Descriptive statistics
Descriptive statistics consist of an exploratory
analysis of the sample and the research variables. Through this analysis, we
will first determine the trend of each variable. The descriptive statistics of
the study model are summarised in Table 3.
Description of the
variables
Table 3 summarises the trend of each variable in the
models. The mean and min and max values of the IR variable are 0 and 1
respectively. 0.595 means that 59.5% of the integrated reports in our study
sample are of good quality. Such a value reflects the importance attached by
the firms in our sample to publishing quality integrated reports.
A mean of the Board size variable of 11.49227 means
that the boards of the firms in our study sample have an average of 11 members.
A mean of the Gender Diversity variable of 19.09709
means that 19% of the board members of the firms in our study sample are
female.
A mean of the Bindep variable of 55.55183 means that
55.5% of the board of directors are represented by independent members.
A mean of the BTenure variable of 7.48 means that the
boards of the firms in our study sample meet 7 times a year.
A mean of the Duality variable of 47.5 means that
47.5% of the firms in our study sample have a duality of functions
A mean of the ESG Perf variable of 55.24, this mean
indicates that the companies in our study sample verify the existence of about
55% of the ESG performance criteria with a maximum of 87.24%.
For the CSR Strategy variable, the average is
49.87285; this average indicates that the companies in our study sample meet
approximately 50% of the social responsibility criteria with a maximum of
98.5%.
In the rest of the analysis of the results we will
interpret the participation of each explanatory variable in the evolution of
the quality of IR.
Correlation analysis:
Bivariate analysis
The correlation analysis aims to identify the
relationships between the variables. Table 4 summarises the correlation
coefficients between the variables in the model for evaluating the quality of
the integrated report discussed in this chapter, using the Pearson test. For
the model used, only the correlation coefficients for the variables Board size
and duality have negative signs, which implies that companies with large boards
of directors and which verify the combination of the functions of the chairman
of the board of directors and the chief executive officer may suffer from an
opaque or mediocre quality of the integrated report distributed. For all other
variables, the positive signs of the correlation coefficients between these
variables and the quality of the integrated report explain that these variables
improve the quality of the integrated reports disseminated by the firms in our
study sample. This reflects the fact that good corporate governance in general,
and board characteristics in particular, provide firms with a better
opportunity to improve the quality of integrated reporting. It is more likely
that a higher proportion of gender diversity is served by women with different
skills and attitudes, including characterised directors, and positively affects
the functioning of other governance mechanisms vis-à-vis the quality of the
integrated report of the firms in our study sample.
Multivariate Analysis:
Results and Interpretations
This part is a possible empirical validation of the
research hypotheses. After an exploratory study that dealt with the
specificities of the sample and the functional relationships between the
variables, we will perform a multivariate analysis.
Econometric tests
applied to the empirical model
???????????????? = ? + ?????BoardSize???????? + ?????????enderDiv???????? + ?????boardInd ???????? + ?????Btenure???????? +
?????Duality???????? + ?????ESGPerf???????? + ?????CSRstrategy ???????? + ????????????
Analysis of the results
of the empirical estimates
It should be recalled that the purpose of estimating
the empirical model is to verify and validate the research hypotheses presented
related to the effects of governance mechanisms on the quality of the
integrated report. The interpretation of the results presented in the table
below allows us to advance some analyses concerning the general characteristics
of the empirical models as well as the validation of the research hypotheses
conducted by the multivariate analysis. Indeed, the value taken by the explanatory
power of the first model R2adjusted= 0.8553 translates a good quality of the
model. This postulate implies that the integration of the different explanatory
variables makes it possible to explain 85.53% of the variation in the quality
of the integrated ratio of the observations in our study sample. Following the
confirmation of the quality of the evaluation model, we will mobilise the
theoretical framework of the argumentation of the results of the multivariate
analysis that justifies the relationship between the quality of the integrated
report and the mechanisms of corporate governance. This postulate is also
confirmed by the result of Fisher's statistics which confirms the capacity of
the independent variables of our econometric model to explain the variation of
the quality of the integrated report which takes a value (F= 327.75) for a
significance level P-value=0.0000. The econometric tests applied to the
explanatory models showed the absence of a heteroscedasticity problem.
