CHAPTER THREE

METHODLOGY

3.1 Research Design

Research design means the structuring of investigation aimed at identifying variables and

their relationships to one another. This is used for the purpose of obtaining data to enable the

researcher test hypothesis or answer research questions (Nnamdi, 2006). The study will made

use of correlational research design. The purpose of correlational research is to determine the

relations among two or more variables. The design is appropriate because of its convenience

in the study and it predicts a statistical relationship between two or more variables such that

systematic changes in the value of one variable are accompanied by systematic changes in the

other.

3.2 Sources of Data

Secondary source of data will be used in the study because of the nature of the study which is

an analysis of the contribution of the agricultural Credit Guarantee Scheme (ACGS) to

poverty reduction. The time series data cover 24 years ranging from 1996-2020. The purpose

of choosing this period is to empirically test the significance or the extent to which

agricultural credit guarantee scheme contributes to poverty reduction in Nigeria after the

sector suffered several years of neglect, and the renewal of effort by government towards

stabilizing the agricultural sector through the scheme. The data were obtained from various

publications of Central Bank of Nigeria (CBN).

3.3 Model Specification

According to Gujarat (as cited by Udonsah, 2012) an econometric investigation begins with

the specification of the econometric model underlying the phenomenon of interest. Also,
Asogwa (as cited by Udonsah, 2012) posit that specification of a model generally is a

function of the theoretical relationship between or among variables, the nature of study

objectives and type of data. This has to do with expressing the model in mathematical and

econometric form which would be used to explore the economic phenomenon. As earlier

mentioned, our independent variables are derived from the credit guarantee funds under the

Agricultural Credit Guarantee Scheme (ACGS) in Nigeria and poverty reduction. Our general

model is thus stated as follows:

Y = f (X1, X2, X3, X4)

Where: Y = Poverty Reduction (PR)

X1 = ACGS loans to individual farmers (LIF)

X2 = ACGS loans to informal groups (LIG)

X3 = ACGS loans to cooperatives (LCO)

X4 = ACGS loans to company (LCY)

The linear form of the regression model is expressed as:

Y = β o+ β 1 X 1+ β 2 X 2+ β 3 X 3+ β 4 X 4+ μ

Where: Y = Poverty Reduction (PR)

β1, β2…β4 = Independent variables coefficient or parameters for LIF,

LIG, LCO, LCY.

β o = The intercept which represents the expected value of the dependent variable (GDP)

when all the independent variables assumed zero as value.

μ = Random or stochastic error term

In order to capture the link between the explanatory variables (LIF), (LIG), (LCO), (LCY)

and their impact on the dependent variable (GDP), the following model is specified to

empirically explain the effect of the independent variables on Poverty reduction (PR).

PHC=β o+ β 1 LIF + β 2 LIG+ β 3 LCO+ β 4 LCY + μ

Where: PHC = Poverty head count used as a proxy for Poverty Reduction

LIF = Loan under ACGS to Individual farmers

LIG = Loan under ACGS to Informal Group

LCO = Loan under ACGS to Co-operative

LCY = Loan under ACGS to Company

β 1 , β 2 , β 3 , β 4 = Independent variables coefficient, that is, the parameters for LIF, LIG,

LCO, and LCY.

β o = Intercept of GDP when the explanatory variables are equal to zero. β1, β2, β3, β4 =

coefficients of explanatory variables.

3.4 Techniques of Data Analysis

3.4.1 Unit Root Test

Since this study deals with time series variables, there is need to test for unit root in each of

the variables employed. The importance of this derives from the fact that estimation in the

presence of non–stationarity in variables usually leads to biased and inconsistent estimates of

the standard errors of the coefficients and this could lead to misleading inference if

appropriate technique is not applied to overcome the problem. The unit root tests will be

carried out using the Augmented Dickey–Fuller (ADF).

