Determinants of Loan Acquisition and Utilization among Smallholder Rice Producers in Lagos State, Nigeria

Abstract

Agriculture is a major contributor to economic development in most developing nations, with smallholder farmers playing a critical role, but their productivity and growth are impeded by a lack of access to agricultural loans. The affordability of loans is critical for sustainable agricultural development. Therefore, this paper investigates farmers’ loan acquisition and utilization, as well as their choice of loan sources using a two-part model and multinomial logit model. A total of 281 smallholder rice farmers were surveyed in Lagos State, Nigeria. The empirical findings show that marital status, farm size, and interest rate were all positive and significant influences on farmers’ loan sources of choice. In addition, annual farm revenue and the interest rate have a significant positive impact on loan access, whereas education, farming experience, farm size, off-farm income, and farm income have a major impact on loan use. The study also reveals that the preferred loan source differs according to the characteristics of farm households. The study concluded that access to loans increases farmers’ income in the region. It was recommended that the socio-economic characteristics of rice farmers should be considered when formulating and implementing policies to improve smallholder farmers’ access to agricultural loans.

1. Introduction

In most developing countries, agricultural credit is the wheel on which agricultural development is based. It is frequently employed to nudge agricultural development in the right way. The agriculture industry uses credit more than any other sector of the economy because of seasonal changes in farmer returns and a shift from subsistence to commercial farming. As a result, introducing accessible and affordable financing is the quickest approach to enhance sustainable agricultural development and livelihood for smallholder farmers in rural parts of Sub-Saharan Africa. Agricultural credit acquisition and use is an important facilitator of the sustainable development of the farming sector [1]. Access to credit will help to develop rural economies, improve rural farmers’ socio-economic conditions, and promote agricultural sustainability, thus addressing the problem of insufficient food, which has become a global trend and a challenge to the world’s population in recent years [2,3]. In the same vein, according to [4], access to agricultural credit encourages the implementation of technologies and increases farm productivity; revenue promotes the formation of capital and improves marketing efficiency. Furthermore, it helps smallholder farmers in purchasing inputs, hiring labor, acquiring equipment, and developing seed types in order to boost agricultural output and food security.

Rice is a staple food crop with a prime income source and employment for smallholder farmers within the agricultural sector in Nigeria. According to Oyaniran [5], the agricultural industry employs 36 percent of Nigeria’s workers and generated around 22 percent of the country’s GDP in the first quarter of 2020. Smallholder farmers account for 80% of Nigerian farmers and contribute over 90% of the country’s agricultural output. The production of rice is mainly in the hands of the smallholder farmers, whose productivity and growth are hindered by limited access to loan facilities, and domestic production has not increased to meet the economy’s demand [6]. Farmers’ productivity and revenue have suffered as a result of the limited availability of finance, which has also hindered them from adopting contemporary technologies and inputs in their farming operations. The provision of credit in the modern farming sector in Nigeria is insufficient, but the efficient use of such credit is a critical aspect in increasing production. Numerous programs have been established by the Nigerian government in an attempt to help smallholder farmers have more access to credit facilities, such as the Anchor Borrower Scheme, Nigeria Incentive-Based Risk Sharing for Agricultural Lending (NIRSAL), and Agricultural Input subsidies to boost farm operation. Despite policies aimed at improving access to formal financial institutions in rural areas in order to promote agricultural production, the problem persists.

There are vast numbers of empirical studies on farmers’ access to credit from financial institutions in many developing countries. For example, Kuwornu et al. [7] conducted research in Ghana to analyze selected maize farmers’ agricultural credit allocation and restriction analyses. The paired sample was used to test for major discrepancies between the amount of credit needed and the amount of credit earned by farmers, showing that the amount of credit received was substantially lower than the amount of credit requested by farmers. The authors examined the determinants of farmers’ credit constraints, and the results revealed that gender, farmers’ household size, farmers’ annual income and farm size had a major effect on farmers’ credit constraints. Additionally, the determinants of the rate of agricultural credit allocated to the farm sector were assessed, and the results showed that age, pre-credit bank visits, and the amount (size) of credit earned had a major effect on the agricultural credit allocation rate for the agricultural sector. Isitor et al. [8] investigated credit utilization and farm income of arable crops farmers in Kwara State, Nigeria, and the empirical results showed that hired labour, cooperative participation, awareness of credit source, past loan size and possession of collateral were significant and positive determinants that influenced farmers’ utilization of agricultural credit, while household size and distance away from the credit source were significant and negative determinants of farmers’ decisions to utilize loans.

