The Outcome of Politically Connected Boards on Commercial Bank Performance in Malaysia

,


Introduction
The association between the board of directors and firm performance is not easy to explain using a single governance theory (Nicholson & Kiel, 2007) (Note 1).Hambrick and Mason's (1984) upper echelons theory has drawn on the disciplines of organizational behaviour and strategic management.The assertion of upper echelons theory identifies observable demographic characteristics, experiences, values, personalities, or strategic behaviours that influence management's value-added decision making styles, profitability variation and firm performance in response to environmental changes.Using the social identity principle (e.g., attractive-selection-attrition), Tajfel and Turner (1986) and Terjesen et al. (2009) explore how individuals seek to surround themselves in specific groups (so-called "groupthink") who share similar demographic profiles, perceptions and principles, which are then reinforced in intra-group communication.Therefore the portrayer of observable demographic profiles and the compositions of upper echelons (top executives or top management team) are determinants of strategic choices and the firm's performance outcomes (see Bantel & Jackson, 1989;Boeker, 1997).
Diverse human capital (Carter et al., 2010), valuable resources and knowledge of board of directors (Hillman et al., 2000;Jackling & Johl, 2009) affect board performance.Within the context of corporate governance, firms act as a socially cohesive group (Westphal & Zajac, 1995) to channel the human capital (Kresner, 1988) and social capital (Pfeffer & Salancik, 1978) of board directors to facilitate advice, information, legitimacy, communication, and power sharing.Demographic diversity may affect how boards govern their firms, and may influence group decisions.We believe that board diversity increases creativity and innovation, produces a more effective problem-solving approach, enhances the effectiveness of corporate leadership, and promotes more effective global relationships (Carter et al., 2003), and facilitates information exchange, discussion and group performance (Kang et al., 2007).We argue that resource dependency, which focuses on access to resources and preferential treatment, usually shows a positive sign.
Political connections are highly relevant in emerging markets.Malaysia is an appealing setting in this study for various reasons.It is a multi-ethnic and cultural country with concentration of ownership in the hands of a few politically connected Government-Linked Companies (GLCs).Ownership of Public-Listed Companies (PLCs) is identifiable along ethnic lines (Yatim et al., 2006).There are three main indigenous groups in Malaysia, namely Malays (Bumiputras), Chinese, and Indians.As of the 2010 census, the population of Malaysia was 28.3 million, and of that number, 50.4% were Malays, 24.6% were Chinese, 7% were Indians, and the remaining 18% were other natives.Each ethnic group differs economically, culturally and socially.The diverse nature of Malaysian society has become a competitive advantage for many Malaysian firms.The issues of sustainability and inter-generational equity are of great concern in order to achieve the Malaysian government's vision to achieve developed nation status by 2020.We argue that demographic incidents of shareholder activism significantly impact politically connected directors' decisions that ultimately affect commercial bank performance.To what extent demographic diversities of politically connected directors influence the performance of the firm, both observable demographic diversity (e.g., age, ethnicity and qualification) and non-observable cognitive diversity (e.g., education, experience, knowledge, values, perception, affection and personality characteristics), is the focus of our interest.
We have selected commercial banks in this study due to differences in the regulatory requirements by the Bank Negara Malaysia (BNM, the Central Bank of Malaysia).Within the context of the commercial bank industry (Note 2), incidents of shareholder activism by local institutional shareholders often take place.If the government has a direct controlling stake in firm X, this allows them to appoint the Board of Directors (BOD) in firm X.Consequently the politic-based BOD of this firm may react in favour of the objectives of government policy and may distinctively influence the firms' performance.For instance, Employees Provident Funds (EPF, the government agency social security and pension funds), and Permodalan Nasional Berhad (PNB, the government's national capital corporation) held 18.58% 2 direct control of the market in the year 2009 in Malayan Banking Berhad (MBB, the largest local financial group), with EPF: 11.95% and PNB: 6.63% stakes.There were also substantial indirect stake controls of 50.38% by Skim Amanah Saham Bumiputra (ASB, one of the government supported Bumiputra agencies) in MBB.This controlling stake exacerbated the crowding out effect and further aggravated the voting rights of minority investors.It may bring some destructive outcomes in the long term, and threaten moving towards nationalization instead of internationalization. Therefore, a study that focuses on the dark side of political connections, emphasizing over-embeddedness and subjective shareholder expropriation is of value.
Studies of the composition of boards using upper echelon theory in the commercial banking sector have not received much attention.This study sheds light on the performance of commercial banks' by examining the demographic diversity of the politically connected upper echelon BODs within the "groupthink".We use a fixed effect panel approach from 2000-2009 and test a lag effect on firm performance that is uncommon in past studies.Results reveal a positive association between age, size of ethnic minority, qualification (the observable demographic diversity), and experience (the non-observable cognitive diversity) of the politically connected directors on firm performance.Interaction effects show that performance gains are greater for elder politically connected non-ethnic minority directors.This paper shows that the independence of politically connected directors on decision making has been compromised.The contribution of this study is to expand the relationship between political connectedness of the governance structures among Malaysian firms, their upper echelons characteristics, and firm performance to address the corporate governance challenges as important drivers for decision making in the context of multi-ethnicity and cultural emerging economy.
The remainder of this study is organized as follows.The second section provides the applicable theories and hypotheses on the relation between board diversity and performance of the firm.The third section discusses the research methodology, sample, data and variables, followed by a section on model specification.The fourth section discusses the results.Finally, limitations are identified, and implication and concluding remarks in section 5.

