V.4.2- Social performance regression analysis
There are three social performance regressions based on the
dependent variables, namely: AL/GDP, CFIR, CFGR
? MODEL SUMMARY: AL dependent variable
Model1: FD, HLAR, COA, ROA, TA
Model2: FD, HLAR, COA, ROA, TA, ROE Model2: FD, HLAR, COA, ROA,
TA, OSS
|
Model
|
R
|
R
Square
|
Adjusted R Square
|
Std. Error
of the
Estimate
|
Change Statistics
|
|
R Square
Change
|
F Change
|
df1
|
df2
|
Sig. F
Change
|
|
1
|
1.000*
|
1.000
|
1.000
|
0.23107
|
1.0000
|
391223.8187
|
5
|
39
|
0.000
|
|
2
|
1.000**
|
1.000
|
1.000
|
0.23386
|
0.0000
|
0.0739
|
1
|
38
|
0.787
|
|
3
|
1.000***
|
1.000
|
1.000
|
0.23050
|
0.0000
|
2.1161
|
1
|
37
|
0.154
|
* Predictors: (Constant), FD, HLAR, COA, ROA, TA
**Predictors: (Constant), FD, HLAR, COA, ROA, TA, ROE
***Predictors: (Constant), FD, HLAR, COA, ROA, TA, ROE, OSS
Table16: ANOVA of AL regression
|
Model
|
|
Sum of
Squares
|
df
|
Mean
Square
|
F
|
Sig.
|
|
1
|
Regression
|
104442.66
|
5.00
|
20888.53
|
391223.82
|
0.00
|
|
Residual
|
2.08
|
39.00
|
0.05
|
|
|
|
Total
|
104444.75
|
44.00
|
|
|
|
|
2
|
Regression
|
104442.67
|
6.00
|
17407.11
|
318278.41
|
0.00
|
|
Residual
|
2.08
|
38.00
|
0.05
|
|
|
|
Total
|
104444.75
|
44.00
|
|
|
|
69
Analysis of microfinances' performance and
development of informal institutions in Cameroon
By Djamaman Brice Gaétan
|
3
|
Regression
|
104442.78
|
7.00
|
14920.40
|
280823.17
|
0.00
|
|
Residual
|
1.97
|
37.00
|
0.05
|
|
|
|
Total
|
104444.746
|
44
|
|
|
|
Table17: AL regression coefficients
|
Model
|
|
Unstandardized Coefficients
|
Standardized
Coefficients
|
t
|
Sig.
|
|
1
|
B
|
Std. Error
|
Beta
|
|
|
|
(Constant)
|
-0.0200
|
0.0737
|
|
-0.2707
|
0.7881
|
|
ROA
|
0.0167
|
0.0053
|
0.0030
|
3.1660
|
0.0030
|
|
TA
|
0.0000
|
0.0000
|
1.0166
|
42.7652
|
0.0000
|
|
COA
|
0.0005
|
0.0005
|
0.0009
|
1.0850
|
0.2846
|
|
HLAR
|
0.0003
|
0.0003
|
0.0006
|
0.8513
|
0.3998
|
|
FD
|
0.0000
|
0.0000
|
-0.0149
|
-0.6274
|
0.5341
|
|
2
|
(Constant)
|
-0.0163
|
0.0758
|
|
-0.2151
|
0.8309
|
|
ROA
|
0.0169
|
0.0054
|
0.0031
|
3.1375
|
0.0033
|
|
TA
|
0.0000
|
0.0000
|
1.0178
|
41.5274
|
0.0000
|
|
COA
|
0.0005
|
0.0005
|
0.0009
|
1.0727
|
0.2902
|
|
HLAR
|
0.0003
|
0.0003
|
0.0006
|
0.8418
|
0.4052
|
|
FD
|
0.0000
|
0.0000
|
-0.0162
|
-0.6594
|
0.5136
|
|
ROE
|
-0.0002
|
0.0007
|
-0.0002
|
-0.2719
|
0.7872
|
|
3
|
(Constant)
|
0.0691
|
0.0950
|
|
0.7272
|
0.4717
|
|
ROA
|
0.0179
|
0.0054
|
0.0033
|
3.3490
|
0.0019
|
|
TA
|
0.0000
|
0.0000
|
1.0005
|
37.1174
|
0.0000
|
|
COA
|
0.0003
|
0.0005
|
0.0005
|
0.6499
|
0.5198
|
|
HLAR
|
0.0002
|
0.0003
|
0.0006
|
0.7627
|
0.4505
|
|
FD
|
0.0000
|
0.0000
|
0.0012
|
0.0459
|
0.9636
|
|
ROE
|
-0.0002
|
0.0007
|
-0.0002
|
-0.3018
|
0.7645
|
|
OSS
|
-0.0006
|
0.0004
|
-0.0013
|
-1.4547
|
0.1542
|
The p-value of F test is 0.000. This means that the overall
model is statistically significant under ANOVA analysis. The R-squared of all
the models is 1.000 meaning that perfectly 100% of the variability of
AL is accounted for by the variables in the model. The
coefficients for each of the variables indicates the value of change that one
could expect in AL given a one-unit change in the value of
that variable, given that all other variables in the model are held
constant.
