Board governance quality and risk disclosure compliance among financial institutions in Uganda

Risk disclosure compliance (RDC) is important to all organizations in developed and developing economies. RDC brings about transparency through adhering with available risk regulations and improving quality of risk information (IASB, 2010). RDC lowers cost of external finance and reduces disclosure costs (Linsley and Shrives, 2006). According to the Association of Chartered Certified Accountants (ACCA, 2014), a number of corporate scandals (Enron, Ahold, Parmalat, Lehman Brothers, WorldCom and Xerox, etc.) faced risk misreporting scandals. As a result, investors found it difficult to assess the market return and risk profiles due to insufficient risk information disclosed (Abraham and Cox, 2007). The concept of RDC is facilitated through mandatory risk disclosure (Tauringana and Chithambo, 2016; Agyei-Mensah, 2017a, 2017b; Sellami and Fendri, 2017) and voluntary risk disclosure (Mokhtar and Mellett, 2013; Hassan, 2014). This paper examines the relation between board governance quality (BGQ) and RDC in Uganda. Empirical evidence in Uganda shows that RDC with the International Financial Reporting Standards (IFRS) requirements is still low (Nalukenge et al., 2018; Nalukenge, 2020). For instance, the COSASE report (2018) indicates compliance with IFRS 7 risk disclosure requirements is low especially with seven defunct banks. At the firm level, the Supervision report (2017) shows that some banks misreport their data submitted in as statutory returns and to credit reference bureau. In addition, World Bank (2014) indicates that financial statements of Savings and Credit Cooperatives Societies did not follow compliance guidelines and format. Therefore, the question that arises is whether BGQ is an important determinant of RDC in Uganda.

Whilst the prior literature shows that RDC in developing countries is still low (Salem et al., 2019; Agyei-Mensah, 2017a, 2017b; Mnif and Tahari, 2017; Tauringana and Chithambo, 2016), Khandelwal et al. (2019) indicate that academic research linking governance mechanisms and corporate risk disclosure are limited and rare. Moreover, relatively less devotion has been to examine a direct relation between BGQ and RDC (Mnif and Znazen, 2020; Salem et al., 2019; Jia et al., 2019; Agyei-Mensah, 2017a, 2017b; Mnif and Tahari, 2017; Juhmani, 2017; Tauringana and Chithambo, 2016). For instance, Tauringana and Chithambo (2016) and Agyei-Mensah (2017) and Juhmani (2017) and Mnif and Znazen (2020) provide evidence that higher proportion of independent directors increase RDC with IFRS 7. While, Salem et al. (2019) establish that board independence and audit committee independence enhance risk disclosure quality. Further, Sellami and Fendri (2017) find that financial expertise enhances corporate risk disclosure. Nevertheless, Jia et al. (2019) document no link between risk management committee meetings and risk management disclosure quality. Also, Sellami and Fendri (2017) and Mnif and Znazen (2020) show no impact existing between meeting frequency and compliance with IFRS. In addition, the agency theory (Jensen and Meckling, 1976) that focuses on principal-agent relationship existing in the separation of ownership and management continue to suggest that managers are critical in reducing agency costs which may arise due to nonRDC. Thus, once managers perceive that RDC roles are maximizing their utility, they will ensure full RDC.

Furthermore, in Uganda, some empirical studies on adoption of and compliance with IFRS disclosure requirements exist (Nalukenge, 2020; Bananuka et al., 2019; Nalukenge et al., 2018). These studies have not focused on RDC in particular. For instance, Nalukenge et al. (2018) indicate that corporate governance, ethical culture and internal controls over financial reporting (ICFR) are significant predictors of compliance with IFRS in Microfinance Institutions (MFIs). Moreover, Nalukenge (2020) find that board role performance, MFI size and membership are important factors to compliance with IFRS disclosure. In addition, Bananuka et al. (2019) indicate that audit committee effectiveness, isomorphic forces and managerial attitude are motivational factors on IFRS adoption in MFIs. Nevertheless, these studies focus on corporate governance mechanisms in general and ignore fundamental aspects such as BGQ. BGQ described in terms of board activity, board independence, board communication and board expertise (Kawaase 2013; Nkundabanyanga et al., 2013; Nkundabanyanga 2016, Nalukenge et al., 2017) are expected to impact RDC. The available limited studies on board governance in Uganda have been linked with service firm performance (Nkundabanyanga et al., 2013; Nkundabanyanga, 2016). Thus, acute scarcity of studies investigating the relation between BGQ and RDC exist in the Ugandan context.

