Testing a model for the term structure of interest rates using time-series analysis (Box-Jenkins approached)

• Main Authors: Torrechante, Aylynn B., Tria Tirona, Ma. Rosario B.
• Format: text
• Language: English
• Published: Animo Repository 1991
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/17187
• Institution: De La Salle University

The study primarily dealth with Time-Series Analysis, the Box-Jenkins Approach. However, it concentrated only with the ARIMA (p, d, q) models. The study applied this topic to a 13-year monthly period of interest rates of loans based on commercial banks for the years 1978-1990. Box-Jenkins Approach involved three steps. Identification stage, to determine whether the series or original observed values were stationary or non-stationary. If the series were found to be non-stationary, then transformations were needed. This may include the natural logarithm transformation, 1/sqrt x transformation and differencing. The next step was about the explanation of models. Proper specification and determination, estimation of parameters and diagnostic checking were all included. Diagnostic checking determines adequacy of the model for future forecasts. Forecasting was the last step in the Time-Series Analysis. This dealt with prediction of future values. The study was able to determine two models that was forecasted for a span of six (6) months. The models were determined to be an ARIMA (1, 1, 0) for the 1n transformation and ARIMA (0, 1, 1) for the 1/sqrt x transformation.