Corporate Governance and Returns on Information Technology Investment:

Evidence from an Emerging Market

Joanna L.Y. Ho* The Paul Merage School of Business

University of California Irvine, California

Anne Wu School of Accountancy

National Chengchi University, Taipei, Taiwan

Sean Xin Xu School of Business and Management

Hong Kong University of Science & Technology Clear Water Bay, Hong Kong

*Contacting author

We would like to thank Daphne Chang for her research assistance. This is a preliminary draft. Please do not quote without consent of the authors.

1

Corporate Governance and Returns on Information Technology Investment:

Evidence from an Emerging Market

Abstract Prior studies have reported mixed findings on the impact of corporate IT investment on firm performance. This study investigates the role of corporate governance, an important management control system, in the IT investment-firm performance relationship in the Taiwanese high-tech industries. Two specific corporate governance factors, i.e., board independence and foreign ownership, are explored across firms of different sizes and in industries whose degrees of competitiveness run a wide gamut. Our results show a positive moderating effect of board independence on the IT investment-firm performance relationship, especially when competition intensifies. Furthermore, we find that the association between foreign ownership and firm performance is negative but not significant. Yet, foreign ownership is a significant positive moderator for small firms, suggesting that foreign investors may bring IT expertise to help small firms reap the benefits of using IT. Keywords: IT investment; Board independence; Foreign ownership; Firm performance; Firm size;

Industry competitiveness

2

1. Introduction

Companies invest substantial financial resources in information technology (IT), and the

significant trend of IT investment is expected to continue, especially in emerging markets (IDC

2006).1 However, it is puzzling that despite significant IT investment, there is no consistent

evidence of IT value. A meta-analysis of 66 prior studies (Kohli and Devaraj 2003) shows that

IT investment improves, reduces, or has no effect on firm performance. Particularly in emerging

markets, positive IT returns are yet to be documented (Dewan and Kraemer 2000). The mixed

evidence on IT returns has triggered significant debate on the IT investment-firm performance

relationship (see Brynjolfsson et al. 2002 for a literature review).

Recent research has largely agreed that IT investments seldom act alone, but work in

conjunction with other corporate resources to create value for shareholders (e.g., Brynjolfsson et

al. 2002; Bresnahan et al. 2002; Wade and Hulland 2004). However, corporate governance

structures, one of the important corporate resources, have received little attention in the literature

on IT returns (e.g., Bresnahan et al. 2002; Brynjolfsson et al. 2002). Corporate governance2

should theoretically play a significant role in the IT investment-firm performance (IT-

performance) relationship. The Committee of Sponsoring Organizations of the Treadway

Commission (1992) argues that there is a strong relationship between management control

philosophy and corporate governance. In the context of IT investment, shareholders expect that

utilizing IT can improve firm performance by increasing operational efficiencies or productivity

(e.g., Hitt and Brynjolfsson 1996; Devaraj and Kohli 2003; Banker et al. 2006). However, IT

managers may over-invest to accumulate excessive IT resources for their own interests (e.g., 1 In the U.S., corporate IT investment amounts to $417.4 billion in 2005, representing 3.35% of GDP or about half of the total capital spending in the private sector; worldwide IT spending is projected to increase at a compound annual growth rate of 6% to reach $1.5 trillion in 2010 (IDC 2006). Among all the regions in the world, the Asia/Pacific Region is experiencing the highest growth rate of IT spending, about 9% in 2006 (IDC 2006). 2 Corporate governance refers to “the ways in which suppliers of finance to corporations assure themselves of getting a return on their investment” (Shleifer and Vishny 1997, p.737).

3

Kauffman and Li 2003; Carr 2004), and they may imitate each other even if imitation deviates

from the optimal investment decision (e.g., Scharfstein and Stein 1990; Graham 1999).

Corporate governance may be important because it can mitigate the above agency problems (e.g.,

La Porta et al. 2000; Bushman and Smith 2001).

Specifically, this study addresses the role of corporate governance in the high-tech

industries in Taiwan, an emerging market. We focus on this research setting for both theoretical

and empirical reasons. Theoretically, corporations in emerging markets generally face weak

judicial systems to protect shareholders’ rights (e.g., Gugler et al. 2003). This may lead these

companies to resort to corporate governance mechanisms (e.g., boards of directors) to align

managerial interests with those of shareholders (e.g., Shleifer and Vishny 1997; La Porta et al.

2000). Moreover, prior research in emerging markets reports wide firm-level heterogeneity in

the forms and the effectiveness of corporate governance (Fan and Wong 2005). Therefore,

corporate governance could be a useful lens to understand different IT returns in emerging

markets.

Empirically, starting in 2001, the Taiwanese government conducted annual surveys of the

top 2,000 companies to better understand their IT investment (e.g., software, hardware, and

training costs). Corporate governance information for all Taiwanese publicly listed firms was

also made available in a Corporate Database. These Taiwanese datasets provide a unique

opportunity to use archival data to assess the role of corporate governance in IT returns.3 This is

different from a number of previous studies using primary survey data to explore the IT-

performance relationship, which could involve biases (Banker et al. 2006). Using a meta-

analysis, Kohli and Devarj (2003) find that studies utilizing primary survey data are more likely

3 In contrast, information concerning IT investment and corporate governance of U.S. companies are not available in public datasets.

4

to report positive IT performance impacts than those using archival data. We use the Taiwanese

datasets partly because they can attenuate the above biases and partly because the Taiwanese

high-tech industries play a leading role in the world market.4

We first investigate board independence—i.e., the presence of outside directors on the

corporate board. It is widely argued that independent boards can better perform a monitoring

role, and thus can reduce agency problems in IT investment (e.g., Kaplan and Reishus 1990;

Dalton et al. 1998; Klein 2002). We extend this line of research by examining whether or not

board independence is contingent on industry competition. Theoretical research implies that

competition may shape the role of corporate governance. That is, competition by itself might

serve as a bonding mechanism to mitigate agency problems, or intense competition may make

the monitoring mechanisms of corporate boards more desirable (e.g., Scharfstein 1988; Schmidt

1997). These competing perspectives motivate us to investigate the role of board independence

across industries.

In addition, considering the specific features of emerging markets, we investigate one

dimension of a firm’s ownership structure, foreign ownership—i.e., the share of foreign

investment in a firm’s common stock. Foreign ownership may play an important role in our

research setting because firms in emerging markets typically lack IT experience and management

expertise, a significant barrier to IT returns (Dewan and Kraemer 2000). Therefore, foreign

investors’ knowledge and experience with global technology may benefit companies in emerging

markets in deploying IT, which is the so-called “spillover” (of IT management expertise) (Aitken

and Harrison 1999). However, foreign investors’ general experience in the host countries may

be domain specific, i.e., not compatible with local management style or not suitable for the local

4 Global outsourcing has made Taiwan the largest producer of desktop PCs, notebooks, displays, and motherboards (Dedrick and Kraemer 2005). To illustrate, by 2004, Taiwan had accounted for over 70% of the world’s notebook PC market (Yang 2006).

5

environment, and may not help improve firm performance. Furthermore, Short (1994) argues

that foreign ownership may increase ownership dispersal and, therefore, exacerbate agency

problems. Motivated by such a tension between IT expertise spillover and exacerbated agency

problems, we attempt to explore the role of foreign ownership in IT returns.

Using a sample of 719 Taiwanese high-tech companies from 2001 to 2005, we find a

positive moderating role of board independence, i.e., the higher the board independence, the

more positive the IT-performance relation. This moderation effect becomes even stronger in

more competitive industries, suggesting that the value of corporate boards’ monitoring function

increases as competition intensifies. Also, we find the impact of foreign ownership on the IT-

performance relationship only in small firms, indicating that foreign investors can mitigate

resource disadvantages of small firms by bringing IT expertise to help them manage IT more

effectively.

Our study adds to the extant literature on IT returns in three different ways. First, the

results offer insights into why the mixed findings are reported in the prior literature on the IT-

performance relationship. Our results show an insignificant relationship between firm

performance and IT investment per se. That relationship, however, is moderated by corporate

governance. Clearly, a better monitoring system can help companies achieve better financial

performance through IT investment. Also, foreign ownership may have a spillover effect to help

small firms attain IT benefits. These results suggest that future research should incorporate

corporate governance in exploring how IT investment affects firm performance.

Second, this work has an important implication for research on board composition.

Based on a review of 54 empirical studies, Dalton et al. (1998) report no association between

board composition and financial performance. Therefore, they conclude “we are not optimistic

6

that future research in the general areas of board composition/financial performance…would be

fruitful. Also, the evidence would not seem to provide much confidence in further examinations

of possible moderating influences on those relationships” (p.284). Yet, this study finds that

while the direct relationship of board independence to firm performance is not clear, the

interaction between board independence and IT investment is significantly associated with better

firm performance. As such, we believe that investigating the role of board composition in

specific investment activities should be a promising avenue for future research.

Third, our results show contingency effects of firm size and industry competition in

examining corporate governance. In particular, we find that the moderation effect captured by

board independence is most robust in intensely competitive industries. One implication is that

controlling principle/agent conflicts through board monitoring may be more beneficial for firms

in more competitive industries. The relationship between competition and monitoring has

received attention in modeling papers (see Karuna 2006 for a literature review), but the literature

fall short of empirical evidence. Our work thus contributes to this burgeoning literature by

relating effects of corporate governance to industry competition.

