CHAPTER FOUR

DATA PRESENTATION AND ANALYSIS

 4.1 Introduction
This chapter presents the results from the analysis of data and its interpretation. The chapter
is divided into four sections. The first section deals with the preliminary analysis of the
sample using descriptive statistics. The second section presents correlation analysis between
the explained and explanatory variables. This is followed by discussion of findings and policy
implications from the findings. The last section ends with a discussion of the various
robustness tests conducted in order to improve the validity of the results.

 4.2 Descriptive Statistics

The summary statistics of the explained and the explanatory variables are presented in Table
4.2 where minimum, maximum, mean, and standard deviation of the data for the variables in
the study are described.

Table 4.2: summary of Descriptive Statistics

Variable Observations Mean Std. Dev. Min Max

ROCE 140 2.87615 0.2766211 2.35416 3.50
2 7
SZE 140 39.81 17.31892 1.20 83.50
EFF 140 5.074693 2.532393 0.871 11.78
LIQ 140 8.857228 2.231658 4.57 14.878
AGE 140 7.11 5.103325 0.80 23.30
Source: Author’s computation using STATA 15.0 Software

The table 4.1 shows the detail account of the descriptive statistics for the explained and
explanatory variables respectively. Profitability as proxied by ROCE shows a mean of about
2.876152 and a maximum and minimum value of 2.354167 and 3.50 respectively. The
standard deviation level of 0.2766211 indicates little variation in the profitability in the listed
insurance company in Nigeria. The mean proportion of size of the listed insurance company
is 39.81 with a minimum and maximum value of 1.20 and 83.50. The standard deviation of
value of 17.31892 implies that there is a large variation across the sample of listed insurance

1
companies in Nigeria. Hence, the highly deviated size may have significant impact on the
profitability of listed insurance firms in Nigeria as this will be reflected in our regression
result.

The efficiency of insurance firms in Nigeria shows an average value of 5.074693 with a
maximum and minimum value of 0.871 and 11.78. Standard deviation value of 11.78 implies
that the efficiency of the firms deviates from its mean value up to 11.78. The analysis of
liquidity ratio shows a mean value of 8.857228 with a minimum and maximum value of 4.57
and 14.878 respectively. It also shows the value of standard deviation of 2.231658 which
implies that liquidity ratio through the analysis of its standard deviation revealed that
liquidity of the firm deviates from its mean value up to 2.23.

Finally, in respect of age of the firm, its average value shows 7.110 and a maximum and
minimum value of 0.80 and 23.30 respectively. It further shows a standard deviation of
5.103325. This implies that there is moderate variation among the values of age due to its
standard deviation. Therefore, this study is conducted to determine the extent to which the
variation in factors affect the financial performance of listed insurance firms in Nigeria.

 4.3 Correlation Matrix

The correlation coefficient represents the linear association or relationship between two
variables (explained and explanatory) and also between the explanatory variables themselves.

Table 4.3 below shows the correlation between the dependent variable which is Return on
Capital Employed (ROCE) and independent variables which are the firm size (SZE),
efficiency, (EFF), liquidity (LIQ), and age of the firms (AGE).

Table 4.3: Correlation Matrix Result

VARIABLE ROCE SZE EFF LIQ AGE

ROCE 1
SZE 0.1850* 1
EFF 0.3820* 0.0207 1
LIQ -0.3806* -0.4352* 0.1592 1
AGE -0.1065 0.1088 -0.0035 -0.2795* 1
Source: Author’s computation using STATA 15.0 Software

Note: (*) denotes 5% level of significance.

2
The significance level of this result is indicated in the table. The coefficient of correlations
between firm size and financial performance of listed insurance firms is 0.1850. This implies
that firm size is positively related to financial performance up to the tune of 18.52%. The
result therefore revealed a positive relationship between firm size and return on capital
employed. The analysis of firm’s efficiency as measured by the ratio of revenue to total asset
revealed a weak positive and significant relationship with return on capital employed. This
result is an indication that firms with high efficiency will report high (ROCE).

The table also shows that financial performance (ROCE) is negatively and significantly
correlated with liquidity as indicated by the correlation coefficient of -0.3806. This implies
that the more liquid the firms are, the less their financial performance. Finally, the analysis of
age shows that the variable is negatively correlated with financial performance of listed
insurance firms in Nigeria as shown by the correlation coefficient value of -0.1065. This
implies that the older the firm, the lower the financial performance.

 4.4 Heteroscedasticity Test

One of the important assumptions of classical linear regression model is that the disturbances
appearing in the population regression are homoscedastic. This means that the variance of the
error term is consistent. A large chisquare indicates presence of heteroscedasticity. In the
result obtained from the heteroscedasticity test conducted in this work, the chi-square value
was large and p-value is small as can be seen in the table 4.4 below, indicating
heteroscedasticity was present and this shows violation of assumption number four of
classical linear regression model which states that there must be constant variance, that is the
disturbances appearing in the population regression function are homoscedastic. Therefore, as
a result of the presence of heteroscedasticity, the researcher decided to conduct fixed and
random effect model which will take care of the individual differences within units. This will
ensure that conclusions reached or inferences made are not misleading

Table 4.4: Heteroscedasticity Test Result

Chi Square value p-value

9453.723 0.003325

Source: Author’s computation using STATA 15.0 Software

Table 4.4: shows the result of the heteroscedasticity test where the high value of the chi
square and the smallness of the probability value shows that presence of heteroscedasticity.

3
Hence, necessitate the conduct of fixed and random effect model which will take care of the
individual differences within units.

4.5 Presentation and Discussion of Regression Result

This section deals with the regression result of the explained variable proxied by Return on
Capital Employed (ROCE) and the independent variables (ZSE, EFF, LIQ, and AGE) of the
study. The results generated from fixed and random effect model were presented first before
Hausman specification test so as to decide the appropriate model from two possible options.

Table 4.5 Fixed effect Model Regression Result

ROCE Coefficient Strd. Error t- value P -value

SZE .004138 .0016206 2.55 0.012

EFF -.0096558 .0076344 -1.26 0.008
LIQ .0003024 .0107968 0.03 0.978
AGE .0000668 .0035411 0.02 0.045
C 2.757267 .1221972 22.56 0.000
R-sq: F(4,122) = 2.03
Within: 0.6250 Prob > F =0.00937
Between: 0.9470
Overall: 0.8850

Source: Author’s computation using STATA 15.0 Software

Table 4.5 above shows the results of fixed effects model. It revealed that firm size, efficiency,
and age are significant while liquidity is not significant. Out of all the variables, three
variables (SZE, EFF, and AGE) are all significant at 5% level of significance. The R 2 within,
below and overall are 62.50%, 94.70% and 88.50% respectively. Within R 2 means that
independent variables explain 62.50% variations in the profitability in this panel from year to
year. Between R2 indicates that independent variables explain the 94.70% variations in
profitability from firm (cross-sectional unit) to other firm. While overall R 2 shows that
independent variables explains 88.50% variations in the whole panel.

 Table 4.6 Random Effects Model

ROCE Coefficient Std. Error z-value p-value
SZE .003927 .0015129 2.60 0.009

4
 EFF .0120788 .0073838 1.64 0.002
LIQ -.0041179 .0103342 -0.40 0.690
AGE 0004103 .0034501 0.12 0.905
C 2.820506 .1304582 21.62 0.000
R-sq: Wald chi2 (4) = 10.12
Within = 0.5950 Prob.> chi2 = 0.0384
Between = 0.4447
Overall = 0.2273
Source: Author’s computation using STATA 15.0 Software

A result of random effect model is provided in table 4.6. Variables such as size, and
efficiency are significant in this model while liquidity and age are not significant. Firm size
and efficiency are significant at 5% level of significance. The within R 2 of this model is
59.50%, between R2 is 44.47% while the overall R2 of the panel is 22.73%. This model is also
significant as indicated by Wald chi2 of 10.12 at 5% level of significance. Within R2 and
overall R2 of random effects model are higher as compared to fixed effects model.
Alternatively, between R2 of fixed effects model is greater than that of the random effects
model. The results of both the fixed and random effect models are significant at 5% level of
significance. As a result, it is hard to choose which model is appropriate. To handle this
problem, the study conducted Hausman‟s specification test in order to decide the appropriate
one from the two possible options. The result of this test is provided in the Table 4.7 below:

 Table 4.7 Hausman Specification Test Result

Random Fixed Difference
SZE .003927 .004138 -.000211
EFF -.0120788 -.0096558 -.002423
LIQ -.0041179 .0003024 -.0044203
AGE -.0004103 .0000668 -.0004771
Chi2 (4) = 5.60
Prob. = 0.2314
Source: Author’s computation using STATA 15.0 Software

The outcome of Table 4.7 above suggests that the most appropriate model is Random Effect
model. This is because the chi2 value of this test is 5.60 which is not significant. Therefore,

5
Hausman specification test proved that random effects model is the more appropriate for this
study. The random effect has two significant variables which include firm size (SZE), and
Efficiency (EFF) while Liquidity and Age (AGE) are insignificant. Therefore, our
interpretation is based on random effect regression model. The summary of the regression
result obtained from random effects model of the study were analyzed and discussed.

