Foreign exchange rate modeling for the Kenyan market - half - life model approach
Date
2015-12
Authors
Ngige, Isabel
Journal Title
Journal ISSN
Volume Title
Publisher
Strathmore University
Abstract
This study is a comparative test of the forecasting ability of the half-life purchasing power parity model and the portfolio-hybrid model used for forecasting exchange rates in Kenya. The United States Dollar and Ugandan Shilling exchange rates are analyzed for the period between 2005 and
2015. The models' forecasting precision assessment is based on three statistical tests; First Diebold and Mariano (1995) null test of equal forecasting ability is used to determine the superior half-life model among the three, four and five year half-life durations of which there is none, the root mean squared forecasting error and Thiel's inequality co-efficient tests are finally used between the three year half-life model for the Ugandan Shilling, the five year half-life model for the United States Dollar and the portfolio-hybrid model. The study concludes that the half-life purchasing power parity model is more precise in forecasting the United States Dollar and Ugandan Shilling exchange rates, albeit in sample. The conclusion of the superiority of the half-life model, which is relatively simple to run, requires no prior forecasts of the economic independent variables and is parsimonious, impacts industry players largely affected by movements in exchange rates who require an efficient exchange rate forecasting model such as portfolio managers.
Description
Research Project submitted in partial fulfillment of the requirements for the Degree of Bachelor of Business Science in Actuarial Science at Strathmore University
Keywords
Half-life purchasing power parity, Forecasting precision, Portfolio-hybrid model