The Effectiveness of Credit Risk Mitigation Strategies Adopted by Ghanaian Commercial Banks in Agricultural Finance

Abstract

Lending to the agricultural sector by commercial banks in Ghana is characterized by high credit risk. Empirical evidence suggests that commercial banks in Ghana have credit risk management (CRM) challenges. This study explores the credit risk mitigation strategies adopted by commercial banks to minimize credit risk in agricultural finance in Ghana. The study adopted a mixed-method approach using a survey questionnaire and interview instruments. The findings indicate that some of the strategies used by commercial banks to mitigate credit risk in agricultural finance do not meet commercial banks’ CRM needs. In addition, Ghanaian commercial banks have not fully adopted some of the recommended strategies that are used to mitigate credit risk associated with agricultural lending. The study unveils some appropriate strategies used to mitigate credit risk exposure in agricultural finance among commercial banks. These strategies include agricultural value-chain financing, collaboration with off-takers, incentive-based and risk-sharing schemes, adoption of a holistic agricultural value chain financing, policy interventions, use of agricultural insurance pool, and the proper structuring of agricultural loans.

1. Introduction and Background

In the 21st century, agriculture is the key to development and growth in every nation (Alvarado et al. 2021). Agriculture plays a significant role in the Ghanaian economy as it provides jobs and serves as the main source of food supply (Ayerakwa et al. 2020). Even though Ghana has the potential to meet its food demand, the country depends largely on imported fresh agricultural goods. From this viewpoint, focusing on the agricultural sector is essential for growth and development in Ghana (MoFA 2023). However, demands for agricultural finance by Ghanaian commercial banks have been severely impeded by high credit risk (MoFA 2023). Even though the credit risk management problem associated with agricultural finance has been confronted head-on by the Ghanaian banking industry in recent times (MoFA 2023), ineffective credit risk mitigation strategies caused insolvency for some banks resulting partly from credit risk concomitant with agricultural finance. These include the revoking of the licenses of UT and Capital banks in August 2017 and the consolidation of Royal Bank, Unibank, Beige Bank, and Sovereign bank in August 2018. Banks in Ghana are expected to increase their support to the agricultural sector, but the painful fact of risk has stalled such an undertaking over the years (BoG 2023). Whilst the finance, service, construction, and transport sectors, among others, received 25.2, 21.1, 11.0, and 21.1 percent of agricultural credit supply in 2017, respectively, the agricultural sector only received 4.1 percent, recording the second-lowest in the season (BoG 2023). Similar trends were recorded from 2018 to 2022. The neglect in the agricultural sector is because commercial banks have, over the years, employed inadequate credit risk management practices to mitigate credit risk associated with agricultural finance, though the sector has been exposed to high lending risks (BoG 2023). The proportion of non-performing loans (NPLs) in the agricultural sector was significant and, on average, higher between 2016 and 2020 as compared to those of the other sectors. The NPL ratio in the agricultural sector was 42 percent in 2016, 43.6 percent in 2017, 35.5 in 2018, 23.1 in 2019, 19.8 in 2020, and 23.7 in 2021 compared to the low rates recorded by other sectors, such as service, transport, and mining, among others, within a similar period (BoG 2023).

The high non-performing loan ratio means commercial banks need to develop appropriate strategies to mitigate credit risk associated with agricultural lending. However, the level of bank failure in Ghana leaves many questions unanswered on credit risk mitigation practices specifically in the area of agricultural finance and this prompted the interest of this research. The key questions are on whether Ghanaian commercial banks have appropriate and effective credit risk mitigation strategies. The existing literature, such as Appiah-Twumasi et al. (2020) and Ayerakwa et al. (2020), highlighted innovative financing and food security whilst completely ignoring credit risk associated with agricultural finance. Similarly, Kessey (2015), Kusi and Opoku-Mensah (2018), and Lagat et al. (2019) provided a general description of credit risk management and credit referencing but have not answered questions relating to credit risk mitigation strategies in agricultural finance, a gap that needs to be investigated. There is a dire need to interrogate these aspects. The immediate implication is that commercial banks might not have appropriate credit risk mitigation strategies (Nyebar et al. 2021). Furthermore, although commercial banks could have credit risk mitigation strategies, some of them might not be effective and there could be concomitant challenges in implementing such strategies.

