Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms

G Model

IBR-1022; No. of Pages 10 International Business Review xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

International Business Review journal homepage: www.elsevier.com/locate/ibusrev

Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms Kang Rae Cho a,1, Chia-Hsing Huang b,1, Prasad Padmanabhan c,1,* a b c

The Business School, University of Colorado Denver, Denver, CO 80217, USA SolBridge International School of Business, Daejeon 300-814, Republic of Korea Bill Greehey School of Business, St. Mary’s University, San Antonio, TX 78228, USA

A R T I C L E I N F O

A B S T R A C T

Article history: Received 20 November 2012 Received in revised form 3 March 2013 Accepted 4 June 2013

We propose an alternate context-based extension to the agency theory-grounded explanation of foreign ownership mode choices proposed in the literature. Using a sample of Taiwanese firms investing in the greater China region over the 2001–2009 period, we show that both economic and non-economic factors influence the choice of foreign ownership mode. In addition, we document that higher institutional ownership percentages motivate Taiwanese firms to select shared ownership in the greater China region. Further, no long term compensation mix/ownership structure link is found. These findings run counter to a theory provided for foreign ownership mode choices of US based firms. Our findings provide support for the validity of stewardship and social capital theory, but not financial incentives-based agency theory, for Taiwanese firms investing in the greater China region. ß 2013 Elsevier Ltd. All rights reserved.

Keywords: Agency theory Corporate governance Foreign ownership mode choice greater China region

1. Introduction In a recent seminal paper, Musteen, Datta, and Herrmann (2009) introduce agency theory into the foreign entry mode literature. Using US firm-level foreign entry mode data from 1991 to 1998, they present evidence that firms prefer full ownership of their foreign operations when there is a high proportion of institutional owners or inside owners in its common equity portfolio. Furthermore, they also document evidence that if the CEO’s compensation mix is biased toward long term performance, then there is a greater propensity toward full ownership. The agency theory-based extensions to the foreign ownership mode literature presented in Musteen et al. (2009) readily apply to firms where corporate strategic decisions are primarily motivated by financial considerations. Higher proportion of long term options in their compensation mix persuades managers to assume greater economic risks (and hence the choice of full ownership over shared ownership) in their foreign entry mode decisions since they provide management with monetary benefits if they pursue long term oriented strategies. Similarly, firms with larger proportion of

* Corresponding author. Tel.: +1 2104312034. E-mail addresses: [email protected] (K.R. Cho), [email protected] (C.-H. Huang), [email protected] (P. Padmanabhan). 1 All authors contributed equally.

institutional investors opt for full ownership over shared ownership because they already are well diversified outside of their investments in the company, and hence can afford to assume greater economic risks. Musteen et al. (2009) also buttress their arguments using the link between corporate R&D investments and the proportion of institutional investments in the firm. They suggest that the fact that high levels of institutional ownership are positively correlated with corporate R&D spending is testimonial that institutional owners discourage strategies that reduce economic risks, allowing them to prefer full ownership. They also provide empirical support that higher levels of insider ownership motivate firms to seek full ownership for their foreign affiliates. They argue that ‘‘. . . when insider ownership is substantial, the financial interests of insiders and those of shareholders tend to converge. . .’’ (p. 326). Furthermore, they suggest that ‘‘. . . the goals and risk preferences of executives with limited equity stakes are likely to be different from those of shareholders. . .’’(p. 326). Again, they are referring to economic risks when they develop their arguments. An important question is whether the agency theory can apply in a different context, i.e., when firms are not motivated primarily by financial considerations when deciding on foreign ownership mode choice. It is well known that Asian firms (documented in Section 2) inject non-economic factors in the strategic corporate decision making process. A well-established body of theoretical and empirical literature on the decision making strategies of Asian firms provides clues that the (financial incentives based) agency

0969-5931/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ibusrev.2013.06.005

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005

G Model

IBR-1022; No. of Pages 10 2

K.R. Cho et al. / International Business Review xxx (2013) xxx–xxx

theory does not work very well in an Asian context (for example, Gedajlovic & Shapiro, 2002; Lee & O’Neill, 2003; Tian & Lau, 2001). What happens to the ownership structure/executive compensation/corporate governance linkages when firms operate in an environment where risks include both economic and noneconomic (network-based) relationship risks? Addressing this question becomes important from an academic perspective because it provides a better understanding of what motivates ownership choices for firms from the greater China region. Practitioners will also benefit from this research since they will have a better understanding of the factors that motive ownership choices of their potential partners in the greater China region. For instance, managers thinking of entering into joint ventures or strategic partnerships in these countries may want to consider local partner inclinations toward shared ownership because of non-agency theoretic considerations. In this paper, we provide empirical evidence that a sample of Taiwanese firms investing in the greater China region over the 2001–2009 period consider non-economic factors over economic ones when deciding on ownership mode choice, offering an alternative explanation of major factors influencing the foreign ownership mode choice to the agency theory-based model (Musteen et al., 2009). The paper is organized as follows. In Section 2, we motivate our paper by presenting a brief theoretical and empirical literature review on the importance of noneconomic factors in an Asian context that affects corporate decision making. This is followed by a brief description of the relevant literature associated with the determinants of foreign ownership mode choice. Next, sample details and characteristics are followed by a presentation of logit regression results. Findings are discussed in the penultimate section, followed by concluding comments and policy implications in Section 6. 2. Importance of non-economic factors in corporate strategic choices in an Asian context The literature is replete with many examples of non-economic factors that influence a variety of strategic choices made by Asian firms. Most of them are related to the importance of network-based relationship in Asian societies. Oftentimes, these non-economic factors play a more important role than economic factors. Most notable example is the importance of guanxi2 in the context of decision making. Park and Luo (2001) suggest that guanxi critically influences firm performance in China. Guanxi is also widely practiced by firms originating from other Chinese dominated Asian countries that include Taiwan, Hong Kong, and Singapore (see, for example, Ai, 2006; Chung, 2006; Hsu & Saxenian, 2000; Huang, Baek, & Min, 2010; Liou, 2009). Concepts deeply embodied in guanxi are the concepts of mianzi (face) and renqing. It is important for firms to maintain mianzi (an intangible form of social currency) to foster/ enhance their guanxi business networks. Renqing is another form of social capital that obligates firms to reciprocate in guanxi networks, with significant social and monetary costs to the firm. The elements of mianzi and renqing make the concept of guanxi unique compared with the Western concept of social capital. In the current context of Taiwanese firms, firms opting for full ownership can expect to incur current and future social/monetary transaction costs of offending a guanxi partner expecting reciprocity since full ownership implies potential partners are shut out of the deal. As Ai (2006) states, [guanxi] business culture ‘‘. . . is . . . long term oriented with more concern for benefit of members of the entire network. . .’’ (p. 116). Empirically, there have been strong links between guanxi based 2 Guanxi refers to ‘‘. . . the concept of drawing on a web of connections to secure favors in personal and organizational relations. . .’’ (Park & Luo, 2001,p. 455). Guanxi can also be viewed as social capital or social networking.

