Retail credit risk and provisioning in late 2022

Interest rates are rising and many individuals and businesses are under stress. On 21 October 2022 the FCA published a snapshot of its Financial Lives survey. The headline is that 7.8 million people are finding it a heavy burden to keep up with bills. To quote from the survey: “One in four UK adults are in financial difficulty or could quickly find themselves in difficulty if they suffered a financial shock, and 4.2 million people have missed bills or loan payments in at least three of the six months before the survey took place.”

Banks benefit from rising interest rates as this generally leads to an improvement in net interest margins (being the difference between the interest they earn on loans and the cost of their funding). But as we approach the 2022 year-end, we wait to see how the stress felt by their customers (borrowers) might affect banks’ credit impairment losses and, consequently, how the benefits arising from improved net interest margins are impacted by any increased credit impairment losses.

IFRS9 defines a credit impairment loss as "the difference between all contractual cash flows that are due to an entity in accordance with the contract and all the cash flows that the entity expects to receive (i.e. all cash shortfalls), discounted at the […] effective interest rate. […] An entity shall estimate cash flows by considering all contractual terms of the financial instrument (for example, prepayment, extension, call and similar options) through the expected life of that financial instrument. The cash flows that are considered shall include cash flows from the sale of collateral held or other credit enhancements that are integral to the contractual terms."

This article firstly explains the factors or risks that affect the cash flows that a bank expects to receive in respect of its retail lending; and whether those factors are in existence in the UK in late 2022. The article then considers how those factors affect the credit impairment loss calculation under IFRS 9 and what financial statement preparers will need to think about. This section has a particular focus on retail mortgages and credit cards. Finally, the article, summarise the implications and challenges for banks and auditors.

Retail credit risk

Credit risk is the risk that borrowers do not meet their obligations when due – i.e., they do not pay the interest and/or principal on a loan when it falls due. Although many factors might lead to an increase in credit risk, a key driver is when there is an increased risk of default. Examples of situations that could put borrowers in financial difficulty and reduce their ability to make payments include:

• Rising interest rates increase the size of the obligation to be paid in terms of higher interest charges. Higher levels of indebtedness are associated with higher levels of credit risk, as the borrower is more exposed to the effects of rising interest rates. As individuals come under financial pressure, they may increase the usage of credit card or other unsecured lending facilities, thereby increasing their indebtedness and future interest costs. As no one wants to lose their home, it’s not uncommon for mortgage borrowers to avoid defaulting on their mortgages through seeking other forms of available credit. This comes at a cost and is likely to adversely impact a customer’s credit score and pose an increased risk to lenders and prospective lenders
• It is the current nature of the UK mortgage market that borrowers typically have a period of fixed interest rates, with the expectation of refinancing with their current or another provider at the end of this period. In times of stress, lenders typically reduce the availability of new loans and demand higher fees and interest charges. Borrowers, particularly those with lower credit ratings, may find it harder to refinance and if they cannot their rate of interest may move onto the standard variable rate (“SVR”) or a bank’s follow-on rate (“FoR”) for that particular mortgage.[1]
• Increasing inflation means higher living costs (especially on energy and food) thereby reducing disposable income for repayment of debt. This becomes more acute where wages are not keeping pace with inflation.
• Rising unemployment dramatically reduces disposable income as state benefits do not provide equivalent compensation. Rising unemployment is typically linked to the performance of the economy and its effect on companies.[2]

Borrowers with lower levels of income, higher levels of debt and lower credit scores are more susceptible to getting into financial difficulty when macroeconomic conditions deteriorate.

Some borrowers have savings. They can provide a buffer to absorb any increased costs delaying the point at which credit risk may crystalize. A recent survey by The Money and Pensions Service found, however, that one in six UK adults has no savings, and a quarter have less than £100. Around 2% of adults have reported using support from charities according to ONS.

In the face of financial difficulties, borrowers may make decisions or look to make savings that could prove to be a false economy and/or may exacerbate their financial difficulty. For example, they may cancel or reduce insurance cover to save on premiums, but then suffer a loss if the previously insured event happens. The use of ‘loan sharks’ may increase and it’s likely there will be an increase in scams seeking to take advantage of an individual’s financial worries and concerns.

UK economic and market conditions in Q3/4 2022

Many of these credit risk factors are starting to manifest in the UK in recent months. The Financial Policy Committee’s (FPC) record of its policy meeting on 30 September 2022 highlighted the current effects of inflation and rising interest rates in the UK:

"In the UK, higher inflation and rising interest rates will weigh on households repaying debt. Rising interest rates will also increase debt-servicing costs for UK corporates, while higher input costs and lower household demand will impact business earnings.

