• 沒有找到結果。

A Study of the Relationship among Family Business, CEO Overconfidence and Corporate Innovation

N/A
N/A
Protected

Academic year: 2021

Share "A Study of the Relationship among Family Business, CEO Overconfidence and Corporate Innovation"

Copied!
36
0
0

加載中.... (立即查看全文)

全文

(1)

Chiao Da M泣nagement Review Vol. 33 NO.1. 2013 pp.105 -140

家族企業、經理人過度自信與創新活

動關條之研究

A Study of the Relationship among Family Business

,

CEO

Overconfidence and Corporate Innovation

張力 Li Chang

世新大學 企業管理學系

Department of Business Adminis甘ation, Shih Hsin University 陳怡珮 Yi-Pei Chen

中原大學 財務金融系

Department of Finance, Chung Yuan Christian University 侯敘縛 Chi-Ping Hou

中國科技大學會計系

Department of Accounting, China University of Technology 林翠蓉1 Tsui-Jung Lin

中國文化大學 財務金融學系

Department of Banking and Finance, Chinese Culture University 李毅志 I-Zhe Lee

世新大學 企業管理學系

Department of Business Administration, Shih Hsin Universi句

摘要:本研究皆在探討家族企業、經理人過度自信與創新活動問之關餘,並

針對 2001 年至 2007 年 692 家上市上櫃之電子產業進行分析。研究結果發現,

家族企業較少創新活動,顯示家族企業由於風險考量、資源不足以因應、龐大 研發投入,因比較少進行創新研發活動;其次,其過度自信傾向之經理人創

1 Corresponding author: Department of Banking and Finance, Chinese Culture University, Taipei City, Taiwan, E-mail: lcr5@faculty.pccu.edu.tw

The Authors Thank the Anonymous Referee for Valuable Comments and Suggestions, and the National Science Council of Taiwan for Financial Support Under Contract No. NSC

(2)

106 A Study 01 the Relationsh伊 αmong F amily Business,

CEO Overconfidence αnd Corporate Innovation

新行為較高,顯示過度自信經理人會高估報酬與低估風險的態度,會使其偏 好高風險的投資計畫(如創新活動) ,以滿足其對未來樂觀的預期;最後,在 經理人態度作為家族所有權與創新活動問調節變項方面,本研究進一步發 現,一旦過度自信的經理人加入家族企業後,會弱化家族特性(保守等)對 創新活動的負面影響,而提高創新活動,產生正向調節效果,尤其當企業之 控制權與所有權偏離程度較低峙,經理人過度自信傾向之影響更為明顯。此 外,當家族企業之經理人為家族成員,更會強化經理人過度自信對創新活動 之正向調節效果。 關鍵詞:家族企業;經理人過度自信;創新活動;調節效果

Abstract : This study examines the relationship among family business, CEO overconfidence and corporate innovation based on analysis of a sample (4,504) of listed electronic firms (692) in Taiwan from 2001 to 2007. The results show 曲的

family businesses are less likely than non-family businesses to engage in corporate innovation. The risk aversion and lack of resources that characterize family businesses can lead to weak corporate innovation. In addition, we find that overconfident CEOs prefer innövation activities. It is suggested that overestimation of returns and underestimation of risk due to CEO overconfidence leads to investment in risky innovation activities in order to meet optimistic expectations. Finally, investigation shows that when an overconfident CEO is employed in a family business, hislher overconfidence will have a positive moderating effect that enhances corporate innovation, especially in the firms with relatively little deviations from control rights and ownership. Furthermore, a family member serving as CEO with overconfidence in a family business will strengthen the positive moderating effect of managerial overconfidence on innovatìon actìvltìes.

Keywords : Fami1y business; CEO overconfidence; Corporate innovation,

(3)

Chiao Da Management Review Vol. 33 No.1, 2013 107

1. Introduction

Innovation, which results in changes in c。中orate strategy and competitive advantages, direct1y influences the strategy and the core competence of a firm. However, previous related studies have focused on organizational and environmental features, rather than the individual properties of top managers. Lin

et al. (2005, 2008) found that the manager's properties, such as optimism2, can affect financing and investment decisions. The managers' background and attitudes have an influence on organizational innovation, structure, culture and procedure (Lewin and Stephens, 1994), suggesting that the individual characteristics of top managers play an important role in corporate innovation strategy.

Previous research results have provided support for the e在ect of ownership heterogeneity on corporate innovation strategies. Those in charge of public pension funds tend to focus on intemal innovation while those in charge of professional investment funds prefer extemal innovation (Hoskisson et al., 2002). In addition, the ownership type, family or non-family, plays a vital role in corporate management around the world. Shanker and As甘achan (1996) have found that about 20 million firms are family-controlled. About one-third of S&P 500 corporations are con甘olled by founding families holding 18% ownership. 67% of firms in Australia and 68% in Italy are family-controlled (Anderson and Reeb, 2003). In East Asian countries over 50% of the firms are con甘olled and managed by families (Tan and F 0仗, 2001).

Family business is often characterized as being controlled by family management. In other words, one of the family members is appointed to be the CEO and will become the dominant information provider (Ahlstrom et al., 2004; Lien et α孔, 2005; Chen, 2001). The individual characteristics of family CEOs influence corporate investment decision more than do those of non-family CEOs,

2 Following Heaton (2002) and Malmendier and Tate (2005), we infer the CEOs' beliefs about the

future perfonnance of their company 企om their personal portfolio 仕ansactions. The tenn

“overconfidence" is used in this paper regardless of whether the biased beliefs derive 企om

(4)

108 A Study 01 the Relationship among Family Business,

CEO Overconfidence and Corporate lnnovation

especially when managerial biases exist, such as CEO overconfidence (Malmendier and Tate, 2008). Recent studies have focused on personal characteristics such as the education background (Forbes, 2005) or institutional logic provided by family (Bhappu, 2000)3. Yet, because they have not examined the effect of CEO overconfidence on the decision making process in family and non-family corporations, the effect of CEO overconfidence has not been ascertained in these studies.

Innovation represents the future growth opportunity and survival of a firm (Ayadi et al., 1996; Hill and Snell, 1988). However, innovation is risky and costly, and leads to longer investment horizons without certain revenue, which induces the managers' personality to affect the corporate innovation strategy. Managerial optimism is significantly related to corporate financing decisions (Lin et 瓜, 2005, 2008). Optimistic managers are more likely to accept negative NPV projects, which damages firm value. In family firms, the classic owner-manager conflict (type 1 agency problem) is not present; family CEOs act as stewards, managing their company based on their own expectations and vision (Davis et al., 1997).

Therefo泊, family CEOs may demonstrate different corporate decision making than non-family CEOs in relation to risky investment, because they feel responsible for the family's reputation and the need to maintain a long-term presence (Amihud and Lev, 1999; Fox and Hamilton, 1994). On the other hand, previous studies indicate that people are more optimistic about outcomes that they believe they can control (Weinstein, 1980). Consistent with Weinstein (1980), evidence reveals that managers underestimate the level of uncertainty

,

believing that they have great control over the firm perfonnance than they actually do (March and Shapira, 1987). Managers are more optimistic about outcomes when they are highly committed to a firm's success (Gilson, 1989). Consequently, this study examines the effects of family ownership, and managerial behavior and attitudes on investment decisions for risky projects-corporate innovation activities.

