The sources of the Kuznets relationship between the COVID-19 mortality rate and economic performance.

This paper discusses the findings of an empirical analysis of the Kuznets, or reverse U-shaped relationship, between the COVID-19 mortality rate and economic performance. In the early stages of economic development, the COVID-19 mortality rate is anticipated to rise with rising economic activity and urbanization. Eventually, the mortality rate decreases at higher economic development levels as people and the government are more capable of investing in disease abatement measures. The quality of political institutions, wealth distribution, urbanization, vaccination rate, and improvements in healthcare systems are hypothesized to affect the COVID-19 mortality rate. Examining this relationship can be effective in understanding the change in the COVID-19 mortality rate at different economic performance stages and in identifying appropriate preventive measures. This study employed the negative binomial regression to model a cross-sectional dataset of 137 countries. Results indicated that the relationship between the per-head gross domestic product (GDP) level and the COVID-19 mortality rate appeared to follow a pattern like the Kuznets curve, implying that changes in institutional quality, healthcare advancements, wealth distribution, urbanization, vaccination rate, and the percentage of the elderly population were significant in explaining the relationship. Improvement of the healthcare system has a notable effect on lowering the COVID-19 mortality rate under more effective government conditions. Additionally, the results suggested that a higher per-head GDP is required to reverse the rising trend of the mortality rate under higher income inequality. Based on these results, preventive measures, and policies to reduce COVID-19 mortalities were recommended in the conclusion section.

Introduction

Coronavirus Disease 2019 (COVID-19) has been a major public health issue worldwide since the end of 2019. The outbreak of COVID-19 has posed a significant threat to human life and a considerable burden on the healthcare systems. The World Health Organization (WHO) reported more than 287 million confirmed COVID-19 cases as of December 31, 2021, accounting for over 5.4 million mortalities during the two-year beginning January 1, 2020 [1]. Meanwhile, it was estimated that 18.2 million mortalities occurred during the same period due to under-reporting, particularly in countries with inadequate death registration systems [2]. This equates to approximately 24,932 mortalities each day, or 229 mortalities per hundred thousand population.

The aforementioned statistics demonstrate that the COVID-19 mortality rates varied considerably across countries with diverse socioeconomic levels. Contrary to popular belief, poorer countries (low-income and lower-middle-income economies) suffered the most from the pandemic due to a lack of resources to invest in mitigation measures; however, wealthier countries (upper-middle-income, high-income, and top 10% economies) accounted for nearly 64% of the total excess mortality. Surprisingly, the upper middle-income countries had the highest mortality rate when compared to the higher-income groups. This is illustrated in Fig. 1 .

Fig. 1

Average Excess COVID-19 mortality rate per hundred thousand population by income group.

As shown in Fig. 1, the non-linear relationship between the COVID-19 mortality rate and economic performance is akin to a reverse U-shape pattern in which the relationship becomes negative after a threshold income level. Kuznets [3] first demonstrated a similar pattern between economic performance and income inequality. Then, several studies later discovered that the rise and fall of road crash mortality rates as income increases followed a similar reverse U-shaped path [[4], [5], [6]]; and 2011; [7,8]. This reverse U-shaped relationship is also commonly known as the Kuznets curve relationship.

Numerous previous studies have presented empirical evidence over the past three decades demonstrating the significant impact of economic performance on mortality due to infectious illnesses. Rechel et al. [9] and Hunter et al. [10] revealed that infectious and chronic disease mortality rates inclined with economic downturn. According to Cutler et al. [11]; infectious disease mortality rates rose about 5%–7% during the economic crisis. In Peru, a decrease in Gross Domestic Product (GDP) by 30% in the late 1980s was associated with an increase of 1.5% in the population infected with cholera [12]. Shilova and Dye [13] revealed that the prevalence and mortality of tuberculosis have increased significantly following the massive financial crisis in the nations of the former Soviet Union and Eastern Europe in the 1990s. Kentikelenis et al. [14] and Vignier and Bouchaud [15] pointed out that travel and migration are particularly at risk of developing severe infectious diseases during the economic downturn. Fan et al. [16] found that economic development reduces the COVID-19 vulnerability. Zilidis et al. [17] suggested that economic crisis has a significant impact on the mortality rate of infectious diseases, but the impact varied by the type of infectious diseases.

On the other hand, previous studies on the association between unemployment rate and infectious diseases revealed that a higher unemployment rate was accompanied by a lower infectious disease mortality rate. Goutte et al. [18] and Zilidis et al. [17] discovered that unemployment and poverty rates are important predictors of COVID-19 mortality. Gerdtham and Ruhm [19] indicated that a 1% increase in the unemployment rate decreased mortalities from influenza and pneumonia by 1.1% in the Organization for Economic Co-operation and Development (OECD) countries in the period of 1960–1997. While in Japan during the years between 1950 and 2002, there was a negative correlation between the unemployment rate and mortalities from pneumonia [20].

Although prior research has examined the effect of economic performance on the mortality rate of infectious diseases, there are several gaps in the related literature. First, as indicated in Fig. 1, there is a reverse U-shaped relationship between the COVID-19 mortality rate and economic performance. However, this reverse U-shaped relationship has not been examined in the infectious disease literature. So, examining this relationship can be effective in understanding the change in the COVID-19 mortality rate at different economic performance stages and in identifying appropriate preventive measures.

Second, there have been several empirical investigations into the relationship between the infectious disease mortality rate and economic performance, such as Cutler et al. [11]; Rechel et al. [9]; Hunter et al. [10]; and Fan et al. [16]. While these studies have enriched our understanding of the association between infectious disease mortality rates and economic performance, the mechanism by which economic performance influences this relationship remains unclear. Therefore, several determinants, such as the quality of political institutions, income inequality, urbanization, vaccination rate, and improvements in healthcare systems, have been taken into consideration in this study to explain the proposed reverse U-shaped relationship.

Third, in estimating a relationship with infectious disease mortalities as the dependent variable, one must utilize a count regression method. This is because the number of mortalities is limited to non-negative integer values, positively skewed, and can have a zero count. Therefore, estimation with other regression techniques, such as ordinary least squares [11,16,21,22], generalized additive regression [10], and autoregressive integrated moving average model [17], are not appropriate to model this sort of data. Therefore, this study suggests using the negative binomial regression technique to address this problem and obtain a more accurate estimation.

Fourth, Shahbazi and Khazaei [23] indicated that there is a higher incidence and mortality rate of COVID-19 in countries with a higher Human Development Index (HDI). However, the HDI is a composite index reflecting several economic and social attributes, such as per-head GDP, life expectancy, and educational attainment. As a result, the study's findings do not directly lead to the development of appropriate policies to solve the COVID-19 health crisis.

