Trump threatens China with new tariffs on another $200 billion of goods

 Fujifilm is suing Xerox for more than $1 billion over failed merger

Global stocks sink as Trump ups the stakes in China trade fight

ZTE stock drops 25% after US Senate pushes to keep ban

Audi CEO Rupert Stadler arrested in Germany

Amazon shareholders call for halt of facial recognition sales to police

Google bets $550 million on Chinese e-commerce firm

Microsoft urges Trump administration to change its policy separating families at border


M-PRESS-CreditRisk: A holistic approach to capital requirements under systemic stress


Regulators are still debating the amount of capital needed to support bank losses in a financial crisis. This column presents a new, pragmatic stress-testing tool that can answer the question under macroeconomic stress scenarios.

The method models inter-sector and inter-country dependence structures between banks in a holistic, top-down supervisory framework. A test of 12 major German banks as of 2013 suggests that while there is enough capital in the system as a whole, capital allocation among the banks is not optimal.

The Global Crisis forced banking supervisors to think about how to reduce systemic risk. With the Basel III framework they took a step back to view the system as a whole, creating a macroprudential overlay to bank regulation.

The aim of traditional regulation is to make sure that individual banks can support their standalone risk. This is the microprudential view. Risk of failure is addressed by setting minimum capital requirements for each bank. The macroprudential overlay complements the microprudential requirements by targeting the resilience of the banking sector as a whole. It does this using a toolkit of capital buffers to address systemic risk.

The amount of capital we need to support losses in a financial crisis, both at individual bank level and for the system as a whole, has created a heated debate for the last ten years (Admati and Hellwig 2013, Dagher et al. 2016). We are joining this debate by suggesting a new holistic approach for supervisors. We propose a pragmatic stress-testing tool that can answer the question under macroeconomic stress scenarios.

We have named our approach M-PRESS-CreditRisk (short for Micro- and MacroPrudential REquirements Systemic Stress Credit Risk). It would help calibrate micro- and macroprudential requirements under systemic stress, with a focus on credit risk in a holistic top-down stress testing framework (Tente et al. 2017).

The core element is a new portfolio model for credit risk assessment called SystemicCreditRisk. It builds on, and extends, the Systemic Risk Monitor of the Oesterreichische Nationalbank (Boss et al. 2006, Elsinger et al. 2006). Our method models inter-sector and inter-country dependence structures within, and between, the credit portfolios of banks. It accounts for many simultaneous borrower defaults, which have been a source of large credit losses in the banking sector.

Building blocks of M-PRESS-CreditRisk

Figure 1 shows the building-block structure of M-PRESS-CreditRisk.

Figure 1 M-PRESS-CreditRisk: Modelling framework

Macroeconomic scenarios

First, we generate three macroeconomic stress scenarios using NiGEM, the global macroeconomic model developed by the National Institute of Economic and Social Research. First we consider two country-specific stress scenarios – a financial crisis scenario and a scenario of fiscal contraction – with hypothetical shocks hitting the economies of Greece, Italy, Portugal and Spain and with the macroeconomic effects spilling over to other countries. Then we consider a global scenario that assumes a temporary surge in the oil price.

Satellite macroeconomic model

Second, we use estimates from an econometric model to make projections for 28 countries (the majority of the EU, plus Australia, Canada, Japan, and the US) of the non-performing loan (NPL) ratio based on macroeconomic variables from the world economy model, including GDP, the unemployment rate, and the long-term and short-term interest rates. The country-specific shocks lead to higher NPL ratios in the affected countries. The oil price shock leads to higher NPL ratios in all economies.

Portfolio model SystemicCreditRisk

Third, projected NPL ratios enter the portfolio model SystemicCreditRisk. For instance, a high ratio of NPLs for Greece and Spain would directly increase systematic risk related to the business cycle for Greek and Spanish assets in banks’ credit portfolios – later on in our study we refer to them as banks’ credit exposures in Greece and Spain. Indirectly, they would also affect exposures in other countries, because all systematic risk factors (or business cycles) are interconnected. Therefore, the probability that borrowers will default increases according to the individual composition of the banks’ credit portfolios.

