Abstract:
Financial volatility forecasting is especially important in financial
econometrics that helps Investors to minimize their losses by understanding the future
financial volatility. To predict financial volatility GARCH models are used which better
reflects the leverage effect, volatility clusters and volatility jumps in a financial time
series. COVID-19 pandemic is one of the most extreme events we have faced recently
which led to sudden fall in stock prices and financial instability. While many research
papers have focused on stock indices particular to relevant countries or emerging
markets, this study examines large-cap stock indices, which are safer to invest in and
covering almost the entire world by considering ten major large-cap stock indices from
nine countries which represent seven regions; Asia, South Asia, Europe, Middle East,
North America, Oceania and South America. Statistical loss functions; MSE, RMSE,
MAE, R2 Log, and QLIKE, were used to determine the best model out of GARCH, GJRGARCH, EGARCH, and PGARCH. Although the literature review suggests using the
High-Low proxy method to capture realized financial volatility, this study used the
OHLC(Open-High-Low-Close) volatility estimator, which considers drift-independence
and is capable of handling opening-pricing-jumps. All tests were conducted using
python programming for significance levels of 1%, 5%, and 10%, and the results
indicated that the study is statistically significant at the 1% level. The results reveal
that, although the stock indices represent different regions, they have shown a similar
impact towards COVID-19, while NZ50 and Nikkei225 indices slightly differ as New
Zealand and Japan are less affected by COVID-19 during the period 11.03.2020 –
10.03.2022. The study concludes that, averagely the best model to forecast financial
volatility on large-cap stock indices which affected from COVID-19 is EGARCH (1,1)
as it is asymmetric and able to grasp the l
