Paper of the week: CEO home purchases and stock performance

Here’s a fun one:

Where are the shareholders’ mansions? CEOs’ home purchases, stock sales, and subsequent company performance by Crocker Liu of Arizona State and David Yermack of NYU.

We study real estate purchases by major company CEOs, compiling a database of the principal residences of nearly every top executive in the Standard and Poor’s 500 index. When a CEO buys real estate, future company performance is inversely related to the CEO’s liquidation of company shares and options for financing the transaction. We also find that, regardless of the source of finance, future company performance deteriorates when CEOs acquire extremely large or costly mansions and estates. …

We look first at whether the CEO sells shares of company stock to finance the home purchase. Although buying a house appears to offer a prima facie personal liquidity reason for CEOs to sell their own shares, most are wealthy enough to acquire homes with other sources of finance. We find a strong temporal pattern of CEOs exercising options and selling shares in the period leading up to their home acquisition dates. These stock sales are often small relative to the CEO’s total investment in the firm, with a mean of about $450,000 and a median of zero. However, they appear to give significant signals about future company performance, a pattern that is all the more surprising due to the apparent personal liquidity rationale for the sales. …

Separate from the decision of whether to sell shares, we look at the size of homes acquired by CEOs. In principal, a large home purchase could indicate either commitment or entrenchment, since reversing the purchase requires significant time and transactions costs. Our analysis supports the entrenchment hypothesis, as we observe an inverse association between the value and size of a CEO’s residence and the returns on his company’s stock. This relation holds cross-sectionally for the year following the effective date of our sample, and it also holds when tested as a hypothetical trading rule implemented at the acquisition date of each residence, for the subsample of CEOs who acquire their homes after taking office.