This example shows the ability of Palisade's PCR to run simulations against a predefined credit data set, connected to a financial institution's database. This database includes customers and their credit rating and current balance, plus tags identifying products, locations and potentially any other category used. The example permits consolidating all this information and performing expected and unexpected loss analyses.
The model calculates the Credit Risk Loss equation which makes reference to the probability of a borrower not being able to comply with his required payments (PD), the LGD (Loss Given Default), which references how much money will be lost if the customer defaults and the EAD (the current balance).
For calculating the Expected Loss, the application simulates customer's behavior using a Bernoulli distribution for the Probability of Default, and also draws samples from another Distribution for the LGD, and sums the potential losses for all customers in each iteration. To calculate the Expected and Unexpected losses, a percentile is selected from the distribution being generated in each calculation's iteration.
The model also permits modifying the rating types, the LGD distributions and adapting the model to the customer's categories such as Branch, City and others.