Stochastic behavior of South African rand exchange rate

Abstract

The study of fluctuations of exchange rates is one of many active research areas in the international finance. This paper is concerned with modeling of exchange rates of the South African Rand, through which the Namibian Dollar accesses the international market. It uses some standard models and tools of quantitative analysis, such as Random walk, Martingale hypothesis, Auto Regressive model (AR), Auto regressive polynomial model (PN), Feed forward artificial neural network (NN), functional coefficient model and non-parametric regression model (NP), to closely look at stochastic behaviors of the South African Rand exchange rates against thirteen currencies of countries that are major trade partners to and within Southern African nations. Due to increasing international trade volumes, Southern African nations are more and more exposed to the international community, and trading in both goods and services that are affected to a large extent by movements in exchange rates.