The rest of this paper is organized as follows. Section 2 discusses related literature and

hypotheses development. Section 3 describes the method, including our sample data and models,

followed by empirical results in Section 4. Lastly, Section 5 provides a summary and conclusion.

2. Hypotheses Development

This study explores two factors concerning corporate governance, board independence

and foreign ownership, which may affect the IT-performance relationship in the Taiwanese

context.

7

2.1. Board independence

The principal/agent literature suggests that managers often act in their self-interest at the

expense of the firm’s value (e.g., Feltham and Xie 1994; Lambert 2001). That is, while

shareholders want managers to work to improve firm performance, managers may maximize

their own utility by over-consuming company resources or selecting suboptimal investments

(e.g., Fama and Jensen 1983).

In our research context, managers may make excessive IT investments, because a large

IT department brings power and high salary (Carr 2004). In addition, the literature of career

concerns suggests that managers may intentionally imitate others’ investment decisions to

enhance their professional reputations (e.g., Scharfstein and Stein 1990; Graham 1999). These

agency problems may occur, because evaluating managers’ IT investment decisions requires

technology-related information (e.g., costs, benefits, risks, implementation complexity, and

future development of the technology, and the company’s own technical capability), but

shareholders often lack such information. As such, shareholders may evaluate IT managers’

decisions by industry consensus (Kauffman and Li 2003). Consequently, managers may take the

advantage of asymmetric information to support their investment decisions (e.g., Fama and

Jensen 1983), and they tend to follow others to maintain their reputations rather than assessing

economic benefits for their companies (Graham 1999). Also, firms in emerging markets seek to

catch up to the global industry by moving along the IT diffusion curve and have been

increasingly allocating financial resources to IT (IDC 2006). This further allows managers to

use the industry consensus of investing in IT to justify their investment decisions. These agency

problems may cause the IT-performance relationship to be insignificant and even negative.

8

The corporate board of directors serves as an important internal mechanism in making

significant managerial decisions and in limiting managerial inefficiencies (e.g., Daily and Dalton

1994; Zahra and Pearce II 1989). The corporate board directly monitors IT investment when the

investment volume is large (Klein 2002). This is supported by prior research on corporate

investment in research and development (R&D) (e.g., Baysinger et al. 1991). In addition, the

board of directors meets regularly with internal and external auditors to review the firm’s

financial statements, audit process, and internal controls. This creates an indirect monitoring

function for IT investment. Thus, the board’s monitoring function helps reduce the agency

problems.

The corporate board consists of inside directors (those employed by the company) and

outside directors (e.g., CEOs of other firms, investment bankers, former governmental officials,

major suppliers). An outsider-majority board is associated with high independence, while an

insider-majority board is associated with low independence (e.g., Dalton et al. 1998; Klein 2002).

It has been argued that an independent board can monitor more effectively (e.g., Fama and

Jensen 1983; Kaplan and Reishus 1990; Klein 2002). One possible explanation is that outside

directors are more objective and have access to external information that is less available to

inside directors (e.g., Sutton and Callahan 1987; Daily and Dalton 1994). Furthermore, Beasley

et al. (2000) argue that outside directors have incentives to protect their reputation and avoid

litigation. Prior studies have shown that an increase in board independence improves earnings

quality (e.g., Dechow et al. 1996), lowers the magnitude of abnormal accruals (e.g., Klein 2002),

and also decreases the likelihood of financial fraud (e.g., Beasley et al. 2000). These studies

support the notion of better monitoring by outsider-majority boards.

9

To sum up, the agency problems may weaken the IT-performance relationship; however,

board independence may reduce the agency problems and thus enhance the IT value. Hence, we

predict a positive moderation effect of board independence as follows:

H1: Board independence positively moderates the relationship between IT investment and firm performance.

2.2. Industry competitiveness and board independence

The above moderation effect (H1) may differ across industries with different degrees of

competitiveness. The industrial organization literature commonly uses industry concentration as

a proxy for industry competitiveness (e.g., Cohen and Levin 1989; Waring 1996). Schumpeter

(1970) stresses that a high level of industry concentration (and therefore low competitiveness)

increases a firm’s profitability due to insulation of competition, and thus decreases the marginal

value of new technology investment. In contrast, industry competitiveness increases the

marginal gains to the firm from new technology (e.g., Levin et al. 1987; Cohen and Levin 1989;

Morrison 1997). Melville et al. (2004) also argue that firms in highly concentrated industries, by

attaining monopoly rents, may achieve superior performance without IT-enhanced efficiency

gains. Conversely, under competitive regimes, firms need to rely on other approaches to greater

profitability, e.g., using IT to increase efficiencies in business processes (e.g., Melville et al.

2004; Banker et al. 2006).

However, as discussed earlier, agency problems must be alleviated in order to convert

IT’s potential benefits to realized performance improvement. This is particularly important for

firms facing intense competition (e.g., Hermalin 1992; Schmidt 1997). That is, the marginal

value to the firm of reducing agency problems is higher when the firm is operating in more

competitive industries than in less competitive industries. Hence, the board’s monitoring would

10

benefit firms more as industry competitiveness becomes more intensive.

In another stream of research on managerial incentives, several theoretical papers imply

that increased competition may work as a bonding mechanism to decrease agency problems, by

providing more performance evaluation information (e.g., Nalebuff and Stiglitz 1983). This is

because intense competition makes firm performance more sensitive to uneconomic investments.

However, other studies suggest that the effect of competition may depend on managers’ utility

function (e.g., Karuna 2006). For instance, Scharfstein (1988) shows that competition actually

increases managerial actions for self-interest when the marginal utility gained from self-interest-

seeking is strictly positive. Taken together, the predicted relationships between competition and

managerial incentives are not clear.

More importantly, it is difficult for shareholders to relate changes in firm performance to

IT investment since disclosure of IT investment is not a requirement. This may weaken the

effect of competition, if any, to constrain principal/agent conflicts in IT investment. In fact,

Abrahamson and Rosenkopf (1997) find evidence suggesting that, under competition, companies

tend to make IT investments by following suit, without a thorough cost-benefit analysis. This

underscores the need for corporate board monitoring. Therefore, we expect the monitoring

function of independent boards, in the IT-performance context, to be more important in more

competitive industries.

H2: The moderation effect of board independence in the IT investment-firm performance relationship is more positive for more competitive industries than for less competitive industries.

2.3. Foreign ownership

Associated with the global sourcing of computers, the shipment of the Taiwanese high-

tech industries has increased significantly on the world platform, which has attracted foreign

11

investors to the Taiwanese market.5 Foreign ownership may influence firm performance in

different ways.

Foreign ownership may be negatively related to firm performance due to ownership

dispersal (e.g., Short 1994). The entry of foreign ownership increases the number of

shareholders, which dilutes share ownership and thereby lessens the incentive for each

shareholder to monitor the agents, consequently lowering firm performance (e.g., Boardman et al.

1997). Also, when foreign investors provide managerial suggestions to top management, their

involvement may not benefit the firm due to their limited knowledge of the local environment,

culture, and management style. Furthermore, Simunic (1980) reports a positive relationship

between foreign ownership and auditing fees. Consequently, firm performance may be lowered

by these additional agency costs, suggesting a negative relation between foreign ownership and

firm performance.

Conversely, Aitken and Harrison (1999) point out that foreign ownership may bring

about a positive moderation (“spillover”) effect in IT investment. Foreign ownership may offer

expertise for deploying IT resources, which helps bridge the gap of IT performance impacts

between developing and developed markets (e.g., Dewan and Kraemer 2000). In the IT

investment setting, foreign investors’ knowledge and experience is technical in nature, which is

not subject to the agency costs associated with their advice in general management. This view

falls into the scope of literature on technological knowledge spillover and foreign investment in

underdeveloped economies (e.g., Aitken and Harrison 1999). For instance, prior research

proposes that “in small economies, especially underdeveloped economies, most technological

knowledge will come from abroad” (Braga and Willmore 1991, p.431), and empirical evidence

5 Starting in 1991, the Taiwanese government approved the Qualified Foreign Institutional Investors (QFII) program, allowing qualified foreign investors to participate in the Taiwanese securities market. Over the past decade, foreign ownership has gained increasing importance in Taiwanese high-tech industries (Chiang and Kuo 2006).

12

supports that foreign investment is one approach to knowledge transfer (e.g., Blomström 1986;

Braga and Willmore 1991; Aitken and Harrison 1999). This discussion leads us to expect a

positive moderating effect of foreign ownership on the IT value.

H3: Foreign ownership positively moderates the relationship between IT investment and firm performance.

2.4. Firm size and foreign ownership

The outcome of new technology deployment may be contingent on certain organizational

characteristics such as firm size (e.g., Damanpour 1996). Compared to small firms, large firms

have more expertise in deploying IT resources, because they can afford to hire major consulting

firms to provide specialized services and offer expert advice (Damanpour 1996). Large firms are

also obliged to have more comprehensive internal controls and auditing systems in place (e.g.,

Klein 2002), which may mitigate agency problems. Therefore, the impacts of foreign ownership

(both the agency costs and the “spillover” benefits) are expected to be insignificant for large

firms.

