Risk minimising portfolios for cryptocurrency and shariah assets using mean variance optimisation model / ‘Afina Syahmah Abdul Gaafar, Nur Athirah Hani Senin, and Nurul Aisah Ahmadi

Portfolio selection is well identified as a crucial problem in finance because the future returns of assets are unknown at the time of the investment decision was made. This research focuses on minimising the risk of portfolios of cryptocurrency and shariah-compliant assets. The objective of this research is to simulate the scenario returns (weekly) for cryptocurrency and shariah-compliant assets. Consequently, we are motivated to construct variance minimising portfolios for both assets and make comparison. Finally, we validate the performance of these portfolios using out of sample analysis. We apply variance as the risk measure and obtain risk minimising portfolios using mean variance model. A set of close prices of 51 companies involving the cryptocurrency and shariah assets are obtained through the Refinitiv website. Simulation is conducted on the collected data using the Microsoft Excel to determine the expected return range. A total of six in sample portfolios are constructed and evaluated across three different target returns representing low, medium and high return. In sample analysis shows that when a low target return is set, shariah portfolios yield lower standard deviation than cryptocurrency portfolios. This means, a shariah asset is less risky than cryptocurrency. The result is similar for portfolios under medium level of target return. However, under a high level of target return, the result shows a contrary pattern because the standard deviation of shariah portfolios start to surpass cryptocurrency portfolios. This may be caused by the classical theory of Markowitz’ portfolio. According to Boiko, Ye, Kononenko , and Goncharov (2021), heavy-tailed profitability in cryptocurrency causes the asset to not be subject to normal distribution. All developed portfolios are validated well because out-of-sample analysis results demonstrate that the realised return and expected target return of the portfolios exhibit similar behaviour. The realised standard deviation of the portfolios for every target return are also congruent to the standard deviation from the in samples analysis.