The appropriate and effective credit risk mitigation strategies can indirectly assist in minimizing the collapse of banks, such as the recent failure of Silicon Valley Bank (SVB) and Switzerland’s Credit Suisse bank. SVB’s failure, though primarily due to interest rate risk and liquidity issues, can be loosely linked to credit mitigation strategies. The bank’s concentrated client base in the tech sector exposed it to high credit risk, which effective mitigation strategies typically aim to avoid through diversification. Additionally, SVB’s failure to manage interest rate risk effectively, a factor closely tied to credit risk, further exacerbated its vulnerabilities. The focus on risky startups might have required more stringent credit assessments, and the resulting liquidity crisis likely weakened the bank’s ability to manage credit risk, creating a feedback loop that contributed to its downfall. Some factors in the SVB’s failure show how weaknesses in risk management, including credit mitigation, can lead to broader systemic failures. Thus, the need for banks to develop effective and implementable credit risk mitigation strategies cannot be overemphasized. This paper contributes to the existing literature by exploring the best credit risk mitigation strategies in agricultural finance to boost the sector as a viable business while protecting banks from possible credit-related failure and ripple effects from other exposures, such as interest rate risk and liquidity issues. Therefore, this research is relevant in identifying the weaknesses of banks’ credit risk mitigation strategies in the aftermath of SVB and other banks’ failures.

The remainder of this article is organized as follows. Section 2 presents the literature review. Section 3 describes the data and model used in the study. Section 4 presents results and discusses the findings of this study. Section 5 concludes the study.

2. Literature Review

2.1. Conceptualization of Credit Risk in the Agricultural Lending Context

Credit risk management can be contextualized by liquidity theory, which has a strong connection with credit risk resulting from borrowers’ defaults, which reduces cash flow and affects liquidity (Abdelaziz et al. 2022). The theory implies that banks must hold a large number of liquid assets against possible demand or payment cushion of readily marketable short-term liquid assets against unforeseen circumstances (Ejoh et al. 2014). Lower liquidity negatively impacts agricultural lending (Maloba and Alhassan 2019). The liquidity crisis of banks sparks the interest in the role of the regulations that could promote resilience in banks (Mohd Amin and Abdul-Rahman 2020). In addition, the adverse selection theory was originated by Stiglitz and Weiss (1981). The theory is underpinned by two main assumptions. Firstly, lenders cannot differentiate between the appetite of borrowers and risk levels (Stiglitz and Weiss 1981). Secondly, credit contracts are subject to limitations (Stiglitz and Weiss 1981). That is, where the debt responsibility is more than the business or project proceeds, borrowers have no responsibility to pay out of their own pockets. The adverse selection theory does not guarantee that the loan applicant secures credit when the interest is high, particularly at a time where loanable funds are insufficient (Vallee and Zeng 2019). Furthermore, credit referencing theory highlights information sharing, which was earlier initiated by Freimer and Gordon (1965) and later extended by Stiglitz and Weiss (1981). The theory states that credit markets do not have adequate, complete, and reliable credit information on individuals (asymmetry of information). For a credit information system to be effective, there should be a reliable national identification system for individuals; an adequate regulatory or policy structure; a framework that supports information sharing; and a system that protects clients’ privacy in Kenya (Saruni and Koori 2020). The purpose of credit referencing theory is to expose all borrowers to a wide range of credit providers through credit bureau institutions. These theories explain the possible behavior of lenders and borrows when faced by credit risk and provides assumptions to use in evaluating the effectiveness of credit risk mitigation strategies.

2.2. Review of Empirical Studies on CRM Strategies

The empirical studies show that credit risk mitigation strategies in agricultural finance comprise factors such as the identification of loans with distress signals, review of loan granting processes, monitoring, maintaining good portfolio quality, review of employees’ skills, and training of credit officers. Yin et al. (2020) investigated loan defaults associated with agricultural finance in China and found that distress loans associated with agricultural lending were higher than non-agricultural related loans and that regular assessment of loans with distress signals significantly reduces loan losses among banks. If distress loans associated with agricultural lending are higher than in other sectors in developed economies such as China, could it be that agricultural lending risk is common among all countries in Africa, especially Ghana? This is not different in Ghana, signaling that credit risk associated with agricultural lending is inevitable and banks must fully embrace and manage it. Similarly, Netzer et al. (2019) found that potential defaulters are most likely to include distress signal words like family relations, mention of God, general hardship of the borrowers, pleading lenders for help, and short-term-focused words. This assertion is relevant, especially in Africa where family relations, mention of God, and pleading with lenders for help are common. These signals do not automatically mean that the agricultural borrower will default. What is most significant is the repayment trend, credit history, cashflow, and the capacity of the borrower to repay the loans.

