“Lessons for Euro markets from the first wave of COVID-19”

Although the coronavirus pandemic hit Europe in the early days of 2020, European stock markets had signaled fluctuations in the days before. This paper assesses the observed volatility on European stock exchanges and searches for its sources during the first four months of 2020. To investigate the issue, a panel VAR model is adopted, and the generalized impulse response function and the variance decomposition methods are used. The estimations show that about 34% of the volatility in European stock markets is due to the Chinese stock market, while 7% is due to international uncertainty, as measured by VIX. The impact of pandemic cases and deaths on European stock markets is negligible, below 1%. This means that the European stock market faced two risk elements: the first is the transmission volatility from the Chinese stock market, and the second is the international uncertainty. The findings also support the view that COVID-19 is more like a systematic risk.


INTRODUCTION
Τhe paper deals with the reaction of the major European stock exchanges to the coronavirus pandemic. The virus started in China and soon spread to Europe and the rest of the world, causing concerns about its effects on both humans' health and the economy. The first economic impact was observed in the stock markets. There was a sharp drop in all stock indices during February and March 2020, but then markets began to recover.
There is little evidence of the effects of the health epidemic on financial markets, but it does indicate a negative response of stock returns toepidemics such as SARS and Ebola (Nippani & & Washer, 2004;Chen et al., 2018;Ichev & Marinč, 2018). This reaction of the market can be justified due to the influence of the disease on the overall behavior of population within the economies affected by the disease (Lee & McKibben, 2004) or because of the fear and anxiety that affected investor decisions (Ichev & Marinč, 2018). Baker et al. (2020) point that COVID-19 caused the biggest stock market volatility compared to other epidemics, while the spread of COVID-19 increases the financial volatility, and its persistence can generate a new episode of international financial stress (Albulescu, 2020).
The first COVID-19 case in Europe was recorded in January 2020 in Italy, while after a few daysCOVID-19 cases occurred in all European countries, forcing one country after another to adopt strict restrictive measures (known as lockdowns) during the months of February and March. European land is heavily affected by the virus in both the first and the second waves. Over the past decade, the European countries have experienced a severe debt crisis, which has threatened their unity and highlighted their fragility. A sequent crisis, such as COVID-19, can cause additional financial stress in this area due to the uncertainty and the fear not only of the pandemic itself but also of financial cost of lockdown measures in some countries. On the other hand, the observed volatility worldwide strengthens the view that COVID-19 creates a systematic risk across markets (Sharif et al., 2020). It would be interesting to assess the volatility in the European stock market and investigate whether systematic or non-systematic risk elements existed from February 2020to March 2020. Understanding the observed volatility in the European stock market is a valuable and useful tool to know if the conditions are being created for a new financial crisis in the Europe, and to make better policy decisions and take appropriate measures towards a future unknown event, such as COVID-19 a year ago.

DATA AND METHODOLOGY
The data set contains daily observations for 16 major European stock indices and the Chinese Stock Index from January3 to April 30, 2020. More specifically, the data set comprises stock closing prices of major stock indices 1 . The indices of each country are presented in Appendix A, Table A1. It also included daily COVID-19 infection data obtained from the European Data Portal 2 . Finally, CBOE Volatility index is used (VIX) 3 , which is the standard measure accounting the market volatility and investors' sentiments.
The study is interested in exploring the degree of volatility during the first wave of COVID-19 in the European stock market. The first wave is defined from January 2020, where the first COVID-19 case was recorded, to April 2020. The first announcements for easing the restrictive measures are considered as an indication that Europe has overcome the first wave of the virus.
where ( ) AL is the polynomial matrix in the lag operator , L it Z is a matrix of endogenous variables, while the vector Z includes the endogenous variables, and it e is the stochastic error terms.
While estimating a VAR model, some important issues must be taking into consideration. The needed conditions for the use of VAR model are the stationarity of variables and the correct selection of the lag length. Panel unit root tests are conducted, as reported in Appendix A, Table A3. The three information criteria of Schwarz (SC), Akaike (AIC) and Hannan-Quinn (HQ) are used, and these criteria indicate the choice of 8 lags.
The data are examined in three different ways. First, the whole time-period is investigated. Second, to be able to study and deal with a crisis, its phases, and its causes every time should be recognized (Philippas and Siriopoulos 2013). According to the volatility analysis, they are recognized as three separate phases: a) the incubation period, which expands from January 3 to February 28, 2020, b) the outbreak period during March 01-30, and c) after-outbreak period. As shown in Figure 1, the most of COVID-19 cases are recorded during the last days of March. However, an increased volatilityinEuropean stock markets is observed, which starts in January and peaks during March (see Figure 2). Third, the sample of countries is classified based on the total deaths per million recorded by each country during the first four months of 2020, and they aredivided into two separate groups (group A and group B). Group A includes the countries with the lowest values, and group B includes those with the highest. The formation of these two groups directly links COVID-19 deaths to stock returns. This analysis makes it possible to more clearly distinguish the effect of the COVID-19 variables.