Christopher et al. [9] conducted a study in Zambia to ascertain the determinants of smallholder farmers’ access to agricultural finance and found that the degree of education of the head of the household, the size of the household and the number of regular meals had a substantial effect on the decision to access funding. Gideon et al. [10] studied the use of agricultural credit among farmers in the northern area of Bole district, Ghana. The authors observed that gender, household size, farmers involved in off-farm income, and farmer-based organization membership are important factors influencing farmers’ access to agricultural loans. In Benue State, Nigeria, [6] conducted a study to identify the socio-economic factors affecting the acquisition of agricultural loans among small-scale rice farmers. The results showed that age, household size, education, farm size, cooperative membership and annual income were significant factors influencing the probability of access to loans for farmers.

Amanuel and Degye [11] examined the determinants of smallholder farmers’ use of microfinance loans in the Lemo district of the Hadiya area, Southern Ethiopia: the case of Omo microfinance. The researcher observed that sex, literacy status, level of income, level of savings, the intent of loan taking and perception of loan repayment were significant factors influencing the number of loans obtained. Additionally, the findings showed that literacy status, household size, and land ownership size were significant factors affecting the use of loans in the study area by smallholder farmers. Shahab et al. [12] conducted a study to identify factors in flood-prone areas of Pakistan that decide the access of subsistence farmers to agricultural credit. The study revealed that education, farming experience, total landholding, monthly income, family size, and proportion of owned land were important factors in farmers’ access to agricultural credit in Pakistan’s flood-hit areas. A study was conducted by Liqiong et al. [13] in China to define rural credit constraints and informal accessibility to rural credit. The results showed that age, family size, annual non-agricultural household income, educational level, and history of informal borrowing had a major effect on credit constraints.

However, in view of previous research, there is no study that has systematically investigated this phenomenon in Lagos State, Nigeria. Hence, little is known about loan acquisition and utilization among smallholder rice farmers in Lagos. This research is timely, especially at this moment, when the government of Nigeria is developing the country’s agricultural industry. In light of this, the goal of this research is to fill a critical knowledge gap by empirically investigating the factors that influence smallholder rice farmers’ access to and utilization of loans in Lagos State, Nigeria by employing a two-part model approach, which is not common in the agricultural economics literature.

2. Conceptual Framework

2.1. Concept of Loan/Credit

The term loan refers to money received from a friend, relative, cooperative, bank or financial institution in exchange for future repayment of the principal with interest. Credit is defined as such assistance given to farmers either in cash, kind or both for the purpose of agricultural production, the repayment of which the beneficiaries are expected to make at a further date with or without an interest rate [14]. Credit was defined as the process of gaining control over the use of cash, goods, and services by Adegeye and Dittoh [15] in return for a promise to repay at a future date. Credit serves as a catalyst, motivates other development factors, and helps farmers achieve economies of scale [16]. Agricultural credit is an essential mechanism not only for fostering agricultural development but improving efficiency and increasing production in the right direction [16].

2.2. Sources of Agriculture Loan Facilities in Nigeria

In Nigeria, the major sources of loan finance available to smallholder farmers are generally classified into three categories, namely formal, semi-formal, and informal sources. Formal sources are those established by law. They can be influenced by government policies and consist of agricultural and commercial banks, such as the Nigerian Agricultural Bank (NAB), the Nigerian Industrial Development Bank (NIDB), the state government-owned credit institutions and micro finance institutions (MFIs), private and merchant banks, and finance houses. Formal credit is assumed to be the most effective source from the point of view of overall agricultural development, although it is associated with the problem of inaccessibility by small-scale farmers [17]. Semi-formal sources of credit include microfinance institutions or NGOs, cooperative societies and support groups, farmers’ associations and the rotating savings and credit associations (ROSCAs). Informal sources and institutions can include moneylenders, informal credit associations such as Esusu and Ajo, relatives, and friends. The informal credit sources give loans in cash or in kind to farmers to be reimbursed in cash or kind, often in agricultural produce. Loan accessed from these sources does not usually require a deposit, and no collateral is required [18].

2.3. Determinants of Loan Access

Numerous socio-economic factors play an important role in deciding individual smallholder farmers’ demand for loans. First, it is a choice that may be affected by variables such as age, gender, marital status, education, size of the farm, size of the household, party membership, and level of income. Secondly, the price or cost of the product (commodity) is a factor. Thirdly, the borrower’s preference among the alternatives available plays a role. According to the theory of demand for goods and services and prices, the purchasing decisions of consumers and the quantity purchased are associated with the prices of commodities demanded [18].