The continued rise in living costs and interest rates will put increased pressure on UK household finances in coming months and make households more vulnerable to shocks. Overnight swap rates, which feed directly into mortgage interest rates, were, at the time of the FPC Policy meeting, priced to peak at around 6%. Assuming rates follow this market implied path, the share of households with high cost-of-living adjusted mortgage debt servicing ratios would increase by end-2023 to around the peak levels reached ahead of the global financial crisis (GFC).

However, households are in a stronger position than in the run-up to the GFC, so UK banks are less exposed to household vulnerabilities. In particular, the ratio of aggregate household debt (excluding student loans) to income is well below the pre-GFC peak and the share of outstanding mortgage debt at high loan-to-value ratios is much lower.

Nevertheless, it will be challenging for some households to manage the projected rises in the cost of essentials alongside higher interest rates."

The Bank of England’s Monetary Policy Committee (MPC) increased the bank rate by 0.75% to 3% on 3 November 2022, the highest rate since the global financial crisis.

In the MPC Report, the Bank’s central projection has GDP continuing to fall throughout 2023 and 2024 (i.e. a recession). As of 11 October 2022 the UK unemployment rate has not yet shown a tendency to rise: at 3.5% it is down on the previous three months and below pre-pandemic levels according to ONS. Nevertheless, as the Bank projects a fall in GDP, it expects unemployment to rise to just under 6.5%, almost doubling from current levels. Companies are not affected evenly by economic downturns: unemployment in some industries and/or UK regions may be worse, whereas in others it could be better.

Approach to calculating a loan loss allowance: a recap of IFRS 9

Banks recognise a loss allowance for their loan exposures by determining an expected credit loss (‘ECL’).

Loan exposures are classified into one of three stages, which determine the nature of the ECL. In stage 1 the loss is a 12-month ECL; in stages 2 and 3 the loss is a lifetime ECL [3]. An exposure moves from stage 1 to 2 if it has experienced a significant increase in credit risk [4] (”SICR”) since initial recognition. Exposures that are impaired are moved to stage 3.

The ECL reflects

i) an unbiased and probability weighted amount that is determined by evaluating a range of possible outcomes;

ii) the time value of money; and

iii) reasonable and supportable information that is available without undue cost or effort at the reporting date about past events, current conditions and forecasts of future economic conditions. (IFRS 9, paragraph 5.5.17).

See the Faculty’s webpages for previous articles on IFRS 9.

Banks approaches to determining expected credit losses

To calculate an ECL, banks typically use models to estimate the probability of default of loans (the “PD”), the likely loss given default (the “LGD”) and the likely exposure value at the point of default (the “EAD”). Models are typically calibrated by reference to past historical data, for example the number of loan defaults over a given period for PD.

To calculate a probability weighted amount, banks typically model the credit losses arising over a number of macroeconomic scenarios, typically four or five for most UK banks – this includes a central base case and a combination of up- and down-side scenarios. The losses arising under each scenario are weighted by the risk or likelihood of the scenario occurring.

Issues that affect all retail loan types

A key feature of ECLs is that it is an estimate of the future and the future is inherently uncertain. In the current economic climate that uncertainty is increased – so making estimates about the future becomes more difficult and challenging requiring even more judgement. There are three notable areas that involve assumptions about the future.

The first is estimating whether there has been a SICR [5], as the measurement of credit risk needs to be forward looking and comprehensive. Banks do not typically monitor retail loans at the individual level unless they are in financial support. The data for the main objective indicators used (which include past due status) typically lags any actual decline in credit worthiness which often only arises much later. This means that identifying whether there has been SICR involves judgement and banks look at other indicators of SICR. Banks’ quantitative criteria are typically based on whether any increase in the lifetime PD since loan origination [6] exceeds a set threshold both in relative and absolute terms. Qualitative criteria, on the other hand, include whether customers have been given forbearance, whether they have defaulted on loans or facilities in the past 1-2 years and other behavioural score indicators. Previous bankruptcy and insolvency might be other qualitative triggers.

An economic downturn adversely affects most business sectors. Some sectors may be more negatively affected than others (e.g., currently those that have high energy usage) and so increases in credit risk may not be uniform. There may be a need to look at different geographic areas and concentrations if banks have exposures to particular cohorts of lending that are higher risk and more susceptible to financial difficulty.

Banks will also need to reflect on the implications of loan contract renegotiation on the change in underlying credit risk, and whether there has been SICR. For example, a renegotiated loan might mean a loan appears to be performing, but the reason for renegotiating might be due to affordability and so should not lead to recognition of an improvement in credit risk [7].

