Descriptive statistics:

Descriptive statistics provide a summary of data set for an explanation of basic features, general
pattern, and trend. It consists of mean, median, maximum, minimum, standard deviation,
skewness, and kurtosis. The details of descriptive statistics are showed in Table 5.1. It showed a
descriptive statistic of the dependent variable and independent variables in Janata Bank from 2010
to 2016.

Table: Descriptive Statistics

GDP INT INF EPS ROA ROE INV2DEP

Mean 1.840297 1.718146 1.988419 3.549934 -0.29284 2.61904 3.475668
Median 1.840045 1.705656 1.983066 3.497753 -0.30902 2.532259 3.616967
Maximum 1.960095 1.930071 2.335052 4.457945 0.350657 3.404193 3.698582
Minimum 1.722767 1.556037 1.708378 2.610805 -1.10866 1.652497 2.999226
Std. Dev. 0.083337 0.131185 0.219451 0.728691 0.507992 0.681577 0.27943
Skewness 0.031133 0.458025 0.340557 0.031952 -0.36423 -0.10161 -0.945
Kurtosis 2.035782 2.287034 2.223388 1.532225 2.31857 1.703519 2.268681

Jarque- 0.233398 0.336867 0.266761 0.539612 0.248752 0.430539 1.026734

Bera
0.889853 0.844987 0.875132 0.763528 0.883048 0.806324 0.598477
Probability

Sum 11.04178 10.30888 11.93052 21.29961 -1.75701 15.71424 20.85401

Sum Sq. 0.034726 0.086048 0.240793 2.65495 1.290278 2.322734 0.390405
Dev.

*ROA (Dep) = Return on Asset (Dependent Variables). *INF= Inflation. * INT=Interest. * ROE=
Return on Equity. * INV2DEP= Investment to Deposit ratio. * GDP= Gross domestic Product.
*EPS= Earnings per Share.

From this Table, it showed that the highest mean is Earnings per Share and the lowest is Return
on Asset which is in negative (0.29284). This result indicated that Banks EPS is at very high level.
Besides that, Return on Asset in Janata Bank is average. For the median, Investment to Deposit
ratio has the highest median (3.616967) while Return on Asset has the lowest median negative
(0.30902). Regarding maximum point, EPS has the highest maximum point (4.457945) while
Return on Asset has the lowest maximum point (0.350657). Due to a high level in Banks
Investment to deposit ratio, it has the highest minimum point (2.999226) while ROA has the lowest
minimum point negative (1.10866). For standard deviation, EPS has the highest standard deviation
(0.728691) while GDP has the lowest standard deviation (0.083337). Among all variables, all show
positive skewness except ROA, ROE and Investment to deposit ratio. ROA has the highest kurtosis
(2.31857) while EPS has the lowest and negative kurtosis (1.532225) according to Table.

Normality Test:
By using the values of the Jarque-Bera and the Probability of the Descriptive statistics we can
describe the normality of the distribution of the information. For this, we need to know the
following hypothesis:

Null Hypothesis : Normal Distribution.

Alternative Hypothesis : Not Normal.

JarqueBera test is a goodness-of-fit test of whether sample data have the skewness and kurtosis
matching a normal distribution. All the values given in the Table explains the fitness of these
variables with the normal distribution.

Table: Normality Test

Variables Jarque-Bera Probability

GDP 0.233398 0.889853
INT 0.336867 0.844987
INF 0.266761 0.875132
EPS 0.539612 0.763528
ROA 0.248752 0.883048
ROE 0.430539 0.806324
INV2DEP 1.026734 0.598477
As the probability of all the variables are more than 5% so we cant reject the null hypothesis rather
we accept the null hypothesis and reject the alternative hypothesis. This means that our distribution
of information are normal.

Unit Root Test:

For unit root test, The Augmented Dicky-Fuller (ADF) test will be carried out to test on the
stationary of dependent and independent variables. The results of ADF for the variables are shown
in the table below. The hypothesis for ADF are:

H0 : Series is non-stationary

H1 : Series is stationary

The null hypothesis will be rejected if the p-value is lesser than , otherwise, we do not reject the
null hypothesis.