Based on the above discussions, this paper aims to examine the relationship between the COVID-19 mortality rate and economic performance. Variables fundamental to a nation's socioeconomic progress, such as income inequality, urbanization, effectiveness of government machinery, democracy level, healthcare services, vaccination rate, and an ageing population, were also accounted for to explain the relationship. The findings of this study can be used to strengthen the understanding of factors affecting the mortality rate caused by the COVID-19 as well as the mechanisms influencing the growth rate. This would enable policy makers to understand the exact mechanism by which changes in per-head GDP could affect the outbreak of COVID-19, and to formulate more effective policies to curb this pandemic in nations with different levels of socioeconomic development.

Kuznets theory of infectious disease mortality rate: hypothesis development

There are three theoretical explanations for the Kuznets hypothesis in the literature. These explanations include (1) the scale of economic activity, (2) changes in the demand for a better quality of life, and (3) an increase in the demand for improved health care and services [5,24,25]. Population mobility promotes contact among human beings and, in turn, increases the transmission of infectious diseases that pose a significant risk to global health [26,27]. The first effect is the scale effect, which is a measure of the level of economic activity. Ng et al. [28,29] indicated that at lower levels of economic performance, the population is concentrated in small geographical areas, and economic activities largely take place on a local scale. Thus, there is a greater need for short-distance travel for dispatching products and services. In view of this, the transmission rate of infectious diseases is expected to be lower under these circumstances. As the economy improves with a rising population, urban sprawl pushes households and business services to the suburbs. This increases the degree of mobility and the need for long-distance travel. As a result, this leads to a higher transmission rate of infectious diseases.

The second effect, known as the transition or composition effect, describes a transition from a poorer to a better quality of life as the economy improves. These include access to better basic needs and infrastructure, for instance, shelter, sanitation, water, electricity, education, and health care services [[30], [31], [32]]. This effect implies a reverse U-shape relationship between the COVID-19 mortality rate and economic performance.

The abatement effect relates to disease mitigation measures, which mirrors the influences from both the demand and supply perspectives. At a lower level of economic performance, individuals are less capable of maintaining good health care services even though there is an apparent need for a better level of health care services and facilities. They are more concerned about necessities such as food and shelter. As economic performance improves over time, the demand for better health care services and facilities increases as costly private medical treatment and insurance protection schemes become affordable to more individuals. On the other hand, governments in low-to-middle income nations place higher emphasis on providing adequate supply of fundamental livelihood needs such as food and social security instead of investing in health care and services. As a result, this raises the disease transmission rate. Governments in higher income nations can nevertheless invest in disease abatement services, causing a reduction in disease transmission rates.

COVID-19 mortality rate, urbanization, and economic performance nexus

COVID-19 mortality rate and economic performance nexus

According to the economic study literature, infectious diseases are one of the most severe risks to a nation's economic performance [[33], [34], [35], [36], [37], [38], [39], [40], [41], [42]]. The COVID-19 pandemic was claimed to have had major economic and financial implications for regional and worldwide economies. According to the World Bank [43]; the pandemic caused the global economy to contract by 3.5% in 2020 as a result of rising unemployment [41,[44], [45], [46]], stock market volatility [[47], [48], [49], [50]], a decline in the efficiency of the energy market [[51], [52], [53]], a reduction in tourism expenditure [54], and falling oil prices [49,55,56]. Since March 2020, UNCTAD has predicted a loss of more than $4 trillion USD in global GDP owing to travel cutbacks and restrictions caused by nations' lockdowns [57]. In response to the disease's continuous spread, particularly the Delta variant, the World Economic Prospects (2021) forecast that economic growth will drop by 0.3%–5.9% in late 2021.

However, findings on the influence of economic performance on COVID-19 mortality rates were rather mixed. For example, in Nigeria, while Hassan et al. [58] discovered that higher GDP was associated with higher mortality rates, Illuno et al. [22] concluded that such a relationship was non-linear. They suggested that reducing the COVID-19 mortality might enhance the nation's economic well-being. Meanwhile, Ivanaj and Oukhallou [21] found that economic variables for 107 nations, for instance, the per-head GDP and unemployment, have no direct relationship with the COVID-19 mortality rates. Institutional variables such as regulatory quality, government effectiveness, and corruption control, on the other hand, demonstrated a significant and stable negative relationship across nations.

Further, Gordon et al. [59] provided a very clear relationship between the COVID-19 mortality rate and economic performance in a study using data from Nordic nations, major Western European nations, and OECD nations. They found that Sweden's decision not to close its air borders during the pandemic's “first wave” in the first half of 2020 partly explains its lacklustre performance in comparison to its Nordic neighbours in terms of mortality rates. The study concluded that poorer COVID-19 performance was associated with weaker economic growth.

COVID-19 mortality rate and urbanization nexus

Urbanization is accompanied by an increase in the urban population due to the social, economic, political, and cultural importance of cities in comparison to rural areas. Demand for products and services is often propelled by factors other than pure population growth as the population grows wealthier. Urbanization is a typical result, and the stimulation of socioeconomic development depends on industry as well as post-industrialization in most nations. In fact, urbanization and economic growth go hand in hand, as demonstrated by prior research [60,61].

Undeniably, urbanization has been shown to influence infectious disease epidemiology. Congestion and crowded populations may facilitate the spread of infectious diseases, particularly respiratory syndrome viruses such as influenza, measles, tuberculosis, and coronavirus [[62], [63], [64], [65], [66], [67]]. Rapid urbanization without proper planning may result in overcrowding in cities, resulting in insufficient housing, sanitation, and basic amenities, which may contribute to an increased risk of infectious disease transmission [62,68,69]. Specifically, Goutte et al. [18] revealed that unworthy housing, household size, and overcrowded housing all contributed to the rise in COVID-19 mortality rates.

A study in the U.S. revealed a high correlation between population density and the COVID-19 transmission and mortality rates. COVID-19 transmission and mortality were, on average, higher in densely populated areas [66]. Moreover, a higher percentage of overcrowded housing, uninsured individuals, and income inequality led to an increased risk of COVID-19 outcomes. In a separate study, Chen et al. [65] discovered that, while large and crowded cities with a higher degree of inter- and intra-city mobility face greater challenges in containing the spread of COVID-19, ensuring the adequacy of healthcare facilities and city management capacity can increase the pandemic control's time efficacy and, consequently, the city's resilience against pandemic.

COVID-19 mortality rate, urbanization level with varying income economies

The COVID-19 mortality rate and urbanization rate are summarised in Table 1 for 190 countries with various income economies. In general, these countries were classified as low-income (28 nations), lower-middle-income (50 nations), upper-middle-income (50 nations), high-income (42 nations), and the top 10% richest nations in the world (20 nations).1 Each income category, on average, has a distinct degree of COVID-19 mortality and urbanization level. Low-income economies had a low average COVID-19 mortality rate of 98 deaths per hundred thousand population and the lowest average urbanization level of 38.92%. The average COVID-19 mortality rates in lower-middle income economies were 157.4 deaths per hundred thousand population, with an average urbanization level of 47.69%. Upper-middle income economies had the highest average COVID-19 mortality rate of 229.3 deaths per hundred thousand population and an average urbanization level of 64.71%. Meanwhile, high-income economies have an average COVID-19 mortality rate of 141.4 deaths per hundred thousand population and a 75.34% urbanization level. The top 10% wealthiest economies had the lowest average COVID-19 mortality rate (75.3 deaths per hundred thousand population) and the highest average degree of urbanization (85.19%).