In other words, when deciding to borrow, an individual farmer looks at the cost of the loan, the available alternatives, the conditions of borrowing from alternatives, and the socio-economic characteristics of borrowers. These put together give bases for consideration to borrow from the alternative sources or not. For example, if borrowing from formal credit sources proves expensive, borrowers are likely to turn to informal sources, and vice versa. This is simply on the basis that if the cost of credit goes up, the marginal utility per unit price raised from that credit goes down. The borrower, therefore, chooses to either not consume or consume less of the credit. The concept of utility and marginal utility explains consumer demand for a commodity. The main objective of any borrower is to maximize satisfaction out of any finances borrowed, given or self-made. As such, the conditions of accessing credit from the alternative markets are taken into consideration before deciding to borrow/access a loan. Figure 1 shows the conceptual framework of farmers accessing loans.

3. Materials and Methods

3.1. Study Area

The study was carried out in Lagos State, Nigeria, between September and December 2019. The state is one of Nigeria’s 36 states, and it is located in the country’s southwest. The state is divided into 20 local governments (LGs) and 37 local development zones. It lies between latitudes of 602′ N and 604′ N and longitudes of 2045/E and 4020/E. The state covers 351,861 hectares of land.

A cross-sectional survey design was used in this research. The study adopted a multi-stage sampling technique. First, we purposefully selected the three local government areas (i.e., Badagry, Epe, and Ibeju-Lekki) that are the major rice-producing areas in Lagos. The second stage involved selecting two communities (villages) at random from each of the three LGAs, totaling six villages (i.e., Itoga, Ganyingbo, Itoikin, Ise/Igbogun, Ibeju, and Lekki), due to the large concentration of smallholder rice farmers in these LGAs. These 6 villages had a total of 350, 246, 330, 333, 449, and 450 rice farmers, respectively, from which a total of 281 respondents were randomly selected. Each component of the questionnaire had sub-sections that allowed the researcher to collect the information she needed from the respondents. The questionnaire’s content represented the study’s objectives, as well as the information needed to solve them. The respondents were given the questionnaires that had been created. Three adhoc enumerators were hired to serve as the go-between the researcher and the respondents in each of the Local Government Areas, for a total of nine adhoc enumerators. The respondents’ socio-economic characteristics were analyzed using descriptive statistics. Percentages and frequency were used to achieve this.

3.2. Empirical Model for the Determinants of Farmer’s Choice of Loan

In order to attain the objectives of this study, several steps of analysis were undertaken. Both descriptive statistics and econometric models were employed to analyze the data. Predictor variables for statistical analysis were drawn based on previous studies [6,8,19,20,21]. Following Gupta et al. [22] and Fikadu [23], this study utilized a multinomial logit model (MLM). This approach was used to estimate the determinant of farmers’ choice of loan sources from financial institutions for the empirical analysis. This model was used because it allows decision analysis over more than two dependent variable categories. In this case, the dependent variable that opts for loan sources has three categories: personal savings (own money), informal loan sources (input providers, friends and relatives), and formal loan sources (agricultural bank, commercial bank). In this case, the logit and probit models should not be used, since they are used only when the dependent variable is binary (i.e., takes two values). Multinomial logistic regression, since the analysis does not presume normality, linearity, or homoscedasticity, is favored over discriminant function analysis. In this analysis, rice farmers had alternatives to three types of loan source options, and the decision was focused on the possibility of optimizing their usefulness according to the factors associated with each loan source. The loan source choice model was expressed as follows:

$$P\left(Y_i=k|x\right)=\frac{exp\left(x_i{\beta }_i\right)}{1+{{\displaystyle \sum }}_{h=1}^{k}exp\left(x_i{\beta }_i\right)} for k=1,2,3\dots K$$

(1)

The response probabilities $P\left(Y_i=k|x\right)$, k = 1, 2 …, k are therefore determined by the factors associated with the source of loan choice, where $x_{ik}$ is the vector of attributes of the source of loan source and β is the vector of parameter coefficients. Multinomial regression coefficients only reflect the direction of the effect of the independent variables on the dependent variable [24]. Thus, neither the real degree of change nor the probabilities are reflected by the estimates. However, the marginal effects are used to calculate the predicted change in probability of a particular technique being selected with respect to a unit change in an independent variable from the mean. The marginal effect was determined by differentiating the average coefficients as follows:

$$\frac{\partial P_k}{\partial x_l}=P_j\left({\beta }_{kl}-{{\displaystyle \sum }}_{k=1}^{K-1}{P}_{kl}\right)$$

(2)

The determinants of farmers’ loan source choice, with the source of finance choice, is described by 1 if the source of the loan is a formal loan source, 2 if the source of the loan is an informal loan source and 3 if the source is personal savings. The independent variables used include the following: $x_1$ = age, $x_2$ = sex, $x_3$ = marital status, $x_4$ = household size, $x_5$ = education, $x_6$ = farming experience, $x_7$ = farm size, $x_8$ = engagement in off-farm, $x_9$ = annual off-farm income, and $x_{10}$ = interest rate.