Table: ADF stationary test on dependent variable and independent variables at level:

Series ADF t statistics P value

ROA -3.077096 0.0814
ROE -2.668182 0.1379
EPS -2.629883 0.1432
INV2DEP -1.796377 0.3488
GDP -1.599606 0.4250
INF -0.417740 0.8425
INT -1.741415 0.3686

The table above showing the result of Unit Root test in E-views where all the variables are non-
stationary including intercept at level because their absolute test statistics is smaller than the
critical value and probability is greater than 5% thats why we cannot reject the null hypothesis.
Table: ADF stationary test on the dependent variable and independent variable at First difference:

Series ADF t statistics P value

ROA -3.781746 0.0284
ROE -4.088469 0.0463
EPS -4.113418 0.0454
INV2DEP -5.528366 0.0158
GDP -6.113801 0.0107
INF -5.925636 0.0022
INT -6.147677 0.0117

But, this table showing that all the variables are stationary including intercept at 1st difference
because their absolute test statistics is greater than the critical value and probability is smaller than
5% thats why we can reject the null hypothesis.

As all the variables are non-stationary at level and stationary at the 1st difference, we can use these
data for our next procedure of the analysis.

Regression Analysis:
The following table reports the results from the regression equation linking the dependent and
independent variables. The F -value for each year in the first model is significant at the 5% level
and the adjusted R square for each of the five years is 94% for the period of 2006-2016 which
means that the data has fitted for the tests. Regression coefficients represent the mean change in
the response variable for one unit of change in the predictor variable while holding other predictors
in the model constant. This statistical control that regression provides is important because it
isolates the role of one variable from all of the others in the model.
Table: Regression Analysis output.

Dependent Variable: ROA

Variable Coefficient Std. Error t-Statistic Prob.

C -3.734706 1.853779 -2.014645 0.0142

ROE 0.019316 0.044845 0.430736 0.0489
EPS 0.015628 0.016632 0.939620 0.0306
INV2DEP 0.017594 0.015853 1.109784 0.4293
GDP 0.331101 0.205288 1.612861 0.1821
INT 0.112868 0.189834 0.594564 0.0341
INF -0.030217 0.078316 0.385832 0.0193

R-squared 0.976351 Mean dependent var 0.370909

Adjusted R-squared 0.940879 S.D. dependent var 1.315481
S.E. of regression 0.319858 Akaike info criterion 0.819244
Sum squared resid 0.409235 Schwarz criterion 1.072450
Log likelihood 2.494158 Hannan-Quinn criter. 0.659633
F-statistic 27.52395 Durbin-Watson stat 2.421854
Prob(F-statistic) 0.003251

ROE is showing positive and significant relationship with the dependent variables. The later
relationship suggest that Return on Asset has positive and significant relationship with ROE. EPS
and Investment to Deposit ratio also showing positive and significant result with the dependent
variables. This means that bank earns more with the higher EPS and more Investment. GDP is
showing positive but insignificant relation with the dependent variable which says that increase in
GDP increases the ROA but it is not significant with the model. Interest has significant positive
relation with ROA and Inflation has significant negative relationship with ROA which implies that
during inflation banks earn less.

Multicollinearity test:
Multicollinearity is a phenomenon in which one predictor variables in a multiple regression model
can be linearly predicted from the others with a substantial degree of accuracy. In this situation the
coefficient estimates of the multiple regression may change erratically in response to small changes
in the model or the data.

To check the multicollinearity, we need to run correlation test only on the independent variables
excluding the dependent variable. And the output of correlation is:
Table: Correlation Analysis

ROE EPS INV2DEP GDP INT INF

ROE 1.000000 0.992748 -0.018417 -0.332456 -0.014174 -0.044556
EPS 0.992748 1.000000 0.027691 -0.334115 0.026893 -0.138246
INV2DEP -0.018417 0.027691 1.000000 0.148068 0.459118 -0.393655
GDP -0.332456 -0.334115 0.148068 1.000000 -0.028983 0.079439
INT -0.014174 0.026893 0.459118 -0.028983 1.000000 -0.462977
INF -0.044556 -0.138246 -0.393655 0.079439 -0.462977 1.000000

We know that multicollinearity creates problem when the correlation exceeds 0.80 and here it is
showing that all the values are less than 0.80 except ROE and EPS. They have high correlation
between them. To adjust this problem we need to exclude one variable. This can be determined by
the probability of both. We will exclude which have higher p-value. And the regression model
shows us ROE has higher p-value (0.0489) than the EPS (0.0306). So, we will skip ROE for the
betterment of the test and solve the multicollinearity problem.