Politically Connected Directors
A politically connected firm could be a group of large shareholders, such as the CEO, president, vice president, chairman or secretary, who control at least 10 percent of voting share, and are connected with a politician, party, minister, or Parliament member (Faccio, 2006).The extant literature documents an association between politically connected firms and Malay ownership corporations (Faccio, 2006;Fraser et al., 2006;Gomez & Jomo, 2002;Gul, 2006;Johnson & Mitton, 2003), but opinions about the link between political connections and firm performance have been mixed.Some studies found a positive and significant relationship i) between Malaysian firm leverage and political patronage (e.g., Fraser et al., 2006), and ii) between politically connected firms and firm performance (e.g., Fisman, 2001); while others form a different conclusion (e.g., Fan et al., 2007;Chizema & Kim, 2010).
Empirical evidence has proved that a bank with politically connected high-ranking officials enrich the business conditions by permeating many barriers (Baum et al., 2008).Politically connected firms receive preferential treatment from government in the form of cheaper sources of funding, favourable tax treatment and access to restricted licenses (Mohamad et al., 2007;Niessen & Ruenzi, 2010).Since the implementation of New Economic Policy in 1970, the non-Bumiputra businessmen have actively solicited and developed ties with politicians or other special interest groups to influence the allocation of financing (Johnson & Mitton, 2003) and to extract rents from government (Olson, 1982).Banks with affiliated political entity connections may obtain a licence to carry out transactions for governmental bodies and public authorities at rents set below-market-rate loans.Politically well-connected firms enjoy insider governmental informational advantage and find it relatively easier to win tenders for privatization projects and handling rights for the transactions of governmental institutions that likely increase the profitability of the connectedness firms (Faccio, 2006;Fisman, 2001;Roberts, 1990).Therefore, we hypothesize that: Hypothesis 1: Having a higher percentage of politically connected directors on board is positively related to bank performance, all else being constant.Johnson and Mitton (2003) found capital control in September 1998 primarily benefited firms with strong ties to the Malaysian Former Prime Minister.Furthermore, some evidence has suggested that politically connected firms were more likely to be rescued from financial turmoil than non-connected peers (Johnson & Mitton, 2003;Faccio, 2006).However there were adverse effects on politically connected firms at the early stage of the financial crisis when government cut the subsidies (Johnson & Mitton, 2003).Furthermore, Mitchell and Joseph (2010) also found that politically connected financial firms performed poorer after the removal of capital controls.The firms of politically connected directors performed poorly due to low financial transparency (Bushman et al., 2004), low professionalism and weak governance (Fan et al., 2007), and incompetent operation as the result of cronyism (Gul, 2006).These findings imply that politically connected firms have a higher probability of receiving government bailouts (Faccio et al., 2006).Therefore, we posit that: Hypothesis 2: Firms that have had high levels of political-connectedness over the later five years performed poorer than the previous years, all else being constant.

