A STRATEGIC CORPORATE GOVERNANCE FRAMEWORK FOR STATE-OWNED ENTERPRISES IN THE DEVELOPING ECONOMY

How to cite this paper: Kaunda, E., & Pelser, T. (2022). A strategic corporate governance framework for state-owned enterprises in the developing economy [Special issue]. Journal of Governance & Regulation, 11 (2), 257–276. shareholders’ power and multiple principals have a negative effect on performance. The study recommended changes to legal form, board operations, and disclosure to improve corporate governance effectiveness. The study has contributed to a body of knowledge in terms of developing a strategic governance framework for SOEs in Malawi. The study has also established that cultural values influence the effectiveness of corporate governance.


INTRODUCTION
The creation of state-owned enterprises (SOEs) in Malawi, like in many other countries, was seen as a strategy of enhancing economic growth (Robinett, 2006;Stambuli, 2002). The Organisation of Economic Cooperation and Development (OECD) defines stateowned enterprises as corporate entities recognised by national law owned by the state or in which the state exercises ownership created for economic purposes. Some of these may include -joint-stock companies, limited liability companies and partnerships limited by shares‖ (OECD, 2015, p. 14). These enterprises may be -wholly owned enterprises or those with minority state ownership‖ (Robinett & Fremond, 2007, p. 1).
In many developing countries, SOEs were meant to boost economic development through industrialisation (Robinett, 2006). Contrary to popular belief that SOEs are an engine of economic growth, many of these companies have performed poorly worldwide, prompting structural reforms to improve their performance (Vagliasindi, 2008). Some of these reforms focused on the change of ownership to the private sector. Private sector companies were regarded to be more efficient than the SOEs. However, with corporate failures associated with private sectors companies worldwide in the past decades, the focus of performance improvement in SOEs has now shifted to reforms while maintaining public ownership (Vagliasindi, 2008).
SOEs in Malawi have undergone several reforms over the years to improve financial performance. These reforms have led to commercialisation as well as privatisation of many SOEs (World Bank, 2003). Reforms have not been a panacea to the performance problems experienced by SOEs. Poor performance has also been attributed to poor corporate governance. It is becoming clear that the performance of these SOEs cannot improve without reforming their corporate governance (World Bank, 2007). Corporate governance (CG), which is defined as a -system by which companies are directed and controlled‖ (The Committee on the Financial Aspects of Corporate Governance, 1992, p. 15), has attracted unprecedented attention following recent corporate failures.
Many reasons have been advanced for the dismal performance of these organisations worldwide, which include, but are not limited to, political interference and poor corporate governance systems. At the centre of this poor performance is a corporate governance framework which, to a large extent, has been regarded as ineffective; characterised by weak boards, lack of disclosure, multiple principals and objectives coupled with unclear ownership policy and -absent fiat‖ from the principals.
SOEs continue to play an important role in Malawi but these organisations' performance cannot improve without incorporating a good corporate governance system. To address SOEs' performance problems, the study seeks to answer the following research questions: RQ1: What factors influence the effectiveness of corporate governance of SOEs in Malawi? RQ2: What has been the impact of corporate governance on the performance of SOEs in Malawi? RQ3: What corporate governance framework is most suitable to address SOEs' performance in Malawi?
Based on study results, researchers developed a strategic corporate governance framework to enhance the performance of SOEs in Malawi. The remainder of the paper is structured as follows. Section 2 discusses the literature review. Section 3 presents the research methodology. Section 4 presents the results. Section 5 proposes a strategic corporate governance framework for SOEs. Section 6 presents the conclusion.

LITERATURE REVIEW
This section analyses the impact of socio-cultural values on corporate governance, theories that form the background to the study, and the influence of corporate governance on SOEs' performance.

