Table of Contents Table of Contents
Previous Page  189 /304 Next Page
Information
Show Menu
Previous Page 189 /304 Next Page
Page Background

189

臺大管理論叢

28

卷第

1

4. Regression Results

4.1 Control Divergence, Legal Institution, and Dividend Payout

This study empirically tests the prediction that the positive relationship between the

ultimate owner’s cash flow-control divergence and dividend payouts is less pronounced

for firms with good investor opportunities in countries in which legal investor protection

is strong. We develop a regression model that includes control divergence (

Diverge

), legal

institutions, growth opportunities, and the interactions between them. It adds firm- and

country-specific characteristics to minimize the confounding effects of potential error in

the measurement of the dividend payouts. The model also controls for the ultimate

owner’s cash flow rights (

Own

) and the interaction between

Own

and country-specific

institutional factors to ensure that the documented relationship between dividend payouts

and

Diverge

is not driven by the absence of

Own

. The generic regression model with firm

and time subscripts omitted takes the following form (with the predicted signs of the

regression coefficients).

PAYOUT

= β

0

+ β

1

Diverge

+ β

2

Institution

+ β

3

Diverge×Institution

+ β

4

GxDiverge

+

(?) (+)

(+)

(+)

(+)

β

5

GxInstitution

+ β

6

GxDiverge×Institution

+ β

7

Own

+ β

8

Own×Institution

(-)

(-)

(?) (?)

+ β

9

G

+ β

10

Size

+ β

11

Leverage

+ β

12

Ln(GNP)

+ β

13

LRes

+ β

14

Dta

+

(-) (?)

(-)

(?)

(-) (+)

error term,

(1)

where

PAYOUT

refers to each of the four dividend payout ratios (before and after the

adjustment for industry effects), and

Institution

represents country-specific legal

institutions such as

Legal, Right

, or

Disclosure

.

G

represents the measure of investment

opportunities. As firms with more investments make lower long-term dividend payouts

(Fama and French, 2002), we control for this by using the decile rank of the average

annual sales growth rate of each firm over the previous five years (

G

).

In addition to cash flow rights (

Own

), we include two other firm-level variables that

may affect a firm’s dividend policy: firm size (

Size

) and leverage (

Leverage

). Cash flow

rights may affect the ultimate owner’s incentive to pay out dividends, although the

empirical evidence is inconclusive (Farinha, 2003; Gugler, 2003). We control for the size

effect, which is defined as the natural logarithm of total assets in U.S. dollar at the end of