Artificial neural network (ANN)-based estimation of the influence of COVID-19 pandemic on dynamic and emerging financial markets

• Published in: Technological forecasting & social change 2023-05, Vol.190, p.122470, Article 122470
• Main Authors: Naveed, Hafiz Muhammad, HongXing, Yao, Memon, Bilal Ahmed, Ali, Shoaib, Alhussam, Mohammed Ismail, Sohu, Jan Muhammad
• Format: Article
• Language: English
• Subjects: COVID-19 pandemic and DNN-based multivariate regression; Currency market; Dynamic emerging financial markets; Futures derivatives market
• ISSN: 0040-1625; 1873-5509; 0040-1625
• DOI: 10.1016/j.techfore.2023.122470

The COVID-19 pandemic is a serious global issue destroying financial markets awfully. The proper estimation effect of COVID-19 pandemic on dynamic emerging financial markets is a big challenge due to a complex multidimensional data. However, the present study proposes a Deep Neural Network (DNN)-based multivariate regression approach with backpropagation algorithm and structural learning-based Bayesian network with constraint-based algorithm to investigate the influence of COVID-19 pandemic on the currency and derivatives markets of an emerging economy. The output shows that the COVID-19 pandemic has negatively influenced the financial markets as indicated by sharply depreciating currency value around 10 % to 12 % and reducing short-position of futures derivatives around 3 % to 5 % for currency risk hedging. The robustness estimation shows that there have probabilistic distributed between Traded Futures Derivatives Contracts (TFDC), Currency Exchange Rate (CER), and Daily Covid Cases (DCC) and Daily Covid Deaths (DCD). Moreover, the output represents that the futures derivatives market conditionally depends on the currency market volatility given percentage of COVID-19 pandemic. This study may help to policymakers of financial markets in decision-making to control CER volatility that may promote currency market stability to enhance currency market activities and boost confidence of foreign investors in extreme financial crisis circumstances. •The Indian Currency market is affected around 10 % to 12 % by COVID-19 pandemic.•The Indian derivatives market is affected around 3 % to 5 % by COVID-19 pandemic.•The Indian Currency market effect spillovers on significant derivatives market.•A Deep neural network (DNN) gives best prediction than a sallow neural network (SNN).•The Bayesian network is an appropriate approach to estimate the conditional effect.