In contrast, smaller firms are afflicted with “resource poverty” in deploying IT (e.g.,

Harrison et al. 1997; Thong 1999). These small firms generally lack IT knowledge and skills,

which is a significant barrier to IT returns. Furthermore, CEOs and directors of small firms are

usually less constrained by organizational systems and structures and therefore may have more

discretion compared to their large firm counterparts (e.g., Eisenhardt and Schoonhoven 1990;

Daily and Dalton 1994). The weak internal control system and lack of a comprehensive audit

system may curtail potential returns on IT investment. Since foreign investors can contribute

their IT expertise and enhance board structure, foreign ownership may strengthen a small firm’s

ability to generate value from deploying IT. Indeed, Aitken and Harrison (1999) find that the

13

spillover effect of foreign ownership is significant for small firms but not large firms. As such,

we expect that small firms benefit more from the “spillover” of IT expertise from foreign

ownership than large firms do. This discussion leads to our last hypothesis:

H4: The moderation effect of foreign ownership in the IT investment-firm performance relationship is more positive for small firms than for large firms.

3. Model Development

3.1. Data

Our sample consists of 719 manufacturing firms in the electronics industry in Taiwan

from 2001 to 2005. The distribution of industry sectors is shown in Table 1. We obtain data

concerning corporate IT investment from a database created by the Taiwanese Institute for

Information Industry, which is funded by the Taiwanese government. Since 2001, to set IT

policies, the Taiwanese government has asked the Taiwanese Institute for Information Industry

to collect IT investments data from the top 2,000 companies. The data is collected through face-

to-face interviews and then verified for accuracy. On average, the response rate is about 45

percent.

We collect financial performance information from the Financial Report Database

compiled by the Taiwan Economic Journal (TEJ), which contains data extracted from the firms’

annual financial reports. Also, we obtain information on firms’ ownership structures and board

composition from the TEJ’s Corporate Database. These two databases cover listed companies on

the Taiwan Stock Exchange, and therefore our sample contains only public companies.

_________________________

Insert Table 1 about here __________________________

14

3.1.1. Variables in hypotheses

Below, we define the variables involved in hypotheses 1-4:

Firm performance (ROA). Firm performance is measured by return on assets, ROA, the

ratio of net income to the year-end book value of total assets (e.g., Johnstone and Bedard 2004;

Fan and Wong 2005; Verrecchia and Weber 2006). This is a widely used proxy to examine how

firm performance relates to IT investment (e.g., Kohli and Devaraj 2003). Efficiency gains from

using IT may lead to higher ROA (e.g., Melville et al. 2004), but uneconomic IT investment,

expensed or capitalized, may lower ROA. Therefore, the relationship between ROA and IT is

shaped by how the uneconomic investment is controlled.

IT investment (ITINV). The amount of IT investment includes hardware, software, and

costs concerning maintenance, personnel, and training. This proxy covers all major components

of corporate IT spending (e.g., Hitt and Brynjolfsson 1996; Bresnahan et al. 2002). Following

previous research, we divide a firm’s IT investment by its total assets for size adjustment.

Board independence (BIND). As described earlier, board independence increases with

the proportion of outside directors. In this study we use three proxy measures. First, outside

directors include affiliated and non-affiliated directors. Affiliated directors include past

managers, relatives of current managers, affiliated attorneys, and consultants to the firm, while

non-affiliated directors are those without such affiliations (e.g., Klein 1998). A conservative

measure is the proportion of non-affiliated directors (NONAFF), which excludes the proportion

of affiliated directors (AFF) (e.g., Yermack 1996; Klein 1998). Other research, however, argues

that affiliated directors may often be substantially independent, and thus both affiliated and non-

affiliated directors should be considered (e.g., Bhagat and Black 2002). Therefore, following

prior studies (e.g., Dalton et al. 1998; Bhagat and Black 2002), we also use the proportion of all

15

outside directors (OUT), i.e., the sum of AFF and NONAFF, to measure board independence.

Third, considering that the presence of affiliated and non-affiliated directors may result in

different degrees of board independence (Klein 1998), we create a composite index by factor

analysis (e.g., Bens 2002), which assigns weights to NONAFF and AFF and gives us the third

proxy: )65.014.1( AFFNONAFFBIND ×+×= . It is clear that OUT treats NONAFF and AFF

as equally important in establishing board independence, while BIND assigns a relatively higher

weight to NONAFF. Below, we first present regression results based on BIND, and then show

the results based on NONAFF and OUT in sensitivity analysis. As will be discussed later, results

based on the three measures are highly consistent.

Foreign ownership (FOR). We measure FOR by the percentage of common stock owned

by foreign investors. This measure is used in the literature on technological knowledge spillover

with foreign ownership (e.g., Blomström 1986; Aitken and Harrison 1999).

Industry competitiveness (CR4). We compute the four-firm concentration ratio (CR4) for

each firm, which equals the percentage of total sales in the firm’s industry sector accounted for

by the four largest firms in the same sector. CR4 is conversely related to industry

competitiveness (e.g., Cohen and Levin 1989; Waring 1996).

Firm size (ASSETS). Following previous studies, we use total assets to measure firm size

(e.g., Johnstone and Bedard 2004; Engel et al. 2006).

3.1.2. Control variables

To evaluate cross-sectional associations between firm performance and board

composition, we need to control for lagged firm performance (e.g., Klein 1998). In addition, we

need to consider other factors that may explain firm performance. We draw upon the empirical

literature in economics that has modeled firm performance and explanatory factors (e.g., Waring

16

1996). We select a set of key factors based on further review of prior studies on IT investment

and firm performance (e.g., Hitt and Brynjolfsson 1996; Kohli and Devaraj 2003). These control

variables are as follows:

Lagged firm performance (ROA-1). A firm’s board composition could be a function of its

past performance. Weisbach (1988), based on observations in a 12-year period of time, found

that inside directors tended to leave the corporate board after a firm experienced a decrease in

stock market returns. However, Weisbach reported low magnitude of the impact of past

performance on board composition. To be conservative and to control for possible endogeneity

problems between board composition and firm performance, we follow prior research (e.g.,

Klein 1998; Peng 2004) by including the one-year lagged performance (ROA-1). This method is

appropriate because multiple regressions allow for correlations among the explanatory variables

(Klein 1998). We expect a positive coefficient of ROA-1 since performance measures are

positively correlated over time.

Firm growth (GROWTH). Prior studies on the IT-performance relationship have shown a

positive association between sales growth and ROA (e.g., Hitt and Brynjolfsson 1996).

Following the literature (e.g., Chaney and Philipich 2002), we measure GROWTH using the one-

year growth rate of sales and expect a positive coefficient of GROWTH.

Leverage (LEV). A firm’s performance may be influenced by its leverage ratio because

debt financing has both costs and benefits (e.g., Flannery and Rangan 2006). We measure LEV

as the book value of total debt divided by the book value of total equity (Flannery and Rangan

2006). Shareholders may have incentives to expropriate bondholders’ wealth by investing in

high-risk, high-return projects. As such, bondholders may demand higher rents by increasing the

costs of debt (Jensen and Meckling 1976). Conversely, debt financing may “urge” managers to

17

maintain their promises to pay out future cash flows. Thus, increased leverage may reduce cash

flows available for spending at the discretion of managers (Jensen 1986). Since previous

research on IT investment reports a negative coefficient for leverage (e.g., Hitt and Brynjolfsson

1996), we expect a negative relationship between firm performance and LEV.

Market share (SHARE). Following prior studies, we also include SHARE, i.e., the ratio

of a firm’s sales to the total industry sale. The literature contains mixed findings regarding both

the sign and the statistical significance of the market share-firm performance relationship (see a

review by Szymanski et al. 1993). This may be because of the inclusion of some proxy variables

for power, which may weaken the statistical association between market share and firm

performance (e.g., Hitt and Brynjolfsson 1996). In addition, firms with high market shares may

derive no extra negotiation power benefits, and as a result, the efforts devoted to increasing

market share (e.g., marketing costs and capacity build-up) may not pay off (e.g., Szymanski et al.

1993). Hence, we cannot predict the sign of SHARE.

Research and development intensity (RD). Research and development intensity is viewed

as a proxy for a firm’s intangible capital, which may enhance firm performance (e.g., Li and

Wong 2005). We measure RD as a firm’s R&D expenditures divided by sales (e.g., Callen et al.

2005; Darrough and Rangan 2005). While some studies show positive market valuation of R&D

expenditures (Hall 1993), recent research on the IT-performance relationship reports R&D to be

negatively related to performance (e.g., Bharadwaj et al. 1999). Thus, a priori the relationship

between R&D intensity and firm performance is not clear.

3.2. Regression Model

Our model builds on a previous cross-sectional analysis regressing firm performance on

board structure and lagged performance (Klein 1998). We extend the previous model in two

18

ways. First, we study two factors concerning corporate governance—board independence and

foreign ownership, both used as independent variables. Second, we add IT investment as a new

independent variable, which is the point of interest in this study. Our regression model is as

follows:

= β1 ITINV + β2 BIND + β3 (ITINV×BIND) + β4 FOR + β5 (ITINV×FOR)

+ β6 ROA-1 + β7 GROWTH + β8 LEV + β9 SHARE + β10 RD + β0 (1)

where )10,...2,1( =kxk is the kth explanatory variable (in vector terms, =],...,[ 1021 xxx [ITINV,

BIND, ITINV×BIND, FOR, ITINV×FOR, ROA-1, GROWTH, LEV, SHARE, ]RD ),

)10,...2,1( =kkβ is the corresponding regression coefficient, and β0 is an intercept.