Table 1

COVID-19 mortality rate and urbanization in different groups of economies.

Economies (income)	Mortality rate	Urbanization	Countries/Places
Low	0.6	62.38	North Korea
Low	19.8	13.71	Burundi
Low	39.5	16.63	Niger
Low	47.0	23.52	Chad
Low	47.3	42.92	Sierra Leone
Low	51.0	30.61	Burkina Faso
Low	51.0	63.33	Eritrea
Low	58.0	42.80	Togo
Low	59.2	45.64	Congo, Democratic Republic
Low	61.7	55.47	Syrian Arab Republic
Low	66.1	20.20	South Sudan
Low	76.1	43.91	Mali
Low	89.8	17.43	Rwanda
Low	90.6	52.09	Liberia
Low	93.5	24.95	Uganda
Low	104.6	21.69	Ethiopia
Low	108.0	37.91	Yemen
Low	108.5	35.25	Sudan
Low	111.0	36.87	Guinea
Low	116.4	44.20	Guinea-Bissau
Low	118.4	42.20	Central African Republic
Low	125.4	38.53	Madagascar
Low	139.2	37.07	Mozambique
Low	144.9	62.58	Gambia
Low	158.9	17.43	Malawi
Low	172.1	46.14	Somalia
Low	211.9	26.03	Afghanistan
Low	274.4	88.28	Venezuela
Average	98.0	38.92
Lower middle	7.4	42.32	Bhutan
Lower middle	9.4	25.52	Vanuatu
Lower middle	29.5	18.71	Sri Lanka
Lower middle	33.6	48.41	Benin
Lower middle	34.7	36.29	Lao PDR
Lower middle	37.0	37.34	Vietnam
Lower middle	37.8	51.96	Nigeria
Lower middle	51.1	31.32	Timor-Leste
Lower middle	52.5	24.23	Cambodia
Lower middle	58.3	57.35	Ghana
Lower middle	61.1	74.35	São Tomé and Principe
Lower middle	63.4	51.71	Côte d'Ivoire
Lower middle	64.5	66.65	Cabo Verde
Lower middle	65.1	73.73	Algeria
Lower middle	70.3	57.56	Cameroon
Lower middle	82.0	13.35	Papua New Guinea
Lower middle	82.9	47.41	Philippines
Lower middle	90.9	68.66	Mongolia
Lower middle	100.5	55.33	Mauritania
Lower middle	100.9	31.14	Myanmar
Lower middle	107.3	66.83	Angola
Lower middle	112.3	50.42	Uzbekistan
Lower middle	117.2	29.38	Comoros
Lower middle	117.2	67.83	Congo, Republic
Lower middle	118.3	57.09	Haiti
Lower middle	126.5	46.02	Belize
Lower middle	127.1	35.23	Tanzania
Lower middle	130.1	48.12	Senegal
Lower middle	133.3	42.78	Egypt
Lower middle	134.7	38.18	Bangladesh
Lower middle	140.7	56.64	Indonesia
Lower middle	152.2	34.93	India
Lower middle	152.6	37.17	Pakistan
Lower middle	160.5	75.87	Iran
Lower middle	160.6	27.51	Tajikistan
Lower middle	164.3	78.06	Djibouti
Lower middle	180.2	36.86	Kyrgyz Republic
Lower middle	181.2	27.99	Kenya
Lower middle	202.5	20.58	Nepal
Lower middle	221.7	69.61	Ukraine
Lower middle	226.8	73.44	El Salvador
Lower middle	228.2	44.63	Zambia
Lower middle	228.3	63.53	Morocco
Lower middle	274.4	59.01	Nicaragua
Lower middle	283.6	32.24	Zimbabwe
Lower middle	297.1	58.36	Honduras
Lower middle	324.2	69.57	Tunisia
Lower middle	562.9	29.03	Lesotho
Lower middle	634.9	24.17	Eswatini
Lower middle	734.9	70.12	Bolivia
Average	157.4	47.69
Upper middle	0.6	61.43	China
Upper middle	24.9	51.43	Thailand
Upper middle	28.2	40.76	Mauritius
Upper middle	30.8	40.67	Maldives
Upper middle	70.0	80.77	Costa Rica
Upper middle	81.4	77.16	Malaysia
Upper middle	87.1	57.25	Fiji
Upper middle	99.9	73.10	Equatorial Guinea
Upper middle	100.7	82.54	Dominican Republic
Upper middle	110.6	53.03	St. Vincent and the Grenadines
Upper middle	114.4	91.42	Jordan
Upper middle	118.6	76.11	Turkey
Upper middle	119.0	90.09	Gabon
Upper middle	120.6	71.09	Dominica
Upper middle	124.1	56.31	Jamaica
Upper middle	126.8	77.19	Cuba
Upper middle	131.5	68.41	Panama
Upper middle	135.7	36.54	Grenada
Upper middle	141.0	57.67	Kazakhstan
Upper middle	152.3	18.84	St. Lucia
Upper middle	159.2	62.18	Paraguay
Upper middle	160.6	52.52	Turkmenistan
Upper middle	166.8	92.11	Argentina
Upper middle	175.7	51.84	Guatemala
Upper middle	186.9	87.07	Brazil
Upper middle	196.0	26.79	Guyana
Upper middle	201.2	81.42	Colombia
Upper middle	228.9	56.45	Serbia
Upper middle	233.1	59.45	Georgia
Upper middle	237.2	66.15	Suriname
Upper middle	245.2	42.85	Moldova
Upper middle	271.5	56.40	Azerbaijan
Upper middle	280.1	70.89	Iraq
Upper middle	292.1	80.69	Libya
Upper middle	293.2	67.35	South Africa
Upper middle	325.1	80.73	Mexico
Upper middle	328.7	54.19	Romania
Upper middle	333.4	64.17	Ecuador
Upper middle	333.9	49.02	Bosnia and Herzegovina
Upper middle	346.5	62.11	Albania
Upper middle	357.0	67.49	Montenegro
Upper middle	364.4	63.31	Armenia
Upper middle	374.6	74.75	Russian Federation
Upper middle	395.6	52.03	Namibia
Upper middle	399.5	70.88	Botswana
Upper middle	416.2	88.92	Lebanon
Upper middle	483.1	79.48	Belarus
Upper middle	528.6	78.30	Peru*
Upper middle	583.6	58.48	North Macedonia
Upper middle	647.3	75.69	Bulgaria
Average	229.3	64.71
High	0.5	86.70	New Zealand
High	4.4	81.41	South Korea
High	21.5	87.28	Greenland
High	26.1	24.43	Antigua and Barbuda
High	28.6	30.84	St. Kitts and Nevis
High	32.2	66.82	Cyprus
High	41.0	91.80	Northern Mariana Islands
High	41.1	100.00	Kuwait
High	44.1	91.78	Japan
High	46.4	84.29	Saudi Arabia
High	51.0	92.59	Israel
High	60.0	57.55	Seychelles
High	60.5	81.56	Canada
High	74.4	100.00	Monaco
High	89.9	94.74	Malta
High	98.0	93.58	Puerto Rico
High	100.7	31.19	Barbados
High	106.6	94.94	Guam
High	108.2	87.73	Chile
High	123.9	100.00	Bermuda
High	124.4	80.98	France
High	126.8	83.90	United Kingdom
High	127.1	79.72	Greece
High	134.7	89.51	Bahrain
High	141.5	86.28	Oman
High	147.6	83.24	Bahamas
High	155.4	95.51	Uruguay
High	179.9	55.12	Slovenia
High	186.7	80.81	Spain
High	198.8	53.21	Trinidad and Tobago
High	202.2	66.31	Portugal
High	205.5	87.92	Andorra
High	226.7	69.23	Estonia
High	227.4	71.04	Italy
High	244.8	74.06	Czech Republic
High	250.4	53.76	Slovak Republic
High	280.3	48.51	British Virgin Islands
High	285.6	57.55	Croatia
High	297.2	60.04	Poland
High	297.8	71.94	Hungary
High	352.0	68.32	Latvia
High	385.0	68.05	Lithuania
Average	141.4	75.34
Top 10%	1.8	78.90	Taiwan, China
Top 10%	4.7	86.24	Australia
Top 10%	5.6	93.90	Iceland
Top 10%	7.2	82.97	Norway
Top 10%	7.4	100.00	Singapore
Top 10%	12.5	63.65	Ireland
Top 10%	15.8	78.25	Brunei Darussalam
Top 10%	27.0	99.24	Qatar
Top 10%	80.8	85.52	Finland
Top 10%	89.2	91.45	Luxembourg
Top 10%	91.2	87.98	Sweden
Top 10%	91.3	87.05	United Arab Emirates
Top 10%	93.1	73.92	Switzerland
Top 10%	94.1	88.12	Denmark
Top 10%	107.5	58.75	Austria
Top 10%	120.5	77.45	Germany
Top 10%	140.0	92.24	Netherlands
Top 10%	146.6	98.08	Belgium
Top 10%	179.3	82.66	United States
Top 10%	189.6	97.50	San Marino
Average	75.3	85.19