Besides, to the author’s knowledge, no single study has examined the relationship between BGQ and RDC especially in Uganda. This paper is motivated by the fact that prior studies on the relationship between BGQ and RDC are still limited both in developed countries and developing countries. Thus, examining how BGQ mechanisms impact RDC in a country setting such as Uganda might provide further evidence to address this research gap and forms a motivation for this study. This is achieved through administering a questionnaire survey and financial statements analysis of 83 Ugandan financial institutions (FIs). This study indicates that the overall RDC level is low. Moreover, the study finds the positive relation between BGQ and RDC. This means that FIs with effective BGQ have higher RDC than those with weak BGQ. Furthermore, the study also finds the positive relation between BGQ mechanisms and RDC.

This study makes contributes to the risk-disclosure literature by providing further empirical evidence that BGQ significantly affects RDC. Moreover, the study reveals the most important BGQ mechanisms in enhancing RDC. Furthermore, the study provides theoretical foundations to understand RDC better based on the agency theory. Besides, the study provides relevant insights into regulators and policy makers with early symptoms of potential problems regarding weak board governance in FIs. Also, policy makers may also use these findings as a guideline tool for improving existing board governance frameworks in place and development of new disclosure policies. In addition, the study provides an input into the review and amendments of existing corporate governance codes for the regulators.

The remainder of the paper is structured as follows. The next section reviews both theoretical, literature review and elaborates hypothesis. Research methodology follows next and section four reports the results and discussion of the study. Then the final section presents the conclusion of the study.

The study used the agency theory (Jensen and Meckling, 1976) to examine the relationship between BGQ and RDC in Ugandan FIs. This theory posits that where there is a separation of ownership and control, agency costs exists (Jensen and Meckling, 1976). In this view, FIs executive management is expected to perform their RDC roles on the behalf of the shareholders. The choice of this theory is to reduce agency costs through establishment of effective BGQ mechanism. BGQ is an instrument used to mitigate information asymmetry in an agency relationship (Krismiaji et al., 2016) and also to resolve agency problems by aligning management and shareholders’ interests. This is achieved through effective monitoring of management’s performance and ensuring compliance with risk disclosure requirements. The study envisages that if there is higher BGQ in place, the extent and quality of risk disclosures will be improved to enhance RDC. In Uganda settings, the theory confirms this argument which is revealed by Nkundabanyanga et al. (2014) and Nkundabanyanga (2016) that effective board governance in terms of board activity, communication, control meetings and organization has a significant impact on service firm performance, and Nalukenge et al. (2018) revealed that corporate governance positively affects IFRS compliance in MFIs. Based on these previous studies, it can be said that BGQ can affect the RDC.

The impact of BGQ on RDC has been investigated and prior research indicates that BGQ is related to RDC (Agyei-Mensah and Buertey, 2019; Oliveira et al., 2018; Agyei-Mensah, 2017a; Tauringana et al., 2016; Krismiaji et al., 2016; Mokhtar and Mellett, 2013). The concept of BGQ is built from a behavioral perspective rather than a structural perspective (Nalukenge et al., 2017; Nkundabanyanga et al., 2013). According to Tauringana and Chithambo (2016) and Agyei-Mensah (2017a), boards that have higher proportion of independent directors will have higher levels of RDC. Thus, BGQ is contributory to RDC activities. However, empirical evidence about the relationship between BGQ and RDC is rather fragmented. Some studies found the positive impact of BGQ on RDC (Agyei-Mensah and Buertey, 2019; Oliveira et al., 2018; Agyei-Mensah, 2017a; Tauringana and Chithambo (2016); Mokhtar and Mellett, 2013). For instance, Agyei-Mensah and Buertey (2019) establish that governance quality mechanisms have a positive influence on corporate risk disclosure. Other studies by Krismiaji et al. (2016) find that board governance has a positive impact on accounting information quality after IFRS adoption. Likewise, empirical evidence from Uganda settings by Nkundabanyanga et al. (2014) and Nkundabanyanga (2016) find that effective board governance has a significant impact on service firm performance while Nalukenge et al. (2017) indicate that behavioral corporate governance has a positive influence on compliance with IFRS. Nevertheless, Majumder et al. (2017) provide evidence that board independence has no impact on corporate social disclosures. Based on previous studies, it is reasonably expected that with higher BGQ in place, the level of RDC by FIs will improve. Therefore, the following research hypothesis is as follows:

Board activity can enhance RDC activities. According to the agency theory, board meeting frequency and involvement of board members into strategic affairs improves effective monitoring which reduces agency costs and information asymmetry. As the board meets more often, members are allocated more time to discuss issues regarding RDC. Thus, the theory predicts that board activity as a measure of BGQ significantly impacts on the RDC levels. Prior research indicates that board activity is related to the extent of disclosure. However, these studies have found inconclusive findings. For instance, Aliyu (2019) and Majumder et al. (2017) establish that board meeting has a significant positive association with corporate environmental disclosures and corporate social disclosures. Moreover, Ahmed and Khan (2016) find that board meeting frequency significantly impacts on MFIs disclosures. In addition, other studies by Nkundabanyanga et al. (2014), (2015) and Nkundabanyanga (2016) in Uganda establish board activity as a significant predictor on firm performance. While, Hemrit (2018) instead find that infrequent board meetings increase the extent of liquidity risk information. This implies that lesser board meetings more liquidity risk disclosures. Nonetheless, other studies find that board activity has no significant impact on risk disclosures (Mnif and Znazen, 2020; Xue and Niu, 2019; Sellami and Fendri, 2017; Saggar and Singh, 2017; Allini et al., 2016). Therefore, these arguments lead to the following hypothesis:

Board independence as a measure of BGQ has a link with RDC. According to the agency theory, the value of independence on the board promotes effective monitoring which reduces agency costs and information asymmetry and in turn enhances RDC (Fama and Jensen, 1983; Peter and Romi, 2015). Further, prior studies on this relationship between two concerned variables have found mixed findings. For instance, Tauringana and Chithambo (2016), Agyei-Mensah (2017a, b), Sellami and Fendri (2017), Juhmani (2017) and Mnif and Znazen (2020) find that board independence has a positive significant impact on compliance with risk disclosure requirements. Moreover, Hemrit (2018) and Neifar and Jarboui (2018) find that proportion of independent nonexecutive members significantly impact on the operational and liquidity risk disclosure level. Nevertheless, other studies by Rahman and Hamdan (2017), Alfraih and Almutawa (2017), Ahmed and Khan (2016) find that board independence insignificantly impacts on the level of RDC while Majumder et al. (2017) provide evidence that board independence insignificantly impact on corporate social disclosures. Accordingly, the following hypothesis attempts to synthesize the previous arguments:

According to the agency theory (Jensen and Meckling, 1976), effective board communication is relevant in reducing agency problems. Maassen (1999) argues that the need for proper flow of information between board structures is very significant. Nkundabanyanga et al. (2013) define board communication as advance delivery of board meeting notices, drafting of clear minutes and delivery of board papers in advance. Existing prior studies have examined board communication not directly linked to RDC. For instance, Nkundabanyanga et al. (2013) find that for boards to be effective, in return for their service, they expect proper flow of communications, advance preparation for board discussions and judicious use of time. In addition, Nkundabanyanga et al. (2014) and Nkundabanyanga (2016) establish that having proper flow of communications, advance preparation for board discussions and judicious use of time stimulates service firm performance. Furthermore, Schwartz-Ziv and Weisbach (2013) argue that board minutes are important in analyzing whether boards and committee meetings have actually discussed issues at meetings well. Similarly, Pugliese et al. (2015) document that use of video recordings during board meetings are key in examining how directors interact with each other across different agenda items. Based on the previous review, it is expected that effective board communication can lead to RDC. Hence, the following hypothesis is formulated:

Board expertise is such a critical aspect of BGQ. According to the agency theory, Nalukenge et al. (2017) and Sultana and Van der Zahn (2015) reveal that board financial expertise improves monitoring ability and due diligence which enhances disclosure compliance. Prior studies indicate board expertise is related to RDC. For instance, Sellami and Fendri (2017) find that having accounting and financial experts with sound knowledge of the company’s business environment and more experienced in the industry enhances corporate risk disclosure. In addition, Nalukenge et al. (2018) find that board expertise positively and significantly affects IFRS compliance in MFIs. Nevertheless, BuckbyGallery and Ma (2015) indicate that board expertise has no significant impact on risk management disclosures. Therefore, these arguments lead to the following hypothesis:

The study was cross-sectional survey design comprising a population of 210 licensed FIs in Uganda. A sample of 138 FIs was determined using Krejcie and Morgan (1970) sampling table. The selection of 138 FIs was done through stratified random sampling technique. Using guidance of Field (2009), a minimum of three respondents were selected through the purposive sampling method. The units of inquiry were risk director, finance director, executive director or board member based on the premise that they had sufficient knowledge. Out of the 138 questionnaires that were distributed, 83 final useable questionnaires were returned (60% response rate). Data were collected from audited financial statements and the questionnaire survey.

As shown in Table 1, majority firms are banking institutions representing 73.5%, 84.3% represents firms in existence for ten years and above, 71.1% has 20 branches and above, 61.4% is foreign owned and 90.4% has more than seven board members on the board. Furthermore, majority respondents are male (56.6%), age bracket between 35 and 45 years (60.6%), job experience ranging from 11 to 15 years (42.5%), 52.3% has bachelor’s degree and 77.2% is risk directors (Table 1). The findings indicate that nature of FIs and respondents in terms of experience and knowledge to interpret and respond to questionnaire items.

(See Table 2). RDC was measured using an existing scale developed by Tauringana and Chithambo (2016) according to mandatory risk disclosure with voluntary risk disclosure scale based on Hassan (2014). This scale has been widely used in previous studies to measure RDC (Agyei-Mensah, 2017a). Following prior studies by Tauringana and Chithambo (2016) and Agyei-Mensah (2017a), a compliance index was computed using an unweight disclosure index (Juhmani, 2017). The disclosure index included risk disclosure items dichotomously scored as one if disclosed or zero otherwise. Furthermore, the risk disclosure items were coded as not applicable if not applicable to avoid penalizing the institution. Data were obtained from audited financial statements for financial year 2016 but accessed in 2017. A financial statement was our basis of assessment to show actual compliance of risk disclosure items. The reliability of the disclosure index was done through random selection of ten financial statements that were scored by coauthors and independent practitioners. The results show no significant differences in the scores between coauthors and practitioners. After scoring, a compliance index was computed by summing all risk items disclosed divided by maximum score of risk disclosures. This percentage level of compliance was converted onto a five-point Likert scale to match the scale for BGQ components. For instance, 1 = “0.0–20.0”, 2 = “20.0–40.0”, 3 = “40.0–60.0”, 4 = “60.0–80.0” and 5 = “80.0–100%”. This method was similar with the previous scholars by Nalukenge et al. (2018) and Nalukenge (2020).

BGQ was measured through board activity, board independence, board communication and board expertise (Kawaase, 2013; Nkundabanyanga et al., 2013; Nalukenge et al., 2017). All items were measured using five-point Likert scales. The questionnaire items were adopted and modified from previous studies. For example, board activity was measured using an existing scale developed by Nkundabanyanga et al. (2013) according to two dimensions: meeting frequency and board involvement. This scale has been used in previous studies to measure board activity (e.g. Nkundabanyanga 2016). Respondents were asked to rate their opinion on a scale from 1 (strongly disagree) to 5 (strongly agree), e.g. “our board makes use of management committee and deliberations reported to the board in full” and “our board always have regular meetings”. Further, board independence was measured using items from Nalukenge et al. (2017) that were originally developed by Kawaase (2013). Respondents were asked the following items: e.g. “our board members have no conflict of interest” and “our board is economically independent” on a scale from 1 (strongly disagree) to 5 (strongly agree). Moreover, board communication was assessed using three items developed by Nkundabanyanga et al. (2013). The respondents were asked to rate their opinion with regard to board meeting notices, board minutes and board paper distribution on a scale from 1 (strongly disagree) to 5 (strongly agree), e.g. “our board makes clear minutes” and “the board documents are delivered to members in advance”. Furthermore, board expertise was assessed using four items from Kawaase (2013), using a scale from 1 (strongly disagree) to 5 (strongly agree), e.g. “our board members have expertise in financial services activities” and “our board members are experienced in the nature of our business”. These variables are described in Table 2.

