Forecasting Monthly Gold Prices using ARIMA Model: Evidence from Indian Gold Market
Rakesh Kumar Sharma1, Anupam Sharma2

1Rakesh Kumar Sharma, Thapar University

2Anupam Sharma, Thapar University

Manuscript received on 09 July 2019 | Revised Manuscript received on 21 July 2019 | Manuscript Published on 23 August 2019 | PP: 1373-1376 | Volume-8 Issue-9S3 August 2019 | Retrieval Number: I32940789S319/2019©BEIESP | DOI: 10.35940/ijitee.I3294.0789S319

Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open-access article under the CC-BY-NC-ND license (

Abstract: In this paper an attempt has been made to give an overview of the Indian gold market so as to develop a model enabling the forecast of gold prices in India. One troy ounce is equal to 31.103 grams. The monthly sample data of gold price (in INR per troy ounce) is taken from December 1997 to December 2017.The entire data has been divided into two segments for estimation and validation sample and to find out the efficiency and accuracy of forecasting models. Since the gold price data series have shown much deviation after March 2006 the first segment of the data is taken from the time period of December 1997 to March 2006 and second segment from April 2006 to December 2017.Due to a larger value and a huge time span of the sample data, the natural logarithm of gold price has been taken to conduct the study and build an effective model to forecast future gold prices. The unit root tests of Augmented Dickey Fuller‖ and Philips Perron have been used to test the gold price series as stationary or non-stationary. It is observed that series are stationary at first difference in both the methods. At first difference the ACFs and PACFs were pattern less and statistically not significant. Box-Jenkins’s Autoregressive Integrated Moving Average of Box-Jenkins methodology has been used for developing a forecasting model of gold price in India. Different models of ARIMA have been used to obtain best suitable model for forecasting using Eviews software 10 for both time periods i.e., December 1997 to March 2006 & April 2006 to December 2017

Keywords: AIC- Akaike Information Criteria, SIC- Schwarz Information criteria, Correglram
Scope of the Article: Marketing and Social Sciences