Determining the optimal Hedge ratio for the exchange rate (dollar) using gold futures contract and its prediction: an artificial neural network modelling approach

Due to the high inflation in recent years in Iran, as well as the uncertainty in environmental conditions, the use of investment risk hedging tools in the capital market has received more attention. What is targeted in this study is to identify the best approach among the existing approaches in calculating and predicting the optimal ratio of risk coverage considering the dynamic nature of this ratio and also environmental uncertainties. Undoubtedly, the performance of approaches based on modelling (parametric) or simulation, such as artificial neural networks, which are formed based on learning as well as previous information, will be affected in a situation where political, economic and social effects dominate a society. But what is targeted in this study is to compare the performance of existing approaches and use the superior approach to estimate this ratio and predict it with a non-parametric approach (an approach that works better in conditions of environmental uncertainty). In this research, determining and predicting the optimal dynamic hedge ratio of exchange rates using gold coin futures contracts in the Iran stock Market is discussed. The approach used in determining this ratio is the minimum variance and the comparison of different econometric models was used in order to optimize this ratio. By using the weekly data of the cash yield of the dollar and gold coin futures from the beginning of 2016 to August 8, 2020, the optimal risk coverage ratio for each model was calculated and by forming a portfolio and evaluating the variance, the effectiveness of the models was examined, the results of which show the superiority of the dynamic model was BEKK-GARCH. Based on the results obtained, a perceptron neural network model was used to predict this time series and it was concluded that the neural network model is a high-performance model in predicting this ratio based on asset returns.