Likewise, Instefjord and Nakata (2022), and Muhammad et al. (2018) found well-established approval processes for granting new loans; monitoring and effective communication of risk management guidelines; evaluating the flow of borrowers’ business; regular review of borrowers reports; on-site visit; a review of employees’ skills; training of credit officers; updating the credit file of borrowers’ ratings assigned at the time of granting the credit; credit reports; credit rating; and portfolio risk control are strategies that are used to reduce credit risk. Marinelli et al. (2022), Monnet and Nellen (2021) found that the regular review of borrowers’ business operations; cash flow verification; loan diversification; credit reminders; review of borrowers’ character, collateral, capacity, and capital conditions; credit appraisal; and collateral monitoring are significant strategies that can be used to mitigate credit risk. However, these findings are generic and are not specifically based on loans granted for agricultural purposes. This study unveiled specific strategies directly aimed at reducing credit risk exposure associated with agricultural lending.

Furthermore, Daum and Birner (2017), Olowa and Olowa (2017), and Sackey (2018) identified collateral monitoring and collaboration with other lenders’ proper documentation to be effective strategies in minimizing the credit risk of the banks. These findings confirm that commercial banks do not lend to agricultural borrowers who do not possess title deeds to lands. It means that Ghanaian commercial banks should finance only those borrowers engaged in agricultural activities to acquire farm equipment and register such equipment in joint ownership under strict monitoring until the loan is fully repaid. Using data gathered from the European Union, Ezike (2021) conducted a study on collateralization, bank loans, and rates. The findings indicated that the collateralized and proper documentation of credit-related transactions among countries such as the United States, United Kingdom, and Australia are good strategies used to mitigate credit risk exposure of banks but must be regularly monitored to make them effective. Ezike (2021) based these findings on banks in developed economies where agricultural borrowers mainly comprise largely organized institutions and individual entities with well-organized collateral that are used to secure loans. This is different in developing countries, particularly Ghana, where most agricultural borrowers are not well organized and do not have the needed collaterals that can attract loans. Be that as it may, the literature evidence is an indication that Ghanaian commercial banks must effectively monitor all collateralized transactions and pay more attention to the documentation of credit-related transactions to minimize defaults among agricultural borrowers.

Netzer et al. (2019) confirmed that the credit departments must regularly review the repayment progress of borrowers to enforce effective repayment of loans on time. Also, Cucinelli et al. (2021) used an unbalanced panel of 73 Italian banks over 7 years and found that establishing a unit or office by banks to constantly manage, assess, and monitor impaired or bad loans increases loan performance and the loan portfolio quality of the banks. Agasha et al. (2020) hold a similar view. It implies that commercial banks must consistently assess the portfolio quality of loans granted to agricultural borrowers to improve the loan performance and reduce the credit risk exposure of the banks. Bilal and Baig (2019), Kessey (2015), and Lagat et al. (2019) found the regular review of credit risk management practices and loan administration process to be very significant strategies used to mitigate credit losses. Even though their findings were based on evidence from savings and credit corporations and not specifically banks, which, in this context, are usually larger and bigger, the results provide open information for banks to emulate and minimize credit losses. This also suggests that a poor review could lead to loan delinquency. There is, therefore, the need to create an interconnected system and interdependent method for deliberate action to curtail risk and uncertainties in credit activities. Therefore, specific processes in credit risk management include the identification of risk factors, assessing the likely consequences of the risk factors that are identified, and the choices of strategies adopted by management in mitigating the effects of identified risk associated with agricultural finance.

Additionally, banks can resort to credit guarantee schemes as part of the credit risk mitigation strategies in agricultural finance. Credit guarantee schemes have been negatively criticized by scholars like Saito and Tsuruta (2018). The examination by Saito and Tsuruta (2018) on information asymmetry in enterprise credit guarantee schemes in Japan indicated that credit risk guarantee corporations cannot differentiate low risk from risky borrowers and that credit guarantee schemes usually attract a larger proportion of risky borrowers, resulting in inefficient resource allocation. Contrarily, Wubin et al. (2020) as well as Yoshino and Taghizadeh-Hesary (2019) identified the regular use of credit guarantee schemes as one of the efficient strategies in mitigating credit risk in agricultural finance. Similarly, Daum and Birner (2017) examined the neglected governance challenges of agricultural mechanization in Africa using Ghana as a case and revealed that, with a high-interest rate of about 35 to 42 percent in the agricultural sector, banks can cooperate with other entities, such as DANIDA, through credit guarantee schemes to address the challenges of credit risk associated with agricultural finance in Ghana. However, a shortfall of this argument is the fact that loans granted under the DANIDA-aided scheme are only limited to farmers providing tractor services to small agricultural farmers and not to general or large agricultural activities. The findings implied that, by regularly using credit guarantee schemes, credit associations and the governments serve as guarantors in the facilitation of the loans to borrowers and agree to pay off lenders a significant part of the loan in the case of default. Unlike developed countries where credit guarantee schemes are effectively working because of their robust systems, this is different in Ghana where successive governments over the past years struggled to meet developmental needs with limited resources. It will sound very irrational and irresponsible in the West African sub-region, particularly Ghana, to defer the government’s development projects just to pay off agricultural loans for defaulters.

