1. Trang chủ
  2. » Luận Văn - Báo Cáo

econometrics analysis of the impact of covid 19 sars cov 2 on international e commerce

46 0 0
Tài liệu đã được kiểm tra trùng lặp

Đang tải... (xem toàn văn)

Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Tiêu đề Econometrics Analysis of the Impact of Covid-19 (SARS-CoV-2) on International E-commerce
Người hướng dẫn Lê Hồng Mỹ Hạnh
Trường học Foreign Trade University
Chuyên ngành International Business
Thể loại Scientific Research
Năm xuất bản 2021
Thành phố Ho Chi Minh City
Định dạng
Số trang 46
Dung lượng 9,68 MB

Cấu trúc

  • 1. Introduction (4)
  • 2. Literature review (5)
    • 2.1. Previous studies (5)
    • 2.2. Overall effect of COVID- 19 (6)
  • 3. Methodology and Data (7)
    • 3.1. Descriptive statistics (7)
    • 3.2. Data description (8)
    • 3.3. Hypotheses (10)
    • 3.4. Method of research (11)
    • 3.5. Model (11)
  • 4. Result (12)
    • 4.1. Pearson correlation coefficients (12)
    • 4.2. OLS model results and problems with the model (12)
    • 4.3. Correction and final result (13)
    • 4.4. Limitations (15)
  • 5. Conclusion and Recommendation (16)
    • 5.1. Conclusion (16)
    • 5.2. Recommendation (17)
  • APPENDIX 1. REFERENCES (19)
  • APPENDIX 2. TABLES (20)
  • APPENDIX 3. STATA RESULTS (24)

Nội dung

In addition, confirmed death cases are also cases in which the patients are already dead when testing reveals the residuals or presence of COVID-19 virus.[Table A.1 Appendix2] 3.2.2 Stoc

Introduction

The E-commerce business has been gaining considerable growth since the appearance of world-changing innovations like the Internet, World Wide Web and more recently and also having a deep influence in our life - smartphones Stepping into the age of technology, traditional methods of commercials are getting outdated and proving to be inefficient Due to its reliance on physical interaction, the reach is, most of the time, limited to a particular geographical area Meanwhile, E-commerce is computed digitally and online, therefore broadening its reach to potentially everyone and everywhere around the globe In short, it is commonly recognized that E- commerce has integrated itself as an inseparable part of the world’s development However, there has been another major event that has changed the world we live in since 2020 - COVID- 19 It is a global consensus that during the pandemic, almost every aspect of our personal life and every field of jobs are negatively affected to a certain extent Noticeably, the decrease in revenue, adjustment in the workforce and job market, disrupted supply chain have all taken a big toll on numerous companies worldwide

Therefore, many people have been speculating whether there is any relationship between one of the most important businesses in our society (E-commerce) and the current COVID- 19 pandemic More specifically, we will be looking into the number of COVID-19 cases and deaths everyday in the year 2020 to determine their effects on the

5 bigge st E-commerce businesses right now - Amazon, Alibaba, Jingdong, Ebay and Rakuten For the companies, we will be looking into their stock prices and interest over time of each company

Understanding the effects of COVID- 19 on these companies’ performance wi ll be important because of two reasons Firstly, it will provide us with a better idea of whether E-commerce businesses are affected negatively or positively by the pandemic Secondly, it will help us predict necessary actions to take in the future in order to minimize the threats (if any) and maximize the potential For instance, the increase in the number of cases and deaths has made the stock price of these firms rise or fall? If we find that there is a positive effect in these companies' financial indicators, we may conclude that the pandemic does not negatively affect these companies but rather supports its business And if the previous statement is true, what further action steps can we take to be more successful and adapt with the new normal in the future when the pandemic still continues

This paper’s analysis is based on the methodology presented in the research paper published by MBA Nguyễn Hoàng Nam (2021)[1] In this paper, he managed to quantify the effect of COVID-19 on Vietnam’s economic activities by constructing six hypotheses Moreover, we are also inspired by the publication of Hung- Ha o Chang and Chad Meyerhoefer (2020) [2], in which they utilized Ubox, the largest business- - to consumer agri-food E-commerce platform in Taiwan, as dependent variables to measure the of COVID- 19 on the demand and growth of online food platforms.

Literature review

Previous studies

According to the study of United Nations Publications [3] , the widespread pandemic caused a sharp deceleration in financial activities for which busine sses were generally ill-equipped However, a noticeable effect of the raging pandemic has been the transition of multiple firms to E-commerce stemming from the requirement for necessity of moving online This report makes a preparatory appraisal of the effect of the COVID- 19 emergency on E-commerce It incorporates overviews of E-commerce businesses and customers and its accomplices within the eTrade for all activity, examinations by other multilateral offices, and inquiries embraced by the United Nation Organization

In addition, the study of Luis Varona and Jorge R Gonzales (2021) [4] analyzed the short-term behavioral dynamics of economic activity, as well as explained the causal relationships in the COVID- 19 pandemic context is based on the baseline amount of COVID- 19 transmission per day Research data was collected with economic variables, including: economic activity index, public expenditure index, real interest rate index, exchange rate index, international dong price index, index stock price numbers of the Lima Stock Exchange, The study uses the ARDL model to measure the impact of COVID- 19 on the economy of Peru The research results show that there is a negative effect with the expected sign, the statistical significance is 1 Moreover, Akbulaev et al.'s study (2020) [5] focuses on the economic impact of COVID-19 The study assessed the impact of COVID- 19 on the economy in many aspects, from the impact on production, employment, import and export, to the analysis of the State's support for producers in the mandatory quarantine

Last but not least, the study by Prince Asare Vitenu-Sackey and Richard Bar (2021) [6] aimed to assess the impact of the pandemic on poverty duction and global reGDP by measuring at the heterogeneous effects of each country Data utilized was collected from 170 countries, using econometric panel techniques such as OLS and squared regression to determine the result Variables included in the st udy were total COVID-19 cases, total confirmed deaths, rigor index, human development index, and gross domestic product per capita The study's findings indicate that the severity of many people's conditions and the rise of the disease have negatively impacted poverty alleviation and economic growth However, the deaths recorded to date positively affect both poverty reduction and economic growth This development signals the nature of controlling population growth as it hinders economic growth and poverty alleviation.

