By José Antonio Núñez and José Juan Chávez*
The use of models and techniques of machine learning (or machine learning) in the banking system is increasingly widespread, especially in risk management and the regulation of financial institutions, especially credit institutions.
Although the machine learning increases the predictive power of credit and financial risk, it also makes these models more complex. It becomes essential to know how to explain its role, directionality and the weight of the variables that intervene in the results of these models, as well as its sensitivity to each of the variables it uses. For this reason, different techniques have been developed and continue to be developed to face this challenge. But how did we get here?
The Basel regulation
The use of internal models (such as those enabled by the machine learning) for financial risk management was widely promoted in the initial stage of the Basel regulation (Basel II), which encouraged risk sensitivity in the allocation of the regulatory capital requirement in each of the portfolios. Let’s see.
In 1988, the BIS or Bank for International Settlements (the “regulator” of financial regulators) defined the minimum capital requirements for banks in its first stage, Basel I. It established standard risk weights for the main credit counterparties (sovereigns, financial institutions, corporations, mortgages, etc.). However, the risk sensitivity of this approach was low. In 2004, Basel II incorporated the use of internal models to improve risk sensitivity in the allocation of capital, mainly in determining the parameters that influence the use of more or less capital. In 2010, Basel III focused on the problems revealed by the 2008-2009 crisis: liquidity; capital supplements to cushion crises and for countercyclical effects -according to the systemic importance of each entity-, in addition to other elements such as large exposures, leverage ratio, counterparty risk, total capital for loss absorption, etc.
machine learning and the complexity
One of the main focuses of Basel II was that the necessary capital was sensitive to the calculation of critical variables of the credit portfolios. Among these variables are the probability of default (default), loss severity given default and exposure given default. In order to capture the key variables that intervene in the risk parameters and define the size of the capital necessary to face this risk, the models become more complex. In fact, the complexity can be such that it becomes undesirable (“undue”, according to the Basel documents), because the comparability of risk between different banks is reduced, not only in one country, but globally. Precisely, seeking a homogeneous metric to allocate capital was one of the Basel Committee’s intentions.
The comparability of risk metrics (and also of capital) is reduced by the different speed with which regulation is implemented in different countries. For example, hehe Basel III components have been implemented with different schedules and at discretionsince countries can make local versions of the standard (the so-called “national discretion”).
What will Basel IV bring us?
With Basel IV, which will be implemented from 2023, we see a regulatory turnaround, as well as an invitation to reduce the complexity of the models just when the use of highly complex models (machine learning) becomes a standard.
The Basel IV regulatory package is the answer to the increase in complexity. One of its objectives is to discourage the use of internal models to reduce undue complexity and improve comparability of regulatory capital. allocated by the global banking system.
Thus, this is the environment that financial risk management models developed with machine learning methodologies must face in order to comply with the regulation.
However, let’s end on a positive note. When the Basel Committee proposed the internal models, these were already a common practice in financial institutions and it was only regulated how to incorporate them into the regulation. The advancement of models machine learning it will continue in the internal processes because they are a powerful lever. Its regulatory recognition will continue on the agenda.
*José Antonio Núñez and José Juan Chávez are research professors at EGADE Business School
Editor’s Note: This text belongs to our Opinion section and reflects only the author’s vision, not necessarily the High Level point of view.
MORE NEWS: