larger performance bonuses) by funneling losses to partnerships operated off the company's books. …