Overall effect of COVID- 19

The COVID- 19 pandemic has devastated the global economy, leaving many people around the world in dire straits COVID- 19 has created some uncertainty regarding economic and social policies From a business perspective, the epidemic has had a negative impact on the finance and various other industries Currently, the potential impact of COVID- 19 on globalization and global health in terms of mobility, trade, travel and the most affected countries is absolutely volatile, which is terrifying The world economic order and activit ies are changing drastically as most countries are going through a period of stay- -home isolation, social distancing and even at nationwide shutdown

Amid the year 2020, the COVID- 19 pandemic has overwhelmed the world financial advance Development confinements has hindered financial activities in most regions and countries However, COVID-19 has had varied effect s on different regions and countries depending on the time, concentration of the infected cases and the nation’s own financial status Developing countries and least developing countries, as well as their population, are more prone to worldwide financial downturns and recessions than ever On the other hand, a few nations, especially those within the Asia-Pacific locale, have experienced moderate contamination rates in later months of the pandemic, permitting for a speedier return to pre-pandemic levels of financial movement than others, especially those in Europe and the Americas, which were anticipated to encounter many more waves of COVID- 19 cases by late 2020 Nevertheless, as stated above, with no absolute evidences nor any solid patterns of the future development of the virus, virologists and economists could only guess what will happen in the future.

Methodology and Data

Descriptive statistics

The variables utilized in this research paper are as follow in this table, with the source and time frame included:

Variable Description Source Expecte d sign Time amzn The stock price of Amazon

NASDAQ.com (The Nasdaq Stock Market)

01/11/2019 - 23/04/2021 baba The stock price of Alibaba jd The stock price of Jingdong ebay The stock price of Ebay rkuny The stock price of Rakuten amznint Amazon’s Interest over time

Google Trend This is left blank as babaint Alibaba’s Interest over time jdint Jingdong’s Interest over time

Variable Description Source Expecte d sign Time ebayint Ebay’s Interest over time we do not intend to study their effects

01/11/2019 - 23/04/2021 rkunyint Rakuten’s interest over time nc Daily new confirmed cases of COVID- 19 worldometer.co m

+/- ttc Total confirmed cases of

+/- nd Daily new confirmed deaths of COVID- 19

+/- ttd Total confirmed deaths of

Data description

We use data from several sources to conduct our analysis, including information about coronavirus cases and death cases worldwide, stock prices of E-commerce companies, and interest over time of those

From a reliable source of information about COVID- 19 on Worldometer.com, we obtain the number of confirmed and death cases over the year of 2020 Confirmed cases were validated by the COVID- 19 test which is different from suspected cases in which patients only have the early symptoms of COVID- 19 Death cases are confirmed immediately after the COVID- 19 patients deceased In addition, confirmed death cases are also cases in which the patients are already dead when testing reveals the residuals or presence of COVID- 19 virus.[Table A.1 (Appendix2)]

Stock price provides one measure of the efficiency of that company operating during the COVID- 19 It also has daily data, which, along with cases and deaths of COVID-

19, assists us in constructing the panel data set on a daily level Thus we chose to study the effect of COVID- 19 on E-commerce companies’ stock prices collected from the Nasdaq Stock market

In the publication of Hung-Hao Chang and Chad Meyerhoefer (2020) [2] , they utilized the financial statistics Ubox, the largest business- -consumer agri-food E- to commerce platform in Taiwan, as dependent variables to measure the of COVID-19 on the demand and growth of online food platforms With their resourceful database, they successfully constructed an extensive and detailed panel data set to efficiently estimate the effect the pandemic has caused on Ubox’s statistics and to draw conclusion about its impact on the industry

As this method proves to be an interesting and efficient way to achieve the desirable results, after referencing various reports and articles, we selected five companies that are consensually deemed as the top ranking ones of the E-commerce industry around the globe as the basis to statistically demonstrate the effect of COVID- 19 These companies are, ranked respectively in order of higher to lower: Amazon, Alibaba, Jingdong, Ebay and Rakuten They will work as the representatives of the industry of E-commerce as a whole for this research; thus, the impact of COVID- 19 on them will then be concluded as the effect to the industry itself The data will be collected from 01/11/2019 to 31/12/2020, as we wish to include the growth of the stock price during the pre-COVID period of time into account for a more precise calculation (*) [Table 1 A.2 (Appendix 2)]

When diseases break out in many countries, social distancing is mandatory in many countries People have to stay at home and thus have difficulty with shopping in stores and markets which have a high rate of coronavirus affection As a result, people turned to shopping online through many different E-commerce platforms We have researched the data for selected companies’ volume of searches on the search engine of Google,

