Master Thesis
MSc in International Financial Management 2018-2020
Earnings management and shareholder activism: The impact of foreign directors
Student Number: S3591697 Student Name: Shenyu Chen Supervisor: Dr. C.L.M. (Niels) Hermes
Second Assessor: Dr. Adri De Ridder Word Count: 7450
Date: 10 January, 2020
Abstract: Earnings management is widely implemented by managers, which may be the result of weak monitoring. This study is based on 140 companies to capture the effect of shareholder activism (proxied by shareholder proposals) on earnings management.
This study adds to the literature by focusing on how the presence of foreign directors on the board moderates this relationship. This study investigates the influence of foreign directors by taking the proportion of foreign directors, as well as by measuring the regulation distance of their home countries to the U. S. The result from the empirical analyses implies that shareholder activism can lead to a lower level of earnings management due to improved monitoring. A higher proportion of foreign directors and a closer regulation distance of their home countries to the U. S do not help shareholder activists to reduce earnings management. Decreasing monitoring effectiveness leads to a higher level of earnings management.
Field keywords: Corporate Governance
Keywords: shareholder activism, earnings management, foreign directors, agency
theory, regulation distance
1
1. Introduction
Shareholder activism has become prevalent since the early 1990s, and it plays a significant role in corporate governance to improve firm performance (Goranova &
Ryan, 2014; Dimitrov & Jain, 2011; Brav et al., 2008; Davis & Thompson, 1994).
Shareholders, as the ultimate owners of public companies, have rights to engage in decision makings such as governance structures, management compensation, or corporate restructuration (Cziraki et al., 2009). To impose their suggestions, shareholders file shareholder proposals to exercise their voice (Hadani et al., 2011).
Some studies argue that shareholder proposals enhance the monitoring mechanism among other things which is a useful tool of corporate governance to reduce agency costs (Dimitrov & Jain, 2011; Renneboog and Szilagyi, 2009; Del Guercio et al., 2008).
Others argue that self-serving problems of shareholders and the nonbinding nature of proposals reduce the benefits of shareholder proposals in corporate governance (Lan &
Heracleous, 2010; Bainbridge, 2006). Sun et al. (2013) indicate the importance to address possible effects that shareholders proposals may have on earnings management.
Shareholder proposals may affect earnings management in two directions. On the one
hand, shareholder proposals may constrain managers' ability to engage in earnings
management because an enhanced monitoring mechanism helps to discipline managers
(Sun et al., 2013). On the other hand, shareholder proposals may create unintended
incentives for managers to manipulate earnings for short-term performance (Hadani et
al., 2011). This study investigates the effect of shareholder activism on earnings
2
management. Specifically, this study examines the aggregate number of shareholder proposals because the aggregate number is one of the most direct manifestations of shareholder activism, which may influence managerial behaviors (Hadani et al., 2011).
As for specific types of shareholder proposals, this study also conducts additional analyses on say on pay proposals and proposals excluding CSR proposals. The additional analyses show similar results.
Although the association between shareholder proposals and earnings management has been investigated massively in the previous papers, this study focuses more on whether foreign directors will have influences on the relationship between shareholder proposals and earnings management.
Board of directors plays an essential role in corporate governance, especially for their monitoring role to managers (Fama & Jensen, 1983). In recent years, the board becomes more and more internationalized, and foreign directors on the board are actively playing their role in corporate decision makings (Suh et al., 2018). Foreign directors are valued for their global visions and specialized advice in corporate decision makings while they are also criticized for their lack of local accounting policies and greater information asymmetry (Masulis et al., 2012).
This study investigates the possible moderating effects of foreign directors on the
relationship between shareholder proposals and earnings management. This study
3
focuses on two dimensions: the proportion of foreign directors on the board, and the regulation distance of the foreign directors’ home countries to the U. S. First, this study argues that a higher proportion of foreign directors on the board is beneficial for shareholder activists to constrain earnings management. Hooghiemstra et al. (2019) argue that foreign directors play a more independent role in monitoring management.
If there is a higher proportion of foreign directors on the board, they may evaluate proposals more independently and objectively. Additionally, they may deliver valuable suggestions on proposals to strengthen the monitoring effects. Enhanced monitoring mechanism helps to scrutinize managers further and restrain managers manipulating earnings for self-serving interests. Second, this study argues that if foreign directors’
home countries have a closer regulation distance to the U. S, foreign directors may help shareholder activists to constrain earnings management. Foreign directors are criticized for their lack of knowledge about local regulations. However, if foreign directors’ home countries have a closer regulation distance to the U.S, which means a more similar regulatory environment as the U.S, foreign directors may not be affected by the regulatory differences. They can be quickly integrated into local regulations and help shareholder activists to monitor managers through proposals, which may prevent managers from engaging in earnings management.
In the sample, there are 798 observations of 140 companies from 2013 to 2018. This
study finds a significant negative relationship between shareholder proposals and the
level of earnings management (measured as discretionary accruals). The result supports
4
that the increasing number of shareholder proposals shows an enhanced monitoring mechanism which restrains managers self-serving behaviors, such as earnings management. This study also finds that foreign directors do not help shareholder activists to monitor managers through proposals, which may lead to a higher level of earnings management. Furthermore, the findings support that a closer regulation distance of foreign directors' home countries to the U. S do not enhance their monitoring role. On the contrary, a longer regulation distance helps foreign directors to exert their monitoring role, which is not in line with the predictions. A longer regulation distance from foreign directors' home countries to the U. S enhances their independent visions and helps shareholder activists to reduce earnings management by proposals.
This study provides evidence for the critical role of foreign directors in corporate operations. By investigating two dimensions of foreign directors, the proportion of foreign directors on the board and their home countries' regulation distance to the U. S, this study may be an extension of previous studies on influences of shareholder proposals on earnings management (Park et al., 2013; Hadani et al., 2011; Prencipe et al., 2008).