The discussed empirical studies open up a broad debate on the need to adopt adequate and the best credit risk management strategies to mitigate credit losses associated with agricultural lending. Even though credit risk mitigation strategies have become an integral part of commercial banks’ overall risk mitigation strategy in minimizing credit risk associated with agricultural lending (Agasha et al. 2020), Ghanaian commercial banks have some credit risk mitigation challenges that need to be improved. Banks must therefore make the best strategic choice to minimize credit losses in agricultural lending and reduce credit risk (Bülbül et al. 2019).

3. Methodology

3.1. Research Approach and Sample Selection

This study adopted a mixed-method approach involving the use of interviews and survey questionnaire data. The study employed the convergent parallel mixed-method through which the researcher administers the questionnaires, reviews policy documents, conducts in-depth interviews simultaneously, and then integrates the findings into the interpretation of the results (Creswell and Clark 2017). Among the 23 commercial banks in Ghana (BoG 2023), 4 commercial banks were purposively selected for this study because they are widely distributed over the country and constitute the leading banks that mostly provide agricultural lending with a market share of 73 percent of the agriculture loan portfolio in Ghana (BoG 2023). The purposive sampling in this study was adopted to select participants who could specifically provide appropriate and relevant information regarding credit risk management in agricultural lending. The selected credit officers and managers work in the banks’ credit risk and agricultural financing units. Due to the nature of their jobs, they have relevant knowledge and could provide adequate information on credit risk management in agricultural lending. Furthermore, the approach was not only cost-effective but was also time-effective in practice. A sample size of 336 participants was representative of the study population size of 1800. However, 319 participants turned up, representing a response rate of 94.9 percent. The questionnaire was validated for content validity, reliability and consistency using Cronbach’s alpha test. Also, the Kaiser–Meyer–Olkin (KMO) sampling adequacy test and Bartlett’s test of sphericity were used to assess the suitability of the data for analysis. Ten credit managers out of the twelve purposively targeted participated in the interviews, representing a response rate of 83 percent. A maximum of 3 staff was considered from each of the four banks since the researcher reached saturation after interviewing 2 to 3 participants. The selection of the interviewed participants considered the assumption that the obtained results supplement the information gathered through policy documents. The NVivo software version 12 was employed to transcribe the qualitative data.

3.2. Data Analysis Method

The study made use of the minimizing projection residual and maximizing variance approach of the principal components analysis (PCA) to examine the major components of commercial banks’ model of CRM practices. This study used the PCA to reduce the original variables into a lower number of orthogonal (non-correlated) variables to visualize the relationship between factors and the original variables and to visualize proximity in mathematical units. The initial extraction of components involved the initial stage of extracting the total number of components, which is equal to the 17 analyzed variables. As the sequential factor selection continues, the factors comprise less variability where the next stage was limited to determining the meaningful components based on Eigenvalue-one criteria (components with eigenvalues >1 were retained and interpreted) (Hatcher and O’Rourke 2013). Eigenvalues are the variances captured by a given component (Hatcher and O’Rourke 2013). The scree test of eigenvalues for each component was plotted with the aim of identifying breaks between large and small eigenvalues. Components that appear before the break were assumed to be more meaningful than those that appear after the break. The study retained all components that accounted for more than 5 percent of the total variance. The interpretability criterion was followed to check the substantive meaning of variables that load on specific components and to confirm whether they share a similar conceptual meaning. Oblimin rotation, which provides an oblique solution that allows the factors to be correlated (Mabel and Olayemi 2020), was used to rotate the factors. A correlation matrix was built to verify the validity of the relevant factors and assess the adequacy of the solutions achieved. The reproduced correlation coefficient has to be close to the original matrix to validate the relevant factors (Hatcher and O’Rourke 2013). Highly correlated items were retained. The first principal component represents the direction in the future space in which the prognoses have the highest variance. Also, the second component represents the direction that maximizes the variance from all the orthogonal directions to the first. Hence, the Kth component = K − 1 component. It means that the Kth factor represents the variance-maximizing orthogonal direction to the preceding K − 1 component. For instance, if the data for this study are centered so that all features have a mean = 0, the data are expressed as Matrix X as:

$$x^x=nV$$

(1)