variables and a firm’s accounting and market performance (Luo, 1997). Park and Luo (2001) show that guanxi is related to sales growth and not to profit growth. In the current context, this is important because agency theory relies on monetary incentives to motive managers. Hence, Asian firms will fail to motivate managers on the basis of profit growth if they (managers) believe that guanxi connections are more important. Another example of non-economic factors for Asian firms is related to the role of insider (or family controlled) and institutionally dominated boards (as is typical for Asian firms) on corporate decision making. Unlike in the US, insider and institutional owners, who are dominated by politically connected members, may prefer strategies that decrease network-based relationship risks, even if it implies the assumption of higher levels of economic risk. For instance, Gedajlovic and Shapiro (2002) document the co-existence of agency effects (caused by economic incentives) and redistribution effects (caused by Japanese social context) when investigating the relationship between ownership concentration and firm performance for Japanese firms. On the one hand, they document evidence of a positive link between ownership concentration (insider dominated boards) and subsequent financial performance for Japanese firms (agency effects). On the other hand, for poorly performing Japanese firms, they find a significant negative relationship between ownership concentration and financial performance. As the Japanese firms engage in the practice of redistribution to promote ‘‘. . .inter-corporate goals of risk reduction and mutual assistance. . .’’ (p. 567) to members of the Keiretsu,3 they find poorly performing Japanese firms with high ownership concentration benefit from the transfer of financial resources from profitable firms (redistribution effects). However, they find that for the entire sample of Japanese firms, the redistribution effects are stronger than the agency effects, indicating that the Japanese ‘social context’ (which emphasizes network-based relationships) matters. Similarly, Lee and O’Neill (2003) explicitly show that relationshiporiented forces dominate market forces when explaining observed relationships between ownership structure and R&D investments in Japan. Finally, Kim (2005) finds evidence that South Korean firms lose financial value if their boards have dense networks, while external network connections of the board enhance firm value. Chen (2001) finds that, unlike Western firms, Chinese and Taiwanese firms tend to place long-term family interests (non-economic factors) ahead of shareholder interests (economic factors) and emphasize the maintenance of long term family prestige over short term profits and preservation/enhancement of shareholder value. In conclusion, the importance of such non-economic factors in Asian corporate decision-making has led many researchers to conclude that the stewardship theory, where managers make decisions based on their role as ‘stewards’ for other principals (for more on stewardship theory, see Fox & Hamilton, 1994; Lane, Cannella, & Lubatkin, 1998), is more relevant/useful than the agency theory in explaining various corporate strategic decisions by Asian firms (Lee & O’Neill, 2003; Peng, Zhang, & Li, 2007; Tian & Lau, 2001).4 In addition, many corporate decision makers (shareholders) in Asian firms have limited ability to achieve economic diversification domestically. It is well known that institutional investors

3 Provision of ‘‘mutual assistance’’ is also consistent with the guanxi business model prescriptions. 4 Here, we argue that stewardship theory can provide one explanation for the importance of non economic factors in influencing foreign ownership choice behavior of Asian firms investing in the greater China region. Other researchers (cited earlier) have confirmed the validity of stewardship theory in an Asian context. For example, Lin (2005), Tian and Lau (2001), and Peng et al. (2007) provide strong evidence affirming the validity of stewardship theory in Asian countries. We are not suggesting that support against the validity of agency theory for greater China countries is de facto support for the validity of stewardship theory or social context theory in this context.

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005

G Model

IBR-1022; No. of Pages 10 K.R. Cho et al. / International Business Review xxx (2013) xxx–xxx

3

a) Musteen, et al. (2009) model: High Institutional/Insider Ownership

(+)

Financial risks and rewards

(+)

High management compensation via stock options

(+)

Financial risks and rewards

(+)

(+)

Non-financial risks and rewards

Full ownership

Full ownership

b) Our study model: High Institutional/Insider Ownership

(+)

Shared ownership

Financial risks and rewards

(+) High management compensation via stock options

(+) ns

Ownership mode

ns Non-financial risks and rewards

Ns = no effect, or uncertain relationship Figure 1. The proposed enriched model linking corporate governance with foreign ownership model in the greater China region.

This paper is designed to extend their analysis using Taiwanese data for many reasons. First, as indicated earlier, it is well known that institutional investors exhibit a ‘home bias’ in terms of their financial holdings. Institutional investors in many smaller countries may be subject (by governmental restrictions) to invest in their home country assets. Taiwanese institutional investors are

less diversified than their US counterparts.5 Second, Musteen et al. (2009) use the R&D expenditure/institutional ownership link to argue for a preference toward full ownership for firms in their sample. As will be presented later, we find evidence that institutional ownership percentages are uncorrelated with R&D expenditures for sample Taiwanese firms. There is also strong direct empirical evidence against the validity of financial incentive based agency theory in Taiwan. For example, Lin (2005) presents evidence that the corporate governance structure of manufacturing companies in Taiwan (1997–1999 period) is better explained by stewardship theory than by agency theory. In particular, she finds that when the CEO is also a chairman of the company, he/she acts as a steward for the company. In addition, she finds that CEO compensation is negatively correlated with the board’s effectiveness of control over the firm. Hung and Chen (2009) also provide evidence that agency theory does not work for small and medium Taiwanese enterprises using data for the 1999–2006 period. They indicate support for the stewardship theory for Taiwanese small to medium enterprises. Next, Yu (2010) presents evidence that a board (that typically includes institutional and other members) may pursue the goal of social objectives over economic objectives in Taiwan. It is not clear that the financial interests of shareholders and insider/ institutional investors converge in the Taiwanese corporate governance system.6 Given these significant differences between the two institutional environments, it is natural to ask whether the

5 If so, according to Musteen et al. (2009), they will be less inclined to opt for full ownership over shared ownership for decisions involving foreign entry modes, even if only economic risks are considered.

6 This convergence is essential to the arguments proposed by Musteen et al. (2009).

exhibit a ‘home bias’ in terms of their financial holdings (see for example, Driessen & Leuven, 2007). Although US based investors may also exhibit such biases, given the size of the US economy, their holdings can be considered as truly well diversified. However, it is difficult to conclude that Asian institutional investors are similarly well diversified. Hence, we argue that the Musteen et al. (2009) theory that relies exclusively on financial motivations to justify the choice of full ownership over shared ownership may not translate in social capital/guanxi/stewardship based Asian societies. Here, we propose a model, different than the one proposed by Musteen et al. (2009), to more closely capture foreign ownership choice behavior of Asian firms investing in the greater China region. While their analysis only accommodates financial risks associated with choice of foreign ownership modes, we allow for financial and nonfinancial risks in our model. The proposed enriched model is presented schematically in Fig. 1. 2.1. Why Taiwan?