The next section reviews the relevant literature and states the hypothesis. Section 3 presents the research design. Section 4 reports the empirical analyses and results.
Section 5 provides robustness check. The last section provides conclusions.
5
2. Literature review and hypothesis development 2.1. Shareholder proposals and earnings management
Shareholders, the ultimate owners of public companies, have the rights to engage in corporate decision makings (Cziraki et al., 2009). The separation of ownership and control gives rise to agency problems between shareholders and managers (Jensen &
Meckling, 1976). To resolve agency problems, shareholders seek different measures.
Shareholder activism became prevalent in the early 1990s, and shareholder activists started to enhance their influences to improve corporate governance (Brav et al., 2008;
Davis & Thompson, 1994). Shareholder activism is defined as explicit actions that can influence corporate governance and performance taken by shareholders (Goranova &
Ryan, 2014; Dimitrov & Jain, 2011; Brav et al., 2008; Davis & Thompson, 1994).
Furthermore, Harris and Raviv (2010) maintain that shareholders should play an
important role in corporate governance and shareholder proposals are useful to reduce
managerial agency problems. Submitting proposals is the principal means implemented
by shareholder activists (Monks et al., 2004). Shareholders activists submit proposals
to impose their suggestions which may enhance the scrutiny towards management
(Goranova & Ryan, 2014; Hadani et al., 2011). However, Bainbridge (2006) indicates
that shareholder proposals can be ineffective because of the nonbinding nature, which
may not be beneficial to resolve agency problems. Sun et al. (2013) also indicate self-
serving problems of shareholders. Submission of proposals is often taken as a negative
sign to the public; therefore, shareholders may impose pressure on managers for their
self-serving interests (Lan and Heracleous, 2010).
6
In the study of Sun et al. (2013), they point out the importance to investigate the effect of shareholder activism on financial reporting, such as earnings management. Although previous literature has investigated massively on the effect of shareholder activism on firm performance and other corporate governance issues, accounting issues are often ignored. Healy and Wahlen (1999) define earnings management as decisions that can influence financial reporting positions. Theoretically, earnings management should be valued because better earnings attract investors (Hadani et al., 2011; Prencipe et al., 2008). However, earnings management is often considered as the management actions to achieve smooth earnings, ultimately to influence contractual outcomes that depend on reported accounting results (Cziraki et al., 2009; Prencipe et al., 2008; Del Guercio et al., 2008). In previous papers, the possible motivations of earnings management concern about compensations, debt covenants, and income smoothing (Prencipe et al., 2008).
Shareholder proposals may affect earnings management in two directions which base
on agency theory. On the one hand, shareholder proposals may discipline managers'
self-serving behaviors such as earnings management. There are diverse issues included
in shareholder proposals, from corporate governance to social policy. Behind these
proposals, there are a variety of shareholders with varying equity ownership,
monitoring ability and professional skills (Thomas & Cotter, 2007). With the
development of SEC rules these years, small investors also have their rights to submit
proposals and impose their suggestions in corporate decision makings. Filing proposals
7
give varying shareholders opportunities to exercise their voice and monitor firm operating performance (Cziraki et al., 2009; Davidson et al., 2004). As discussed previously, earnings management may affect firm value and produce financial reporting outcomes that are reversed to shareholder anticipations. Following agency theory, shareholders may submit more proposals to improve the monitoring effectiveness and reduce information asymmetry with managers. Enhanced monitoring mechanism may discipline managers behaviors and restrict them from manipulating earnings for self- serving interests. Although proposals have the nature of being unbinding, the behavior of submitting proposals itself can be a signal of improved monitoring. Managers may receive the signal and feel the monitoring pressure from shareholders. As a result, managers may refrain themselves from manipulating earnings. This line of reasoning suggests the following hypothesis:
Hypothesis 1a. There is a negative association between the aggregate number of all types of shareholder proposals and the level of earnings management.
On the other hand, shareholder proposals may provide incentives for managers to
engage in earnings management. Sun et al. (2013) indicate that those shareholder
proposals may incur unintended consequences that managers are more incentivized to
manipulate earnings. Furthermore, shareholder proposals that are largely associated
with managers' self-interests, such as compensations, may also create incentives for
managers to achieve smooth earnings and diversify risk by selling their equity holdings
8
for better compensations (Healy, 1985). Additionally, Hadani et al. (2011) reveal that instead of only relying on the support of shareholder proposals, shareholder activists also press managers to obtain their well-organized activism attempts. In most cases, these attempts are public, which can draw massive attention from media and other investors. Shareholder activists propose proposals to intensify the public scrutiny that managers face. In this sense, the conflict between shareholders and managers becomes substantially severe, and managers may react negatively by engaging in subsequent earning management. This line of reasoning suggests the following hypothesis:
Hypothesis 1b. There is a positive association between the aggregate number of all types of shareholder proposals and the level of earnings management.
2.2. The moderating effect of foreign directors
The board of directors plays a critical role in firms' corporate governance, which has two significant roles: monitoring role and advisory role (Masulis et al., 2012). The performance of the monitoring and advisory role can primarily affect the corporate decision making and shareholder value creation, and the foreign directors are one unique class which can enhance or weaken board decision making (Masulis et al., 2012).
Within the globalization, the proportion of foreign directors on the board is increasing
and today’s corporations are actively designing international development strategy and
expand their products and services, relying a lot on foreign directors for their global
visions, diverse perspectives, and independent status (Suh et al., 2018). However, there
9
are two contradictory views on the possible influences of foreign directors on corporate decision makings. On the one hand, foreign directors are valuable for their global visions and expert advice to firms, and also, they are more independent from management. They may deliver their suggestions in a more open and less restricted way on the corporate operation. Additionally, a higher proportion of foreign directors on the board may increase independence in the monitoring mechanism within the firm (Klein, 2002). Because foreign directors have a longer distance to management, they will put the interests of the firm at the priority instead of politeness towards other board members (Chiu et al., 2016). When shareholders file proposals to express their suggestions on the corporate operation, foreign directors may inspect these proposals more independently and may also give shareholder activists better suggestions which may improve monitoring effectiveness. Under the improved monitoring mechanism, managers need to take a higher risk of manipulating earnings and may reduce earnings management.