The model specifies direct associations of firm performance (ROA) with IT investment

(ITINV) and the two corporate governance variables (BIND and FOR). Following the literature,

we expect the coefficients of IT investment (β1), board independence (β2), and foreign ownership

(β4) to be insignificant (e.g., Kohli and Devaraj 2003), insignificant (e.g., Dalton et al. 1998;

Bhagat and Black 2002), and significant and negative (e.g., Short 1994), respectively. To

evaluate the moderation of corporate governance, the model incorporates interaction terms

between IT investment and corporate governance (ITINV×BIND and ITINV×FOR). The model

also controls for lagged performance (ROA-1) and other factors (GROWTH, LEV, SHARE, and

RD) as discussed above.

According to this model, the relationship between firm performance and IT investment is

a function of board independence (BIND) and foreign ownership (FOR), expressed by the

following first-order derivative:

0

10

1ββ += ∑

=kkk xROA

FORBINDITINVROA

531 βββ ++=∂∂

19

where β1 represents the direct IT-performance relationship, and β3 and β5 represent moderating

effects of board independence and foreign ownership, respectively. According to the hypotheses

proposed earlier, β3 is expected to be significantly positive (H1) and β5 is to be significantly

positive (H3).

In H2 and H4, we are concerned with differential moderating effects of corporate

governance across industry sectors and size classes. Chow test (Chow 1960) can be used to test

for whether the relationships in a linear model remain stable across sub-samples. We employ the

dummy variable regression approach to Chow test (Gujarati 1988). Specifically, we create a

dummy C: C=1 for firms with CR4 above the sample median; C=0 for firms with CR4 below the

sample median. Then, we estimate the following regression model:

where )10,...2,1( =kxk denotes the explanatory variable as before (defined in Model (1)), and Ckβ

and Ck−1β are regression coefficients for less competitive industries (C=1) and more competitive

industries (C=0), respectively. The dummy variable regression also specifies the differential

intercept (by CC 0α ), which indicates how much the intercept term in the less competitive

industries differs from the base intercept ( 0α ). To see the implications of Model (2), we obtain:

where C3β and C−1

3β gauge the moderating effect of board independence in less competitive

industries and more competitive industries, respectively. As stated in H2, we expect C−13β to be

more positive than C3β .

)2())(1()( 00

10

1

110

1 CxCxCROA C

kk

Ck

kk

Ck ααββ ++−+= ∑∑

=

−

=

FORBINDITINVROA Custries titive indMore compe

FORBINDITINVROA Custries titive indLess compe

CCC

CCC

−−− ++=∂∂=

++=∂∂=

15

13

11

531

:)0(

:)1(

βββ

βββ

20

We create another dummy A: A=1 for firms with ASSETS above the sample median; A=0

for firms with ASSETS below the sample median. Then, we estimate the following model:

where )10,...2,1( =kxk is the explanatory variable as before, and Akβ and A

k−1β are regression

coefficients for large firms and small firms, respectively. AA 0γ is the differential intercept and

0γ is the base intercept. We obtain:

where A5β and A−1

5β represent the moderating effect of foreign ownership for large firms and

small firms, respectively. H4 proposes that A−15β is more positive than A

5β .

4. Results

4.1. Summary Statistics

Table 2 summarizes the descriptive statistics of the key variables on both the full sample

and sub-samples. Table 3 shows Pearson correlations. The ANOVA results in Table 2 show that

small firms in our sample have greater ROA, consistent with the nature of the electronics industry

(Kraemer and Dedrick 1998). The production of commodity electronic products (e.g., modems,

motherboards, flat-panel displays) is characterized by asset-intensive, high-volume, but low-

return manufacturing. In contrast, applications providers and firms focused on design and

services enjoy high profitability, while their operations require relatively fewer assets.

The ANOVA also shows that foreign investors target firms in less competitive industries,

possibly because investments in these firms are less risky. The mean board independence is also

higher in less competitive industries. It might be because firms in these industries are more

FORBINDITINVROA A firms Small

FORBINDITINVROA A firms Large

AAA

AAA

−−− ++=∂∂=

++=∂∂=

15

13

11

531

:)0(

:)1(

βββ

βββ

)3())(1()( 00

10

1

110

1 AxAxAROA A

kk

Ak

kk

Ak γγββ ++−+= ∑∑

=

−

=

21

inclined to rely on board independence to immunize themselves against managerial incentives

for self-interest (Nalebuff and Stiglitz 1983). Or, it may simply be that BIND is positively

correlated with FOR (see Table 3) and the mean FOR is higher in less competitive industries.

Because our focus is on how BIND moderates IT value and the corporate governance variables

are correlated with several control variables (Table 3), it is difficult to draw conclusions based

solely on the univariate analysis and we need to proceed to a multivariate regression.

Table 2 also shows that small firms have higher BIND, consistent with prior research

showing a negative relationship between firm size and audit committee independence (Klein

2002). Klein discusses that “[l]arger firms have stronger internal controls systems than smaller

firms…If the firms’ internal controls act as in-house monitoring mechanisms, then larger firms

require less alternative monitoring” (Klein 2002, p.440). A similar rationale can be used to

explain the observation in our sample. Also, our results show that firms operating in less

competitive industries or having a large size have higher GROWTH than those that are in more

competitive industries or are smaller in size.

In Table 3 we note two high correlations. First, as expected, our dependent variable

(ROA) is highly correlated (r=0.835, p<0.01) with a control variable ROA-1. Second, a control

variable SHARE is highly correlated (r=0.724, p<0.01) with CR4, which is used to divide the

sample into less and more competitive industries. The high correlation is intuitive; on average,

firms in more concentrated industries have higher market shares.

Finally, it is worth noting that the pair-wise correlations between the explanatory

variables in our regression model are not high, with the largest absolute value equaling 0.204

(between BIND and RD). Hence, multicollinearity is not likely to be a problem. We also

22

calculate variance inflation factors, which in our study are all below 1.5, suggesting no harmful

mutilcollinearity (Greene 2000).

__________________________________

Insert Tables 2 and 3 about here __________________________________

4.2. Full-Sample Results

Table 4 presents estimates for Model (1), showing the relationships of ROA with the ten

explanatory variables on the full sample. The overall model is significant (F= 221.07, p<0.01).

The coefficient of ITINV is insignificant (t=0.27), supporting that the direct relationship between

IT investment per se and firm performance is not clear (Kohli and Devaraj 2003).

The coefficient of BIND is insignificant (t=-0.62), consistent with the non-correlation

between board independence and firm performance reported by previous studies (e.g., Dalton et

al. 1998; Bhagat and Black 2002). Nonetheless, we find a positive and significant interaction

between ITINV and BIND (β3=0.084, t=3.84), indicating that the IT investment-ROA relation

would be more positive for firms with higher board independence. This finding supports H1.

The coefficient of ITINV×FOR is insignificant (t=0.78), thus showing no support for H3.

The coefficient of FOR is also insignificant (t=-0.89), which is different from our expectation. A

plausible explanation is that large firms, with sound internal controls and comprehensive

auditing systems in place, do not suffer from agency costs due to foreign investment. This

dilutes the influence of FOR on the full sample. To further check this explanation, we need to

look into the sample split by firm size, which is discussed in Section 4.4.

_________________________

Insert Table 4 about here _________________________

23

Among the controls, the coefficients of ROA-1 and GROWTH are significant and positive,

consistent with our expectations. Debt financing has both benefits and costs, and here we find a

significant and negative coefficient of LEV (-0.013, p<0.01). This is consistent with results in

previous studies using different samples (e.g., Agrawal and Knoeber 1996; Hitt and Brynjolfsson

1996). SHARE is insignificant, possibly because GROWTH taps increased negotiation power.

This interpretation is consistent with evidence in prior research (e.g., Hitt and Brynjolfsson 1996).

Another explanation is that firms in our sample may lack buying power (Szymanski et al. 1993).

Finally, the coefficient of RD is insignificant. This may be because ROA does not effectively

capture impacts of the intangible assets proxied by R&D expenditures.

4.3. More Competitive versus Less Competitive Industries

As shown in Table 5, estimates of Model (2) include two sets of regression coefficients,

i.e., for less competitive industries (C=1) and for more competitive industries (C=0). The

significant Chow test (11.21, p<0.01) rejects equality between the two sets of coefficients. A

notable difference is that the significant coefficient of ITINV×BIND in the full-sample (Table 4)

becomes insignificant in less competitive industries ( C3β =0.0093, t=0.27), while remains

significant and of higher magnitude ( C−13β =0.17, t=4.60) in more competitive industries. This

difference ( C−13β vs. C

3β ) provides support for H2.

In less competitive industries, there is no significant direct or indirect IT-performance

relationship through interacting with corporate governance. These results suggest that in less

competitive industries, firm profitability may result from sources other than new technologies,

such as monopoly rents. By contrast, firms operating in more competitive industries need to rely

more heavily on new technologies to improve performance. The significant interaction term

24

between ITINV and BIND (0.17, p<0.01) shows that, in conjunction with board independence, IT

investment creates significant value in intensely competitive industries.