Age of Directors
The age and tenure of the directors "may determine his or her effectiveness in managing the firms" (Cornett et al., 2008, p.360).Alderfer (1986) argues that top board members with little experience may limit the effectiveness of firm, as it takes time to gain an in-depth knowledge of the firm and suggests that the more senior (or the longer the tenure of) the directors, the more likely their knowledge of the industry enhance firm performance.Age diversity of board members may be considered from various perspectives.Kang et al. (2007) believe that young board members are more energetic and are therefore more likely to aim for a successful future; middle board members are suitable to hold important positions in companies because they have the ability to integrate new ideas and confidence when making decisions; and senior board members can share their wisdom by giving valuable experience to management and evaluating decisions more accurately.These studies suggest the more senior the board members and the deeper the knowledge of the firms, the better the firm's performance, and hence a senior politically connected director may positively impact firm performance.We also examine the interaction term to test whether bank performance changes for elder politically connected directors.Based on these arguments, we posit that: Hypothesis 3: The age of directors is positively related to bank performance, all else being constant.
Hypothesis 4: The elderly politically connected directors are positively associated with firm performance, all else being constant.

Ethnicity of Directors
Theories from interdisciplinary sources provide a relationship between the ethnic board diversity and firm performance.However, the empirical evidence on this link is mixed.Haniffa and Cooke (2002) examine the relationship between board composition, culture and voluntary disclosure in Malaysia and conclude that Malays are less professional and secretive, and more uniform and conservative in providing disclosures, when compared to Chinese.In contrast, Ramasamy et al. (2007) claim that Malay CEOs showed higher corporate social performance than Chinese CEOs.Carter et al. (2010) examines the relationship between ethnic minority directors in a sample of U.S. boards and firm performance using fixed effect single regression equation, and they find evidence of a positive and significant relationship between the number of ethnic minority on board and return-on-asset (ROA), but no significant relationship is found between ethnic minority director and financial performance.Nevertheless, Erhardt et al. (2003) reveal a significant positive relationship between ethnic minority on both ROA and return-on-equity (ROE) for U.S. firms.Carter et al. (2003) also find a positive and significant relationship between ethnic minority directors on the board and Tobin's Q. Zahra and Stanton (1988), however, reveal no relationship between ethnic minority on the board and firm performance for U.S. firms.Based on Carter et al.'s (2010) and Adams and Ferreira's (2009) conjectures, we argue that ethnic minorities promote board independence because a board which has a diverse ethnicity background might be a more active board in considering issues that are not considered by a homogeneously composed board.We extend these studies by using an interaction term to examine whether bank performance changes for the elderly, ethnic minority directors with political connections.Therefore, we state the following hypotheses: Hypothesis 5: The number of ethnic minority directors on the board is positively associated with bank performance, ceteris paribus.
Hypothesis 6: The elderly politically connected ethnic minority directors change firm performance, all else being constant.

Qualifications of Directors
The Malaysian Code on Corporate Governance (MCCG) (Revised, 2007, p.11, Note 3) recommends that all directors should be equipped with talent, proficiency, and know-how to demonstrate the credibility and accountability of the management team and to promote the corporate image (Haniffa & Cooke, 2002).Higher education of top management directors in organizational contexts is positively related to receptivity to innovation, creativity, and better strategic decision making (Bantel, 1993;Becker, 1970).Therefore innovation has become a key firm strategy to gain competitive advantage (Hitt et al., 1996) and the qualifications of directors are positively related to firm performance (Morbey, 1988).This relationship suggests that education level and a diverse functional background are positively related to organizational performance (Bantel, 1993;Haniffa & Cooke, 2002), especially in business-related background such as in Economics, Law, Marketing, Accounting, Management, or Finance.In addition, Gray (1988) and Wallace and Cooke (1990) suggest that higher education level may increase demand for political awareness and corporate accountability.Hence, less qualified directors are possibly less effective than directors with business qualifications (Ferreira, 2009).Drawing on this strand, we put forward the seventh hypothesis: Hypothesis 7: The qualification of directors with business backgrounds increases bank performance, all else being constant.

Experience
The "groupthink" cohort is a group of top managerial teams who have some relevant data in common such as socioeconomic background, employment and industry, societal experience, functional track, and training, and have been imprinted on its members and shaped their values and perceptions.While directors are chosen based on "the old boy" network or "people like us", they may have their own priorities in decision making, but the politically connected directors have to support the agenda politically if the director perceives his adherence to the continued of directorship tenure with a firm.It is generally believed that homogeneous managerial teams, manifested as cohesiveness and insularity, will make strategic decisions quicker than heterogeneous teams (Hambrick & Mason, 1984).Janis (1972), however, argues that homogeneity leads to inferior decision making (cited in Hambrick & Mason, 1984).In a stable environment or routine situation, problem solving is best handled for a homogeneous group, but in a chaotic environment, problem solving is best managed by a heterogeneous group since diversity of skill allows more airing of alternatives (Filley et al., 1976).Therefore, the following hypothesis is posited: Hypothesis 8: The degree of peripheral-function experience of directors is positively related to the firm's performance, all else being constant.