On the other hand, foreign directors are criticized for their weak monitoring role for
several reasons which include geographic distance problem, lack of soft information,
unfamiliarity with local accounting rules and laws, and also language obstacles
(Hooghiemstra et al., 2019; Masulis et al., 2012). First, the geographic distance is an
obstacle that can generate massive costs and undermine the passion of attending board
meetings, which may weaken the monitoring function. Second, foreign directors can be
excluded from obtaining soft information for their fewer appearances and become more
10
"bovine" to local information and operations (Masulis et al., 2012; Coval & Moskowitz, 2001). Third, the unfamiliarity with local accounting rules and laws can also create problems, which can undermine foreign directors' ability and bring problems to managerial decisions (Masulis et al., 2012). Last, Piekkari et al. (2015) indicate that with the acquisition of new foreign directors on the board, corporate boards may need to change their working language which impedes some board members from actively engage in discussion in board meetings. Hooghiemstra et al. (2019) also point out that these discussions may be less nuanced, and board members may not understand each other appropriately. There are apparent differences between native speakers and non- native speakers in understanding and interacting with other members. Hence, being impeded by these obstacles, foreign directors may behave poorly in their monitoring role and may not give valuable advice to enhance the monitoring role of shareholder proposals towards earnings management. Based on the discussion above, this line of reasoning suggests the following hypothesis:
Hypothesis 2. Foreign directors will moderate the association between the aggregate number of all types of shareholder proposals and the level of earnings management, such that the association will be strengthened (weakened) in the presence of foreign directors.
Only focusing on the proportion of foreign directors does not make a whole story. Peetz
(2015) introduces the concept of regulation distance which is defined as the extent of
11
regulation differences. He maintains that a closer regulation distance means high regulation proximity. Following this line, this study focuses on the regulation distance of foreign director' home countries to the U. S. If foreign directors come from countries with long regulation distance to the U. S, they may act poorly and unqualified in monitoring and advisory role. Due to a longer regulation distance to the U. S, foreign directors may be unfamiliar with the U. S laws, regulations, and governance standard (Masulis et al., 2012). Furthermore, the regulations in foreign directors' home countries can even be the opposite of the U. S regulations. Because of this unfamiliarity, foreign directors may have the inadequate capability to inspect and evaluate shareholders proposals properly and enhance the possible monitoring effects of proposals on managers' self-serving behaviors, earnings management. Under the weak monitoring mechanism, managers may engage in more earnings management to serve self-interests.
If a foreign director comes from a country with a regulation similar to the U. S, they may have similar accounting rules, and laws as the U. S. Foreign directors can serve better in their monitoring and advising roles in shareholder proposals and help shareholder activists to impose greater monitoring on managers through the submission of proposals. As a result, managers may refrain them from engaging in earnings manipulation. Accordingly, this line of reasoning suggests the following hypothesis:
Hypothesis 3. Foreign directors will moderate the association between the aggregate
number of all types of shareholder proposals and the level of earnings management,
such that the association will be strengthened (weakened) in the presence of closer
12
regulation distance.
3. Research design
3.1. Data source and sample
This study drew the sample from ProxyMonitor.org for shareholder proposals received by the largest 250 U.S. public companies ranked by the Fortune magazine according to revenue as reported on yearly frequency. The initial sample includes all recorded 3,654 proposals at 284 companies from 2013 to 2018. The company number is bigger than 250 observed firms because of the stretched selection criterion of the Fortune.
For the empirical analysis, the shareholder proposals data were merged with financial data from Compustat; the information of foreign directors comes from MSCI (formerly KLD and GMI); the regulation distance from foreign directors’ home country to the U.
S comes from govindicators.org. Companies with missing data are excluded from the sample. As firm-level analysis required, firm-year observations with missing financial data, and with 0 total assets are dropped from the dataset. Also, given the econometric methodology, firms with less than three years are excluded from the dataset (Camara, 2012). Finally, this led to the final sample of 798 observations in 140 companies.
3.2. Variables
3.2.1. Dependent variable
This study employs the absolute value of the discretionary accruals as a proxy for
13
earnings management (Hooghiemstra et al., 2019; Chen et al., 2015; Peek et al., 2013).
This study identifies the non-discretionary accruals using the Healy model, which is one of the most influential models in measuring earnings management (DeAngelo, 1986). The non-discretionary accruals are calculated as follows:
𝑁𝐷𝐴
𝜏=
∑ 𝑇𝐴𝑡 𝑡𝑇
(1) where 𝑁𝐷𝐴
𝜏denotes the non-discretionary accruals in year 𝜏, 𝑇𝐴
𝑡the total accruals scaled by lagged total assets in 𝑡 years, T years included in the estimation period. The discretionary accruals are residual from the total accruals. The absolute value of the discretionary accruals is the dependent variable (EM_ABSOLUTE).
3.2.2. Explanatory variables
The first explanatory variable is the aggregate number of shareholder proposals (ALL_PROP). This variable intends to measure the impact of all types of shareholder proposals on the level of earnings management. Supported by the study of Cziraki et al.
(2009), the aggregate number of shareholder proposals shows increasing monitoring effectiveness and may have a negative impact on the level of earning management.