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005

G Model

IBR-1022; No. of Pages 10 K.R. Cho et al. / International Business Review xxx (2013) xxx–xxx

4

Musteen et al. (2009) findings transport easily to such an environment. With respect to the issue of CEO compensation structures in a Taiwanese context, Musteen et al. (2009) contend that (U.S. based) CEOs may opt for the full ownership mode if the proportion of long term compensation to total compensation is high. They provide compelling arguments that CEOs may prefer short term, low risk strategies if their compensation mix is tied to short term performance. While this may be true for US data, recent evidence using Taiwanese data suggests that there is a large disconnect between long term financial compensation and the financial performance of Taiwanese firms. For example, Yu (2010) documents strong evidence that if the corporate boards in Taiwanese firms are characterized by individuals who are, directly or indirectly, politically connected, then the long term compensation packages for top management tends to be lower than if there are no such connections. Yu (2010) also makes the point that top management compensation packages in Taiwan are not independently decided by shareholders7 and that the board exerts significant control at shareholder meetings. In addition, Yu (2010) concludes the policies that pursue the maximization of social objectives rather than maximization of shareholder wealth reduce the effectiveness of incentive based compensation schemes, which serve as the keystone for the Musteen et al. (2009) arguments. Empirically, as documented later, there is no correlation between return on assets and the long term compensation mix for sample Taiwanese firms. To conclude, there are vast differences in the institutional and risk perceptional environment of Taiwanese and U.S. based firms. For the Taiwanese cases, the concept of risks can (and will) include non-economic risks. From this perspective, it can be argued that higher levels of shared ownership extend their guanxi relationships, provide new relationships, but also minimize their noneconomic transaction costs. These would be clearly prominent for Taiwanese firms investing in other guanxi-practicing countries of the greater China region. 2.2. Hypotheses Based on these rationalizations, we postulate the following context-specific hypotheses which are different from the agency theory-based hypotheses of Musteen et al. (2009): Hypothesis 1. The larger the percentage of institutional owners in the firm, the more likely is it that Taiwanese firms would opt for shared ownership of their foreign (in other guanxi-practicing regions) investments. Hypothesis 2. There is a positive relationship between CEO compensation mix and full ownership for Taiwanese foreign (in other guanxi-practicing regions) investments. Hypothesis 3. Higher percentages of insider ownership motivate Taiwanese firms to prefer shared ownership of their foreign (in other guanxi-practicing regions) investments.

7 This is in contrast to the situation in the United States. Yu (2005) also states that shareholders ‘‘. . .often waive their right to reject CEO compensation in shareholder meetings either due to their overly small shareholding or because there is a lack of cohesiveness among individual shareholders. . .’’ (Yu, 2005, p. 6). In addition, Solomon, Lin, Norton, and Solomon (2003), using survey evidence from Taiwan, document that the board of directors constitute the most important instrument in Taiwanese corporate governance. However, survey respondents also indicated that the presence of outside directors increased corporate accountability to shareholders. We submit that this increased accountability is less useful if shareholders waive their rights to reject CEO compensation packages.

3. Major determinants of foreign ownership mode choice The literature related to the determinants of foreign ownership mode is well established in the international business field (for a most recent comprehensive review, see Brouthers & Hennart, 2007). The choice between the two alternative ownership structures (full or shared ownership) basically involves trade-offs related to the level of resource commitment (Stopford & Wells, 1972), the degree of control (Caves, 1982; Davidson, 1982; Root, 1987), the specification and assumption of risks and returns, and the degree of global rationalization (Franko, 1971; Hladik, 1985; Kim & Hwang, 1992; Stopford & Wells, 1972). Primarily, transaction cost-based arguments have been utilized successfully to rationalize a firm’s preference for ownership mode for its foreign affiliate. Major determinants for the choice include, among others, firm size, relative size of the foreign investment, R&D and/or advertising intensity, experience (general international, host country, and prior decision-specific), cultural distance, country risk, investment risk, and host country size and growth (Brouthers & Hennart, 2007). Recent notable additions to the determinants of foreign ownership mode include, as discussed before, the seminal works of Musteen et al. (2009), who invoke agency theory and corporate governance issues to further refine our understanding of how firms make foreign ownership structure decisions. Specifically, they recognize the role of ownership structure aimed at controlling opportunistic behavioral tendencies of managers (i.e., institutional ownership and insider managerial equity ownership) and incentive mechanisms designed to align interests of managers with those of shareholders (i.e., long-term performance-linked managerial reward structure) in the choice. They reason that the longterm orientation and greater risk-taking ability of institutional shareholders, the closer alignment of goals and interests between insider managerial equity owners and managers (i.e., reduced agency costs), and long-term performance-oriented CEO compensation contracts would favor a full ownership mode. Using US manufacturing firm data, they provide strong evidence supporting their contention. Since the primary focus of this study is to investigate whether a firm’s equity ownership structure and CEO compensation mix would play similar roles in the case of Taiwanese firms’ foreign ownership mode choices, we use the same set of variables (dependent, independent and control) as used in Musteen et al. (2009) to test our hypotheses. This is done in order to maintain consistency. Briefly, referencing Musteen et al. (2009), a firm will select full ownership (defined as foreign affiliate ownership in the 95–100% range) over shared ownership (ownership in the 10–95% range) if (1) it is large in firm size, (2) it is highly experienced in international business, (3) its pre-entry firm profitability is high, (4) it possesses high R&D intensity, (5) it faces a host country environment relatively accommodating of full ownership (i.e., large size and rapid growth), (6) it enters a culturally dissimilar country, (7) managers are rewarded for long term performance in terms of stock options, (8) the firm has a large proportion of inside directors, and (9) the firm has a large proportion of institutional shareholders. As rationalized earlier, we expect that (1) a Taiwanese firm is more likely to select shared ownership over full ownership if it has a large proportion of institutional or insider shareholders for their foreign (in other guanxi-practicing regions) investments (Hypotheses 1 and 3) and (2) there would be no relationship between CEO compensation mix and its ownership mode choice (Hypothesis 2), contrary to Musteen et al. (2009). For the roles of Musteen et al. (2009)’s other variables, we expect to find similar relationships.

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005

G Model

IBR-1022; No. of Pages 10 K.R. Cho et al. / International Business Review xxx (2013) xxx–xxx