Together with the ANOVA in Table 2, these regression results suggest an interesting

paradox: The mean board independence is lower in more competitive industries (Table 2), which

seems to be consistent with the rationale that competition, in general, could serve as a bonding

mechanism to monitor managerial incentives (e.g., Nalebuff and Stiglitz 1983). Nonetheless, as

shown in Table 5, the bonding mechanism of competition does not seem to be effective in the

context of IT investment. In the more competitive industries, the coefficient of ITINV is positive,

but numerically small and insignificant. IT investment relates to higher ROA only through

interacting with board independence. A plausible explanation, as stated in hypothesis

development, is that shareholders lack information about IT investment, which limits their ability

to assess managers’ investment decisions. Alternatively, shareholders need to rely on corporate

boards to monitor IT investment.

As to the controls, we observe qualitatively similar results except for the insignificant

LEV in less competitive industries. Since firms in less competitive industries generally possess

greater free cash flow, they benefit more from the control function of debt financing (Jensen

1986). The increased control function thus weakens the negative influence of LEV.

_________________________

Insert Table 5 about here _________________________

4.4. Large Firms versus Small Firms

Table 6 shows estimates for Model (3), including regression coefficients for large and

small firms. The significant Chow test (2.27, p<0.01) rejects equality between the two sets of

coefficients. Again, we find insignificant coefficients of ITINV and BIND. Consistent with the

25

full sample, the coefficient of ITINV×BIND is significant for both large and small firms. We

conclude that the moderating effect of board independence is robust across size classes.

The coefficient of FOR is negative, significant for small firms (t=-2.13) and insignificant

for large firms (t=-0.37). This suggests that compared with large firms, small firms suffer from

the agency problem attributed to dispersed ownership as suggested by Short (1994). The

interaction ITINV×FOR is significant and positive for small firms ( ,26.015 =−Aβ t=1.84) and

insignificant for large firms ( ,024.05 =Aβ t=0.45). This difference ( A−15β vs. A

5β ) supports H4

that small (instead of large) Taiwanese firms benefit from the IT expertise “spilled over” from

foreign investors.

_________________________

Insert Table 6 about here _________________________

4.5. Sensitivity Analysis

We conduct the following tests to examine the robustness of the results to measures and

model specifications. These additional tests demonstrate the robustness of our results.

Alternative measures for board independence. Earlier, we discussed alternative

measures for board independence (OUT and NONAFF). Panel A of Table 7 presents the results

of regressions using NONAFF as a proxy for board independence. As seen in Panel A, the

coefficients of all independent variables are virtually the same. The independent variable that

appears to be most affected by using the alternative proxy is ITINV×BIND, which turns out to be

significant at the 5% level for small firms (compared with 10% in Table 6). Panel B of Table 7

presents the results of regressions using OUT as a proxy for board independence. The results are

highly consistent with the results in Tables 4-6.

26

Alternative measure for industry competitiveness. We use an alternative measure for

industry competitiveness, entry costs (ENTCOST), which refer to sunk costs that must be

incurred for operating in one industry (Porter 1985). The higher the entry costs, the higher the

industry concentration and the lower the industry competitiveness. Following the literature (e.g.,

Karuna 2006), ENTCOST is defined as the weighted average gross value of property, plant, and

equipment for firms in one industry, weighted by each firm’s market share in the industry. We

rank all firms in our sample by ENTCOST and divide them at the median. Firms with ENTCOST

above median form the less competitive industries, and firms with ENTCOST below median

form the more competitive industries. As seen in Panel C of Table 7, we observe qualitatively

similar results using this alternative measure for industry competitiveness (Table 5).

Industry effects. Next, we control for industry effects. We re-estimate regression (1) by

including industry dummies as additional controls. Only one industry dummy (the systematic

product sector) has a significant coefficient (at 10% significance level), and jointly the industry

dummies are insignificant. We then add the industry dummies into regressions (2) and (3) and

find no significant industry dummies. As shown in Panel D of Table 7, including industry

dummies produces highly consistent results for all regressions.

Two-year lagged performance. We examine the sensitivity of our regressions (1)-(3) by

adding an additional control variable, the two-year lagged performance (ROA-2). As shown in

Panel E of Table 7, the coefficient of ROA-2 is significant only for large firms (in the “A=1”

column). The coefficients of all the other independent variables are qualitatively similar to those

reported in Tables 4-6. Also, for small firms (in the “A=0” column), the significance level of the

coefficient for ITINV×BIND increases (i.e., from 10% in Table 6 to 5%).

27

Partial models. We also estimate partial models. We check for the robustness of

regressions (1)-(3) by dropping the moderation effects (BIND, ITINV×BIND, FOR,

ITINV×FOR). The coefficient of ITINV remains insignificant in each regression (not reported),

consistent with the finding that the direct IT -performance relationship is not clear.

We then re-estimate regressions (1)-(3) with all control variables dropped. As reported in

Panel F of Table 7, the significance levels of two coefficients are affected by the removal of

controls, while there were no significant coefficients with a change in sign. The coefficient of

FOR becomes significant at the 10% level for small firms (compared with 5% in Table 6), and

the coefficient of ITINV×BIND becomes significant at the 1% level for large firms (compared

with 5% in Table 6). Taken together, we observe qualitatively similar results.

Dummy variable regression. Finally, we consider the robustness of the dummy variable

regression. We use the sample mean of CR4 as the cutoff to classify more and less competitive

industries. We obtain estimates qualitatively similar to those reported in Table 5 (where the

sample median of CR4 is used as the cutoff). We do not use the sample mean of ASSETS to

classify large and small firms because this variable is skewed.

________________________

Insert Table 7 about here ________________________

5. Concluding Remarks

Information technology has the potential to improve firm performance. Yet, researchers

are puzzled by the mixed empirical findings concerning returns on IT investment. Using a

sample of manufacturers in high-tech industries in Taiwan, we investigate how corporate

governance shapes the IT investment-firm performance relationship. We find moderating effects

28

of board independence and foreign ownership on the IT value on firm performance. The

moderation effects differ in significance across size classes and industries.

Specifically, we find no moderating effect of corporate governance on the IT investment-

firm performance relationship in less competitive industries. In contrast, in more competitive

industries, there is a significant interaction effect between IT investment and board independence,

which is positively related to firm performance. Information technologies help firms improve

performance under competition, while the economic returns on IT investment are positively

related to board independence. This supports the conjecture that the corporate board plays a

monitoring role in new technology investment. Our findings show that board monitoring

benefits intensely competitive industries more. Clearly, the relationships among IT returns,

board independence, and industry competitiveness are rather complex. While we have not

completely disentangled the effects, our results offer insights into the role of corporate

governance in achieving returns on IT investment.

Moreover, we find a negative direct association between foreign ownership and firm

performance for small firms, possibly due to additional agency costs incurred and foreign

investors’ limited local knowledge in offering advice for general management. However, we

observe a significant and positive interaction between IT investment and foreign ownership in

the regression for small firms. Foreign investors may inject IT expertise that is likely to be

applicable to both developed and emerging markets. Hence, foreign ownership helps small firms

more effectively deploy IT. As an implication for research on corporate governance, our results

suggest that an analysis of governance structures needs to be situated within specific contexts

such as firm size and industry competitiveness. Before discussing the managerial implications of

our results, we briefly point out some limitations of our study.

29

First, this study is based on a sample from an emerging market, and thus questions must

be raised about the generalizability of the findings to developed markets, e.g., the U.S. Second,

this study is focused on two specific corporate governance factors, and we view our work as an

initial attempt to analyze the moderation of governance structures on IT value. One line of

future inquiry is to investigate other dimensions of corporate governance (e.g., board size, major

stockholders’ ownership, deviation between major shareholders’ control rights and cash claim

rights). Third, although our analysis of spillover follows a common method in economics (e.g.,

Blomström 1986; Braga and Willmore 1991; Aitken and Harrison 1999), future studies can

cross-validate our results and delve deeper into these issues by using other methodologies. For

example, field research may detail how foreign investors are involved in making IT investment

decisions. Also, more studies are needed to better understand what type of knowledge and

experience needs to be adapted to the local environment and culture in emerging markets and

what type can be more directly transferred from foreign ownership. Notwithstanding these

limitations, our study makes significant contributions to both research and practices.

As an important implication for firms considering IT investment, our results suggest that

competitive firms should be more proactive in using IT to increase their business value. Yet IT

alone does not hold the answer to performance improvement and must be combined with better

and more objective monitoring to reap the benefits. Our results also imply that foreign

ownership can help smaller firms transcend barriers to IT return. Managers need to realize that,

with foreign investors involved, their firms are more likely to attain performance improvement

through IT. However, the evidence should not be interpreted as recommending no IT investment

at firms with little or no foreign ownership, because there are approaches other than spillover that

are considered in acquiring IT-related knowledge. Overall, an important message to managers is

30

that management control systems play a crucial role in IT impacts across different structures of

markets and firms.

In conclusion, the contribution of this study lies in its evaluation of the moderation

effects of corporate governance in IT value creation. Our study provides one theoretical lens

through which different outcomes of IT are analyzed. The evidence also shows the usefulness of

studying the role of corporate governance in a specific investment activity, rather than linking it

directly to overall firm performance. These directions may lead to fruitful areas of future inquiry.