Sample and Variables
The sample consisted of an unbalanced panel of all post-merged domestic commercial banks in Malaysia from 2000-2009 (Note 4).Data was extracted from the companies' websites, companies' Annual Reports, and Bankscope.Panel data analysis was employed to avoid unobservable heterogeneity problems in empirical analyses.
Dependent Variable.We measured firm performance using ROE, a commonly used accounting-based measure of firm performance.Following Fich and Shivdasani (2006), we did not use Tobin's Q because if financial or liquidity constraints cause some firms to under invest, the potential value of unexploited investments may lead to a high marginal Tobin's Q.If underinvestment is pervasive, formulation would erroneously treat a high market-to-book ratio, which indicates good governance.We obtained the data on ROE from Bankscope.
Explanatory variables.We define politically connected directors (POL) as the percentage of directors serving as current or former members of a political party, or current or former delegates from the ruling government (Fan et al., 2007;Niessen & Ruenzi, 2010).Alternately POL could be defined as informal ties between Malaysian politicians and firms run by Malay, Chinese and Indian (Gul, 2006).We measured age (AGE) as tenure in years held by directors.Director age captures the level of expertise built up by the top executive regarding the organization (Alderfer, 1986;Cornett et al., 2008).We expected that senior or longer-tenured directors were less likely to adversely affected performance.Ethnic minority (EM) was measured using the percentage of minority directors on the board (i.e., the Chinese and Indian), while qualification (QUA) measures the percentage of directors with qualifications in business backgrounds (i.e.either in Economics, Law, Marketing, Accounting, Management, or Finance).Experience (EXP) measures the percentage of directors who have experience in the banking industry.To allow for the possibility of partial adjustment to changes in other variables, we included a lagged dependent variable in the performance equation.The one-year lagged firm performance [ROE(-1)] was employed as stakeholders use previous year firm performance to estimate future performance.It is expected that higher value creation may lead them to expect satisfaction of their interests (Delgado-Garcia et al., 2010).Following Bhagat and Bolton (2008), we used the lagged-one value of ROE as an instrument in our estimations, but we lost data due to the one-year lag.To measure whether there was an association between age, political-connected directors and firm's performance, an interaction term POL*AGE was tested.To measure whether there was an association between age, ethnic minority and political-connected director on the firm's performance, another interaction term POL*AGE*EM was used.Other interaction terms, such as POL*AGE*EXP and POL*AGE*QUA, were also tested, but results were insignificant (results are not reported here).
Control variables.We controlled for firm and director characteristics through firm fixed effects regression methods.The effects of size were controlled for using the natural log of total assets, as it is related to market returns (Fama & French, 1992).We measured firm size using a natural log of the total assets of the firm.We controlled for board size, measured as the natural log of the number of directors elected on the board.Study considering the effect of board size and financial performance has developed contradictory arguments.Past studies have predicted a positive relation between board size and financial performance (e.g.Dalton et al., 1998;Forbes & Milliken, 1999;Jackling & Johl, 2009).Proponents of this view argue that larger boards provide a larger pool of information, and greater depth of intellectual knowledge may improve the quality of strategic decisions that in turn translate into better performance.While larger firms are argued to have better corporate performance, smaller firms are expected to be less controlled in the distribution of firm value (Delgado-Garcia et al., 2010), thus reducing firm performance.Nonetheless, evidence supports the assertion that larger boards encumber communication, coordination and eventually influence the decision-making competencies of the board (Cornett et al., 2008;Jensen, 1993), and smaller boards are more effective monitors than large boards (Yermack, 1996), suggesting that board size is inversely related to performance.Firm performance may drop as firms grow older (Loderer & Waelchli, 2010) and older firms tend to be larger after merger and acquisition; takeover hazards could increase with firm age, while firm size and age are negatively related to takeover hazard (Palepu, 1986).Therefore we controlled for firm age, measured as the number of years since the firm's first incorporation.The fiscal year-end was controlled for using a dummy variable, with a value of 1 for firms with a 31 December fiscal year-end and zero otherwise (Gul, 2006).Most firms in the sample had a December year-end and all commercial banks had been audited by large audit firms.

