A study on seasonal and cyclical movements of 91-, 182-, and 364-day Philippine treasury bills for the period 1984-1999

• Main Authors: Bunggo, Andrew, Elbinias, Jose Romeo, Gonzales, Katrina Eleanor, Sy, Stephanie Katrina
• Format: text
• Language: English
• Published: Animo Repository 2000
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/16685
• Institution: De La Salle University

The objective of this study is to be able to find seasonal and cyclical movements in t-bill rates. In order to determine if 91-, 182-, and 364-day t-bill rates do have seasonal patterns, a tabulation of their proportions and significance tests concerning these proportions were done in order to prove that seasonality exists. The results show that for the 1st week concerning the said t-bill rates, a significant proportion is found to be considerably highest, as for the week with the lowest rate-there were no significant results to validate this hypothesis. The months with the highest proportion was January, but after the test of proportions was done, no significant differences were discovered between January and other months. The same can be said for determining the month that has the lowest average t-bill rate. The test of the significance of seasonality was then used in order to determine if the seasonality enough to affect the t-bill rates. This is due to the fact that if historical data were used to determine future values, the presence of seasonality will greatly affect the predictive power of these rates. The Kruskal-Wallis test of seasonality or the H-test was used to prove that in the t-bills do have significant seasonality. The 91-, 182, and 364-day Philippine t-bills are proven to contain seasonality. When these data are used to predict seasonality, they will greatly affect the forecasted rates, so it would be best to use the latest data in order to predict the weekly and monthly rates for the first quarter of the year 2000. The next step is to determine if there are cyclical movements in t-bills, it was proven that cyclical movements do exist. There were three cyclical movements found and each was affected by extraneous variables. These cyclical movements ranged from a minimal period of three years to a maximum of seven years. The three cycles that were found started from January 1984 and ended on February 1987, then began again February 1987 to October 1994 and the last cycle was from October 1994, to July 1999. The study then used the historical data to attempt to predict future t-bill rates or their averages (both monthly and weekly) to determine if historical data can be used to predict the future values of t-bill rates. Application of various forecasting models were then implemented in order to determine what mathematical model would best explain the movements of the data gathered. The naive model was used and it proved that historical data could not be used to effectively forecast t-bill rates. The results of the forecasts showed that they were inaccurate. Because when they were compared against the available actual weekly and monthly rates for the specified period, the predicted rates fell below the actual rates. It was only for the 182-day monthly did the forecast hold true. These results show that the historical data of the t-bills cannot be used to forecast their future movements.