3.3. Empirical Model for the Determinants of Loan Acquisition and Utilization

A two-part model approach was used for the empirical analysis. The model consists of two parts. The first model is used to estimate the probability of whether or not the individual farmer has access to loans. Then, the second part is used to estimate how much the individual farmer spends on the farm’s operation if he or she has access to a loan. The two-part model refers to situations in which an event’s outcome can or may not occur. One observes a positive random variable as it happens. If it does not, the result observed takes a value of zero and thus becomes a zero-censored variable. The two-part model allows for separate investigation of the effect of covariates on the extensive margin (logit model, if any access) and on the intensive margin (amount received or spent, if any). The two-part model has a long history in empirical analysis, and it is mostly used in a variety of empirical work in health service research [25,26,27,28]. It is important to know that the two-part model is not common in agricultural economics research work. The two-part model is expressed as follows.

In the first stage, we used logit estimation of the probability of having positive outcomes (y = 1, if y > 0). The selection equation is specified as follows:

$$\varnothing \left(y>0\right)=Pr(y>0 | X)$$

(3)

where x is a vector of explanatory variables (equal to 1 if the farmer has access to credit and 0 if otherwise). The second stage is used for estimating a continuous outcome (i.e., the positives outcome (y = 1 if y = 0); the conditional equation can be written as:

$$E[ y | y=1,X]=E[ y | y>0, X]$$

(4)

where X includes the determinants of the dependent variable y (i.e., loan utilization). There are two parts of the prediction of the dependent variable y, with the first part resulting from the first phase (1), $Pr(y>0)$, and the second part being the conditional expectation $E[ y |y>0]$ from the second phase (2):

$$\begin{array}{c}E\left[ y \right]=Pr\left(y>0\right)\times E[ y | y>0\left]+Pr\left(y=0\right)\times E[ y | y=0\right]\\ =Pr\left(y>0\right)\times E[ y | y>0]\end{array}$$

(5)

4. Results and Discussion

4.1. Descriptive Statistics Results

Table 1. Summary description of variables included in the models.

Variable	Description	Mean	SD
Dependent
Access to loan	1 if farmer has access to loan and 0 otherwise	0.88	0.32
Amount of loan	The amount of loan obtained by farmers from the financial institutions (in NGN/USD)	104,192.2	60,335.83
Choice of loan sources	Category (1 if formal source, 2 if informal source and 3 if personal savings)	1.37	0.68
Explanatory
Age	Continuous variable indicating age of the farmers (years)	38.37	10.47
Sex	1 if male and 0 otherwise	0.65	0.477
Marital status	1 if single and 0 otherwise	0.23	0.42
Household size	Continuous variable indicating number of people on respondent	3.62	2.31
Education	Continuous variable indicating number of years spent in formal education (years)	12.31	3.33
Farming experience	Continuous variable indicating the period of time a farmer has been growing rice (years)	4.87	1.73
Farm size	Continuous variable indicating the total land area under rice production (hectares)	1.57	0.49
Membership in a cooperative	1 if farmer is member of cooperative and 0 otherwise	0.89	0.32
Contact with extension agent	1 if famer has access to extension service and 0 otherwise	0.028	0.1666
Interest rate	Interest rate as a percentage	15.41	6.659
Annual off-farm income	Continuous variable indicating the annual off-farm income of the respondent farmer (in NGN/USD)	312,438.4	175,204.5
Annual farm income	Continuous variable indicating the annual farm income of the respondent farmer (in NGN/USD).	462,098.9	213,997.2

Note: USD 1 = NGN 360 (the official currency of Nigeria is Naira (NGN). Source: authors’ computations.

summarizes the descriptive statistics of the variables used in the regression models. The mean, standard deviation, minimum, and maximum values of the variables used in the models are highlighted in this summary. The study analyzed data collected from 281 smallholder rice farmers, of whom 88.3% of respondents had access to credit, while 11.7% of rice farmers used personal funding sources. The age range for rice farmers is between 18 and 68 years, with an average age of about 38 years and 4 months. The study shows that 65% of the respondents are male, and 35% of them are female, while 23% are single and 77% are married. In addition, the household size range of respondents ranges from 0 to 12, and the average household size is 4 people. The results also showed that the respondents spent a minimum of 5 years in school and a maximum of 20 years in school, while the average years of formal education obtained was 12 years. The results of the review of Table 1 show that rice farmers have acquired a minimum of 2 years and a maximum of 12 years of agricultural experience in rice cultivation or production in the study field. The farm size varies from 1 to 2 hectares for respondents, and the average farmland cultivated is 1.57 hectares, while 88% of rice farmers are cooperative members and 2% have communication with extension agents.