 Firm size (SZE) and Financial Performance (ROCE)

The random effects regression result revealed that firm size as shown in table 4.6 has a z-
value of 2.60 and a coefficient value of .003927 with a significant value of 0.009. This
signifies that firm size has a positive and strong influence on financial performance of listed
insurance firms in Nigeria, that is, the greater the size of a firm, the higher the reported
ROCE. The positive effect of firm size is consistent with our a priori expectation. This
implies that for every 1% increase in firm size, the ROCE will increase by 0.3927%. The
finding is in line with the resource based theory which articulates a positive and significant
relationship between firm size and financial performance of a firm.

 Efficiency (EFF) and Financial Performance (ROCE)

The table 4.6 above shows that z-value for efficiency is 1.64 and coefficient of .0120788 with
a p-value of 0.002 which is therefore significant at 5% level of significance. This signifies
that efficiency is positively and significantly influencing the financial performance of listed
insurance firms in Nigeria. This implies that the higher the level of efficiency, the higher the
financial performance of listed insurance firm proxied by ROCE. The positive effect of the
value of efficiency is consistent with our a priori expectation, implying that every 1%
increase in value of efficiency causes financial performance to increase by 1.21%.

 Liquidity and Financial Performance

Liquidity ratio as measured by the ratio of current assets to current liabilities (CA/CL) shows
a z-value of -0.40 with a coefficient value of -.0041179 and an insignificant p-value of 0.690.
This implies that liquidity ratio is negatively influencing the financial performance of listed
insurance firms in Nigeria within the study period. This means that the higher the liquidity
position of the firm, the less the reported financial performance as proxied by ROCE. The
negative effect of liquidity is inconsistent with our a priori expectation. This implies that for
every 1% increase in liquidity ratio, the financial performance of listed insurance firms in

6
Nigeria will decline by 0.41%. Therefore, it can be concluded that the more liquid firms are,
the less the return on capital employed.

 Firms Age (AGE) and Financial Performance (ROCE)

The age of insurance firms shows a z-value of 0.12 and a coefficient value of .0004103 with
p-value of 0.905 which is statistically insignificant. This shows that age of insurance
companies has a positive coefficient but insignificant in explaining and predicting the
financial performance of listed insurance firms in Nigeria within the study period. The
relationship which is not line with the expected sign implies that older firms are more
experienced and are not prone to the liabilities of newness, and can therefore enjoy superior
performance. Older firms may also benefit from reputation effects, which allow them to earn
a higher margin on sales. herefore, it can be concluded that the age of insurance firms still
explains the variation in profitability of insurance companies positively.

The R2 (0.2273) which is the multiple coefficient of determination gives the percentage or
proportion of total variation in the dependent variables measured by ROCE explained by the
independent variables jointly. Hence, the result of R 2 signifies that 22.73% of total variation
in the profitability measured by ROCE is caused by firm size, efficiency, liquidity, and age of
the firms. This indicates that the model is fit and the explanatory variables are properly
selected, combined and used.

Therefore, Wald chi2 of 10.12 which is significant at one percent (5%) indicates that financial
performance and firm specific characteristics model is fit. This means that there is 95%
probability that the relationship among the variables is not due to mere chance.

4.6 Robustness tests

This section presented the result of robustness test conducted to improve the validity of the
statistical inferences for the study. The problem of multicollinearity is discussed based on the
result generated for the purpose of the study. Multicollinearity is investigated using variance
inflation factor (VIF) value. The result of multicollinearity test is presented in Table 4.8
below:

Table 4.8 Multicollinearity Test

VARIABLES VIF
SZE 1.25

7
 EFF 1.04
LIQ 1.37
AGE 1.09

Source: Author’s computation using STATA 15.0 Software

The Variance Inflation Factors (VIF) for all the variables as shown in Table 4.8 are found to
be consistently smaller than 10, indicating an absence of multicollinearity. This shows the
appropriateness of the model of the study with the four explanatory variables.

4.7 Implication of Findings

Findings of this study have several implications for investors and insurance firms. Viewing
evidence about what drives insurance firm’s profitability will help business understand which
factors are critical to track and analyze in order to attain operational success. It is paramount
to note that, if insurance firms know the factors that can boost performance, it will culminate
into creating increased competition in the market place. Investors can use the knowledge
derived from the findings of this study to take care of their investment. The outcomes of this
study contribute towards a better understanding of the firm specific characteristics that affect
financial performance of listed insurance firms in Nigeria.