Preliminary analysis of average mortality rates and urbanization levels showed that fast urbanization without adequate planning in less developed and developing countries contributed to increases in the COVID-19 mortality rates compared to other countries in the same economic group. For example, in Gambia, Somalia, and Venezuela in the low-income economies; Ukraine, El Salvador, Morocco, Nicaragua, Honduras, Tunisia, and Bolivia in the lower-middle-income economies; Mexico, Russian Federation, Botswana, Lebanon, Belarus, Peru, and Bulgaria in the upper-middle-income economies.

Methodology

The number of COVID-19 mortalities is a non-negative count response variable with a positive skew and a zero-count possibility. As a result, the conventional least squares technique is ineffective for modelling COVID-19 mortalities. Under these circumstances, the Poisson regression technique is more appropriate to model this type of data. However, a disadvantage of Poisson regression analysis is the assumption of equal-dispersion, which limits the mean of the dependant variable to an equal variance [70]. If over-dispersion was not considered, the standard error would be smaller than usual.

It is commonly known that the negative binomial regression technique is more suitable for handling over-dispersion. This is because the negative binomial regression model permits over-dispersion of the dependent variable. The log-likelihood ratio (LR) test was employed in this study to determine if Poisson or negative binomial regression models were more suitable. In essence, the LR test detects the existence of over-dispersion. The null hypothesis specifies the mean and variance to be equal, and, therefore, rejection of the null hypothesis suggests the negative binomial regression technique is more appropriate. The estimated results shown in Table 6 indicated that the more appropriate technique for all models was the negative binomial regression. The negative binomial regression model used can be expressed as,where λ denotes the expected number of COVID-19 mortalities, POP is the population (in hundred thousand), GDP is the per-head GDP, URB represents urbanization levels, AGE64 is the percentage of the population aged above 64, POLITY2 represents democracy level, GE represents government effectiveness, GINI represents income inequality, LEB denotes life expectancy at birth, β is the model coefficient, α is the model intercepts, and ε denotes the error term. Several explanatory variables were transformed to the natural logarithm to reduce heteroskedasticity and improve the model's fitness. These variables can be interpreted as the relative elasticity values of the estimated coefficients.

Table 6

Kuznets relationship for COVID-19 mortalities.

DV: COVID-19 mortalitiesIV:	Model
	A	B	C	D	E	F
ln(GDP)	3.518 (4.28) ***	3.956 (5.21) ***	2.425 (2.57) ***	2.198 (2.29) **	2.365 (2.48) **	2.396 (2.52) **
[ln(GDP)]2	−.192 (−4.24) ***	−.239 (−5.68) ***	−.149 (−2.80) ***	−.109 (−2.02) **	−.135 (−2.58) ***	−.137 (−2.63) ***
URB		.008 (1.91) *	.009 (1.99) **			.017 (1.24)
AGE64		.064 (4.59) ***	.064 (4.43) ***		.061 (3.45) ***	.059 (3.36) ***
GE			−.782 (−2.85) ***	−.859 (−3.21) ***	26.541 (2.31) **	26.725 (2.41) **
VACCINE				−.008 (−1.88) *	−.009 (−2.08) **	−.009 (−2.09) **
POLITY2				.038 (3.35) ***	.019 (1.51)	.018 (1.44)
GINI					−.009 (−0.81)	.018 (0.71)
GINI* URB					.0002 (1.92) *	−.0002 (−0.61)
LEB					−.005 (−0.27)	−.003 (−0.15)
GE*LEB					−.334 (−2.38) **	−.336 (−2.51) **
Constant	−7.766 (−2.11) ***	−8.929 (−2.61) ***	−2.501 (−0.60)	−2.626 (−0.62)	−2.031 (−0.51)	−3.422 (−0.82)
Number of nations	137	137	137	137	137	137
Probability of LR Chi-squared	All models' likelihood ratio test is significant at 1%
Pseudo R-squared	0.0048	0.0111	0.0136	0.0114	0.0181	0.0185
Turning point (USD)	9522	3929	3420	23,922	6369	6276

***Significant at 0.01, **Significant at 0.05, *Significant at 0.1 (z-statistics are shown in parentheses).

The ln(POP) variable was utilized as an offset variable2 to normalize the population exposure effect on the number of COVID-19 mortalities. We presumed that the number of COVID-19 mortalities would be higher in a country with a larger population. equation (1) can be re-expressed as

In this way, the explanatory variable coefficient can be defined as the impact on the rate. This is identical to applying the COVID-19 mortalities per hundred thousand population as the dependent variable, which permits the model error structure to follow the negative binomial distribution.

