Loan Loss Scenarios for Individually Assessed Provisions

When a loan or credit instrument is in default, a estimate of the likely the recoverable amount is required. On many occasions defaulting customers ultimately make good on their obligations, but even if a strict workout through repossession and sale of collateral is required the ultimately amount recovered is unknown until the end of the workout. Accounting requirements and good risk management require a lender to consider a variety of potential workout pathways and collateral valuations in determine the current value of a financial instrument in default.

Realised loss calculator

For an impaired facility, the factor determining whether a loss occurs or does not occur depends on:

• the recoverable value of the property;
• the time taken to realise the proceeds of sale and the present value of those proceeds;
• the interest accrued over the intervening period;
• workout costs and overheads incurred.

These factors are reflected in the following calculation:

where:

• is a discount factor;
• is the workout period, and w is each month throughout the workout period;
• are the monthly collections costs;
• is the customer interest foregone over the workout period;
• is the recovery proceeds, the colateral sale at the end of the workout period (W).

Loan amount and terms

Applying this calculation first requires a reference to the expected loan value, the principal and interest foregone to the time of default and expected cashflows over the workout period. These are derived from the loan contract and the current balance, including capitalised interest.

Confidence interval

In the following analysis the requirement for multiple scenarios is achieved through parameterising the distribution of the unknown quantities (property values and workout times) and considering a good, mid and bad scenario defined according to a required confidence interval. The confidence interval allows for scenarios to be explored in the extremes of the distribution of potential values and workout times. In the following analysis the confidence interval used is:

Uncertainty in property values

A mortgage possession is a possible outcome, the value of the property on the market is an approximation at the best of times and may be further depressed by the poor standard of property given the financial difficulties of the borrower. The FSD is used to represent the uncertainty inherent in the valuation.

As the seller of the property sale costs are also incurred including marketing costs, agents commission and stamp duty. These costs need to be covered before the remaining proceeds of sale can be shared between the customer and the bank. Property sales costs, both commissions and duties are usually expressed as a percentage of the property value.

Uncertainty in workout times

A loan workout takes time and incurs costs, the provision allowance should consider the recurring costs incurred and their expected timing. In this calculator costs are incurred on a simple constant monthly basis, for example month till the workout is finalised. In practice costs might vary according to the workout strategy and the likely methods of recovery:

• cures are cheap and over quickly;
• restructures can be complex require some evidence that the customer has ulitmately recovered, will therefore take longer;
• mortgage possession and court orders are expense, require formal notification periods and therefore take longest and cost the most.

The workout assessment requires scenarios and financial value should be expressed net present value terms. The effective interest rate is the reported measure of loan's ongoing financial value and is therefore used as the discount rate for any future recoveries and costs.

Nominal holding costs over the workout period, based on the three possible workout scenarios are:

The following table summarises the range of potential total loss provision amounts by workout period and property value, assuming the property valuation outcome and the workout period are independent. The provision amounts are greatest for the longest workout with the lowest property price outcomes, and lowest for short workouts with strong property prices.

The probability distribution for each of scenario is derived from the outer matrix multiplication of the probability distributions for each price scenario and each workout period, for example:

The matrix of probabilities and the matrix of losses for each scenario can be combined to provide the probability weighted expected credit loss estimate for the defaulted facility:

In this example the probability weighted and un-weighted mid scenario expected loss amounts are summarised in the following table. The difference between these two loss estimates is often attributed to convexity, that the reduction in loss on the better than expected scenarios is not equivalent to the increase in loss on the worse than expected scenarios. This is due to:

• over-recovery, where the collateral is worth more than the odutstanding loan amount is paid to the customer;
• loss amounts compound through interest accrual over the workout period.

The non-linearity of the escalation of losses is the primary motivation behind the inclusion of a probability weighted loss estimation process in the accounting standards.

The probability weighted loss is: .