4. Data, dependent and independent variables, and methodology 4.1. Data The M&A data for Taiwan foreign investments by publicly traded manufacturing firms were obtained from Zephyr data for the period 2001–2009.8 Other corporate financial data were obtained from the Taiwan Economic Journal and annual reports of public firms. GDP information and other host country economic data were collected from Datastream. In view of the focus of the study on regions where guanxi is important, the sample foreign investments are restricted to China, Hong Kong, and Singapore. These regions (and including Taiwan) are known to follow guanxi practices (Ai, 2006), and constitute the greater China region.9 Such targeted sample selection and additional analyses help us overcome data limitations which do not allow us a direct measurement of the extent of firm-level guanxi, enabling us to test our hypotheses and to draw valid conclusions regarding the role of non-economic factors in foreign ownership mode choice. 4.2. Dependent variable: full or shared ownership Following previous studies on ownership mode choice, this study uses the 95 percent cutoff point to capture the alternative ownership structures (for example, Gatignon & Anderson, 1988; Gomes-Casseres, 1989, 1990; Hennart, 1991; Musteen et al., 2009; Padmanabhan & Cho, 1996). A Taiwanese parent firm’s ownership share over 95 percent of equity is defined as full ownership (ACQMODE = 1), whereas ownership share in the 10–95 percent range (inclusive) represents shared ownership (ACQMODE = 0). 4.3. Independent variables Following the leads of past research in the area, the independent variables selected for this study are outlined below. Following Musteen et al. (2009), institutional ownership INSTPER is defined as the ratio of shares held by Taiwanese institutional investors to the total number of outstanding common shares in the fiscal year just prior to the event year.10 Insider ownership, INSO is captured as the ratio of the holdings of inside directors to total outstanding common shares in the fiscal year just prior to the event year. Next, LTCMIX is defined as the ratio of long term compensation (via options, etc.) to total compensation provided to top management in the fiscal year prior to the event year. The Black–Scholes options pricing method is utilized.11 We proxy residual government ownership GOVPER, as the ratio of common shares owned by the Taiwanese government to the total outstanding common shares in the fiscal year prior to the event year. We also introduce many control variables that have been shown in the literature to influence the ownership mode decision. Following Musteen et al. (2009), we proxy firm size (AEMP) by the natural logarithm of the 3 year average number of employees in 8 The start date of 2001 is selected because this is the year when Taiwanese firms were required by law to report long term compensation packages for senior executives. See Yu (2005) for more information. 9 The other Asian firms do not practice guanxi and hence are excluded from the analysis. Hence, from the original sample of 348 firms, 67 firms were excluded from the analysis. In China there is no restriction of full ownership for foreign companies, including companies from Taiwan. Taiwan companies are free to choose full ownership or shared ownership in China. See: http://www.gwytb.gov.cn/gjstfg/jjfl/ touzi/201101/t20110123_1724092.htm for more information on these rules. 10 Unless otherwise stated, all institutional ownership data and subsequent data referenced in the paper originate from the Zephyr data base, the corporate financial reports, the Taiwan Economic Journal, or Datastream. 11 This is unlike the procedure used by Musteen et al. (2009), who use the present value method to compute this ratio for US data.

5

the period prior to the event year. Likewise, firm level international experience (FSPER) is captured appropriately by the ratio of foreign sales to total sales in the year prior to the entry year. Pre entry firm profitability (PEROA) is captured as the 3 year average of pre entry ROA, prior to the event year. RND, a measure of R&D intensity, is captured as the ratio of R&D expenditures/total sales in the year prior to the event year. Cultural distance (CD) was measured using Hofstede’s (1980) measures.12 Finally, GDP (GDP of the host market), GDPCAP (GDP per capita of the host market), and GDPGRTE (growth rate in host GDP in the year prior to entry), respectively controlled for the size of the host and the level of economic development of the host. These statistics are obtained from International Financial Statistics Yearbooks. To maintain consistency, all values where a currency unit of measurement is required, are measured in US dollars. Exchange rate information is obtained using International Financial Statistics Yearbooks. 4.4. Methodology Since this study involves a study of the variables that serve as important discriminators of the ownership variable, and following prior literature, we use logit analysis since the dependent variable is a 0–1 variable (1 for full ownership, and 0 for shared ownership). This technique is well suited for the case in hand, where the phenomenon to be explained can only be measured as [0/1] alternatives. Logit analysis determines the probability (p) of full ownership as a function of a set of independent variables. A positive sign for the coefficient implies that the variable increases the likelihood of full ownership. The model can be expressed as a function of selected predictor variables x1, x2, . . ., xn as:

pðxi Þ ¼ pðy ¼ 1jxi Þ ¼

1 1 þ eðaþbxi Þ

where y is the dependent variable, xi is the vector of independent variables (outlined in Section 4.3) for the ith observation, a is the intercept parameter, and b is the vector of regression coefficients (Altman, Avery, Eisenbeis, & Sinkey, 1981). The logistic regression model can be written as:

y ¼ a þ b1 x1 þ b2 x2 þ    þ bn xn þ ei where y, a dummy that captures the ownership mode (full ownership = 1, shared ownership = 0); xi, the list of independent variables; a, intercept; bi, regression coefficients associated with xi. 5. Presentation and discussion of results 5.1. Presentation of results We use LOGIST procedure in the SAS program package to estimate the parameters of the logit regression model (Harrell, 1986). Table 1 (part A) presents the sample descriptive characteristics for the greater China sample. Host country profile for the complete foreign investment sample (including the greater China sample) is presented in Table 1, part B. Next, Table 1, part C, presents the industry classifications for the greater China 12 Various conceptual and measurement problems are suggested regarding the Hofstede (1980) indices and the Kogut and Singh (1988) measure, most recently and notably in Shenkar (2001). Consequently, caution must be exercised in interpreting the effects of cultural distance in this study. However, as documented in Malhotra, Sivakumar, and Zhu (2011), many studies have confirmed the current validity of the original constructs, and the validity of the Hostede constructs over time. See also Padmanabhan and Cho (1996).

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005

G Model

IBR-1022; No. of Pages 10 K.R. Cho et al. / International Business Review xxx (2013) xxx–xxx

6

Table 1 Summary sample characteristics: Taiwan manufacturing firms investing in greater China region. (A) Data characteristics – greater China investmentsa Variable

N

Mean

Std. Dev.

Minimum

Maximum

PEROA GDP GOVPER GDPCAP GDPGRTE INSTPER LTCMIX CD ACQMODE INSO AEMP FSPER RND

281 281 281 281 281 281 281 281 256 281 281 281 256

7.7578885 14.7316458 0.7652669 6.0725271 0.2438605 40.8176512 0.0205022 1.2760754 0.7578125 20.3720641 6.4177150 0.6230513 0.0286730

6.9252313 0.8966958 1.8565333 0.9948122 0.0941593 21.7067585 0.0999520 0.1185663 0.4292464 12.5764348 1.5191655 0.3251040 0.0314716

14.0400000 11.3578225 0 4.9182655 0.0737184 0.9400000 0 0.8004650 0 3.0000000 2.6390573 0 0.0000321

37.2933333 15.3375839 8.5600000 10.2482612 0.3556425 91.7700000 0.7539809 1.5933180 1.0000000 67.3900000 9.9255915 0.980000 0.1942917

(B) Host country profile of Taiwanese foreign investmentsb Country Australia Brazil Canada China The Czech Republic Germany Spain Great Britain Hong Kong India

Frequency

Country

Frequency

1 3 1 257 1 3 1 3 12 1

Japan Mexico Malaysia Netherlands Philippines Singapore Thailand United States Vietnam Total

10 1 1 3 2 12 4 24 8 348

(C) Sample industry profile – greater China investments SIC

Category

Frequency

SIC

Category

Frequency

20 21 22 23 24 25 26 27 28 29 Total industries Total frequency

Food and kindred Products Tobacco products Textile mill products Apparel Lumber and wood not furniture Furniture and fixtures Paper and allied products Printing, publishing Chemical and allied products Petroleum refining and ind.

