The severity of the pandemic and the measures taken during the crisis will have a major impact on global economic development. To help prepare the financial industry, Strategy& conducted a COVID-19 stress simulation. We created three possible economic scenarios based on different shock and recovery patterns: the V, U, and L scenarios. Since mid-March 2020, we have revised these scenarios based on dynamic market sentiments and changing projections. At the same time, we have continually validated the perspectives through conversations with economists and industry representatives.
The “mild” or “optimistic” V-scenario assumes a lockdown period limited to several weeks, and a swift economic recovery after the shock. This scenario seems unlikely, since the lockdown has lasted over two months in most European countries before starting the relaxation of some measures.
Currently, we consider the “severe” U-scenario as the most likely. It assumes that overall growth is affected for at least two years. Finally, the “drastic” L-scenario involves continued widespread infection and repeated lockdowns, thereby condemning the economy to a prolonged recession.
In Germany, we expect a GDP contraction of 6.4%, 8.7%, and 10.9% respectively for the V, U, and L scenarios in 2020. We have considered other important macroeconomic drivers when modelling Non-performing Loan (NPL) ratios, including unemployment, consumer prices, and interest rates. The corporate and SME segments, as well as other asset classes such as retail loans, each show a specific sensitivity to those drivers, which is likewise taken into account.
In all of the simulated scenarios, corporate and SME credit will be the largest source of new NPLs in 2020 accounting for almost two third of the NPL ratio increase. The impact of COVID-19 on corporate credit varies between industries. We expect the default rates of companies in the passenger transport, travel and hospitality, services, and entertainment and media sectors to rise the most steeply through the COVID-19 crisis. The 2020 probability of default for companies in these sectors is projected to rise by a factor between 1.4 and 1.7 when compared to 2019 values. The second biggest source of new NPLs is retail credit contributing around 30% of the NPL ratio increase.
Actual and reported values may deviate and be lower due to regulatory measures around forbearance (e.g. repayment moratorium) as well as other government support measures. However, this could only represent a time lag effect until these relief measures are taken back.
It is vital that financial institutions are prepared for the months ahead. Despite the fiscal countermeasures announced by governments throughout the world, German banks will need to closely monitor the effectiveness of these measures on their clients and judge whether country-specific risks will start to materialize on a global scale.
Reporting from the first quarter of 2020 offers an initial indication of what most banks can expect over the course of the year and provides a basis for validating our simulation results.
The results published by a set of large US banks have shown an overall growth of +9% in total loan volumes. In the United States, this is largely a consequence of drawdowns of commercial credit lines. At the same time, provisions for loan losses skyrocketed to levels not seen since the global financial crisis. In comparison to the first quarter of 2019, provisions for loan losses increased by +59%.
These US quarterly reports, together with the sheer magnitude of the increase in provisions, confirm the general direction of our simulation results. Moreover, even though loan loss provisions have increased so steeply, they might even underestimate future losses if the macroeconomic scenario worsens.
With +5% on loan loss provisions, certain large European banks have increased provisions to a significantly lesser degree than their US peers. This difference can largely be explained by accounting rules, as for European banks under IFRS9, lifetime expected losses on loans are first considered with a transfer of loans to Stage 2.
Given the discretion that European banks have with these transfers, and the regulators’ guidance not to transfer in a mechanistic way, as the uncertainties of COVID-19 remain significant, the full effect will likely phase in over the next quarters.
The scale of its impact, and the compressed timeframe in which loans are expected to default, make this crisis very different from previous ones. While the main trigger has been an external shock rather than assets being misevaluated or borrower quality being overestimated, governmental protection and stimulus measures may not mitigate the damage for all borrowers. Banks, therefore, need to prepare to deal with a higher level of non-performing exposure.
What makes this situation so critical for European banks is that, unlike for US institutions, the required buildup of risk provisions cannot be covered by retained earnings. As insufficient retained earnings are available, banks will need to tap into their capital buffers or mitigate the damage in other ways.
Banks need to take a step back and ask whether their risk assumptions are still valid. Are downside scenarios sufficiently pessimistic? Are decision-making processes robust enough? If these are debatable questions, front-office units may be even more hesitant to relinquish client relationships and transfer them to workout units. However, the situation may also present an opportunity to review the approach to client selection and the profitability of client relationships. In a heightened risk environment, banks must carefully manage these client relationships.
Strategy& has identified key action items along three dimensions — business, risk and NPL management, and capital — that risk managers should carry out in order to strengthen financial resilience and navigate forthcoming NPL challenges.
Overall, we anticipate a potential second-round impact on the NPL burden from slow economic rebounds or stalling government support programs. This all means that risk managers have some busy times ahead.
Florian Balke and Clemens Bürgel also contributed to this article.