The results also showed that the choice of loan sources by the respondents varies from 1 to 3, and an average of 1.37 individuals have variations in the study region in the choice of loan sources. The minimum and maximum loan interest rates obtained by respondents from various loan sources vary from 0 to 25%, and the average loan interest rate obtained is 15.4%. In addition, 84% of respondents are engaged in both off-farm business and farming activities, while 26% of respondents are engaged in farm business only. The annual off-farm income of respondents engaged in an off-farm company varies from NGN 0 to NGN 840,000 (USD 0 to USD 2333) with an average income of approximately NGN 312,438.4 (USD 868), while the annual farm income of respondents in the study area ranges from NGN 15,000 to NGN 922,500 (USD 42 to USD 2563), with an average income of NGN 462,098.9 (USD 1284). Finally, the results show that a maximum loan amount of NGN 300,000 (USD 833) was disbursed to smallholder rice farmers, and an average of approximately NGN 104,192.2 (USD 289) was received by the respondents in the study area with regard to the total amount of loan released.

4.2. Descriptive Statistics of Farmers’ Sources of Loan

Farmers’ loan sources are summarized in

Table 2. Sources of loans.

Loan Sources	Frequency	Percentage (%)	Interest Rate (%)
Personal savings	32	11.39	-
Formal sources	112	39.86	9–15
Informal sources	137	48.75	20–25

Source: Authors’ computations.

. The table shows that informal sources are the major source of loans available to rice farmers in the study area, followed by formal sources. Analysis of

Table 3. Diagnostic check for multicollinearity and heteroscedasticity.

Variable	VIF	Tolerance
Multicollinearity test
Farm size	3.92	0.2549
Household size	3.54	0.2759
Marital status	2.92	0.3428
Annual off farm income	2.28	0.4394
Membership	2.27	0.4410
Interestrate	2.17	0.4614
Age	1.89	0.5282
Farming exp	1.47	0.6815
Education	1.20	0.8365
Sex	1.17	0.8521
Extension agent	1.16	0.8648
Mean VIF	2.30
Breusch–Pagan test for heteroscedasticity
Chi-square		31.84 ***

Note: *** denotes 1% significant level. Source: authors’ computations.

reveals that the majority (48.75%) of the rice farmers in the study area sourced their loans from informal sources, with an interest rate of 20–25 percent. This is followed by formal sources (39.86%) and those that used personal savings (11.39%). This implies that the major sources of loans among respondents in the study were informal sources. This is supported by [20], who found that informal sources’ high patronage could be related to cheaper interest rates and the lack of a requirement for collateral security. Farmers may now obtain loans from informal sources far more easily and quickly than they do from legal institutions.

4.3. Econometric Results

One of the assumptions made about the explanatory variables is that they are unrelated. When this assumption is violated, a multicollinearity problem arises, resulting in inaccurate regression coefficient estimations with incorrect signs and magnitudes. As a result, inferences about the link between the dependent and explanatory variables may be incorrect. A multicollinearity test was carried out with the variance inflation factor (VIF) as the primary criterion. Table 3 displays the results of the multicollinearity test. One of the assumptions made regarding the explanatory variables is that they do not correlate with each other. The violation of this assumption creates a multicollinearity problem, which results in incorrect signs and magnitudes of regression coefficient estimates. This may lead to inaccurate conclusions regarding the relationship between the dependent and explanatory variables.

A multicollinearity test was performed, specifically using the variance inflation factor (VIF). The results of the multicollinearity test are shown in Table 3. According to Maddala [29], multicollinearity is present in a model if the VIF is larger than 10 and the tolerance is close to 0. None of the variables had a VIF larger than 10, according to the results. The entire mean VIF is 2.30, which is much less than 10. This shows that the explanatory variables are not associated, and hence, the multicollinearity problem is not present. We also used the Breusch–Pagan test to see if there was any heteroscedasticity. The Breusch–Pagan test’s chi-squared value (31.84) is statistically significant at the 1% level, indicating that the model possesses heteroscedasticity (i.e., the error term has non-constant variance). If this problem is not addressed, estimations may become skewed and inaccurate. As a result, a robust estimating strategy was used to estimate the standard errors, as described by [29]. The heteroscedasticity issue was solved via robust estimation.