31

Table 1. Industry Sectors

Industry sector Frequency % of sample Electronic components 135 18.8 Photoelectric products 131 18.2 Motherboard scheme 110 15.3 Integrated circuits 95 13.2 Electronic channel 81 11.3 Software applications 37 5.1 Network modem 34 4.7 General electronics 30 4.2 Consumer electronics 26 3.6 Communication technologies 21 2.9 Systematic product and others 19 2.7 Total 719 100.0

32

Table 2. Descriptive Statistics Less competitive vs. more competitive industries Large firms vs. small firms

Full sample CR4 above

median CR4 below

median ANOVA ASSETS above

median ASSETS below

median ANOVA Mean S.D. Mean S.D. Mean S.D. t-statistic Mean S.D. Mean S.D. t-statistic

ROA 0.151 0.114 0.158 0.118 0.152 0.120 0.66 0.132 0.104 0.170 0.121 -4.61 ***ITINV 0.037 0.617 0.004 0.005 0.094 1.033 -1.39 0.003 0.05 0.070 0.871 -1.48 BIND 16.348 18.749 28.117 17.453 1.511 4.440 27.48 *** 14.474 17.715 18.217 19.572 -2.69 ***FOR 5.490 9.726 7.010 10.744 4.175 8.857 3.56 *** 7.920 10.922 3.066 7.645 6.90 ***ROA-1 0.169 0.122 0.166 0.124 0.184 0.128 -1.74 * 0.145 0.108 0.193 0.131 -5.34 ***GROWTH 0.111 0.363 0.148 0.299 0.063 0.408 2.81 ** 0.150 0.380 0.073 0.342 2.84 ***LEV 0.673 0.618 0.607 0.617 0.737 0.631 -2.53 ** 0.681 0.561 0.665 0.670 0.33 SHARE 0.322 1.729 0.382 1.888 0.201 0.427 1.75 * 0.346 1.607 0.299 1.844 0.36 RD 0.028 0.041 0.024 0.042 0.040 0.043 -4.44 *** 0.025 0.033 0.032 0.047 -2.09 **ASSETS 9.583 23.711 9.493 18.842 10.773 32.434 -0.61 17.380 31.705 1.807 0.812 9.30 ***CR4 0.208 0.069 0.223 0.087 0.188 0.001 7.64 *** 0.202 0.406 0.214 0.088 -2.47 **

***p<0.01; **p<0.05; *p<0.10.

Data from the Taiwanese Institute for Information Industry: ITINV = annual IT spending (hardware, software and costs concerning maintenance, personnel and training) divided by total assets

Data from TEJ’s Corporate Database: BIND = 1.14×NONAFF+0.65×AFF

NONAFF = proportion of non-affiliated directors (%) AFF = proportion of affiliated directors (%) FOR = proportion of common stock owned by foreign investors (%)

Data from TEJ’s Financial Report Database: ROA = net income ÷ yearend book value of total assets

ROA-1 = ROA in the previous year GROWTH = one-year growth rate of sales

LEV = book value of total debt ÷ book value of total assets SHARE = sales ÷ total industry sales (industry sectors shown in Table 1)

RD = research and development expenses ÷ sales ASSETS = total assets (US$ billion, 2001 constant dollar)

CR4 = sales of the four largest firms in the industry ÷ total industry sales

33

Table 3. Pearson Correlations

ROA ITINV BIND FOR ROA-1 GROWTH LEV SHARE RD ASSETS CR4ROA 1 ITINV -0.002 1 BIND 0.114*** -0.047 1 FOR -0.010 -0.029 0.084** 1 ROA-1 0.835*** -0.009 0.077** -0.008 1 GROWTH 0.185*** -0.037 0.074** -0.007 -0.043 1 LEV -0.168*** 0.003 -0.112*** -0.120*** -0.159*** 0.147*** 1 SHARE 0.017 -0.008 -0.059 0.011 0.028 -0.002 0.013 1 RD 0.178*** -0.026 -0.204*** 0.043 0.199*** 0.017 -0.082** 0.087** 1 ASSETS -0.149*** -0.018 -0.109 0.308*** -0.119*** 0.014 -0.016 0.138*** 0.093** 1 CR4 0.008 -0.016 0.110*** -0.012 -0.006 -0.034 -0.109*** 0.724*** 0.046 -0.043 1 ***p<0.01; **p<0.05; *p<0.10. See Table 2 for variable definitions.

34

Table 4. Regression Results: Full Sample

= β1 ITINV + β2 BIND + β3 (ITINV×BIND) + β4 FOR + β5 (ITINV×FOR)

+ β6 ROA-1 + β7 GROWTH + β8 LEV + β9 SHARE + β10 RD + β0

Coefficient βk Predicted Sign (t-statistic)

ITINV ? 0.0011 (0.27)

BIND ? -0.000085 (-0.62) ITINV × BIND + 0.084*** (3.84) FOR – -0.00020 (-0.89) ITINV × FOR + 0.010 (0.78) Controls ROA-1 + 0.77*** (42.49) GROWTH + 0.073*** (12.23) LEV – -0.013*** (-3.64) SHARE ? -0.00066 (-0.53) RD ? 0.018 (0.34) Intercept 0.019*** (3.43) F-statistic 221.07*** Adjusted R2 0.754

***, ** and * denote significance at the 1%, 5% and 10% levels, respectively. See Table 2 for variable definitions.

0

10

1ββ += ∑

=kkk xROA

35

Table 5. Regression Results:

More Competitive Industries vs. Less Competitive Industries

Less competitive industries (C=1)

More competitive industries (C=0)

Coefficient Ckβ Coefficient C

k−1β

(t-statistic) (t-statistic) ITINV -0.066 0.0032 (-0.67) (0.72) BIND 0.00014 -0.00021 (0.58) (-1.07) ITINV × BIND 0.0093 0.17*** (0.27) (4.60) FOR -0.00079 -0.000067 (-1.50) (-0.17) ITINV × FOR 0.107 -0.027 (0.74) (-0.52) Controls ROA-1 0.72*** 0.79*** (18.06) (36.01) GROWTH 0.069*** 0.095*** (6.67) (10.53) LEV -0.012 -0.014*** (-1.47) (-3.10) SHARE 0.00080 -0.00049 (0.42) (-1.48) RD -0.10 0.013 (-0.64) (0.20) Base intercept 0.0084 0.0084 (1.10) (1.10) Differential intercept 0.026** -- (2.03) Chow test 11.21*** F-statistic 94.88*** Adjusted R2 0.764

***, ** and * denote significance at the 1%, 5% and 10% levels, respectively. C=1 for firms with above-median CR4; C=0 for firms with below-median CR4. See Table 2 for variable definitions.

C

kk

Ck

kk

Ck CxCxCROA 00

10

1

110

1))(1()( ααββ ++−+= ∑∑

=

−

=

36

Table 6. Regression Results:

Large Firms vs. Small Firms

Large Firms

(A=1) Small Firms

(A=0) Coefficient A

kβ Coefficient Ak−1β

(t-statistic) (t-statistic) ITINV -0.0097 -0.0091 (-0.27) (-1.34) BIND -0.000013 -0.00016 (-0.08) (-0.65) ITINV × BIND 0.082** 0.059* (2.33) (1.84) FOR -0.00012 -0.0023** (-0.37) (-2.13) ITINV × FOR 0.024 0.26* (0.45) (1.84) Controls ROA-1 0.81*** 0.70*** (34.57) (24.36) GROWTH 0.061*** 0.094*** (8.42) (9.12) LEV -0.014*** -0.011** (-2.96) (-2.03) SHARE -0.0015 0.00058 (-0.80) (0.35) RD 0.011 -0.0065 (0.15) (-0.08) Base intercept 0.037*** 0.037*** (3.82) (3.82) Differential intercept -0.025** -- (-2.13) Chow test 2.27*** F-statistic 108.56*** Adjusted R2 0.759

***, ** and * denote significance at the 1%, 5% and 10% levels, respectively. A=1 for firms with above-median ASSETS; A=0 for firms with below-median ASSETS.See Table 2 for variable definitions.