Culture and corporate governance
Corporate scandals that have rocked the world in the past two decades have heightened the importance of corporate governance (Zalewska, 2014). However, despite the overwhelming interest, the level of application and adoption of corporate governance codes by different countries has cast doubt on the universality as well as their effectiveness. Different governance frameworks have emerged with marked variations. These variations have been attributed to differences in national culture (Li & Harrison, 2008). There is growing interest to study the effect of culture on corporate governance. The challenge of cultural studies arises from the varied definitions and disciplines where culture originates. Studies on culture and corporate governance have used cultural dimensions of Hofstede's (Franke & Nadler, 2008;Bae, Chang, & Kang, 2012) and Schwartz's data (Desender, Castro, & De Leon, 2011;Licht, Goldschmidt, & Schwartz, 2005). However, Hofstede's work has been considered to be the most popular (Bae et al., 2012). This section reviews some empirical work on culture and corporate governance that has brought some mixed results.
A study by Chan and Cheung (2012) across different countries revealed that high individualism (IDV), low uncertainty avoidance index (UAI), low masculinity (MAS), and three control variables: log (GDP per capita), common-law, and market-to-book ratio had higher CG scores. Similar results were obtained in another study conducted by Griffin, Guedhami, Li, Kwok, and Shao (2018), who found that high IDV and low UAI were related to better company-level good governance practices. In this study, the authors used the Governance Metric International index as a measure of good corporate governance practice. In an earlier study, Li and Harrison (2008) examined the influence of ownership structure and national culture on corporate governance. The study found that high power distance index (PDI), high IDV, high MAS tend to prefer consolidated leadership or CEO duality. The study also found that high IDV was related to a small board.
In addition to cultural dimensions, other studies have also found that religion has an influence on corporate governance practices. Studies by Kim and Daniel (2016) found that religiosity was related to better governance practices. Religion as a construct has been studied widely but as observed by Nadler (2002), its effect on human behaviour is yet to be conclusively established.
The influence of religion on an individual's behaviour is a result of one's desire to comply with the social norms of a particular group (Kim & Daniel, 2016). Compliance comes as a result of fear of being ostracised by the society or the group to which an individual is affiliated (Rashid & Ibrahim, 2008). Religion defines the values and culture of individuals in a particular community (Nadler, 2002). Johnstone (as cited in Kum-Lung & Teck-Chai, 2010) defines religion as -a system of beliefs and practices on how people respond and interpret what they feel is supernatural and sacred‖ (p. 226). Baxamusa and Jalal (2014) observe that -religion consists of beliefs, values and behaviour‖ (p. 114). In relation to culture, religion is considered to be one of the most important proxies of culture (Stulz & Williamson, 2003).
While religion has been defined as a set of beliefs and values, religiosity, on the other hand, is defined as an individual's -commitment to follow principles set by God‖ (Vitell, 2009, p. 156). According to Kum-Lung and Teck-Chai (2010, p. 226), religiosity defines the degree of one's commitment to the religion and teachings that one professes. This commitment is reflected by the individual's attitudes and behaviours.
There has been a growing interest to study religiosity and its effect on human behaviour. Most of these studies have been carried out in developed countries and emerging economies. Studies on religiosity and corporate governance practices have been influenced by agency theory. Research conducted by Stulz and Williamson (2003) found that more religious societies offer stronger protection to creditors. Kum-Lung and Teck-Chai (2010) observed religiosity positively related to business ethics. This is because highly religious people are expected to be more ethical in their behaviour which is defined by their beliefs. These findings are consistent with the findings of Donahue (1985). On the contrary, Rashid and Ibrahim (2008) argued that high religiosity does not in itself translate to high ethical values. This is due to other cultural variables. Kim and Daniel (2016) observed that while agency-principal conflict is associated with agency problems in developed markets, this conflict was also prevalent in emerging markets. Developed markets were also associated with protestant religion and formal institutions. In emerging markets, the prevalence of informal institutions has resulted in weak corporate governance practices. These markets are also characterised by a lack of regulatory environment and enforcement of laws where laws exist. Where there is a lack of formal institutions, Boytsun et al. (2011) propose that informal constraints such as religion can work as an alternative mechanism to improve corporate governance practices. While