4.4. Determinants of Rice Farmers’ Choice of Loan Sources from Financial Institutions

The results of the choice of loan source from the multinomial logit model are presented in

Table 4. Estimates of the multinomial logit for the determinants of farmer choice of loan sources.

Explanatory Variables	Formal Source			Informal Source
	Personal Savings (Base Outcome)
	Coeff	SE	ME	Coeff	SE	ME
Age	0.067	0.060	0.0073	0.037	0.059	−0.0070
Sex	−0.30	0.89	−0.015	−0.24	0.88	0.014
Marital status	1.62	1.54	−0.25	2.72 *	1.51	0.26
Household size	−0.028	0.26	−0.044	0.15	0.25	0.045
Education	0.20	0.15	0.029	0.083	0.15	−0.028
Farming experience	−0.37	0.31	−0.023	−0.28	0.31	0.021
Farm size	4.30 **	1.76	−0.018	4.42 **	1.76	0.043
Engage in off farm	−0.84	1.80	0.025	−0.95	1.78	−0.029
Annual off farm income	1.12 × 10−6	3.92 × 10−6	0.219 × 10−6	0.240 × 10−6	3.92 × 10−6	−0.215 × 10−6
Interest rate	0.40 ***	0.081	0.00080	0.40 ***	0.080	0.0015
Constant	−11.45 ***	4.13		−9.88 ***	4.07
Observation	281
LR chi2 (22)	172.23 ***
Prob > chi2	0.0000
Pseudo-R2	0.3178

Note: *, **, ***, indicate 10%, 5%, and 1% statistical significance. Source: authors’ computations.

. The results showed that marital status, farm size, and interest rate were positive and significant factors that influenced farmers’ choice of loan sources. The likelihood ratio test (Chi² (22) = 172.23; p > 0.000) was significant at 1 percent level, thus rejecting the null hypothesis that the socio-economic characteristics of rice farmers in the study region do not have a significant effect on their choice of loan sources. The pseudo-R square was 0.3178, suggesting a 31.78 percent relationship between the predictors and the predictions. In other words, 31.78 percent of the likelihood of rice farmers’ choice of loan sources is explained by the independent variables.

The results in Table 4 reveal that the coefficient of marital status is statistically significant at the 10 percent level and is shown to have a positive impact on the choice of loan source. This means that if a farmer is single, compared to using his/her personal savings, he/she is 26 percentage points more likely to source loans from an informal source. Formal financial institutions see married farmers as comparatively more secure, accountable, and willing to repay borrowed funds, making them the only source of formal loans to farmers. Our finding is consistent with that of Aladejebi et al. [30], who observed that single farmers acquired less agricultural credit compared to married farmers.

The coefficient of farmer based on farm size shows a positive effect and is statistically significant at the 5 percent level. This indicates that farmers whose farm sizes increase by 1 ha are 1.8 percentage points more likely to use formal loan sources and 4.3 percentage points more likely to use informal loan sources than their personal savings. With an increase farm size, the demand for loans from formal sources is also increased to match the increased farm inputs and labour that will be needed. However, farmers’ access to agricultural loans from formal sources grows as farm size increases, but after a period of time, the likelihood of farmers receiving formal loans diminishes as farm size increases. Farmers with more farming experience are better managers and hence less likely to need credit from formal sources. This finding is also in line with the results of [31,32], who found that farm size has a positive relationship with farmers’ access to credit from banks.

The empirical results show that the coefficient of the interest rate was significant at 1% and positively related to the choice of loan source. This implies that as interest rates increase, farmers are 0.08 percentage points more likely to use formal sources of loans and 0.15 percentage points more likely to use informal sources of loans compared to their personal savings. Informal sources of loans are attractive to farmers owing to the lower interest charged from these sources. Our finding collaborates with [33], who opined that interest rates could negatively influence credit demand from formal sources.

4.5. Determinants of Loan Access and Utilization

Table 5. Determinants of farmers’ access to loans and utilization.

