Desafíos de la industria bancaria en la estimación de la severidad de pérdida (LGD - LOSS GIVEN DEFAULT).

Abstract

The banking industry in Ecuador is governed by the regulatory standards of the Superintendency of Banks, so entities must establish efficient schemes for risk control through PD, EAD and LGD. This study focuses on the study of the LGD, or Loss Given Default, which corresponds to a fraction of a loan that is not recoverable in case it is in default or due to factors external to the entity (macroeconomic factors). There are different estimation techniques using different variables, in this case microeconomic factors provided by the Superintendency of Banks and the Central Bank of Ecuador (BCE) in a quarterly period of time from 2017 to 2020, these indicators will be used for the construction of the data set and thus use statistical techniques such as regression analysis to estimate LGD. Currently, in Ecuador there are no public studies to estimate the LGD with statistical models, therefore, this project is an efficient contribution to the introduction of more statistical techniques to the financial system.

The panel data method, a technique that controls for effects at the individual and time levels, was used. The fixed effects model helped us to estimate the LGD and to be able to conclude that macroeconomic factors such as the CPI, CCI and real GDP influence financial entities to have a higher LGD, which causes lower recovery rates. Hence more losses. Including these variables allows us to fit a successful backtested model with short-term predictions of the dependent variable.