3.4.2 Ordinary Least Square (OLS)

The method of data analysis is the Ordinary Least Square (OLS) Multiple Regression Model

(MRM). Data estimation will be carried out using Statistical Package of Social Sciences

(SPSS), version 23 and results will be analyzed using the values of the Pearson correlation

coefficients, coefficient of determination (R2), and the coefficients of the independent

variables – the t-test or t-statistics.

 CHAPTER FOUR

DATA PRESENTATION, ANALYSIS AND DICUSSION

1.1 Data Presentation

Data presented in this chapter are secondary source of data. Secondary source of data was

used in the study because of the nature of the study which is an analysis of the impact of

agricultural credit guarantee scheme on poverty reduction in Nigeria. Data that had been

generated are required for this type of study. The time series data cover 24 years ranging

from 1996-2020. The purpose of choosing this period is to empirically test the significance or

the extent to which of agricultural credit guarantee scheme contributes to poverty reduction in

Nigeria. The data were obtained from various publications of Central Bank of Nigeria (CBN).

The data used for the analyses is provided in Table 4.1) which contains variables as used in

the estimation.

Table 4.1: Cumulative Loans Guaranteed under ACGS Operations from 1996 to 2020 -
Category Basis [AMT: =N= ('000)]

Loans under Agricultural Credit Guarantee Scheme (ACGS) to:

Poverty
Head Informal
Count Groups Cooperatives Company
Year (Y) Individual Farmers (LIF) X1 (LIG0 X2 (LCO) X3 (LCY) X4
1996 0.36 86,213.30 - 5,933.50 6,347.60
1997 0.37 69,932.90 - 5,338.70 6,835.80
1998 0.33 75,549.10 - 6,858.30 5,624.40
1999 0.34 68,257.90 - 9,048.50 3,538.40
2000 0.62 86,451.90 - 9,000.10 7,734.00
2001 0.68 132,778.30 - 19,285.80 12,098.00
2002 0.73 179,824.70 - 34,425.40 11,252.40
2003 0.77 184,686.00 7,516.50 34,379.00 15,456.70
2004 0.77 190,305.20 1,705.00 8,960.00 14,727.00
2005 0.76 193,501.00 1,351.00 42,325.50 8,905.00
2006 0.69 324,187.40 9,995.00 22,928.00 4,340.00
2007 0.65 727,945.40 - 100 500
 2008 0.64 1,025,575.80 10,594.00 14,170.00 1,250.00
2009 0.50 1,106,456.40 30,774.00 16,230.00 11,000.00
2010 0.58 2,017,344.70 21,180.00 31,620.00 13,600.00
2011 0.59 2,969,096.70 20,036.50 38,537.30 19,068.00
2012 0.59 3,984,895.50 82,661.00 171,963.80 23,540.00
2013 0.57 4,145,410.50 228,200.00 27,751.30 24,500.00
2014 0.57 6,157,288.40 289,219.00 165,475.00 109,092.20
2015 0.55 7,495,288.30 128,674.00 586,992.00 138,555.00
2016 0.67 7,370,945.60 43,274.00 249,703.10 76,585.00
2017 0.78 9,375,403.90 384,641.00 305,171.40 124,388.00
2018 0.46 9,128,295.30 27,987.00 267,309.50 283,169.30
2019 0.41 8,673,464.90 181,535.00 372,810.00 196,640.00
2020 0.43 9,977,834.20 160,872.00 400,860.00 130,600.00
Source: Development Finance Department, Central Bank of Nigeria (CBN) 2021, Poverty head count
- sources: World Bank, Global Poverty Working Group. Data are compiled from official government
sources/computed from World Bank staff using national i.e. country-specific poverty lines.