3 Bhappu (2000) have found that the Japanese family provldes the institutionallogic for Japanese

co中orate networks and Japanese management practices. Furthermore, the historical actions of individuals in Japanese corporations are indicative of a strategy to sustain and nurture their social capital.

(5)

Chiao Da Management Review Vol. 33 No.1, 2013 109

In addition, the moderating effect of managerial overconfidence on family ownership and innovation activities is 伽ther tested. The results should fill a gap in the related literature.

Our study makes the following contributions. First, previous studies have emphasized the influence of family ownership on firm performance, rather than on c。中orate innovation. Firm value mainly results 企om innovation activities which enhance competitive advantages and operating performance. However, the uncertainty and substantial capital expenditure required for innovation are inconsistent with other characteristics of family firms such as conservatism and lack of resources. Therefore, it is meaningful to investigate the relation between family ownership and corporate innovation. Furthermore, most firms in Taiwan,

an area famous for its technological orientation, are family-controlled. Our findings should be helpful to assist government authorities and management to obtain a better understanding of corporate innovation.

Second, research related to behavioral finance has focused on the investors' expectation, instead of the managers' considerations. However, recent studies indicate the impact of managers' attitudes on corporate decision. Heaton (2002) have explored the relation between managerial optimism and corporate decision making in efficient capital markets. Based on Heaton (2002), Lin et al. (2005) have found that, in more financing constrained firms, optimistic managers exhibit higher investment-cash flow sensitivity than do non-optimistic managers. Managerial optimism can also lead to pecking order preference in financing decisions (Lin et al., 2008). In addition, Goel and Thakor (2008) have examined the effect of CEO overconfidence on corporate govemance, suggesting that the board will lay off CEOs who demonstrate excessive diffidence and excessive overconfidence. We attempt to examine the impact of CEO overconfidence on corporate innovation, since the individual characteristics of CEOs play an important role in corporate decision making. To the best of our knowledge, we are the first in Taiwan to explore whether CEO overconfidence is one of the dominant factors affecting c。中orate innovation activities.

Third, we mainly focus on the moderating effect of CEO overconfidence on the relation between family business and c。中orate innovation. Regarding innovation

(6)

110 A Study 01 the Relationsh伊 among F amily Business,

CEO Overconfidence and Corporate Innovation

activities, the goals of family business, family mission and inheritance of family descendants, may be different 企om the traditional goal: maximization of stockholder wealth. The related evidence is scarce, and therefore we a仕empt to explore how CEO overconfidence affects innovation activities in fami1y firms.

The sample inc1udes 692 listed and OTC electronic firms from 2001-2007. The final sample consists of 4,504 observations. The results show that family businesses are less likely than non-fami1y businesses to engage in corporate innovation, indicating that this risk aversion and lack of resources of family businesses lead to weakc。中orate innovation. In addition, we find that an overconfident CEO prefers more innovation activities, especially when there is relatively low divergence between control right and ownership, which is consistent with Papadakis and Bourantas (1998). It is suggested that CEO overconfidence causes investment in risky projects (e.g., innovation) in order to meet optimistic expectations. Finally, consistent with the arguments of Zahra (2005), CEO overconfidence will have a positive moderating e旺ect that enhances corporate innovation when they are employed in a fami1y business, especially when there is no or relatively low deviation 企om con仕01 rights and ownership. Furthermore

,

a fami1y member serving as CEO that is overconfident would strengthen his/her positive moderating effect on innovation activities in family firms. We infer that family members have strong incentives to engage in innovation activities in order to create and sustain family wealth (Z ahra , 2005). Moreover, since family CEOs pay much attention to the family's reputation and performance

,

they are likely to be optimistic (Miller and Le Breton-Miller

,

2005). The evidence also shows that non-family CEOs in family business do not have a positive moderating effect on innovation, because they may feel monitored by fami1y members, although family input does mitigate the problem of owner-manager co1iflict in fami1y business. This is consistent with the arguments of Solomon et al. (2003)4.

4 Solomon et al. (2003) argue th剖 it is likely that many company directors in Taiwan are

uncomfortable with high levels of family control. Although they themselves are often selected by family members, they may not be related. They may therefore feel that their position and decision-making ability is hindered by family interference.

(7)

Chiao Da Management Review Vol. 33 No.1, 2013 111

The rest of the paper is organized as follows. In Section 2 we review the literature and develop our hypotheses. In Section 3 we describe our data and methodology. The main results and an analysis of the robustness are presented in Section 4 while some conclusions are offered In Section 5.

2. Literature Review and Hypothesis Development

Innovation activities require long-term investment horizons (Bange and De Bondt

,

1998) and have a high degree of uncertainty (Pisano

,

1989). Also

,

the outcomes are difficult to value. Previous Studies found the retum on R&D investment to be about 330/0, and that it takes at least five years to gradually gain

the profits. Furthermore, expenditure is substantial during the period of R&D development. This ambiguity involved in innovation enhances the managers' power of decision making in R&D, and increases the degree of infonnation asymmetry between owners and managers, resulting in agency problems (Ryan and Wiggins, 2002). In addition, R&D expenditure also worsens information asymmetry and reduces corporate profits, which act as a negative signal to the market to decrease the firm value. It is also possible that the benefits of R&D will not be reco伊拉ed if firms encounter the financial distress or if they have relatively high financialleverage, even though, as suggested by Bhagat and Welch (1995), R&D activities result in the fu仙re benefits. Thus, uncertainty is one of the dominant factors in R&D decision making.

Family-con仕olled firms are prevalent in Asia. Claessens et α1. (2000) find that about 50% of the firms are controlled by families. Concentration of family ownership induces different agency problem (~泊, 2005). Compared to westem family firms, in Chinese society, management in family firms is heavily inf1uenced by Confucianism (Yen, 1994). Chinese are highly committed to their family and the corporation is like an extension of the family system (Zapalska and Edwards, 2001). The relationships between relatives are quite important in Chinese family firms (Chen, 2001). Consequently, the usage of corporate resources is inf1uence by personal familial re1ationships, loyalty and altruism, and these t尬的 elements create value in family firms (James

,

1999a). Therefore

,

(8)

/

112 A Study 01 the Relationsh伊 among F amily Business,

CEO Overconfidence and Corporate lnnovation

families are in an uncommonly powerful position to exert influence and control over the firm, leading to potential innovation differences 企om nonfamily firms.

2.1. The Effect of Family Ownership on Corporate Innovation

We examine the innovation activities in family firms 企om the perspectives of type II agency problem arising from the conflict between majority and minority shareholders

,

altruism and resource-based theory.

Ali et al. (2007)5 suggest that ownership concen甘ation would mitigate type 1 agency prQblems but worsen type II agency problems. Yeh (2005) indicates that when families not only owns but also operate the company, it alleviates type 1 agency problems. Zahra (2005) also reports the same evidence, that when the large shareholder is present in the family firm as management,可pe 1 problems wil1 be mitigated. The alignment of interests between management and shareholders reduces agency costs, which al10ws for the input of more resources into innovation activities. In this scenario, family firms' ownership structure also leads to continuity, encouraging investment in long-term development, rather than short-term retums (Zahra et a孔, 2008). However, it has been argued in several studies that the concentration of ownership in family firms induces a tradeoff between type 1 and type II agency problems (Anderson and Re的, 2003; Tsai et al叮

2006; Vil1alonga and Am泣, 2006). The large shareholder may use their controlling position in the firm to extract private benefits at the expense of the small shareholders, especial1y from short-term and risky investments. Therefore, family firms are less likely to engage in long-term R&D activities when ownership is concentrated and there is divergence between control rights and ownership.

