The Impact of Board Diversity, CEO Characteristics, and Board Committees on Financial Performance in the Case of Romanian Companies

Abstract

This paper examines the impact of board diversity, CEO characteristics, and board committees on the financial performance of the companies listed on the Bucharest Stock Exchange (BSE). In order to test the influence of these characteristics, detailed data on more than 70 firms are collected by hand, for the 2016–2020 period, and comprehensive regression models are estimated. The findings show that there are positive effects of board diversity especially with regard to the independent board members. In terms of the board committees, the audit committee is found to have a favourable influence. The regression coefficients imply that a 10% increase in the share of independent board members would be associated with a 0.93% increase in ROE. Based on these findings, it can be argued that improving the corporate governance practices of the companies listed on the BSE would increase the performance and the value of these firms.

1. Introduction

Corporate governance has a great influence in determining the efficient management of businesses, with a focus on balancing and reconciling the interests of different stakeholders surrounding companies (Solomon 2020). Depending on the focus and perspective of these stakeholders, the specific dimensions of corporate governance can cover very broad topics and issues. In this context, the board characteristics and board diversity are investigated widely in the corporate governance literature in terms of their effectiveness in the monitoring capacity and in addressing agency problems (Rutherford and Buchholtz 2007; Jermias and Gani 2014). Similarly, some studies examine the effects of certain CEO characteristics such as CEO duality (i.e., CEO holding the position of the Chairperson as well), age, gender, and ethnicity (Manner 2010; Kaplan et al. 2012). Moreover, the literature also examines the effects of board committees such as the audit committee or risk committee. For example, Spira and Bender (2004) argue that “The establishment of board sub-committees has been strongly recommended as a suitable mechanism for improving corporate governance, by delegating specific tasks from the main board to a smaller group and harnessing the contribution of non-executive directors” (p. 489). Hence, committees can become another important corporate governance dimension. In addition to these board and CEO-related characteristics, corporate governance also includes other topics such as minority rights (Ginevri 2011), investor relations (Crifo et al. 2019), executive pay (Sarhan et al. 2019), and corporate social responsibility (Widiatmoko 2020). Within this broad context of corporate governance, a major question emerges: “Which provisions, among the many provisions firms have and outside observers follow, are the ones that play a key role in the link between corporate governance and firm value?” (Bebchuk et al. 2009, p. 783). Then, it is possible to focus on different dimensions of corporate governance and their effects on firm performance. The present paper focuses on three main dimensions of corporate governance identified in the literature and examines their effects on the performance of firms on the BSE. These areas are the board diversity (in terms of the shares of the non-executive, independent, women, and foreigner board members), CEO characteristics (e.g., CEO duality, gender, and ethnicity), and board committees. In this way, the paper provides a comprehensive account of the effects of these corporate governance dimensions on firm performance in Romania.

By examining the case of a transition country, i.e., Romania, with an evolving corporate governance structure, this paper contributes to the relevant literature in terms of conducting a comprehensive examination regarding the effects of various corporate governance characteristics on the financial performance of companies listed on the Bucharest Stock Exchange. There are some studies that examine similar topics on the effects of board characteristics and other corporate governance issues such as Vintila and Gherghina (2013) and Borlea et al. (2017). While these papers provide valuable information on the research topic, the current study aims to conduct a very comprehensive analysis by incorporating many dimensions of corporate governance based on a recent period of time, that is, 2016–2020. Therefore, the present paper greatly expands the existing literature. The results of the paper are also important for policy and managerial purposes in the sense that the findings of the papers (such as the positive performance effects of independent board members and audit committees) produce important policy recommendations—namely, these results indicate that improving corporate governance practices by having independent board members and audit committees would be favourable for the financial performance of companies in the Bucharest Stock Exchange. In return, the favourable performance of the stock market and high standards of corporate governance practices would be important factors supporting the stock market development, financial development, and economic development of the country.

The paper is organised as follows: The next section provides a review of the prior literature and puts the present paper in context. Then, the third section provides the details of the dataset and research methods used in the empirical analysis. The findings are presented in the fourth section, while the fifth section provides a discussion of limitations and future research. Finally, the last section concludes the paper.