Agency theory, corporate governance, and performance
Studies on culture, religiosity and corporate governance have been influenced by agency theory. This section critically analyses agency theory in relation to corporate governance. Agency theory is defined as a contractual relationship between the principal and the agent (Perrow, 1986). The principal, in this case, is the employer or shareholder who has ownership rights of the property and the agent is the employee or manager who is entrusted with the responsibility of taking care of an enterprise on behalf of the principal (Donaldson & Davis, 1991). According to Perrow (1986), agency theory is based on three assumptions. Firstly, -individuals tend to maximise their own interest‖. Secondly, -social life is a series of contracts that are governed by competitive self-interests‖ and lastly, -that monitoring of contracts is costly and ineffective‖ (p. 12).
Agency theory finds its origin in the theory of the firm by Jensen and Meckling (1976). According to Jensen and Meckling (1976), organisations are considered as -legal fictions which serve as a set of contracting relationship among individuals‖ (p. 8). The contracting individuals are the principals (owners) on the one hand and the agents (managers/employees) on the other hand. Where there is a separation of ownership and control, individuals identified above have divergent interests. In order to protect their interests, the principals incur agency costs to monitor contracts entered with agents. Some of the agency costs that the principal incurs in the contractual relationship include -monitoring expenditures by the principal; bonding expenditures by the agent; and the residual loss‖ (Jensen & Meckling, 1976, p. 6).
Consistent with agency theory, Liu, Miletkov, Wei, and Yang (2015) found that the appointment of independent directors had a positive impact on company performance. This was one of the comprehensive studies covering the role of independent directors in performance improvement. While the study covered more variables of board composition and board structure like board size, CEO duality, and board meetings, not all board composition variables were included. In addition to variables covered by Liu et al. (2015), the current study also discussed the effect of director interlocking, the appointment of public servants, board committees, and board evaluation on SOEs performance which was not covered by their study. Similar results were obtained by Bhat, Chen, Jebran, and Bhutto (2018) on the Pakistani state and nonstate-owned companies. Thenmozhi and Sasidharan (2020) found that independent directors in Chinese and Indian SOEs' boards improve performance through better monitoring mechanisms. Contrary to these results, a study by Shao (2019) found no relationship between independent directors and company performance. In addition, the study found a positive relationship between ownership concentration and performance.
In a study on Canadian SOEs, Bozec (2005) found that board independence is negatively related to company performance where SOEs are subjected to competition. The study also found a negative relationship between board size and CEO duality on performance. While good corporate governance codes recommend that civil servants should not serve on SOEs boards, this study found that there is a positive relationship between the involvement of public servants on the SOE board and performance. On directors affiliated with a political party, a study by Heo (2018) on Korean SOEs found that political appointee directors are negatively correlated with SOE performance. Dragomir, Dumitru, and Feleagă (2021) attributed the poor performance of a Roman Airline to a lack of independence by directors due to political interference. While agency theory advocates the majority of non-executive directors (NEDs) in boards, a study by Kakabadse, Yang, and Sanders (2010) found that NEDs were not effective monitors of SOEs in China owing to a lack of quality information to discharge their monitoring responsibilities.
Director's experience, tenure, interlocking, and functional background have attracted governance scholarship interest. A review of the literature shows that much of this work had been conducted in private sector companies. Literature on SOE corporate governance for these board attributes is still in infancy and inconclusive. Kim, Mauldin, and Patro (2014) found that longer tenure for outside directors is useful for their role of advising and monitoring. This was supported by a study conducted by Kuzman, Talavera, and Bellos (2018) on Yugoslavia's SOEs, who found that long board tenure is positively related to SOEs' performance. However, Chamberlain (2010) noted the value of tenure diminishes as the director serves for a much longer period. On director's interlocking, McIntyre, Murphy, and Mitchell (2007) found director's business is associated with a decrease in company performance. This was supported by Mishra and Kapil (2018) in their study on Indian companies, who observed that directors' interlocking has a negative effect on company performance. While good corporate governance practice advocates that directors should have the necessary professional background, empirical evidence cast doubt on the impact of professional background on company performance (Zandstra, 2002).
Agency theorists posit that the use of debt as a financing source is also treated as a governance control mechanism. Empirical evidence on the effect of capital structure on company performance has raised a number of enduring debates on the role of debt. A study on Pakistan companies by Ahmed Sheikh and Wang (2013) found that capital structure was negatively related to company performance. The authors noted that the increase of debt in the capital structure had a negative influence on company performance. The study was conducted in Pakistan, where the market for corporate control is neither developed nor efficient. The results of this study are supported by the study conducted by Le and Phan (2017) on Vietnamese listed companies. A study by Dawar (2014) in India found that capital structure significantly negatively influences company performance. Findings by these studies are contrary to agency theory which posits that debt can be used as a disciplinary device on managerial behaviour. The authors noted that environmental factors may have influenced the findings of the study. For instance, the studies cited above were conducted in India, Pakistan, and Vietnam, which are emerging markets. Pakistan and Indian financial markets are underdeveloped and dominated by SOEs financial institutions, which are considered ineffective control devices to limit managerial -discretionary behaviour‖.
Other studies on capital structure, however, are in support of agency theory as a mitigating factor in the conflict between managers and shareholders. Kyereboah-Coleman (2007) found that high leverage has a positive effect on the performance of microfinance companies in Ghana. Most of the studies on capital structure have been conducted in developed and emerging markets. The current study was be conducted in Malawi, an least developed country (LDC) where literature is not available on the effect of capital structure and company performance.
Disclosure is as important to privately-owned companies as it is to publicly-owned companies. Disclosure promotes transparency which in turn ensures accountability. Robinett (2006, p. 19) has grouped disclosure into two categories: ex-ante reporting and ex-post reporting. These disclosure categories have been used in formulating Disclosure Index scores. Empirical evidence on the relationship between disclosure and financial performance reveals that disclosure positively influences financial performance. In a study on Indian companies listed on the Securities Board of India (SEBI), Assankutty, Fatima, and Kuntluru (2019, p. 10) found that the Corporate Governance Index (CGI) was related to company performance measured by Tobin Q, but the study found a negative relationship between CGI and financial performance measured by accounting measure of ROE. The study used a self-constructed CGI using disclosure made in annual reports. However, the study did not account for the influence of other corporate governance variables on disclosure which would impact company valuation. Consistent with Assankutty et al. (2019, p. 10), the current study investigated the impact of disclosure on performance. Since the unit of analysis is SOEs, which inherently face challenges of availability of CGI scores, the study used a selfconstructed governance index developed from Malawi Code II of best practice of corporate governance.