A

kk

Ak

kk

Ak AxAxAROA 00

10

1

110

1))(1()( γγββ ++−+= ∑∑

=

−

=

37

Table 7. Sensitivity Analysis

Panel A: Use NONAFF to measure board independence Full Sample C=1† C=0† A=1‡ A=0‡

Coefficient kβ Coefficient Ckβ Coefficient C

k−1β Coefficient A

kβ Coefficient Ak−1β

(t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) ITINV 0.00079 -0.060 0.0012 -0.027 -0.0088

(0.19) (-0.61) (0.27) (-0.77) (-1.30) NONAFF -0.00019 0.00012 -0.00037 0.0000094 -0.00040

(-0.72) (0.26) (-0.99) (0.03) (-0.92) ITINV × NONAFF 0.16*** 0.063 0.24*** 0.13** 0.15**

(3.47) (0.91) (3.51) (1.94) (2.20) FOR -0.00020 -0.00072 -0.00018 -0.00019 -0.0023**

(-0.87) (-1.37) (-0.45) (-0.63) (-2.13) ITINV × FOR 0.012 0.098 0.019 0.049 0.25*

(0.90) (0.68) (0.38) (0.98) (1.81) Controls in Tables 4-6 Included Included Included Included Included

Panel B: Use OUT to measure board independence Full Sample C=1† C=0† A=1‡ A=0‡

Coefficient kβ Coefficient Ckβ Coefficient C

k−1β Coefficient A

kβ Coefficient Ak−1β

(t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) ITINV 0.0010 -0.070 0.0031 -0.010 -0.0097

(0.25) (-0.71) (0.69) (-0.28) (-1.44) OUT -0.00016 0.000091 -0.00022 -0.000037 -0.00027

(-0.86) (0.29) (-0.79) (-0.16) (-0.84) ITINV × OUT 0.10*** 0.0041 0.19*** 0.093** 0.070*

(3.65) (0.08) (4.29) (2.12) (1.66) FOR -0.00019 -0.00079 0.000093 -0.00011 -0.0024**

(-0.85) (-1.15) (0.23) (-0.35) (-2.18) ITINV × FOR 0.010 0.11 -0.023 0.024 0.27*

(0.78) (0.78) (-0.44) (0.46) (1.95) Controls in Tables 4-6 Included Included Included Included Included † “C=1” and “C=0” denote the two subsamples as defined in Model (2), i.e., less competitive and more competitive industries, respectively. ‡ “A=1” and “A=0” denote the two subsamples as defined in Model (3), i.e., large firms and small firms, respectively. ***, ** and * denote significance at the 1%, 5% and 10% levels, respectively. NONAFF and OUT are the shares of non-affiliated and outside directors in the board, respectively; see Table 2 for variable definitions.

38

Table 7. Sensitivity Analysis (continued)

Panel C: Use ENTCOST to measure industry competitiveness Less Competitive Industries

(ENTCOST above median) More Competitive Industries

(ENTCOST below median)

Coefficient Coefficient (t-statistic) (t-statistic) ITINV -0.040 0.0018

(-0.08) (0.44) BIND -0.00013 -0.0000090

(-0.55) (-0.05) ITINV × BIND 0.031 0.13 ***

(0.81) (4.27) FOR -0.00038 -0.00030

(-1.10) (-0.76) ITINV × FOR 0.041 0.0081

(0.81) (0.62) Controls in Tables 4-6 Included Included

Panel D: Control for industry effects Full Sample C=1† C=0† A=1‡ A=0‡

Coefficient kβ Coefficient Ckβ Coefficient C

k−1β Coefficient A

kβ Coefficient Ak−1β

(t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) ITINV 0.0012 -0.072 0.0035 -0.013 -0.0090

(0.28) (-0.71) (0.78) (-0.37) (-1.29) BIND -0.00011 0.00016 -0.00024 -0.000015 -0.00019

(-0.82) (0.66) (-1.21) (-0.09) (-0.80) ITINV × BIND 0.091*** 0.0092 0.17*** 0.078** 0.064*

(3.91) (0.24) (4.45) (2.20) (1.80) FOR -0.00017 -0.00079 -0.00016 -0.000066 -0.0024**

(-0.72) (-1.47) (-0.39) (-0.21) (-2.15) ITINV × FOR 0.0092 0.115 -0.032 0.029 0.25*

(0.72) (0.77) (-0.62) (0.56) (1.76) Controls in Tables 4-6 Included Included Included Included Included Industry dummies Included Included Included Included Included † “C=1” and “C=0” denote the two subsamples as defined in Model (2), i.e., less competitive and more competitive industries, respectively. ‡ “A=1” and “A=0” denote the two subsamples as defined in Model (3), i.e., large firms and small firms, respectively. ***, ** and * denote significance at the 1%, 5% and 10% levels, respectively. ENTCOST = weighed average gross value of property, plant and equipment for firms in one industry; see Table 2 for variable definitions.

39

Table 7. Sensitivity Analysis (continued)

Panel E: Include lagged 2-year performance measure Full Sample C=1† C=0† A=1‡ A=0‡

Coefficient kβ Coefficient Ckβ Coefficient C

k−1β Coefficient A

kβ Coefficient Ak−1β

(t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) ITINV 0.00095 -0.066 0.0033 -0.012 -0.0086

(0.23) (-0.67) (0.73) (-0.32) (-1.27) BIND -0.00011 0.00014 -0.00021 -0.000083 -0.000059

(-0.75) (0.56) (-0.99) (-0.46) (-0.23) ITINV × BIND 0.083** 0.0094 0.17*** 0.073** 0.063**

(3.72) (0.27) (4.56) (2.06) (1.96) FOR -0.00021 -0.00079 -0.000071 -0.00013 -0.0023**

(-0.91) (-1.49) (-0.17) (-0.42) (-2.10) ITINV × FOR 0.010 0.107 -0.027 0.026 0.26*

(0.78) (0.74) (-0.52) (0.49) (1.84) Controls in Tables 4-6 Included Included Included Included Included ROA-2 0.014 -0.0013 -0.0047 0.066* -0.0039 (0.59) (-0.03) (-0.15) (1.93) (-1.09)

Panel F: Estimate partial models Full Sample C=1† C=0† A=1‡ A=0‡

Coefficient kβ Coefficient Ckβ Coefficient C

k−1β Coefficient A

kβ Coefficient Ak−1β

(t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) ITINV 0.0011 -0.124 0.0039 0.048 -0.020

(0.13) (-0.63) (0.43) (0.67) (-1.54) BIND 0.00011 0.00058 0.00010 0.00022 -0.00036

(0.42) (1.23) (0.27) (0.67) (-0.79) ITINV × BIND 0.19*** 0.11 0.22*** 0.26*** 0.10*

(4.51) (1.58) (3.09) (3.73) (1.66) FOR -0.00025 -0.0016 -0.00029 0.00038 -0.0039*

(-0.56) (-1.52) (0.37) (0.63) (-1.82) ITINV × FOR -0.0038 0.18 -0.083 -0.070 0.47*

(-0.15) (0.53) (-0.79) (-0.69) (1.74) Controls in Tables 4-6 Dropped Dropped Dropped Dropped Dropped † “C=1” and “C=0” denote the two subsamples as defined in Model (2), i.e., less competitive and more competitive industries, respectively. ‡ “A=1” and “A=0” denote the two subsamples as defined in Model (3), i.e., large firms and small firms, respectively. ***, ** and * denote significance at the 1%, 5% and 10% levels, respectively.

40

References:

Abrahamson, E., Rosenkopf, L., 1997. Social network effects on the extent of innovation diffusion: A computer simulation. Organization Science 8 (3), 289-309.

Agrawal, A., Knoeber, C.R., 1996. Firm performance and mechanisms to control agency problems between managers and shareholders. Journal of Financial & Quantitative Analysis 31(3), 377-390.

Aitken, B.J., Harrison, G., 1999. Do domestic firms benefit from direct foreign investment? Evidence from Venezuela. American Economic Review 89 (3), 605-618.

Banker, R.D., Bardhan, I.R., Chang, H., Liu, S., 2006. Plant information systems, manufacturing capabilities, and plant performance. MIS Quarterly 30 (2), 315-337.

Baysinger, B.D., Kosnik, R.E. Turk., T.A, 1991. Effects of board and ownership structure on corporate R&D strategy. Academy of Management Journal 34(1), 205-214.

Beasley, M.S., Carcello, J.V., Hermanson, D.R., Lapides, P.D., 2000.Fraudulent financial reporting: Consideration of industry traits and corporate governance mechanisms. Accounting Horizons 14(4), 441-454.

Bens, D.A., 2002. The determinants of the amount of information disclosed about corporate restructurings. Journal of Accounting Research 40 (1), 1-20.

Bhagat, S., Black, B., 2002. The non-correlation between board independence and long-term firm performance. Journal of Corporation Law 27 (2), 231-273.

Bharadwaj, A.S., Bharadwaj, S.G., Konsynski, B.R., 1999. Information technology effects on firm performance as measured by Tobin's q. Management Science 45 (7), 1008-1024.

Blomström, M., 1986. Foreign investment and productive efficiencies: The case of Mexico. Journal of Industrial Economics 35 (1), 97-110.

Boardman, A.E., Shapiro, D.M., Vining, A.R., 1997. The role of agency costs in explaining the superior performance of foreign MNE subsidiaries. International Business Review 6 (3), 295-317.

Braga, H., Willmore, L., 1991. Technological imports and technological effort: An analysis of their determinants in Brazilian firms. Journal of Industrial Economics 39 (4), 421-432.

Bresnahan, T.F., Brynjolfsson, E., Hitt, L.M., 2002. Information technology, workplace organization, and the demand for skilled labor: Firm-level evidence. Quarterly Journal of Economics 117 (1), 339-376.

Brynjolfsson, E., Hitt, L.M., Yang, S., 2002. Intangible assets: Computers and organizational capital. Brookings Papers on Economic Activity: Macroeconomics (1), 137-199.

Bushman, R.J., Smith, A.J., 2001. Financial accounting information and corporate governance. Journal of Accounting and Economics 32 (1-3), 237-333.

Callen, J.L., Hope, O.-K., Segal, D., 2005. Domestic and foreign earnings, stock return variability, and the impact of investor sophistication. Journal of Accounting Research 43 (3), 377-412.

Carr, N., 2004. Does IT Matter: Information Technology and the Corrosion of Competitive Advantage. Cambridge: Harvard Business School.