Variables	First Part Model	Second Part Model	Marginal Effect
Age	0.19	−0.00089	23.81
	(0.20)	(0.0027)
Sex	−0.87	−0.0028	−829.96
	(1.81)	(0.046)
Marital status	3.38	0.058	8134.39
	(3.94)	(0.085)
Household size	−0.82	0.017	1292.62
	(0.66)	(0.017)
Education	0.15	0.026 ***	2771.96
	(0.24)	(0.0067)
Farming experience	0.64	0.028 **	3331.04
	(0.67)	(0.014)
Farm size	−1.74	0.45 ***	45,395.16
	(4.71)	(0.067)
Annual off-farm income	−0.000012	−2.76 × 10−7 **	−0.04
	(9.21 × 10−6)	(1.18 × 10−7)
Annual farm income	0.000055 **	6.62 × 10−7 ***	0.10
	(0.000022)	(1.70 × 10−7)
Interest rate	0.76 *	−0.0069	−258.32
	(0.39)	(0.0049)
Constant	−23.01 *	10.22 ***
	(13.56)	(0.19)
Observation	281	248
LR Chi2(10)/F-value	187.51 ***	33.33 ***
Prob > Chi2/Prob > F	0.0000	0.0000
Peudo-R2/R-squared	0.9222	0.5844
Adjusted R-squared		0.5669

Note: *, **, ***, indicate 10%, 5%, and 1% statistical significance. Source: authors’ computations.

presents the estimation results from the two-part model. The results show that annual farm income and interest rate were significant factors influencing farmers’ access to loans using the logit model (first part model), while the results of the second model using least squares reveal that education, farming experience, farm size, annual off-farm income, and annual farm income were significant factors influencing the use of the loan.

Analysis of the logit part of the model in Table 5 shows that the coefficients of annual farm income and interest rate were significant. The annual farm income had a coefficient that was positive and significant at the 5 percent level, indicating that the farmers are more likely to access loans from financial institutions than their counterparts. The marginal effect of 0.10 suggests that farmers are 1.0 percent more likely to access loans as farm income rises. As farm income increases, the perception of farmers not being able to repay collected loans decreases, and this motivates farmers to request loans. Our finding is confirmed with a recent study by [6,34]. The researchers ascertained that increasing the income of farmers decreases the risk perception and thus increases the possibility of access to credit.

Interest rate is positively related to the probability of farmers accessing loans from financial institutions. The marginal effect of −258.32 indicates a 10 percent increase in interest rate, suggesting that farmers are more likely to access loans as the interest rate rises. Farmers are discouraged from using a loan by high interest rates. However, those farmers with a high level of education demonstrate that their likelihood of accessing loans increases with a higher interest rate. Farmers with a high level of education are more likely to realize the credit benefits of modern farming, to understand the expansion of credit sources, and, thus, to increase their demand for loans. These results are corroborated by [6,35], who recorded that education increases the knowledge and awareness of the need for loans for farmers and thus leads them to search for credit facilities.

The results of the second part of the two-part model are shown in Table 5. The least-square estimates reveal that the coefficients of education, farming experience, farm size, annual off-farm income, and annual farm income were significant. The coefficient of education was significant at the 1 percent level and positively related to loan utilization. This indicates that as a farmer’s years of education increase by 1 year, his/her loan utilization increases by NGN 2771.96 (USD 8), as seen in the marginal effect column. Education impacts the farmer’s ability to allocate loans more efficiently, and thus, farmers with more years of formal education will allocate more of their loan to their farm enterprises. This finding agrees with [4], who posited that educated farmers allocate farm resources more efficiently than uneducated farmers. However, it contradicts the findings by [20], who found that education does not have a significant impact on farm loan usage.

The coefficient of farming experience is statistically significant at 5 percent and is shown to have a positive impact on loan utilization. This implies that as the years of farming experience of the farmer increase by 1 year, his/her loan utilization increases by NGN 3331.04 (USD 9.25), as seen in the marginal effect column. Farmers with more years in farming are likely to allocate more of their loan to their farm enterprises than farmers who are inexperienced. Our finding is in accordance with Adekoya [36], who indicated that farmers with more years of experience in farming have better knowledge of farming and the ability to utilize necessary farm inputs efficiently.

The empirical results revealed that the farm size coefficient was positively significant in relation to loan utilization at 1 percent. The positive sign of the coefficient conforms to the a priori assumption, indicating that as the farmer’s farm size increases by 1 ha, the utilization of his/her loan increases by NGN 45,395.16 (USD 120.09), as seen in the column of marginal effect. Farmers with larger farm sizes will need more capital to take care of the increased inputs and labour required for production, and hence, they are more likely to have a higher level of loan utilization. This finding agrees with the recent study by Okeke [37], who posited that the level of credit utilization is higher among farmers with large farm sizes when compared with farmers of smaller farm sizes.

Annual off-farm income is significant and negatively related to the use of loans at the 5 percent level. This indicates that as annual off-farm income rises by NGN 1 (USD 0.003), as seen in the marginal effect column, the amount of loan utilization of the farmer decreases by 4 kobo. An increase in annual income from off-farm activities reduces farmers’ perception of not being able to repay collected loans, thus increasing their demand and utilization of loans. However, farmers whose level of loan utilization decreases with an increase in annual off-farm income are those with a high dependent ratio due to a large household size. Such a large number of dependents increases the likelihood to allocate a portion of the borrowed loan to family upkeep. This was confirmed by [30], who revealed that meeting family needs was one of the most common reasons farmers diverted the loans they obtained.