Data sources and variables

Numerous data sources were compiled to establish a country-level dataset for this study. The data sample is cross-sectional in nature and includes 137 countries.3 The countries included in the analysis are shown in Table 2, along with the number of COVID-19 mortality rates, per-head GDP, and other explanatory variables. The dependent variable, the COVID-19 mortality rate, was retrieved from a study that estimated the excess mortality caused by the COVID-19 pandemic in 191 countries and territories, between January 1, 2020, and December 31,20214 [2]. All explanatory variables' data were obtained from various data sources published in 2019 or 2020. The population statistics came from the World Bank's database of World Development Indicators (WDI). It was used as the risk exposure variable for the COVID-19 mortality.

Table 2

List of countries.

No.	Country	Mortality rate	GDP	URB	POLITY	GE	GINI	AGE64
1	Afghanistan	211.9	1978.96	26	−1	−1.46	31	2.65
2	Albania	346.5	13295.4	62	9	−0.06	33.17	14.70
3	Algeria	65.1	10681.7	74	2	−0.52	27.62	6.74
4	Argentina	166.8	19686.5	92	9	−0.09	40	11.37
5	Armenia	364.4	12592.6	63	7	−0.07	34.43	11.80
6	Austria	107.5	51935.6	59	10	1.49	27.5	19.20
7	Azerbaijan	271.5	13699.7	56	−7	−0.14	26.55	6.74
8	Bahrain	134.7	40933.4	90	−10	0.3	59.6	2.65
9	Bangladesh	134.7	4818.09	38	−6	−0.74	32.39	5.23
10	Belarus	483.1	19148.2	79	−7	−0.18	25.24	15.58
11	Belgium	146.6	48210	98	8	1.03	25.1	19.25
12	Benin	33.6	3323.14	48	7	−0.44	47.76	3.28
13	Bhutan	7.4	10909.1	42	7	0.31	37.44	6.20
14	Bolivia	734.9	7931.75	70	7	−0.7	39.3	7.49
15	Botswana	399.5	16040	71	8	0.43	53.33	4.51
16	Brazil	186.9	14064	87	8	−0.19	53	9.59
17	Bulgaria	647.3	22383.8	76	9	0.34	49.02	21.47
18	Burkina Faso	51.0	2160.51	31	6	−0.76	35.3	2.41
19	Burundi	19.8	4544.02	14	8	−1.15	30.7	6.24
20	Cabo Verde	64.5	6045.06	67	10	0.29	42.38	4.79
21	Cambodia	52.5	4191.85	24	−4	−0.58	37.9	4.85
22	Canada	60.5	45856.6	82	10	1.73	42.7	18.10
23	Chad	47.0	1519.91	24	−2	−1.57	43.32	2.50
24	Chile	108.2	23324.5	88	10	1.06	44.44	12.24
25	China	0.6	16410.8	61	−7	0.52	46.5	11.97
26	Colombia	201.2	13441.5	81	7	0.07	51	9.06
27	Costa Rica	70.0	19679.3	81	10	0.42	53.2	10.25
28	Croatia	285.6	26465.1	58	9	0.41	29.2	21.25
29	Cyprus	32.2	37655.2	67	10	0.99	31.1	14.41
30	Czech Republic	244.8	38319.3	74	9	0.89	24	20.14
31	Denmark	94.1	55938.2	88	10	1.94	27.5	20.16
32	Djibouti	164.3	5481.11	78	3	−0.71	41.59	4.71
33	Dominican Republic	100.7	17003	83	7	−0.36	43.7	7.53
34	Ecuador	333.4	10329.2	64	5	−0.4	44.1	7.59
35	Egypt	133.3	11951.4	43	−4	−0.42	31.53	5.33
36	El Salvador	226.8	8056.54	73	8	−0.47	39	8.65
37	Estonia	226.7	35638.4	69	9	1.17	30.5	20.37
38	Ethiopia	104.6	2296.83	22	1	−0.63	34.99	3.54
39	Fiji	87.1	10997.5	57	2	0.2	36.7	5.82
40	Finland	80.8	47260.8	86	10	1.93	26.2	22.55
41	France	124.4	42025.6	81	10	1.38	29.2	20.75
42	Gabon	119.0	14399.9	90	9	−0.9	38.02	3.53
43	Gambia	144.9	2159.44	63	4	−0.63	35.92	2.53
44	Germany	120.5	50922.4	77	10	1.59	29.7	21.69
45	Ghana	58.3	5304.98	57	8	−0.21	43.52	3.14
46	Greece	127.1	27287.1	80	10	0.41	31	22.28
47	Guatemala	175.7	8393.28	52	8	−0.68	48.28	5.04
48	Guinea	111.0	2670.82	37	4	−0.78	33.73	2.95
49	Guinea-Bissau	116.4	1847.47	44	6	−1.51	35	2.89
50	Guyana	196.0	18680	27	7	−0.39	35	7.00
51	Haiti	118.3	2773.08	57	1	−2.02	41.1	5.17
52	Honduras	297.1	5138.39	58	7	−0.61	43.8	4.97
53	Hungary	297.8	31007.8	72	10	0.5	28	20.16
54	India	152.2	6118.36	35	9	0.17	35.9	6.57
55	Indonesia	140.7	11445	57	9	0.18	40.05	6.26
56	Iran, Islamic Rep.	160.5	12433.3	76	−7	−0.55	40.8	6.56
57	Iraq	280.1	9255.26	71	6	−1.34	48.99	3.44
58	Ireland	12.5	89689	64	10	1.28	28.3	14.58
59	Israel	51.0	38341.3	93	6	1.33	44.4	12.41
60	Italy	227.4	38992.1	71	10	0.46	32.8	23.30
61	Jamaica	124.1	8741.55	56	9	0.5	37.48	9.08
62	Japan	44.1	41380.1	92	10	1.59	50.4	28.40
63	Jordan	114.4	9816.55	91	−3	0.1	30.82	3.95
64	Kazakhstan	141.0	25337.2	58	−6	0.12	27.45	7.90
65	Kenya	181.2	4220.44	28	9	−0.38	40.78	2.51
66	Korea, Republic	4.4	42251.4	81	8	1.38	40.2	15.79
67	Kuwait	41.1	49853.7	100	−7	0.02	36.02	3.04
68	Kyrgyz Republic	180.2	4706.57	37	7	−0.68	27.69	4.73
69	Lao PDR	34.7	7805.8	36	−7	−0.78	36.39	4.26
70	Latvia	352.0	29932.5	68	8	1.11	35.2	20.69
71	Lebanon	416.2	11649	89	6	−0.83	31.83	7.55
72	Lesotho	562.9	2279.9	29	8	−0.83	44.88	4.95
73	Liberia	90.6	1353.84	52	7	−1.38	35.27	3.32
74	Libya	292.1	10282.3	81	0	−1.92	30.2	4.53
75	Lithuania	385.0	36732	68	10	1.04	35.4	20.62
76	Luxembourg	89.2	110261	91	10	1.73	32.3	14.39
77	Madagascar	125.4	1510.14	39	4	−1.14	42.65	3.10
78	Malawi	158.9	1486.78	17	6	−0.75	44.69	2.64
79	Malaysia	81.4	26435.2	77	7	1	41.04	7.18
80	Mali	76.1	2216.77	44	5	−1.06	33.04	2.48
81	Mauritania	100.5	4983.22	55	−2	−0.5	32.62	3.18
82	Mauritius	28.2	19469.5	41	10	0.87	36.76	12.52
83	Mexico	325.1	17887.8	81	8	−0.16	45.38	7.62
84	Moldova	245.2	12324.7	43	9	−0.38	25.66	12.49
85	Mongolia	90.9	11470.7	69	10	−0.19	32.74	4.31
86	Montenegro	357.0	18278.7	67	9	0.16	34.1	15.77
87	Morocco	228.3	6916.35	64	−4	−0.12	39.55	7.61
88	Mozambique	139.2	1229.08	37	5	−0.82	51.1	2.86
89	Myanmar	100.9	8893.81	52	6	0.1	59.07	3.59
90	Namibia	395.6	3800.07	21	7	−1.05	32.84	5.83
91	Nepal	202.5	54209.6	92	10	1.8	27.5	20.03
92	Netherlands	140.0	5280.14	59	6	−0.77	46.16	5.68
93	Nicaragua	274.4	1196.88	17	5	−0.8	34.28	2.60
94	Niger	39.5	4916.72	52	3	−1.09	35.13	2.74
95	Nigeria	37.8	15848.4	58	9	0	30.7	14.48
96	North Macedonia	583.6	63585.9	83	10	1.86	25.4	17.53
97	Norway	7.2	27294.6	86	−8	0.26	30.7	2.51
98	Oman	141.5	4622.77	37	7	−0.68	33.45	4.35
99	Pakistan	152.6	25381.8	68	9	0.07	46.5	8.54
100	Panama	131.5	12335.5	62	9	−0.53	43.6	6.81
101	Paraguay	159.2	11260.8	78	9	−0.07	39.5	8.73
102	Peru	528.6	7953.58	47	8	0.05	40.11	5.51
103	Philippines	82.9	32238.2	60	10	0.6	28.5	18.74
104	Poland	297.2	32181.2	66	10	1.15	31.9	22.77
105	Portugal	202.2	28832.6	54	9	−0.28	34.8	19.23
106	Romania	328.7	26456.4	75	4	0.15	37.47	15.51
107	Russian Federation	374.6	11470.7	69	10	−0.19	32.74	4.31
108	Rwanda	89.8	2098.71	17	−3	0.19	43.71	3.12
109	Saudi Arabia	46.4	44328.2	84	−10	0.31	42.2	3.50
110	Senegal	130.1	3300.09	48	7	−0.06	40.29	3.11
111	Serbia	228.9	18210	56	8	0.02	33.3	19.06
112	Sierra Leone	47.3	1648.05	43	7	−1.13	35.69	2.93
113	Singapore	7.4	93397	100	−2	2.22	47.3	13.35
114	Slovak Republic	250.4	30330	54	10	0.67	22.8	16.70
115	Slovenia	179.9	36547.7	55	10	1.08	23.9	20.74
116	Somalia	172.1	829.611	46	6	−2.24	36.31	2.90
117	South Africa	293.2	11466.2	67	9	0.37	61.64	5.51
118	Spain	186.7	36215.4	81	10	1	33	19.98
119	Sri Lanka	29.5	12536.9	19	6	−0.11	39.81	11.23
120	Sudan	108.5	4022.87	35	−4	−1.62	34.24	3.67
121	Sweden	91.2	50683.3	88	10	1.83	27.6	20.33
122	Switzerland	93.1	68393.3	74	10	1.95	30.6	19.10
123	Tajikistan	160.6	3657.57	28	−3	−1.05	34	3.18
124	Tanzania	127.1	2635.34	35	3	−0.88	40.49	2.64
125	Togo	58.0	2107.88	43	−2	−0.92	43.06	2.91
126	Trinidad and Tobago	198.8	23728.2	53	10	0.1	51.4	11.51
127	Tunisia	324.2	9727.5	70	7	−0.1	32.82	8.87
128	Turkey	118.6	28385	76	−4	0.05	41.7	8.98
129	Uganda	93.5	2177.6	25	−1	−0.59	42.75	1.99
130	Ukraine	221.7	12377	70	4	−0.3	26.07	16.95
131	United Kingdom	126.8	41627.1	84	8	1.44	35.08	18.65
132	United States	179.3	60235.7	83	5	1.49	48.4	16.63
133	Uruguay	155.4	21608.4	96	10	0.7	39.4	15.09
134	Uzbekistan	112.3	6994.17	50	−9	−0.51	39.7	4.79
135	Vietnam	37.0	8200.33	37	−7	0.04	35.72	7.87
136	Zambia	228.2	3270.04	45	6	−0.68	57.14	2.13
137	Zimbabwe	283.6	2744.69	32	4	−1.21	44.34	3.01

