CC BY
44
Section 4. World economy
https://doi.org/10.29013/EJEMS-22-2-37-51
Andrew Zhe Chen, Canterbury School, CT
COVID IMPACT ON GLOBAL ECONOMY
Abstract. Millions of people lost their jobs and small businesses, hospitals and medical care professionals experienced unprecedented hardships, children all over the world face the loss of family members and the psychological consequences of the isolation from remote learning... just to name a few aftermath caused by the pandemic. This paper aims to draw insights from analyzing the pandemic data. The data used in this study includes 210 countries concerning the impact of covid-19 on the global economy. 6 feature indicators including 4 economic indicators and COVID-19 Total Cases (TC) and Total Deaths (TD). The economic indicators are Human Development Index (HDI), Stringency Index (STI), Population (POP), and Gross Domestic Product per capita (GDPCAP). The Organization for Economic Co-operation and Development (OECD) unemployment data is also included in the analysis, which includes 38 member countries' unemployment values from January 2020 to October 2020. From these key dimensions of global development, lessons can be learned to aid government officials in response to future pandemics.
Keywords: Covid19, economic indicator, impact, employment, global economy.
1. Introduction including 4 economic indicators and COVID-19
The COVID-19 pandemic disrupted almost Total Cases (TC) and Total Deaths (TD). The eco-
every corner of the world in countless ways for the nomic indicators are Human Development Index
past two and half years. Millions of people lost their (HDI), Stringency Index (STI), Population (POP),
jobs and small businesses, hospitals and medical care and Gross Domestic Product per capita (GDPCAP).
professionals experienced unprecedented hardships, The HDI is the composite statistic of life expectancy,
many children all over the world face the loss of fam- education, literacy, and income indices used to rank
ily members and the psychological consequences of countries into four stages of human development. the isolation from remote learning. just to name The stringency Index represents the system-
a few. While we can't solve or alleviate any of these atically collected information on several different
problems by looking back at the pandemic data, in- common policy responses governments have taken,
sights and lessons for the future can be drawn by measured, and aggregated. STI is a composite mea-
harvesting the power of data. sure of nine of the response metrics calculated by
The data used in this study includes 210 coun- The Oxford Coronavirus Government Response
tries concerning the impact of covid-19 on the global Tracker (OxCGRT) project. The metrics include
economy. 50418 data entries and 6 feature indicators school closures; workplace closures; cancellation
of public events; restrictions on public gatherings; closures of public transport; stay-at-home requirements; public information campaigns; restrictions on internal movements; and international travel controls.
The Organization for Economic Co-operation and Development (OECD) unemployment data is also included in the analysis, which includes 38 member countries' unemployment values from January 2020 to October 2020. The OECD is a group of developed countries with high HDI that commit to democracy and market economy intending to stimulate economic progress and world trade.
Although we can't paint a complete picture of the world with only these economic indicators, financial hardship is an integral part of people's suffering and an important metric for governments' performances.
2. Data analysis and Results
The data (.csv file) is imported as a DataFrame using Python programming language. Figure 1 shows the result from the.info() function from the Pandas library. The HDI column has 6202 null values according to figure 1. No obvious missing values in other columns and three columns have object data
types while the rest are all numeric. Figure 2 visualizes the dispersed missing values in the HDI column.
<class 'pandas.core.frame.DataFrame'> Rangelndex: 5041& entries, 0 to 50411 Data columns (total 9 columns 1:
# Column Non-Hull Count Dtype
0 CODE 50418 non- -null object
1 COUNTRY 50418 non- -null object
2 DATE 50418 non- -null object
3 HDI 44216 non- -null float64
4 TC 50418 non- -null float64
5 TD 50418 non- -null float64
6 STI 50418 non- -null floats4
1 POP 50418 non- -null float64
a GDPCAP 50418 non- -null float64
dtypesî float64 ft), object(3) memory usage: MB
Figure 1. Information of the dataset
To aid the further investigation, the DATE column was transferred to the Date Time type using the to_datetime function from the Pandas library. The time frame of the dataset is from 2019-12-31 to 2020-10-19. The month information is extracted and formed a new column called "MONTH".
