CHAPTER THREE Dnial

CHAPTER THREE
RESEARCH METHODOLOGY
3.1Research Design
In this research, ex-post facto research design shall be employed to examine the
effect of fiscal policy on economic growth. The adoption of ex-post facto
research design is based on the justification that the data for the are
mainly quantitative built on predetermined variables.
3.2Sources of Data Collection
The dataset for this research are time series data, secondarily sourced from
Central Bank of Nigeria (CBN) Statistical Bulletin, spanning from 1986 to
2021. The period are selected to reflect diverse policies initiated in Nigeria to
promote the advancement of credit to different sector of the economy.
3.3Model Specification
In this study, multiple econometric models is employed to model the effect of
bank lending on economic growth in Nigeria. In this research, the empirical
model of Adedigba and Obi (2020) which captured the effect of bank credit for
agriculture, forestry and fishery loan, bank credit for manufacturing loan, bank
credit for mining and quarying loan and bank credit for real estate and
construction loan on economic growth was adapted. As a modification, the
main independent variable employed in this study is aggregated bank credit to
private sector as opposed to the disaggregated variable employed
by Adedigba and Obi (2020). In addition, total deposit and total asset of banks
are used as control variables. Hence, the model for this study is given as:
GDP=f(CPS, FIND,IFR,INR,LNR,MNF)
InGDP = Bo+ B,InCPS + 32InFIND+ B-/nIFR+ Ba/nINR+ Bs/nLNR+
BaInMNF + µi
Where;
InGDP = Gross Domestic Product
InCPS = Bank Credit to Private Sector
InFIND = Financial Deepening
InIFR = Inflation Rate
InINR = Interest Rate
InLNR = Lending Rate
InMNF = Manufacturing Capacity
In= Natural Logarithm
Bo = Intercept
B, - B6 = Coefficients of explanatory variables
µi = Error term
3.4Method of Data Analysis
From the cross-sectional model specification above, we have pooled time

series regression model. In other to estimate the parameters of the model,
Ordinary Least Square { OLS}will employed to estimate the effect of the
independent variables and dependent variable. Ordinary LEAST square
technique is an econometric method that is used to determine and estimate the
parameters of the linear regression model. Among the several methods under
ordinary least squares, this study adopted the error correction model (ECM)
method of analysis.The analysis procedure which were followed by this study
are stated below:
3.4.1 Pre-estimation Test
The pre-estimation tests include descriptive statistics test, correlation

test and unit root test. Descriptive statistics test provides tips as to the
statistical behavior of the variables of study particularly on the mean, standard
deviation, kurtosis, normality of the series, among others. Correlation test
based on Pearson test is also applied to the time series to ascertain the nature of
relationship between variables of interest. Unit root test, the estimation of
variable-series that are non stationary may lead to estimates that are spurious
and thus render the coefficients unreliable for policy prescription and usage.
Hence investigation was conducted on conventional unit root tests on each of
the variables used in this analysis.
Furthermore, to avoid spurious regression results and also as a key test that

should precede the estimation of model,this study applied the Augmented
Dickey-Fuller (ADF) test in evaluating the stationarity property of each of the
variables of study.Co-integration is the test of long run relationship among
variables of study. The co-integration test is performed using the F-Bounds test
of co-integration considered suitable for the I(1) and I(0) times series.Co
integration is then the link between integrated process and steady state
equlibirum. Nonstationary series are said to be cointegrated if a linear
combination of the series result in a stationary time series.
3.4.2 Estimation Techniques

Autoregressive Distributed Lag (ARDL) model
Autoregressive Distributed Lag (ARDL) model is estimated to determine the
effect of credit to private sector, total deposit and total asset on gross domestic
product.
GDP, = do + L'-10,GDP.-I + 2P -102 CPS-1 + 2P -103 FIND-I + LPi-LNR +-
1+2P+-INR-I+ Z'i-106IFR+1+ InLNR+-1+ InINR
Where;
InGDP = Gross Domestic Product

InCPS = Bank Credit to Private Sector
InFIND = Financial Deepening
InIFR = Inflation Rate
InINR = Interest Rate
InLNR = Lending Rate
Granger Causality Test
Granger causality test is performed to estimate the direction of causality among

the variables. Furthermore, in order to determine whether a specific variable or
group of variables play any roles in the determination of other variables
process, Pairwise Granger Causality approach as employed.
Diagnostics Test
The reliability and robustness of the estimated result were tested by conducting
a diagnostic test on the model using Jargue-Bera Statistic to test the normality
of the variables, Breusch-Godfrey Serial Correlation Test, Breusch-Pagan
Heteroscedaticity and Stability test using Cusum stability test.
3.4.3 Post estimation Techniques

Auto correlation Test
One of the assumptions of the classical least square techinque is the assumption
of no autocorrelation. Autocorrelation may be defined as “ correlation between
members of series of observation ordered in time or space” ( Gujarati 2003).
The classical linear regression assumes that such correlation does not exist in
the residual term t that is the error terms of different observations ( j j) are
dependent
Residual .normality test
Normally test is carried out to check whether the error term follows the normal
distribution. The normality test adopted is the Jarque-Bera (JB) Test of
normality. The JB test of normality is an asymptotic, or large- sample, test and
it is based on the OLs residuals and uses the chi-square distribution (Gujaratii,
2004:148).
Test for heteroskedasticity
This test is basically focused on the variance of the error term. The test helps
to ascertain whether the variance of the error term is constant.
Ho: Homoskedasticity
Hi: Heteroskedasticity
Decision Rule: Reject Ho if x2> x2 0.05 and accept if otherwise.

CHAPTER THREE Dnial

Uploaded by

Copyright:

Available Formats

CHAPTER THREE Dnial

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CHAPTER THREE Dnial

Uploaded by

Copyright:

Available Formats

3.4Method of Data Analysis

From the cross-sectional model specification above, we have pooled time

3.4.1 Pre-estimation Test

The pre-estimation tests include descriptive statistics test, correlation

Furthermore, to avoid spurious regression results and also as a key test that

3.4.2 Estimation Techniques

GDP, = do + L'-10,GDP.-I + 2P -102 CPS-1 + 2P -103 FIND-I + LPi-LNR +-

1+2P+-INR-I+ Z'i-106IFR+1+ InLNR+-1+ InINR

InGDP = Gross Domestic Product

Granger causality test is performed to estimate the direction of causality among

3.4.3 Post estimation Techniques

You might also like