F test: we can observe that all the 3 models have a very
higher value of F test. This is very significant: therefore we can conclude
that financial performance indicator is useful to estimate the dependent
variable Average Loan/GDP. In fact, there is positive link between AL and
70
Analysis of microfinances' performance and
development of informal institutions in Cameroon
By Djamaman Brice Gaétan
financial performance variables ceteris paribus. This implies the
availability of fund by the microfinance institution. Thus, an allowance
resulting from good economic figures brings in an improvement of SP (H2a)
? MODEL SUMMARY: CFIR dependent variable
Model1: FD, HLAR, COA, ROA, TA
Model2: FD, HLAR, COA, ROA, TA, ROE Model2: FD, HLAR, COA, ROA,
TA, OSS
|
Model
|
R
|
R Square
|
Adjusted R Square
|
Std. Error of
the Estimate
|
Change Statistics
|
|
R Square
Change
|
F Change
|
df1
|
df2
|
Sig. F
Change
|
|
1
|
0.117*
|
0.014
|
-0.116
|
22.67354
|
0.0137
|
0.1052
|
5
|
38
|
0.990
|
|
2
|
0.212**
|
0.045
|
-0.110
|
22.61250
|
0.0311
|
1.2054
|
1
|
37
|
0.279
|
|
3
|
0.213***
|
0.046
|
-0.140
|
22.91574
|
0.0007
|
0.0272
|
1
|
36
|
0.870
|
* Predictors: (Constant), FD, HLAR, COA, ROA, TA
**Predictors: (Constant), FD, HLAR, COA, ROA, TA, ROE
***Predictors: (Constant), FD, HLAR, COA, ROA, TA, ROE, OSS
Table18: ANOVA for CFIR regression
|
Model
|
|
Sum of
Squares
|
df
|
Mean
Square
|
F
|
Sig.
|
|
1
|
Regression
|
270.494806
|
5
|
54.0989612
|
0.10523262
|
0.99047007
|
|
Residual
|
19535.3916
|
38
|
514.089252
|
|
|
|
Total
|
19805.8864
|
43
|
|
|
|
|
2
|
Regression
|
886.860095
|
6
|
147.810016
|
0.28907252
|
0.93839558
|
|
Residual
|
18919.0263
|
37
|
511.325034
|
|
|
|
Total
|
19805.8864
|
43
|
|
|
|
|
3
|
Regression
|
901.168592
|
7
|
128.73837
|
0.24515475
|
0.97053821
|
|
Residual
|
18904.7178
|
36
|
525.131049
|
|
|
|
Total
|
19805.8864
|
43
|
|
|
|
71
Analysis of microfinances' performance and
development of informal institutions in Cameroon
By Djamaman Brice Gaétan
Table19: CFIR regression
coefficients
|
Model
|
Unstandardized Coefficients
|
Standardized
Coefficients
|
t
|
Sig.
