Modelo de pronóstico para el tipo de cambio - promedio (S/ por US$) - interbancario – venta en el Perú, periodo 2018-2023
No Thumbnail Available
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad Nacional de Trujillo
Abstract
El objetivo de esta investigación es determinar el modelo de pronóstico más adecuado para el tipo de cambio promedio interbancario (S/ por US$) en Perú durante el periodo de enero de 2018 a diciembre de 2023, utilizando la metodología Box-Jenkins. Se llevó a cabo un análisis de la tendencia, estacionariedad y estacionalidad de la serie de tiempo. El modelo seleccionado abarca los enfoques AR(1), AR(2), AR(3) y AR(4), elegidos en función de los criterios de Akaike, Schwarz, el error cuadrático medio (ECM), el error absoluto medio (EAM) y el MAPE. Este modelo se demostró como el más preciso, con coeficientes estadísticamente significativos, y cumplió con los supuestos de normalidad, homocedasticidad e independencia. Se identificó una tendencia creciente en el tipo de cambio interbancario desde 2018 hasta 2021, seguida de una estabilización en 2022 y 2023. Aunque la serie original no era estacionaria, presentando una raíz unitaria, se logró la estacionariedad mediante la primera diferenciación. No se observaron diferencias significativas al aplicar transformaciones logarítmicas, lo que hizo que la diferenciación resultara ser el método más efectivo. Las pruebas de validación, como la de Jarque-Bera, confirmaron la normalidad de los residuos, así como la ausencia de heterocedasticidad y autocorrelación. El modelo final mostró un ECM de 0.2364, un EAM de 0.2101 y un MAPE de 5.43%, lo que indica una buena capacidad predictiva
Abstract The objective of this research is to determine the most appropriate forecast model for the average interbank exchange rate (S/ per US$) in Peru during the period from January 2018 to December 2023, using the Box-Jenkins methodology. An analysis of the trend, stationarity and seasonality of the time series was carried out. The selected model covers the AR(1), AR(2), AR(3) and AR(4) approaches, chosen based on the Akaike, Schwarz criteria, the mean squared error (MSE), the mean absolute error ( EAM) and MAPE. This model was shown to be the most accurate, with statistically significant coefficients, and met the assumptions of normality, homoscedasticity and independence. An increasing trend in the interbank exchange rate was identified from 2018 to 2021, followed by stabilization in 2022 and 2023. Although the original series was non-stationary, presenting a unit root, stationarity was achieved through the first differentiation. No significant differences were observed when applying logarithmic transformations, which made differentiation the most effective method. Validation tests, such as the Jarque-Bera test, confirmed the normality of the residuals, as well as the absence of heteroscedasticity and autocorrelation. The final model showed an ECM of 0.2364, an EAM of 0.2101 and a MAPE of 5.43%, indicating good predictive capacity
Abstract The objective of this research is to determine the most appropriate forecast model for the average interbank exchange rate (S/ per US$) in Peru during the period from January 2018 to December 2023, using the Box-Jenkins methodology. An analysis of the trend, stationarity and seasonality of the time series was carried out. The selected model covers the AR(1), AR(2), AR(3) and AR(4) approaches, chosen based on the Akaike, Schwarz criteria, the mean squared error (MSE), the mean absolute error ( EAM) and MAPE. This model was shown to be the most accurate, with statistically significant coefficients, and met the assumptions of normality, homoscedasticity and independence. An increasing trend in the interbank exchange rate was identified from 2018 to 2021, followed by stabilization in 2022 and 2023. Although the original series was non-stationary, presenting a unit root, stationarity was achieved through the first differentiation. No significant differences were observed when applying logarithmic transformations, which made differentiation the most effective method. Validation tests, such as the Jarque-Bera test, confirmed the normality of the residuals, as well as the absence of heteroscedasticity and autocorrelation. The final model showed an ECM of 0.2364, an EAM of 0.2101 and a MAPE of 5.43%, indicating good predictive capacity
Description
Keywords
Tipo de Cambio; pronostico, interbancario