Introduction and literature review

The world of stock markets is no stranger to surprises, and recent years have seen a flurry of research into how these unexpected events shake things up. Uncertainty plays a big role here, affecting future dividends (usually for the worse) and expected rates of returns (often for the better) until the uncertainty is resolved (Brown et al, 1988). One standout in this field is the work of Baker, Bloom, and Davies (2016), who cooked up an index to catch those moments of intense economic policy uncertainty, reaching a peak during the first wave of COVID-19 (Baker et al, 2020).

Now, while we won’t dive deep into the uncertainty-market relationship, we’re more interested in comparing two popular methods used to study how this uncertainty affects stock markets. First up is the event study methodology, a veteran in the game, used to analyse how unexpected events impact stock returns and other variables, like elections and referendums. Early on, Phan and Narayan (2021) suggested using the event study method to tackle the impact of COVID-19 on stock markets, leading to a flurry of research in this area.

Many studies have reported abnormal returns and volatility during the pandemic, offering insights into how different events affect different markets (Ashraf, 2020b; Fernandez-Perez et al, 2021; Pandey and Kumari, 2021). On the other hand, there’s the time series analysis of COVID-19 cases and deaths, examining how investors react to pandemic-related information. This method has shown how an increase in cases can negatively impact stock prices worldwide (Ashraf, 2020a; Seven and Yilmaz, 2021; Pandey and Kumari, 2021).

Our research aims to bridge these methods, providing a global and regional analysis of the pandemic’s impact on stock markets. By comparing results, we hope to shed light on when each method is most appropriate.

Data and methods

In this research, we embark on a journey to explore how the stock markets of various countries reacted to the unprecedented events of the COVID-19 pandemic. Our adventure begins with the collection of 80 daily stock market indices in local currencies, ensuring a robust sample size. Our methodological journey splits into two distinct paths: the traditional event study and an analysis of how returns react to increases in COVID-19 cases. For the former, we employ a multivariate system called seemingly unrelated regressions, which allows us to estimate abnormal returns with ease and account for cross-sectional correlation (Zellner, 1962; Karafiath, 1988).

To delve deeper, we enhance the traditional market model to accommodate autocorrelation of returns and lagged market variables, acknowledging empirical evidence of such phenomena across various markets worldwide. This extended model not only aligns with theoretical expectations but also exhibits superior explanatory power compared to its predecessor. Then we delve into a second experiment aimed at shedding light on market behaviour in response to daily fluctuations in new cases. This endeavour not only offers a comparison of two different approaches but also acts as a reliability check for our previous study while delving into the efficiency of markets in processing daily news. We also identify which countries and regions remain sensitive to changes in case numbers, even after accounting for various risk factors.

Results

Traditional Event Study:

We adopt two distinct approaches for interpreting the data. One approach anchors the analysis to the week of the first COVID-19 case in each country, allowing for direct comparison across nations. The other approach centres around the week when the World Health Organization declared the pandemic, providing a broader context for understanding market reactions. Both approaches yield valuable insights into market behaviour during the crisis.

For the first approach, during the initial weeks of the crisis, we witnessed a significant concentration of negative abnormal returns, indicating a swift and pronounced market reaction to the emerging threat. By the end of the study period, global markets had incurred substantial losses, with certain regions faring worse than others.

Market Sensitivity to New Cases:

This analysis sheds light on how daily fluctuations in new COVID-19 cases influence market performance.

On a global scale, we observed a pronounced negative sensitivity to increases in new cases, indicating a widespread investor apprehension in the face of rising infection rates. This trend persisted across most regions, highlighting the global nature of investor anxiety.

Winners and Losers: Model Comparison:

By examining cumulative abnormal returns and sensitivity coefficients, we identify the countries most affected during the pandemic’s first wave. Asian countries emerged as both significant gainers and losers, reflecting the region’s economic volatility during the crisis. In addition, Central and Eastern Europe are the most adversely affected regions, while Oceania and North America are the least affected.

Conclusion

Our study delves into the financial impact of the COVID-19 pandemic through two distinct experiments. We find a pervasive negative effect on global stock markets, with notable variations across regions. While certain areas, like Europe and South America, suffer disproportionately, investor apprehension is widespread globally.

Overall, our research contributes valuable insights into the pandemic’s financial implications, paving the way for further exploration into the underlying factors driving market dynamics during this unprecedented crisis. Understanding these complexities is crucial for informing effective strategies to navigate future challenges.