2. Prior Literature

Corporate governance is a very broad topic that covers many different dimensions of business management and relations among different stakeholders. Becht et al. (2003) state that “Corporate governance is concerned with the resolution of collective action problems among dispersed investors and the reconciliation of conflicts of interest between various corporate claim holders” (p. 1). Hence, corporate governance is interested in the problems that arise from agency relations and collective actions surrounding companies. These issues are examined extensively in the literature in terms of both theoretical approaches and empirical analyses. One of the leading theoretical perspectives informing the corporate governance issues is the agency theory. In terms of agency problems, there can be several stages of agency issues within companies. For example, executives are agents of the shareholders or owners, which are the principal. Then, executives can pursue their own interests at the expense of shareholders, which would create principal–agent problems (Sappington 1991). This type of problem leads to the creation of corporate governance mechanisms that would try to address the agency problems and increase the monitoring capacity of shareholders. In this context, the creation of the board of directors, the presence of independent board members, and the establishment of various board committees, such as audit and risk committees, can be leading corporate governance mechanisms. At another level, the firm can be considered as an agent of the society, which would be the principal. Then, in the cases of social and environmental sustainability, the private interest and costs of the companies might conflict with the interests of society and the environment. In this case, another theoretical perspective, i.e., a stakeholder approach would be needed to address such issues surrounding companies. In the stakeholder theory, the firm would not only consider the effects of its actions in terms of profits but would also consider the effects on people and planet, or society and the environment (Freeman and Reed 1983; Freeman 2015). While from a narrow and short-term perspective, the stakeholder approach might look anti-competitive and hurt profitability, given that the social awareness and regulatory expectations on these societal and environmental issues increase, following a stakeholder approach can be favourable for survival, brand image, reputation, and profitability over the long term. Overall, these discussions show that various theoretical approaches such as the agency theory and stakeholder theory can be used to study corporate governance issues. In addition, these theories also produce some testable hypotheses about the relationship of board characteristics (such as independent board members and board diversity), CEO characteristics, and board committees. The relevant hypotheses are examined extensively by empirical studies, as discussed below.

Within the broad context of corporate governance, the present paper focuses on board diversity, board committees, and CEO characteristics. The relevant empirical literature shows that these points can matter for the monitoring effectiveness of boards, as well as the firm value and financial performance (Carter et al. 2003; Adams and Ferreira 2009; Knyazeva et al. 2013; Krause et al. 2014; Kolev et al. 2019). For example, Carter et al. (2003) examine the case of Fortune 1000 companies in terms of women and minority board members. Their empirical results indicate strong positive associations of gender and ethnic diversity with the financial performance of companies. Regarding this positive association, the authors note that board diversity would increase board independence, and in return, board independence would be a positive factor in terms of increasing board monitoring capacity and effectiveness. In their regression models, the authors use Tobin’s Q as the dependent variable. In the regression results, board size and CEO duality are found to be negatively associated with firm performance. In addition, the share of internal or executive board members is also negatively related to Tobin’s Q. Then, the presence of both women and minority members on boards is found to be positively associated with firm performance. Moreover, the variable of the average age of the board members does not have a statistically significant regression coefficient. This paper provides a useful regression framework, which is followed in the present paper as well. In another detailed study, Adams and Ferreira (2009) find that female board members are more active participants in board meetings and audit committees. However, the authors note that mandated quotes on female members can create negative effects on firm value. Knyazeva et al. (2013) to control the issue of endogeneity by using the local labour market conditions of the independent board members as an instrument, confirming their positive effects on firm performance. Overall, this literature shows that board diversity in terms of independent board members has positive effects on firm performance, while the effects of other diversity characteristics are mixed. Based on these findings, the first research hypothesis is postulated as follows:

Hypothesis 1.

Independent board members have positive effects on firm performance.

Another important corporate governance area is the issue of CEO duality, and there is a large body of literature examining this issue. The relevant results are generally mixed. Krause et al. (2014) note that the duality of CEO and Chairperson positions can be examined from different theoretical perspectives such as agency approach, stewardship approach, and managerial power approach. However, these theories do not provide a clear answer on the effects of CEO duality. The empirical studies also produce mixed effects. For example, Baliga et al. (1996) examine 375 companies from Fortune 500 covering the 1980–1991 period and do not find any effects of CEO duality. However, Ballinger and Marcel (2010) examine the case of 540 events for the S&P companies during the 1996–1998 period and find that the negative impacts of interim CEO changes are weakened by CEO duality. In another detailed empirical study, Krause and Semadeni (2013) study the case of 1053 companies from S&P 1500 and Fortune 1000 and show that the separation of CEO and Chairperson positions have adverse effects after the strong performance but positive impact after the weak performance. Hence, these findings do not find a clear effect of CEO duality on firm performance. Based on these results, the second research hypothesis is stated as follows:

Hypothesis 2.

CEO duality has no effects on firm performance.

The third corporate governance area that the study examines is the board committees. The board of directors is expected to conduct monitoring and supervisory tasks so that the actions of managers are in line with the interests of shareholders (Khan 2011; Pande and Ansari 2014; Alhossini et al. 2021). However, some of these tasks such as risk management, auditing, and remuneration can require more specific expertise. In this context, boards started to for committees to evaluate these dimensions of their companies. For example, Kolev et al. (2019) provide a detailed literature review and conclude that board committees, such as audit committees, can have favourable effects on firm performance. In another recent study, Lee (2020) examines the case of public US companies for the 2005–2015 period and finds that when independent board members are active in board committees, the firm performance measured by ROA improves. Based on this newly developing literature, the third research hypothesis is given as follows:

Hypothesis 3.

Audit committees have positive effects on firm performance.