The government effectiveness (GE) index was used as a proxy for the regulatory governance process in terms of consistency, accountability, and openness. This index is composed of five components: discernment of the quality of public services; evaluation of the quality of the civil service and its extension of independence from political pressures; the quality of policy formulation and implementation; and the credibility of the government's commitment to such policies. This index is calculated on a scale of −2.5 to 2.5, with −2.5 revealing the least effective government and 2.5 revealing the most effective government. The GE index was classified in this study as −2.5 to 1.2 (labelled as 0, indicating low government effectiveness) and 1.2 to 2.5 (labelled as 1, which indicates high government effectiveness). A more effective administration is anticipated to be more competent and efficient in enacting health measures to combat the COVID-19 epidemic. The data for this variable was drawn from the Worldwide Governance Indicators (WGI).

We utilized the POLITY2 democracy index to assess if a country is authoritarian or democratic. This variable's data comes from the Polity V database [71]. This index runs from −10 (extremely autocratic) to +10 (extremely democratic). We believe that a more democratic country is more susceptible to disease infection because of its increased exposure to globalisation. Due to the interconnectivity of the world, an infected person travelling from one country to another could initiate the spread of the virus in a population. Thus, a more open country is at a higher risk of its population getting infected compared to a less democratic country especially when new variants appear in any part of the world.

The COVID-19 vaccines have been shown in several previous studies to help reduce the transmission of the virus and, in the long run, halt it by establishing herd immunity and lowering mortality risks [[72], [73], [74], [75], [76], [77]]. Data on the number of people fully vaccinated per hundred population from the Our World in Data5 database was used to investigate the correlation between the COVID-19 vaccination rates and mortality rates.