The coefficient of annual farm income was significant at 1% and positively related to loan utilization. This result implies that as annual farm income increases by NGN 1 (USD 0.003), the farmer’s level of loan utilization increases by NGN 0.10, as seen in the marginal effect column. Increased farm income is a function of the efficient utilization of the loan received. Thus, farmers with higher farm income would want to utilize the loan received efficiently so as to be able to secure another loan. Farm revenue has a favorable impact on a farmer’s access to agricultural financing since farmers can utilize farm income to manage loan risks and implement new agricultural technologies. This finding is also consistent with the findings of [31,32], who discovered a positive association between farm income and farmers’ access to non-institutional loans. However, it contradicts the findings [20], who report that farmers’ decreased loan utilization with increasing farm income is probably those with large household sizes.

5. Conclusions and Recommendations

Smallholder rice farmers need access to agricultural credit and financial services to improve their welfare, especially in developing nations like Nigeria, where agricultural development is growing. The availability of agricultural loans is essential to the sustainable development of the agriculture sector. This paper has investigated farmers’ access to loans and sources of loan choice using a two-part model approach and a multinomial logit model, respectively. This study used a multinomial logit model to identify factors that influenced a farmer’s choice of loan sources. Through the use of a two-part model, we also pinpoint the factors that influence farmers’ access to agricultural loans and use of loans (utilization). This study was conducted in the rain-fed rice production areas of Lagos State, Nigeria, with a sample size of 281 rice farmers. Evidence from this study has shown that rice farmers in Lagos State have a very high affinity for loan acquisition and utilization. Most of the rice farmers, however, were unable to access and utilize loans from formal institutions compared to informal institutional sources due to constraints imposed by socio-economic factors. Consequently, the results of the study rejected the null hypothesis that socio-economic characteristics of rice farmers do not significantly influence their loan acquisition, utilization, choice of loan sources and farm income and accepted the alternative hypothesis.

Our findings show that most of the respondents were male, married, and members of a cooperative society with an average age of 38 years. This could imply that because most rice farmers are young and responsible, they are more likely to access loans from financial institutions. Furthermore, being a member of a cooperative allows them to obtain loans from a financial institution at a reasonable rate. The majority of the rice farmers, on the other hand, had no contact with an extension agent. Most of the farmers lived in households with fewer than four persons and had less than five years of farming experience. Therefore, most of the rice farmers in the area were young, educated, and earned above minimum wage on a monthly basis. The most preferred source of loans among rice farmers in the study area was an informal source (produce buyers). This is because it is easier and quicker for farmers to get loans with no insistence on collateral from informal sources compared to formal sources. Furthermore, this research revealed that the farmer’s education, farming experience, farm size, off-farm income, farm revenue, and interest rates are all criteria that serve as both boundaries and deciders for the amount of loan acquired and used by smallholder rice farmers. This study also discovered that marital status, farm size, and interest rate all have an impact on smallholder farmers’ loan sources in Lagos State, Nigeria.

Based on the key findings of the study, the following policy recommendations are proposed. First, a deliberate policy and program should be developed and implemented to improve extension services to farmers in all rice-producing regions in the state of Lagos. Secondly, policies that will encourage the establishment of more produce buyers (inform institutional sources of loans) in the state should be put in place. The state government, lending institutions, and NGOs in the state should take into cognizance the marital status, farmer’s education, farming experience, farm size, off-farm income, farm revenue, and interest rate when designing credit schemes for rice farmers in the state. These variables should also be taken into consideration in the effort to get these farmers to apply for a designed loan scheme. Thirdly, policies and programmes aimed at improving loan access and utilization among rice farmers in the state should include the socio-economic characteristics of the farmers in their formulation and implementation.

A deliberate policy should be implemented to encourage financial institutions to provide rice farmers with the exact amount of loans they need at a low interest rate. Land tenure policies should, as a result, adopt land reforms that expand farmers’ access to land while also securing their land rights. It is expected that the effective implementation of these policy recommendations will stimulate the sustainable development of the agricultural sector in Lagos State.

Due to limited resources and time, the study focused on Lagos, which is only one state in the western region of Nigeria. The rice value chain in Nigeria has producers (on small, medium, and large scales), processors, and marketers. However, due to resources and time, the study only considered the smallholder segment of the value chain.