It needs to be noted that more variables and more research hypotheses can be developed given the extensive nature of the literature on corporate governance. However, in order to have a focused scope, the present study is focused on the above three dimensions on board diversity (specifically independent board members), CEO duality, and board committees (specifically audit committees). In addition, as a transitioning country, Romania has been developing its corporate governance codes in line with the EU and OECD practices. For example, the 2015 Code of Corporate Governance document recommends the majority of the non-executive members be independent (BSE 2015). In addition, it suggests that the committees (such as the audit committee) not be chaired by the Chairperson of the board but by an independent member. Specifically, it states that “The Board should set up an audit committee, and at least one member should be an independent non-executive. The majority of members, including the chairman, should have proven an adequate qualification relevant to the functions and responsibilities of the committee … The audit committee should be chaired by an independent non-executive member” (BSE 2015, p. 6). Overall, it is seen that Romania is developing important governance codes on the above dimensions. Then, it becomes important to check whether these corporate governance factors produce similar effects in the case of Romania, a transitioning country.

The above topics on board diversity, board committees, and CEO characteristics are also examined in the context of the Bucharest Stock Exchange. For example, Vintila and Gherghina (2013) examine the effects of board independence and CEO duality. The authors collect firm-level data covering the 2007–2011 period and use Tobin’s Q as their dependent variable. Their results indicate that board independence has a negative and non-linear effect in the case of the OLS regression model, whereas there are no statistically significant effects in the case of the fixed-effects regression model. In another study, Vintila et al. (2015) conduct a more detailed study and find positive effects of board diversity. A more recent study by Borlea et al. (2017) also examines various board characteristics (such as independent board members and audit committees) and their effects on firm performance in the case of Romanian public companies. They note that these specific board characteristics can have positive effects on the financial performance of companies, as they improve the monitoring efficiency of boards and alleviate the corresponding agency problems. The authors use data only for 2012 and do not find any statistically significant results. The use of only one year in the sample and the relatively small sample size (i.e., only 55 observations) are possible factors in these weak empirical findings. There are also studies that examine similar topics in the case of central and eastern European (CEE) countries (Primecz et al. 2019). For example, Bistrova and Lace (2011) examine the case of CEE countries in terms of leading corporate governance dimensions, including independent board members and CEO duality, and find that there is a positive association with higher governance scores and better stock market performance. In another study, Firtescu and Terinte (2019) examine the case of firms from 11 CEE countries for the 2004–2013 period using the Orbis dataset and find that “independent internal audit committee … has a positive sign on firm’s profitability” measured by ROE and ROA (p. 114). Hence, these studies on CEE also provide quantitative evidence on the importance of corporate governance. The present paper improves over this relevant literature by conducting a more comprehensive empirical analysis covering the 2016–2020 period for the companies listed on the BSE.

3. Data and Research Methods

3.1. Sample Selection and Variable Description

The presents study focused on board diversity (the number of non-executives, independents, females, and foreigners), board committees, and CEO characteristics in Romania. The relevant variables are explained in

Table 1. Description of variables.

Variable	Definition	Relevant References
Return on equity	Net profits as a ratio to shareholder equity	Firtescu and Terinte (2019)
Return on assets	Net profits as a ratio to total assets	Firtescu and Terinte (2019)
Tobin’s Q	Market value as a ratio to total asset	Vintila and Gherghina (2013)
Board size	The number of members on the board	Vintila et al. (2015)
Board age	The average age of board members	Ferrero-Ferrero et al. (2015)
Firm size	Firm turnover or revenue	Vintila et al. (2015)
Non-executive share	The number of non-executive members as a ratio to board size	Vintila et al. (2015)
Independent share	The number of independent members as a ratio to board size	Vintila et al. (2015)
Women share	The number of women members as a ratio to board size	Vintila et al. (2015)
Foreign share	The number of foreign members as a ratio to board size	Masulis et al. (2012)
Board committees	The number of board committees	Vintila et al. (2015)
Committee members	The number of members in different board committees	Vintila et al. (2015)
Ceo age	The age of the CEO	Vintila et al. (2015)
CEO duality	Takes a value of 1 if the CEO has the dual duty of CEO and Chairperson	Vintila et al. (2015)
CEO women	Takes a value of 1 if women CEO	Vintila et al. (2015)
CEO foreign	Takes a value of 1 if foreign CEO

. The third column of the table provides some references that use the same variables in their analysis. Regarding the sample selection, data for more than 70 firms in the Bucharest Stock Exchange were collected by hand on these variables—namely, all the firms in the stock market were included in the analysis depending on data available in their annual reports. The sample period covered the last five years of 2016–2020. Hence, the sample size and period coverage are relatively large, to obtain robust findings on the recent corporate governance developments in the Bucharest Stock Exchange.