1 (*) Because the Nasdaq stock market do not open on Saturdays and Sundays, the timeline will have missing values as it is the most popular engine, from November 2019 to April 2021 Utilizing the extension Google trend, which is specified in measuring the Internet users’ interest in a topic or a search term in a period of time, we obtained the interest over time (**) 2 of the chosen companies

The fact that companies have a high interest over time percentage means they are increasing in popularity in that time range, which in turn means they are becoming more favorable in purchasing Consequently, we believe it would be a factor affecting the firms’ stock prices [Table A.3 (Appendix 2)]

Hypotheses

Conducting this research, we based our analysis off the methodology presented in the research paper published by MBA Nguyễn Hoàng Nam (2021) [1] In this publication, he constructed six hypotheses in order to measure the impact of COVID-

19 on Vietnam’s economic activities These hypotheses are as follow

- H1: COVID- 19 has an effect on Vietnam’s Exchange rate

- H2: COVID- 19 has an effect on gold price

- H3: COVID- 19 has an effect on oil price,

- H4: COVID- 19 has an effect on silver price

- H5: COVID- 19 has an effect on cooper price,

- H6: COVID- 19 has an effect on VN -index

Utilizing his model, he successfully managed to numerically quantify the effect COVID-19 has on the growth of these financial indicators, thus proving that COVID-

19 has an influence on Vietnam’s economic activities’ development

Inspired by Mr.Nam's methodology and models, we established our own hypotheses in order to measure the impact of COVID- 19 on the state of the worldwide E- commerce industry These hypotheses, which would serve as the foundation for our final claim and results in this paper, are:

2 (**) Interest over time: Numbers represent search interest relative to the highest point on the chart for the given region and time A value of 100 is the peak popularity for the term A value of 50 means that the term is half as popular A score of 0 means there was not enough data for this term (Source: Google Trend)

- H1: COVID- 19 has influenced Amazon’s financial development

- H2: COVID- 19 has influenced Alibaba’s financial development

- H3: COVID- 19 has influenced Jingdong ’s financial development

- H4: COVID- 19 has influenced Ebay’s financial development

- H5: COVID- 19 has influenced Rakuten’s financial development.

Method of research

This research employed the method of Ordinary Least Square (OLS) Estimation to measure the correlation of the selected E-commerce companies' stock price, as the indicators of their business status, with other two factors: cases or deaths of COVID- 19 and the company's interest over time According to the OLS estimation method, we would check the p_value of the explanatory factors to verify their significance to the models With the null hypothesis H0: B =0, which indicates that a variable is i statistically insignificant, the lower the p_value of each independent variables is, preferably below 0.05 and still acceptable if below 0.1, the higher the probability that we can reject this hypothesis and conclude that the variable is vital to the model Afterwards, we would run a few tests to detect some possible problems in our models and database: VIF was used to test for Multicollinearity; Breusch-Pagan test, White test and Park test were to detect Heteroskedasticity; lastly, Durbin-Watson test and Breusch-Godfrey test were employed to check for Autocorrelation Afterwards, if any problem were present in our models, we would use different methods to fix the models, mitigate the unwanted effects and achieve the optimal models and results for this research.

Model

We collected and used a panel database, which is enclosed in the Appendix A, to identify whether COVID- 19 has an influence on the selected companies, specifically if it had affected the development of the companies’ stock price In addition, as a method to reduce the disturbance of the model, we decided to include the interest over time of each company in their own models on the Internet

The Ordinary Least Squares (OLS) estimation model is:

In this model, Y i represents the ith company’s daily close/last stock price within the period of 2020 The variable is an explanatory variable standing for the number of X daily new confirmed cases (nc), total confirmed cases (ttc), daily new confirmed death (nd) and total death (ttd) Lastly, the variable Int i is an independent variable of the th i company’s interest over time.

Result

Pearson correlation coefficients

Using the Pearson correlation coefficients method, we obtain the results as in the table A.4 and A.5 (Appendix 2) The significance of each pair of variables must be less than 0.05 for the variables to have a statistically significant correlation As observed from the tables, all of the dependent variables and its explanatory variables have statistically significant correlation.

OLS model results and problems with the model

After performing the OLS model mentioned above, we obtained its results, which are displayed in the tables A.6, A.7, A.8 and A.9 (Appendix 2) The variables in the results all have p_values < 0.05, which means they are statistically significant to the models We also do not include any results that has an insignificant variable’s coefficients, as its result is quite unreliable

We then perform the VIF test to see if our models have the Multicollinearity problems With all the VIF values < 5, we can safely conclude that our models are not affected by the problem of Multicollinearity

In the same table, you will notice that our models failed the White test for Heteroskedasticity and the Durbin-Watson (White, 1980) [7] test for Autocorrelation The former, our p_values were less than 0.05, which indicated the presence of Heteroskedasticity The latter, our d-statistic was outside the acceptance range, which proved our models had Autocorrelation problems.