This study creates two variables (NX_FOREIGN and REG_FOREIGN) to investigate
possible effects of foreign directors. First, NX_FOREIGN intends to measure the
proportion of foreign directors. Second, REG_FOREIGN intends to measure the
regulation distance to the U. S of the foreign directors’ home countries. To calculate the
average regulation distance to the U. S of firms, the sum of the regulation distance of
14
individual foreign director’s home countries to the U. S is divided by the number of foreign directors. Both variables are implemented to test whether they will strengthen or weaken the relation between shareholder proposals (ALL_PROP) and earnings management (EM_ABSOLUTE).
This study controls firm-fixed effect 𝑓
𝑖, and year-fixed effect 𝑦
𝑡to capture unobserved heterogeneity across firms and time which can affect earnings management as several empirical studies do (e.g., Bourveau and Schoenfeld, 2017; Peasnell et al., 2005).
To test Hypothesis 1, the regression is formulated as follow:
𝐸𝑀_𝐴𝐵𝑆𝑂𝐿𝑈𝑇𝐸
𝑖,𝑡= 𝛼
0+ 𝛼
1𝐴𝐿𝐿_𝑃𝑅𝑂𝑃
𝑖,𝑡+ 𝛽𝑋
𝑖,𝑡+ 𝑓
𝑖+ 𝑦
𝑡+ 𝜀
𝑖,𝑡(2) where 𝐸𝑀_𝐴𝐵𝑆𝑂𝐿𝑈𝑇𝐸
𝑖,𝑡represents the absolute value of discretionary accruals. The firm is indexed by 𝑖 and time is indexed by 𝑡. ALL_PROP represents the aggregate number of shareholder proposals. 𝑋
𝑖,𝑡is a vector of firm-level control variables.
To test Hypothesis 2, the regression is formulated as follows:
𝐸𝑀_𝐴𝐵𝑆𝑂𝐿𝑈𝑇𝐸
𝑖,𝑡= 𝛼
0+ 𝛼
1𝐴𝐿𝐿_𝑃𝑅𝑂𝑃
𝑖,𝑡+ 𝛼
2𝑁𝑋_𝐹𝑂𝑅𝐸𝐼𝐺𝑁
𝑖,𝑡+
𝛼
3𝐴𝐿𝐿_𝑃𝑅𝑂𝑃
𝑖,𝑡∗ 𝑁𝑋_𝐹𝑂𝑅𝐸𝐼𝐺𝑁
𝑖,𝑡+ 𝛽𝑋
𝑖,𝑡+ 𝑓
𝑖+ 𝑦
𝑡+ 𝜀
𝑖,𝑡(3) where 𝐸𝑀_𝐴𝐵𝑆𝑂𝐿𝑈𝑇𝐸
𝑖,𝑡represents the absolute value of discretionary accruals. The
firm is indexed by 𝑖 and time is indexed by 𝑡. ALL_PROP represents the aggregate
number of shareholder proposals. NX_ 𝐹𝑂𝑅𝐸𝐼𝐺𝑁
𝑖,𝑡represents the proportion of
15
foreign directors. 𝑋
𝑖,𝑡is a vector of firm-level control variables.
To test Hypothesis 3, the regression is formulated as follows:
𝐸𝑀_𝐴𝐵𝑆𝑂𝐿𝑈𝑇𝐸
𝑖,𝑡= 𝛼
0+ 𝛼
1𝐴𝐿𝐿_𝑃𝑅𝑂𝑃
𝑖,𝑡+ 𝛼
2𝑅𝐸𝐺_𝐹𝑂𝑅𝐸𝐼𝐺𝑁
𝑖,𝑡+
𝛼
3𝐴𝐿𝐿_𝑃𝑅𝑂𝑃
𝑖,𝑡∗ 𝑅𝐸𝐺_𝐹𝑂𝑅𝐸𝐼𝐺𝑁
𝑖,𝑡+ 𝛽𝑋
𝑖,𝑡+ 𝑓
𝑖+ 𝑦
𝑡+ 𝜀
𝑖,𝑡(4) where 𝐸𝑀_𝐴𝐵𝑆𝑂𝐿𝑈𝑇𝐸
𝑖,𝑡represents the absolute value of discretionary accruals. The firm is indexed by 𝑖 and time is indexed by 𝑡. ALL_PROP represents the aggregate number of shareholder proposals. REG_𝐹𝑂𝑅𝐸𝐼𝐺𝑁
𝑖,𝑡represents the regulation distance to the U. S of foreign directors’ home countries. 𝑋
𝑖,𝑡is a vector of firm-level control variables.
3.2.3. Control variables
Foreign directors may shareholder proposals’ influence on earnings management.
Previous studies also verify many factors that influence the level of earnings management. Therefore, this study controls several important firm-level determinants.
Firm size (FSIZE) is measured by the logarithm of total assets (Hooghiemstra et al., 2019). According to agency theory, there may be a higher agency cost between shareholders and managers in large firms, relative to smaller firms. In order to mitigate agency costs, shareholders from large firms are more incentivized to strengthen monitoring towards managers, which may lead to less earnings management.
Profitability (PROFIT) is measured by the ratio of operating income before depreciation
16
and amortization to the total assets (Park et al., 2013). Profitability should be negatively related to earnings management that profitable firms may contribute more to manage earnings for better performance. Leverage (LEVER) is defined as total debt to total assets (Sun et al., 2013). Firms with higher leverage may engage in more earnings management out of default avoidance. Tangibility (TANG) is measured by the ratio of net property, plant, and equipment to total assets (Titman and Wessels, 1988). Higher tangibility is related to better investment opportunities in the future. In this context, firms with higher tangibility may manage earnings in order to obtain future investment opportunities. R&D intensity (RDINTEN) is defined as the ratio of R&D expenditures to total assets, which is the proxy for growth opportunities (Prencipe et al., 2008).
