182 /304
投資機會和投資者保護機制對控股股東派息的影響
182
firm-year observations, which appears to represent the initial sample for the period 1990
to 2000.
4
3.2 Variables
We calculate four dividend payout ratios to measure a firm’s dividend policy –
dividends/sales, dividends/operating cash flows, dividends/market value of equity, and
dividends/earnings – to avoid the results being biased by an individual dividend payout
measure (La Porta et al., 2000; Faccio et al., 2001). The average dividend payout ratios are
calculated using the 3-year averages, rather than annual figures, of dividends, sales,
operating cash flow, market value of equity, and net income to smooth out noise and
transitory factors. As different industries are at different stages of maturity and growth,
which determine their dividend policies, the average dividend payout ratios are adjusted
for industry effects. For each average dividend payout ratio, we first compute the
worldwide median for each industry (based on the 2-digit primary SIC code), and then
subtract a firm’s average dividend payout ratio from the industry median to obtain the
industry-adjusted dividend payout ratio. Our main tests are performed on dividend payout
ratios both before and after the adjustment of industry effects.
The extent to which the ultimate owner’s voting rights diverge from the cash flow
rights (
Diverge
) is defined as one minus cash flow rights divided by voting rights, in
accordance with previous studies (Fan and Wong, 2002; Haw et al., 2004). Thus, the
closer the value of
Diverge
to 1, the more the ultimate owner’s control rights are detached
from the cash flow rights. An ultimate owner is defined as a shareholder who has
determining voting rights in the firm and who is not controlled by anyone else. The
ultimate owner’s voting rights level is set at 50% and is not traced any further, as an
absolute majority is assumed beyond 50% of voting rights. The maximum proportion of
cash flow rights of any ultimate owner is also set at 50% for the sake of consistency with
the voting rights measure. If a firm does not have an ultimate owner, then it is considered
to be widely held and the value of
Diverge
is set to zero.
4 To ensure that the final sample (6,093 observations) is representative of the initial sample in terms of
ownership data (51,876 observations), the means and medians of the ownership variables and firm-
specific variables of the final sample were compared with those of the initial sample. The unreported
results show that the means and medians of cash flow-control divergence (
Diverge
), cash flow rights
(
Own
), sales growth rates (
G
), size (
Size
), and leverage (
Leverage
) of the final sample are similar to
those of the initial sample.