12 0 5 6 0 0 6 0 17 1

30 31 32 33 34 35 36 37 38 39

Rubber and plastic products Leather and leather products Stone, clay, glass and concrete Primary metal industries Fabricated metal Machinery and computer Eq. Electronics exc computer Eq. Transportation equipment Measuring, analyzing, and contrl. Miscellaneous manu. ind.

8 0 12 11 4 53 127 4 14 1 15 281

a PEROA is the pre entry firm profitability, measured as the average return on assets over the three years prior to the entry year. This variable is expressed as a percentage. GDP is the nominal gross domestic product of the host market in the year prior to entry, expressed in US dollars. Logs are used. GOVPER represents the percentage equity of the entity owned by the Taiwan government in the year prior to entry. GDPCAP is the per capita gross domestic product of the host country in the year prior to entry, expressed in US dollars. Logs are used. GDPGRTE is the growth rate in GDP of the host country in the year prior to entry, expressed as a percentage. INSTPER is the percentage of equity owned by institutional owners in the year prior to entry, expressed as a percentage. LTCMIX is the ratio of long term compensation (via long term options) provided to top management in the year prior to entry, expressed as a percentage. CD is a composite index showing the overall cultural distance of each host country from the parent country (the United States) using Hofstede’s four indices (Hofstede, 1980). ACQMODE is a dummy which assumes a value of 1 if the entry is via full ownership (ownership percentage = >95%), and 0 otherwise. INSO is the percentage of equity holdings by inside directors in the year prior to entry, expressed as a percentage. AEMP is a proxy for firm size and is measured as the logarithm (base 10) of the average number of employees over a 3 year period prior to the entry year. FSPER is the ratio of foreign sales to total sales measured in the year prior to the entry year. RND is the ratio of parent’s research and development expenses expressed as a function of total sales in the year prior to the year of entry, expressed as a percentage. b The greater China region includes China (257 cases), Hong Kong (12 cases), Singapore (12 cases), and Taiwan.

investment sample. Correlation coefficients are presented in Table 2.13 Next, Table 3 presents the results of paired difference tests between the two groups (full ownership and shared ownership) across the selected set of independent variables for foreign investments in the greater China region.14 Paired difference

13 Based on these correlation coefficients, it does appear that there are possible multicollinearity problems that need to be addressed for some of the independent variables. These issues will be addressed later in the paper. 14 Since variables GDP, GDPGRTE, and GDPCAP were found to be highly correlated with each other, Table 3 contains paired difference test results using the original variables and using factor scores.

results are designed to detect the presence of independent variables that possess the ability to discriminate between full ownership and shared ownership mode. Table 4A reports regression results for the greater China sample with all variables included in the logit regression analysis. Since the independent variables GDP, GDPGRTE and GDPCAP are highly correlated with each other (as evidenced by the correlation coefficients presented in Table 2 and by the high variance inflation factors reported in Table 4A), we reran the regressions using factor scores generated using GDP, GDPGRTE, and GDPCAP. This variable is labeled as ECONFACT, and is used to capture the host country’s ‘economic’ attractiveness for investments. Results are reported in Table 4B.

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005

G Model

IBR-1022; No. of Pages 10 K.R. Cho et al. / International Business Review xxx (2013) xxx–xxx

7

Table 2 Correlation coefficients: Taiwanese manufacturing firms – greater China investments. Vars

ACQMODE

PEROA

GOVPER

GDPCAP

GDP

GDPGRTE

INSTPER

LTCMIX

INSO

FSPER

AEMP

RND

PEROA

0.0713 (0.2556) 0.1408 (0.0243) S0.1079 (0.0849) S0.0377 (0.5487) S0.1857 (0.0029) S0.1974 (0.0015) 0.0118 (0.8518) S0.0598 (0.3403) 0.0165 (0.7923) S0.1207 (0.0538) 0.0801 (0.2233) 0.0081 (0.8979)

–

–

–

–

–

–

–

–

–

–

–

0.0195 (0.7451) 0.0014 (0.9810) 0.0227 (0.7053) 0.0117 0.8444 0.2293 (0.0001) 0.0174 (0.7715) S0.03197 (0.5522) 0.11826 (0.0274) 0.04670 (0.3851) 0.21401 (0.0001) 0.1275 (0.0327)

–

–

–

–

–

–

–

–

–

–

0.0539 (0.3682) S0.1936 (0.0011) S0.2090 (0.0004) 0.1979 (0.0008) 0.0711 (0.2348) S0.2398 (<0.0001) S0.2298 (0.0001) 0.4832 (<0.0001) S0.0228 (0.7166) 0.1882 (0.0015)

–

–

–

–

–

–

–

–

–

S0.7772 (<0.0001) S0.2412 (<0.0001) 0.0156 (0.7945) 0.0473 (0.4301) S0.0953 (0.1111) 0.1342 (0.0245) 0.0240 (0.6882) 0.2247 (0.0003) 0.0490 (0.4131)

–

–

–

–

–

–

–

–

0.7868 (<0.0001) S0.0420 (0.4832) 0.0636 (0.2879) 0.1368 (0.0218) S0.0270 (0.6520) S0.2014 (0.0007) S0.0680 (0.2781) 0.1186 (0.0469)

–

–

–

–

–

–

–

S0.0558 (0.3512) 0.1405 (0.0185) 0.1267 (0.0338) 0.0760 (0.2038) S0.2758 (<0.0001) 0.0832 (0.1846) 0.1794 (0.0025)

–

–

–

–

–

–

0.0596 (0.3197) 0.2177 (0.0002) 0.0084 (0.8891) 0.5333 (<0.0001) 0.0065 (0.9178) 0.0749 (0.2107)

–

–

–

–

–

S0.0483 (0.4205) 0.0742 (0.2150) 0.0292 (0.5867) S0.0324 (0.5881) 0.0472 (0.4311)

–

–

–

–

0.0822 (0.1693) S0.1695 (0.0044) S0.0348 (0.5797) S0.0842 (0.1591)

–

–

–

S0.1993 (0.0008) 0.0864 (0.1683) S0.1081 (0.0705)

–

–

S0.0869 (0.1659) 0.0991 (0.0974)

–

GOVPER GDPCAP GDP GDPGRTE INSTPER LTCMIX INSO FSPER AEMP RND CD

0.1990 (0.0014)

(Coefficients (p values) significant at the 5% or higher level are bolded).

Table 3 Paired T-test comparison between full ownership and shared ownership samples across independent variables: greater China investments. Variable

Paired difference T test full – shared using all vars

Paired difference T test full – shared using factor scores

Dependent variable: ACQMODE (ACQMODE). ACQMODE = 1 if the firm has = >95% ownership, and zero otherwise GOVPER S1.947* S1.947* AEMP S1.756* S1.755* FSPER 0.2499 0.2499 PEROA 1.3129 1.3129 RND 1.476 1.476 CD 0.1187 .1171 INSO S0.9319 S0.9319 GDP/FACTOR S0.5447 S0.6495 * GDPCAP S1.6743 S GDPGRTE S2.7923*** S INSTPER S3.1726*** S 3.1726*** LTCMIX 0.1862 0.1864 Dependent variable: ACQMODE = 1 if the firm has >= 95% ownership, and zero otherwise. Variable definitions can be found in the legend under Table 1. ***, **, and * represent significance at 1%, 5%, and 10% levels, respectively.

