from a low of 7% in 1992 to nearly 10% in 1999. There are substantially more
owner CEOs within the S&P 1500 firm universe (Panel B). Approximately
every second firm has a CEO who is invested in the firm, and well above
10% of firms have a CEO who owns more than 5% of the firm in each sample
year. There are only a few CEOs that own more than 50% of a S&P 1500
firm, and there is only one instance in which a manager owns more than
50% of a S&P 500 firm.
2.2 Construction of Portfolios
We construct portfolios based on ownership data in order to test whether
or not these portfolios would have earned abnormal returns. Portfolios are
constructed based on publicly available information about managerial own-
ership and are reset at the beginning of each year. For each year t, our initial
full universe to choose firms from is the constituency list of the S&P 1500
(or S&P 500) at the end of year t - 2. For example, a firm qualifies to be
in a portfolio in the year 1994 if it was a member of the S&P 500 index
at the end of 1992. Firms that were members of the S&P 500 at the end
of 1992 filed their ownership data during 1993. We only start to invest at
the beginning of 1994. This ensures that the ownership information for the
universe of investable firms is public information. Thus, all portfolios are
constructed using public information only. This ensures that our results are
not driven by announcement effects.
Using the methodology described above, we construct value-weighted
portfolios consisting of firms in which the manager with the highest owner-
ship owns more than a specific cutoff fraction of the company’s stocks. We
use 5%, 7.5%, 10%, 12.5%, and 15% of managerial ownership as alternative