According to, such a problem implies that the hazards do not have the same
variance, the variance-covariance matrix cannot be estimated systematically and
the generalized least squares estimator, which is an efficient estimator,
cannot be calculated [85].
The board size variable
The regression coefficients of the variable board
size, denoting the variable board size, is positive for the evaluation model by
a value of (t-student= 0.44; P = 0.657) not significant. This indicates that
the size of the board of directors is not significant in determining the
quality of the integrated report issued by the company.
Hypothesis (H1) not confirmed, this means that a very
large board is generally ineffective compared to smaller boards due to
communication and coordination issues [86].
The gender diversity
variable
The regression coefficients of the variable gender
diversity, designating the gender diversity within the board of directors, are
negative and significant at the 1% level (t-student= -5.39; P = 0.000). An
inverse sign to the expected one implies that the existence of a significant
proportion of women on the board deteriorates the quality of the integrated
report, such a value is justified by the low representation of women on the
board of the companies studied.
The unconfirmed hypothesis (H2), the greater
orientation towards actors who distinguish that there is no positive
relationship between the presence of women in the board and reporting quality
[87].
The B Indep variable
The results provide evidence for the acceptance of the
hypothesis of board member independence. Indeed, an expected and significant
coefficient sign of this variable leads to the conclusion that the percentage
of independent directors significantly reduces the quality of the reports
integrated by the companies in our sample (t-student= 13.93; P = 0.000). The
results found highlight the significance of the effect of the Bindep variable
on the quality of integrated reports.
Hypothesis (H3) is confirmed at the 1% significance
level, most of the actors distinguish that boards with a higher proportion of
independent directors are more effective in monitoring and controlling
management and are more successful in steering management towards long-term
value [88].
BTenure variable
The interpretation of the results presented in the
table below concerning the variable Btenure allows us to state that the
hypothesis related to this variable is confirmed at a significance level of 1%
for the model of evaluation of the quality of the integrated report. Indeed,
the fact of having values of (t-student= 4.99; P = 0.000), allows us to detect
the implication of this postulate in the improvement of the quality of the
integrated reports disseminated by the companies.
Hypothesis (H4) is confirmed at the 1% significance
level, this shows that we have obtained the same results from the research on
which we relied.
The Duality variable
With regard to the result of the estimation of the
evaluation models, the results provide elements of acceptance of the hypothesis
of the duality of the functions. Indeed, an expected sign of the coefficient of
this variable leads to the conclusion that the duality of the functions of
chairman and chief executive officer affects negatively and significantly at
the 1% threshold and reduces the quality of the integrated report of the
population of our study (t-student= -6.74; P = 0.000). Indeed, several studies
show a significant effect that the separation of the functions of chairman and
chief executive officer can exert.
Hypothesis (H5) is confirmed at the 1% significance
level, this result being consistent with research by [89].
The ESG Perf variable
The regression coefficient of the ESG Perf variable,
which refers to the percentile score ranking that takes into account a
company's ESG performance indicators, is positive but not significant for all
three models (t-student= 8.25; P = 0.000). This postulate implies that a high
level of performance contributes to the improvement of the quality of the
integrated reports processed in our study.
Hypothesis (H6) is confirmed at the 1%, % significance
level, this shows that we obtained the same results from the research we relied
on.
The SR strategy variable
With regard to the result of our estimation, we can
point out that the regression coefficient for the variable strategy, denoting
the CSR index, is positive and significant at the 1% level (t- student= 7.22; P
= 0.000). This hypothesis implies that an increase in the CSR index is worth
the improvement in the quality of integrated reports disseminated by companies listed
on the London Stock Exchange FTSE100.
Hypothesis (H7) is confirmed at the 1% significance
level, this shows that we obtained the same results from the research we relied
on.