|
|
B
|
Std. Error
|
Beta
|
|
|
|
1
|
(Constant)
|
3.731
|
7.309
|
|
0.510
|
0.613
|
|
ROA
|
0.047
|
0.539
|
0.018
|
0.086
|
0.932
|
|
TA
|
0.000
|
0.000
|
-0.495
|
-0.092
|
0.927
|
|
COA
|
-0.032
|
0.053
|
-0.105
|
-0.612
|
0.544
|
|
HLAR
|
0.004
|
0.030
|
0.022
|
0.137
|
0.892
|
|
FD
|
0.000
|
0.001
|
0.509
|
0.095
|
0.925
|
|
2
|
(Constant)
|
2.281
|
7.408
|
|
0.308
|
0.760
|
|
ROA
|
-0.029
|
0.542
|
-0.011
|
-0.054
|
0.957
|
|
TA
|
0.000
|
0.000
|
-1.631
|
-0.300
|
0.766
|
|
COA
|
-0.032
|
0.052
|
-0.104
|
-0.608
|
0.547
|
|
HLAR
|
0.004
|
0.030
|
0.022
|
0.135
|
0.893
|
|
FD
|
0.000
|
0.001
|
1.604
|
0.295
|
0.770
|
|
ROE
|
0.076
|
0.069
|
0.183
|
1.098
|
0.279
|
|
3
|
(Constant)
|
1.297
|
9.584
|
|
0.135
|
0.893
|
|
ROA
|
-0.043
|
0.556
|
-0.017
|
-0.078
|
0.939
|
|
TA
|
0.000
|
0.000
|
-1.176
|
-0.191
|
0.850
|
|
COA
|
-0.029
|
0.055
|
-0.095
|
-0.527
|
0.601
|
|
HLAR
|
0.004
|
0.031
|
0.024
|
0.144
|
0.886
|
|
FD
|
0.000
|
0.001
|
1.148
|
0.186
|
0.854
|
|
ROE
|
0.076
|
0.070
|
0.183
|
1.086
|
0.285
|
|
OSS
|
0.007
|
0.041
|
0.032
|
0.165
|
0.870
|
The entire models have the p-value of F test higher than their
F value. Thus model 1, 2 and 3 of this regression is not statistically
significant. For example, under model 2 the significance level is 0.938 which
is greater than F value 0.289.
Under T test, model 1 gives the following results: T (0.05;
38) is 2.024. When we compare it with the T Student of first regression
coefficient model we observed that it is greater than all T test regression
coefficients. Therefore, we could conclude that all independent variables used
in this model cannot serve to estimate the dependent variable CFIR. Thus
financial performance variables of that model do not influence the Commitment
in Favour of Individual Related. This reflects to H2b: «financially
powerful companies are the worse in terms of SP because of their leaders'
greed, who do not share the margin»
72
Analysis of microfinances' performance and
development of informal institutions in Cameroon
By Djamaman Brice Gaétan
We can notice that this model is similar to the last two models
and therefore have the same conclusion.
? MODEL SUMMARY: CFGR dependent variable
Model1: FD, HLAR, COA, ROA, TA
Model2: FD, HLAR, COA, ROA, TA, ROE Model2: FD, HLAR, COA, ROA,
TA, OSS
|
Model
|
R
|
R
Square
|
Adjusted
R Square
|
Std. Error of the
Estimate
|
Change Statistics
|
|
R Square
Change
|
F
Change
|
df1
|
df2
|
Sig. F
Change
|
|
1
|
0.096
|
0.009
|
-0.118
|
43.22994
|
0.0092
|
0.0722
|
5
|
39
|
0.996
|
|
2
|
0.241
|
0.058
|
-0.091
|
42.69970
|
0.0489
|
1.9746
|
1
|
38
|
0.168
|
|
3
|
0.244
|
0.059
|
-0.118
|
43.24205
|
0.0013
|
0.0528
|
1
|
37
|
0.820
|
* Predictors: (Constant), FD, HLAR, COA, ROA, TA
**Predictors: (Constant), FD, HLAR, COA, ROA, TA, ROE
***Predictors: (Constant), FD, HLAR, COA, ROA, TA, ROE, OSS
Table20: ANOVA OF CFGR REGRESSION
|
Model
|
|
Sum of
Squares
|
df
|
Mean
Square
|
F
|
Sig.