Model Specification
The empirical specification aims to explain the pace in commercial banks' performance using a wider demographic board characteristic.The regression used for testing was as follows: where i equals each firm, t equals time, and it  is the random error for each observation, which is independent and identically distributed with mean zero and variance σ 2 .PERF is the financial performance of the firm measured by commonly used ROE in governance; POL is the percentage of political connected directors in the board; AGE is the average age of the director; EM is a measure of percentage of ethnic minority on board, i.e., total number of Chinese and Indian directors on the board; QUA is the percentage of qualified directors; it 1 PERF  is the lagged value of ROE; EXP is the percentage of directors with prior experience in banking sector; POL* AGE and POL* AGE * EM are two interaction terms; FS is firm size, as the natural logarithm of the total assets; BS is board size, measured as the natural logarithm of the number of directors on the board; FA is firm age, measured as the number of years since its operation; FY is the fiscal year, as a dummy variable of 1 if financial ended 31 December, zero otherwise.

Panel Regression Method
A linear model of Equation ( 1) was estimated on a pooled data sample from 2000-2009 with a constant using fixed effect model estimation using Eview software.The general panel model is written as follows: where it PERF denotes the dependent variable (ROE it ) for country i at period t,  is kX i coefficient vector, and it X is the i-th observation on the k number of explanatory variables [k=12, see Eq.( 1)].The term i  is the cross-sectional unit residual, and the term t v is the unobservable time specific residual to account for firm effect and period effect, respectively.The term it  is the usual error term after removing individual and period effects.These residuals are components of the error term from Equation (1) where Panel regression offers flexibility in modelling heterogeneity in country-specific behaviour, and temporal changes in the country's macroeconomic condition.Pooling both time series and cross-section data provides more information, greater variation, less collinearity, greater degree of freedom and more efficiency (Gujarati & Porter, 2009).Fixed-effects models are not without their disadvantages.The constant covariates within individuals cannot be measured in this setting.The fixed effects model allows the unobserved firm effects to be correlated with the deterministic variables.If there is a correlation between the fixed effects and the determinants, the random effects model is preferred.The variance component of the dependent variable can be decomposed into: 2) is merely a simple pooled regression (with no firm or period effects).If the firm and period effects are insignificant, both i  and t v are zero; in this case OLS provides consistent and efficient estimates for both  and  .

Descriptive Statistics and Correlation Matrix
The Malaysian BODs in the commercial banking sector comprised of 6-15 directors, headed by an independent male "chairperson" and included not more than three female directors.The mean age of BOD is 58.99.Table 1 reports descriptive statistics for all variables.The measures of firm performance indicate that the firms were on average financially sound over the 10 years period.The log of return of equity (ROE) is 14.54, which suggests the net income was higher than the book value of total equity of the firm.The average number of ethnic minority directors (EM) on board was 44.89%, with 22.34% being politically connected directors (POL), 63.77% being directors with qualifications (QUA) in business, and 70.76% of the directors having background experience (EXP) in the banking industry over the sample period.For 78% of the sample firms the fiscal year ended on December 31.The log of the total assets of the firms (FS) is 10.83, with a board size (BS) of 2.29.Table 2 provides the bivariate correlation matrix for variables.The correlation coefficients between the measures of performance and board characteristic are within the acceptable ranged from 0.0008 and 0.6492.These coefficients confirm the choice of both deterministic variables and instruments.We found an inverse relationship for the percentages of ethnic minority and ROE.

Unit Root Tests
Augmented Dickey Fuller (ADF) -Fisher chi-square test, PP Fisher, and Breitung t-stat unit root tests were employed to examine stationarity of the series.The results of the unit root test are reported in Table 3. Results reveal that all series are I(1) variables, and the unit root test shows that on average pooled data of all variables are stationary at first difference.

