Optimal Asset Allocation Using Predicting Stock and Coin outputs in the Iranian Capital Market

Leila Torki, Mahmoud Botshekan & Soheila Mohamadghasemi


One of the most important factors in deciding on investment is the amount of risk and output on capital assets. Choosing a set of optimal assets is often done by exchanging between risk and output, the higher the risk, so investors expect higher outputs. Portfolio optimization is about choosing the best combination of assets to maximize output on investment and minimize risk as much as possible. Therefore, one of the important steps in portfolio formation is to determine the optimal ratio or weight of assets to reduce the risk of investment portfolio. This important step is made by choosing the right strategy. The present study investigates the optimal allocation of assets (coins and stocks) using macroeconomic variables. The purpose of this study is to compare the performance of a predictability-based portfolio with a strategy-based portfolio (1/N). The data of this study were collected from internet databases stock exchange and central bank of Iran. The data are collected monthly from the beginning of March 2001 until the end of March 2017. In order to form predictability-based portfolios, first, out-of-sample stock and coin return forecasts were made by recursive and rolling regression models by five (24, 48, 60, 90, 120) month windows regression models. Then, by comparing the predictive power of the models within the sample, the optimal model is selected to predict the next periodic output and to predict the output on both stocks and gold. Then, using the predicted output on assets per month and in each window, two investment portfolios based on the variance mean strategy (investment strategy in tangent portfolio and variance-mean investment strategy with 3 and 5 aversion coefficient and another portfolio based on the minimum variance strategy have been performed and the performance of each of these portfolios with equal weight portfolios has been investigated through means comparison test, variance comparison test and Sharp Ratio. The results of the comparative test of variances and the Sharp ratio showed that the strategy of mean variance with a specified risk aversion coefficient (three and five) in all windows was able to defeat the strategy (1/N). The reason for the better performance of the mean variance strategy is that the underlying decision making is the predictability of asset outputs, and the weighting of each asset is based on the projected maximum output per month. The weighting of each asset per month is based on the maximum expected output.