Models down to earth, and up to heaven
April 2024

What risks and real world characteristics must models be able to address in the future in order to be a valid risk management and steering tool?

Life has become complex, and so has risk management, risk quantification and risk pricing. We are confronted nowadays with diverse and heterogeneous challenges which require very different approaches and modelling techniques to address and capture them. For example, Claudia Buch, Chair of the Supervisory Board of the European Central Bank, talked in a recent speech (ECB 2024) which she held at the annual conference of a large financial industry firm about the kind of risks banks may be confronted with in the future and how banks can prepare for them. The risks she drew attention to reach from geopolitical risks, climate and environmental risks, and other global trends like demographic change and digitalisation, to interest rate and credit spread risk. She emphasized that the addressed global factors will strongly influence the evolution of the real economy and the financial sector, and at the same time those developments and the associated risks are afflicted with a high degree of uncertainty. In other words: We often don´t know what is going to happen.

Some practical applications from these insights, as they are derived in the quoted speech, are:

• There is a high degree of uncertainty around the baseline scenario. The range of scenarios has widened, and it is more difficult to attach probabilities to them.
• Stress tests have to take into account multiple scenarios (simulate a broad spectrum of potential risks and impacts).
• For forward-looking risk management banks need to work with scenarios and have to challenge assumed (or accepted) patterns.
• Overall, it is necessary to go beyond traditional risk management practices in order to cope with the unpredictable nature of risks and driving factors that lie ahead of us.

Therefore — and against the backdrop sketched above — we would like to present in the following a short example of one of the model applications which we expect to gain increasingly more importance under the new conditions in which we live. The model category we are going to put the focus on is suitable for the analysis of different possible future realization paths (incl. variability per path) of a selected set of variables. The future realization paths can be estimated based on the assumption of different (also multidimensional) occurrences or with a focus to selected topics/constellations etc. This primarily refers to the materialization of specific stress scenarios, to other relevant scenarios to be investigated or to forecasts under regular (baseline) conditions.

Specifically, we are talking about the broad spectrum of possibilities offered by global vector autoregressive-type models (GVAR models) for investigating a variety of risk management, risk quantification and pricing topics. GVAR models typically consider a set of macroeconomic variables (such as gross-domestic product, the stock market, interest rates, etc.) and allow to jointly model these variables for several countries. In more detail the GVAR approach tries to avoid the curse of dimensionality which shows up in a joint vector autoregressive model when considering multiple countries and variables, by first estimating model parameters in relatively small country specific models for each country on a separate basis. Then, after parameter estimation, the country specific models are aggregated to the global model which allows to describe all the variables considered simultaneously. Finally, an impulse-response or a scenario analysis can be performed in a straightforward way, by which we observe how shocks propagate forwards amongst the countries and variables considered.

Endowed with different scenarios how some events instantaneously affect economic activity for various countries, the GVAR model can be used to derive forecasts how the different scenarios impact the variables considered. For example, suppose that we work with a GVAR model where an interest rate variable is included for each country and that the Federal Reserve changes the interest rate. The interest rate variable considered for the U.S. included in the GVAR is directly impacted. The GVAR estimates the impact of this interest rate shock on all the other interest rates considered and the impacts on all the other variables such as economic activity and the stock market. Then, these estimates e.g. can be imputed into a stress-test or risk management tool to infer the corresponding impact on a bank’s credit portfolio.

A second example are joint shocks in the price levels of the countries considered, as — for example — observed approximately two years ago. Since consumer prices indices are usually part of the macroeconomics variables included in a GVAR model, the model can be used to perform predictions of the impact of these shocks on all the other variables considered in the model, and to investigate how these shocks propagate forward through the various economies. Third, given that different scenarios of how environmental events affect economic activity (such as gross domestic product) are available, the GVAR model can be used to derive forecasts how the different scenarios impact the variables considered.

Applicability of the methodology is not restricted to macoeconomic data, but it can be used on other data systems as well (e.g. for financial data or real estate data).

The task of scenario definition (for stress tests or scenario analysis) is itself a complex and multidimensional problem setting and therefore lies beyond the scope of the present short note. Nontheless, this should in no way diminish the importance and relevance of this modeling step, which we spare here for the moment.

At the end let us get back to where we started from, to Claudia Buchs´s ECB (2024) speech, from which we would like to highlight two more statements for concluding here: Under the current and envisaged market environment, banks have to be adaptable, and to be prepared to recalibrate their strategies and operational practices to new conditions as they arise. Furthermore, underlying vulnerabilities may take time to work through the system as a whole or the balance sheet positions until they become openly visible as threatening or materialized risks.

References
European Central Bank (ECB 2024), Speech by Claudia Buch, Chair of the Supervisory Board of the ECB, at the Morgan Stanley annual conference, Bridges to the future: managing bank risk amid uncertainty, 12 March 2024.

 

Disclaimer
Ed.: LBMS Advisory Services GmbH, 1070 Vienna, Austria, FN 417881g, office@betainside.com, https://betainside.com.
The contents of this document are for general information and marketing purposes only and do not contain any recommendations whatsoever. In particular, the information does not constitute any offer or legal, tax, investment or financial advice and is not suitable as a substitute for individual business advice. The content has been compiled with the greatest possible care. However, we accept no liability whatsoever for the accuracy, completeness and up-to-dateness of the information and data contained herein, nor for any loss or damage of any kind (whether direct or indirect) arising from or in connection with this document. The information does not constitute a decision-making aid. The content created and published by the publisher is protected by copyright. All rights are reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. Unauthorized use of such content is not permitted.