3.2. Econometric Specification

With reference to the research methods, the literature review section showed that regression models are commonly used to examine the effects of different corporate governance dimensions on firm performance. For example, Carter et al. (2003) use a regression model with the dependent variable of Tobin’s Q and independent variables of various diversity indicators and CEO characteristics. The present paper also follows a similar regression approach and estimated the following regression model:

$$\begin{gathered} RO{A}_{it}orRO{E}_{it}\mathrm{or}Tobi{n}^{\prime }s{Q}_{it}={\beta }_0+{\beta }_{1}Ln{(FirmSize)}_{it}+{\beta }_{2}BoardSiz{e}_{it}+ \\ {\beta }_{3}BoardDiversityMeasure{s}_{it}+{\beta }_{4}CEOCharacteristic{s}_{it}+ \\ {\beta }_{5}Board Committee{s}_{it}+{\epsilon }_{i,t} \end{gathered}$$

(1)

In the above regression model, i refers to the company and t refers to the year. The dependent variable was chosen among three performance indicators of ROA or ROE or Tobin’s Q. Then, Ln(firm size) and board size were utilised as the main control variables. The regression model also included three sets of independent variables corresponding to the three research hypotheses on the effects of board diversity measures, CEO characteristics, and audit committees. Overall, the paper provides a very rich set of regression models and empirical evidence. In terms of empirical strategy, the pooled OLS methods were utilised. Given the data limitations such as the relatively small number of cross sections (i.e., the number of companies) and low variation on some board characteristics, the panel data methods or dynamic methods were not utilised. These limitations can be addressed in future research as more data points become available.

4. Findings

4.1. Descriptive and Correlation Analyses

The board age is estimated at 52 years, with a standard deviation of 8 years. With reference to the diversity measures, the share of non-executive board members is 68%, while the share of independent board members is 38%, the share of female board members is 21%, and the share of foreign board members is 18%. Therefore, the lowest levels of board diversity are observed on foreigners and women, while the largest diversity level is observed in the non-executive members.

In terms of CEO characteristics,

Table 2. Summary statistics.

Descriptive Statistics—Board Characteristics
Variable	Obs	Mean	Std. Dev.	Min	Max
Board Size	345	4.971	1.73	1	11
Board Age	260	52.199	8.272	31.667	69.7
Executive Share	345	25.403	24.09	0	100
Non-Executive Share	345	67.585	29.602	0	100
Independent Share	295	38.423	30.427	0	100
Women Share	340	20.466	23.14	0	100
Foreigner Share	345	17.938	30.999	0	100
Descriptive Statistics—CEO Characteristics
Variable	Obs	Mean	Std. Dev.	Min	Max
CEO Age	245	52.531	11.465	30	71
CEO Duality	320	0.297	0.458	0	1
CEO Women	330	0.106	0.308	0	1
CEO Foreign	360	0.125	0.331	0	1
Descriptive Statistics—Board Committees
Variable	Obs	Mean	Std. Dev.	Min	Max
Number of Committees	305	1.656	1.277	0	5
Audit Members	200	2.75	0.776	1	5
Remuneration Members	110	3.136	0.818	2	5
Nomination Members	105	3.048	0.955	1	5
CSR Members	15	4.667	0.488	4	5
Stakeholder Members	5	3	0	3	3
Risk Members	20	3.25	0.444	3	4
Descriptive Statistics—Firm and Performance Variables
Variable	Obs	Mean	Std. Dev.	Min	Max
ROA	319	3.151	9.645	−44.49	77.05
ROE	320	7.662	14.711	−52.8	86.19
Tobins Q	119	0.677	0.664	0	3.94
Size	303	5.33 × 108	1.65 × 109	236,000	1.48 × 1010
ln_Size	303	18.282	1.97	12.37	23.41

shows that the average CEO age is 53 years, with a standard deviation of 12 years. 30% of CEOs in the sample hold the Chairperson position as well. In addition, 11% of CEOs are women and 13% are foreigners. Hence, the gender and ethnic diversity of CEOs is lower, compared with the board diversity, in both dimensions. Table 2 displays that the average number of committees is 1.66, with a standard deviation of 1.28. This variable ranges from 0 to 5 in the sample. The recent set of relevant data can be valuable, as there are no studies that examine the effect of board committees and their members on the firm performance of the companies listed on the BSE, to the best knowledge of the authors. In terms of specific board committees, the data collection process was able to recover information on the existence of an audit committee for 40 companies, so that there were 200 observations in five years of the sample. Then, the average member size in the audit committee is estimated as 2.75, with a standard deviation of 0.78. In the case of other board committees, the total number of observations is smaller, with 110 observations (i.e., for 22 firms) in the case of remuneration committees and 105 observations (i.e., for 21 firms) in the case of nomination committees. In the case of other board committees, the number of observations is smaller than 5 firms (or 20 observations), thereby making statistical analysis infeasible. The last panel of Table 2 presents the summary statistics for the performance indicators. It is found that the average ROA is 3.2%, while the average ROE is 7.7% and the average level of Tobin’s Q is 0.677.

Table 3. Correlation coefficients of firm performance.