The GINI index was employed in this research to quantify the extent of inequality in the distribution of wealth in a country. This index is a numeric value between 0 and 100. A GINI index of 0 implies complete equality when everyone owns the exact same amount of money, whereas a GINI index of 100 indicates perfect inequality when one individual owns all the nation's income. The indices were collected from the Standardized World Income Inequality Database (SWIID). Earlier health research studies did incorporate this variable to account for the effect of income inequality [[78], [79], [80], [81]].

The remaining four variables were derived from the WDI database. These indicators include per-head GDP, the percentage of urbanization (URB), the life expectancy at birth (LEB), and the proportion of the population over the age of 64 (AGE64). Per-head GDP was employed as a proxy for economic performance. It was described as “purchasing power parity-converted gross domestic product per capita at constant prices in 2017". By dividing the entire urban population by the total population, the URB was calculated. This variable was used to determine the degree of population concentration. Urbanization is expected to increase human interaction, which may result in an increase in COVID-19 transmission.

Numerous earlier investigations have shown that the severity of numerous infectious diseases is higher in older people than in younger ones [[82], [83], [84]]. Considering this, the AGE64 was used to evaluate the likelihood that elderly populations would be exposed to COVID-19. Life expectancy at birth (LEB)6 was used as a proxy to determine the level of healthcare systems. Ho and Hendi [85] explained that a country's life expectancy mirrors the social, economic, and quality of public health and healthcare infrastructure. Table 3, Table 4, Table 5 summarise and describe all the data utilized in the study.

Table 3

Description of factors and data source.

Variable	Definition	Source
MORTALITY	COVID-19 cumulative excess mortalities (up to December 31, 2021)	COVID-19 Excess Mortality Collaborators
POP	Population (in hundred thousand people)	WDI
GDP	Purchasing power parity converted gross domestic products per capita at 2017 constant prices	WDI
URB	Percentage of urban to total population	WDI
AGE64	Percentage of the population over the age of 64	WDI
LEB	Life expectancy at birth	WDI
VACCINE	People fully vaccinated per hundred population	Our World in Data
GE	Government effectiveness, −2.5 (lowest effectiveness) and +2.5 (highest effectiveness)	WGI
GINI	Income inequality	SWIID
POLITY2	Democracy index, −10 (strongly autocratic) and +10 (strongly democratic)	Polity V Database

Table 4

Descriptive statistics.

Variable	Obs.	Mean	Std. Dev.	Min	Max	85th percentile
MORTALITY	137	127895.50	386723.90	108	4070000	180200
POP	137	52.94	172.91	0.56	1400	60
GDP	137	19754.80	20057.24	731.06	110261.2	40156.9
URB	137	60.52	21.74	14	100	83.6
AGE64	137	9.66	6.79	1.98	28.39	18.22
LEB	137	72.82	7.36	54.2	84.35	81.11
GE	137	0.006	0.95	−2.24	2.22	1.11
VACCINE	137	43.82	26.26	0.03	87.09	73.36
GINI	137	37.56	8.08	22.8	61.64	45.18
POLITY2	137	5.01	5.63	−10	10	10

Table 5

Correlation matrix.

Variable	GDP	URB	POLITY2	GE	VACCINE	GINI	LEB	AGE64
GDP	1.0000
URB	0.618	1.0000
POLITY2	0.227	0.181	1.0000
GE	0.824a	0.583	0.356	1.0000
VACCINE	0.672	0.589	0.169	0.776	1.0000
GINI	−0.229	−0.008	−0.061	−0.156	0.121	1.0000
LEB	0.724	0.676	0.243	0.784	0.109	−0.238	1.0000
AGE64	0.677	0.507	0.482	0.761	0.098	−0.3701	0.756	1.0000

Correlation value greater than 0.8.

Results and discussions

The descriptive statistics for all explanatory variables and the correlations between them are shown in Table 4, Table 5, respectively. The correlation between the variables indicated that there is a strong correlation between government effectiveness and per-head GDP (0.824). However, the estimated coefficients for these variables in Models C and D were significant (Table 6), indicating that multicollinearity is not present. The estimated results in Table 6 indicate that there is a significant reverse U-shaped relationship between the COVID-19 mortality rate and per-head GDP for all the models (the coefficient for per-head GDP is positive, whereas the coefficient for the square term of per-head GDP is negative). This implies that a rise in per-head GDP is, at the beginning, accompanied by a rise in the COVID-19 mortality rate, but that this rate decreases once the per-head GDP passes a specific income value (turning point). Fig. 2 illustrates the reverse U-shaped relationship of Model F with the turning point for the COVID-19 mortality rate estimated at the ln (GDP) value of 8.74 (US $6276). This was in line with the Kuznets curve hypothesis for infectious disease that is explained in Section 2. This is a novelty finding of this study as previous research have yet to examine this reverse U-shaped relationship.

Fig. 2

The effect of per-head GDP on the COVID-19 mortality rate (Graph plotted using Model F in Table 6; all explanatory variables, except per-head GDP, are substituted with the average value of each variable).

A separate analysis was conducted on two samples categorized by income inequality levels7 with the purpose of examining the effect of income inequality on the relationship between economic performance and the COVID-19 mortality rate. The estimated results in Table 7 suggest that the mortality rates for the country groups with lower income inequality start to decrease at a lower per-head GDP. As illustrated in Fig. 3 , the turning point for countries with lower income inequality is US $8813 compared to US $9685 for countries with higher income inequality. These results indicate that a higher per-head GDP is required to reverse the rising trend of the mortality rate under conditions of higher income inequality. This can be attributed to the fact that poorer people are less capable of accessing better quality food, water, and health care services. Therefore, they are more exposed to health risks, including infectious diseases. The existing literature investigating the effect of income inequality on health outcomes consistently points out that health is worsening in more unequal societies [[78], [79], [80], [81],[86], [87], [88]]. These results explained that this is partly due to sustained poverty conditions among most of the population, which lead to a higher number of materially deprived individuals.

Table 7

Estimates of COVID-19 mortality rate at lower and higher income inequality levels.

DV: COVID-19 mortalitiesIV:	Model
	A	A1 (GINI ≤32)	A2 (GINI >32)
ln(GDP)	3.518 (4.28) ***	6.268 (3.28) ***	2.882 (2.87) ***
[ln(GDP)]2	−.192 (−4.24) ***	−0.345 (−3.44) ***	−0.157 (−2.79) ***
Constant	−7.766 (−2.11) ***	−19.825 (−2.21) **	−5.023 (−1.13)
Number of nations	137	36	101
Probability of LR Chi-squared	All models' likelihood ratio test is significant at 1%
Pseudo R-squared	0.0048	0.0174	0.0186
Turning point (USD)	9522	8813	9685

***Significant at 0.01, **Significant at 0.05, *Significant at 0.1 (z-statistics are shown in parentheses).

Fig. 3

The effect of per-head GDP on the COVID-19 mortality rate under different income inequality levels (The graphs were plotted using Models A1 and A2 in Table 7).

