You are on page 1of 4

Credit Risk: How to Calculate Expected Loss

& Unexpected Loss


When measuring and managing credit risk, It is important to have a clear understanding of common terms such as
expected loss and unexpected loss. As stated in Credit Risk Assessment, credit risk is defined as the potential that a
bank borrower or counterparty will fail to meet its obligations in accordance with agreed terms. This article will
explain the calculations for expected loss (EL) and unexpected loss (UL), for both an individual asset and for a
portfolio. Additionally, risk contribution (RC) will be defined.

To begin, lets review the common terminology used in these calculations:


Terminology

Probability of Default (PD) is the likelihood that a loan will not be repaid and will fall
into default. It must be calculated for each borrower. The credit history of the borrower
and the nature of the investment must be taken into consideration when calculating PD.
External ratings agencies such as Standard & Poors or Moodys may be used to get a PD;
however, banks can also use internal rating methods. PD can range from 0% to 100%. If
a borrower has 15% PD it is considered a less risky company vs. a company with a 30%
PD.

Adjusted Exposure (AE) is equal to outstanding loan amount (OS) plus the percentage
of unused loan commitment (COM) drawn-down by the borrower; known as the usage
given default (UGD). AE is calculated as follows:
AE = OS + (COM OS) * (UGD)

Recovery Rate (RR) the proportion of a bad debt that can be recovered.

Loss Given Default (LGD) the credit loss incurred if an obligor of the bank defaults. LGD = 1- RR

Expected Default Frequency (EDF) refer to Probability of Default

Expected Loss (EL) referring back to Expected Loss Calculation, EL is the loss that can be incurred
as a result of lending to a company that may default. It is the average loss in value over a specified period.

Unexpected Loss (UL) it is known as the variation in expected loss. UL is typically larger than EL
but they are both equal to zero when PD is zero.

Expected Loss (EL)


EL for a single asset is calculated by using the following formula:
EL = AE * LGD * EDF
To calculate EL for a portfolio we must add the expected losses of the individual assets; formula below:
ELP = ELi
Unexpected Loss (UL)
UL for a single asset is calculated by using the following formula:
UL = AE * SQRT [(EDF * LGD) + (LGD * EDF)]
To calculate the UL for a portfolio, we need to use a more complex formula. Below is the formula for a 2 asset
portfolio:
ULP = SQRT [ULi + ULj + 2 (i,j) (ULi) (ULj)]
Note: The UL of the portfolio will be less than the sum of the ULs for the individual assets due to diversification
benefits.
Risk Contribution
Each asset in a portfolio contributes a portion of its UL to the portfolio UL; known as risk contribution (RC). In
other words, it is the incremental risk that the exposure of a single asset contributes to the total risk of the portfolio.
The formula for calculating the RC of asset i and j for a two asset portfolio is illustrated below:
RCi = ULi * [ULi + (i,j) * (ULj) / ULp]
RCj = ULj * [ULj + (i,j) * (ULi) / ULp]
If we take the sum of RCi + RCj we will get the same value as the unexpected loss on the portfolio.

Correlation () plays an important role in measuring the potential UL in the portfolio. If correlation between the
assets increases, the potential UL will increase.
Sources:

Expected Loss (EL) Calculation


Lending institutions need to understand the loss that can be incurred as a result of lending to a
company that may default; this is known as expected loss (EL). EL can be expressed as a

simple formula:
EL = PD * LGD * EAD
The total exposure to credit risk is the amount that the borrower owes to the lending institution at
the time of default; the exposure at default (EAD). Generally, EAD will not be larger than the
borrowing facility.
PD & LGD are risk metrics employed in the measurement and management of credit risk. The
metrics are used to calculate EL.
The probability of default (PD) is the likelihood that a loan will not be repaid and will fall into
default. It must be calculated for each borrower. The credit history of the borrower and the nature
of the investment must be taken into consideration when calculating PD. External ratings
agencies such as Standard and Poors or Moodys may be used to get a PD; however, banks can
also use internal rating methods. PD can range from 0% to 100%. If a borrower has 50% PD it is
considered a less risky company vs. a company with an 80% PD. For example:
A borrower (Company X) takes out a loan from Bank ABC for $10 million (EAD). Company X
pledges $3 million collateral against this loan (for simplicity, lets say the collateral is cash). The
Companys PD is determined by analyzing their credit risk aspects (evaluate the financial health
of the borrower, taking into account economic trends, borrower relationship with the bank, etc.)
For Company X, lets say the PD is 0.99. This means that the Company is extremely risky; the
probability of them defaulting on the loan is 99%.
Loss given default (LGD) is the fractional loss due to default. Continuing from the previous
example:
If Company X defaults (is unable to pay back the $10 million to Bank ABC), the Bank will be
able to recover $3 million (this is the cash-secured collateral).
So, how do we calculate the actual loss given default (LGD)?

LGD = 1 Recovery Rate (RR)


The Recovery Rate (RR) is defined as the proportion of a bad debt that can be recovered. It is
calculated as:
RR = Value of Collateral/Value of the Loan
Back to our example, the recovery rate for Bank ABC = $3 million/ $10 million = 30%
So % LGD= 1- 0.30 = 0.70 or 70%.
$ LGD= 70% of a $10 million (EAD) loan is equal to $7 million.
$ LGD = $7 million
Expected Loss (EL) is what a bank can expect to lose in the case that their borrower defaults. It
is calculated below:
EL = PD * LGD * EAD
EL= 0.99* 70% * $10 million
EL = $6.93 million
Bank ABC can expect to lose $6.93 million.
Sources:
Financial Risk Manager Handbook: (Fifth Edition) Philippe Jorion Introduction to Credit Risk
(pg 434)
Wharton: What Do We Know About Loss Given Default? Til Schuermann
http://fic.wharton.upenn.edu/fic/papers/04/0401.pdf