As for altruism 6, founding families view their firms as assets to pass on

5 Ali et al. (2007) suggested that the non-family firrns have more serious type 1 agency problems but less serious type II agency problems than family firrns do. Specifically, the difference in agency costs between family and non-family firrns due to Type 1 agency problems dominated the difference in agency costs across family and non-family firrns due to Type II agency problems.

6 Schulze et al. (2002) defmed altruism as a moral value that motivates individuals to undertake actions that benefit others without any expectation of extemal reward. Altruism thus compels parents to transfer resources to their children, since to refrain from doing so would harrn the altruist's welfare.

(9)

Chiao Da Management Review 均1.33 No.1, 2013 113

down to their descendants, instead of wealth to consume. The survival of the finn is thus an important concem for fami1ies (Zahra et al., 2000; Anderson and Re鉤,

2003), giving them the potential to advocate long tenn value maximization. Altruism induces common beliefs among family members and strengthens the link between the children's welfare and the family's. This belief encourages family members to align managerial attitudes toward long-tenn investment with growth opportunities, instead of short-term benefit (Bruton et 仗, 2003). However, family firms might wel1 reduce investment in R&D to avoid the loss of family wealth under inheritance considerations (Shanna et al., 1997). Yeh et al. (2001) also argue that fami1y-controlled Taiwanese companies tend to be conservative and therefore less likely to invest when risk is high, which might harm the firm value. Furthermore

,

family members are highly committed to their corporate mission

,

employees and shareholders (Mil1er and Le Breton-Miller

,

2005). 1n short

,

the goal of the chairman and management is organizational and maximization of family wealth. Family firms tend to evaluate investment projects prudent1y in order to achieve the common vision among family members

,

thereby reducing R&D investment (Donaldson and Davis, 1991; Davis et al., 1997; Fox and Hamilton, 1994).

From the resource-based perspective, substantial human and financial resources are needed for innovation

,

but this does not guarantee the achievement of a competitive advantage. It is also necessary to create corporate core competence and promote investment efficiency. Family frrms suffer weak innovation as a result of the lack of human and financial resources. Semkow's (1994) findings show that when non-family employees are interfered with by the family or when family members serve in management, this has a negative impact on firm value. Family management seeks to avoid extemal finance in order to maintain their positions and control rights (Górriz and Fum訟, 2005; Sirmon and Hi仗, 2003), thereby decreasing their capital resources. James (1999b) has suggested that family companies rely on intemal resources, instead of extemal market resources, for investment. Therefore, family firms tend to utilize their funds more prudent1y when involved in risky investment-innovation.

(10)

114 A Study of the Relationship among Family Business,

CEO Overconfidence and Corporate Innovation

To sum up

,

according to resource-based theory

,

family firms su旺er weak innovation due to the lack of human and financial resources. On the other hand, from the perspectives of agency theory and altruism, family firms experience both positive and negative incentives to be involved in innovation. However, the regulatory and cultural environments also have an influence on c。中orate

investment decisions. Weak investor protection exists in Taiwan because multiple share classes with pyramids and crossholdings are prevalent in Taiwanese family firms (Yeh et 仗, 2001; Lee and Ma, 2006), and the c。中orate govemance mechanism is immature (Hung, Chen and Ke, 2005). In this type of situation,

agency conflicts between controlling and minority shareholders worsen (Wei and Zhang, 2008), which facilitates the expropriation of minority shareholders. Thus,

family firms are less likely to engage in long-term and risky innovation activities.

Hl: Family businesses are less like砂 than non-family businesses ω engage

ìn corporate ìnnovatìon.

2.2. The Effect of CEO Overconfidence on Corporate Innovation

Behavioral approaches are now common in asset pricing, but little work has been done in the area of corporate finance (Heaton, 2002). Several recent studies have applied behavioral approaches to the process of making corporate financial decisions. According to Heaton (2002) and Lin et a1. (2005, 2008) with managerial optimism

,

there is a tradeoff effect between underinvestment and overinvestment related to free cash f10w without invoking asymmetric information or rational agency costs.

The degree of managerial overconfidence depends on a self-serving bias,

including biased inferences or assumptions. The link between managers' decisions and firm benefits induces a self-serving bias for en甘epreneurs starting their own business. Therefore, overconfidence is embodied not only in the way a business is started but also the way corporate decisions are made. There are three features of overconfidence. First, managers are more optimistic about the outcomes that they believe they can control (Weinstein

,

1980)

,

and when they have greater control over firm performance (March and Sha中間, 1987). Second, managers are

(11)

ChiaoDaManα!gement Review Vol. 33 No.1, 2013 115

general1y highly committed to the firm's success due to their wealth, professional reputation, and employability (Gilson, 1989; Weinstein, 1980). Third, managers are prone to overconfidence and exaggerate their skills when sources of private information are too abstract or too ambiguous to be confmned (Alicke et 仗, 1995; Larwood and Whittaker

,

1977). Jensen and Meckling (1976) and Rol1 (1986) have suggested that managers tend to be overconfident due to self-hubris.

According to Heaton (2002),企ee cash flow results in overinves伽ent

because optimistic managers think that this type of cash flow is flexible

,

and they believe that the capital market underestimates firm value, which however is overestimated by optimistic managers. In other words, managerial overconfidence leads to investment in negative NPV projects when 企ee cash flow is sufficient.

Innovation motivates corporate growth, but the process is complex and success if difficult to achieve (Avermaete et al., 2003). However, overconfident managers believe that they can con甘01 this high degree of uncertainty (Weinstein, 1980) so are willing to engage in innovation. In addition

,

optimistic managers are confident of positive outcomes (Weinstein, 1980). Their professional reputation relies on the firm's success. Hence

,

overconfident managers prefer innovation because they would like to maintain and create individual wealth and reputation, and they believe they can control the outcome of innovation.

H2: An overconfident CEO prefers more innovation activities.

2.3. The Relationship among Family Ownership

,

CEO

Overconfidence and Corporate Innovation

Family CEOs evaluate long-term investment projects prudently based on altruism, and therefore will not pursue short-term projects with high returns but high risk. Moreover, family firms emphasize good s甘ategy development for firm survival and ensured inheritance. They may suffer from fmancial cons甘aints

because they are reluctant to raise funds external1y in order to maintain their

con仕ol1ing positions. Consequently

,

family firms are less likely to engage in innovation.

(12)

116 A Study 01 the Relationship among Family Business,

CEO Overcon月dence and Corporate Innovation

However, overconfident CEOs believe that the outcomes are actually under their control and are therefore willing to be involved in innovation activities (Weinstein, 1980). The employment of overconfident CEOs has an impact on the family firm's strategy and management style and enhances innovation investment.

We thus hypothesize that CEO overconfidence positively moderates the influence of family business on corporate innovation.

H3: CEO overco昕'dencepositively moderates the influence of family business on corporate innovαtion.

3. Research Method

3.1 Data Sampling and Modeling

The firm-year sample in this study is comprised of listed electronics industry firms in Taiwan 企om 2001 to 2007. The financial information is acquired from the Taiwan Economic Journal (TEJ) database; patent infonnation is from the Intellectual Property Office ofthe R.O.C.

R&D and innovation have widely been recognized as the driving factors for creating value and sustaining competitiveness in the electronics indust哼" as well as important input for management and business performance. Therefore, the intension of R&D in the electronics industry in Taiwan is higher than 3.5% (Lee and Su, 2009).