Pairwise Correlations with Board Diversity Measures
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
(1) ROA	1.000
(2) ROE	0.405 *	1.000
(3) Tobin’s Q	−0.064	0.251 *	1.000
(4) Board Size	0.140 *	0.055	0.097	1.000
(5) Board Age	0.025	−0.024	−0.030	0.135 *	1.000
(6) Non_exec Share	−0.035	−0.13 *	0.079	0.261 *	−0.17 *	1.000
(7) Indep Share	−0.100	0.053	0.181	−0.039	−0.14 *	0.437 *	1.000
(8) Women Share	0.063	−0.074	−0.054	−0.14 *	−0.15 *	−0.16 *	−0.15 *	1.000
(9) Foreigner Share	−0.20 *	0.194 *	0.046	0.013	−0.24 *	0.124 *	0.069	−0.15 *	1.000
Pairwise Correlations with CEO Characteristics
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)
(1) ROA	1.000
(2) ROE	0.405 *	1.000
(3) Tobin’s Q	−0.064	0.251 *	1.000
(4) CEO Age	0.125	−0.049	0.028	1.000
(5) CEO Duality	0.023	−0.011	−0.040	0.487 *	1.000
(6) CEO Women	−0.034	−0.043	0.084	−0.104	−0.118 *	1.000
(7) CEO Foreigner	−0.192 *	0.111 *	−0.000	−0.063	0.131 *	−0.137 *	1.000
Pairwise Correlations with Board Committees
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)
(1) ROA	1.000
(2) ROE	0.405 *	1.000
(3) Tobin’s Q	−0.064	0.251 *	1.000
(4) Number of Committees	0.128 *	0.005	0.029	1.000
(5) Audit Member	0.083	−0.010	0.214	0.339 *	1.000
(6) Remuneration Member	0.431 *	0.225 *	0.427 *	−0.038	0.780 *	1.000
(7) Nomination Members	0.263 *	0.019	0.345 *	−0.057	0.691 *	1.000 *	1.000

* shows significance at the 0.05 level.

presents the correlations of the three sets of independent variables (i.e., board diversity measures, CEO characteristics, and board committees). In the case of the board diversity indicators, only one of the performance indicators, i.e., return on equity (ROE), has statistically significant correlations with some indicators—namely, ROA and Tobin’s Q do not display any statistically significant bivariate relationship with the diversity measures. However, in the case of ROE, the shares of both independent and foreign members are positively associated with this performance indicator at the 5% statistical significance level. In the case of the CEO characteristics, ROE is again positively associated with the presence of foreign CEOs. Finally, in the case of board committees, the number of board committees is not correlated with any of the performance indicators in a statistically significant way. However, ROA and Tobin’s Q display positive correlations with the number of members in different board committees. Overall, the correlation results in Table 3 provide some initial insights on the possible effects of board diversity in terms of independent and foreign board members and the different committees. Figure 1 and Figure 2 provide some graphical evidence in support of these findings.

4.2. Regression Analysis

This section presents the regression results. Since the number of independent variables is numerous (i.e., there are 14 independent variables in Table 2), including all of them at the same time can lead to large declines in the degrees of freedom in the regression estimations. In order to lessen this problem, the three sets of independent variables were separately included in the regression estimations. Then, a final regression model was estimated with the selected variables from each group.

Results for the board diversity measures in

Table 4. OLS regression results with the board diversity measures.

Linear Regression for ROA
ROA	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	1.128	0.256	4.40	0.000	0.623	1.633	***
board_size	0.330	0.360	0.92	0.360	−0.379	1.040
non_exec_share	0.022	0.015	1.53	0.127	−0.006	0.051
indep_share	−0.019	0.022	−0.87	0.387	−0.061	0.024
women_share	0.048	0.023	2.08	0.039	0.002	0.093	**
foreigner_share	−0.056	0.025	−2.26	0.025	−0.104	−0.007	**
Constant	−19.56	4.962	−3.94	0.000	−29.335	−9.780	***
Mean dependent var	3.572		SD dependent var		9.225
R-squared	0.144		Number of obs		233.000
F-test	7.047		Prob > F		0.000
Akaike crit. (AIC)	1673.493		Bayesian crit. (BIC)		1697.650
Linear Regression for ROE
ROE	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	1.054	0.494	2.13	0.034	0.081	2.027	**
board_size	0.633	0.621	1.02	0.309	−0.591	1.857
non_exec_share	−0.073	0.026	−2.79	0.006	−0.125	−0.022	***
indep_share	0.059	0.030	1.98	0.049	0.000	0.118	**
women_share	−0.014	0.030	−0.48	0.629	−0.073	0.044
foreigner_share	0.096	0.039	2.47	0.014	0.019	0.173	**
Constant	−13.438	8.025	−1.68	0.095	−29.249	2.374	*
Mean dependent var	7.875		SD dependent var		12.232
R-squared	0.120		Number of obs		236.000
F-test	4.165		Prob > F		0.001
Akaike crit. (AIC)	1834.358		Bayesian crit. (BIC)		1858.605
Linear Regression for Tobin’s Q
Tobin’s Q	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	0.048	0.031	1.57	0.120	−0.013	0.110
board_size	0.021	0.048	0.44	0.659	−0.074	0.116
non_exec_share	0.000	0.003	−0.07	0.945	−0.005	0.005
indep_share	0.005	0.003	1.51	0.135	−0.002	0.011
women_share	−0.001	0.003	−0.25	0.805	−0.008	0.006
foreigner_share	0.000	0.004	0.09	0.928	−0.007	0.007
Constant	−0.382	0.510	−0.75	0.456	−1.397	0.633
Mean dependent var	0.788		SD dependent var		0.734
R-squared	0.061		Number of obs		85.000
F-test	2.043		Prob > F		0.070
Akaike crit. (AIC)	196.343		Bayesian crit. (BIC)		213.441