where V represents the covariance matrix of the data. Therefore, the score for each participant of a particular component was computed. A total of 17 items were developed and loaded into the PCA. These include the identification of loans with distress signals, regular review of loan granting processes, portfolio quality, credit administration process, borrowers’ credit reports, borrowers’ performance profile, flow of borrowers’ business, training of credit officers, and monitoring of collateralized transactions. It also includes loan guarantees, checks on repayment of loans, identification of loans with potential credit weakness, and risk management practices. For instance, four components were identified, and each participant had four scores, with one score representing each of the components. The researcher weighed the participants’ actual scores on the 17 questionnaire items and summed them up for the estimation of scores for the given components. This was expressed as follows:

$$M_1=W_1\left(X_1\right)+W_2\left(X_2\right)+W_3\left(X_3\right)+\dots +W_{1P}\left(X_P\right)$$

(2)

where M₁ = score of participants on principal components (1st component extracted), W₁ = the coefficient (weight) of observed variables, and X_p = score of the participants on the observed variable. Therefore, if factor 1 in this study was ‘identification of loan distress signals’, the researcher estimated the scores of each participant by using the data and formula below:

$$M_1=W_1\left(X_1\right)+W_2\left(X_2\right)+W_3\left(X_3\right)+W_4\left(X_4\right)+\dots +W_{17}\left(X_{17}\right)$$

(3)

In the above, the X variables are the observed variables that refer to the participant’s responses to the 17 questions on CRIMs, where: X₁ = question 1; X₂ = question 2; X₃ = question 3; X₄ = question 4; and so on. Hence, subsequent components were separately measured by different constructs using different equations with different coefficients. By this estimation technique, meaningful components were retained, indicating the most frequent methods used by commercial banks to identify credit risk and minimize credit risk exposure in agricultural financing. In line with Braun and Clarke (2006), a thematic analysis was also employed to categorize and present themes extracted from the interviews. The two models were used to synergize with each other to achieve the full rigor of the objective and probe more into strategies used in mitigating credit risks by Ghanaian commercial banks and also validate and complement the findings that were obtained through the questionnaire.

4. Results and Discussion

4.1. Interview Results of CRM Strategies

The interview aimed to explore the best possible strategies that can be used by commercial banks to mitigate credit risk effects in agricultural financing. The results from the interview data suggest that the strategies that can be used to mitigate credit risk exposure in agricultural lending is bank-specific. In Bank A, a participant revealed that the agricultural sector generally has a high risk, making most banks relax in financing the sector’s activities. It was posited that commercial banks can minimize credit risk in agricultural lending by collaborating with incentive-based risk-sharing systems, such as the Ghana Incentive-Based Risk-Sharing System for Agricultural Lending Projects (GIRSAL), to encourage banks to lend more to the agricultural sector. The loan guarantee project aims to enhance agricultural financing and sector transformation through increased capital inflows. In Bank A, a participant again recommended the government intervene in the provision of inputs to farmers, the subsidizing of loans for players in the agricultural value chain, and the recruitment of graduates who studied agriculture in all tertiary institutions. Likewise, the recruitment of graduates from agricultural training institutions for all districts by the government would enable players in the agricultural value chain to obtain regular technical and expert advice to mitigate the credit risk exposure of commercial banks. For example, a participant indicated that,

‘…So, if the government’s intervention in agriculture is high, it brings the number of challenges down’.

This implies that the role of government cannot be overemphasized in ameliorating the challenges in agricultural financing. It was further indicated that commercial banks should set up units purposely in charge of agricultural financing and these units should be managed by technical people who have an agriculture-training background. A participant in Bank B asserted that:

‘…Commercial banks should either have technical units or have an outsourced technical unit or standby that assists farmers who take facilities from them’.

Technical units are required in commercial banks to completely eradicate the challenges of credit risk management associated with agricultural financing. Therefore, technical units are in a better position to closely monitor and assist agricultural borrowers to minimize credit risk exposure in agricultural financing.

It was also suggested in Bank B that commercial banks should collaborate with international bodies to identify and access a cheaper and long-term source of funding to finance agribusiness. In another instance in Bank B, it was elucidated that commercial banks can collaborate with funding bodies, such as the Ghana Exim Bank, African Development Bank, Agence Française de Développement (AFD), and then Out grower and Value Chain Fund (OVCF). Also, group financing, value chain financing, and collaboration with off-takers were highly recommended by all the banks as effective strategies for minimizing credit losses. A participant asserted:

‘…look at the key players within a certain value chain so that if you are financing rice growers, then you look at the chain in totality. …the bank, now links the farmers to buyers. So that we don’t have a situation where you lend and the project is successful but because of lack of marketing the farmer is unable to pay back the loans.’