Liu et al. (2005) study the R&D performance of listed companies in Taiwan.

They find the number of accumulated patents in the elec甘onics indus甘Y to be 14,847 but there are only 2,757 in the other industries. This indicates that the innovation input in the electronics indu的y is seven times higher than in other industries. Therefore

,

we select the electronics industry as our sample to examine the relationship among family business, CEO overconfidence and corporate lnnovat1on.

Previous literature suggests that most family electronics firms are not innovation-oriented (Upton et al., 2001; Morck and Yeung, 2003). On the other hand, Zahra (2005) finds that family firms are more likely to involve in

(13)

ChiaoDα Management Review Vol. 33 No.1, 2013 117

innovation. These inconsistent results are induced by indus甘y differences. Therefore

,

in this study we test family innovation only in the elec仕onics indus甘y

to avoid cross-industry interference.

This study examines the relationships among family business, CEO overconfidence and corporate innovation, following the work carried out by Tidd (2001), Liu et al. (2005) and Wang et al. (2008). The fixed year effect is also controlled (Galasso and Simcoe, 2011). Our model can be specified as follows:

INNOit=B

o

BjFBitB20CitB3FB*OC 手B4GRit 手B5ROEit

BJ1Tit 手B7Tobin 's Qit 手B8DEBTit 手B9}告。ri +εit ••• A 、‘,/

/'EK

INNOit=Co CjFBCEOitC20Cit C3FBCEOitxOCit C4GRit

+C5ROEit C♂Tit+C7Tobin

s

Qu 手 Csf)EBTu 7ι C9}切ri 爭的 (2) where i and t represent firm and year, respectively;εit and πit are the residuals of the model; while εit and 1tit '-""N(O, 1).

Model (1) is used to test hypothesis 1, which examines the effect of fami1y business and non-family business on innovation. We expect a negative coefficient for B1 and a positive one for B2 . The positive moderating e宜的t{B3) is predicted.

Model (2) is used to veri命 the effect of overconfident fami1y CEOs on innovation. According to the degree of ownership divergence, we test whether the deviation of control 企om cash flow rights will affect innovation in family finns. First, the sample is divided into two groups according to zero or non-zero deviation. Furthermore, in the non-zero deviation group, the sample is divided into high and low deviation subgroups based on the median.

3.2 Dependent and Independent Variables

(1) Innovation (INNO)

We use the research and development expenditures divided by net total sales as a measure of innovation, followed by Tidd (2001); Liu et al. (2005); Wang et al. (2008).

(2) Family business (FB)

(14)

118 A Study 01 the Relationsh加 among Family Business,

CEO Overcon月dence and Corporate Innovation

directors occupied by family members exceeds 50%7 and the sum of the voting power held by family members exceeds 20%8. The criteria for judging who has

ultimate control" are derived from those used by La Porta et al. (1999) and Yeh

(1999).

(3) CEO overconfidence (OC)

Following Malmendier and Tate (2005)

,

our measures of CEO

overconfidence are built on CEOs' personal investment decisions to cons甘uct a proxy for overconfidence, or systematic over-estimation of the retums to holding stock in their own firm. Specifically, we consider the subsample of CEOs who keep their position for at least 3 years. CEOs are identified as overconfident if they are net buyers of company equity during these years, that is, if they are net buyers of stocks in more years than they sell during this time period. CEO overconfidence (OC) is indicated by the dummy variable 叮";"。"otherwise.

(4) Family CEO (FBCEO)

If a family member serves as CEO in the family firm, then we set FBCEO to be 1; "0" otherwise.

3.3 Control Variables

We inc1ude a variety of control variables obtained 企om the literature review in our model, including growth (GR), firm profitability (ROE), patents (PT), firm value (Tobin's Q), debt ratio (DEBT), the ratio of deviations of control from cash flow rights (DEV).

(1) Growth (GR)

Anthony and Ramesh (1992) show that capital expenditure is a function of firm life cycle. Growth (GR) controls for the effect of firm life cycle on innovation. The variable is defined as the growth rate of the realized gross profit margin (Anthony and Ramesh, 1992; PO間, 1980).

(2) Firm profitabili可 (ROE)

7 Following Lee and Ma (2006), most ofthe spouses involves in the management or the boards of directors in Taiwanese family groups.

8 Following La Porta et al. (1999), we use the summation of direct and indirect ownerships to measure this voting power. The indirect ownership represents the ownership in the end of the controlling chain.

(15)

ChiaoDaA也nagement Revi的v Vol. 33 No.1, 2013 119

Branch (1974) find that profits may influence subsequent R&D, and R&D may influence subsequent profits. Profitability is measured as net income after taxes divided by total equity (Young et α1. , 2008).

(3) Patents (PT)

Lev (2001) argues that patents may prevent other firms from copying and are needed in order to maximize profits of R&D. Therefore

,

patents are the middle product derived‘from R&D and will push businesses to engage in more corporate

R&D and innovation. However

,

Liu et al. (2005) indicate that the output of R&D can feedback to the original factor, then be revised and improve management factor which are helpful for increasing performance.

(4) Firm value (Tobin's Q)

Firms with be此er growth opportunities should exhibit higher expenditures for R&D. We use Tobin's

Q

to measure firm value, which represents the firm's market value of equity plus its book value of debt less its book value of current asset divided by the book value of total assets (Lang and Stulz, 1994; Yermack,

1996; Tu et 瓜, 2002; Wang et 仗, 2008). (5) Debt ratio (BEBT)

Zantout (1997) has pointed out that there is a positive re1ation between debt ratio and R&D expenditure, and supports the hypotheses of debt as a monitor. We control for the effect of debt on innovation using total debt divided by total asset (Wang et 仗, 2008; Young et 仗, 2008; Zantout, 1997).

(6) Deviations of control 企om cash flow rights (DEV)

Following La Porta et al. (1999)

,

Yeh and Woidtke (2005)

,

we estimate the deviation of control 企om cash flow rights captured as control rights less cash flow rights.

3.4 The moderating Effect

A moderator is a quantitative or qualitative variable which impacts the direction and/or strength of the relation between an independent variable and a dependent variable. Moreover, a moderator is a variable which alters the direction of the relation between a predictor and an outcome (Baron and Kenny,

(16)

120 A Study 01 the Relationsh加 among Family Business, CEO Overco,!月dence and C。中orate Innoνation

The study investigates the impact of the type of ownership on innovation activities. The interaction term, FB *OC, is used as a moderator to predict the effect of managerial overconfidence on innovation activities. Weinstein (1980) argues that overconfident managers tend to believe that fu仙re outcomes are under their control and so are likely to engage in innovation activities. If an overconfident manger is employed in a family business, this can change strategy and management and affect innovation activities. Managerial overconfidence can moderate the effects of ownership style on innovation. Therefore, we expect a positive coefficient for the interaction term FB*OC and B3 to be positive.

In addition, we want to test the specific effect of CEO overconfidence in family firms. When family members serve as CEO, the FBCEO is set to be 1; 0 otherwise. In family businesses, most non-family CEOs are employed by family directors and therefore they are often monitored by family members, which hinders their attitudes and personality. On the other hand

,

family CEOs operate family business for firm survival, thus they will emphasize firm reputation and performance. Therefore, they will tend to be overconfident about their decisions (Miller and Le Breton-Miller, 2005). We expect the fami1y CEOs with overconfidence to have a positive effect on innovation, that is, C3 is positive.