RESEARCH METHODOLOGY
The current research has used critical realism applied multi-methodology. Critical realism does not accept the pure deductive method of positivists and the pure inductive method of interpretivism. As a pluralist paradigm, critical realism is also highly pluralistic in terms of research methods (Miller & Tsang, 2011). The paradigm uses multimethodology and mixed design in its methodology (Mingers, 2006). This research which is based on critical realism combined quantitative and qualitative methods.
A three-phased approach was used to conduct a mixed-method retroductive research design. These phases include the appreciation phase that predominately uses the quantitative method, the retroduction phase, which involves the qualitative method, and the assessment phase, which includes both qualitative and quantitative inferences. This study has adopted the same retroductive research design, guided by the critical realism approach. This study was guided by the conceptual framework presented in Figure 1.  The religiosity questionnaire had ten items. The researcher performed a Cronbach's alpha to test the reliability of the measurement instrument. The results of the Cronbach's alpha test (see Table 1) revealed that the instrument used in the first research question (RQ1) is highly reliable thereby passing the reliability test with a test result of 0.90.
The second research question (RQ2) is aimed at investigating the effect of corporate governance variables on SOEs' performance. This study has identified the following corporate governance variables: legal form, board attributes, capital structure, and disclosure.
To investigate the effect of corporate governance variables on performance, the study developed the following hypothesis: H2: Good corporate governance practices have a positive influence on SOEs' performance.
The study used multiple regression analysis to test the above hypothesis. The following model was used: (1) Stepwise regression was applied to the entire model both as a dimensionality reduction technique as well as to eliminate multicollinearity in the independent variables. A final model was selected that best explains the impact between governance variables and SOEs' performance. The final model assumed a dynamic relationship between corporate governance variables and SOEs' performance, consistent with Nguyen, Locke, and Reddy (2014). To identify mechanisms and structures that caused the observed performance, an intensive design was undertaken on a selected case and this was achieved through the use qualitative case study.
The sample frame for the extensive design included all commercial SOEs registered with the Registrar of Companies and established through the Acts of Parliament, which were active between 2000 and 2016. However, regulatory, training, and commercial, financial institutions were excluded. The final sample included 13 listed and non-listed commercial SOEs. Data used for the period from 2000 to 2016 was collected from annual reports and other organisational documents. Out of the 13 SOEs that were active between 2000 and 2013, only 9 SOEs had complete data. In a study that has a small population, Easton (2010) argues that cases studies with a suitable small number of entities and where the objective of the study is to seek causality are available. Yin (2009) advocates the use of case study where the objective of the study is to investigate -a contemporary phenomenon in depth and within its real-life context, especially when the boundaries between phenomenon and context are not clearly evident‖ (p. 18).
The current study used a purposive sampling technique to select cases from all non-financial commercial SOEs for the period from 2000 to 2016. Yin (2009) recommends the use of a multiple-case design against a single-case design. The multiplecase design can enhance external validity through replication logic (Yin, 2009, p. 41). For the purpose of the current study, four cases were chosen from a population of 13 for a detailed analysis. These cases were selected based on performance, complexity, diversity, as well as their relevance to the phenomenon under study, and a criterion that is in line with Stake (2006, p. 23).
The study used quantitative data analysis to analyse the relationship between corporate governance variables and performance. In addition, the study analysed the relationship between sociocultural variables and the quality of corporate governance. The quantitative method employed a longitudinal panel study. The use of panel data analysis using fixed effects and random effects estimation was applied for initial regression analysis. Hausman specification test was carried out to check the suitability of random effects model consistent with the study by Ibrahimy and Ahmad (2012). Fixed effects estimation has its weaknesses as it suffers from biases one of which is the assumption that current observations are not dependent upon past events (Wintoki, Linck, & Netter, 2012). To minimize this bias, the study considered the use of a dynamic generalised method of moments (GMM) estimation (Wintoki et al., 2012). This was achieved by running Bover/Blundell-Bond estimation (Blundell & Bond, 1998) in STATA 13. To test the suitability of GMM, the study conducted a Durbin-Wu-Hausman (DWH) test that is used to check for the presence of endogeneity among regressors, this is in line with the study by Nguyen et al. (2014).
The second stage involved qualitative analysis using the critical realism case study method. Collected data has been categorised according to themes or research questions. Codes were assigned to data for ease of reduction. The critical realism data analysis used retroduction to identify structures and generative mechanisms which is a similar approach used by Miller and Tsang (2011).

