• 沒有找到結果。

Additional Analyses

Among all the shareholders of listed companies, institutional investors are major monitors of firms’ policies, and they play an important role in shaping companies’

external governance. Institutional investors in general prefer more public disclosures (Bushee & Noe, 2000; Bird & Karolyi, 2016). After the implementation of Say-on-Pay,

institutional investors may require companies to disclose compensation policies in more details to facilitate their assessment of the company’s executive pay scheme. In addition,

institutional investors are considered as sophisticated investors in prior studies. For example, they are more capable of analyzing the persistence of current accruals for future earnings (Collins, Gong, & Hribar, 2003), and their monitoring is effective in reducing future stock price crash risks (Callen & Fang, 2013). Their demand for more detailed compensation related information combined with their higher sophisticates may contribute to the increasing trend in the difficulty of CD&A readability.

In this section, I investigate whether institutional investors’ monitoring is associated

with the decreasing CD&A readability after the Say-on-Pay provision. I use the ownership by a firm’s top five institutional investors and by a firm’s dedicated

institutional investors to measure institutional investors’ monitoring. I employ the following two models to examine how institutional investors’ monitoring affects the

CD&A readability after the Say-on-Pay implementation:

READj,t = β0 + β1POSTj,t2HIGH_TOP5j,t 3POST*HIGH_TOP5j,t 4OVERPAIDj,t5UNDERPAIDj,t6PPSj,t7SIZEj,t8MTBj,t

9SEGMENTj,t10FIRM_AGEj,t + INDUSTRY DUMMYj,t + ej,t (5)

READj,t = β0 + β1POSTj,t2HIGH_DEDj,t 3POST*HIGH_DEDj,t 4OVERPAIDj,t5UNDERPAIDj,t6PPSj,t7SIZEj,t8MTBj,t

9SEGMENTj,t10FIRM_AGEj,t + INDUSTRY DUMMYj,t + ej,t (6)

Where the HIGH_TOP5 in equation (5) is a dummy variable equals to one if the annual average ownership of the firm’s top five institutional investors is greater than the

sample mean calculated using the average ownership from the year 2011 to 2013, and zero, otherwise. I obtain the ownership of the firm’s top five institutional investors from the Thomson Reuters Institutional Holdings databases.

The HIGH_DED in equation (6) is a dummy variable equals to one if the annual average ownership of the firm’s dedicated institutional investors is greater than the sample

mean calculated using average ownership from year 2011 to 2013, and zero otherwise.

The definition of dedicated institutional investors is based on Bushee and Noe (2000) and Bushee (2001). Dedicated institutions have large average investments in invested firms and extremely low turnover, consistent with a relationship investing role and a

Other variables in equation (5) and (6) are as defined in equation (1). The coefficient

on HIGH_TOP5 (HIGH_DED) measures the incremental CD&A readability of HIGH_TOP5 (HIGH_DED) firms in the pre-SOP period relative to firms with lower

top-five (dedicated) institutional ownerships. The coefficient on the interaction term between HIGH_TOP5 and POST in equation (5) measures the incremental change of CD&A

readability for HIGH_TOP5 firms in the post-SOP period. Similarly, the coefficient on the interaction term between HIGH_DED and POST in equation (6) measures the incremental CD&A readability for HIGH_DED firms in the post-SOP period.

The estimation result for equation (5) is shown in Table 6-1. The coefficient on

HIGH_TOP5 is not significant, suggesting that the CD&A readability of HIGH_TOP5 is

not significantly different from that of firms with low top 5 institutional ownership in the pre-SOP period. In addition, the coefficient on POST is not significant, indicating that the CD&A readability of firms with low top 5 institutional ownership in the post-SOP period does not deviate from their CD&A readability in the pre-SOP period. However, the coefficient on the interaction between HIGH_TOP5 and POST is positive and significant (Coefficient = 0.511; t-statistics = 1.87), meaning that the CD&A readability of HIGH_TOP5 firms has reduced after Say-on-Pay implementation.

The estimation result for equation (6) is also shown in Table 6-1. The coefficient on

HIGH_DED is not significant, suggesting that the CD&A readability of HIGH_DED firm

is not significantly different from that of firms with low dedicated institutional ownership in the pre-SOP period. The coefficient on POST is positive and significant (Coefficient = 0.322; t-statistics = 2.14), indicating that the CD&A readability of firms with low dedicated institutional ownership has reduced in the post-SOP period compared with the pre-SOP period. More interestingly, the coefficient on the interaction between HIGH_DED and POST is positive and significant (Coefficient = 0.599; t-statistics = 1.98),

meaning that the CD&A readability of HIGH_DED firms has also decreased after the implementation of Say-on-Pay, and the magnitude of decrease in CD&A readability in high dedicated firms is greater than that of firms with low dedicated institutional ownership.

Taken together, the above results suggest that the sophisticated institutional investors

with important voting blocs and long-term investment horizons might demand more information on the CD&A section to facilitate their assessment of firms’ compensation

policies for voting purpose. This might contribute to the decreasing CD&A readability after the implementation of Say-on-Pay.

Table 6-1

Regressions of CD&A readability factor score- High versus Low Top 5 and Dedicated Institutional Ownership

Equation (5) Equation (6)

Variable READ Variable READ

HIGH_TOP5 0.040 HIGH_DED 0.300

(0.16) (1.03)

POST 0.128 POST 0.322**

(0.73) (2.14)

POST*HIGH_TOP5 0.511* POST*HIGH_DED 0.599**

(1.87) (1.98)

OVERPAID 0.442 OVERPAID 0.486

(1.00) (1.08)

UNDERPAID 0.015 UNDERPAID 0.015

(0.87) (0.83)

FIRM_AGE -0.036*** FIRM_AGE -0.036***

(-4.08) (-4.12)

INDUSTRY Y INDUSTRY Y

INTERCEPT 2.585** INTERCEPT 2.807**

(2.03) (2.20)

N 1098 1098

adj. R2 0.299 0.299

t statistics in parentheses. * p<0.1, ** p<0.05, *** p<0.01. See Table 4-3 for variables definition.

相關文件