4. Empirical Results

4.1 Descriptive Statistics

Our sample incIudes data on 692 electronics firms listed in Taiwan 企om

2001 to 2007. Excluding fmns with missing data, we have a total of 4,504 observations. It can be seen in Table 1 that the mean of family firm (FB) is 26% meaning presents that 26% of the firms in the electronics indus甘y are family firms. The mean of the CE的 overconfidence (OC) 的 34%, indicating that 34% of CEOs in the electronic indus甘y are overconfident. The mean (median) ratio of innovation activities (INNO) is 5% (3%) meaning that electronics firms rely on innovation to enhance competitive advantage. The mean of deviation level (DEV) between con仕01 rights and cash f10w rights is 6.67%. Also, the minimum ratio is 0 (no deviation), while the maximum one is 95.59%. Moreover, 19.400/0 (874/4,504)

(17)

Chiao Da Management Reνiew Vol. 33 No.1, 2013 121

of finns in our sample have a zero deviation level meaning that around 80% of electronics finns have a deviation from control and cash flow rights. The mean of growth (GR) is 15%. The mean of profitability (ROE) 的 6% which is similar to the innovation ratio. The mean (median) of the number of patents (PT) per year is 51.91 (2.00). This indicates that electronics finns focus on applying patents. The mean (median) of finn va1ue and debt ratio is 0.71 (0.73) and 0.39 (0.38),

respectively.

Table 1

Descriptive Statistics

Variable Mean Med. Min. Max. Std.

Family Businesses (FB) 0.26 0.00

1.00 0.44

Overconfidence (OC) 0.34 0.00

1.00 0.47

Innovation (INNO) 0.05 0.03

7.00 0.14

De吋ation (DEV) 6.67 2.18

95.59 11.23

Growth (GR) 0.15 0.18 -1.82 0.99 0.29

Firm Profitability (ROE) 0.06 0.11 -9.04 1.90 0.41

Patents (PT) 51.91 2.00 0.00 15374 501.55

Firm Value (Tobin's Q) 0.71 0.73 0.04 1.00 0.15

DebtRatio (DEBT) 0.39 0.38 0.00 1.89 0.17

Note: Innovation represents the ratio of R&D to Net Sales; Patent's measurement unit is number; deviation's measurement unit is %.

In addition, the sample is classified into two groups based on median of

con仕ol-ownership deviation. The mean value of deviation in family finns is 28.240/0 with high deviation and 2.11 % in family finns with low deviation. This

indicates that the deviation of control 企om cash flow rights exists in family firms in the electronics industry. Furthermore

,

according to previous studies

,

family firms are usual1y associated with type 11 agency problems (La Porta et al., 1999; Tan and Fock, 2001). Kao et al., (2006) find that the deviation of control 企om

cash flow rights is large in the electronics indus甘y and families acquire con仕ol1ing positions through complex ownership structure and pyramid structure. To sum up, although only 26% of finns in the elec仕onics industry are family firms

,

the deviation of control 企om cash flow rights can be at甘ibuted to family businesses in the electronics indus甘y.

(18)

122 A Study 01 the Relationship among Family Business,

CEO Overcof1:,月dence and Corporate lnnovation

Table 2

Pearson's Correlation

OC FB INNO DEV GR ROE

OC 1 FB -0.04* 1 INNO 0.07料 -0.06** 1 DEV 0.01 0.10** 0.38* 1 GR 0.02 -0.04料 0.02 0.04** 1 ROE 0.00 0.02 -0.06** -0.02 0.06** 1 PT 0.06** 0.05** -0.00 0.05** 0.00 -0.01 1 PT Tobin's Q DEBT Tobin's Q 0.06 料 -0.01 0.17 串串 0.01 -0.01 0.19** 0.04* 1 DEBT -0.08** 0.02 -0.17** -0.02 -0.01 -0.26** -0.03 -0.83** 1 Note: ***, **, and * indicate the significance level at the 1 %, 5%, and 10%, respectively.

Table 2 shows that overconfidence (OC) is significantly and positively correlated with innovation (INNO). There is a significantly negative relation between family business (FB) and innovation (INNO) 的 the 5% level

,

consistent with our hypothesis. In addition

,

the deviation level (DEV) is significantly and positively linked to family business (FB), providing an important relation between deviation and family businesses. Most correlation coe宜icients of con甘01 variables (except the DEBT and Tobin's Q) are low, showing there are lesser multi-collinearity problems. Moreover

,

the study examines the Variance Inflation Factors (VIF) of variables; the results reveal that all values are less than 10,

indicating negligible multi-collinearity problems.

4.2 Univariate Test

It can be seen in Table 3 that the mean difference in innovation activity (CEO overconfidence) between family and non-family business is 0.35 (0.31),

significant at the 5% leve1. Family businesses have less enhanced innovation and hire less overconfident CEOs than nonfamily businesses. These results are consistent with previous findings in Goel and Thakor (2008). Furthermore, family businesses prefer to employ family members to be their CEOs, which do not consider their ability and performance (Schulze et al., 2003). In contra泣,

(19)

Chiao Da Management Review Vol. 33 No.1, 2013 123

Table 3

Univariate Analysis - Variable Comparison

Variable Family business Non-family business Mean difference (t-value) OC 0.31 0.35 -0.04

**

(-2.57) INNO 0.03 0.05 -0.01

***

(-4.83: DEV 8.52 6.02 2.50

***

(5.24: GR -5.55 21.42 -26.96

***

ROE 0.06 0.05 -0.01 (1.29 PT 93.69 40.96 52.73

***

(2.99) Tob妞 's Q 0.70 0.71 -0.01 ( -0.28: DEBT 0.38 0.37 0.07 (1.41)

Note: t-values are presented in parentheses; ***, **, and

*

indicate the significance level at the 1 %, 5%, and 10%, respective1y.

based on ability and expe此 knowledge. Therefore

,

there are more overconfident

CEOs in non-family businesses than in family businesses.

In addition

,

the results from the effect of ownership style on innovation

show that the innovation ratio in family businesses is 0.03, which is significantly

lower than the 0.05 of non-family businesses at the 1 % level. These results

support our hypothesis. This suggests that family businesses engaged in less R&D

,

because of an aversion to risk and scare resources. The control variable results

indicate that there is a significant difference in growth (GR)

,

patents (PT) and

deviation (DEV) between family and non-family samples. This means that

differences in ownership structure will affect management, especially, family and

nonfamily businesses9• More importantly, the deviation level (DEV) of control

and cas:tJ. f10w rights is larger in family than nonfamily businesses. This suggests a

significant difference in the deviation between family and nonfamily businesses.

Thus, we go a further step to examine the deviation level.

(20)

124

4.3 Tests of the Hypotheses

A Stu咚J 01 the Relationsh伊 among Family Business, CEO Overconfidence and Corporate lnnovation

An OLS regression. is used with a time fixed effects approach to estimate

the effect of CEO overconfidence on the relation between family ownership and

lnnovat1on.