Analysis
Table 4 reports the fixed effect regression models.The association between firm performance and POL, AGE, EM, QUA, EXP is tested in both Model 1 (with no control variable) and Model 2 (with control variable -FS and BS).By adding two interaction terms to Model 1, the interaction effects are tested in Model 3. In Model 4, we include lag-one firm performance to Model 1 to test the past year effect on today's performance.In Model 5, we test all variables in Model 4 plus one interaction term to test the interaction effect between POL*AGE on firm performance.By including the control variables to Model 5, a robustness test is carried out in Model 7. In Model 6, we include two interaction terms to test the effects e.g., POL*AGE*EM on firm performance, and lastly the full model is tested in Model 8 (with control variables and interaction terms).The probability of the F-statistic for the overall regression relationship for all independent variables is at least at p<.05, except for Model 7 at p<0.1.Overall, results reveal significant and positive signs in the predicted direction for the coefficients for POL, AGE, QUA, and EXP.The coefficient for POL is positive and significant for all models at p<.05, thereby providing support for Hypothesis 1.There is evidence suggesting that a higher percentage of politically connected directors on boards enhance bank performance by 3%, holding all other things constant (see Model 8).Therefore having politically connected directors would be an important aspect of corporate performance.This finding implicitly shows that directors representing investors' interests or non-political professionals might obstruct politicians' agendas.
The study reveals age diversity is significant and positively associated with ROE for four out of eight models.This study reveals that senior directors gain more experience and obtain adequate knowledge in banking industry, and thus are more effective in managing the firm.However, these senior directors are compromised due to their political affiliation.Firms with government-linked and politically connected directors need allies on the board to reinforce their objectives and policies that ultimately lead to better persuasive governance, increase professionalism and more attractive firm performances.This finding suggests that the more senior the firm's director, the greater experience, wisdom and understanding of the banking industry, and the better the firm performance.This finding provides weak evidence to support Hypothesis 3 that the age of directors is positively associated with bank performance.Therefore our result is consistent with the assertion that age diversity increases creativity and problem-solving capability Cornett et al. (2008) and Kang et al. (2007).Senior directors are more conservative in pursuing a firm's strategies and tend to focus on business activities that yield immediate profits in the short term during their service periods, ultimately improving firm performance (see Cox, 1993;Li et al., 2011;Milliken & Martins, 1996).
The results show a positive relationship between performance and the qualifications of directors.This finding supports the belief that top management teams who have qualifications in business-related disciplines such as Marketing, Management, Finance, Law or Economics, improved higher banks performance, and therefore supports Hypothesis 7.These qualified directors chose to increase firm performance to promote corporate image, and demonstrate accountability and credibility within the management team.On the other hand, four out of eight models show that directors with relevant business experience are significantly and positively associated with performance, weakly supporting Hypothesis 8.
Results, however, reveal neither ethnic minority nor past performance significantly impact firm performance.
We have no evidence to support the notion that firms with directors predominantly by Chinese and Indian outperform the Bumiputra directors in commercial banking industry.Hence, Hypothesis 5, i.e. the assertion that the number of ethnic minority directors on a board is positively associated with bank performance, is not supported.The nature of the relationship between ethnic minority on the board and firm performance is merely a public policy issue.On the other hand, there is no statistical difference when previous year firm performance is added to Models 4-8.This finding suggests that they are more likely to be under the control of politically connected directors for post-merging banks.
For the control variables, results reveal a positive relationship between performance and firm size at p<1.This finding is consistent with the argument that larger firms are expected to perform better because they are able to diversify their risks (Ghosh, 1998).In addition, directors are closely monitored by more analysts and thus they are under more pressure to perform well (Haniffa & Hudaib, 2006).However, there is no evidence to support the assertion that larger boards hinder communication, coordination and eventually the performance of the firm.
The coefficient of the interaction term of POL*AGE is negative and significant (e.g.,  =-0.754, p<.01 in Model 8).From Equation (1), the significant coefficient of the sum of 1  and 7  is 2.311 (3.065-0.754),providing evidence that senior political-connected directors enhance firm performance by approximately 2.3% compared to non-connected peers (supporting Hypothesis 4).When the ethnic minority factor is taken into account, the significant coefficient for POL*AGE*EM (  =0.0004, p<.05) of -0.7536 [the sum of  7 and  8 , i.e., -0.754+0.0004suggests that elderly political-connected ethnic minority director reduces firm performance by 0.75% (supporting Hypothesis 6).This result suggests that the market does react to the ethnic minority factor.In other words, expectations about merger-related gains are greater for senior politically connected non-ethnic minority directors.-2009 2000-2009 2000-2009 2001-2009 2001-2009 2001-2009 2001-2009  To test whether politically connected firms performed poorer in the later five years, we divided the sample periods into early years (2000)(2001)(2002)(2003)(2004), and later years (2005)(2006)(2007)(2008)(2009).Year 2004 has been chosen as the cut-off year because this was the last year the 4 th prime minister held official position in the government.Results in Table 5 reveal that the mean value for low politically connected firms shows significant improvement in performance during the later years; meanwhile the mean value for high politically connected firm is relatively lower during the later five years.Hence, we conclude that firms that are politically connected performed poorer in the later five years, thereby supporting Hypothesis 2. These firms performed poorly due to inefficient operation as a result of cronyism and the changing hand among politicians.