The study also included variables such as industry type, branch numbers, ownership type, firm age, firm size and board size (see Table 2). Previous studies document that such variables have an impact on RDC. Specifically, Juhmani (2017) and Sarea and Al Dalal (2015) find a positive significant link between industry type and compliance level with IFRS disclosure. Nevertheless, Appiah et al. (2016) find no association with compliance with IFRS disclosures. Besides that, Bananuka (2019) find that the number of branches has no impact on Internet financial reporting. In addition, Agyei-Mensah and Buertey (2019) find that ownership type significantly improves corporate risk disclosure. Furthermore, prior studies by Demir and Bahadir (2014) find a positive significant impact of firm age with mandatory IFRS disclosure requirements (Demir and Bahadir, 2014). Moreover, Grassaa et al. (2020) find a positive relationship between bank size measured in terms of total assets and risk disclosure behavior. Meanwhile, Agyei-Mensah (2017a), Juhmani (2017) and Agyei-Mensah and Buertey (2019) found no significant relationship between firm size RDC with IFRS7, while a negative significant association with IFRS disclosure (Appiah et al., 2016) and no association with corporate compliance with IFRSs (Juhmani, 2012). Finally, previous studies by Mnif and Znazen (2020) and Agyei-Mensah (2017a, b) find that board size significantly impacts on compliance with IFRS. Nevertheless, Salem et al. (2019) find no significant association between board size and corporate risk disclosure quality. These variables are expected to have a positive and significant impact on RDC.

The study analyzed data using SPSS (version 23) and partial least squares structural equation modeling (PLS–SEM) (SmartPLS3.2.7). The choice for the PLS–SEM was based on its suitability to estimate parameters of small sample size (Chin, 1998; Hair et al., 2019) and secondary data (Hair et al., 2019). The measurement model was analyzed using PLS algorithm in SmartPLS3.2.7 software to verify the reliability, convergent validity and discriminant validity of the constructs.

The reliability of constructs was evaluated through outer loadings and composite reliability index (Hair et al., 2019). The results in Table 3 show that all indicator loadings attained reliability, with all their outer loading values exceeding the 0.708 threshold (Hair et al., 2019). Additionally, Table 3 shows that all composite reliability values range from 0.851 to 0.891, above minimum 0.7 cut-off value, satisfying prerequisite of constructs reliability (Hair et al., 2019). Moreover, convergent validity was assessed using average variance expected (AVE) (Fornell and Larcker, 1981). The results in Table 3 shows that all variables achieved convergent validity with AVE values ranging from 0.611 to 0.698 above minimum 0.50 cut-off value (Hair et al., 2019).

Furthermore, discriminant validity was evaluated through Fornell and Larcker (1981) procedure and heterotrait–monotrait ratio of correlations (HTMT) (Henseler et al., 2015). The results in Table 4 show that the square root of the AVE for each construct was greater than its correlation with other constructs, demonstrating the discriminant validity (Fornell and Larcker, 1981). In addition, discriminant validity was also evaluated using HTMT (Henseler et al., 2015). This was so because Henseler et al. (2015) argue that Fornell–Larcker criterion and crossloadings are no longer sufficient methods. Table 4 shows that all variables satisfied discriminant validity, with HTMT values below minimum 0.85.

Descriptive statistics findings are presented in Table 5. The mean value of RDC is 3.679 while the SD is 0.405. This mean value approximates to 52.9%. Compared with prior studies of Sarea and Al Dalal (2015) and Hewaidy and Al Mutawaa (2010), compliance levels ranging from 40 to 60% were rated low. This implies that Ugandan FIs have not fully attained full RDC levels. The findings are consistent with previous studies of Salem et al. (2019), Tauringana and Chithambo (2016), Agyei-Mensah (2017a) and Mnif and Tahari (2017). Additionally, Table 5 also shows mean values of the independent variables range from 3.42 to 3.67 with SD values from 0.52 to 0.72. Compared with Field (2009), smaller standard deviations relative to their mean values imply that data points are close. This finding is similar with Nalukenge et al. (2018) and Bananuka (2019). Furthermore, the results show that the maximum value of board communication and board expertise is 5.00, while the minimum value is 1.50 and 1.33, respectively, indicating higher differences exist in terms of board communication and expertise in Ugandan FIs. Moreover, the results also indicate that smaller differences exist (maximum value = 4.89 and minimum value = 3.00) in terms of board independence in Ugandan FIs. This finding is in line with the stipulated best corporate governance practices in the FI Act (2004) and Corporate Governance Regulations (2005).