Table 4

Family Ownership

,

CEO Overconfidence and Innovation (modell) - Full Sa旦~ Variables I 11 111 IV C 0.07** 0.06 0.06* 0.04** (2.20) (1.59) (5.39) (2.33) FB -0.02*** -0.02*** -0.01 *** ( -3.58) (-3.11) (-4.39) OC 0.02*** 0.02*** 0.00 (2.98) (2.87) (0.59) FBxOC 0.01 ** (2.04) GR 0.00 0.00 0.00 0.00*** (1.20) (1.14) (1.02) (18.93) ROE -0.04*** -0.04*** -0.04*** -0.05*** ( -6.41) (-5.78) (-5.71) (-14.82) PT 0.00 0.00 0.00 0.00* (0.79) (0.49) (0.60) (1.89) Tobin's Q 0.06** 0.06* 0.06** 0.02* (2.08) (1.86) (1.92) (1.71) DEBT -0.13*** -0.13*** -0.13*** -0.06*** (-4.72) (-4.14) (-4.06) (-4.79)

YEAR Yes Yes Yes Yes

Adj.R2 0.05 0.05 0.05 0.27

F-value 13.97 11.12 11.04 64.47

Obs. 3081 2481 2481 2481

Note: This table presents the regression results for family frrm and CEO overconfidence on innovation for 692 TSE electronics frrms over the period 2001-2007. The dependent variable is innovation, measured as R&D expenditure divided by net total sales. The independent variable is FB, which equals 1 if the number of seats on the board of directors occupied by the family members exceeds 50% and the sum of the voting power held by family members exceeds 20%; OC, equalsl if CEOs were net buyers of company equity for three years; FB*OC, represents the moderating effect of CEO overconfidence between family business and innovation; GR is the growth rate or realized gross profit margin; ROE,

net-income after taxes divided by total equity; PT indicates the number of patents; the Tobin's Q is measured by the market value of equity plus the book value of debt less the book value of current assets divided by the book value of total assets; DEBT is total debt divided by total assets; YEAR indicates the year fixed effect; t-values are presented in parentheses; *** ** and * indicate significance at the 1%, 5%, and 10% levels,

(21)

Chiao Da Management Review Vol. 33 No.1, 2013 125

Table 4 shows the results for model 1. In column 1, we can see that there is

a significantly negative relationship between family business (FB) and innovation

(INNO). The coefficient ofFB is -0.02 and significant at the 1 % level, suggesting

that family businesses engage in less innovation. This results support Hypothesis 1. Our evidence shows that family businesses cannot afford to make huge expenditures on innovation because of their lack of human and financial resources.

On the other hand, they prefer to have family members serve in management.

This can result in management.with the lack of professional knowledge and skills.

Moreover, innovation ambiguity and uncertainty mean that in pursuit of firm

survival family firms hesitate to invest in long-term innovation activities.

Column II of table 4 shows that the relation between overconfidence (OC)

and innovation (INNO). OC (0.02) is positively significant at the 1 % level,

suggesting CEO overconfidence. This result supports Hypothesis 2, which posits

that overconfident CEOs are more likely to engage in innovation activities

because they are confident that their decision will lead to the making of huge

profits for the firm. As shown in column II1 of Table 4

,

FB is negatively related to

INNO at the 1 % significance level, and the relation between OC and INNO is

significantly positive at the 1 % level. These results are similar to those in column

II and confirm the robustness. Finally, we test the moderate effect of

overconfidence (OC) on the relation between family business (FB) and innovation

(INNO). The coe旺icient of FB*OC is 0.01 and is significant at the 50/0 level,

which supports H ypothesis 3. The evidence shows that CEO overconfidence enhances the degree of innovation in family business.

We also control for firm characteristics. 1n column 1V, it can be seen that firm growth (GR) is positive and significant at the 1 % level. This suggests that growing firms need more innovation in order to reach their goals and to create

core value as well as to obtain competitive advantages in the elec甘onics industry.

Firm profitability (ROE) is negatively related to firm innovation at the 1 %

significance level. Branch (1974) shows that profits may influence R&D and vice

versa. Sougiannis (1994) also suggests that R&D has a leading effect on

subsequent net income and stock prices for the next 7 years. Therefore

,

the effect

(22)

126 A Study 01 the Relationship among Family Business,

CEO Overconfidence and Corporate lnnovation

high expenditure and periods of low profit. Thus, innovation had a negative effect on profits in the current year. Lev (2001) argues that patent protection is useful for encouraging innovation and can help to avoid being copied by other firms. However

,

the performance of innovation

,

namely patents

,

can feed back to revise and improve the original impact factor, leading to better performance. That 芯,

patents have a negative effect on innovation (Liu et a孔, 2005). The results show a positive relation between patents (PT) and innovation (INNO) , indicating that patents can increase innovation. Firm value (Tobin's Q) is positively associated with innovation, suggesting that a higher level of firm value will1ead to a higher level of innovation in the electronics industry. The coefficient of DEBT is negative and significant at 1 0/0, indicating that a higher debt ratio represents a

higher probability of bankruptcy which cons甘ains business resources needed for lnnovat1on.

Table 5 shows the relation among family ownership, overconfidence and innovation in family business. Based on the degree of deviation, we also divide the sample into 3 subsamples (no deviation, low deviation and high deviation). Each subsample has 441

,

1

,

078 and 952 firm-year observations

,

respectively.

In table 5, we find that family firms (FB) with no deviation, high deviation and low deviation all have significantly negative effects on innovation (INNO) at the 50/0 level. 1n addition, the coefficients on FB*OC are 0.02 and 0.02 in column

1 (no deviation) and column II (low deviation), significant at the 10% and 5% levels, respectively. The results indicate that CEO overconfidence moderates the relation between family business and innovation, especially when there is no deviation or low deviation in family firms. 1n family firms, owner-manager agency problems are slight, and Type II agency problems between m句 ority and minority shareholders are slight when there is no deviation and low deviation. This be仕er ownership structure leads to the alignment of interests between managers and shareholders, which reduces agency costs and increases family willingness to invest more resources in innovation. These results are consistent with the arguments of Zahra (2005)10.

10 Zahra (2005) found that fami1y members have stronger incentives to engage in innovation

(23)

Chiao Da Management Review Vol. 33 No.1, 2013 127

Table 5

Family Ownership

,

CEO Overconfidence and Innovation (Model 1)

- Full Sample for Different Deviation Levels

Variables I 11 111 IV

No devia位on Low deviation High deviation High deviation

C 0.06* 0.04* 0.05** 0.09*** (1.92) (1.67) (2.52) (4.54) FB -0.02*** -0.01 ** -0.01 ** (-2.84) ( -2.37) (-2.61) FBCEO -0.03*** (-4.89) OC -0.01 0.01 0.01 -0.01 (-1.42) (0.48) (0.94) (-1.27) FBxOC 0.02* 0.02** -0.01 (1.83) (1.97) (-0.96) FBCEOxOC 0.02*** (2.90) GR 0.00*** 0.00*** 0.00*** 0.00*** (5.35) (9.60) (20.64) (3.70) ROE -0.04*** -0.04*** -0.16*** -0.09*** (-5.89) ( -7.61) (-17.56) (-10.78) PT 0.00 0.00 0.00 0.00 (0.19) (0.07) (0.06) (0.06) Tobin's Q 0.02 0.01 * -0.01 0.04** (0.65) (0.59) (-0.46) (2.29) DEBT -0.08*** -0.07*** -0.07*** -0.15*** (-2.89) (-3.28) (-3.49) (-8.55)

YEAR Yes Yes Yes Yes

Adj.R2 0.21 0.19 0.47 0.22

F-value 9.60 19.12 60.25 32.11

Number of

441 1078 952 952

Obs.

Note: This table represents the regression results for family firm and CEO overconfidence on

innovation. N 0 deviation indicates that control rights are equal to cash flow rights; the

subsamples for lower deviation and higher deviation are classified as the median. The

dependent variable is innovation, measured as R&D expenditure divided by total net sales.