4. Results
4.1. Descriptive Statistics and correlations
Table 1
Descriptive Statistics
Variable Obs Mean Stdev. Min Max
EM_ABSOLUTE 798 0.022 0.028 0 0.157
ALL_PROP 798 2.228 1.728 0 9
NX_FOREIGN 798 0.171 0.186 0 0.8
REG_FOREIGN 798 0.057 0.179 0.21 0.969
TANG 798 0.594 0.463 0.056 2.079
PROFIT 798 0.151 0.078 -0.112 0.385
RDINTEN 798 0.016 0.029 0 0.147
FSIZE 798 10.166 1.018 8.189 12.755
LEVER 798 0.285 0.154 0.001 0.814
Notes: This table presents the descriptive statistics of key variables. See Appendix A for all variable definitions.
Table 1 presents the descriptive statistics of key variables, including firm-level
variables for the full sample. This study has 798 observations. The average number of
all types of shareholder proposals of the sample firms is 2.228, with values ranging
from zero to nine, which suggests distinct quantitative differences across companies.
17
Table 1 also shows that the maximum proportion of foreign directors is 80 percent in sample companies which is relatively high. Furthermore, the regulation distance of foreign directors' home countries to the U. S has a large range from 0.21 to 0.969, which suggests regulatory differences to the U. S.
Table 2 provides the correlations for the main variables and shows no severe multicollinearity issues. However, some significant correlations are noticeable. The correlation coefficient of earnings management and shareholder proposals is significant at the 5 percent level. The coefficient of shareholder proposals and the proportion of foreign directors is significant at the 10 percent level. R& D intensity, firm size, and leverage all show significant correlations with earnings management.
4.2. Testing hypotheses
4.2.1. Testing hypotheses 1,2, and 3
Table 3 presents the results with absolute discretionary accruals based on the Healy
model as the dependent variable. This study implements a fixed-effects model with
standard errors adjusted for firm-year. All regressions include firm and year fixed
effects. Column 1 shows that the coefficient of shareholder proposals is -0.0017, which
is statistically significant at the 5 percent level. Following agency theory, earnings
management, as a manifestation of agency problems, maybe one of the main targets of
shareholder proposals to reduce agency costs and then improve firm performance
(Marquardt & Wiedman, 2016; Masulis et al., 2012; Jensen & Meckling, 1976).
18
Shareholders may submit more proposals to constrain managers' behaviors of manipulating earnings.
In column 2 of Table 3, coefficients of shareholder proposals and its interaction with the proportion of foreign directors are both negative and not statistically significant.
Column 3 shows that the coefficient of the main relationship between the number of all types of shareholder proposals and earnings management is -0.0013, which is statistically significant at the 10 percent level. This may indicate that the increasing number of shareholder proposals enhance the monitoring effectiveness, which may refrain earnings management.
In column 3, the coefficient estimator of interaction between regulation distance and shareholder proposals is -0.0114, which is significant at a 10 percent level. That is, the negative relationship between the number of shareholder proposals and the level of earnings management enhances marginally with the regulation distance by 0.0114 units.
This coefficient may indicate that if foreign directors come from countries with longer regulation distance to the U. S, they may exert their monitoring roles better and help shareholder activists to reduce the level of earnings management through proposals.
Among control variables, only leverage shows a constant positive and significant
relationship with the level of earnings management.
19
20
4.3. Additional analyses 4.3.1. Say on pay proposals
Shareholders submit different proposals on different issues, including corporate governance, executive compensation, and social policy. In consideration of the fact that say on pay proposals make up 41% of the sample, this study investigates whether the possible association between this specific type of shareholder proposals, say on pay proposals, and the level of earnings management will be consistent with the results of the main analyses.
There is a massive amount of empirical studies also focus on a specific type of shareholder proposals which is the say on pay proposals (Sun et al., 2013; Ertimur et al., 2011; Prencipe et al., 2008). Sun et al. (2013) conclude that say on pay proposals may have the strongest relationship with earnings management. Say on pay proposals relate managers' interests with shareholders' interests and then reduce the incentives of managers to manipulate earnings. With this alignment of interests, managers may refrain themselves from earnings management for better anticipations of compensation.
Additionally, say on pay proposals have the characteristic as other all other proposals
that are the enhancement of monitoring effectiveness. However, this strong alignment
of interests may incentivize managers to engage in earnings management. As discussed
in the previous section, managers may diversify risks and achieve smooth earnings for
good compensations (Cziraki et al., 2009; Prencipe et al., 2008; Healy, 1985).
21
Table 3Regression analysis of impact of all types of shareholder proposals on earnings management.
Notes: Dependent variable is the absolute value of the discretionary accruals based on the Healy model (EM_ ABSOLUTE). Firm-level variables (TANG, PROFIT, RDINTEN, FSIZE, LEVER) are controlled in each regression. See Appendix A for variable definitions.
Standard errors are robust to clustering within each firm. The t-statistics based on standard errors adjusted for firm-year are shown in the table below the coefficient estimates. ***, ** and * indicate significance at the 1%, 5% and 10% levels respectively.
EM_ABSOLUTE 1 2 3 ALL_PROP -0.0017
(-2.02) **
-0.0076 (-0.64)
-0.0013*
(-1.47)
NX_FOREIGN -0.0010
(-0.06) ) NX_FOREIGN
*ALL_PROP -0.0047
(-1.12)
REG_FOREIGN 0.0818
(3.55) **
REG_FOREIGN
*ALL_PROP -0.0114
(-1.97) *
TANG 0.0183
(1.31)
0.0190 (1.36)
0.0196 (1.41)
PROFIT -0.0213
(-0.83)
-0.0202 (-0.78)
-0.0172 (-0.67) RDINTEN 0.0464
(0.21) )
0.0529 (0.23)
0.0771 (0.34)
FSIZE 0.0066
(1.10)
0.0065 (1.08)
0.0076 (1.28)
LEVER 0.0685
(4.50) **
0.0708 (4.60) **
0.0674 (4.47) **
Firm-fixed effects Yes Yes Yes
Year-fixed effects Yes Yes Yes
No. of observations 798 798 798
R-square 0.0026 0.0026 0.0073
There is a massive amount of empirical studies also focus on a specific type of
shareholder proposals which is the say on pay proposals (Sun et al., 2013; Ertimur et
al., 2011; Prencipe et al., 2008). Sun et al. (2013) conclude that say on pay proposals
may have the strongest relationship with earnings management. Say on pay proposals
relate managers' interests with shareholders' interests and then reduce the incentives of
managers to manipulate earnings. With this alignment of interests, managers may
22
refrain themselves from earnings management for better anticipations of compensation.