To examine extent to which variables are correlated, the Pearson correlation analysis was performed. Table 6 shows that correlation coefficient between variables was relatively weak and no pair-wise correlation was above ±0.8, indicating absence of multicollinearity in the model. Further analysis in Table 6 show a positive significant association between BGQ and RDC (r = 0.568^**, p < 0.01). This means that higher governance quality on the board and its subcommittees enhances RDC. Moreover, the results show that RDC is positively and significantly associated with board activity (r = 0.413^**, p < 0.01), board independence (r = 0.477^**, p < 0.01), board communication (r = 0.561**, p < 0.01) and board expertise (r = 0.505^**, p < 0.01), respectively. This implies that RDC is likely to improve more when on the board is more independent, holds regular productive meetings, effective communication and has more knowledgeable and experienced members. Furthermore, the results show that RDC is positively correlated with ownership type (r = 0.258^*, p < 0.05), firm size (r = 0.342^**, p < 0.01) and board size (r = 0.244^*, p < 0.05). However, industry type (r = 0.055, p > 0.01), number of branches (r = −0.190, p > 0.01) and firm age (r = −0.021, p > 0.01) show an insignificant relationship with RDC. This indicates that industry category, existing branch network and years in existence may have no influence on FIs RDC levels. Therefore, the results provide preliminary evidence that variables are associated with RDC.

This study used Smart PLS 3.2.7 software with PLS bootstrapping (5000 resamples) to test the study hypotheses through structural equation modeling. The study first evaluated model’s in-sample predictive power (Hair et al., 2019). Table 7 shows that model predictive ability, with adjusted R² value of 0.491 above minimum 0.10 (Falk and Miller, 1992). In addition, Table 7 shows no multicollinearity issues in the model, with highest VIF value (2.014) below acceptable threshold of three (Hair et al., 2019).

Further, the study posits a positive and significant relationship between BGQ and RDC as H1. Table 7 shows a positive and significant coefficient (β = 0.346, t = 3.980, p < 0.001); thus, H1 is supported. This finding suggests that high levels of BGQ lead to an increase in FIs RDC. The finding is essential as some earlier studies in the Ugandan context found evidence to suggest that board governance enhances firm performance (Nkundabanyanga, 2016), corporate governance enhances IFRS compliance (Nalukenge et al., 2018) and board performance improves IFRS compliance (Nalukenge, 2020). In addition, the COSASE (2018) reveals that the collapse of some banks from 1993 to 2016 in Uganda was driven by weak BGQ which exposed FIs to higher nonRDC levels. In fact, the focus on the agency theory is of high relevance in explaining RDC after the collapse of the banks. For example, PwC (2016) shows that some banks did not have approved board charter; some board members without tertiary certificate and some lacked of proper board minutes in place. A case in point is CBL, one of the collapsed banks, which was found to have board members without tertiary certificate, had no approved board charter and some board meetings appeared not to have taken place. Furthermore, the findings are consistent with previous studies (Agyei-Mensah and Buertey, 2019; Oliveira et al., 2018; Agyei-Mensah, 2017a; Mokhtar and Mellett, 2013) which found that some governance mechanisms significantly affect the level of RDC. The current findings confirm the significant role of BGQ in enhancing RDC of Ugandan FIs.

