Why Managers Hold Shares of Their Firms: An Empirical Analysis

3.2.3 Temporal Stability

Our sample period contains the years of the technology bubble. To exam-
ine whether our results are driven by the building up of the bubble or its
bursting, we split our sample in two parts. The first part contains observa-
tions from 1996 to February 2000, the month in which the markets peaked,
and the second part contains observations from March 2000 till the end of
2004.¹⁰ Results are presented in Panel C. They show that our results are
not driven by a specific time period like the building up or the bursting of
the high-tech bubble.¹¹

3.2.4 Alternative Sample

While merging our different data sources, some firms could not be matched
(see Section 2.1). Although the number of non-matched firms is very small,
this could still create some kind of selection bias. To control for this, we
alternatively use data from Dlugosz, Fahlenbrach, Gompers, and Metrick
(2006) (DFGM).¹² They provide ownership information on 7,873 firm years
over the time period from 1996 to 2001. Their database is drawn from the
universe of firms covered by publications of the Investor Responsibility Re-
search Center (IRRC). DFGM provide information on all blockholders, i.e.,
¹⁰Results are very similar if we split our sample period in two subsamples of equal
length.

¹¹For reasons of direct comparability with our results from the broader S&P 1500 uni-
verse, we also compute abnormal returns for managerial ownership portfolios drawn from
the S&P 500 universe for the years 1996 to 2004. They are presented in the last row in
Panel C and are very consistent with results obtained using the slightly longer 1994 to
2006 time period.

¹² We thank Andrew Metrick for providing this data on his Web page
http://finance.wharton.upenn.edu/ metrick/data.htm.

13

<<   13   14   15   16   17   18   19   >>

More intriguing information