Correction and final result

As we can observe from the testing result, our models and results are influenced by heteroskedasticity and autocorrelation Thus we employed the Newey-West (1987)

[8] test to attempt to mitigate and correct these two problems simultaneously

Firstly, we will try to find the relationship between COVID- 19 daily new confirmed cases and 5 companies’ stock The effect of COVID-19 daily cases on five companies’ stock can be indicated in the table below

Variable Beta Newey West – standard error t value p-value amzn 001913 0001334 14.34 0.000 ebay 0000229 4.06e-06 5.63 0.000 jd 0000696 5.94e-06 11.72 0.000 baba 0000879 000012 7.30 0.000

Table 2 Newey-West result of the effect of Daily new confirmed cases

From this table the COVID-19 has the largest impact on Amazon’s stock price when the Beta is as high as 0.001913, which can be explained by its domination in the E- commerce market When the new COVID- 19 cases increase to 1000 cases, Amazon stock price will likely increase by 1.13 Dollars Ebay, Alibaba and have their stock jd increased when there are more and more COVID- 19 cases occurring but with less acceleration compared to Amazon Rakuten’s stock price seems to be affected by COVID-19 slightly All results have a very small amount of p-value which indicates that COVID- 19 clearly has an impact on the company's stock price at 5% significance level (p-value = 0.000 < 0.05) All regression models do not run in multicollinearity problems due to their VIF values being smaller than 2

Variable Beta Newey West – standard error t value p-value jd 3.59e-07 3.71e-08 9.70 0.000 ebay 1.38e-07 2.62e-08 5.25 0.000

Table 3 Newey-West result of the effect of Total confirmed cases

Next , in the table above, we can observe the impact of total COVID- 19 cases on the stock price From this, we can conclude that although the p-value of these estimators are statistically significant, compared to New COVID- 19 cases the effect of Total COVID-19 cases has a relatively small impact on the 5 companies’ stock price

Variable Beta Newy-west standard err or t-value p-value amzn 0986512 0064846 15.21 0.000 ebay 0011214 0001528 7.34 0.000 jd 0035041 0002168 16.16 0.000 rkuny 0001928 0000293 6.57 0.000

Table 4 Newey-West result of the effect of Daily New confirmed deaths

Beside determining the effect of total confirmed cases, we also conduct an estimation on how Daily COVID-19 deaths can lead to changes in companies’ stock price The estimation’s result is shown in table The test also shows very small p- values (p-value < 0.05), the COVID- 19 daily death cases have a significant impact on the selected companies’ stock prices beside Daily New COVID- 19 cases When we try to run an estimation on how COVID-19 Total death cases relate to our five companies’ stock price, we predict that the result would be the same when we conduct it on Total COVID-19 confirmed cases As predicted, table 14 shows that although the effect of Total COVID-19 death cases on companies’ stock price is not as strong as Total COVID-19 cases and Daily New COVID- 19 deaths but it still be more influence than Total COVID- 19 cases with a very high reliable p-value (p-value < 0.000)

Variable Beta Newy-west standard error t-value p-value amzn 0004996 0000382 13.08 0.000 jd 0000182 1.74e-06 10.47 0.000

Table Newey-West result of the effect of Total confirmed deaths5.

According to the research result, we can conclude that COVID-19 had a positive impact on Amazon, Alibaba, , Rakute jd n and Ebay’s stock prices with high statistical significance Amazon’s stock seems to be influenced by COVID-19 cases the most All our hypotheses are accepted; COVID-19 had an effect on all selected’s businesses’ stock price

When more and more COVID-10 cases occur, countries from all over the world tend to close their borders and preserve their citizen’s health as much as they can This action can restrict people from moving and therefore cut their basic daily needs as well When the opportunity comes, E-commerce companies overwhelmingly thrive from these bad situations when they can provide foods and products to customers while complying with social distance rules from these countries

The detailed Stata processes of these calculations can be view in Appendix

Limitations

First of all, it is glaringly apparent that our model has too few independent variables, which can drastically increase the disturbance of the model This means we might u have not considered or included all the factors contributing to the growth of the companies’ stock price, such as other financial statistics, demand of traders, etc Unfortunately, due to our inability to find any other relevant variable with daily values, we had to accept this defection in our model Nevertheless, the R-squared values of our OLS models are relatively acceptable, with only one exception of the smallest being 0.2384 As a result, we are convinced that we have selected two of the major variables affecting the stock prices of chosen companies

Secondly, another major drawback of our model is the overgeneralization stemming from the fact that we have chosen to analyze the effect of COVID- 19 on the global E- commerce industry through its influence on the stock prices of five top companies in the industry This can be deemed as biased because their relations with COVID-19 might not be the best measurement of the pandemic’s effect on the industry as a whole, because the effect might be tremendously different when we consider smaller companies, even being opposed to the result we have collected from this research This is due to the fact that the size, market share, business capability and other factors of a firm play major roles in how they are affected by COVID- 19 Without taking the smaller companies into account, the models can generate bias results

Additionally, choosing stock price as the main indicator of a company’s financial performance can be a limitation to our model Other financial statistics, such as net profit (loss), net cash flow, expense, etc can serve as better measurement for a company’s business status However, as explained above, we were unable to compile a usable database with these statistics due to the low number of total observations and rare occurrence of COVID- 19 in that model Inevitably, we must utilize stock prices as a means to estimate the selected companies’ financial performance, as they have a sufficient number of values on a daily basis for us to create a large scale database that can effectively demonstrate the effect of COVID- 19

Despite these limitations, we believe this research paper can contribute as a basis to further research in the future into the effect of COVID- 19 to E-commerce and other industries as well.

Conclusion and Recommendation

Conclusion

Because of the widespread of COVID- 19 pandemic, social distancing and staying at home are expected to push shoppers to shop online A large number of suppliers facing a reduction in leisure purchases, supply chains and an increase in purchases of basic toiletries, basic supplies and other items are also affected by the pandemic Due to the supply chain instability and global customer requirements, the COVID- 19 incident has had a huge impact on the E-commerce market

The analysis conducted in this paper proves that COVID- 19 has a positive effect on the stock price of all 5 E-commerce companies: Amazon, Alibaba, Jingdong, Ebay and Rakuten

This result implies that even though other businesses suffer from the pandemic, E- commerce is still consistent or rather doing better than before the pandemic Due to its contactless nature while having a potential reach of everywhere on Earth, this aligns extremely well with the compulsory situation we have to place ourselves in - quarantine Unlike traditional commercial, E-commerce is relatively safer for users as they can order things virtually and do not have to physically contact anyone The number of rising cases and deaths resulting in higher stock values and interest over time also implies that more and more people are aware of the level of threat COVID- 19 can pose on one’s health As a result, they are bound to stay at home more often, if not to say all the time, and transferring all of their normal physical work to online: from their job to their hobbies and in this case, their shopping platform More people staying at home helped the E-commerce business to be more profitable than ever before, due to the suitable characteristics of the business There is only 1 major negative effect that COVID-19 had on E-commerce business: supply-chain The shipping/delivering process is slowed down by the pandemic However, consumers are still buying because they have no other alternatives.

Recommendation

A strategy for COVID and post-COVID situation is a must, and the crisis plan will have to be conducted fast and reactively according to many scenarios

Firstly, finding ways to increase shipping speed or rather improve the customers’ experience while waiting for their items to arrive is important and should be focused on because it is the supply chain that tends to hinder E-commerce growth Logistics and postal services have been slowed in many countries, due to new COVID- 19 related safety guidelines and government recommendations Specifically, there are recommendations that should be used by the majority of businesses:

- Simplifying internal processes (be mindful and not turn a simple task into a complicated and end up with a slow process in the end, along with side effects such as frustration If your process is long and tedious, consider starting from scratch and building up a more simple, quick and efficient process

- Using Electronic data interchange, which is a technique that replaces the paperwork with electronic data in order to speed up the process as a whole

- Provide reliable, helpful online tracking information By letting customers have plenty of options for package tracking, such as on-site tracking, links to the carrier’s site or the ability to track their orders from mobile devices can greatly enhance the online experience

Secondly, firms should get creative on many channels outside of the ones currently in use to stay engaged with many communities, at the same time prepare to support those in difficult times as they will be more likely to support us back Connection and support is very important in such uncertain times Some recommended channels: Search engines like Google, Bing, Yahoo, DuckDuckGo Comparison shopping engines: Idealo, PriceGrabber,etc Email marketing is another good way to capture those shoppers who are on the edge of making a purchase, with this channel one can target different behavior-based audiences, for example: those who abandon shopping carts when shopping online may get a reminder, or those who have a birthday in May will get a discount, etc There are many other channels that E-commerce businesses can try and expand on for their own companies in order to reach out more target audiences and fulfill their goals.

TABLES

Number of observations Maximum Minimum Mean

Standard deviation ttc 371 1.46e+08 0 3.81e+07 4.49e+07 nc 371 903747 0 279755.1 256828.8 ttd 371 3086756 0 938929 949495.1 nd 371 17906 0 6214.814 4839.212

Table A.1 Summary of COVID-19’s statistics

Variable Number of observations Maximum Minimum Mean

Standard deviation amzn 371 3531.45 1676.61 2689.488 593.5119 baba 371 317.14 176.34 235.722 33.57524 jd 371 106.88 31.49 63.857 21.57024 ebay 371 64.93 26.34 47.457 9.974053 rkuny 371 14.12 6.22 9.538 1.394147

Table A.2 Summary of selected companies’ stock prices

Variable Number of observations Maximum Minimum Mean

Standard deviation amznint 371 100 63 77.10512 8.841074 babaint 371 100 46 57.89757 8.749688 jdint 371 100 27 50.97305 13.10722 ebayint 371 99 65 81.72507 7.973937 rkunyint 371 100 64 79.79784 9.015578

Table A.3 Summary of selected companies’ interest over time

Table A.4 Summary of selected companies’ interest over time amzn baba ebay rkuny jd ttc nc ttd nd amzn 1 baba 0.8137* 1

0 0 0 0 amzn baba ebay rkuny jd amznint babaint ebayint jdint rkunyint amzn 1 baba 0.8137* 1

0.0003 0.6944 0.0249 0.0358 0 0 0.2573 0.0008 0.0003 amzn baba ebay rkuny jd ttc nc ttd nd ttc 0.7047* 0.3982* 0.7579* 0.7784* 0.8234* 1

Table A.5 Correlation between stock price and COVID-19’s statistics

Variable Beta R squared Adj R- squared p-value VIF White test d-statistic amzn 001913 0.6940 0.6924 0.000 1.00 0.0000 137305 ebay 0000229 0.6073 0.6051 0.000 1.83 0.0000 0809741 jd 0000696 0.7548 0.7535 0.000 1.30 0.0000 1565603 baba 0000879 0.3852 0.3819 0.000 1.18 0.0000 0956151

Table A.6 OLS result of the effect of Daily new confirmed cases

Variable Beta R squared Adj R- squared p-value VIF White test d-statistic ebay 1.38e-07 0.5902 0.5879 0.000 2.22 0.0000 0309073 jd 3.59e-07 0.7396 0.7381 0.000 1.09 0.0000 0684753

Table A.7 OLS result of the effect of Total confirmed cases

Variable Beta R squared Adj R- squared p-value VIF White test d-statistic amzn 0986512 0.6543 0.6524 0.000 1.00 0.0000 2689033 ebay 0011214 0.6261 0.6241 0.000 1.42 0.0000 1510583 jd 0035041 0.7340 0.7325 0.000 1.22 0.0000 3299079 rkuny 0001928 0.3845 0.3811 0.000 1.20 0.0000 1768873

Table A.8 OLS result of the effect of Daily new confirmed death

Variable Beta R squared Adj R- squared p-value VIF White test d-statistic amzn 0004996 0.6123 0.6102 0.000 1.06 0.0000 0270502 jd 0000182 0.7916 0.7905 0.000 1.13 0.0000 0683064 baba 0000185 0.2384 0.2343 0.000 1.16 0.0000 044816 lkruny 1.24e-07 0.6452 0.6433 0.000 1.05 0.0027 0997054

Table A.9 OLS result of the effect of Total confirmed death

STATA RESULTS

_cons 2732.258 151.0458 18.09 0.000 2435.237 3029.279 nc 001913 0000666 28.70 0.000 0017819 002044 amznint -7.495359 1.936121 -3.87 0.000 -11.30261 -3.688109 amzn Coef Std Err t P>|t| [95% Conf Interval] Total 130334871 370 352256.407 Root MSE = 329.19 Adj R-squared = 0.6924 Residual 39879750 368 108368.886 R-squared = 0.6940 Model 90455120.6 2 45227560.3 Prob > F = 0.0000 F(2, 368) = 417.35 Source SS df MS Number of obs = 371 reg amzn amznint nc

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 30.37 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Table set B.2 Stata results o : f Ebay’s stock price (ebay) and Daily new confirmed cases (nc):

_cons 2732.258 246.9758 11.06 0.000 2246.597 3217.919 nc 001913 0001334 14.34 0.000 0016506 0021753 amznint -7.495359 3.094574 -2.42 0.016 -13.58063 -1.410091 amzn Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 102.83 Regression with Newey-West standard errors Number of obs = 371 newey amzn amznint nc, lag(4)

_cons 66.58112 4.862528 13.69 0.000 57.0193 76.14295 nc 0000229 1.72e-06 13.33 0.000 0000195 0000262 ebayint -.3123024 0552535 -5.65 0.000 -.4209545 -.2036502 ebay Coef Std Err t P>|t| [95% Conf Interval] Total 36808.2426 370 99.4817368 Root MSE = 6.2676 Adj R-squared = 0.6051 Residual 14456.0249 368 39.2826764 R-squared = 0.6073 Model 22352.2177 2 11176.1089 Prob > F = 0.0000 F(2, 368) = 284.50 Source SS df MS Number of obs = 371 reg ebay ebayint nc

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 73.67 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Table set B.3: Stata results of Jingdong’s stock price (jd) and Daily new confirmed case (nc):

_cons 66.58112 10.84026 6.14 0.000 45.26451 87.89774 nc 0000229 4.06e-06 5.63 0.000 0000149 0000309 ebayint -.3123024 117451 -2.66 0.008 -.5432618 -.081343 ebay Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 120.04 Regression with Newey-West standard errors Number of obs = 371 newey ebay ebayint nc, lag(4)

_cons 37.92766 2.290082 16.56 0.000 33.42437 42.43095 nc 0000696 2.47e-06 28.16 0.000 0000648 0000745 jdint 1264752 0484567 2.61 0.009 0311885 221762 jd Coef Std Err t P>|t| [95% Conf Interval] Total 172151.766 370 465.275043 Root MSE = 10.71 Adj R-squared = 0.7535 Residual 42213.6715 368 114.711064 R-squared = 0.7548 Model 129938.094 2 64969.0472 Prob > F = 0.0000 F(2, 368) = 566.37 Source SS df MS Number of obs = 371 reg jd jdint nc

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 58.88 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

_cons 37.92766 3.030001 12.52 0.000 31.96937 43.88595 nc 0000696 5.94e-06 11.72 0.000 000058 0000813 jdint 1264752 0632818 2.00 0.046 002036 2509145 jd Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 151.00Regression with Newey-West standard errors Number of obs = 371 newey jd jdint nc, lag(4)

Table set B.4: Stata results of Alibaba’s stock price (baba) and Daily new confirmed case (nc):

_cons 257.7734 9.454845 27.26 0.000 239.1811 276.3657 nc 0000879 5.81e-06 15.13 0.000 0000765 0000993 babaint -.8055151 1705173 -4.72 0.000 -1.140826 -.4702045 baba Coef Std Err t P>|t| [95% Conf Interval] Total 417099.741 370 1127.2966 Root MSE = 26.397 Adj R-squared = 0.3819 Residual 256417.049 368 696.785459 R-squared = 0.3852 Model 160682.692 2 80341.3458 Prob > F = 0.0000 F(2, 368) = 115.30 Source SS df MS Number of obs = 371 reg baba babaint nc

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 37.78 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Table set B.5: Stata results of Ebay’s stock price (ebay) and Total confirmed case (ttc):

_cons 257.7734 17.37655 14.83 0.000 223.6036 291.9432 nc 0000879 000012 7.30 0.000 0000642 0001115 babaint -.8055151 2974926 -2.71 0.007 -1.390514 -.2205163 baba Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 26.84 Regression with Newey-West standard errors Number of obs = 371 newey baba babaint nc, lag(4)

_cons 61.32121 5.41466 11.33 0.000 50.67365 71.96876 ttc 1.38e-07 1.10e-08 12.45 0.000 1.16e-07 1.59e-07 ebayint -.2338041 0622176 -3.76 0.000 -.3561506 -.1114575 ebay Coef Std Err t P>|t| [95% Conf Interval] Total 36808.2426 370 99.4817368 Root MSE = 6.4027 Adj R-squared = 0.5879 Residual 15085.8475 368 40.9941508 R-squared = 0.5902 Model 21722.3951 2 10861.1976 Prob > F = 0.0000 F(2, 368) = 264.95 Source SS df MS Number of obs = 371 reg ebay ebayint ttc

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 129.53 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

_cons 61.32121 12.74249 4.81 0.000 36.26397 86.37844 ttc 1.38e-07 2.62e-08 5.25 0.000 8.60e-08 1.89e-07 ebayint -.2338041 1393344 -1.68 0.094 -.5077956 0401874 ebay Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 135.63Regression with Newey-West standard errors Number of obs = 371 newey ebay ebayint ttc, lag(4)

Table set B.6: Stata results of Jingdong’s stock p rice (jd) and Total confirmed case (ttc):

_cons 28.41409 2.309995 12.30 0.000 23.87165 32.95654 ttc 3.59e-07 1.33e-08 26.93 0.000 3.33e-07 3.86e-07 jdint 426456 0457385 9.32 0.000 3365144 5163976 jd Coef Std Err t P>|t| [95% Conf Interval] Total 172151.766 370 465.275043 Root MSE = 11.038 Adj R-squared = 0.7381 Residual 44836.5644 368 121.83849 R-squared = 0.7396 Model 127315.201 2 63657.6007 Prob > F = 0.0000 F(2, 368) = 522.48 Source SS df MS Number of obs = 371 reg jd jdint ttc

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 74.74 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Table set B.7: Stata results of Amazon’s stock price (amzn ) and Daily new confirmed deaths (nd):

_cons 28.41409 4.028629 7.05 0.000 20.49207 36.33611 ttc 3.59e-07 3.71e-08 9.70 0.000 2.86e-07 4.32e-07 jdint 426456 0841233 5.07 0.000 2610333 5918786 jd Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 95.10 Regression with Newey-West standard errors Number of obs = 371 newey jd jdint ttc, lag(4)

_cons 2784.168 160.4011 17.36 0.000 2468.75 3099.586 nd 0986512 003764 26.21 0.000 0912495 1060529 amznint -9.1794 2.060261 -4.46 0.000 -13.23076 -5.128039 amzn Coef Std Err t P>|t| [95% Conf Interval] Total 130334871 370 352256.407 Root MSE = 349.9 Adj R-squared = 0.6524 Residual 45054905.4 368 122431.808 R-squared = 0.6543 Model 85279965.1 2 42639982.6 Prob > F = 0.0000 F(2, 368) = 348.28 Source SS df MS Number of obs = 371 reg amzn amznint nd

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 36.58 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Table set B.8: Stata results of Ebay’s stock price (ebay) and Daily new confirmed deaths (nd):

_cons 2784.168 243.0473 11.46 0.000 2306.232 3262.104 nd 0986512 0064846 15.21 0.000 0858997 1114028 amznint -9.1794 3.126701 -2.94 0.004 -15.32784 -3.030958 amzn Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 118.67 Regression with Newey-West standard errors Number of obs = 371 newey amzn amznint nd, lag(4)

_cons 76.3245 4.177383 18.27 0.000 68.10997 84.53904 nd 0011214 0000783 14.33 0.000 0009675 0012752 ebayint -.4384939 0474889 -9.23 0.000 -.5318775 -.3451104 ebay Coef Std Err t P>|t| [95% Conf Interval] Total 36808.2426 370 99.4817368 Root MSE = 6.1151 Adj R-squared = 0.6241 Residual 13761.373 368 37.3950353 R-squared = 0.6261 Model 23046.8696 2 11523.4348 Prob > F = 0.0000 F(2, 368) = 308.15 Source SS df MS Number of obs = 371 reg ebay ebayint nd

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 50.96 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

_cons 76.3245 8.240813 9.26 0.000 60.11951 92.5295 nd 0011214 0001528 7.34 0.000 0008208 0014219 ebayint -.4384939 0888825 -4.93 0.000 -.6132752 -.2637127 ebay Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 180.35Regression with Newey-West standard errors Number of obs = 371 newey ebay ebayint nd, lag(4)

Table set B.9: Stata results of Jingdong’s stock price (jd) and Daily new confirmed deaths (nd):

_cons 30.03892 2.339888 12.84 0.000 25.43769 34.64015 nd 0035041 0001322 26.50 0.000 003244 0037641 jdint 2362232 0488208 4.84 0.000 1402206 3322259 jd Coef Std Err t P>|t| [95% Conf Interval] Total 172151.766 370 465.275043 Root MSE = 11.156 Adj R-squared = 0.7325 Residual 45796.6242 368 124.447348 R-squared = 0.7340 Model 126355.142 2 63177.5708 Prob > F = 0.0000 F(2, 368) = 507.67 Source SS df MS Number of obs = 371 reg jd jdint nd

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 86.96 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Table set B.10: Stata results of Rakuten’s stock price (rkuny) and Daily new confirmed deaths (nd):

_cons 30.03892 4.041314 7.43 0.000 22.09196 37.98589 nd 0035041 0002168 16.16 0.000 0030777 0039304 jdint 2362232 0883591 2.67 0.008 0624712 4099753 jd Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 281.65 Regression with Newey-West standard errors Number of obs = 371 newey jd jdint nd, lag(4)