Figure 2. Missing values in the dataset
Figure 3. Histograms of each feature, color-coded by continent
With the pycountry_convert library, continent information was created from each country name. The following countries resulted in an "Unknown" continent: 'Bonaire Sint Eustatius and Saba', «Cote d'lvoire», 'Curacao', 'Democratic Republic of Congo', 'Faeroe Islands', 'Kosovo', 'Sint Maarten (Dutch part)', 'Timor', 'Vatican'.
Figure 3 shows the histograms for every feature column in "stacked" mode color-coded by the Continent. Histograms visualize the distribution of data
with the x-axis being the "bins" and the y-axis is the count of observance for the data that fall inside each range (bin). Missing values are omitted in the histograms. In the HDI histogram, a small amount of"Un-known" Continent data is close to zero which means potential missing data as well. The same conclusion can be drawn for the STI and GDPCAP columns according to the COVID-19 stringency index data from Our World in Data and the GDP per capita data from the World Bank.
Figure 4. Heatmap of the correlation matrix of the features
After merging the unemployment data with the previous DataFrame, the correlation matrix can be calculated and shown as a heat map in figure 4. A heat map is a visualization tool that uses color intensity to represent the magnitude of numbers. From figure 4, HDI and GDP per capita are highly correlated with
a correlation coefficient of 0.85. Total cases and total deaths, total cases, and STI are highly correlated as well. TD and STI, POP and TC, POP and TD, POP and STI, POP and GDPCAP, GDP and TD are all moderately correlated. HDI and unemployment, GDPCAP, and unemployment are negatively correlated.
Figure 5. Pair plot of features to show pair-wise relations
HDI (human development index) Figure 5 shows the pairwise relationships between each feature column. It is color-coded by the continent column as well. TC and TD, GDPCAP and HDI both have a linear relationship with each other.
After dropping missing values in the HDI column, 182 countries remained. The index also stayed constant during the time frame in the dataset. From figure 6, the top 5 countries with the highest HDI are Germany, Ireland, Australia, Switzerland, Nor-
way; the bottom 5 countries with the lowest HDI are Kosovo, Niger, Central African Republic, South Sudan, Chad.
Figures 7, 8, 9, 10, 12, 14, 15, 16, 17 are generated by the lineplot() function from the Seaborn library which automatically calculates the average value of the y-axis at any given point on the x-axis and generates a colored range for the y values.
Figure 7 shows that European countries have the highest average HDI, the second-highest continent is Australia. The lowest HDI continent is the "Unknown" category.
CODE
25104 OWID KOS
COUNTRY CONTINENT HDI
Kosovo Unknown 0.000
33784 NER Niger AF 0.354
9180 CAF Central African Republic AF 0.367
42568 SSD South Sudan AF 0.388
9398 TCD Chad AF 0.404
17652 □ EU Germany EU 0.936
22604 IRL Ireland EU 0.938
2434 AUS Australia OC 0.939
4429 E CHE Switzerland EU 0.944
34494 NOR Norway EU 0.953
Figure 6. Top 5 and bottom 5 countries ranked by HDI
Figure 7. Line plot of HDI in each continent
Total Cases (TC) and Total Deaths (TD) The COVID-19 total cases and total deaths data recorded 210 countries for most of 2020. Figure 8 shows the time-series data of total cases color-coded by continent. There is a sharp increase worldwide in March 2020 and South America has the biggest
increase since then. Figure 9 confirms that South America has the most total cases with Europe being the second. Figure 10 shows the time-series data of total deaths color-coded by continent. South America has the most total deaths again. Oceania has the lowest case and death count.
Figure 8. Line plot of total cases over time, color-coded by continent
Figure 9. Total cases in each continent
27 41 20D
Figure 10. Line plot of total deaths over time, color-coded by continent
Figure 11 lists the top and bottom 5 countries ranked by total cases. Hong Kong, Solomon Islands, Anguilla, Falkland Islands, and Bonaire Sint Eusta-tius and Saba have the least amount of total cases. Spain, Italy, Brazil, China, and United States have the highest count of total cases. Figure 12 shows the time series plot of total cases in 10 countries with the most COVID-19 cases. China had the earliest increase but contained the situation rapidly and steadily, while the United States had the fastest increase and the highest count from April to November 2020.