Additionally, say on pay proposals have the characteristic as other all other proposals that are the enhancement of monitoring effectiveness. However, this strong alignment of interests may incentivize managers to engage in earnings management. As discussed in the previous section, managers may diversify risks and achieve smooth earnings for good compensations (Cziraki et al., 2009; Prencipe et al., 2008; Healy, 1985).
As discussed in the previous section, foreign directors play a significant role in the possible relationship between say on pay proposals and the level of earnings management. Foreign directors may spot the possible unintended consequences of the say on pay proposals and orient say on pay proposals towards long-term visions instead of short-term, which may reduce managers' incentives to engage in earnings management (Chiu et al., 2016). However, they often have higher level of information asymmetry, lack of knowledge in local accounting rules, and interaction difficulties which may impede foreign directors' ability of judgement on say on pay proposals and may not be able to deliver valuable advice on proposals, which may weaken the monitoring role of shareholder proposals (Hooghiemstra et al., 2019; Masulis et al., 2012). The regulation distance of foreign directors' home countries to the U. S may also affect the relationship between say on pay proposals and the level of earnings management. Foreign directors from countries with closer regulation distance to the U.
S means closer regulation proximity with the U. S. This proximity may help them adapt
to local regulations and accounting rules efficiently and reduce possible obstacles
23
brought by other factors.
This study constructs a PAY_PROP variable which is the number of say on pay proposals as a new explanatory variable. This variable intends to measure the impact of say on pay proposals impact on the level of earnings management.
Table 4 presents the results with absolute discretionary accruals based on the Healy model as the dependent variable. The fixed-effects model is implemented with standard errors adjusted for firm-year. All regressions include firm and year fixed effects.
Column 1 in Table 4 shows that the number of say on pay proposals is statistically
insignificantly associated with earnings management. Specifically, say on pay
proposals is negatively but statistically insignificantly associated with the absolute
value of discretionary accruals. In column 2, the coefficient of the number of say on
pay proposals is -0.0214, which is statistically significant at the 1 percent level. This
result is in line with hypotheses development that the increasing number of say on pay
shareholder proposals may enhance monitoring mechanism and constrain managers'
ability to manipulate earnings management. The coefficient of the interaction term
between the proportion of foreign directors and say on pay proposals is 0.0842, which
is statistically significant at the 5 percent level. This coefficient may indicate that the
proportion of foreign directors weakens the negative relationship between the number
of say on pay proposals and earnings management, which is consistent with the
24
argument that a higher proportion of foreign directors may have more disadvantages for lack of local accounting rules, more information asymmetry, and language issues.
These obstacles weaken the monitoring role of foreign directors to inspect say on pay proposals properly, and they may not give valuable advice to help shareholder activists to reduce earnings management by submitting proposals.
Column 3 of Table 4 shows that the coefficient of the interaction between regulation distance and say on pay proposals is -0.1334, which is statistically significant at the 1 percent level. That is, the negative relationship between say on pay proposals and the level of earnings management enhances marginally with regulation distance by 0.1334 units. This coefficient may indicate that regulation distance to the U. S may help foreign directors have better visions and exert their monitoring role properly. Among control variables, leverage shows a constant positive and significant relationship with the level of earnings management. Tangibility shows a positive and significant relation to earnings management in column 3.
4.3.2. Shareholder proposals excluding corporate social responsibility (CSR) proposals According to Monks et al. (2004), there are three broad categories of proposals:
corporate governance, corporate social responsibility, and crossover. Although CSR
activists have the same interests as traditional corporate governance activists in
enhancing transparency and accountability in decision making, the primary topics of
CSR proposals are different from traditional corporate governance proposals. During
25
the six years, of the 798 proposals, 24% are CSR related proposals. There are sixteen subcategories in CSR proposals, such as human rights, environmental issues, employment discrimination, and political spending. Earnings management are used by managers to affect financial reporting outcomes which should not be included in the content of CSR. (Hooghiemstra et al., 2019; Hadani et al., 2011; Healy & Wahlen, 1999).
Table 4
Regression analysis of impact of say on pay proposals on earnings management.
Notes: Dependent variable is the absolute value of the discretionary accruals based on the Healy model (EM_ABSOLUTE). Firm-level variables (TANG, PROFIT, RDINTEN, FSIZE, LEVER) are controlled in each regression. See Appendix A for variable definitions. Standard errors are robust to clustering within each firm. The t-statistics based on standard errors adjusted for firm-year are shown in the table below the coefficient estimates. ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
EM_ABSOLUTE 1 2 3 PAY_PROP -0.0100
(-1.56)
-0.0214 (-2.65) ***
-0.0077 (-1.19)
NX_FOREIGN -0.0941
(-2.42) **
) NX_FOREIGN
*PAY_PROP 0.0842
(2.31) **
REG_FOREIGN 0.1956
(3.67) **
REG_FOREIGN
*PAY_PROP -0.1334
(-2.70) ***
TANG 0.0221
(1.59)
0.0221 (1.59)
0.0243 (1.76) *
PROFIT -0.0145
(-0.56)
-0.0127 (-0.50)
-0.0118 (-0.46) RDINTEN 0.0438
(0.19) )
0.0849 (0.37)
0.0713 (0.32)
FSIZE 0.0066
(1.10)
0.0073 (1.22)
0.0078 (1.31)
LEVER 0.0672
(4.42) **
0.0718 (4.67) **
0.0635 (4.21) **
Firm-fixed effects Yes Yes Yes
Year-fixed effects Yes Yes Yes
No. of observations 798 798 798
R-square 0.0009 0.0013 0.0053
26
For additional analyses, this study includes a new explanatory variable, EXCSR_PROP.