Additionally, the study predicts a positive and significant relationship between board activity and RDC as H1a. Table 7 shows a positive and significant coefficient (β = 0.211, t = 2.490, p < 0.01); thus, H1a is supported. This means that FIs whose boards meet more frequently and also involve their members into strategic affairs, the quality and depth of financial statements improve, which leads to higher RDC levels. In fact, the findings are consistent with previous studies by Aliyu (2019), Majumder et al. (2017) and Ahmed and Khan (2016) and inconsistent with the previous studies by Mnif and Znazen (2020), Xue and Niu (2019), Sellami and Fendri (2017), Saggar and Singh (2017) and Allini et al. (2016). This finding is consistent with our expectations that reflect the realities of the Ugandan market. For example, the enabling regulations on board meetings (FIs Act, 2004; Corporate Governance Regulations, 2005) simply require these FIs to hold meetings not less than once in every quarter of the financial year but are silent on how on a productive board meeting should be held. In this regard, FIs with higher productive board meeting frequency compared to others help board members understand available risk disclosure frameworks, standards and policies in place, which increases their vigilance toward compliance with risk disclosure requirements. This is also consistent with the agency theory which states that boards meet more regularly and frequently to enhance their capacity toward monitoring the management of FIs.

Moreover, the study predicts a positive and significant relationship between board independence and RDC, as H1b. Table 7 shows a positive and significant coefficient (β = 0.330, t = 3.704, p < 0.001); thus, H1b is supported. This suggests that FIs boards with no conflict of interest, perform their oversight role without any due influence, are truly independent minded and have financial independence increases their RDC levels. The findings are consistent with previous studies by Tauringana and Chithambo (2016), Agyei-Mensah (2017a, b), Sellami and Fendri (2017), Juhmani (2017), Oliveira et al. (2018) and Mnif and Znazen (2020), while it is inconsistent with studies of Rahman and Hamdan (2017) and Alfraih and Almutawa (2017). The findings confirm the significant role of board independence in boosting RDC in Ugandan FIs. Furthermore, within the situation of boards characterized by conflict of interest, management due influence and financial dependence, nonRDC costs will increase resulting into collapse (PwC, 2016). This study shows that Ugandan FI boards with higher board independence comply more with risk disclosure requirements than other parties. The findings are also consistent with the agency theory which states that an independent board mitigates agency costs and improves the firm’s compliance in terms of comprehensiveness and quality of risk disclosure.

Besides, the study predicts a positive and significant relationship between board communication and RDC as H1c. Table 7 shows a positive and significant coefficient (β = 0.348, t = 3.882, p < 0.001); thus, H1c is supported. The findings suggest that advance delivery of board meeting notices, drafting of clear minutes and delivery of board papers in advance are critical aspects that affect RDC of FIs. This is consistent with findings of Nkundabanyanga et al. (2014) and Nkundabanyanga (2016) that board communication predicted to influence service firm performance. This finding confirms the significant role of effective communication as an engine of RDC in the case of Ugandan FIs. The study suggests that FIs whose boards make advance delivery of board meeting notices, record proper board minutes and send board papers in advance are expected to handle more RDC issues than others.

Furthermore, the study predicts a positive and significant relationship between board expertise and RDC as H1d. Table 7 shows a positive and significant coefficient (β = 0.412, t = 5.168, p < 0.001); thus, H1d is supported. It suggests that better board expertise can be taken as a proxy for enhanced BGQ. These results support the requirement of the Institute of Corporate Governance of Uganda (2008) which recommends FI boards to be composed of qualified individuals with diversity of training, experience and background. The agency theory also confirms this argument that was revealed by Nalukenge et al. (2017) and Sultana et al. (2015) that board financial expertise improves monitoring ability and due diligence which enhances disclosure compliance. Furthermore, the findings are consistent with previous studies of Sellami and Fendri (2017), while it is inconsistent with the studies of Buallay and Al-Ajmi (2019) that found board financial expertise reduces sustainability reporting.