RESULTS
To identify factors that influence the effectiveness of corporate governance of SOEs in Malawi, a model was developed from a hypothesis (H1) meant to test the influence of religiosity on corporate governance. A model was developed as shown below: (2) A structured questionnaire was administered to all 13 SOEs to obtain scores for the quality of governance. Another structured questionnaire was administered to obtained data for religiosity. Only 9 SOEs responded with complete data. These 9 SOEs represented all major industries which include tourism, water utilities, telecommunications, energy, and property development.   Table 3 presents the Pearson correlation matrix between the quality of corporate governance score as the dependent variable and religiosity. The results show a weak and insignificant negative relationship between religiosity and governance score. However, governance score is positively related to legal form, leverage, and company size but the relationship is weak and insignificant.  Table 4 presents results from regression analysis for factors that influence the quality of corporate governance measured by governance score. Four groups of corporate governance variables were identified as legal form, board attributes, capital structure, and disclosure and transparency. Two SOEs' performance measures were identified as dependent variables and these are EBIT and ROA. Data was collected from 9 SOEs through document review of annual reports, interviews using a semistructured questionnaire which was used as a guide.

Factors impacting corporate governance
Stepwise OLS was applied to the initial model to reduce the number of independent variables in order to establish those that have better explanatory power. The other reason for applying stepwise was to avoid multicollinearity. This is consistent with the study conducted by Hussain, Rigoni, and Orij (2018). After applying stepwise, the following model was developed: (3) Some independent variables were dropped from the initial models due to various reasons which include multicollinearity, similar scores for all SOEs, and unavailability of data.  On ownership structure, results show that legal form (LF) had an average of 0.28 with a standard deviation of 0.45. Over 80% of the sampled SOEs are established under the Acts of Parliament and are wholly owned by the government.
The current study found that on average the board size (BSize) is 11.72 and its standard deviation is 2.95 revealing that most of the boards of directors are close to the average. The maximum board size is 19 while the minimum size is at 5. While the minimum and average is within the recommended board size, the results show that 36.8% are between 13 and 19, which is above the recommended. Overall, the results show compliance with good corporate governance principles in terms of board size. Results also show that an average of 57% of board members are affiliated with governing political parties, a standard deviation of 14%. The minimum is 20% and the maximum 80%. The results show that the board representation is highly politicised with over 65% of the board above average. A politicised board leads to political interference and consequently to poor corporate governance practice. On boards with risk management committees (Risk) in their board, on average 10% of the boards have risk management committees with a standard deviation of 30%, and 90% of the board did not have a risk management committee. Results show that meetings were held as per requirements. On average 4.04 board meetings were held per SOEs per annum, with the maximum and minimum of 7 and 4, respectively. The standard deviation for the frequency of board meetings (Freq) is 0.31. About 98% of the meetings were held as per the requirement that stipulates 4 sittings per year. Board tenure (Tenure) was averaged at 2.15 years with a standard deviation of 1.37 while the maximum was 5 years. The results show that the boards had a short tenure with about 68% having tenure of 2 years or less. Civil (Civil) represents a percentage of the board with the representation of directors from civil service. The results show that an average of 33% of the board were appointed from the civil service with a standard deviation of 0.7% while the maximum and minimum were 58% and 20%, respectively.
On annual reports (AnRpt), the results show that on average 36% of the sampled SOEs prepared their annual reports with a standard deviation of 48%. Approximately 64% of the period, annual reports were not prepared by some SOEs. On third party disclosure (Third), results show that on average 14% of the SOEs had third party disclosure with a standard deviation of 35%. Results show that 86% did not have third-party disclosure in their annual reports or financial statements. On disclosure of conflict of interest (Conflict), results show that an average of 11% had disclosure of conflict of interest with a standard deviation of 31%. Approximately 89% of the annual and financial reports had no disclosure of conflict of interest. Quality of disclosure is a measure of compliance and accountability. As shown from the results above, most of the SOEs do not comply with corporate governance principles because of a lack of disclosure. On the capital structure, results show that leverage (Lev) had an average of 1.73 with a standard deviation of 7.89 while the maximum and minimum were 65.99 and -35.19, respectively. The results show that the SOEs are highly leveraged.
Correlation coefficients of key variables for 133 observations of the study for the period from 2000 to 2016 for 9 SOEs were calculated. Accounting performance measures of EBIT and ROA have a weak positive correlation of r = 0.48 indicating that these performance measures cannot be used interchangeably. Results show that EBIT has a significant positive correlation at p-value of < 0.05 with legal form, risk management committee, annual performance, and company size. However, EBIT was significantly and negatively correlated with political party affiliation. Regarding ROA, results show that performance is significantly and positively correlated at p-value of < 0.05 with independent variables of legal form, risk management committee, tenure, annual report, third party disclosure, conflict of interest, disclosure leverage, and company size. In addition, ROA is also negatively related to political affiliation and civil servants' presence in the board at r = -0.25 (p = 0.005). Correlations results for both EBIT and ROA are in support of H2 that corporate governance practices have a positive influence on SOEs' performance.
