Credit Risk Modeling in Python

Teaching you the programming behind how banks decide who should get a loan. You will learn risk modeling theory and advance your Python modeling skills.

Hours

Lessons

Quizzes

Assignments

Create Your Free Account

Join over 900,000 students

Our graduates work at exciting places

Course description

Credit risk modeling is the place where data science and fintech meet. It is one of the most important activities conducted in a bank and the one with the most attention since the recession. This course is the only comprehensive credit risk modeling course in Python available right now. It shows the complete credit risk modeling picture, from preprocessing, through probability of default (PD), loss given default (LGD) and exposure at default (EAD) modeling, and finally finishing off with calculating expected loss (EL).

Nikolay Georgiev

Introduction

We start by explaining why credit risk is important for financial institutions. We also define ground 0 terms, such as expected loss, probability of default, loss given default and exposure at default.

Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment

Setting up the environment

Here you will learn how to set up Python 3 and load up Jupyter. We’ll also show you what the Anaconda Prompt is and how you can use it to download and import new modules.

Quiz Assignment
Quiz Assignment

Dataset description

Our example focuses on consumer loans. Since there are more than 100 potential features, we've devoted a complete section to explain why some features are chosen over others.

Quiz Assignment
Quiz Assignment

General preprocessing

Each raw datasets has its drawbacks. While most preprocessing is model specific, in some cases (like missing values imputation), we could generalize the data preparation.

Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment

PD model: data preparation

Once we have completed all general preprocessing, we dive into model-specific preprocessing. We employ fine classing, coarse classing, weight of evidence and information value criterion to achieve the probability of default preprocessing. Conventionally, we should turn all variables into dummy indicators prior to modeling.

Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment

PD model estimation

Having set up all variables to be dummies, we estimate the probability of default. The most intuitive and widely accepted approach is to employ a logistic regression.

Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment

PD model validation (test)

Since each model overfits the training data, it is crucial to test the results on out-of-sample observations. Consequently, we find its accuracy, its area under the curve (AUC), the Gini coefficient and the Kolmogorov-Smirnov test.

Quiz Assignment
Quiz Assignment
Quiz Assignment

Applying the PD model for decision making

In practice, banks don't really want a complicated Python-implemented model. Instead, they prefer a simple score-card which contains only yes and no questions that could be employed by any bank employee. In this section, we learn how to create one.

Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment

PD model monitoring

Model estimation is extremely important, but an often-neglected step is model maintenance. A common approach is to monitor the population stability over time using the population stability index (PSI) and revisit our model if needed.

Quiz Assignment
Quiz Assignment
Quiz Assignment

LGD and EAD models

To calculate the final expected loss, we need three ingredients: probability of default (PD), loss given default (LGD) and exposure at default (EAD). In this section, we preprocess our data to be able to estimate the LGD and EAD models.

Quiz Assignment
Quiz Assignment
Quiz Assignment

LGD model

LGD models are often estimated using a beta regression. To keep the modeling part simpler, we employ a two-step regression model, which aims to simulate a beta regression. We combine the predictions from a logistic regression with those from a linear regression to estimate the loss given default.

Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment
Quiz Assignment

EAD model

The exposure at default (EAD) modeling is very similar to the LGD one. In this section, we take advantage of a linear regression to calculate EAD.

Quiz Assignment
Quiz Assignment

Calculating expected loss

After having calculated PD, LGD, and EAD, we reach the final step: computing expected loss (EL). This is also the number which is most interesting to C-level executives and is the finale of the credit risk modeling process.

Quiz Assignment

Credit Risk Modeling in Python

Our graduates work at exciting places

Introduction

What does the course cover (5 min)

What is credit risk and why is it important? (5 min)

Expected loss (EL) and its components: PD, LGD and EAD (4 min)

Capital adequacy, regulations, and the Basel II accord (5 min)

Basel II approaches: SA, F-IRB, and A-IRB (10 min)

Different facility types (asset classes) and credit risk modeling approaches (9 min)

Setting up the environment

Setting up the environment

Installing the relevant packages (1 min)

Dataset description

Our example: consumer loans. A first look at the dataset (3 min)

Dependent variables and independent variables (6 min)

General preprocessing

Importing the data into Python (4 min)

Preprocessing few continuous variables (13 min)

Preprocessing few discrete variables (7 min)

Check for missing values and clean (3 min)

PD model: data preparation

How is the PD model going to look like? (4 min)

Dependent variable: Good/ Bad (default) definition (5 min)

Constructing independent variables (6 min)

Information value (5 min)

Data preparation. Splitting data (8 min)

Data preparation. Preprocessing one discrete variable (8 min)

Data preparation. Preprocessing discrete variables: automating calculations (6 min)

Data preparation. Preprocessing discrete variables: visualizing results (10 min)

Data Preparation. Preprocessing Discrete Variables: Creating Dummies (part 1) (7 min)

Data Preparation. Preprocessing Discrete Variables: Creating Dummies (part 2) (11 min)

Data preparation. Preprocessing continuous variables: automating calculations (5 min)

Data preparation. Preprocessing continuous variables: creating dummies (part 1) (7 min)

Data preparation. Preprocessing continuous variables: creating dummies (part 2) (14 min)

Data preparation. Preprocessing continuous variables: creating dummies (part 3) (13 min)

Data preparation. Preprocessing the test dataset (4 min)

PD model estimation

The PD model. Logistic regression with dummy variables (8 min)

Loading the data and selecting the features (6 min)

PD model estimation (4 min)

Build a logistic regression model with p-values. (11 min)

Interpreting the coefficients in the PD model (6 min)

PD model validation (test)

Out-of-sample validation (test). (7 min)

Evaluation of model performance: accuracy and area under the curve (AUC) (11 min)

Evaluation of model performance: Gini and Kolmogorov-Smirnov. (10 min)

Applying the PD model for decision making

Calculating probability of default for a single customer (5 min)

Creating a scorecard (13 min)

Calculating credit score (6 min)

From credit score to PD (3 min)

Setting cut-offs (9 min)

PD model monitoring

PD model monitoring via assessing population stability (5 min)

Population stability index: preprocessing (12 min)

Population stability index: calculation and interpretation (11 min)

LGD and EAD models

LGD and EAD models: independent variables (6 min)

LGD and EAD models: dependent variables (5 min)

LGD and EAD models: distribution of recovery rates and credit conversion factors (6 min)

LGD model

LGD model: preparing the inputs (3 min)

LGD model: testing the model (5 min)

LGD model: estimating the accuracy of the model (5 min)

LGD model: saving the model (3 min)

LGD model: stage 2 – linear regression (4 min)

LGD model: stage 2 – linear regression evaluation (4 min)

LGD model: combining stage 1 and stage 2 (3 min)

EAD model

EAD model estimation and interpretation (6 min)

EAD model validation (4 min)

Calculating expected loss

Calculating expected loss (16 min)