Pronóstico de la Tasa de Siniestralidad de un Seguro de Desempleo en Colombia Mediante Modelos Lineales Dinámicos.

Abstract

Unemployment is one of the most analyzed and expected monthly variables since it implies deciding sions on public policy that impact production and the country’s growth model, hence the It is important that insurers offer unemployment insurance that covers this contingency, for this reason reason, this study proposes to implement statistical methods, to generate a forecast of the rate of claims of an unemployment insurance in Colombia, describing the level of exposure that will have the product of unemployment and thus contribute to companies having the capacity to face and mitigate a possible crisis. Within the classical forecasting methods, dynamic linear models were used, the root mean square error (RMSE) and the Rhat used for comparison, selection of the best model and string convergence check; In addition, the backtesting determined the efficiency of the model in an atypical scenario.