By excluding CSR proposals from the aggregate number of all types of proposals, this study investigates whether the number of corporate governance and crossover proposals will affect the level of earnings management. The proportion of foreign directors and regulation distance are also included in this regression. Table 5 shows the results.
Column 1 shows that the effects of the number of proposals excluding CSR proposals on the earnings management are not statistically significant. In column 2, none of the coefficients is statistically significant. In column 3, the interaction term between regulation distance and shareholder proposals excluding CSR proposals is also not statistically significant. Among all control variables, only leverage shows a positive and significant relationship with the level of earnings management.
5. Robustness check
To further verify the robustness of the previous findings, this study uses an alternative
measure for discretionary accruals. This study uses the DeAngelo Model to check the
robustness. The DeAngelo Model is also one of the most influential models in
measuring earnings management (Jones, 1991). The model separates total accruals into
discretionary and nondiscretionary accruals and defines the “abnormal” total accruals
as the difference between current total accruals and total accruals from a prior period.
27
Table 5Regression analysis of impact of all types of proposals excluding CSR proposals on earnings management.
Notes: Dependent variable is the absolute value of the discretionary accruals based on the Healy model (EM_ABSOLUTE). Firm-level variables (TANG, PROFIT, RDINTEN, FSIZE, LEVER) are controlled in each regression. See Appendix A for variable definitions.
Standard errors are robust to clustering within each firm. The t-statistics based on standard errors adjusted for firm-year are shown in the table below the coefficient estimates. ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
EM_ABSOLUTE 1 2 3
EXCSR_PROP -0.0023 (-1.33)
-0.0012 (-0.55)
-0.0020 (-1.12)
NX_FOREIGN -0.0087
(-0.53) ) NX_FOREIGN
*EXCSR_PROP -0.0049
(-0.69)
REG_FOREIGN 0.0593
(2.84) ***
REG_FOREIGN
*EXCSR_PROP 0.0014
(0.13)
TANG 0.0197
(1.41)
0.0201 (1.44)
0.0204 (1.47)
PROFIT -0.0188
(-0.73)
-0.0175 (-0.68)
-0.0175 (-0.68) RDINTEN 0.0735
(0.32) )
0.0852 (0.37)
0.0915 (0.40)
FSIZE 0.0074
(1.23)
0.0075 (1.23)
0.0079 (1.31)
LEVER 0.0690
(4.51) **
0.0709 (4.58) **
0.0670 (4.39) **
Firm-fixed effects Yes Yes Yes
Year-fixed effects Yes Yes Yes
No. of observations 798 798 798
R-square 0.0015 0.0015 0.0052
The abnormal accruals are calculated as follows:
𝑁𝐷𝐴
𝜏= 𝑇𝐴
𝜏−1(5)
where 𝑁𝐷𝐴
𝜏denotes the non-discretionary accruals in year 𝜏 , 𝑇𝐴
𝜏−1denotes the
difference of total accruals between year 𝜏 and 𝜏 − 1. The discretionary accruals are
residual from the total accruals. The absolute value of the discretionary accruals is the
dependent variable (DeEM_ABSOLUTE). Other variables are as defined in the previous
study. Generally, the results support previous findings based on the analyses using the
28
Healy model. The results are reported in Appendix B.
6. Conclusion
This study has examined the association between shareholder activism and earnings management. By using shareholder proposals as the proxy of shareholder activism, this study has investigated the possible effects of shareholder proposals on the level of earnings management. As supported by previous studies, shareholder activists play a critical role in influencing earning management. Furthermore, because of the participation of foreign directors, this relationship becomes more attractive to investigate. Through examining 798 samples of 140 companies during 2013 to 2018, this study provides empirical results showing that both the number of all types of proposals and say on pay proposals are associated with lower levels of earnings management. There are negative relationships between the number of shareholder proposals and earnings management. Specifically, this study has investigated the possible moderating effect of foreign directors in two dimensions: the proportion of foreign directors on the board, and the regulation distance of foreign directors' home countries to the U. S. This study finds that a higher proportion of foreign directors impair the monitoring mechanism. The lack of knowledge about local accounting rules or a higher level of information asymmetry can be the reasons behind. Under a loose monitoring mechanism, managers may engage in more earnings management.
Furthermore, this study finds that regulation distance to the U. S from foreign directors'
home countries impedes foreign directors from exerting their monitoring role, which
29
may reduce the monitoring function of shareholder proposals on earnings management.
By constructing new explanatory variables and using an alternative measure of earnings management, this study finds similar results.
This study also has limitations as any study. First, the small sample size is one of the main limitations of this study. Due to limited availability of data, this study only involves 798 samples in 140 U. S companies. For further researches, it needs more comprehensive data and multinational analysis. Second, as mentioned by previous studies, potential measurement errors cannot be prevented in measuring earnings management (Hooghiemstra et al., 2019; Dechow et al., 2010). Third, there are formal and informal proposals. This study is only able to investigate the formal proposals.
Informal proposals, such as private negotiations, can also have specific impacts (Hadani
et al., 2011; Prevost & Rao, 2000).
