Forecasting Gold Price Using Geometric Random Walk Growth Model

Abstract

Gold price forecast is important to ascertain the performance of gold as a precious commodity in money and capital markets. This paper addresses the applicability of a geometric random walk model also known as ARIMA(0,1,0) with constant and log transformation as a forecasting tool and analyze the performance of forecast for a short-term and long term horizon. Findings suggest that use of a geometric random walk model to gold price data is valid and comparatively better than other regular ARIMA models. In this study, both in sample and out sample and combined sample forecasts were studied. A forecast for the short-range period is developed and validated through the measures of accuracy of the forecast.