In view of the possible multicollinearity between predictor variables, we run and report the stepwise regression results for the greater China sample are presented in Table 5. The greater China investment sample highlighted in Table 1A suggests that sample Taiwanese manufacturing firms have recorded 281 cases of investments in the greater China region over the 2001–2009 period. The number of employees (expressed in log millions) range from a minimum of 2.639 to a maximum of 9.926. The institutional ownership ranges from a low of 0.94 percent to a high of 91.77%. Insider ownership also ranges from a low of 3% to a high of 67.39%. Finally, long term compensation as a percent of total salary for the CEO ranges from a low of 0% to a high of 75.40%. Next, sample host country presented in Table 1B indicates that over 73.85% of all Taiwanese manufacturing firm investments are targeted toward greater China. This fact should not be surprising since greater China and Taiwan share a long

history of business relationships. For the greater China targets, 257 investments were made in Greater China, with 12 cases each for Hong Kong and Singapore. Sample industry classifications are presented in Table 1C. The sample includes 15 different industries as categorized by 2-digit SIC codes. In terms of target industries, most investments (127 cases) occur under SIC 36 code (electronics, excluding computer equipment), followed by 53 cases under SIC code 35 (machinery and computer equipment). Chemicals and allied products (SIC code 28) is the next biggest sector for target investments with 17 cases. The other cases are generally evenly spread out over the rest of the manufacturing spectrum. 5.2. Discussion of results An examination of correlation coefficients (Table 2) reveals that some independent variables for the greater China sample are highly correlated with each other, enough to possibly cause problems in logit regressions. For instance, GDP is highly correlated with GDPCAP and GDPGRTE. PEROA is also highly correlated with INSTPER, FSPER, RND, and CD. GOVPER is significantly correlated with many independent variables.15 Other sample independent variables are also correlated with each other. Also, ACQMODE (the dependent variable that captures the incidence of full vs. shared ownership) is significantly correlated with GOVPER, GDPCAP, GDPGRTE, INSTPER, and AEMP. In view of the strong correlations between many independent variables included in the study, and the low signal to noise ratio associated with studies of this nature, we also run stepwise regressions on all independent variables. 5.2.1. Preliminary indicators To provide some preliminary evidence related to the importance of non-economic factors in entry mode ownership decisions, we examine the correlations from Table 2. RND seems to be 15 We report variance inflation factors (VIFs) for all independent variables (Tables 4 and 5) in view of possible multicollinearity problems. All VIFs seem to be within acceptable limits.

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005

G Model

IBR-1022; No. of Pages 10 K.R. Cho et al. / International Business Review xxx (2013) xxx–xxx

8

Table 4 Logit regression results: ownership mode determinants for Taiwanese manufacturing firms: greater China investments. Est.

Std. error

Wald Chi-SQ

Pr > CHSQ

Var inf.a

Odds ratiob

(A) Regression using original variables INTERCEPT 1 GOVPER 1 AEMP 1 FSPER 1 PEROA 1 RND 1 1 CD INSO 1 GDP 1 GDPCAP 1 GDPGRTE 1 INSTPER 1 LTCMIX 1

S2.1803 S0.1774 S0.1795 0.3061 0.0331 8.0906 2.1242 S0.0119 0.4574 S0.2635 S11.7945 S0.0136 1.5400

31.1543 0.1119 0.1627 0.5994 0.0269 6.2998 1.5341 0.0145 1.8654 1.1358 11.1081 0.0115 1.7111

0.0049 2.5161 1.2176 0.2607 1.5192 1.6493 1.9171 0.6749 0.0601 0.0538 1.1274 1.4120 0.8100

0.9442 0.1127 0.2698 0.6096 0.2177 0.1990 0.1662 0.4113 0.8063 0.8165 0.2883 0.2347 0.3681

– 1.4657 2.3332 1.1663 1.2433 1.1974 1.3171 1.3474 95.737 40.908 35.684 2.1773 1.1053

– 0.837 0.836 1.358 1.034 >999 8.366 0.988 1.580 0.768 <0.001 0.986 4.665

Var

Est.

Std. error

Wald Chi-SQ

Pr > CHSQ

Var Inf.a

Odds ratiob

2.1617 0.1038 0.1517 0.5469 0.0259 6.4651 1.5590 0.0135 0.1889 0.0110 1.6704

0.4484 2.9756 0.0799 0.1953 2.7291 0.5078 0.1622 0.3709 1.3207 2.7574 0.2668

0.5031 0.0845* 0.7775 0.6586 0.0985* 0.4761 0.6872 0.5425 0.2505 0.0931* 0.6055

– 1.4105 2.2312 1.1185 1.2306 1.1362 1.1317 1.3191 1.1305 2.1232 1.1093

– 0.836 0.958 0.785 1.042 100.2 1.873 0.992 0.805 0.982 2.370

Var

DF a

DF

b

(B) Regression using host economic factor INTERCEPT 1 1.4476 GOVPER 1 S0.1790 AEMP 1 S0.0429 FSPER 1 S0.2417 PEROA 1 0.0428 RND 1 4.6069 CD 1 0.6278 INSO 1 S0.00825 ECONFACT 1 S0.2170 INSTPER 1 S0.0179 LTCMIX 1 0.8628

ECONFACT are the factor scores developed from GDP, GDPCAP, and GDPGRTE. Dependent variable: ACQMODE. ACQMODE = 1 if the firm has = >95% ownership, and zero otherwise. N = 233; cases with full ownership = 178; cases with shared ownership = 55. Overall Chi-Sq = 16.3925, 10 df, p < 0.10. *** ** , , and * represent significance at the 1%, 5%, and 10% levels, respectively. Variable definitions can be found in Tables 1 and 2. a The variance inflation factors were generated using the PROC REG procedure in SAS. b The odds ratio captures the impact of a one unit change in the predictor variable (increase in probability), assuming that the other variables are held constant. This is akin to calculating the marginal effect.