Brazil 2913.256895 China 3128.587021 UflIted state S 3343.05075 5
COUNTRY TC
87 Hong Kong 0.000000
173 Solomon Islands 3.988984
5 Anguila 224.574488
65 Falkland Islands 493.373289
24 Bonaire Sint Eustatius and Saba 497.356421
178 Spain 2843.820343
97 Italy 2873.314489
Figure 11. Top 5 and bottom 5 countries in total cases
Figure 12. 10 countries with the most COVID cases
28 countries had zero total deaths during the time in the data. Some examples are Monaco, Timor, Liechtenstein, British Virgin Islands, and Gibraltar as shown in figure 13. The five countries that had the most deaths due to the pandemic are France, China, the United Kingdom, Italy, and United States. Figure 14 shows the time series plot of total deaths in 10 countries with the most deaths from the pandemic. Similar conclusions can be drawn as the total cases in figure 12.
STI (stringency index)
The OxCGRT's STI is a composite measure of nine of the response metrics rescaled to a value from 0 to 100 (100 = strictest). It represents the different common policy responses governments have taken. There are 210 unique countries in the dataset, 30 of those have STI of zero which are missing values. Figure 15 shows the time series plot of STI color-coded by continents. Australia, Africa, and South America have the strictest pandemic responses.
COUNTRY TD
127 Monaco 0.000000
190 Timor 0.000000
113 Liechtenstein 0.000000
28 British Virgin Islands 0.000000
75 Gibraltar 0.000000
68 France 2203.766710
41 China 2205.740166
199 United Kingdom 2209.519314
97 Italy 2327.710204
200 United States 2480.576064
Figure 13. Top 5 and bottom 5 countries in total deaths
2020-01 2020-02 2020-03 2020-04 2020-05 2020-06 2020-07 2020-OB 2020-09 2020-10 2020-11
DATE
Figure 14. 10 countries with the most deaths caused by the pandemic
2020-01 2020-02 2020-03 2020-04 2020-05 2020-06 2020-07 2020-06 2020-09 2020-10 2020-11
DATE
Figure 15. Line plot of stringency index over time, color-coded by continent
Figure 16. United States STI over time
Figure 17. China STI over time
Figures 16 and 17 illustrate the STI of the US and pression. Figure 18 shows the five strictest countries
China over time. US's response to covid is slower towards the pandemic are Panama, Uganda, El Sal-
and more gradual than China's. However, after April vador, Honduras, Eritrea. 2020, they had similar STIs despite the common im-
CONTINENT COUNTRY STI
209 Unknown Vatican 0,000000
127 EU Malta 0,000000
129 EU Monaco 0,000000
130 EJ Montenegro 0,000000
123 EU Liechtenstein 0,000000
170 MA Panama 4,374658
49 AF Uganda 4,406101
160 MA El Salvador 4,414523
165 MA Honduras 4,424304
15 AF Eritrea 4,507309
Figure 18. Countries ranked by their average STI value over the time in the data
Population (POP)
Population in the 210 countries stayed constant within the time frame in the dataset. The top five countries are India, China, the United States, Indonesia, and Pakistan. And the five countries with the smallest population are the Vatican, Falkland Islands, Montserrat, Anguilla, and Bonaire Sint Eustatius and Saba. And Anguilla, Falkland Islands, and Bonaire Sint Eustatius and Saba are among the five countries with the least amount of total cases.
The Gross Domestic Product per capita (GDP-CAP)
As justified earlier, the countries with the GDP-CAP of zero are missing data. Out of the 210 countries in the dataset, 27 countries have GDPCAP=0. From figure 20, the richest countries are San Marino, Switzerland, Norway, Kuwait, United Arab Emirates, Ireland, Brunei, Singapore, Luxembourg, and Qatar.
COUNTRY POP
49D62 Vatican 6.695799
1565 G Falkland Islands 8.155649
313G3 Montserrat 8.516993
1251 Anguilla 9.615939
5945 Bonaire Sint Eustatius and Saba 10.174316
35082 21722 48121 21428 9841
Pakistan 19.213186
Indonesia 19.426899
United States 19.617637
India 21.045353
China 21.087439
Figure 19. Top 5 and bottom 5 of the countries ranked by population
CONTINENT COUNTRY GDPCAF
39208 EU San Marino 10.948373
4429 G EU Switzerland 10.957977
34494 EU Norway 11.079062
25324 AS Kuwait 11.090272
47533 AS United Arab Emirates 11,116819
22604 EU Ireland 11,117440
7072 AS Brunei 11.181769
40821 AS Singapore 11.356685
27693 EU Luxembourg 11.454003
37464 AS Qatar 11.669379
Figure 20. Top countries in GDP per capita Unemployment in OECD countries Even before the pandemic hit, global economic growth had slowed. COVID-19 put a big dent in the world economy. Millions of jobs were lost and the worst recession since the Great Depression happened in 2020.
