UK retail banks switched to the IFRS9 accounting standards for expected loss calculations in 2018. While IFRS9 has many facets which banks globally are coming to terms with, recent developments from the COVID-19 pandemic have brought questions about its impact and the challenges of shock scenarios to center stage.

The current situation has multiple aspects that a bank’s regulatory teams must keep in mind as part of ECL estimation in the near term:

  1. The current and future state of the UK economy in terms of its post-2008 financial crisis status, vulnerabilities related to Brexit, rising non-mortgage consumer indebtedness, and concerns around house prices bubbles, especially in London and South of England.
  2. Regulatory guidance provided by the Prudential Regulatory Authority and Bank of England recommending banks make well-balanced and consistent decisions in accounting practices.
  3. Recent measures taken on the fiscal and monetary front and their “cushion effect” in the near term.
  4. The length and persistence of lockdowns and their short- and long-term impacts on ECL estimation.

All of these factors will play an important role in defining ECL strategy and forecasting volatility in the near term.

Moreover, some or all of the factors can be expected to evolve going forward in ways that can make ECL estimation even more challenging, especially when it is key to remain more accurate than conservative.

The Pre-COVID-19 Macro Situation in the UK Here we would take a quick look at the state of the UK economy just before entering into the COVID-19 shock. Post-recovery from 2008 financial crisis, we clearly see a trend of strong economic growth, low unemployment, and benign inflation from 2013 onwards.

Unemployment rates in 2019 were at a historic low of around 3.8% in UK. Interestingly, some of the trends towards temporary employment in the aftermath of the financial crisis had reversed in the last eight years of continuous and strong economic performance.


GDP growth rates have been in the positive territory ever since the beginning of 2010, however the growth was modest in 2019 due to Brexit related uncertainties. We expect the full year GDP growth of -2% to -3%.

While modelling for COVID-19 scenarios, banks need to consider key aspects of the current economic and regulatory environment to establish a consistent and robust framework. Such a framework is expected to deliver accurate outcomes even as the COVID-19 situation changes rapidly.


Looking at the level of consumer indebtedness, a notable de-leveraging happened in the last decade, with household debt-to-income hovering around 127% as of late 2019.


However, we observe significant increase in nonmortgage consumer credit over the last five years, which is a concern for lower income groups who may struggle to repay this debt and may have a knock-on effect on mortgages as well. This is the first risk in the macro picture of the pre-COVID scenario.


Inflation and interest rates also looked benign, keeping the last two decades in consideration. This leads to the second macroeconomic challenge of the pre- COVID-19 scenario - with low interest rates, there may be limited room which monetary policy can play to boost investment and consumer demand.

Despite a decade of stable and strong economic performance, there are high levels of shortterm indebtedness, inflated price levels for housing stock and limited opportunities to decrease interest rates, creating key macroeconomic risks in the pre-COVID scenario of early 2020.


Another risk, this time of little longer-term nature, was house prices. Over the last ten years, house prices, especially in England, have gone up significantly, with London and South England outpacing the rest of U.K.


As the property market stalls and incomes get squeezed from the coronavirus, the possibility of a 10-15% price correction opens up over the next year. A falling property market can lead to those booked in last three-to-five years pushed into negative equity and increases the risk of delinquency – the third and final risk observed.

Coming to the fiscal situation and the spending room with government, the fiscal deficit for 2019 was equivalent to 1.8% of GDP – interestingly, the third consecutive year where the deficit has been below the target of 3.0%. Hence, there seems to be headroom for the government on fiscal front. The fall in tax revenue due to a halt in economic activity, combined with spending on fiscal support, is expected to push government debt up significantly, something that would be seen in the later part of 2020.


COVID-19 Related Governmental and Regulatory Measures

Apart from monetary policy actions taken by the Bank of England to manage liquidity in the financial space, we see three set of actions which will have significant bearing on commercial bank’s performance and ECL:

  • Government imposed lockdown: Lockdowns have been put in place for two-to-three months in many cities which are leading to disruption in economic activity. This is likely to cause a drop in GDP in the near term. Both demand and supply side are expected to be adversely impacted for at least two quarters.
  • Central bank measures: One of the key regulatory guidance includes payment holidays and moratoriums (for three months) to be rolled out by lenders. This is aimed at preventing retail and commercial borrowers from immediately going into delinquency and payment hardship. Additionally, the regulator has requested banks for suspensions on dividends and bonus payouts as well as share buybacks in the current year. These measures are aimed at preserving bank’s capital.
  • Budgetary measures and fiscal support: Measures from government primary consists of loan guarantee schemes for businesses costing exchequer approximately GBP 330 billion. For a large part of these government backed loans (CBILS/CCCF), as much as 80% these loans are guaranteed by the government which would limit loss given defaults. Other measures include the various fiscal packages costing approximately GBP 60 billion spanning across unemployment benefits (including self-employed), business support in terms of grants and tax cuts, and other additional spending. Note that the size of this “relief package” can increase further if the COVID-19 related disruptions become deeper.

