<![CDATA[The expected credit loss framework established under International Financial Reporting Standards Foundation 9 is structured around three core components: probability of default, loss given default and exposure at default. Among these three elements, the probability of default component most frequently lacks a coherent mechanism for embedding macroeconomic dynamics into the estimation process, a deficiency that carries particular weight in Latin American developing economies where economic volatility is a persistent structural feature. This article applies a novel five-step macroeconomic scalar methodology for dynamically adjusting the probability of default through the systematic integration of forward-looking macroeconomic information, with empirical application to unsecured retail portfolios in Brazil and Mexico. Unsecured retail lending portfolio datasets sourced from regional banking institutions in Brazil and Mexico provide the empirical basis through which the proposed methodology is validated across two economically distinct Latin American environments. The methodology advances through five sequential stages: research and planning; data preparation; model development; scalar calculation; and model validation. Comparative modelling draws on multiple regression, generalised linear models with logit and probit specifications, and machine learning techniques encompassing feedforward neural networks, random forests and gradient boosting. Model performance is evaluated through mean absolute error, mean absolute percentage error and mean squared error. Data collection extends to December 2024. The macroeconomic scalar produced consistent and economically coherent probability of default adjustments within the expected credit loss model for both Brazil and Mexico. Each modelling technique contributed distinct analytical insights, and the scalar demonstrated reliable capacity to improve expected credit loss forecasts across environments characterised by interest rate volatility, persistent inflation and exchange rate depreciation. Embedding a macroeconomic scalar within the expected credit loss framework constitutes a disciplined and auditable method for incorporating forward-looking information while preserving the model interpretability that bank boards, auditors and regulators require in Latin American credit markets. This article delivers a replicable approach for macroeconomic probability of default adjustment in expected credit loss models calibrated specifically to Latin American economic conditions. Structured implementation guidelines are provided for practitioners operating under International Financial Reporting Standards Foundation 9 in Brazil and Mexico.]]>