The OLS results in Table 6 show that there is a significant relationship between both EBIT and ROA and the independent variables. However, before a further analysis could be performed, it was noted that two of the independent variables of corporate governance namely, LF and Conflict had very high variance inflation factor (VIF) indicating that they were correlated with other corporate governance variables. These variables were then dropped, and a further regression test was conducted to check multicollinearity on the remaining corporate governance variables. Below is the model that was adopted after controlling for multicollinearity: (4) Below are the results for regression analysis for corporate governance and SOEs' performance using OLS, fixed effects, random effects, and GMM estimations. Table 7 presents OLS regression analysis for corporate governance and SOEs' performance.  Table 8 for EBIT and ROA of the fixed effects model revealed a significant relationship between corporate governance and SOEs' performance at p-value of less than 5% level. To test for the appropriateness of the fixed effects estimation in the regression, the Hausman specification test was performed. The test was conducted to find out if there is a correlation between unique errors and regressors. The p-value of less than 0.05 would mean that the null hypothesis would have been rejected thereby accepting fixed effects estimation as the most appropriate model. Hausman specifications tests in Table 9 show that the p-value for both EBIT and ROA are 0.1767 and 0.72, respectively. These p-values are above the threshold of 0.05; therefore, the study accepted the null hypothesis in both cases. Random effects estimation was found to be the most appropriate estimation. One of the endogeneity problems not addressed by fixed effects and random effects estimations is the one that arises from the effect of past actions on current performance. To address this bias, the study tested for the presence of endogeneity in all regressors as per prior studies (Nguyen et Table 10 and Table 11. The DWH results for EBIT in Table 10 fails to reject the null hypothesis for EBIT indicating that variables are exogenous therefore traditional static models of OLS, fixed effects, and random effects are efficient and consistent. Results for ROA presented in Table 11 show p-values of less than 0.05. In the case of ROA, the study shows that there is a concern of endogeneity therefore a dynamic model had to be applied.   The final results for the EBIT accounting measure are presented in Table 12. The regression results show that EBIT is positively and significantly associated with the risk management committee in support of H2 of our study but is negatively and significantly associated with third-party disclosure which is contrary to H2 of the study. On the control variables, EBIT is negatively and significantly associated with industry and competition at a 5% level. The rest of the hypotheses are not supported by the results for the accounting measure of EBIT using both static models of random effects and OLS. 0.000 0.000 Notes: Table 12 presents results of regression analysis used to measure the relationship between corporate governance and SOEs' performance using OLS and random effects estimations. Both models reveal that the independent variables have a significant influence on performance at p-value of less than 5%. Performance in this model has used an accounting measure EBIT. The results obtained for the DWH endogeneity test for ROA in Table 11 reveal that variables are endogenous therefore traditional static models of OLS and random effects were considered biased. In order to address the problem of endogeneity -due to unobserved heterogeneity, simultaneity and reverse causality‖ (Shao, Table 13 presents results for Sargan and Basmann test. The p-value is above 0.05 indicating that the instruments used in the model are valid and consistent with a system-GMM model.  The GMM results in Table 14 show that PAf is negatively and significantly associated with ROA at 10% level supporting the assertion that political affiliation of the directors leads to poor corporate governance practice. The findings of this study are also consistent with the results obtained by Cong Phuong, Dinh Khoi Nguyen, and Phuoc Vu (2020) on Vietnamese state-owned companies. Their study revealed that the presence of politicians on the board increases the degree of political interference in the SOEs. Results also show that Risk is positively and significantly related to ROA. The inclusion of risk management committee in the board leads to an effective board structure. Board structure is one of the elements of an effective corporate governance framework. The results of Risk support H2 good corporate governance practices have a positive influence on SOEs' performance. Civil is negatively and significantly associated with ROA at 5% for OLS and random effects models but the level of significance changes to 10% when dynamic models are employed. The results show that the inclusion of civil servants in the board affects the effectiveness of corporate governance and consequently has a negative effect on performance. Third-party disclosure (Third) is negatively and significantly related to ROA at a 5% level for the GMM model contrary to H2. Third-party disclosure is meant to be a building block of effective corporate governance practice. Results obtained in the study may mean that third-party disclosure is done for the sake of legitimation purposes.
On the capital structure, results show that Lev is positively and significantly related to ROA at a 5% a level in all models in Table 14 supporting a postulation that capital structure leads to better performance because of the control that this external governance mechanism exerts on agents. Control variables of industry and competition are negatively related to performance, but age is negatively related to ROA at a 5% level on OLS and random effects models. However, these control variables do not have any influence when dynamic models are applied suggesting that the relationship may be as a result of the spurious correlation between corporate governance and ROA when static models of OLS and random effects are applied and which consequently disappear after employing dynamic models.
Results from both static and dynamic models confirm that there is a significant relationship between corporate governance and performance. Correlation results confirmed that there is a significant correlation between effective corporate governance practice and company performance. On regression, EBIT and corporate governance had a weak significant relationship than ROA, which used a dynamic model. To understand the causes of performance observed during the quantitative data analysis, the study employed qualitative analysis consistent with the critical realism research paradigm. The next section presents results from the qualitative data analysis.