While there measures can provide a significant cushion in case COVID-related disruptions are short term, a conventional modeling approach to measure the positive impacts of these interventions might be difficult and inaccurate. Hence, it is important to account for the same in the form of management adjustments (more on this in subsequent sections).

IFRS9 Related Regulatory Guidance for ECL Estimation

Before examining ECL estimations and impacts, recent recommendations and guidance from the regulators should be considered. PRA and the Bank of England came up with guidance in late March to reflect on adjustments in IFRS9 and capital related estimates for banks. Three key points from the letter might be of use to IFRS9 risk and analytics professionals.

  • Consistent and robust IFRS 9 accounting and the regulatory definition of default, with due stress given to not overstate ECL in view of support from the central government and bank. The letter also explicitly recommends not to treat payment holidays as a reason to allocate an account to higher stage within IFRS9. A similar approach is recommended in case of exception convent breaches.
  • Reliance on adjustment and “overlays” for estimating ECL in the short term rather than using many model based scenarios or changing probability weightages or model adjustments in general. Reasonable stress has been given to the governance around estimation and the application of overlays.
  • Create a reasonable distinction between COVID-19 impacts on portfolio which are short term versus long-term deterioration in credit profile (SICR) wherever possible.

Potential Scenarios and Strategy for ECL Estimation

For the purpose of this study and given the evolving nature of the Covid-19 situation, two general scenarios have been proposed. These are distinguished based on key assumptions about the path of the pandemic, the extension of lockdowns, changes in regulatory guidance, reactions of households and businesses, macroeconomic impacts, and the path of recovery.

Base scenario of a short-term recession followed by a quick recovery: The best case scenario is one where the pandemic is short-lived, including any secondary waves, with insignificant number of new cases by the end of May and lockdowns lifted before the end of the second quarter.

This scenario would be characterized by a quick “V-shape” macroeconomic recovery by late 2020 or early 2021, with no major change in retail bank portfolio characteristics or risk except for low credit score segments. Government measures, relief packages, and moratoriums currently rolled out would benefit in the second half of 2020 by preventing any long-term impact. The financial and real economy gets back on track in next 12-15 months in this scenario. However, due to COVID scenario, unemployment would be expected to shoot up to 4.5% - 5% by the end of 2020.

Adverse Scenario of a prolonged recession followed by a slow recovery: This is an adverse view of the pandemic wherein it does not resolve in short term despite lockdowns and social distancing measures. This may force the authorities to temporarily suspend and reimpose lockdown protocols to keep the essential economy going, causing spurts in infections.

Under this scenario, the pandemic is more likely to be controlled only by the development of a vaccine and administration to the larger population. Over this period, the recession deepens with accompanied rising unemployment, indebtedness and decreasing investment. Fiscal stimulus may be increased, however its amount will be limited by falling tax collections, and its impact on boosting demand will be limited as well. Bank moratoriums cease to exist while those rolled out in March and April will translate into defaults at a significant level. This will put sudden strain on banking sector P&L and capital position. Adverse macroeconomic conditions will prevail for at least 1.5-2 years. The recovery may either be slow (“U–shaped”) or laggard (“L-shaped”), or somewhere in between, but starting only after 2021.

Impact on IFRS9 Model Components and Approach for ECL Estimation

IMPACT ON ECL: BASE SCENARIO

In this scenario, the expected impact on key loss rate components like PD and LGD would be limited. This is because short-term bank and regulatory measures in the presence of improved liquidity would minimize
 

While at this stage, ECL estimation may need to be based on management adjustments rather than a pure model outputs, there is a significant role of industry and sectoral segmentation to understand where the risk lies. Analysis of risk profile of customers who opt for payment holidays will also be useful.


any impact on delinquencies and longer-term credit risk of the customers. Hence, stage migrations in this scenario are also expected to be limited. Moreover, key macroeconomic indicators like GDP, unemployment, inflation, and indebtedness that form a part of ECL models may show only temporary and short-term deterioration followed by a sharp recovery to long-run trends.