This is not so in Ghana, since there is no ready market for agricultural products. It is denoted that commercial banks must be involved in value-chain financing by proving already market for agricultural borrowers to generate more income and minimize credit losses. The interview also revealed that commercial banks should invest more time in ‘knowing your customers’ (KYC) before granting loans for agricultural purposes to reduce credit risk exposure. Contrarily, the quantitative results indicate that the credit history of borrowers is not most frequently preferred as the first-option strategy factor in the credit risk management policy implementation process. This is dangerous for the banks because the inability to explore the credit trend of agricultural borrowers will make it difficult to spot dishonest and competent borrowers. The identified strategies, such as agricultural value-chain financing, collaboration with off-takers, incentive-based and risk-sharing schemes, and the proper structuring of agricultural loans excluded in the questionnaire, were some of the strategies unveiled during the interview and significantly complemented the results.

4.2. Analysis of Quantitative Results on Credit Risk Mitigation Strategies

In addition to the qualitative results, the Credit Risk Mitigation Strategies used by Ghanaian commercial banks in agricultural finance were grouped into three components using PCA, as per

Table 1. Total variance analysis of credit risk mitigation strategies.

Component	Initial Eigenvalues			Extraction Sums of Squared Loadings			Rotation Sums of Squared Loadings
	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %
1	5.655	33.264	33.264	5.655	33.264	33.264	4.379	25.759	25.759
2	2.886	16.975	50.239	2.886	16.975	50.239	3.042	17.893	43.652
3	1.106	6.503	56.743	1.106	6.503	56.743	2.225	13.091	56.743
4	0.829	4.875	61.617
5	0.752	4.421	66.038
6	0.733	4.310	70.349
7	0.643	3.780	74.129
8	0.612	3.600	77.729
9	0.591	3.474	81.203
10	0.549	3.232	84.435
11	0.494	2.905	87.340
12	0.407	2.395	89.735
13	0.395	2.325	92.060
14	0.383	2.253	94.313
15	0.359	2.114	96.427
16	0.306	1.802	98.229
17	0.301	1.771	100.000

Extraction Method: Principal Component Analysis. Source: Author’s estimates (2024).

. The variance of the three components accounted for 56.74 percent, a p-value of 0.048 (Bartlett’s test), KMO = 0.892, and Cronbach’s alpha ranged from 0.75 to 0.88, indicating the adequacy and reliability of the data gathered.

In conjunction with the eigenvalue-one criteria, the proportion of variance accounted for, and scree plot, three relevant components were retained, as indicated in Table 1. The scree plot in Figure 1 consists of a list number of components on the horizontal axis and a list of eigenvalues on the vertical axis. Components with eigenvalues greater than 1 indicated on the plot in Figure 1 were retained. Components 1 to 3 produced eigenvalues greater than 1 and were retained for discussion. The first, second, and third components contributed 33.26, 16.98, and 16.98 percent, respectively, to the total variance.

The correlation between the components extracted and the individual variables were reviewed. The review established the constructs measured by each of the three components. For easy interpretation and analysis, the components were rotated. The varimax rotation with the Kaiser normalization approach was performed and the results are displayed in

Table 2. Component transformation matrix.

Component	1	2	3
1	0.840	−0.240	0.486
2	0.184	0.970	0.161
3	0.510	0.046	−0.859

Extraction method: Principal component analysis. Rotation method: Varimax with Kaiser normalization. Source: Author’s estimates (2024).

.

The first components, namely loan review and the documentation in

Table 3. Factor analysis of CRM strategies.

Factors	Loadings
Component 1: (Loan review and Documentation)
(Eigenvalues: 5.655; Variance: 33.26%; Cronbach’s alpha: 0.88)
Credit officers regularly identify loan distress signals	0.711
The loan granting process is regularly reviewed	0.748
There is a regular review of loan portfolio quality	0.779
The credit administration process is regularly reviewed	0.697
Borrowers’ credit report is regularly reviewed	0.747
Borrowers’ performance profile is regularly reviewed	0.714
All credit-related transactions are properly documented	0.586
The overall quality of the loan portfolio is assessed on a timely basis	0.628
Component 2: (Credit Skills Review and Monitoring)
(Eigenvalues: 2.886; Variance: 16.98%; Cronbach’s alpha: 0.80)
Employees’ credit skills are regularly reviewed	0.725
Collateralized transactions are regularly monitored	0.620
Credit officers are regularly trained	0.773
Risk management practices are regularly reviewed	0.781
The flow of the borrowers’ business is regularly monitored	0.720
Loans are guaranteed by the Government or credit associations	0.578
Component 3: (Credit Department Checks)
(Eigenvalues: 1.106; Variance: 6.50%; Cronbach’s alpha: 0.75)
Credit department checks that loans are repaid on time	0.726
The credit department regularly identifies loans with potential credit weaknesses that can cause repayment problems	0.766
Credit risk management guidelines are regularly communicated	0.653
Kaiser–Meyer–Olkin Measure of Sampling Adequacy	0.892
Bartlett’s Test of Sphericity (p-value)	2133.55 (0.048)
Percentage of the total variance explained	56.74%

Author’s estimates (2024).