Regarding control variables, there are positive significant impacts of ownership type (β = 0.341, t = 3.732, p < 0.001), firm size (β = 0.265, t = 3.247, p < 0.01) and board size (β = 0.223, t = 2.038, p < 0.01) on RDC (See Table 7). This means that foreign-owned FIs, bigger and larger boards comply more with risk disclosures than locally-owned FIs and smaller boards. These findings are consistent with that of Agyei-Mensah and Buertey (2019) who found institutional ownership improving risk disclosure, whereas Mnif and Znazen (2020) and Agyei-Mensah (2017a) who found that larger boards enhance IFRS compliance. Furthermore, in line with firm size, Grassaa et al. (2020) found that larger banks in terms of total assets have a higher risk disclosure behavior compared to smaller ones while Trong and Nguyen (2020) found a negative significant relationship between firm size and firm performance. Further, Agyei-Mensah (2017a), Juhmani (2017) and Agyei-Mensah and Buertey (2019) found no significant relationship between firm size and RDC with IFRS 7. Nevertheless, Salem et al. (2019) found no significant impact between board size and quality of corporate risk disclosure. Moreover, the study found that industry type (β = 0.043, t = 0.476, p > 0.05), number of branches (β = −0.071, t = 0.641, p > 0.05) and firm age (β = −0.080, t = 0.807, p > 0.05) are statistically insignificantly related with RDC (See Table 7). These findings suggest that industry type, number of branches and firm age have no significant impact on the level of RDC. These findings are consistent with previous studies of Appiah et al. (2016) that found that industry effects are inconsequential to compliance levels with IFRS; and Bananuka (2019) who found that the number of branches has no impact on Internet financial reporting. However, there are inconsistent with studies of Juhmani (2017) which documented that industry type improves compliance levels with IFRS disclosure and Appiah et al. (2016) who found that younger firms comply more with IFRS disclosures. Therefore, after controlling for variables that may be related to RDC, this study finds that ownership type, firm size and board size remain statistically significant to RDC.

The main purpose of this study was to examine whether the BGQ mechanisms have a significant impact on RDC in Uganda. The findings of descriptive statistics indicate that that overall level of RDC in Ugandan FIs is low. Further, the statistical findings show that BGQ has a positive significant impact on the level of RDC in the selected Ugandan FIs. Furthermore, the findings support the notion that board activity, board independence, board communication and board expertise can lead to enhanced RDC in Ugandan FIs. Also, the findings show that ownership type, firm size and board size can influence the level of RDC.

This study offers important academic and practical implications. It contributes to the academic research by complementing to the previous studies (Tauringana and Chithambo, 2016; Agyei-Mensah, 2017a, 2017b) by expanding on the RDC existing literature, especially in the African experience. It is now evident that BGQ has a significant impact on RDC. Its focus on factors for effective BGQ adds new insights into the functionality of the predominantly unitary boards and this contributes to the scarcity of BGQ literature, especially in developing countries. Additionally, it adds to the theoretical foundations of BGQ in the risk disclosure knowledge.

Furthermore, this study revealed more than ever the vital role of board activity, independence, effective communication and board expertise on RDC. It seems that this study will make aware policy makers and managers of this fact that board governance qualities will be effective in the extent of RDC in developing countries like Uganda. In fact, it provides relevant insights into regulators and policy makers with early symptoms of potential problems regarding weak board governance in FIs. Further, policy makers can use these findings as a guideline tool to improve the existing board governance frameworks into their structures and development of disclosure policies. For example, FIs would handle RDC concerns when they have regular productive meetings in time. Additionally, the findings are helpful to regulators in providing as an input into the review and amendments of corporate governance codes to solve the problem of ineffective boards as evidenced by bank collapses in Uganda. Moreover, regulators in , for example BoU, UMRA and IRAU in Uganda, the issue of low RDC should be paid attention to by imposing stringent corporate governance reforms that incorporate these better governance qualities on their boards.

The study has a number of limitations. First, the study was restricted to a single financial year analysis and a single developing economy. Future studies should be expanded by using many financial year analyses across different countries and sectors. Second, the study only considered quantitative data through crosssectional research design settings and a positivistic approach. Future studies should be expanded by adopting mixed methods and longitudinal designs to enrich the understanding of how these BGQ mechanisms impact RDC. Third, the study did not perform robustness analysis for the measurement model and structural model (Hair et al., 2019). Future studies should be expanded by including holistic robustness analysis in PLS–SEM like confirmatory tetrad analysis and unobserved heterogeneity to check for the stability of the findings. Fourth, the study considered only certain BGQ mechanisms to assess the impact on RDC and explained only 49.1%. However, other governance mechanisms might have an impact on the RDC levels. Future studies should be expanded by incorporating other BGQ mechanisms to examine their relationship to RDC.