|
|
1
|
Regression
|
674.8437379
|
5
|
134.968748
|
0.07222109
|
0.99603927
|
|
Residual
|
72884.26737
|
39
|
1868.82737
|
|
|
|
Total
|
73559.11111
|
44
|
|
|
|
|
2
|
Regression
|
4275.056303
|
6
|
712.509384
|
0.3907877
|
0.88031162
|
|
Residual
|
69284.05481
|
38
|
1823.2646
|
|
|
|
Total
|
73559.11111
|
44
|
|
|
|
|
3
|
Regression
|
4373.752018
|
7
|
624.821717
|
0.33415167
|
0.93316901
|
|
Residual
|
69185.35909
|
37
|
1869.87457
|
|
|
|
Total
|
73559.11111
|
44
|
|
|
|
Table21: CFGR regression
coefficients
|
Model
|
Unstandardized Coefficients
|
Standardized
Coefficients
|
t
|
Sig.
|
|
1
|
|
B
|
Std. Error
|
Beta
|
|
|
|
(Constant)
|
17.364
|
13.792
|
|
1.259
|
0.216
|
|
ROA
|
-0.210
|
0.987
|
-0.046
|
-0.213
|
0.833
|
|
TA
|
0.000
|
0.000
|
0.034
|
0.006
|
0.995
|
|
COA
|
-0.048
|
0.094
|
-0.094
|
-0.517
|
0.608
|
73
Analysis of microfinances' performance and
development of informal institutions in Cameroon
By Djamaman Brice Gaétan
|
HLAR
|
0.003
|
0.058
|
0.010
|
0.060
|
0.953
|
|
FD
|
0.000
|
0.002
|
-0.097
|
-0.018
|
0.986
|
|
2
|
(Constant)
|
13.919
|
13.841
|
|
1.006
|
0.321
|
|
ROA
|
-0.385
|
0.983
|
-0.083
|
-0.391
|
0.698
|
|
TA
|
0.000
|
0.000
|
-1.398
|
-0.262
|
0.795
|
|
COA
|
-0.049
|
0.093
|
-0.094
|
-0.527
|
0.601
|
|
HLAR
|
0.003
|
0.057
|
0.009
|
0.057
|
0.955
|
|
FD
|
0.001
|
0.002
|
1.285
|
0.241
|
0.811
|
|
ROE
|
0.184
|
0.131
|
0.229
|
1.405
|
0.168
|
|
3
|
(Constant)
|
11.388
|
17.827
|
|
0.639
|
0.527
|
|
ROA
|
-0.415
|
1.005
|
-0.090
|
-0.413
|
0.682
|
|
TA
|
0.000
|
0.000
|
-0.784
|
-0.130
|
0.897
|
|
COA
|
-0.043
|
0.097
|
-0.083
|
-0.438
|
0.664
|
|
HLAR
|
0.004
|
0.058
|
0.011
|
0.070
|
0.944
|
|
FD
|
0.000
|
0.002
|
0.671
|
0.111
|
0.912
|
|
ROE
|
0.185
|
0.133
|
0.230
|
1.391
|
0.172
|
|
OSS
|
0.018
|
0.078
|
0.044
|
0.230
|
0.820
|
Under ANOVA of CFGR we can say that all the models are not
statistically significant because the p-value of F test is higher than their F
value (example of model 1: 0.996 is greater than 0.072). We can notice that
most of the regression coefficients under this model are not significant.
The T test of the first model gives the following results:
T(0.05; 39) is 2.022. When we compare it to the T Student of first regression
coefficient model we observed that it is greater than all the T test regression
coefficients. We can conclude by rejecting the entire hypotheses reflected by
this model. This means that the hypothesis under which financial performance
variables influence the commitment in favour of global related is false ceteris
paribus.
The first model can be generalized to the next two models,
because all T tests of those models are less than the empirical T tests.
| 
