3.2.7 Treatment of Missing Returns
Shumway (1997) argues that some asset pricing anomalies could be caused
by problems due to missing returns for some months and firms in the CRSP
database. Firms with missing returns are usually excluded when construct-
ing portfolios. To check whether the abnormal returns we find are due to
these exclusions, we replace every missing return with -1 and include the
respective stock in our portfolio if it qualifies. Results using this extreme as-
sumption to calculate portfolio returns are presented in Panel G of Table 4.
They show that the critique of Shumway (1997) does not apply in our case.
Our results are not driven by the influence of missing return observations.
3.2.8 Constant Portfolio Size
The incentives of a manager to work hard and increase a firm’s value are
primarily determined by the absolute value of her shareholdings.14 This is
the reason why we mainly focus on portfolios in which thresholds of share-
holdings are used as the most important selection criterium. Nevertheless,
some objections may be raised about using specific thresholds since this
creates portfolios that differ over time with respect to the number of firms
included. To counter these objections, we also construct portfolios using the
rank of firms with respect to the shareholdings of the officer with the highest
managerial ownership. Panel H in Table 4 shows that these portfolios also
produce similar abnormal returns to the ones using percentage sharehold-
14To give a precise measure of the manager’s incentives, this number alternatively could
also be related to the manager’s overall personal wealth. Unfortunately, we do not have
information on the manager’s personal wealth outside the firm. However, especially for
the large shares of managerial ownership of 5% or more of the whole company, it is very
likely that the investment in her own firm clearly dominates the personal portfolio.
16