Desarrollo de un modelo cuantitativo para la medición del riesgo de liquidez en función del GAP de liquidez, stress y backtesting

Abstract

The purpose of the study is to develop a quantitative model to measure liquidity risk, validating the GAP assumptions through backtesting, comparing time series methodologies and designing stress tests. A quantitative and descriptive approach was implemented, with analysis of 2024 and 2025 data, without manipulation of variables, using Excel and Python. Models such as ARIMA, Holt-Winters and Exponential Smoothing were selected to project assumptions, improving the accuracy of the analysis. Backtesting evidenced significant deviations, especially in time deposit renewals, highlighting the need to use quantitative methods. The stress tests, based on historical scenarios, showed sensitivity of the GAP to changes in time deposits, although without compromising liquidity. It is concluded that the institution maintains a stable and resilient liquidity position, recommending the distribution of liabilities in the time bands, diversification of funding sources, improvement of interest assumptions and new retrospective validations.