Findings from Random Effects Models

As previously discussed in section 5.2.1, local new COVID-19 cases are expected to impact cumulative abnormal returns negatively. The findings of the random-effects models, shown in Table 25, provide evidence of a significant negative relationship between local COVID-19 cases and CAR for all model specifications at a 1% significance level. Specifically, a new COVID-19 case decreases cumulative abnormal returns by 0.005 percentage points in Scandinavia. Hereunder, a new COVID-19 case in Norway decreases Norwegian companies’ CAR by 0.031 percentage points, while a new case in Sweden decreases the CAR of Swedish companies by 0.005 percentage points. Lastly, a new COVID-19 case in Denmark decreases the CAR of the local companies by 0.029 percentage points. Thus, the results of the regressions indicate that the local spread of COVID-19 has depressed investors’ expectations for companies’ value.

The extraordinary monetary policy measures introduced by the local central banks were expected to have a significant positive impact on CAR, as they indicate that the liquidity challenges for companies are mitigated (discussed in section 5.2.1). The panel regression results indicate that the local monetary policies significantly affect Scandinavian companies’ cumulative abnormal returns, according to the baseline specification and the Denmark specification. The baseline model indicates that introducing a new extraordinary monetary policy increases CAR by 0.664%. Thus, our results indicate that the monetary policy measures announced by the local central banks have helped create confidence regarding companies’ future among stock investors.

However, it should be noted that the country-specifications for Norway and Sweden find no significant relationship between CAR and monetary policies.

Similar to the monetary policies, the extraordinary fiscal policies were also expected to have a significant positive impact on CAR, given that policies help companies survive the pandemic (discussed in section 5.2.1).

We find no significant relationship between fiscal policy announcements and cumulative abnormal returns for all model specifications except for Sweden. The model specification for Sweden indicates that a new fiscal policy announced by the Swedish government on average increases the CAR of local companies by 1.301% at a 1% significance level. Hence, the results suggest that investors’ concerns regarding companies’ performance have only in Sweden been lessened by the fiscal policies.

As earlier mentioned, the introduction of local restrictions is expected to have had a significant impact on CAR, but the direction of this impact is ambiguous (discussed in section 5.2.1). As expected, the findings presented in Table 25 provide evidence of a significant relationship between local restrictions and CAR for all model specifications, at minimum at a 5% significance level. However, the direction of the impact on CAR differs across the model specification. The findings suggest that the introduction of new restrictions has generally in Scandinavia and in Sweden specifically had a negative impact on CAR. More specifically, the result suggests that a new local restriction announcement, on average, decreases CAR by 0.556% for

Scandinavian companies generally and 2.139% for Swedish companies. On the other hand, for Denmark and Norway specifically, the findings indicate that new restrictions announcement increase CAR on average by 1.253% and 3.847% for Danish and Norwegian companies, respectively. Hence, the random-effects model indicates that the reaction to restriction announcements has differed across the Scandinavian countries.

As previously discussed in section 5.2.1, the volatility index, VIX, is expected to significantly negatively impact stock returns, as a higher VIX value indicates higher uncertainty among investors. The VIX index is significantly negative at a 1% significance level according to all model specifications. Hence, our findings provide evidence that increased global volatility significantly decreases cumulative abnormal returns in Scandinavia. We find that a unit increase in the VIX index generally decreases cumulative abnormal returns in Scandinavia by 0.390 percentage points. More specifically, an increase in the VIX index leads to a 0.157, 0.388, and 0.409 percentage points decrease in CAR in the Danish, Norwegian, and Swedish stock markets.

Hence, the results show that the increased volatility in the global stock market created significant concerns for companies’ performance among investors.

In section 7.1, the event study analysis indicated that the impact of the pandemic on stock performances differed across sectors. Hereunder, the event study suggested that the Health Care sector, which is the base of the regression models, was one of the sectors that had the best stock performance during the first wave.

Supporting the event study, the random-effects models generally indicate that CAR is larger in the Health Care sector than in the other sectors.

The random-effects models find significant differences between CAR in the Health Care sector and the Consumer Discretionary, Real Estate, Energy Utilities, Financials, and Industrials and Basic Materials sectors.

More specifically, the regression results indicate that, on average, CAR is approximately 10-14% lower for a company in the Consumer Discretionary sector than in the Health Care sector over the event period.

Furthermore, for companies in the Energy and Utilities sector, the results indicate that CAR is on average approximately 18-30% lower than for Health Care companies. For the Financials sector, only the baseline specification and Norway specification find a significant difference in CAR. Hereunder, the baseline specification suggests that on average financial companies had approximately 11% lower CAR than Health Care companies in Scandinavia. In contrast, the Norway specification suggests that this difference is 30% for Norwegian companies.

The results further suggest that Industrials and Basic Materials companies in Scandinavia, on average, had between 8% and 20% lower CAR than Health Care companies. Lastly, the results indicate that, on average, CAR is around 16-23% lower for companies in the Real Estate sector than those in the Health Care sector.

On the other hand, for the Consumer Staples and Technology and Telecommunications sector, the panel regression finds no significant difference in CAR from the Health Care sector. Hence, the random-effects models further support the event study, which concluded that these three sectors were the best performers over the event period.

The country dummies included in the baseline specification indicate that cumulative abnormal returns over the event period are not significantly different between the three Scandinavian stock markets. Hence, it is not possible to conclude that the stock performance over the event period differed across the three countries.

7.2.1. Random Effects Models Sub-conclusion

This section of the thesis introduced and analyzed the results of the random-effects models. It became evident that the model specifications found several explanatory variables to be statistically significant and explained at least 12% of the variation of the cumulative abnormal returns during the event period. More specifically, this section concluded that an increase in new COVID-19 cases significantly decreased stock performance, indicating that the spread of COVID-19 induced fear among stock investors.

Moreover, the random-effects models concluded that the implemented extraordinary monetary and fiscal policies positively impacted CAR in certain parts of Scandinavia. The random-effects models further all provided evidence that the implementation of local restrictions significantly impacted CAR in Scandinavia.

However, the impact of the restrictions differs across the Scandinavian countries. While the regressions find a new restriction announcement to improve stock performance in Denmark and Norway, they indicate that new restrictions worsen stock performance in Sweden. The models also provide evidence that increased global expected volatility significantly decreases CAR in Scandinavia.

Furthermore, the random-effects models found significant differences in CAR between sectors. More specifically, the models concluded that the Health Care sector generally had higher CAR over the event period than all other sectors, except for Technology and Telecommunications, and Consumer Staples. Thus, the results confirmed that the Health Care, Consumer Staples, and Technology and Telecommunications sectors were the best performing sectors over the event period. Lastly, the random-effects models conclude that there is no significant difference between CAR in the three Scandinavian countries over the event period.