Chaney, P.K., and Philipich, K.L., 2002. Shredded reputation: The cost of audit failure. Journal of Accounting Research 40 (4), 1221-1245.

Chiang, Y., Kuo, C., 2006. Characteristics of Firms Owned by Foreigners on the Taiwanese Stock Market: An Analysis. International Journal of Management 23 (3), 721-728.

Chow, G.C., 1960. Tests of equality between sets of coefficients in two linear regressions. Econometrica 28 (3), 591-605.

41

Cohen, W.M., Levin, R.C., 1989. Empirical studies of innovation and market structure. In: Schmalensee, R., Willig, R. (Eds.). Handbook of Industrial Organization. Amsterdam: The Netherlands, 1059-1107.

Committee of Sponsoring Organizations of the Treadway Commission 1992. Daily, C. M., Dalton, D.R., 1994. Bankruptcy and corporate governance: The impact of board

composition and structure. Academy of Management Journal 37 (6), 1603-1617. Dalton, D.R., Daily, C.M., Ellstrand, A.E., Johnson, J.L., 1998. Meta-analytic reviews of board

composition, leadership structure, and financial performance. Strategic Management Journal 19 (3), 269-290.

Damanpour, F., 1996. Organizational complexity and innovation: Developing and testing multiple contingency models. Management Science 42(5), 693-716.

Darrough, M., Rangan, S., 2005. Do insiders manipulate earnings when they sell their shares in an initial public offering? Journal of Accounting Research 43 (1), 1-33.

Dechow, P.M., Sloan, R.G., Sweeney, A.P., 1996. Causes and consequences of earnings manipulation: An analysis of firms subject to enforcement actions by the SEC. Contemporary Accounting Research 13 (1), 1-36.

Dedrick, J., Kraemer, K.L., 2005. The impacts of IT on firm and industry structure: The personal computer industry. California Management Review 47 (3), 122-142.

Devaraj, S., Kohli, R., 2003. Performance impacts of information technology: Is actual usage the missing link? Management Science 49 (3), 273-289.

Dewan, S., Kraemer, K.L., 2000. Information technology and productivity: Evidence from country-level data. Management Science 46(4), 548-562.

Eisenhardt, K.M., Schoonhoven, C.B., 1990. Organizational growth: Linking founding team, strategy, environment, and growth among U.S. semiconductor ventures, 1978-1988. Administrative Science Quarterly 35 (3), 504-529.

Engel, E., Hayes, R.M., Wang, X., 2006. The Sarbanes-Oxley Act and firms' going-private decisions. Journal of Accounting and Economics In Press, Corrected Proof.

Fama, E.F., Jensen, M.C., 1983. Agency problems and residual claims. Journal of Labor Economics 26 (2), 327-349.

Fan, J.P. H., Wong, T.J., 2005. Do external auditors perform a corporate governance role in emerging markets? Evidence from East Asia. Journal of Accounting Research 43 (1), 35-72.

Feltham, G., Xie, J., 1994. Performance measure congruity and diversity in multi-task principal/agent relations. The Accounting Review 69 (3), 429-453.

Flannery, M.J., Rangan, K.P., 2006. Partial adjustment toward target capital structures. Journal of Financial Economics 79 (3), 469-506.

Graham, J., 1999. Herding among investment newsletters: Theory and evidence. Journal of Finance 54(1), 237-269.

Greene, W.H., 2000. Econometric Analysis. Upper Saddle River, NJ: Prentice Hall. Gugler, K., Mueller, D.C., Yurtoglu, B.B., 2003. The impact of corporate governance on

investment returns in developed and developing countries. The Economic Journal 113, F511-F539.

Gujarati, D.N., 1988. Basic Econometrics. 2 ed. New York NY: McGraw-Hill. Hall, B.H., 1993. The stock market's valuation of R&D investment during the 1980's. American

Economic Review 83 (2), 259-264.

42

Harrison, D.A., Mykytyn, P.P., Jr., Riemenschneider, C.K., 1997. Executive decisions about adoption of information technology in small businesses: Theory and empirical tests. Information Systems Research 8(2), 171-196.

Hermalin, B., 1992. The effects of competition on executive behavior. Rand Journal of Economics 23(3), 350-365.

Hitt, L.M., Brynjolfsson, E., 1996. Productivity, business profitability, and consumer surplus: Three different measures of information technology value. MIS Quarterly 20 (2), 121-142.

IDC, 2006. Worldwide IT Spending 2006-2010 Forecast by Vertical Market. International Data Corporation, Framingham, MA.

Jensen, M.C., 1986. Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review 76 (2), 323-329.

Jensen, M.C., Meckling, W.H., 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics 3 (4), 305-360.

Johnstone, K.M., Bedard, J.C., 2004. Audit firm portfolio management decisions. Journal of Accounting Research 42 (4), 659-690.

Kaplan, S.N., Reishus, D., 1990. Outside directorships and corporate performance. Journal of Financial Economics 27 (2), 389-410.

Karuna, C., 2006. Industry product market competition and managerial incentives. Forthcoming, Journal of Accounting and Economics.

Kauffman, R.J., Li, X., 2003. Payoff externalities, informational cascades and managerial incentives: A theoretical framework for IT adoption herding. Proceedings of the 2003 INFORMS Conference on IS and Technology, Atlanta, GA.

Klein, A., 1998. Firm performance and board committee structure. Journal of Law and Economics 41 (1), 275-303.

Klein, A., 2002. Audit committee, board of director characteristics, and earnings management. Journal of Accounting and Economics 33 (3), 375-400.

Kohli, R., Devaraj, S., 2003. Measuring information technology payoff: A meta-analysis of structural variables in firm-level empirical research. Information Systems Research 14 (2), 127-145.

Kraemer, K.L., Dedrick, J., 1998. Globalization and increasing returns: Implications for the U.S. computer industry. Information Systems Research 9 (4), 303-322.

Lambert, R.A., 2001. Contracting theory and accounting. Journal of Accounting and Economics 32 (1-3), 3-87.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny, R.W., 2000. Investor protection and corporate governance. Journal of Financial Economics 58 (1), 3-27.

Levin, S.G., Levin, S.L., Meisel, J.B., 1987. A dynamic analysis of the adoption of a new technology: The case of optical scanners. Review of Economics and Statistics 69 (1), 12-17.

Li, F., F.M.H. Wong, 2005. Employee stock options, equity valuation, and the valuation of option grants using a warrant-pricing model. Journal of Accounting Research 43 (1), 97-132.

Melville, N., Kraemer, K.L., Gurbaxani, V., 2004. Information technology and organizational performance: An integrative model of it business value. MIS Quarterly 28(2), 283-322.

Morrison, C.J., 1997. Assessing the productivity of information technology equipment in U.S. manufacturing industries. Review of Economics and Statistics 79 (3), 471-481.

43

Nalebuff, B., Stiglitz, J., 1983. Information, competition, and markets. American Economic Review 73(2), 278-283.

Peng, M.W., 2004. Outside directors and firm performance during institutional transitions. Strategic Management Journal 25 (5), 453-471.

Porter, M.E. 1985. Competitive advantage: Creating and sustaining superior performance. New York: Free Press.

Scharfstein, D., 1988. Product market competition and managerial slack. Rand Journal of Economics 19(1), 147-155.

Scharfstein, D., Stein, J., 1990. Herd behavior and investment. American Economic Review 80(3), 465-479.

Schmidt, K., 1997. Managerial incentives and product market competition. Review of Economic Studies 64(2), 191-213.

Schumpeter, J.A., 1970. Capitalism, Socialism, and Democracy, 3rd ed. New York: Columbia University Press.

Shleifer, A., Vishny, R.W., 1997. A survey of corporate governance. Journal of Finance 52 (2), 737-783.

Short, H., 1994. Ownership, control, financial structure and the performance of firms. Journal of Economic Surveys 8 (3), 203-249.

Simunic, D.A., 1980. The pricing of audit services: Theory and evidence. Journal of Accounting Research 18 (1), 161-190.

Sutton, R.I., Callahan, A.L., 1987. The stigma of bankruptcy: Spoiled organizational image and its management. Academy of Management Journal 30 (3), 405-436.

Szymanski, D.M., Bharadwaj, S.G., Varadarajan, P.R., 1993. An analysis of the market share-profitability relationship. Journal of Marketing 57 (3), 1-18.

Thong, J.Y.L., 1999. An integrated model of information systems adoption in small businesses. Journal of Management Information Systems 15 (4), 187-214.

Verrecchia, R.E., Weber, J., 2006. Redacted disclosure. Journal of Accounting Research 44(4), 791-814.

Wade, M., Hulland, J., 2004. Review: The resource-based view and information systems research: Review, extension, and suggestions for future research. MIS Quarterly 28 (1), 107-142.

Waring, G.F., 1996. Industry differences in the persistence of firm-specific returns. American Economic Review 86 (5), 1253-1265.

Weisbach, M.S., 1988. Outside directors and CEO turnover. Journal of Financial Economics 20(1/2), 431-460.

Yang, Y., 2006. The Taiwanese notebook computer production network in China. Center on China's Transnational Relations Research Workshop.

Zahra, S.A., Pearce II, J.A., 1989. Board of directors and corporate financial performance: A review and integrative model. Journal of Management 15 (2), 291-334.