*** p < 0.01, ** p < 0.05, * p < 0.1.

are presented in three panels. The upper panel shows the results for the performance variable of ROA, while the middle panel presents the results for ROE and the lower panel for Tobin’s Q. In the case of the ROA variable, the firm size has a positive and statistically significant coefficient, thereby indicating that larger firms perform better in this indicator. As additional results, it is found that the share of foreigners on boards is negatively associated with firm performance, whereas the share of women on boards is positively associated. The relevant regression coefficients are statistically significant at the 5% level. In the case of the middle panel with the performance variable of ROE, it is found that both independent share and foreign share are positively related, with a significance level of 5%. Finally, in the case of Tobin’s Q, the lower panel of Table 4 shows that no regression coefficient is statistically significant at the 10% level. When the results across these three performance indicators are compared, it is found that the upper panel with ROA has the highest R square value of 14.4%. Overall, these findings provide supportive evidence on the favourable effects of independent board members, thereby supporting the first research hypothesis. This finding is also consistent with the results in the literature about the positive performance effects of independent board members, such as Knyazeva et al. (2013) and Vintila and Gherghina (2013).

Table 5. OLS regression results with the CEO characteristics.

Linear Regression for ROA
ROA	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	1.063	0.312	3.40	0.001	0.448	1.679	***
board_size	0.257	0.386	0.67	0.507	−0.504	1.018
ceo_duality	1.578	1.168	1.35	0.178	−0.722	3.878
ceo_women	0.441	1.910	0.23	0.818	−3.322	4.204
ceo_foreign	−6.695	2.092	−3.20	0.002	−10.815	−2.58	***
Constant	−17.316	6.487	−2.67	0.008	−30.092	−4.54	***
Mean dependent var	3.304		SD dependent var		9.510
R-squared	0.116		Number of obs		254.000
F-test	7.225		Prob > F		0.000
Akaike crit. (AIC)	1844.753		Bayesian crit. (BIC)		1865.977
Linear Regression for ROE
ROE	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	0.775	0.646	1.20	0.231	−0.496	2.046
board_size	0.092	0.695	0.13	0.895	−1.276	1.460
ceo_duality	−0.132	1.903	−0.07	0.945	−3.880	3.615
ceo_women	0.210	2.580	0.08	0.935	−4.871	5.291
ceo_foreign	5.197	4.126	1.26	0.209	−2.930	13.324
Constant	−7.729	11.118	−0.69	0.488	−29.627	14.168
Mean dependent var	7.743		SD dependent var		14.224
R-squared	0.025		Number of obs		255.000
F-test	1.121		Prob > F		0.350
Akaike crit. (AIC)	2082.180		Bayesian crit. (BIC)		2103.428
Linear Regression for Tobin’s Q
Tobin’s Q	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	0.028	0.040	0.69	0.491	−0.052	0.108
board_size	0.038	0.047	0.82	0.414	−0.055	0.131
ceo_duality	−0.028	0.174	−0.16	0.874	−0.374	0.318
ceo_women	0.207	0.234	0.88	0.379	−0.258	0.672
ceo_foreign	0.015	0.303	0.05	0.961	−0.587	0.617
Constant	−0.004	0.809	−0.01	0.996	−1.611	1.604
Mean dependent var	0.719		SD dependent var		0.696
R-squared	0.027		Number of obs		96.000
F-test	0.609		Prob > F		0.693
Akaike crit. (AIC)	211.293		Bayesian crit. (BIC)		226.679

*** p < 0.01.

repeats the same set of regression estimations for the independent variables of CEO characteristics. In the case of the upper panel with the ROA indicator, the size variable is positively associated with this performance measure, whereas having a foreign CEO is negatively associated. In the case of the ROE and Tobin’s Q indicators, the results of the middle and lower panels indicate that none of the CEO characteristics is statistically significant at the 10% level. This is an interesting result, which might arise from the exclusion of important board characteristics in the regression model.

Table 6. OLS regression results with board committees.