Qualitative data analysis
In line with the critical case study design, the second part of the study had to do with the qualitative part of the study to identify governance structures and mechanisms that emerged from various themes. Results were obtained through document reviews and personal interviews. Results from the interviews are summarised in Table 15 and were applied to four cases using the replication model. Table 15 presents the frequency of themes and governance mechanisms summarised from interviews collected from 9 SOEs. The results in Table 15 reveal that corporate governance influences corporate performance. Effective legal form and board of governance have a positive influence on performance while captured board, soft budget constraints, lack of disclosure, and socio-cultural values have a negative effect on performance. After applying the above mechanism to selected cases, the following results were obtained. Table 16 presents a list of selected cases chosen based on their performance, and Table 17 presents summarised results from case analysis following replication logic.  There was a lack of disclosure but there was no relationship between disclosure and performance.
There was a lack of disclosure as stipulated by the act but there is no relationship between disclosure and performance.

Cronyism
No evidence of cronyism.
Evidence of cronyism was prevalent and this contributed to poor performance.
There was evidence of cronyism which contributed to poor performance.
There was evidence of cronyism and this had some effect on performance.

Large power distance
While there was evidence of large power distance, this did not affect performance due to an effective monitoring framework.
Large power distance had a negative effect on good corporate governance.
Large power distance had a negative effect on good corporate governance practices.
Large power distance affected good corporate governance practices.

Materialism
There is no effect of materialism on corporate governance. Responding to a question about the ideal legal form, respondents from ESCOM and Sunbird said that if these organisations are to perform well, the government should not own more than 50% of the shares. The findings of this study are consistent with the study conducted by Lin and Fu (2017) on Chinese companies listed on Shanghai Stock Exchange (SHSE) and Shenzhen Stock Exchange (SZSE), who found that government-controlled companies are related to poor performance. One way of constraining state power is through listing on the stock exchange.
Other highlights from the interviews have revealed that board size does not significantly impact the SOEs' performance.
However, respondents noted that political party leaders use SOEs to reward their political protégés. Responding to the question about the effect of tenure on the performance of SOE, some respondents felt that the short term has a negative effect on the stability of the SOE (Department of Statutory Corporations, DSC). In contrast, others felt that when people overstay in a position, they lose innovation and creativity (ESCOM). The short term was attributed to poor performance and the eventual collapse of Malawi Development Corporation (MDC). Where the state is a sole shareholder, the appointment process is flawed, the board is captured, and hence the committees become ineffective. Committees as board structures in this scenario are used for legitimation to meet the minimum requirements or to conform to the rules and regulations.
Transparency and disclosure are the key elements of corporate governance. However, interviews and documents reviewed reveal that there is a prevailing culture of secrecy that pervades Malawian corporate society. This culture of secrecy promotes non-disclosure of conflict of interests and compromise levels of compliance. It was also noted that non-compliance issues are related to cronyism which is prevalent in a legal form where the government is the sole shareholder. The study also noted that most of the institutions did not bother to disclose their financial results and other material effects to the general public except for Sunbird, the listed SOE. A lack of transparency and disclosure is a characteristic of a large power distance society. This finding agrees with the study of Qu and Leung (2006) who noted that large power distance cultures are less transparent and very secretive in terms of disclosures. Large power distance is an antecedent to culture of cronyism. Boateng, Wang, Ntim, and Glaister (2020) found that large power distance has an influence on favouritism and nepotism which form part of the elements of cronyism.
Cronyism can either be horizontal or vertical. Horizontal cronyism is where peers share favours, while in vertical cronyism, superiors extend favours to their subordinates (Shaheen, Bari, Hameed, & Anwar, 2019). It is the vertical cronyism that is prevalent in Malawi. Related to cronyism, the study observed a growing culture of materialism. Materialism has been associated with increased incidents of corruption. Lu and Lu (2010) found that higher materialism was related to lower ethical values in Indonesia. Several respondents attributed materialism to lower levels of ethical standards in Malawi's corporate society.