_cons 10.37738 5285373 19.63 0.000 9.33805 11.41671 nd 0001928 0000129 14.93 0.000 0001674 0002182 rkunyint -.0255289 0069327 -3.68 0.000 -.0391616 -.0118961 rkuny Coef Std Err t P>|t| [95% Conf Interval] Total 719.148596 370 1.94364485 Root MSE = 1.0967 Adj R-squared = 0.3811 Residual 442.648947 368 1.2028504 R-squared = 0.3845 Model 276.499649 2 138.249824 Prob > F = 0.0000 F(2, 368) = 114.94 Source SS df MS Number of obs = 371 reg rkuny rkunyint nd

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0009 chi2(5) = 20.70 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Table set B.11 : S tata results of Amazon ’s stock price ( amzn) and Total confirmed deaths (ttd):

_cons 10.37738 1.025693 10.12 0.000 8.360426 12.39434 nd 0001928 0000293 6.57 0.000 0001351 0002505 rkunyint -.0255289 0133919 -1.91 0.057 -.0518631 0008054 rkuny Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 22.12 Regression with Newey-West standard errors Number of obs = 371 newey rkuny rkunyint nd, lag(4)

_cons 1738.423 179.5358 9.68 0.000 1385.378 2091.468 ttd 0004996 0000209 23.93 0.000 0004585 0005406 amznint 6.251247 2.241942 2.79 0.006 1.842622 10.65987 amzn Coef Std Err t P>|t| [95% Conf Interval] Total 130334871 370 352256.407 Root MSE = 370.54 Adj R-squared = 0.6102 Residual 50525431.5 368 137297.368 R-squared = 0.6123 Model 79809439 2 39904719.5 Prob > F = 0.0000 F(2, 368) = 290.64 Source SS df MS Number of obs = 371 reg amzn amznint ttd

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 77.30 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

_cons 1738.423 296.5347 5.86 0.000 1155.308 2321.538 ttd 0004996 0000382 13.08 0.000 0004245 0005747 amznint 6.251247 3.760562 1.66 0.097 -1.14364 13.64613 amzn Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 90.51Regression with Newey-West standard errors Number of obs = 371 newey amzn amznint ttd, lag(4)

Table set B.12: Stata results of Jingdong’s stock price (jd) and Total confirmed deaths (ttd):

_cons 29.92442 2.069455 14.46 0.000 25.85498 33.99386 ttd 0000182 5.76e-07 31.60 0.000 0000171 0000193 jdint 3307192 0416922 7.93 0.000 2487344 4127041 jd Coef Std Err t P>|t| [95% Conf Interval] Total 172151.766 370 465.275043 Root MSE = 9.8727 Adj R-squared = 0.7905 Residual 35869.3359 368 97.4710215 R-squared = 0.7916 Model 136282.43 2 68141.215 Prob > F = 0.0000 F(2, 368) = 699.09 Source SS df MS Number of obs = 371 reg jd jdint ttd

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 126.10 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Table set B.13: Stata results of Alibaba’s stock price (baba) and Total confirmed deaths (ttd):

_cons 29.92442 3.404436 8.79 0.000 23.22983 36.61901 ttd 0000182 1.74e-06 10.47 0.000 0000148 0000216 jdint 3307192 0762279 4.34 0.000 1808223 4806162 jd Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 113.34 Regression with Newey-West standard errors Number of obs = 371 newey jd jdint ttd, lag(4)

_cons 249.4461 10.50122 23.75 0.000 228.7962 270.0961 ttd 0000185 1.73e-06 10.67 0.000 0000151 0000219 babaint -.5367121 1879607 -2.86 0.005 -.9063239 -.1671004 baba Coef Std Err t P>|t| [95% Conf Interval] Total 417099.741 370 1127.2966 Root MSE = 29.38 Adj R-squared = 0.2343 Residual 317645.248 368 863.166434 R-squared = 0.2384 Model 99454.4926 2 49727.2463 Prob > F = 0.0000 F(2, 368) = 57.61 Source SS df MS Number of obs = 371 reg baba babaint ttd

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0000 chi2(5) = 32.54 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Table set B.14: Stata results of percentage of change of Rakuten’s stock price (lkruny) and Total confirmed deaths (ttd):

_cons 249.4461 18.39126 13.56 0.000 213.281 285.6113 ttd 0000185 3.59e-06 5.15 0.000 0000114 0000255 babaint -.5367121 3249103 -1.65 0.099 -1.175626 1022016 baba Coef Std Err t P>|t| [95% Conf Interval] Newey-West

Prob > F = 0.0000 maximum lag: 4 F( 2, 368) = 13.38 Regression with Newey-West standard errors Number of obs = 371 newey baba babaint ttd, lag(4)

_cons 2.21201 0402024 55.02 0.000 2.132955 2.291065 ttd 1.24e-07 4.85e-09 25.59 0.000 1.14e-07 1.34e-07 rkunyint -.0010474 0005104 -2.05 0.041 -.0020511 -.0000436 lrkuny Coef Std Err t P>|t| [95% Conf Interval] Total 7.714085 370 020848878 Root MSE = 08624 Adj R-squared = 0.6433 Residual 2.73677765 368 007436896 R-squared = 0.6452 Model 4.97730735 2 2.48865368 Prob > F = 0.0000 F(2, 368) = 334.64 Source SS df MS Number of obs = 371 reg lrkuny rkunyint ttd

Ngày đăng: 10/08/2024, 20:04

TÀI LIỆU CÙNG NGƯỜI DÙNG

TÀI LIỆU LIÊN QUAN