Endogeneity Test
The study of the upper echelons board-performance link is plagued with an endogeneity problem, and this problem may bias the results on firm performance.Instances of firms performing well could be due to the political connectedness of the upper-echelons.It could also be due to reputation that these delegates only chose the better performing firms to work for.That is, if the explanatory variable (X) and the error terms were correlated, the OLS estimate would likely be attributed to the X of some of the variation in the dependent variable that actually comes from the error term.In this paper we deal with this endogeneity problem by examining the issue of whether political connectedness affect firm performance or whether firm performance affects political-connectedness.Therefore, formally testing whether or not these explanatory variables are endogeneous is a matter of pre-requisite.We attempt to control for the possible reverse causality by using a time-lagged approach between politically connectedness and firm performance.Specifically, we use the following simple models to determine whether x ik in the model is endogenous or, more precisely, whether it is correlated with the error term u i , where x ik is a k-element row vector whose last element, x k .
First we test if beta is equal to zero (it means that x i has no predictability to y i ), save its residual as the first error term; and then X i,t is regressed on X i,t-1 , again save its residual as the second error term.Finally we test if the first error term is correlated with the second error term.Results reveal that we cannot reject the hypothesis (coefficient= 0.572, t-stat=1.070,P=0.288), which implies that error terms do not correlate to each other.Hence the endogeneity problem has been sorted out.

Discussion, Conclusion and Policy Implication
This study employs upper echelon theory to examine the impact of governance and demographic characteristics of politically connected board members on firm performance.We investigate the impact of political connections, age, ethnic minority, qualification, and experience of board members on commercial bank performance.Results reveal that bank performance of politically connected firms improves compared to the non-connected peers.Overall, we conclude that having a relatively higher number of politically connected directors, qualified senior directors, and peripheral-function experience of directors on the board will lead to higher bank performance, ceteris paribus.Consistent with upper echelon theory, the directors' greater knowledge base, creativity and innovation were a competitive advantage to improve bank performance.Hence, there is a need to legitimise the selection of directors with higher educational qualifications.Furthermore, diverse group dynamics are likely to impact controlling function and minimize potential agency issues (Erhardt et al., 2003).The finding in this study is consistent with the argument by Johnson and Mitton (2003) and Gul (2006).While neither ethnic minority nor past performance significantly impacts firm performance, the interaction effects show that ethnic minority plays a role in determining firm's performance, in which there is a positive association between elder non-ethnic minority politically connected directors and bank performance.
The findings suggest that corporate governance can be improved through better use of the entire aptitude capital of the directors.An interaction term that embraces age, ethnic diversity, and political connections is a good practice and serves as a business strategy for creativity and innovation that could strengthen firm performance.
The innovation needed now is not to eliminate the long implanted special connections practices within directors, but we need to delve into how politically connected directors can make a difference to improve firm performance as an injection booster to leap forward onto the next podium of economic growth and position Malaysia to become a high income nation by 2020.The Malaysian government has embraced an entire new paradigm via Economic Transformational Program and New Economic Model, particularly focused on private sector's accessibility to capital to boost and rejuvenate investment, facilitate modern business, encourage competition, and attract foreign direct investment in domestic banking institutions.
The results of this study imply that appropriate human capital for directorship brings positive value-added to boards.Therefore, reengineering GLCs, reenergizing the support to private sector, and creating a competitive domestic economy are of utmost importance.GLCs should engineer socioeconomic change through wealth redistribution, and should not crowd out private investment.Trust is an important element amongst different ethnic groups for fundamental economic network integration.Malaysia's major obstacle is the absence of fairer competition to raise competitiveness within the nation.This lack of fairness has caused difficulty for global foreign investors to invest in the nation.Continuous and significant regulatory framework improvements needs to be embedded into government practice with respect to transparency, non-discrimination, objective-based and performance-based criteria, market-based mechanisms and accountability to reap independence, integrity and soundness of policies and regulations.These informational levels will then be prescribed by mechanisms put in place within the competitive governance and legal environment to ensure rational management decision making processes.
Notwithstanding its contributions, this study has limitations.There are only nine commercial banks in Malaysia in the total population at the time of study, and we have included all nine banks data in this study; this sample remains too small to substantiate claims and may cause firm-specific bias.However the sample is restricted to Malaysian commercial banking industry because other industries have different regulation structures, cultural environments and other factors that may result in different board compositions and firm performances.Future research may undertake gender perspective in multiple directorships to measure firm performance within banking industry.Future study may consider both short-and long-term impact of corporate governance reforms on firm performance.
Note 2. The Asian Financial Crisis caused liquidity and insolvency problems for banking industry following excessive deficient collateral property lending and inattentive exposure to share-based lending, leaving them burdened with high Non-Performing Loans (NPLs).BNM enforced a merging scheme with the objective to enhance shareholder value by reinforcing the efficiency and profitability of the financial institutions vis-a-vis to minimize post-integration costs.BNM announced on 29 th July, 1999 that the existing 55 financial institutions (i.e., 20 commercial banks, 23 finance companies, and 12 merchant banks) were to form six-anchor banking groups by the end of September 1999 (Poon, 2008).These six-anchor banks were nominated by the central bank and their selection was politically driven (Mitchell & Joseph, 2010), and linked to allegations of cronyism (Chong, Liu, & Tan, 2006).Some objections arose pertaining to the composition of the groups.The government responded by giving flexibility to the banking institutions to form their own merger groups and the number of groups rose from six to ten, and the merging exercise deadline was extended till the end of January 2000.
Note 3. The guideline for banking sector was published by Basel Committee (1999;2006), and it lists the responsibility of BOD on operations and financial soundness of the banks.The independent director requirement was announced in 1998, and 1999 was the first year when such regulations were enforced.During the last decade, the government has introduced significant reforms in corporate governance, including the enactment of the revised Malaysian Code on Corporate Governance (MCCG) 2007 as the "Best Practice Guide" to promote transparency and encourage self-regulatory behavior among PLCs."Best practice" advocates that the board composition should have a majority of independent non-executive directors (MCCG Revised 2007, p.11).

