Exposure at Default (EAD)

Exposure at default (EAD) is an estimate of the total amount a bank will be owed by a borrower or counterparty at the time of default. For a simple term loan it is close to the outstanding balance. For facilities that the borrower can draw on at will, such as overdrafts, credit cards and revolving credit lines, it also includes the part of the unused limit that the borrower is expected to draw before defaulting. For derivatives it reflects how the value of the contracts might move before the counterparty fails.

EAD is one of the three main parameters of credit risk measurement, with probability of default and loss given default. Together they determine expected loss, and they are used in bank capital rules under the Basel framework, in the calculation of expected credit loss for accounting, and in pricing and limit setting.

For a facility with an undrawn portion, the usual calculation is:

Here CCF is the credit conversion factor, the share of the unused commitment expected to be drawn by the time of default. Expected loss then follows as:

Suppose a company has a credit line of 1,000,000 and has drawn 600,000. With a conversion factor of 40%, EAD is $600{,}000 + 0.40 \times 400{,}000 = 760{,}000$. If the probability of default over the next year is 2% and loss given default is 45%, the expected loss is $0.02 \times 0.45 \times 760{,}000 = 6{,}840$.

Borrowers in difficulty tend to draw down available credit before they default, since other sources of funds dry up. Studies of defaulted facilities consistently find that balances at default are higher than balances a year earlier. A bank that measured exposure only by the current drawn amount would understate its risk on any credit line. The conversion factor captures this behaviour, and it varies with the type of product, the borrower's quality, how tightly the bank can cancel or reduce the credit limit, and how far the borrower is from default.

Off-balance-sheet items other than commitments, such as guarantees and letters of credit, are also converted into credit equivalents, because they become claims on the customer if called.

Under the standardised approach for credit risk, supervisors prescribe the conversion factors. In the final Basel III standards, commitments that the bank can cancel unconditionally at any time receive 10%, other commitments 40%, transaction-related contingent items such as performance bonds 50%, short-term self-liquidating trade letters of credit 20%, and direct credit substitutes such as financial guarantees 100%.

Under the internal ratings-based approach, banks on the foundation version use supervisory conversion factors, while those on the advanced version may use their own estimates of EAD, subject to floors; the final reforms restricted own estimates largely to revolving facilities. In every case EAD is multiplied by a risk weight to give risk-weighted assets. Eligible collateral and netting recognised as credit risk mitigation can reduce the exposure amount.

For derivatives, the amount at risk is not a fixed sum but the replacement cost of the contracts if the counterparty fails, which changes with market prices. This is the domain of counterparty risk. The Basel standardised approach for counterparty credit risk calculates EAD as a multiple of the sum of current replacement cost and an add-on for potential future exposure, taking account of close-out netting and margin. Banks with supervisory approval may instead model the expected exposure profile over time. Repos and securities lending are measured after haircuts on the securities exchanged.

Under IFRS 9, the loss allowance is a probability-weighted estimate of credit losses, and most banks build it from PD, LGD and EAD for each future period. The accounting measure differs from the regulatory one in several ways. It must reflect current and forecast conditions and not a downturn; it follows the expected path of the balance, including scheduled repayments and prepayments; and for exposures with a significant increase in credit risk it extends over the remaining life of the facility. For revolving products such as credit cards, the standard allows the expected life to exceed the contractual notice period, because banks do not in practice cancel limits immediately.

Banks estimate EAD from their own history of defaulted facilities, comparing balances at default with balances and limits at earlier reference dates. The work is harder than estimating default rates. Defaults are few, limits change, and results are sensitive to the reference date chosen. Estimates for facilities that are already almost fully drawn can be unstable. Behaviour also shifts in a crisis: in March 2020 many companies drew their credit lines at once as a precaution, far beyond normal patterns. Regulators therefore require conservative margins, downturn estimates and regular back-testing of the models.

• Bank for International Settlements: Basel Framework