, indicating the most frequently used strategies adopted by commercial banks in agricultural finance to minimize credit risk, consist of eight factors with an eigenvalue of 5.655, variance explained of 33.26 percent, and a Cronbach’s alpha of 0.88. The factors identified included the regular identification of loans with distressing signals with factor loads of 0.711, regular review of loan granting process with factor loads of 0.748, review of loan portfolio quality with factor loads of 0.779, regular review of credit administration process with factor loads of 0.697, and regular review of borrowers’ credit reports with factor loads of 0.747. Others include a regular review of borrowers’ performance with factor loads of 0.714, the proper documentation of all credit-related transactions with factor loads of 0.586, and the assessment of the overall quality of loan portfolio on a timely basis, with factor loads of 0.628.

The second component, given the name of credit skills reviewed and monitored, comprises six factors with an eigenvalue of 2.886, variance explained of 16.98 (shown in Table 1), and a Cronbach’s alpha of 0.80. The second most frequently used factors were identified as the regular review of employees’ credit skills with factor loadings of 0.725, the regular monitoring of collateralized transactions with factor loadings of 0.620, and the regular training of credit officers with factor loadings of 0.773. The rest includes the regular review of risk management practices with factor loadings of 0.781, regular monitoring of the flow of borrowers’ business with loadings of 0.720, and loan guarantees by government or credit associations. The third component, identified as credit department checks, comprises three factors with an eigenvalue of 1.106, variance explained of 6.50 percent, and a Cronbach’s alpha of 0.75. The factors identified include regular checks by the credit department for the repayment of loans on time, which has factor loads of 0.726, regular identification by the credit department for loans with potential credit weakness that can cause repayment problems (loads of 0.766), and the regular communication of credit risk management guidelines, with factor loads of 0.653.

4.3. Discussion of the Results

The first and most frequently used strategies are the factors identified as part of the first components adopted to mitigate credit risk. The results from the questionnaire data indicate that the identification of loans with distress signals; portfolio quality of loans; regular review of the credit administration process and borrowers’ performance; review of borrowers’ performance and proper documentation are strategies frequently used by commercial banks to mitigate credit risk in agricultural finance. These findings are consistent with the views of Yin et al. (2020), Marinelli et al. (2022), and Dlugosch et al. (2018). The findings also agree with the liquidity theory of credit risk and the adverse selection theory, indicating that borrowers cannot repay loans and interest out of their pocket and that credit reports are very critical in mitigating credit losses. The result implies that commercial banks have effectively adopted reviewing the performance of borrowers as well as credit reports as one of the most significant strategies to mitigate credit risk in agricultural lending. This has a positive implication for commercial banks and the borrowers since it improves the cashflow of the banks and is likely to increase agricultural lending support. These results confirm the findings of Olowa and Olowa (2017), who indicated that an effective banking system ensures the repayment of loans by borrowers to reduce credit losses.

The second most frequently used strategy by commercial banks in agricultural finance represents the factors identified as part of the second component (credit skill review and monitoring). The results indicate that the regular review of employees’ credit skills, regular training of credit officers, monitoring of risk management practices, credit guarantee schemes, and review of borrowers’ business were identified as the second most preferred credit risk management strategies adopted by commercial banks in agricultural finance. Furthermore, the interview results indicate that, even though commercial banks consider experience and training on the job for credit officers to some extent, these banks appoint credit officers from different fields of study, such as Sociology, Science, Accounting, Psychology, or Finance. The problem most frequently encountered stems from the fact that officers with backgrounds other than agriculture are likely to lack technical expertise, unlike credit officers with an agricultural background who have already acquired knowledge on agriculture, as indicated by the interview results. The findings do not conform to the results of Daum and Birner (2017), Monnet and Nellen (2021), and Wubin et al. (2020). They further disagree with the adverse selection theory.