4.1. 1 Descriptive statistics of the variables

This sub-section provides a descriptive analysis to describe the main characteristics of the

data utilized for the study. The summary of descriptive statistics includes the median, mean,

minimum, maximum, standard deviation, and observations in the analysis. The summary

statistics for all the variables utilized in the impact of agricultural credit guarantee scheme on

poverty reduction in Nigeria, spanning from 1996 to 2020. Before embarking on the details of

empirical issues, it is important to examine whether the data exhibits normality. Most

economic data is skewed (non-normal), possibly due to the fact that economic data has a clear

floor but no definite ceiling. Also, it could be because of the presence of outliers. The Jarque-

bera statistics test is used to test normality of the series. It utilizes the mean based coefficients

of skewness and kurtosis to check normality of the variables used. Skewness is the tilt in the

distribution and should be within the -2 and +2 range for normally distributed series. Kurtosis

put simply is the peakedness of a distribution and should be within -3 and +3 range when data

is normally distributed. Normality test uses the null hypothesis of normality against
alternative hypothesis of non-normality. If the probability value is less than Jarque-Bera chi-

square at the 5% level of significance, the null hypothesis is not rejected

Table 4.2: Descriptive statistics of variables of the model of the study

PHC(Y) LIF(X1) LIGO(X2) LCO(X3) LCY(X4)
Mean 1.221676 -0.633200 0.939948 0.845520 0.876273
Median 1.318261 -0.690000 0.995854 0.521085 0.975085
Maximum 1.719428 0.040000 1.553387 2.737730 1.704583
Minimum 0.000000 -1.700000 0.000000 -0.312436 -0.498655
Std. Dev. 0.365154 0.441881 0.292207 0.915370 0.514602
Skewness -2.334117 -0.060088 -1.142687 0.755947 -0.636368
Kurtosis 8.350558 2.001381 5.425245 2.195136 2.754893
Jarque-Bera 262.6088 5.269179 57.83725 15.27932 8.749665
Probability 0.000000 0.071748 0.000000 0.000481 0.012590

Sum 152.7095 -79.15000 117.4935 105.6900 109.5341

Sum Sq. Dev. 16.53383 24.21212 10.58776 103.9000 32.83707
Observations 24 24 24 24 24
Source: Researcher’s Compilation, 2023.

The distribution properties of the time-series data were analyzed using summary statistics

including mean, median and standard deviation. Skewness, kurtosis and Jacque-Bera

statistics were also used to examine the normality of the data.

1.2 Analyses of Data

The empirical results obtained from the regression estimation were interpreted using; the

Pearson correlation coefficient which serves to measure the strength of linear relationship

between variables, the t-test coefficient of the independent variables which attest to the

individual significance of the independent variables and the coefficient of determination,

otherwise referred to as adjusted R square (R 2).The respective results are shown and analyzed

below.
Table 4.3 Pearson Correlation Coefficient Matrix of Variables
PHC LIF LIG LCO LCY

PHC 1.00

LIF .901 1.00

LIG .630 .755 1.00

LCO .805 .895 .614 1.00

LCY .768 .855 .521 .792 1.00

The above table shows that the correlation coefficients of all the variables are high. This

indicates that there exists a strong linear relationship between them. In other words, there is

the existence of strong linear relationship between PHC and the independent variables;

ACGS loan to individual farmers (LIF), Informal Group (LIG), Co-operative (LCO) and

Company (LCY). Also, there exists a strong relationship between the independent variables

(LIF, LIG, LCO and LCY). The least between the dependent and independent variables is

that of PHC and ACGS loan to Individual Group (LIG) which stood at .630; and between the

independent variables is LIG and LCY which stood at .521; this represent a linear

relationship of 63% and 52.1% between PHC and independent variables; and between

independent variables respectively.

On the other hand, the highest linear relationship viz-a-viz PHC and the independent

variables is .901, indicating a 90.1% linear relationship. This relatively high correlation

coefficients between the variables above, indicate that a strong linear relationship exist

between ACGS loans and Poverty reduction (PHC) in Nigeria.

4.3 Test of Hypothesis

The regression results were analyzed using the t-test coefficients which attests

to the significance of each of the independent variables and the coefficient of

determination, otherwise referred to as adjusted R square (R 2) which measures

the proportion of variation explained by the independent variables in the

regression model. These empirical results which were employed to explain and

test the Null hypothesis formulated are presented in Tables 4.4, and 4.5 below.

