Modelling And Forecasting Exchange Rates In Social Statistics
Exchange rates are a significant reflector of the well-being of an economy of
importance to investors, government and also policy makers. Exchange rates have
however shown significant fluctuation and volatility depicting high degree of
unpredictability. This study is aimed to develop a tool for predicting the future exchange
rate movements with the application to Kenyan shilling against the United states dollar
exchange rate data observed from August 2005 to August 2016. The data was obtained
from the Central bank of Kenya Forex statistics. Time series models are one of the
most popular and explanatory models with regard to exchange rate data, especially
autoregressive integrated moving average (ARIMA) models. The study utilized daily
exchange rate data to come up with a suitable ARIMA model that fits the time series
data as well as provide adequate daily Ksh/USD exchange rate forecast. The study
intends to utilize statistical techniques through the use of some R-programming
language packages related in the analysis of time series data. From the analysis and
model formulation process, different candidate models will be proposed. The relatively
best model will be selected from the proposed models relying on results from goodness
of fit measures. These measures to be used by this study include Akaike information
criterion (AIC), Bayesian information criterion (BIC) and the mean error measure (ME).