What it means
The idea is close to an allowance for bad debt but broader and more anticipatory. It covers loans, lease receivables, debt securities held to collect and ordinary trade receivables, and it rests on expectations about the future rather than on evidence of trouble already apparent.
That forward-looking approach was the major accounting change of the last decade. Under the older incurred-loss model, lenders could not provide for a downturn until it had started to bite, which meant provisions arrived late and amplified the very cycle they were meant to absorb.
It matters because the allowance is often the single largest estimate in a lender's accounts and can move profit dramatically. A modest change in assumed default rates across a large portfolio can swing reported earnings by tens of millions, so analysts examine the assumptions as closely as the reported result.
The mechanics usually combine three inputs: exposure at default, the probability that a borrower defaults, and loss given default, which is the share of the exposure not recovered after any security has been sold. Multiplying the three produces an expected credit loss, which is then adjusted for forecasts of unemployment, interest rates and other economic conditions.
The nuance most people miss is that the allowance is not a prediction about any single borrower. It is a probability-weighted estimate across a portfolio, so a lender can be broadly right in aggregate while being wrong about every individual account within it.
In practice
Real-world examples
Example 1
A regional bank increases its allowance ahead of an expected slowdown in construction. Reported profit falls that quarter even though arrears have not risen, which the bank explains to investors as a forward-looking adjustment rather than a deterioration in the current book.
Example 2
An equipment leasing company splits its portfolio into three risk stages. Performing leases carry a twelve-month expected loss estimate, while leases showing a significant increase in credit risk carry a lifetime estimate, which materially raises the allowance on that slice.
Example 3
A large manufacturer applies the same expected loss thinking to trade receivables using a simplified matrix by customer segment and ageing bracket. Export customers in one region attract a higher loss rate because payment delays there have historically been longer.
Think of it
Imagine baking pies and selling them to neighbours on credit. You know from experience that about one in ten people will forget or be unable to pay you back. Instead of acting shocked every time it happens, you set aside a slice of flour upfront to cover the losses.
Formula
Calculation
Expected credit loss = Exposure at default x Probability of default x Loss given default
Worked example: a specialist lender holds a portfolio of small business loans with a total exposure of $10,000,000. Based on its own history, it assesses a 3% probability of default over the next twelve months, and it expects to recover 60% of any defaulted balance through security and collections, giving a loss given default of 40%.
Expected credit loss = $10,000,000 x 0.03 x 0.40 = $120,000.
That $120,000 is the allowance, deducted from the loan balance so the portfolio is carried at $10,000,000 - $120,000 = $9,880,000.
Now suppose the economic forecast worsens and the lender raises its probability of default to 5% while keeping loss given default at 40%.
Revised expected credit loss = $10,000,000 x 0.05 x 0.40 = $200,000.
The allowance must therefore rise by $200,000 - $120,000 = $80,000, and that $80,000 is charged to profit in the current period even though not one borrower has yet missed a payment.Case study
Seen in the real world
This is an illustrative example using a fictional lender. Ashborne Credit Partners, an invented provider of asset finance to small hauliers, carried a $10,000,000 book with an allowance of $120,000, reflecting a 3% default probability and 40% loss given default. Fuel prices then rose sharply and two of its larger customers began paying late.
Rather than waiting for actual defaults, Ashborne raised its assumed default probability to 5%, lifting the allowance to $200,000 and taking an $80,000 charge in the quarter. Its board found the timing painful, since profit fell before any loss had crystallised, but the earlier recognition gave the credit team a mandate to tighten lending criteria and restructure two accounts. When defaults did arrive the following year at close to the revised estimate, the results were far less volatile than they would have been under the old approach of providing only once the loss was already apparent.
Watch out
Common mistakes
- Treating the allowance as a prediction of specific defaults rather than a probability-weighted estimate across a whole portfolio.
- Leaving loss assumptions unchanged when the economic outlook shifts, which defeats the entire purpose of a forward-looking model.
- Reading an increase in the allowance as proof that borrowers are already failing, when it often reflects a changed forecast rather than a changed book.
Questions
People also ask
How is this different from an allowance for bad debt?
It is the same underlying idea applied more widely and more formally, covering loans and other financial assets and requiring expected future losses rather than losses already incurred.
What does loss given default mean?
It is the percentage of an exposure that a lender expects not to recover after selling any security and pursuing collection, so a 40% figure implies 60 cents in the dollar is eventually recovered.
Does a bigger allowance mean the lender is in trouble?
Not necessarily, since a larger allowance can simply reflect a cautious view of the economy, and a lender that provides early is usually better prepared than one that provides late.