Among OECD members, the Czech Republic and Japan have the lowest unemployment rate
around 2 and 2.5. Spain, Greece, and Colombia have Australia on average have lower unemployment
the highest unemployment rate at around 16 and 20. rates. North America is more spread out while South
Figure 21 shows the color-coded stacked histogram American countries typically have higher unemploy-
of the unemployment data. European countries and ment rates.
Figure 21. Distribution plot of unemployment rate in OECD countries color-coded by continent
Figure 22. Distribution plot of unemployment rate in OECD countries color-coded by month
Figure 22 shows the stacked histogram of unemployment data color-coded by month. While some countries don't seem to be affected by the pandemic, more countries have higher unemployment rates during the height of the pandemic.
Figure 23 shows the line plot of the unemployment time series data color-coded by continent. The
"Unknown" category experienced the most dramatic change in February and June compared to other continents. Most countries had a sharp increase in March and a gradual fall around June except African and European countries which had a slow increase in most of 2020.
Figure 23. Line plot of unemployment rate in OECD countries color-coded by continent
3. Conclusion
Unsurprisingly, HDI and GDP per capita are highly correlated with a correlation coefficient of 0.85. Total Covid cases and deaths are closely related. And STI is highly correlated with total cases meaning most countries impose more stringent policies in response to higher case counts. The unemployment rates in OECD countries are negatively correlated with HDI and GDP per Capita.
COVID-19 brought the worst recession since the Great Depression in 2020. Most countries have higher unemployment rates during the height of the pan-
demic. European countries and Australia have the highest average HDI and lower unemployment rate on average in comparison with other countries. Most countries had a sharp increase in the unemployment rate in March and a gradual fall around June except African and European countries which had a slow increase in most of 2020. Among OECD members, Spain, Greece, and Colombia have the highest unemployment rate at around 16 and 20.
A sharp increase in total Covid cases also happened worldwide in March 2020 and South America has the biggest increase since then. Spain, Italy, Brazil,
China, and United States have the highest count of total cases. China had the earliest increase but contained the situation rapidly and steadily, while the United States had the fastest increase and the highest count from April to November 2020. The five countries that had the most deaths due to the pandemic are France, China, the United Kingdom, Italy, and United States.
Australia, Africa, and South America have the strictest pandemic responses. US's response to covid is slower and more gradual than China's. However, after April 2020, they had similar STIs despite the
common impression. It can be concluded that the initial response is crucial in fighting the pandemic.
The limitations of this study can come from the explicit and inexplicit missing data in almost all the variable columns. Also, during the data analysis stage, some countries were classified as the "Unknown" continent which has the lowest HDI and experienced the most dramatic change in the unemployment rate in February and June of 2020. The countries in this category may need further investigation.
References:
1. Human Development Index (HDI) // URL: https://hdr.undp.org/en/content/human-development-index-hdi/ (Access date March 12, 2022).
2. A global panel database of pandemic policies (Oxford COVID-19 Government Response Tracker) // URL: https://www.nature.com/articles/s41562-021-01079-8 A global panel database of pandemic policies/ (Oxford COVID-19 Government Response Tracker) (Access date March 12, 2022).
3. OECD. From Wikipedia, the free encyclopedia // URL: https://en.wikipedia.org/wiki/OECD/ (Access date March 12, 2022).
4. COVID-19 Stringency Index, Mar 30, 2022 // URL: https://ourworldindata.org/grapher/covid-strin-gency-index/ (Access date March 30, 2022).
5. GDP per capita (current US$). World Bank national accounts data, and OECD National Accounts data files // URL: https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?end=2020&start=2007/ (Access date March 12, 2022).
6. Impact of Covid-19 Pandemic on the Global Economy. Emphasis on Poverty Alleviation and Economic Growth // URL: https://www.kaggle.com/shashwatwork/impact-of-covid19-pandemic-on-the-global-economy/ (Access date March 20, 2022).