A simulation algorithm

This delivers loss distributions for individual banks and the system as a whole. These distributions contain information about rare, extremely high losses in the system. We suggest that the relevant measure of systemic risk is the expected shortfall of the banking system. This shows how much loss the banking system’s aggregate credit portfolio would incur in the event that the loss exceeded the system’s value at risk. We calculate the expected shortfall as the mean over 1% of the largest realisations of portfolio credit losses. Our approach helps to identify which individual banks contribute most to the system-wide losses.

Capital requirements and buffers for systemic risk

Figure 2 summarises the measures that inform supervisors in a gradual manner on how to build the capital requirements for individual banks, taking into account systemic risk.

Figure 2 Model-based common equity tier 1 (CET1) capital requirements and loan loss provisions

At first, we generate simulations for individual banks without any macroeconomic developments. Mean credit losses of a bank from this simulation can thus guide forward-looking loan loss provisioning. Losses that exceed this expected amount can inform decisions on how much capital a bank should have as a minimum to cover unexpected losses.1

In the next step, we take into account the baseline scenario from the model of the world economy that provides projections as to how macroeconomic variables could develop in the future, in the absence of additional shocks. In such a case, an additional requirement to support losses from changing macroeconomic conditions may be posed under ‘Pillar 2’ in the overall framework. Specifically, the difference between the amount that a bank contributes to the system’s expected shortfall under the baseline projections, and the bank’s unconditional value-at-risk (VaR) can be charged as a Pillar 2 add-on.

On top of the previous measures which have a bank-specific, microprudential focus, the Pillar 2 add-on takes a system-wide view on capital adequacy. This perspective would be central to the application of macroprudential instruments which focus on the resilience of the entire banking system, such as:

The system as a whole should have enough capital to cover the difference between the system’s expected shortfall in the severest stress scenario and its expected shortfall under the baseline conditions. This amount of capital is allocated to individual banks proportionally to the banks’ mean expected shortfall contributions and could be charged as the systemic risk buffer.

A bank’s share in the system’s expected shortfall can inform decisions on the capital buffers for other systemically important institutions. Generally, only the maximum from the two buffers would apply for an institution.

A test run

We tested the approach on the 12 major German banks that the German Federal Financial Supervisory Authority identified in 2013 as systemically important institutions. Our test run, based on detailed data from the credit register of loans of €1.0 million or more (Evidenzzentrale für Millionenkredite) as of the end of 2013, suggested that there was enough capital in the system as a whole to withstand the adverse macroeconomic developments we considered. In our three stress scenarios, the system’s expected shortfall never exceeded 18% of the reported credit risk weighted assets, or 88% of the available common equity tier 1 capital requirements.

Nevertheless, the heterogeneous results at the bank level shown in Figure 3 lead us to conclude that, from the systemic point of view, capital allocation among the banks was not optimal. In the worst case, five banks would have had difficulties satisfying the combined micro- and macroprudential capital requirements calibrated to the model that ranged between around 6.3% and 27.2% of the reported credit risk weighted assets.

Figure 3 Model-based capital requirements per bank

Note: The figure summarises the maximum model-based micro- and macroprudential capital requirements for each bank in terms of credit RWA (left-hand side) and CET1 capital (right-hand side) and builds the averages across the banks for the whole system. For the Pillar 2 add-on, only the maximum values observed during the 28 quarters of projections were taken. The systemic risk buffer was simulated for the most severe stress scenario (the oil-price shock). The buffer for the other systemically important institutions was based on an average over all scenarios. For each bank, only the maximum of both – either the systemic risk buffer or the buffer for other systemically important institutions – is shown.

M-PRESS-CreditRisk: A sophisticated portfolio model

M-PRESS-CreditRisk is a sophisticated portfolio model that captures a significant part of systemic credit risk. It can help supervisors to calibrate different micro- and macroprudential capital instruments as a logically linked chain of requirements, in a one holistic top-down supervisory framework.

Needless to say, due to model uncertainty, our numerical results should be treated with caution. Also, we concentrated on credit risk from correlated exposures. We did not touch on other channels of systemic risk. Nevertheless, M-PRESS-CreditRisk is an important step forward in designing capital requirements in a way that takes a systemic view of risk.

-Natalia Tente, Natalja von Westernhagen, Ulf Slopek

Leave a Comment