PROPOSED CORPORATE GOVERNANCE FRAMEWORK AND RECOMMENDATIONS
Based on the aforementioned findings, the study would like to make the following recommendations to strengthen the corporate governance framework. Figure 2 shows that management reports directly to the board, which reports to shareholders. In addition, Malawi Stock Exchange (MSE) serves as an external governance monitoring mechanism. Line ministry advised on policy matters, but it does not exercise ownership rights on behalf of shareholders. We propose that listed SOEs use this structure because it has proved effective.   The state should establish an ownership entity that has the capacity to exercise its shareholders' rights.
 The study recommends that the DSC can be capacitated to be an ownership entity for the state.
 The appointment process of directors and management for wholly-owned government SOEs should be outsourced to independent job placement companies. This will ensure that there is transparency and that only qualified directors are recruited.
 The board tenure should be a minimum of 3 years subject to renewal on satisfactory performance for additional 3 years. However, the tenure of Chairperson should be a minimum of 4 years and subject to renewal for an additional one term. Tenure should not be tied to political changes.  The reporting system should be streamlined. Management should report to the board and the board should report to the ownership entity, which should report to the parliamentary committee on public accounts.

Recommendations to the National Assembly
Recommendations to the National Assembly are as follows:  The Parliament should amend Acts to create harmony. Clauses that give the minister power to appoint and dismiss directors or CEOs should be amended.
 The Parliament, through the Public Appointments Committee, should confirm all candidates appointed to serve as directors through a transparent appointment process.
 The Parliament, through the Public Accounts Committee, should review performance reports from SOEs and hold the board accountable.
 The Parliament should amend roles that give independence to the governance bodies like ACB, NGO, and National Audit Office.

Recommendations to the boards of directors
The board is the highest internal governance institution of an SOE tasked with making performance improvement decisions. The following recommendations are made for the board:  The board should perform its fiduciary responsibility with due care and act independently without external influence.
 The board should be the only authority in appointing CEOs for SOEs.
 The board should submit performance reports to shareholders, the ownership entity on behalf of the shareholders, and the public accounts committee of the National Assembly timely.

CONCLUSION
The study concludes that large power distance, cronyism, and materialism negatively affect the quality of corporate governance. The study also concludes that legal form and ownership arrangements influence the performance of SOEs measured by ROA. Increased state ownership hurts performance, and decentralised arrangements negatively affect monitoring and control and invariably the performance of SOEs. In addition, the study concludes that the increased shareholder's power of state nature without external monitoring mechanism leads to non-commercial expediency, which results in poor performance. The study further concludes that qualified and independent directors positively affect SOEs performance; on the contrary, a captured and ingratiated board has a negative effect on performance. The study has found that board effectiveness is influenced by the level of state ownership. Increased state ownership leads to a captured and ingratiated board.
Increased state ownership also negatively affects board effectiveness through the appointment process of the board and management. However, reduced state ownership through listing positively influences board quality and independence. The study concludes that shorter board tenure due Parliament DSC Board of directors CEO and top management of SOE Line ministry to increased state ownership has a negative effect on the performance of SOEs. On board structures, the study concludes that board committees positively influence performance. However, the effectiveness of the committees mirrored that of the board. Committees like the board are a function of state ownership. SOEs which had risk management committees performed better than their counterparts.
The study further concludes that civil servants and ex-officio members on the board do not add value to SOEs. On the contrary, their presence affects board independence and regresses the SOEs into government departments. In addition, their presence increases the conflict of interest. On the capital structure, the study concludes that leverage has a positive and significant influence on the performance of SOEs. It has further been concluded that soft budget constraint does not have any influence on performance. Finally, the study concludes that transparency and disclosure influence performance because it promotes accountability. Disclosure is moderated by legal form and cultural variables. Increased state ownership leads to low levels of disclosure.
The current study is a pioneering work that has combined socio-cultural values, corporate governance, and the performance of SOEs. The main contribution of this paper is the development of a strategic corporate governance framework for SOEs for a developing country to enhance performance. In addition, the authors have also established the influence of socio-cultural values on the effectiveness of corporate governance in less developed markets. The study has not exhausted the issue of socio-cultural values and their impact on corporate governance. Future studies should expand to private sector companies to understand how socio-cultural values impact corporate governance and company performance. This study focused on commercial SOEs; further studies should include all statutory bodies.
The other limitation is that this study is based on a case study of SOEs from one country. The use of the case study method has raised the debate of generalisation.
Whereas positivists view generalisation of the observed events in the empirical domain (Zachariadis, Scott, & Barrett, 2013), in other words, generalising findings from a sample to a population (Yin, 2016, p. 104), critical realists, on the other hand, argue that generalisation can only be that of generative mechanism (Zachariadis et al., 2013). While this study does not claim statistical generalisation, the use of replication logic has been used to achieve analytical generalisation (Yin, 2009).