Thursday, December 19, 2013

The CEO Investment Cycle

In the world of economic theory, and sometimes in the minds of employees, the chief executive officer of a company is an all-seeing, all-knowing planner with a ruthless eye on the bottom line. However, it appears that the average CEO has a typical career trajectory: spend a few years cleaning up the past investment mistakes of the firm, then start expanding the firm's investments again, and finish off by over-expanding--so that the next income CEO has a bunch of investment mistakes to clean up, and the cycle can start anew. Yihui Pan, Tracy Yue Wang, and Michael S. Weisbach provide the evidence in "CEO Investment Cycles," which was published as national Bureau of Economic Research Working Paper #19330. NBER working papers are not freely available online, although some readers will have access through library subscriptions, but a readable summary of the paper is available here.

Pan, Wang, and Weisbach collect data on 5,420 CEOs who took office in 2,991 publicly traded US firms between 1980 and 2009. They gather data on whether a departing CEO does so for reasons of age or health, or under some particular pressure. They also have data on whether the CEO is hired from inside or outside the firm. They don't look at the profits reported by firms, in part because profit data is at least as much an outcome of past decisions rather than current ones, and in part because profit data is often so massaged by corporate tax attorneys that it actual economic meaning can be unclear. Instead, they look at the annual patterns of investment by firms.

"We estimate the magnitude of the CEO cycle in terms of the differences in disinvestment, investment, and firm growth, between the first three years of a CEO’s tenure and the later years, holding other factors constant. The magnitude of the changes in firm investment and growth over the CEO cycle is substantial. For example, the annual investment rate (investment-to-capital-stock ratio) tends to be 6 to 8 percentage points lower and the asset growth rate tends to be 3.2 percentage points lower in the first three years of a CEO’s tenure than in his later years in office. Given that the median investment rate in our sample is 24% and the median asset growth rate is 7.6%, the differences in investment and growth between the earlier and the later parts of the CEO cycle are clearly non-trivial. The effect of CEO cycle on investment is also of the same order of magnitude as the effects of other factors known to influence investment such as the business cycle, political uncertainty, and financial constraints."
Their preferred explanation is that when CEOs are first appointed, they are under pressure from the board of directors to be the ruthless profit-seekers of legend. But over time, as the CEO appoints more members of the board of directors, the CEO feels less of this pressure and has a tendency to over-invest.

"First, when a CEO takes office, he will have incentives to divest poorly performing assets that the previous CEO established and was unwilling to abandon. Second, for many reasons, CEOs usually prefer their firms to grow, potentially at the expense of shareholder value maximization. The board of directors is an important constraint on CEOs’ ability to deviate from the shareholders’ interest. However, as the CEO becomes more powerful in the firm over time, he will have more sway over his board and will be able to undertake investments that maximize his utility, potentially at the expense of value. Eventually, when the CEO steps down, the process is repeated by the next CEO. ... We measure the CEO’s capture of the board by the fraction of the board that is appointed during his tenure, and find that the increasing CEO influence on the board over his tenure explains the positive relation between CEO tenure and investment. ... In addition, we find that the quality of a firm’s investments, measured by the market reaction to acquisition announcements, decreases with CEO tenure and becomes negative during the later portion of his time in office. The deteriorating investment quality is also related to the CEO’s control of the board."

They also seek to evaluate some of the possible alternative explanations. But they find, for example that this pattern holds even when the turnover of the CEO is due to death, illness, or retirement--and thus it just that boards fire CEOs who aren't performing well. They also find this effect both for CEOs hired from inside and from outside the firm, and for firms where the industry is being hit by big positive or negative surprises.

To me, the analysis makes CEOs sound a bit like coaches of sports teams: they arrive to clean up the mistakes of the past regime, but over time many of them gradually drift into their own set of mistakes. It also suggests that firms should think seriously about the independence of their boards of directors and about rotating the CEO on a semi-regular basis. One suspects that the cozy relationships between CEOs and the directors they have appointed is not just manifested in the firm's investment choices, but may well show up in executive compensation and other firm decisions, too.