Table 5 Logit stepwise regression results: ownership mode determinants for Taiwanese manufacturing firms: greater China investments. Var INTERCEPT PEROA INSTPER

DF 1 1 1

Est. 1.8161 0.0462 S0.0229

Std. error 0.3774 0.0243 0.00761

Wald Chi-SQ 23.1523 3.6333 9.0411

Pr > CHSQ

Var inf.a

Odds ratiob

1.05256 1.05256

1.047 0.977

***

<0.0001 0.0566* 0.0026***

Dependent variable: ACQMODE. ACQMODE = 1 if the firm has = >95% ownership, and zero otherwise. N = 233; cases with full ownership = 178. Cases with shared ownership = 55. Overall Chi-Sq = 11.0435, 2 df, p < 0.0001. Correct ratio = 71.7% * represents significance at the 10% level. ** represents significance at the 5% level. *** represents significance at the 1%level. a The variance inflation factors were generated using the PROC REG procedure in SAS. b The odds ratio captures the impact of a one unit change in the predictor variable (increase in probability), assuming that the other variables are held constant. This is akin to calculating the marginal effects.

uncorrelated with INSTPER indicating that higher levels of institutional ownership do not foster higher percentage investments in research and development. Next, if the Musteen et al. (2009) argument is valid for sample firms, return on assets should be positively associated with the long term compensation mix for the sample firms. Examination of reported correlations indicates that LTCMIX is uncorrelated with PEROA for the sample firms. There is no link between the firm’s financial profitability and the long term compensation to total compensation ratio for senior managers for the sample firms. Larger proportions of long term compensation for managers are not associated with full ownership for Taiwanese sample firms.

Finally, it is also interesting to determine whether there are any significant differences in means between the full ownership and shared ownership groups for each of the predictor variables. Any significant differences between key predictor variables will offer preliminary clues to the links between institutional/insider/long term compensation structure and entry ownership mode. Hence, for the greater China sample, and for each predictor variable, we performed a simple t-test to determine if there are any significant differences in means between the two groups. Results of these runs are reported in Table 3. From Table 3, we detect significant differences between the two groups for GOVPER (negative, 10% level), AEMP (negative, 10% level), INSTPER (negative, 1%),

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005

G Model

IBR-1022; No. of Pages 10 K.R. Cho et al. / International Business Review xxx (2013) xxx–xxx

GDPCAP (negative, 10%), and GDPGRTE (negative, 1%). Larger firms and firms with higher percentages of institutional owners tend to prefer shared ownership. These preliminary findings are inconsistent with the findings reported by Musteen et al. (2009) for US data. Finally, firms opt for shared ownership, ceteris paribus, even when the GDP per capita of the host country is larger. Clearly, even though the economic motivation for full ownership may be present (large GDP per capita) Taiwanese firms seem to prefer shared ownership.16 Do these results carry over when we conduct a formal regression analysis? We now examine the validity of the various hypotheses in a formal manner in Section 5.2.2. 5.2.2. Regression results For the overall regressions presented in Table 4A, the logit regressions do not have significant overall explanatory power, with model chi squares of 16. 3925 (p < 0.10). In addition, none of the sample variables are significant. We speculate that these insignificant results are driven by the weak signal to noise ratio, and the multicollinearity between sample independent variables. The variance inflation factor (VIF) is used to detect the multicollinearity problem. The variance inflation factors for GDP, GDPCAP, and GDPGRTE are 95.737, 40.908, and 35.684, respectively. Results using factor scores for GDP, GDPCAP, and GDPGRTE reported in Table 4B suggest that GOVPER (negative, 10%), PEROA (positive, 10%), and INSTPER (negative, 10%) are found to be important determinants of ownership mode choice for Taiwanese firms investing in the greater China region. However, before making conclusions with these results, it is prudent to check whether similar results are found when stepwise regressions are used. Stepwise regression results presented in Table 5 indicate that PEROA (10%) and INSTPER (1%) are found to be important determinants of ownership mode choice for Taiwanese firms.17 Taiwanese manufacturing firms prefer shared ownership when their return on assets is large, and when there is a high percentage of institutional ownership. The results presented in Table 5 are interesting because they run counter to the US evidence; the (economic based) agency theoretic explanations presented by Musteen et al. (2009) do not seem to work for Taiwanese investments in the greater China region. For instance, Musteen et al. (2009) find evidence that a firm with a high percentage of institutional owners prefer full ownership; we find evidence that Taiwanese institutional owners prefer shared ownership. In addition, we fail to find support for the hypothesis that higher proportions of long term stock options motivate management to seek full ownership for their foreign acquisitions. However, government residual ownership does affect ownership mode negatively–a higher percentage of government ownership motivates firms to opt for shared ownership. The agency theoretic explanations that work well for US data do not work at all for Taiwanese foreign investments in the greater China region. But, our findings confirm the alternate explanations (stewardship theory, social capital theory and guanxi) for Taiwanese manufacturing investments in other guanxi-practicing countries of the greater China region. Although our results indicate that some economic factors (embedded in ECONFACT, and in PEROA) motivate the full ownership form, the non-economic guanxi related factors are stronger than the economic factors, i.e., social costs associated with guanxi connections are also important. Overall, we cannot reject that institutional ownership is associated with shared ownership mode for the greater China sample 16 These results remain unchanged when factor scores using GDP, GDPCAP, and GDPGRTE are used. 17 We selected a criterion of 10% for inclusion and exclusion of variables in the stepwise regression.

9

(Hypothesis 1). Similarly, we can reject that there is a positive relationship between CEO compensation and full ownership for Taiwanese firms operating in the greater China region (Hypothesis 2). Finally, we can reject that insider ownership is related to shared ownership for the greater China sample (Hypothesis 3).

6. Conclusions and implications Based on results presented in this study, the following conclusions can be made. For Taiwanese manufacturing firms, and unlike for US manufacturing firms, long term executive compensation schemes seem to be unassociated with ownership structure for foreign investments in the greater China region. Further, insider and institutional ownership percentages are selectively related to ownership structure, but in a direction opposite to what was proposed by the agency theoretic explanations provided by Musteen et al. (2009). We submit that social (and monetary) costs associated with breaking existing guanxi networks persuade firms to consider shared ownership even if the economic factors dictate that they opt for full ownership. An alternate theory is that strategic decision making in Taiwan are based less on agency theory than on stewardship theory or social capital theory. Finally, the lack of support for the agency-theoretic relationship between ownership entry mode choice and firm ownership structure and executive compensation scheme for Taiwanese firms is indicative that the postulated relationship is more complex than what was originally envisioned. We provide strong evidence that non-financial risks can be important to explain the linkages between foreign ownership mode, executive compensation structure and corporate governance/ownership structure. Finally, the model proposed here can be viewed as an enriched version of the Musteen et al. (2009) model. Our results provide some policy implications for foreign firms aspiring to invest in Taiwan and other regions where noneconomic considerations are important. For instance, they should recognize the unique importance of guanxi connections when they seek local partners. As others have reported (for example, Chen et al., 2010; Luo, 1997), the preferred mode of entry into these countries is through a joint venture. Our findings reinforce this assessment. In terms of theory, attention needs to be paid to the importance of considering both economic and non economic risks when agency theoretic links are explored in the context of the full vs. shared ownership decision (or any strategic decisions made by firms from similar countries). Since other countries have similar network based relationships (for example, Keiretsu in Japan and Chaebol in South Korea), further insights may be gained by examining the entry mode decisions by manufacturing firms originating in these countries.