Linear Regression for ROA
ROA	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	0.403	0.308	1.31	0.195	−0.212	1.019
number_committees	−0.827	0.819	−1.01	0.316	−2.464	0.809
audit_member	2.624	0.760	3.45	0.001	1.105	4.142	***
remuneration_members	1.229	0.343	3.58	0.001	0.543	1.915	***
Constant	−10.386	5.073	−2.05	0.045	−20.526	−0.247	**
Mean dependent var	6.373		SD dependent var		5.357
R-squared	0.415		Number of obs		67.000
F-test	10.940		Prob > F		0.000
Akaike crit. (AIC)	388.150		Bayesian crit. (BIC)		399.174
Linear Regression for ROE
ROE	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	1.139	0.559	2.04	0.046	0.021	2.256	**
number_committees	−0.520	1.389	−0.38	0.709	−3.297	2.256
audit_member	3.342	0.877	3.81	0.000	1.590	5.095	***
remuneration_members	0.161	0.561	0.29	0.775	−0.960	1.281
Constant	−21.963	6.867	−3.20	0.002	−35.686	−8.241	***
Mean dependent var	8.692		SD dependent var		8.053
R-squared	0.275		Number of obs		68.000
F-test	10.403		Prob > F		0.000
Akaike crit. (AIC)	463.839		Bayesian crit. (BIC)		474.936
Linear Regression for Tobin’s Q
Tobin’s Q	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	−0.030	0.045	−0.66	0.514	−0.122	0.063
number_committees	−0.029	0.073	−0.40	0.692	−0.180	0.121
audit_member	0.056	0.127	0.44	0.663	−0.205	0.316
remuneration_members	0.205	0.100	2.06	0.050	0.000	0.411	*
Constant	0.673	0.731	0.92	0.366	−0.832	2.179
Mean dependent var	0.809		SD dependent var		0.434
R-squared	0.215		Number of obs		30.000
F-test	2.131		Prob > F		0.107
Akaike crit. (AIC)	36.767		Bayesian crit. (BIC)		43.773

*** p < 0.01, ** p < 0.05, * p < 0.1.

presents the OLS regression results for the independent variables related to board committees. The number of committees is included as a relevant variable, along with the number of members in the audit and remuneration committees. Since there were not enough observations in the case of other committees, they were not included in the regression model. The results indicate that the audit committee has a positive and statistically significant effect on firm performance in the case of ROA and ROE, while the remuneration committee has a positive and statistically significant effect In the case of ROA and Tobin’s Q. These are novel findings in the literature for the Bucharest Stock Exchange. Overall, the results of Table 6 indicate that having audit committees with enough members would improve the financial performance of companies in Romania. Hence, these findings support the second research hypothesis on the positive performance effects of the audit committees. This favourable effect of audit committees is also consistent with the studies in the literature that find similar positive performance effects, such as Aldamen et al. (2012) and Ghafran and O’Sullivan (2013).

Finally,

Table 7. OLS regression results with full set of independent variables.

Linear Regression for ROE
ROE	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	0.857	0.609	1.41	0.161	−0.345	2.059
board_size	1.223	0.826	1.48	0.140	−0.406	2.853
non_exec_share	−0.170	0.059	−2.86	0.005	−0.287	−0.053	***
indep_share	0.093	0.048	1.96	0.052	−0.001	0.188	*
women_share	0.003	0.043	0.07	0.947	−0.082	0.088
foreigner_share	0.031	0.049	0.65	0.519	−0.064	0.127
ceo_duality	−4.661	2.776	−1.68	0.095	−10.138	0.816	*
ceo_women	−1.659	2.690	−0.62	0.538	−6.966	3.648
ceo_foreign	0.891	5.678	0.16	0.875	−10.311	12.094
number_committees	−1.589	1.018	−1.56	0.120	−3.599	0.420
Constant	−1.584	10.595	−0.15	0.881	−22.487	19.318
Mean dependent var	8.339		SD dependent var		12.260
R-squared	0.105		Number of obs		196.000
F-test	1.744		Prob > F		0.074
Akaike crit. (AIC)	1538.018		Bayesian crit. (BIC)		1574.077
Linear Regression for ROE
ROE	Coef.	St. Err.	t-Value	p-Value	[95% Conf	Interval]	Sig
ln_size	−0.386	0.766	−0.50	0.617	−1.925	1.153
board_size	−0.481	0.857	−0.56	0.577	−2.204	1.241
non_exec_share	0.122	0.087	1.41	0.166	−0.053	0.297
indep_share	−0.048	0.065	−0.75	0.457	−0.178	0.081
women_share	−0.049	0.102	−0.48	0.633	−0.254	0.156
foreigner_share	0.229	0.130	1.77	0.084	−0.032	0.490	*
ceo_duality	−0.413	2.640	−0.16	0.876	−5.722	4.896
ceo_women	−4.922	2.707	−1.82	0.075	−10.364	0.520	*
ceo_foreign	−6.773	9.209	−0.74	0.466	−25.289	11.742
number_committees	1.518	1.516	1.00	0.322	−1.530	4.567
audit_member	2.713	1.155	2.35	0.023	0.390	5.036	**
remuneration_members	0.126	1.600	0.08	0.938	−3.092	3.343
Constant	−1.134	17.627	−0.06	0.949	−36.576	34.308
Mean dependent var	8.898		SD dependent var		8.466
R-squared	0.351		Number of obs		61.000
F-test	12.793		Prob > F		0.000
Akaike crit. (AIC)	432.325		Bayesian crit. (BIC)		459.766

*** p < 0.01, ** p < 0.05, * p < 0.1.

shows the results of the full set of independent variables. To save on the degrees of freedom, the leading variables from the other regression models are included in this table. The upper panel has 196 observations, while this number falls to 61 when the audit and remuneration committee variables are included in the regression equation. In addition, this table is estimated only for the ROE variable. In the upper panel, it is found that the share of independent board members is positively related to the firm performance, whereas CEO duality is negatively related. The regression coefficient of 0.093 for the variable of independent board members implies that a 10% increase in the share of independent board members would be associated with a 0.93% increase in ROE. Hence, this effect is economically significant given that the average ROE was 3.2% in the sample. In the lower panel of Table 7, the results indicate that the audit committee has a positive effect on firm performance. Overall, these results also provide supportive evidence on the positive performance effects of independent board members and audit committees, thereby validating both research hypotheses.