The independent variable is FB, which equals 1 if the number of seats on the board of

directors occupied by the family members exceeds 50% and the sum of the voting power held by family members exceeds 20%; OC equals 1 if CEOs were net buyers of company equity during the previous three years; FB*OC shows the moderating etfect of CEO overconfidence between family business and innovation; GR is the growth rate of realized gross profit margin; ROE indicates net-income after taxes divided by total equity; PT is the

number of patents; Tobin's Q is measured by taking the market value of equity plus the

book value of debt less its book value of current asset divided by the book value of total assets; DEBT indicates total debt divided by total assets; YEAR is the year fixed effect;

t-values are presented in parentheses; ***, **, and * indicate significance at the levels of

(24)

128 A Stu吋1 of the Relationship among Family Business,

CEO Overconfidence and Corporate Innovαtion

Table 6

Identity of Family CEO

,

CEO Overconfidence and Innovation (Model 2) -Full Sample with Different Deviation levels

Variables I 11 111 IV

N o-deviation Low deviation High deviation

C 0.05* 0.09** 0.05 0.07** (1.77) (2.17) (1.07) (2.09) FBCEO -0.01 0.01 0.02** -0.01 ( -0.l7) (1.27) (1.99) (-1.11) OC -0.02*** -0.01 -0.01 -0.02** (-2.62) ( -1.26) (-0.64) (-2.l8) FBCEOxOC 0.05*** 0.03** 0.04** 0.03*** (4.74) (2.10) (2.33) (2.86) GR 0.01 *** 0.01 *** 0.01 *** 0.01 *** (9.89) (5.98) (6.10) (9.67) ROE -0.04*** 個0.07*** -0.03*** -0.14 牢牢牢 ( -6.18) (-3.65) (-3.74) ( -8.33) PT 0.01 -0.01 -0.01 0.01 (1.41 ) (-0.84) (-0.69) (1.16) Tobin's Q 0.01 -0.02 -0.04 -0.03 (0.11) (-0.72) (-0.87) (-0.86) DEBT -0.09*** -0.10*** -0.15*** -0.07** ( -4.16) (-2.67) ( -3.46) ( -2.48)

YEAR Yes Yes Yes Yes

Adj. R2 0.26 0.29 0.30 0.41

F-value 22.85 8.33 10.88 14.84

Number of 858 247 329 282

Obs.

Note: This table represents regression of family firms and CEO overconfidence in relation to innovation for 692 TSE electronics firms over the period 2001-2007. No deviation means that control rights are equal to cash flow rights; the subsamples with lower deviation and high deviation are classified as median. The dependent variable is innovation, measured as R&D expenditure divided by total net sales. The independent variable is FBCEO, and is equal to 1 if a family member serves as CEO in a family business; OC equalsl if CEOs were net buyers of company equity for a period of three years in our sample; FB*OC represents the moderating effect of CEO overconfidence between family business and innovation; GR is the growth rate of realized gross profit margin; ROE indicates net-income after taxes divided by total equity; PT is the number of patents; Tobin's Q is measured as the market value of equity plus the book value of debt less the book value of current assets divided by the book value of total assets; DEBT is total debt divided by total assets; YEAR is the yearly fixed effect; t-values are presented in parentheses; ***, **, and * indicate significance levels of 1 %, 5%, and 10%, respectively.

In column III of table 5, the coefficient on FB*OC is negative but insignificant. That 站, there is no evidence to show the moderating e旺ect ofOC on the relation between innovation (INNO) and family (FB) with high deviation. We

(25)

Chiao Da Management Review Vol. 33 No.1, 2013 129

try to take the identity ofthe fami1y CEO (FBCEO) into consideration. The results are shown in column IV. The coefficient onFBCEO*OC is 0.02 and statistically significant at the 1 % level. This suggests that overconfident family CEOs will engage in more innovation. Non-family CEOs may therefore feel their position and decision-making ability to be hindered by family interference (Solomon et a止,

2003), especially when there is higher level of deviation in family firms.

In column 1 and column IV of table 6

,

the coefficients on OC are -0.02 and -002, and are significant at the 1 %and 5% levels, respectively suggesting that overconfident CEOs engage in less innovation especially in high

con甘ol-ownership deviated family businesses. This result is consistent with Solomon et al. 's arguments (2003). 1n column II1, the coefficient for FBCEO is 0.02 and significant at the 50/0 level. That 白, when deviation is low, family CEOs

will more likely to engage in innovation. Low deviation leads to interest alignment in fami1y businesses, which suppo此s family CEOs being involved in more innovation (Astrachan

,

2003).

According to the results in table 6

,

the coefficients of FBCEO*OC are all positive and significant at the 50/0 level. It is suggested that an overconfident

family CEO is a vital factor of influence for innovation and plays a positive moderating role between fami1y business and innovation. We infer that the family tends to support the decisions of fami1y CEOs. 1n order to maintain family reputation and long run operation, fami1y CEOs remain optimistic when making decisions (Miller and Le Breton-Miller

,

2005).

4.4 Robustness Testing

1n this section, we discuss a number of robustness tests undertaken to provide additional evidence in support of our Hypotheses11•

1nstead of R&D expenditure divided by net income, we use R&D expenditure divided by the number of employees as the proxy for innovation (INNO). The results are similar to our main results. That is, CEO overconfidence has a positive effect on innovation. This proxy represents the average R&D

(26)

130 A Study 01 the Relationship among Family Business,

CEO Overconfidence and Corporate Innovation

expenditure per employee

,

which helps measure the impact of CEOs' attitude-overconfidence, on innovation.

As in Malmendier and Tate (2005)

,

CEOs who keep their position for at least 5 years are used to measure CEO overconfidence (OC) in our sample. In models (1) and (2), family business (FB) is found to have a negative effect on innovation and is significant at the 5% level; CEO overconfidence has an insignificantly positive effect on innovation; the FBCEO results are consistent with our main results.

As for the measure of CEO overconfidence, we also use outsider perceptions, cash dividends and eamings forecasting. Following Malmendier and Tate (2008)12, we hand-collected data on how the press por甘ays CEOs during the sample period. We searched for artic1es referring to the CEOs in the KMW and udndata.com news banks. For each CEO and sample year

,

we recorded and analyzed a number of artic1es as containing specific words classified CEO as overconfident or non-overconfident. We analyzed 1,702 articles and constructed three indicators:

Press1"

,“

Press 2" and

both KMW and udndωatωa.c∞om news banks;

Press 2" and

"Press 3" are constructed 企om KMW and udnda前ta.c∞om, respectively. Consistent with our main results

,

the results for model (1) show that family business (FB) has a negative effect on innovation. The results for model (2) are similar to our main results. However

,

the coefficients on FB*OC and FBCEO*OC are all positive but insignificant.

Finally, the change in cash dividend payout ratio is used as a measure of CEO overconfidence. DeAngelo et al. (1996) found that managers tend to overestimate future firm performance and therefore may distribute more cash dividends

,

as the result of CEO overconfidence. Furthermore

,

as in Lin et al.

(2005), we use managers' eaming forecasts as a proxy for CEO overconfidence. After re-estimation we find that the results for models (1) and (2) are consistent with our main results. The coefficients of FB are negative and significant at the

1 % leve1. OC, FB*OC and FBCEO*OC are positive but insignificant.