Table 4: 4 Coefficients a
Model Unstandardized Coefficients Unstandardized T Sig.
Coefficients
B Std. Error Beta
Constant 68625.97 2079720.706 0.33 0.974
LIF 7.097 1.776 1.103 3.996 0.000
LIG -32.47 30.308 -0.141 -1.071 0.293
LOC -5.891 27.114 -0.04 -0.217 0.83
LCY -23.939 55.821 -0.07 -0.429 0.671
Source: Researcher’s computation using IBM SPSS Statistics 20; 2023

Table 4.5 Model Summaryb

Mode R R Square Adjusted Std. Error of the Durbin-
R Square Estimate

l Watson
1 .905a 0.818 0.793 9837900.706 0.856
64104800000
a. Predictors: (Constant), LCY, LIG, LOC, LIF
b. Dependent Variable: GDP
Source: Researcher’s computation using IBM SPSS Statistics 20; 2023

4.4 Discussion of Findings

The empirical results obtained from our regression estimation shows that the coefficient of

ACGS loan to individual farmers (LIF) which stood at 3.996 is the best in the model, which

indicates that ACGS loan to individual farmers impacted significantly on poverty reduction

(PHC). The positive sign of the LIF coefficient is in conformity with apriori expectation

which explains that ACGS loan to individual farmers is required and plays an important role

in the production of agricultural products (goods and services) both for subsistence and

commercial purposes respectfully. This impact may not be unconnected with the

concentration of ACGS loans on individual farmers; that is, the cumulative size of loans

received within the period of study (92.7% out of 100%). In this regard, this finding agrees

with the conclusion of Nwankwo (2018) that the size of loans to the agricultural sector has

significant impact on poverty reduction (GDP) in Nigeria.

On the other hand, the coefficient of the variables of ACGS loan to Informal Groups (LIG),

Co-operatives (LCO), and Companies (LCY) from the empirical result showed negative signs

of -1.071, -.217 and -.429 respectively. This evidence with respect to these variables did not

conform to apriori expectation, which means that they do not have significant impact on

poverty alleviation. It appears these results may be reflecting the negative impact of

underfunding or lack of appropriate funding of ACGS loan to the affected categories.

Olaitan (2016) reiterated that the lack of access to credit (finance) impedes growth amongst

farmers and raise poverty levels in Nigeria, making them to be endangered species. He

therefore, called for transformative efforts to address the problem. A large proportion of the

rural population depends on agriculture for their main source of sustenance and livelihood,

yet the supplies of credit still leave a wide gap of rural access to finance. This means that, the

lack of access to finance create poverty and socio-economic problem for agricultural
development; which affect the overall growth of the economy. The ACGS fund were

provided to make access to finance much easier and accessible, since it guarantees credit

facilities from the bank to farmers at 75 percent of total fund borrowed without any security,

which is expected to alleviate poverty and contribute meaningfully to improve the livelihoods

of farmers (all categories; whether subsistence or commercial), and emerging entrepreneurs

in the agriculture sector. It is pertinent to state that this expectation is yet to be fully realistic

judging from the underfunding experienced by the three categories in the agricultural sector.

Generally, the regression analysis results appear to suggest that ACGS loans to Informal

groups (LIG), Co-operatives (LCO) and Companies (LCY) are not relevant in formulating

policies that may impact significantly on poverty reduction in Nigeria. Though the three

variables from the analysis appear not to have significant impact, but obviously show the

effect of the small size of loans received (7.3% out of 100%) within the period (1996-2020).

This result therefore agrees with conclusion of Nwankwo 2013, that the size of loan

influences the level of impact a variable can have on economic growth.

However, the coefficient of determination otherwise known as R square (R 2) which gives

information about the goodness of fit of the model stood at .818. This indicates that the

explanatory variables were able to explain 81.8% of the dependent variable; approximately,

the explanatory variables such as LIF, LIG, LCO and LCY explained 82% of the dependent

variable (GDP).