References Ai, J. (2006). Guanxi networks in China: Its importance and future trends. China and World Economy, 14(5), 105–118. Altman, E., Avery, R., Eisenbeis, R., & Sinkey, J. (1981). Application of classification techniques in business, banking and finance. Contemporary studies in economics and financial analysis (Vol. 3). Greenwich, Connecticut: JAI Press. Brouthers, K., & Hennart, J. F. (2007). Boundaries of the firm: Insights from international entry mode research. Journal of Management, 33(3), 395–425. Caves, R. (1982). Multinational enterprise and economic analysis. New York: Cambridge University Press. Chen, M-J. (2001). Inside Chinese business: A guide for managers worldwide. Boston: Harvard Business Review Press. Chen, H., Tian, Y., Ellinger, A., & Daugherty, P. (2010). Managing logistics outsourcing relationships: An empirical investigation in China. Journal of Business Logistics, 31(2), 279–299. Chung, C-N. (2006). Beyond guanxi: Network contingencies in Taiwanese business groups. Organization Studies, 27(4), 461–489. Davidson, W. (1982). Global strategic management. New York: John Wiley & Sons.

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005

G Model

IBR-1022; No. of Pages 10 10

K.R. Cho et al. / International Business Review xxx (2013) xxx–xxx

Driessen, J., & Leuven, L. (2007). International portfolio diversification benefits: Crosscountry evidence from a local perspective. Journal of Banking and Finance, 31(6), 1693–1712. Fox, M. A., & Hamilton, R. T. (1994). Ownership and diversification: Agency theory of stewardship theory. Journal of Management Studies, 31, 69–81. Franko, L. (1971). Joint venture survival in multinational corporations. New York: Praeger. Gatignon, H., & Anderson, E. (1988). The multinational corporation’s degree of control over subsidiaries: An empirical test of transaction cost explanation. Journal of Law, Economics, and Organization, 4(2), 305–336. Gedajlovic, E., & Shapiro, D. M. (2002). Ownership structure and firm profitability in Japan. Academy of Management Journal, 45(2), 565–575. Gomes-Casseres, B. (1989). Ownership structures of foreign subsidiaries. Journal of Economic Behavior and Organization, 11, 1–25. Gomes-Casseres, B. (1990). Firm ownership preferences and host government restrictions: An integrated approach. Journal of International Business Studies, 21(1), 1–22. Harrell, F., Jr. (1986). The LOGIT procedure. In R. P. Hastings (Ed.), SUGI supplemental library user’s guide (5th ed.). Cary, NC: SAS Institute. Hennart, J. F. (1991). The transaction costs theory of joint ventures: An empirical study of Japanese subsidiaries in the United States. Management Science, 37(4), 483–497. Hladik, K. (1985). International joint ventures: An economic analysis of U.S. foreign business partnerships. Lexington, MA: Lexington Books. Hofstede, G. (1980). Culture’s consequences: International differences in work related values. Beverly Hills, CA: Sage. Hsu, J.-Y., & Saxenian, A. (2000). The limits of guanxi capitalism: Transnational collaboration between Taiwan and the USA. Environment and Planning, 32(11), 1991–2005. Huang, L-Y., Baek, Y., & Min, S. (2010). Key drivers of guanxi in China among Taiwanese small to medium-sized firms. International Business Research, 3(1), 136–146. Hung, J-H., & Chen, H-J. (2009). Minimum shareholding requirements for insiders: Evidence from Taiwanese SMEs. Corporate Governance: An International Review, 17(1), 35–46. Kim, Y. (2005). Board network characteristics and firm performance in Korea. Corporate Governance, 13(November (6)), 800–808. Kim, W. C., & Hwang, P. (1992). Global strategy and multinationals’ entry mode choice. Journal of International Business Studies, 23(1), 29–54. Kogut, B., & Singh, H. (1988). The effect of national culture on the choice of entry mode. Journal of International Business Studies, 19(3), 411–432. Lane, P. J., Cannella, A. A., & Lubatkin, M. H. (1998). Agency problems as antecedents to unrelated mergers and diversification: Amihud and Lev reconsidered. Strategic Management Journal, 19, 555–578.

Lee, P. M., & O’Neill, H. M. (2003). Ownership structure and R&D investments of U.S. and Japanese firms: Agency and stewardship perspectives. Academy of Management Journal, 46(2), 212–225. Lin, Y-F. (2005). Corporate governance, leadership structure and CEO compensation: Evidence from Taiwan. Corporate Governance, 13(November (6)), 824–835. Liou, D.-Y. (2009). Globalization with guanxi for Taiwanese high-tech industry to China: Panacea or pandora’s box. paper presented in Portland investment conference on management of engineering & technology. Luo, Y. (1997). Guanxi and performance of foreign-invested enterprises in China: An empirical inquiry. Management International Review, 37(1), 51–70. Malhotra, S., Sivakumar, K., & Zhu, P. (2011). Curvilinear relationship between cultural distance and equity participation: An empirical analysis of cross-border acquisition. Journal of International Management, 17(4), 316–332. Musteen, M., Datta, D., & Herrmann, P. (2009). Ownership structure and CEO compensation: Implications for the choice of foreign entry modes. Journal of International Business Studies, 40(2), 321–388. Padmanabhan, P., & Cho, K. R. (1996). Ownership strategy for a foreign affiliate: An empirical investigation of Japanese firms. Management International Review, 36(1), 45–65. Park, S. H., & Luo, Y. (2001). Guanxi and Organizational dynamics: Organizational networking in Chinese firms. Strategic Management Journal, 22(5), 455–477. Peng, M. W., Zhang, S., & Li, X. (2007). CEO duality and firm performance during China’s institutional transitions. Management and Organization Review, 3(2), 205–225. Root, F. (1987). Entry strategies for international markets. Lexington, MA: DC Heath. Shenkar, O. (2001). Cultural distance revisited: Towards a more rigorous conceptualization and measurement of cultural differences. Journal of International Business Studies, 32(3), 519–535. Solomon, J. F., Lin, S. W., Norton, S. D., & Solomon, A. (2003). Corporate governance in Taiwan: Empirical evidence from Taiwanese company directors. Corporate Governance: An International Review, 11(3), 235–248. Stopford, J., & Wells, L. (1972). Managing the multinational enterprise. New York: Basic Books. Tian, J. J., & Lau, C-M. (2001). Board composition, leadership structure and performance in Chinese shareholding companies. Asia Pacific Journal of Management, 18, 245–263. Yu, H. Y. (2005). Corporate governance, political involvement and firm performance: An empirical investigation in Japan and Taiwan, International Conference on Business, Management and Economics, Turkey Reprinted in IJBME Conference Proceeding Vol. 2. Yu, H. Y. (2010). Politically connected boards and the structure of CEO compensation packages in Taiwanese firms. Asia-Pacific Journal of Financial Studies, 39(5), 553–686.

Please cite this article in press as: Cho, K. R., et al. Foreign ownership mode, executive compensation structure, and corporate governance: Has the literature missed an important link? Evidence from Taiwanese firms. International Business Review (2013), http:// dx.doi.org/10.1016/j.ibusrev.2013.06.005