12 Following Malmendier and Tate (2008), we record the number of articles containing the words

“confident" , “confidence" , “optimistic", “optimism" or containing the words “reliable" , “cautious" ,“conservative" ,“practical" ,“企ugal", or “steady".

(27)

Chiao Da Management Review Vol. 33 No.1, 2013 131

5. Conclusion and Management Implications

5.1 Conclusion and Discussion

This purpose of this study is to examine the relationships among family business, CEO overconfidence and corporate innovation. CEOs of family business are separated into two types: family and non-family CEOs. We

investigate (1) how family ownership affects corporate innovation, (2) how CEO

overconfidence impacts innovation activities, and (3) how CEO overconfidence moderates the relationship between family ownership and c。中orate innovation. The results show that innovation activities are less in family business than those in nonfamily businesses, which is consistent with our expectations. The attitude of risk aversion and a lack of resources can lead to weak corporate innovation in family business. Secondly

,

the evidence reveals that CEO overconfidence prefers more innovation activities, especially in firms with relatively low deviation 企om

control rights and ownership; these results are consistent with our hypothesis and the findings of Papadakis and Bourantas (1998). The implication is that (1) high risk of corporate innovation satisfies the individual properties of an overconfident CEO; (2) the huge investment 企om innovation processes causes CEO to be free rider to reach personal interests; and (3) the high risk of innovation will not directly affect an overconfident CEO's own benefits. Last1y

,

CEO overconfidence has a positively moderating effect on family ownership on innovation activity,

which is consistent with our hypothesis. This shows that CEO overconfidence will enhance c。中orate innovation, when the overconfident CEO is employed in a conservative family business. This evidence mainly comes from family businesses with no or relatively low deviations 企om control rights and ownership. This agrees with the results obtained 企om Zahra (2005). This is because an overconfident CEO tends to make multidimensional and creative decisions (innovation)

,

which is different from conservative family businesses; the perspective agrees with Englmaier's results (2007). This implies that when a company is faced with a high degree of competition or changeable environment,

family businesses may hire an overconfident CEO to provide more faith and trust to fight for competitors.

(28)

132 A Study 01 the Relationship among Family Business,

CEO Overconfidence and Corporate lnnovation

In addition, the results reveal that an overconfident family CEO promotes corporate innovation. CEO overconfidence has a positive moderating effect on innovation activity. However, non-family CEOs are still con仕ol1ed by family businesses (Solomon et 仗, 2003). The special ownership structure of the family business reduces traditional agency problems. Therefore, an overconfident non-family CEO does not increase innovation activity. Overall' family CEOs in the family income focus on reputation and perforrnance

,

and then they will have more confidence to make an aggressive decision. Therefore, an overconfident family CEO will positively moderate relation between a family firrn and lnnovat1on act1vlty.

5.2 Management Implications

This paper applies agency theory to family businesses. Our results can be applied to the practice of employment and monitoring systems. They suggest that it is a good policy for the family business to hire an overconfident CEO to overcome big challenges and strong conipetition. However, this action will worsen agency problems between controlling family shareholders and minority stockholders. If the family business does not have a sound monitoring system,

hiring an overconfident CEO will increase risks. In addition, our evidence shows that CEO overconfidence is the key factor affecting corporate innovation strategy. When investors find targets, they should not only examine their tìnancial condition

,

but also the CEO' s individual character

,

such as

,

overconfidence.

5.3 Limitations and Suggestions

This paper is limited to data for the electronics industry. It is difficult to obtain detailed information about the number of patents in other industries. In

fu仙re research the database could be extended to other industries. In addition, this study adopts the proportion of CEO' s shares as a proxy of overconfidence. However, the CEO's share holdings are affected by other factors. Also, the family maintains a controlling position. Even though the number of CEO held shares are few, the CEO still has controlling power because ofthe entrenchment effect.

(29)

Chiao Da Management Reνiew Vol. 33 No.1, 2013 133

References

Ahlstrom, 缸, Young, M. 瓜, Chan, E. S. and Bruton, G. D. (2004),“Facing Constraints to Growth? Overseas Chinese Entrepreneurs and Traditional Business Practices in East Asia," Asia Pacific Journal 01 Management, 21(3), 263-285.

Ali, A., Chen, T. Y. and Radhakrishnan, S. (2007),

Family 曰lrmF 虹ms," Journal of Accωoun叫lti均n咚~g a仰ndEcωonom1的Cααs, 44(1-2), 238必6.

Alicke, M. D., K.lotz, M. L., Breitenbecher, D. L., Yurak, T. J. and Vredenburg, D. S. (1995),

Ef能f臼ec吼t丸," JhOzu4rmFn1αalolPer.閃'sωO仰nali妙砂砂y αF叫ld8品'ocial Psy戶J旭ch仰0/扣ogy, 的 (5) , 804-825.

Amihud, Y. and Lev, B. (1999),“Does Corporate Ownership Structure Affect Its S仕的gy towards Diversification?" Strategic Management Journal, 20(11), 1063-1069.

Anderson, R. C. and Reeb, D. M. (2003),“Founding Family Ownership and Firm Perfonnance: Evidence from the S&P 500," Journal 01 Finance, 58(3), 1301-1328.

Anthony, J. H. and Ramesh, K. (1992), “Association between Accounting Perfonnance Measures and Stock Prices: A Test of the Life Cycle Hypothesis, " Journal 01 Accounting and Economics, 的(2-3), 203-227.

Astrach侃, J. H. (2003),

F

曰ield," Journ縛7ηnαalofBuω4ι衍Sl切neωss f抬陶台切nt仰ur仰例/.rr叫-i切i的n啥~g, 18(5), 567-572.

Avermaete, T., Viaene, J., Morgan, E. J. and Crawford, N. (2003),“Determinants of Innovation in Small Food Firms

,"

European Journal 01 Innovation Manα~geme.,呵, 6(1), 8-17.

Ayadi, O. F., Dufrene, U. B. and Obi, C.P. (1996),“Firm Perfonnance Measures: Temporal Roadblocks to Innovation?" Managerial Finance, 22(8), 18-32. Bange, M. M. and De Bondt, W. F. M. (1998),“R&D Budgets and Corporate

Eamings Targets," Journal olCorporate Finance, 4(2), 153-184.

Baron, R. M. and Kenny, D. A. (1986),“The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual

,

Strategic

,

and Statistical Considerations," Journal 01 Personality and Social Psychology,

參考文獻

相關文件

Chen, The semismooth-related properties of a merit function and a descent method for the nonlinear complementarity problem, Journal of Global Optimization, vol.. Soares, A new

In summary, the main contribution of this paper is to propose a new family of smoothing functions and correct a flaw in an algorithm studied in [13], which is used to guarantee

In conclusion, we have shown that the ISHE is accompanied by the intrinsic orbital- angular-momentum Hall effect so that the total angular momenttum spin current is zero in a

Over there, there is a celebration of Christmas and the little kid, Tiny Tim, is very ill and the family has no money to send him to a doctor.. Cratchit asks the family

In terms of “Business Model Canvas,” the Value Proposition of Humanistic Buddhism is “to establish the Buddha’s vocation in the world.” Given that a specific target audience

Teacher starts the lesson with above question and explains to students that making business decision is one of the basic functions of a

Key concepts :personal growth (family roles) , family relationship, family problems, social welfare system, interpersonal relationship, communication among family members,

* All rights reserved, Tei-Wei Kuo, National Taiwan University, 2005..