Since the 1990s the pay of corporate executive officers has skyrocketed. How to determine CEO performance and thus pay is far from a simple question. In an ideal world, the interests of a company’s leader and its shareholders would be perfectly aligned. In reality, their relationship reflects what is known as the principal-agent problem: The CEO acts on behalf of the shareholders, but their priorities can differ. For example, public firms may underinvest in their companies in order to post better results, as a 2013 paper by scholars from NYU and the Harvard Business School suggests.

One way to address this problem is to more closely connect CEO compensation with company performance. This issue has sparked a battle at Oracle, where some of the company’s shareholders are trying to tie CEO Lawrence J. Ellison’s pay — $96.2 million for the company’s 2012 fiscal year, the top among U.S. CEOs — to metrics such as return on equity or industry benchmarks. (Another possible option would be to limit the ratio of CEO pay to that of the median worker in the corporation — data that the SEC is now proposing that companies disclose.)

To better understand these issues, a 2013 working paper for the National Bureau of Economic Research, “CEO Investment Cycles,” investigates the relationship between corporate investment and disinvestment rates and CEO tenure in the United States. The authors — Yihui Pan of the University of Utah, Tracy Yue Wang of the University of Minnesota, and Michael S. Weisbach from Ohio State University — based their work on a sample of compensation, turnover and investment data from 5,420 CEOs at 2,991 firms between 1980 and 2009.

They find significant variation in average investment and disinvestment rates over a CEO’s time in office, even when controlling for economic, political and financial factors. In particular, the study finds that:

• Investment is much lower at the beginning of a CEO’s tenure: “The annual investment rate (investment-to-capital-stock ratio) tends to be 6 to 8 percentage points lower … in the first three years of a CEO’s tenure than in his later years in office.” This is a substantial difference: the median investment rate over the sample was 24%.
• The effect of the CEO cycle on investment is similar to that of other economic factors. For example, the effect of a CEO’s being in the first three years of tenure is “approximately the same as being in recession or facing financial constraints, and more than twice the effect of being in an election year.”
• The longer a CEO’s tenure, the less the market reacts to acquisition announcements by the firm, suggesting that these investment decisions are not maximizing profits. “CEOs have many reasons to prefer more investment than is optimal from a value-maximization perspective. As the CEO acquires more influence over the board, his ability to overinvest increases, leading to increasing investment with CEO tenure.”
• Disinvestments are much higher during early periods in a CEO’s time with a firm: “CEOs are reluctant to divest or reoptimize on bad investments that they have made due to private benefits or career concerns. It often takes a new CEO to enforce optimal disinvestment, leading to high disinvestment intensity shortly after CEO turnover.”
• The likelihood of disinvesting from poorly performing assets decreases significantly if the old CEO or former executives stay on the board. “While departing CEOs can certainly play an important advisory role when they are chairmen of the board, this finding suggests that there is also a negative side to this practice in that it can hinder error correction subsequent to CEO turnover.”
• As CEO tenure increases, his or her power over the board increases. This in part is a function of the number of directors appointed during their tenure, something that tends to increase with the CEO’s time in office.
• The higher the number directors appointed by the CEO, the higher the level and the lower the quality of the investment undertaken.

These findings have important implications for decisions on CEO appointment, performance and compensation. They show that agency problems — empire-building and unwillingness to change past investment decisions — have a major impact on corporate investment, generating a “CEO investment cycle.” The magnitude of this cycle is affected by the power balance between the CEO, the board and the previous management. The authors suggest that a “policy of regular management turnover,” with frequent changes in senior management and limited retention of old management in advisory roles, could be valuable to limit poor-quality overinvestment by long-tenure CEOs.

Keywords: economy, corporate investment, Q theory, private companies, managerial incentives, agency costs, short-termism, managerial myopia, IPOs