The above findings are closely related to the theoretical and empirical studies in the literature. The favourable impacts of having independent members on boards and establishing audit committees are consistent with the agency theory and the resource-based view of the firm (Lockett et al. 2009; Panda and Leepsa 2017; Raimo et al. 2021). In addition, the positive effects of independent board members and audit committees in the case of Romania are consistent with the relevant literature that finds similar positive effects of both variables in other countries and Romania (Knyazeva et al. 2013; Vintila and Gherghina 2013). Similar findings on the effects of board independence and audit committees are also shown in the case of CEE countries by Bistrova and Lace (2011) and Firtescu and Terinte (2019). Hence, the results of the present paper provide supportive evidence about the importance of various corporate governance indicators in the case of Romania.

The findings of the present paper have also important managerial and policy implications. Regarding corporations, the implementation of best corporate governance practices such as the employment of independent board members and the establishment of audit committees. The shareholders can strive to implement these policies, as they would greatly benefit from these corporate governance indicators in the case of independent board members and audit committees. However, in the case of CEO duality, there is no robust evidence in terms of negative or positive effects. This finding is also consistent with the results in the literature. From a broad perspective, the results indicate that improving corporate governance practices in the Bucharest Stock Exchange would create positive performance effects for public companies. Hence, the 2015 Code of Corporate Governance by BSE makes useful suggestions about independent board members and audit committees. These recommendations can be strengthened and some specific corporate governance strategies can be made mandatory based on the findings of the present study. In return, higher performance would increase investor attention and inflows, thereby supporting stock market development, financial development, and economic development in the country.

5. Limitations and Future Research

This paper utilises a unique and valuable dataset on the joint-stock companies in Romania to see the possible effects of three corporate governance dimensions (board diversity in terms of independent board members, CEO duality, and audit committees) on financial performance. The paper also produces important quantitative results on the importance of corporate governance for Romania. Hence, it contributes to the relevant literature with evidence from a transitioning country. However, the study also suffers from some limitations that can be addressed in future research. As the data were collected by hand, the number of years and the number of control/independent variables were restricted to some extent. Expanding these dimensions would lead to a larger sample size, as well as more control variables such as liquidity and leverage that can affect firm performance. Moreover, stock performance indicators can also be used as additional dependent variables. Future research can make effort to address these data issues by collecting larger datasets on more variables. In addition to these data limitations, the use of only pooled OLS methods stands out as a methodological limitation. Normally, the data have a panel data nature, which would allow the implementation of various panel data methods, such as fixed effects and dynamic GMM estimations. However, the cross-sectional dimension in the dataset is relatively small, and there is not much variation in the board characteristics of the sample companies. Therefore, it is difficult to obtain strong results in the present sample with other estimation methods. This limitation also restricts the analysis in terms of addressing endogeneity issues. These methodological limitations can also be addressed in future research by using more data years and information on more board characteristics. Another possible extension in future research can be to conduct a cross-country analysis with other transition countries in the same region, in order to investigate if the findings from Romania could be generalised to them.

6. Conclusions

This paper examined the impact of board diversity, CEO characteristics, and board committees on the financial performance of the companies listed on the Bucharest Stock Exchange. The relevant literature on these corporate governance dimensions highlights the finding that board diversity and committees can play important roles in terms of increasing the effectiveness of board monitoring and improving firm value. In addition, the CEO duality might have negative effects on firm performance, as it can restrict the ability of the board to monitor and supervise the activities of executives. In order to test the relevance of these arguments for the case of the companies listed on the BSE, detailed data were collected by hand on these indicators for more than 70 firms covering the 2016–2020 period. Then, descriptive, correlation, and regression analyses were conducted to document the possible effects of board diversity, CEO characteristics, and board committees. The relevant results indicate that there are positive effects of board diversity in terms of independent members on boards. Regarding board committees, the audit committee is found to have positive performance effects. In terms of quantitative sizes, the regression coefficient of 0.093 for the independent board members in the case of the dependent variable of ROE implies that a 10% increase in the share of independent board members would be associated with a 0.93% increase in ROE. In conclusion, the economic effects of the empirical findings are also significant. Based on these findings, it can be argued that improving the corporate governance practices of the companies listed on the BSE would improve the valuation and performance of these firms. There are some research limitations in the present study that can be addressed in future research. The empirical analysis relies on the OLS regression models, while future research can implement more advanced regression methods such as panel data regressions and IV/GMM regressions (Wooldridge 2010). In addition, the dataset can be further improved by incorporating additional firm and board variables, as well as extending the sample period to earlier years.