The findings imply that commercial banks have not been effective in monitoring collateralized transactions to improve loan repayment and reduce the credit risk associated with agricultural lending. The interviews confirmed that there is high monitoring costs in agricultural finance and that it is difficult for commercial banks to obtain credit officers in all the regions who can monitor the performance of borrowers as a result of the scattered nature of borrowers across Ghana. This has the potential of increasing loan defaults, which affects cashflows in commercial banks and can practically reduce lending to agricultural sector. The results also indicate the ineffective use of credit guarantee schemes by commercial banks to mitigate credit risk in agricultural lending. The results do not follow the Credit Guarantee Policy indicated by the Borrowers and Lenders Act 2008 (Act 773), where commercial banks are required to request collateral security or guarantors from borrowers as part of the credit agreements before loans are granted. In addition, the results do not support the findings of Wubin et al. (2020), who considered the regular use of credit guarantee schemes as one of the best strategies used to mitigate credit risk and cautioned banks to pay more attention to third-party guarantees to reduce credit risk and prevent losses resulting from credit risk exposure. An important implication of this finding is that the inadequate use of credit guarantee schemes will not only reduce loan recovery rates but will most importantly reduce the level of agricultural credit supply to agricultural borrowers.

The third most frequently used strategies adopted by commercial banks in agricultural finance to minimize credit risk are the factors that are identified as part of the third component (credit department checks). These include checks by the credit department for the repayment of loans on time, identification of loans with potential credit weaknesses that can cause repayment problems, and communication of credit risk management guidelines, among others. It implies that commercial banks do not mostly adopt these strategies in mitigating credit risk. This disagrees with the adverse selection theory, which stresses regular checks for repayment as a good strategy for minimizing loan losses. Also, the results do not align with the findings of Instefjord and Nakata (2022), who posited that risk management guidelines should be communicated to all individuals involved in the risk management process to mitigate credit risk. The findings also contradict the credit referencing theory, which emphasizes information sharing through the effective communication of credit risk to mitigate the credit risk exposure of banks. This is because information sharing on credit risk mitigation cannot be complete without first communicating the nature of credit risk faced by banks regarding agricultural lending.

The interview results also found that commercial banks should collaborate with incentive-based risk-sharing systems, such as the Ghana Incentive-Based Risk-Sharing System for Agricultural Lending Projects (GIRSAL), to mitigate credit risk in agricultural finance. This is in line with the opinion of Muhammad et al. (2018), who posited that banks must collaborate with risk-sharing systems to serve as an incentive to reduce credit risk. Additionally, the interview results suggest that commercial banks should collaborate with international bodies, such as the Ghana Exim Bank (GEB), African Development Bank, and the AFD, to identify and access a cheaper and long-term source of funding for agricultural value-chain financing. The interview results further suggest that commercial banks should make use of the Ghana Agricultural Insurance Pool, which provides insurance against possible loan losses. Commercial banks must adopt value-chain financing and collaborate with off-takers in agricultural financing to effectively manage credit risk to an acceptable level of tolerance. It was also found that commercial banks should invest more in knowing their customers to avoid loans being granted to regular bad borrowers in agricultural financing. As pointed out by the interview results, agriculture is a highly specialized area that needs special attention and investment to develop to effectively reduce credit risk resulting from bad loans.

5. Conclusions

The study investigated the effectiveness of the strategies used by Ghanaian commercial banks to mitigate credit risk in agricultural financing. The findings confirm that some of the strategies that have proved effective in minimizing credit risk are not frequently used by commercial banks to mitigate credit risk in agricultural finance. These include: checks by the credit department for the repayment of loans on time; identification of loans with potential credit weaknesses that can cause repayment problems; information sharing; the use of credit bureau institutions; and communication of credit risk management guidelines, among others. Instead, Ghanaian commercial banks favor strategies such as the regular review of loan granting process, portfolio quality of loan, borrowers’ performance profile, and documentation of credit-related transactions, among others. It was further found that commercial banks need robust credit risk management strategies to mitigate credit risk in agricultural finance. Also, most Ghanaian commercial banks specifically lack technical units and technical people with agricultural training backgrounds to manage the credit related to transactions in agricultural finance. The study contributes to the existing literature on credit risk management practices of commercial banks by closing the empirical gap in the strategies adopted by Ghanaian commercial banks to minimize credit risk in agricultural finance. Lastly, the study provides valuable information for regulators and policy promulgators, such as the government and BoG, to pay critical attention to agricultural finance since agriculture remains the foundation of the Ghanaian economy. The study further recommends commercial banks to collaborate with incentive-based risk-sharing systems, establish technical units purposely in charge of agricultural finance, seek government intervention, adopt a holistic agricultural value-chain financing strategy, and make use of the Ghana Agricultural Insurance Pool to minimize credit risk. As credit risk is associated with a poor loan performance, searching for the major determinants of loan performance in agricultural finance should be the basis of future research.