The second is estimating the base case, upside and downside economic scenarios for the purpose of determining the probability weighted amount. Ongoing economic changes mean that future projections will likely need updating to reflect new estimates. It seems likely that economic scenarios will be weighted more towards the downside in the current environment. The febrile economic environment also makes it more difficult to project the path of future economic variables. Rapid changes in UK Government fiscal policy increases the difficultly, though it is to be assumed that the recent Autumn Statement signals an end to these changes.

The PRA also considers it essential that banks develop capabilities to perform more comprehensive economic sensitivity analysis more quickly. From its Written Auditor Reporting it finds limitations in banks’ abilities to respond to events shortly before the reporting period ends, and to develop scenarios that explore vulnerabilities in specific sectors or segments.

A common assumption in economic projections is ‘mean reversion’ – i.e. after an initial shock to the economic variables (e.g. jump in oil prices), they revert to a long-term average. Banks will need to consider over what period to model the shock, how quickly to revert to the long-term average, and what the long-term average is. Any fundamental shifts in the economics are likely to mean a change in the future potential long-term average.

The third is calibrating models based on historical data given the past is no longer a reasonable proxy for the future, whether that be individual parameters (e.g. inflation) or the correlations (i.e. relationships) between parameters. Models may need to be adjusted. More likely, post- model adjustments or overlays may be required to compensate for model and data weaknesses, which is likely to mean more judgement is involved in the estimate. To ensure estimates are as ‘reliable’ as possible good processes and governance will be required.

Issues to consider affecting mortgages

ECL is an estimate of future cash flows. Many mortgages have a fixed period, after which the mortgage reverts to SVR or FoR. It is common for borrowers to refinance when the fixed period ends – either with the same provider or a different one. As noted above, the availability of refinancing or remortgaging may be curtailed and moving to SVR or FoR is likely to result in higher interest charges. There is a potential change in activity or behaviour that banks will need to factor into estimates of future cash flows (i.e. capture through their EAD models). Such behavioural changes may also require banks to review assumptions in respect of their effective interest rate (“EIR”) methodologies which also make assumptions about customer behaviour.

A bank’s recovery strategies for problem and delinquent loans will also affect the future cash flows: whether that be repayment of principal and interest due or the costs a bank may incur in recovering a loan. The recovery strategies may work well in normal circumstances but may need to be adapted in stressed conditions. Banks will need to consider whether their assumed recovery strategies are feasible in their projected economic scenarios. For example, in the case of mortgages, banks’ scenarios will need to take account of changes in house prices.

In benign times a bank works with its customer to resolve any issues but in some cases may ultimately need to take possession of the property and sell it to recover its loan. In a stressed period where there are multiple defaults, multiple possessions and sales may lead to further house price contraction, potentially worsening the cash recoveries (also to the detriment of the borrower). Banks may look to provide more forbearance, at least in the short-term, to avoid triggering a downward cycle [8].

Separately, rising interest rates and increased borrower hardship feeds into reduced housing demand and house prices falls. As the property is the security for a mortgage loan in the event of borrower default, when house prices fall, there is an increased risk that the value of the property is insufficient to cover the outstanding balance on the mortgage, leading to an increased loss on default that LGD models will need to capture.

Issues to consider affecting credit cards

Credit cards do not have a contractual maturity date. Banks have to estimate the life of a credit card for the purpose of estimating a lifetime loss. This is typically based on past customer behaviour. If customers reduce periodic payments or take longer to pay because of financial difficulty, the behavioural assumptions about maturity may need to be adjusted.

Customers may also increase their credit card usage or draw upon their credit card limits as a means of accessing additional funds more quickly. Assumptions about the outstanding amount (or exposure) at default may require adjustment.

Credit cards typically carry higher interest rates. Continued usage of credit cards without paying down the principal potentially adds to the financial difficulty faced by borrowers. There may be second order effects that need to be captured.

Conclusion

It is likely that banks will report an increase in ECLs in their 2022 year-end financial statements due to the economic uncertainty and the increased weighting of scenarios towards the downside. Banks face a number of challenges in estimating ECLs as noted above. As a result, more judgement will be needed. It is important that good governance can be evidenced over this, including over what information was used and which alternatives were evaluated as well as the rationale for reaching conclusions.

Bank auditors will also face a challenge when auditing ECL. The uncertainty of the current economic climate increases the inherent audit risk associated with estimates. Where an estimate involves greater use of judgement, there is typically less reliable audit evidence available. More than ever auditors will need to review and test banks’ internal processes and governance arrangements to help ensure that judgements involved in ECL estimates are subject to robust scrutiny and challenge. Where control deficiencies are identified, assessed risk levels will be higher and typically there would be an increase in the level of independent substantive testing performed.