In the Harvard Business School cafeteria, students sometimes play a game. It’s called “what’s your number?” Players take it in turn to nominate how much money they are aiming to make in their career, before they stop working hard and relax on a beach. For some, it’s $5 million, for others it’s $70 million. But the principle is the same: given enough money, even students at the world’s most famous business school will swap dividends and buyouts for daiquiris in the Bahamas.
All well and good for those lucky enough to make it. But for shareholders, this poses quite a different problem. Those of us owning shares would like to know what happens to the firm when the CEO reaches his or her “number”. But without knowing the number, how can we answer the question?
In a recently-released working paper, Crocker Liu (New York University) and David Yermack (Arizona State University) have come up with a clever way of answering the question. Using data from the Fortune 500 companies, they investigate the impact on a company’s sharemarket performance when its CEO buys real estate. The idea is simple: if one of the CEO’s goals in life is to secure a grand home, then perhaps he or she won’t try so hard when the dream comes true.
Since a dataset of the real estate holdings of senior executives isn’t sitting on the Internet waiting to be downloaded, Liu and Yermack have to engage in the kind of clever sleuthing that has come to characterise many of the most interesting papers in empirical economics today.
Combining real estate databases, voter registration data, aerial photographs, and Google Maps, they are able to estimate the value and characteristics of 98 per cent of the CEOs in America’s largest firms. The typical house of a Fortune 500 CEO has 11 rooms, 4 bathrooms, and is worth US$2.7 million (about ten times the US median house price).
So how does a CEO’s house relate to the performance of the firm? Comparing share market returns in recent years, Liu and Yermack find that firms whose CEOs have the most expensive mansions significantly underperform firms whose CEOs live in more modest residences. They find that a portfolio of “small-home CEO” firms would have outperformed a “large-home CEO” portfolio by about 15 per cent per year.
Knowing that big mansions are correlated with small firms leads naturally to the question of why CEO pay has increased so dramatically in recent years.
According to Liu and Yermack, many of the homes of America’s top CEOs feature swimming pools, tennis courts, boathouses, separate guest houses and servants’ quarters. At least one estate includes private polo fields and an equestrian ring. For the very top executives, the past generation has been another Gilded Age, with CEO salaries increasing sixfold in the United States over the past 20 years, and fourfold in Australia over the past decade.
Until now, two explanations have been put forward to explain the rise in CEO pay. One set of economists have attributed it to the widespread adoption of high-powered incentives (particularly stock options). Related to this was the increase in CEO turnover, which meant that managers had to be compensated for a higher risk of termination.
The other popular explanation for rising CEO pay has been the “skimming view”. Proponents of this approach have argued that rising remuneration has been driven by the loosening of social norms against generous pay packages, and a clubbish relationship between CEOs and those who advise on executive remuneration.
Yet social norms evolve slowly, whether it be in the boardroom or on the street. While both the incentive theory and the skimming theory predict some increase in CEO pay, neither is able to account for the meteoric rise over recent decades.
By contrast, a new theory - propounded (PDF 933KB) by Xavier Gabaix (MIT) and Augustin Landier (New York University) - claims to be able to fully account for the rise in CEO pay. Their point is startlingly simple: CEOs are paid proportionately to the size of their companies, and a wave of mergers has seen the largest companies become larger still. Over the past two decades, they point out, there has been a six-fold increase in market capitalisation of large US companies, which perfectly explains the six-fold increase in CEO pay.
So next time someone tells you that CEO pay is all Gordon Gecko’s fault, point out to them that the “greed is good” decade also saw a major rise in firm size. Today’s top executives are now managing substantially bigger companies than their predecessors. But regardless of how large their firm is now, if the boss buys a mansion, it probably means you should sell your shares.