Note 4 .
There are nine domestic commercial banks in Malaysia in the post-merged era.There are seven firms in 2000, eight firms from 2001-2002, and nine firms from 2003-2009.

Table 1 .
Summary statistics, annual data (2000Summary statistics, annual data ( -2009)   ) Notes: AGE is the average age of the director, EM is the total number of minority (Chinese and Indian, i.e., non-Bumiputra) directors on the board (%), QUA is Total number of directors with qualification either in Law, Business and Management, or Economics (%), POL is the number of political connected directors (%), ROE is the natural log of return to equity, EXP is Number of directors who have background experience in banking sector (%), FS is firm size, measured as the natural log of the total assets of the company, BS is board size, measured as the natural log of total number of directors on the board, FA is firm age, measured as the natural log of the number of years since its operation.The null hypothesis is that the series are non-stationary or contains a unit root.The rejection of null hypothesis for ADF test is based on the MacKinnon critical values.Lag length selection is based on automatic selection: SIC.Spectral estimation is based on Barlett Kernel, Bandwidth selection is based on automatic Newey-West.*, **, and *** denote the rejection of the null hypothesis of non-stationary at p<.10, p<.05, and p<.01 significance level.Probabilities for Fisher tests are computed using an asymptotic Chi-square distribution.All other tests assume asymptotic normality.

Table 4 .
Fixed effect regression estimates of the relationship between firm performance and board characteristic All regressions are estimated using panel least squares.Dependent variable: ROE, the natural log of return to equity.POL is the number of political connected directors (%), AGE is the average age of the director, EM is the total number of minority (Chinese and Indian, i.e., non-Bumiputra) directors on the board (%), QUA is Total number of directors with qualification either in Law, Business and Management, Economics, and etc. (%), EXP is Number of directors who have background experience in banking sector (%), FS is firm size, measured as the natural log of the total assets of the company, BS is board size, measured as the natural log of total number of directors on the board, FA is firm age, measured as the natural log of the number of years since its operation, and FY is the Fiscal year as indicator variable, 1 for fiscal year ended 31 December; 0 otherwise.Probability values are based on a t-statistic for a two-tailed test of significance.The first number of each cell is the regression coefficient and figures in the parentheses are standard errors.R 2 is the explanatory power of the regressor.*, ** and *** denotes statistically significance at p<10, p<.05 and p<.01 level, respectively.

Table 5 .
Mean estimates between political connection and period