Project – Credit Risk Assessment and Portfolio Management in Nigerian Banks Using Logistic Regression and Copula Models. A Study of GTB

CHAPTER ONE

INTRODUCTION

1.1 Background of the Study

Credit risk—the risk that borrowers may fail to meet their financial obligations—is a critical concern for banks and financial institutions worldwide. Effective credit risk assessment and portfolio management are essential to ensure bank stability, profitability, and compliance with regulatory frameworks (Altman & Saunders, 1998). In emerging economies like Nigeria, where the banking sector faces challenges such as high default rates, economic volatility, and limited access to reliable borrower information, accurate credit risk modeling is vital (Adeleye & Olowe, 2020).

Guaranty Trust Bank (GTB), one of Nigeria’s leading banks, serves a large customer base and offers various credit products, including personal loans, business loans, and corporate financing. Managing credit risk in GTB’s portfolio requires sophisticated analytical tools that can capture the probability of default and the dependencies among multiple borrowers or asset classes. Traditional approaches, such as credit scoring and basic financial ratios, may fail to account for the complex and interdependent nature of defaults in diversified portfolios (Crouhy, Galai, & Mark, 2014).

Recent advancements in statistical modeling and financial analytics have introduced techniques such as logistic regression and copula models, which provide a robust framework for predicting defaults and modeling dependency structures between financial assets. Logistic regression allows estimation of the probability of default based on borrower-specific financial and behavioral characteristics, while copula models capture nonlinear dependencies and joint default probabilities in a credit portfolio (Embrechts, McNeil, & Straumann, 2002).

The application of these methods in Nigerian banks, particularly GTB, has the potential to enhance risk management, optimize capital allocation, and improve portfolio performance. By integrating logistic regression with copula-based models, the bank can identify high-risk borrowers, assess portfolio-level vulnerabilities, and implement effective mitigation strategies. This study aims to investigate credit risk assessment and portfolio management in GTB using these advanced statistical techniques.

1.2 Statement of the Problem

Credit risk remains one of the major threats to the stability of Nigerian banks. Despite regulatory frameworks such as the Central Bank of Nigeria (CBN) guidelines on risk management, high default rates and non-performing loans continue to challenge banks, including GTB (Olalekan & Adekunle, 2021). Many banks rely on basic credit scoring or manual assessment procedures that may not capture the complexities of borrower behavior, leading to inaccurate predictions and mismanaged portfolios.

Portfolio managers face difficulty in assessing the interdependence between assets and borrowers, which can result in concentrated risks that amplify losses during economic downturns. Furthermore, the lack of reliable predictive models can hinder banks from optimizing capital allocation, leading to financial inefficiencies and regulatory non-compliance (Adegbite, 2020).

The existing literature indicates that logistic regression and copula models can improve predictive accuracy and provide insights into portfolio-level dependencies. However, their application in Nigerian banks, particularly GTB, remains limited, with few empirical studies addressing local borrower characteristics, economic volatility, and data quality issues. Therefore, there is a need to apply robust statistical techniques to assess credit risk accurately and inform portfolio management decisions in Nigerian banks.

1.3 Research Objectives

The main objective of this study is to assess credit risk and improve portfolio management in Nigerian banks using logistic regression and copula models. The specific objectives are:

1. To identify key factors influencing the probability of borrower default in GTB.

2. To develop logistic regression models for estimating default probabilities of individual borrowers.

3. To apply copula models for assessing dependencies among loans and quantifying portfolio-level credit risk.

4. To evaluate the effectiveness of integrating logistic regression and copula models for credit risk assessment and portfolio management.

1.4 Research Questions

The study seeks to answer the following questions:

1. What are the major determinants of borrower default in GTB?

2. How can logistic regression be used to estimate the probability of default for individual borrowers?

3. How do copula models capture dependencies between loans in GTB’s portfolio?

4. How effective is the combination of logistic regression and copula models in managing credit risk and optimizing portfolios?

1.5 Research Hypothesis

H₀: Logistic regression and copula models do not significantly improve credit risk assessment and portfolio management in GTB.

H₁: Logistic regression and copula models significantly improve credit risk assessment and portfolio management in GTB.

1.6 Significance of the Study

This study is important for several reasons:

1. Banking Industry: By providing evidence-based methods for credit risk assessment, GTB and other Nigerian banks can reduce defaults, optimize loan portfolios, and enhance financial stability.

2. Policy and Regulation: Insights from the study can guide regulators such as the Central Bank of Nigeria (CBN) in developing policies for risk management, loan classification, and capital adequacy.

3. Academic Contribution: The study contributes to the limited empirical literature on the application of logistic regression and copula models in Nigerian banking, providing a foundation for further research.

4. Risk Management Practices: Portfolio managers can use the findings to identify high-risk borrowers, manage dependencies among loans, and implement mitigation strategies.

1.7 Scope of the Study

The study focuses on Guaranty Trust Bank (GTB) and its credit portfolio between 2018 and 2025. It covers both retail and corporate loans. Data will include borrower financial information, repayment history, loan characteristics, and macroeconomic variables affecting default probabilities. Methodologically, the study employs logistic regression to model individual defaults and copula models to capture dependencies in the loan portfolio. Geographically, the study covers branches and borrowers across Nigeria.

1.8 Definition of Terms

Credit Risk: The risk of financial loss due to a borrower’s failure to meet contractual loan obligations (Altman & Saunders, 1998).

Portfolio Management: The process of selecting, monitoring, and managing a collection of financial assets to maximize returns while controlling risk.

Logistic Regression: A statistical model used to predict the probability of a binary outcome, such as default or non-default, based on explanatory variables.

Copula Models: Statistical methods that model and simulate the dependence structure between multiple random variables, often used in finance to understand joint default probabilities.

Probability of Default (PD): The likelihood that a borrower will fail to repay a loan within a specified period.

Non-Performing Loan (NPL): A loan in which the borrower has failed to make scheduled payments for a specified period, usually 90 days or more.

Risk Mitigation: Strategies implemented to reduce potential financial losses due to credit risk.

Project – Credit Risk Assessment and Portfolio Management in Nigerian Banks Using Logistic Regression and Copula Models. A Study of GTB