30 Appendix A. Variable definitions and source
Variable name Measures Data Source
Earnings Management (EM_ABSOLUTE) Discretionary accruals / Total assets Compustat Earnings Management (DeEM_ABSOLUTE) Discretionary accruals / Total assets Compustat All types of shareholder Proposals
(ALL_PROP)
The number of all types of shareholder proposals ProxyMonitor.org
Say on pay proposals (PAY_PROP) The number of say on pay proposals ProxyMonitor.org Proposals excluding CSR proposals
(EXCSR_PROP)
The number of all types of shareholder proposals excluding CSR proposals
ProxyMonitor.org
Foreign Directors – Nationality Mix (NX_FOREIGN)
The proportion of foreign directors MSCI (formerly KLD and GMI)
Foreign Directors – Regulation Distance (REG_FOREIGN)
Regulation distance to the U. S govindicators.org
Firm Size (FSIZE) Logarithm of total assets Compustat
Profitability (PROFIT) Operating income before depreciation and amortization / Total assets
Compustat
Tangibility (TANG) Property, plant and equipment / Total assets Compustat R&D Intensity (RDINTEN) Research and development expenses / Total assets Compustat Leverage (LEVER) Short-term and long-term debt / Total Asset Compustat
31
Appendix B. Robustness results
Appendix B1
Robustness checks using alternative measures for earnings management.
Notes: Dependent variable is the absolute value of the discretionary accruals based on the DeAngelo model (DeEM_ABSOLUTE). Firm- level variables (TANG, PROFIT, RDINTEN, FSIZE, LEVER) are controlled in each regression. See Appendix A for all variable definitions. Standard errors are robust to clustering within each firm. The t-statistics based on standard errors adjusted for firm-year are shown in the table below the coefficient estimates. ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
DeEM_ABSOLUTE 1 2 3 ALL_PROP 0.0011
(1.08)
-0.0007 (-0.47)
0.0009 (0.83)
NX_FOREIGN -0.0003
(-2.07) **
(-0.02) NX_FOREIGN )
*ALL_PROP 0.0096
(1.86) *
REG_FOREIGN 0.0160
(0.55) REG_FOREIGN
*ALL_PROP 0.0078
(1.07)
TANG 0.0331
(1.90) *
0.0321 (1.85) *
0.0329 (1.89) *
PROFIT -0.0317
(-0.99)
-0.0308 (-0.96) (-0.73)
-0.03315 (-1.03) RDINTEN 1.2159
(4.32) **
1.2263 (4.36) **
1.2231 (4.34) **
FSIZE 0.0090
(1.21)
0.0081 (1.08)
0.0091 (1.22)
LEVER 0.0101
(0.53)
0.0118 (0.62)
0.0087 (0.46)
Firm-fixed effects Yes Yes Yes
Year-fixed effects Yes Yes Yes
No. of observations 798 798 798
R-square 0.0286 0.0269 0.0295
32
Appendix B2Robustness checks using alternative measures for earnings management.
Notes: Dependent variable is the absolute value of the discretionary accruals based on the DeAngelo model (DeEM_ABSOLUTE). Firm- level variables (TANG, PROFIT, RDINTEN, FSIZE, LEVER) are controlled in each regression. See Appendix A for all variable definitions. Standard errors are robust to clustering within each firm. The t-statistics based on standard errors adjusted for firm-year are shown in the table below the coefficient estimates. ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
DeEM_ABSOLUTE 1 2 3 PAY_PROP -0.0056
(-0.70)
-0.0149 (-1.48)
-0.0063 (-0.77)
NX_FOREIGN -0.0927
(-1.91) * ) NX_FOREIGN
*PAY_PROP 0.0693
(1.52)
REG_FOREIGN 0.0161
(0.24) REG_FOREIGN
*PAY_PROP -0.146
(0.23)
TANG 0.0317
(1.83) *
0.0319 (1.84) *
0.0319 (1.84) *
PROFIT -0.0344
(-1.08)
-0.0321 (-1.01)
-0.0341 (-1.07) RDINTEN 1.2052
(4.27) **
1.2464 (4.41) **
1.2163 (4.31) **
FSIZE 0.0086
(1.15)
0.0088 (1.18)
0.0088 (1.18)
LEVER 0.0119
(0.63)
0.0177 (0.92)
0.0113 (0.59)
Firm-fixed effects Yes Yes Yes
Year-fixed effects Yes Yes Yes
No. of observations 798 798 798
R-square 0.0268 0.0252 0.0277
33
Appendix B3Regression analysis of impact of all types of proposals excluding CSR proposals on earnings management.
Notes: Dependent variable is the absolute value of the discretionary accruals based on the DeAngelo model (DeEM_ABSOLUTE). Firm- level variables (TANG, PROFIT, RDINTEN, FSIZE, LEVER) are controlled in each regression. See Appendix A for variable definitions.
Standard errors are robust to clustering within each firm. The t-statistics based on standard errors adjusted for firm-year are shown in the table below the coefficient estimates. ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
DeEM_ABSOLUTE 1 2 3 EXCSR_PROP -0.0009
(-0.43)
-0.0034 (-1.21)
-0.0008 (-0.37)
NX_FOREIGN -0.0323
(-1.59) ) NX_FOREIGN
*EXCSR_PROP 0.0125
(1.42)
REG_FOREIGN 0.0287
(1.10) REG_FOEIGN
*EXCSR_PROP 0.0016
(0.12)
TANG 0.0306
(1.76) *
0.0302 (1.74) *
0.0309 (1.78) *
PROFIT -0.0306
(-1.14)
-0.0363 (-1.13)
-0.0359 (-1.12) RDINTEN 1.2186
(4.31) **
)
1.2134 (4.29) **
1.2258 (4.32) **
FSIZE 0.0090
(1.20)
0.0073 (0.98)
0.0092 (1.22)
LEVER 0.0125
(0.66)
0.0150 (0.78)
0.0115 (0.60)
Firm-fixed effects Yes Yes Yes
Year-fixed effects Yes Yes Yes
No. of observations 798 798 798
R-square 0.0251 0.0252 0.0262