However, some segments in different exposure classes may be especially susceptible to liquidity stress related defaults and increase in riskiness, such as retail student loan portfolios, debt-ridden credit card revolvers and customer serving specific industries such as travel and tourism. It is pertinent for banks to proactively identify such segments and pockets by assimilating and assigning employment, designation and industry affiliation of customers.

The efficacy of existing models in translating the macroeconomic impact accurately into ECL estimates may be very limited in this scenario. Hence, banks may have to rely on well-thought management overlays to account for any inaccuracy in ECL estimations without being too conservative.

Some considerations in deciding a relevant management adjustment are as follows:

  • Granularity of overlays:
    1. Model component based overlays. For example, LGD overlays may be based on the expectation of impacts on collections efforts, while PD overlays may be based on the best estimate of macroeconomic-related changes to risk in the short term.
    2. Segment or sector overlays based on categories including risk grade, delinquency buckets, industry, and other factors. Segment level overlays would be very effective for account-level ECL estimations and where customer application data provides ready information around employment and sectoral linkages.
  • Management of forecast volatility: Given the evolving situation, risk management teams will have to balance out volatility in forecasts. It may be tricky to capture accurately the impact of the pandemic versus the risk profile of a customer. There may be concerns whether impact should be apportioned in today’s charge or averaged out over a year.
  • Analysis of customers who take payment holiday support: Banks need to analyze the credit worthiness of customers who avail to a payment holiday facility. Such tracking needs to be done on a daily or weekly basis. This can help understand segments, which have historically performed badly in making regular payments against those who are genuinely impacted by the COVID situation.

IMPACT ON ECL: ADVERSE SCENARIO

In this scenario, the expected impact on PD and LGD will to be severe, especially for short term lending portfolios. This scenario would be marked by a prolonged recession and deterioration in credit risk profiles both for both short term and lifetime. In such case, a significant increase in stage two or stage three migrations that may be attributed to COVID-19 would be expected. Since the changes to macroeconomic factors and risk profiles would be significant, it is expected that models should be able to accurately translate the impact after the next 12 months or so.

Component-wise, the impact on IFRS models for the two scenarios under consideration has been summarized below:


Recommendations and Concluding Remarks

The following highlights the direction risk communities within banks globally and in UK should be heading towards to manage and estimate ECL:

  1. Focus on data capabilities, especially with respect to daily portfolio tracking. This includes capabilities to understand industry association and employment impact for customers.
  2. Modelling exercises with focus on conventional stress testing, keeping 2008 crisis as a benchmark or a comparison to the current situation.
  3. Perform an in-depth analysis of credit risk for customers who use the moratorium option.
  4. Digitalization of collection activities to minimize significant pressure on collection machinery within banks after the payment holiday period ends.

This study of the UK’s macro situation highlights areas of concern in the pre-COVID19 world. Despite efforts from central bank and the government, it is expected for ECL to increase from high single digits in a base case or by 18-25% in an adverse case. It is not recommended for banks to have a heavy reliance on models for ECL estimating purposed for short-term estimates given the higher volatility in the economic model drivers and the complexity of measuring the true impact of fiscal and regulatory measures on portfolio riskiness. It is expected that risk managers in financial institutions will manage this increase in volatility using appropriate management overlays. It might be worthwhile for banks and building societies to start investing in granular data and industry linkages, as well as tracking moratorium take-ups that would be important for analysis and forecasting going forward. Investing in collection technologies for the post-moratorium situation will also help manage LGD components. In case of a major recession, ECL estimates will have to be model driven and use granular data for managing accuracy and volatility. The coming months should be very busy for risk managers, including modeling teams, and operations will be required to constantly re-evaluate the performance, methodologies and forecasts for ECL.

Written by

Kartik Sahni,
Engagement Manager, BFSI, EXL Analytics, EXL

Manish Dureja
Senior Engagement Manager, BFSI, EXL Analytics, EXL

Samant Kacker
Engagement Manager, BFSI, EXL Analytics, EXL

Contributors

Alok Rustagi
Director, BFSI,EXL Analytics, EXL

Contact US