The GIRF method
This section presents the responses of European stock returns to shocks derived from the COVID-19 variables, VIX and China returns. Figure 3 shows the impulse responses of European stock returns when an innovation of other variables occurs over the next 15 days. The solid lines are the impulse responses, and the dotted lines illustrate 95% confidence intervals.
The response of stock returns is close to zero whenever there is a shock in the COVID-19 variables. This is an indication that COVID-19 cases

Variance decomposition method
The forecast error variance decomposition of European stock returns is estimated.

RESULTS PER SUB-PERIOD
The sample is divided into three sub-periods, and the data are re-run to have a deeper insight about the development of the detected effects overtime. Selective results are reported, which add value to the analysis (Figure 4). All results are available upon request.
When investigating the sub-periods, the influence of the Chinese market on the European market is confirmed. The detected influence is from the outbreak period. It is also observed that VIX causes negative returns in the incubation period, while it is a source of volatility for European returns in the subsequent sub-periods. The response of returns to COVID-19 cases is slightly insignificant apart fromthe outbreak period during which some limited fluctuations are displayed.

Ranking countries
In this section, the countries are classified according to the index of total deaths per million recorded by each country during the first four months of 2020, and two separate groups are formed (group A and group B). Group A includes the countries with the lowest values, and group B includes those with the highest. This allows one to have a deeper investigation of the relationship between COVID-19 and stock returns, and the robustness of the results is also checked. Table 2 presents the groups of countries.    The GIRF and the variance decomposition are calculated for each group separately. Figure 5 1 2 3 4 5 6 7 8 9 10 11 12 13 14       In conclusion, the results are consistent with the previous section about COVID-19 variables. Some different patterns are detected between the two groups;the countries of group A are more susceptible to reported cases than those of group B, which appear to be most affected by deaths, although these effects are small. In addition, VIX has a greater impact on the group A countries than on countries from group B, while China's influence is greater in group B.

DISCUSSION
This study examines the influence of the COVID-19pandemic on European stock markets. The empirical evidence shows no indication that stock market fluctuations were derived from the recorded COVID-19 cases or deaths during the first wave of the virus. This paper highlights how the negative consequences of the health crisis can be strengthened through the financial sector and the importance of financial transmission channels when an external shock occurs.
COVID-19 is an exogenous shock that has a significant impact on the economy and financial mar-

CONCLUSION
Market volatility is the reaction of investors to uncertainty. Indeed, the pandemic has caused uncertainty and fear for both human health and the economy. This paper focuses on the Europe stock market, assesses the market volatility and explores its source. The study of the pandemic in Europe is a challenge. Europe has been a common market for decades and recently faced a severe sovereign debt crisis. This paper contributes to understanding the financial risks in this sensitive area during the current health crisis as there is little evidence.  Notes: Table A2 shows the COVID-19cases and deaths and the market risks. Market risks are approximated by calculating the standard deviation of the main indices daily returns per country from January to April 2020, for sixteen European countries and China. COVID-19 cases and depths is the average of daily records.