Позитивные изменения. Том 3, № 4 (2023). Positive changes. Volume 3, Issue 4(2023) - Редакция журнала «Позитивные изменения» Страница 62

Table 1. The general matrix of life

Source: Field Survey, 2023.

Data obtained is analyzed via the multiple regression technique. The model specified for the study is stated as follows:

SRQi,t = β0 + β1SCSIZEi,t + β2SCINDEPi,t + β3SCDIVi,t + β4SCDILi,t + β5FSIZEi,t + εi,t (1)

where:

SRQ = Sustainability reporting quality;

SCSIZE = Sustainability committee size;

SCINDEP = Sustainability committee independence;

SCDIV = Sustainability committee diversity;

SCDIL = Sustainability committee diligence; CSIZE = Size of total assets;

ε = Error term;

β0 = Constant/intercept;

β1 — β5 = Coefficients of explanatory variables; and

i,t = company i for time period t.

Furthermore, the a priori expectation of the study is that sustainability committee size, independence, diversity and diligence positively impact sustainability reporting quality of the sampled banks. This is expressed as SCSIZE > 0; SCINDEP > 0; SCDIV > 0; and SCDIL > 0.

RESULTS AND DISCUSSION Descriptive and correlational statistics are provided in Tables 2 and 3 respectively for the variables being examined in this study: SRQ (N, mil.), SCSIZE (N, mil.), SCINDEP (mil.), SCDIV (mil.), SCDIL (mil.) and CSIZE (mil.)

Table 2 presents the minimum and maximum values of the indicator for sustainability reporting quality (SRQ), which are 0.264 and 0.868, respectively, and shows that the data for SRQ of sampled companies are narrowly dispersed from the mean of 0.601, given the standard deviation of 0.144.

On the other hand, the average sustainability committee size (SCSIZE) for the independent variables is 1.78, with a standard deviation of 2.64, indicating that the data is widely dispersed from the mean, as seen by the minimum and maximum values of 0 and 9 respectively. Given that the lowest and maximum values are 0 and 1, SCINDEP also shows a mean of 0.205 and a standard deviation of 0.325, showing a modest departure from the mean value. This means that even though at least one of the sampled companies succeeded in disclosing sustainability committee director’s independence to the fullest extent possible, the quality of disclosure among the sampled banks in both Nigeria and South Africa does not have significant variations because the majority of the companies only disclose a small portion of the information that is required to be disclosed.

The same is true for SCDIV data, which shows an average disclosure of 7.5 percent and a standard deviation of 0.14, both of which indicate that the data is not significantly out of the range of the mean value as shown by the lowest and maximum values of 0 and 0.5, respectively. Although the minimum number of 0 suggests no disclosure at all, this shows that one business of the companies studied achieved a 50 % disclosure of SCDIV. Similarly, the data for SCDIL has a mean value of 1 (which represents disclosure of 100 % of annual board meeting frequency information disclosable) and a standard deviation of 1.54, which suggests that the data for SCDIL is not significantly scattered from its mean value. It indicates that at least one bank from each of the two countries under review reports having attended no sustainability committee meetings during any of the research years (minimum value of 0), with the highest number of meetings per year being five (maximum value of 5).

Table 2. Results of Descriptive Statistics of Model Variables

Source: STATA 17.0 Output, 2023.

Table 3. Correlation Matrix of SRQ and Explanatory Variables

Source: STATA 17.0 Output, 2023.

Legend: P-value < 0.05

In addition, CSIZE, a control variable, has a mean value of 8.585 and a standard deviation of 0.765 when calculated using the natural logarithm of the sampled banks’ total assets. As seen by the minimum and maximum CSIZE values of 6.64 and 9.75, respectively, this demonstrates that the data for CSIZE is widely scattered from its mean.

High-quality sustainability reporting aids investors in making informed decisions about their investments’ longterm sustainability.

According to Table 3, the coefficients for SRQ’s positive relationships with SCSIZE, SCINDEP, SCDIL, and CSIZE are 0.147, 0.344, 0223, and 0.472, respectively; at the same time, SCDIV has a negative value of -0.007. However, only the associations with SCINDEP, SCDIL and CSIZE are significant. In addition, all the explanatory and control variables have moderate-to-strong positive and significant association with each other, except for the two positive associations between CSIZE/SCSIZE and CSIZE/SCDIV (0.177 and 0.023 respectively) which are not significant. These coefficients indicate that improvements in sustainability committee attributes are positively correlated to each other. Additionally, the control variable CSIZE exhibits a positive correlation with every research variable, with coefficients ranging from a low of 0.0227 (SCDIV) to a high of 0.4715 (SRQ). This suggests that as a result of the increase in asset size for the companies, the quality of their sustainability reporting also rises.

The variance inflation factor (VIF) for multicollinearity is used to further confirm that there is no multicollinearity among the explanatory variables. According to Murray, Nguyen, Lee, Remmenga and Smith (2012), continuous variables calculated over time (timevariant) are bound to have some elements of multicollinearity. Thus, the use of VIF measures the extent to which explanatory variables explain themselves in the models. Based on the position of Cohen, Cohen, West and Aiken (2013) and Akinwande, Dikko and Samson (2015), VIFs greater than 5 with the tolerance levels drawing closer to 0 show evidence of high multicollinearity among explanatory variables. The 8th column in Table 3 presents the individual VIFs of the explanatory variables with the lowest at 1.36 (CSIZE) and the highest at 7.86 (SCDIL). This implies that there is no perfect