Models B and C in Table 6 show that urbanization level is positively related to the COVID-19 mortality rates, suggesting that a higher level of urbanization would raise the COVID-19 mortality rates. Several previous studies have shown that the migration of the population from rural areas to urban areas has significant implications for public health, such as infectious diseases like malaria, influenza, and HIV [62,[89], [90], [91]]. This is because urbanization could increase the infectious disease transmission rate through increases in human contact and changes in socioeconomic conditions. This result can be illustrated in Fig. 4 .

Fig. 4

The effect of urbanization on the COVID-19 mortality rate (The graph was plotted using Model B in Table 6; all explanatory variables, except urbanization, are substituted with the average value of each variable.).

The results in Models E and F in Table 6 show that the coefficient for GINI was statistically insignificant, but the coefficient for its interaction with urbanization was significantly positive in Model E. This estimated result indicated that the positive impact of income inequality on the COVID-19 mortality rate was larger with an increase in urbanization. The intuitive explanation for this result is that the poorest population living in urban slums is more vulnerable to infection because physical separation is impossible in cramped living conditions. To our knowledge, this is the first study in the literature that investigates the interaction effect between income inequality and urbanization on infectious disease mortalities.

Two political-related variables, government effectiveness and democracy level, were used to examine the effect of the quality of political institutions on the COVID-19 mortality rate. The coefficients for government effectiveness in Models C and D were significantly negative,8 indicating that a higher level of government effectiveness would reduce the COVID-19 mortality rate. This is in line with the notion that a higher quality of government machinery is more capable of implementing effective health measures and policies, which in turn reduces disease transmission rates and mortalities. In contrast, democracy level had a significant positive effect on the COVID-19 mortality rate, which was inconsistent with a previous study [92].9 This result indicated that a higher democracy level would raise the COVID-19 mortality rate. One possible explanation is that democratic nations are more open to the world, particularly in terms of economic activities, making them more vulnerable to a pandemic.

The life expectancy at birth was used as a proxy for the quality of the healthcare system in this study. As can be seen in Models E and F, the coefficient for life expectancy at birth was negative but statistically insignificant, whereas the coefficient for its interaction with government effectiveness was significantly negative. This finding suggests that the improvement of the healthcare system has a notable effect on lowering the COVID-19 mortality rate under more effective government conditions. This result is consistent with previous studies that indicated improvements in healthcare service could reduce the number of mortalities, such as traffic mortalities [6,93] and maternal mortalities [94]. However, this study further explores the healthcare effect on the COVID-19 mortality rate under different government effectiveness levels.

Another variable studied was the percentage of the population aged over 64 (AGE64). This variable was employed to examine the vulnerability of the elderly population during the COVID-19 pandemic. The coefficient for AGE64 had a significant positive impact on the COVID-19 mortality rate in Models B, C, E, and F. Previous studies have shown that the mortality rate for elderly patients aged 60 years and above infected with COVID-19 is higher than for patients in other age groups [[95], [96], [97], [98]]. According to models D to F in Table 6, the COVID-19 vaccination rate is negatively related to COVID-19 mortality rates, implying that increasing the COVID-19 vaccination rate reduces the risk of COVID-19 mortality.

Each country's risk factors (the explanatory variables) were assessed based on the estimated results. When the level surpasses the 85th percentile, the country's risk factors are highlighted. Table 4 displays the 85th percentile values for each risk factor. The vaccination rate is highlighted if it falls below the 15th percentile (19.35). The assessment results are listed in Table 8, Table 9, Table 10, Table 11, Table 12 . In general, the prevalence of risk factors is low in low-income countries. Mozambique's income inequality is more than the 85th percentile, yet its urbanization rate is low. As previously explained, greater income inequality will only result in an increase in the COVID-19 mortality rate under conditions of greater urbanization. Nonetheless, most countries in this group have vaccination rates below the 15th percentile.

Table 8

Risk Factor Assessment for low-income countries.

Country	Polity 2	GE	LEB	URB	GINI	AGE64	VACCINE+
Burundi							*
Chad							*
Guinea-Bissau							*
Mali							*
Madagascar							*
Sudan							*
Burkina Faso							*
Malawi							*
Uganda							*
Niger							*
Sierra Leone							*
Somalia							*
Guinea							*
Ethiopia							*
Afghanistan							*
Gambia							*
Togo							*
Liberia							*
Mozambique					*		*
Lesotho
Rwanda

*15th percentile of vaccination rate was used to identify the level of critical risk.

Table 9

Risk Factor Assessment for low-middle income countries.

Country	Polity 2	GE	LEB	URB	GINI	AGE64	VACCINE+
Haiti							*
Nigeria							*
Tanzania							*
Senegal							*
Zambia					*		*
Ghana							*
Kenya							*
Djibouti							*
Benin					*		*
Algeria							*
Kyrgyzstan							*
Mauritania							*
Zimbabwe							*
Egypt
Burma
Ukraine
Pakistan
Tajikistan
Bangladesh
Uzbekistan
Nepal
Bolivia
Indonesia
India
Honduras
Nicaragua					*
Philippines
Cabo Verde	*
Tunisia
Laos
Iran
Morocco
Sri Lanka
El Salvador
Mongolia	*
Fiji
Vietnam
Bhutan
Cambodia

*15th percentile of vaccination rate was used to identify the level of critical risk.

Table 10

Risk Factor Assessment for upper-middle-income countries.

Country	Polity 2	GE	LEB	URB	GINI	AGE64	VACCINE+
Gabon				*			*
Libya							*
Namibia					*		*
Iraq					*		*
Jamaica							*
Armenia
South Africa					*
Lebanon				*
Guatemala					*
Bulgaria					*	*
Guyana
Moldova
Belarus
North Macedonia
Jordan				*
Albania
Romania						*
Paraguay
Botswana					*
Montenegro
Kazakhstan
Serbia
Azerbaijan
Russia
Dominican Republic
Colombia					*
Mexico					*
Turkey
Panama					*
Peru
Brazil				*	*
Costa Rica	*				*
Mauritius	*
Ecuador
Argentina				*
Malaysia
China					*

*15th percentile of vaccination rate was used to identify the level of critical risk.

Table 11

Risk Factor Assessment for high-income countries.

Country	Polity 2	GE	LEB	URB	GINI	AGE64	VACCINE+
Trinidad and Tobago	*				*
Slovakia	*
Croatia						*
Cyprus	*
Poland	*
Slovenia	*		*			*
Oman				*
Czech Republic						*
Hungary	*					*
Estonia		*				*
Greece	*		*			*
Israel		*	*	*
Lithuania	*					*
Saudi Arabia				*
Latvia		*				*
Bahrain				*	*
United Kingdom		*	*	*
France	*	*	*			*
Italy	*		*			*
Kuwait				*
Uruguay	*			*
Japan	*	*	*	*	*	*
Canada	*	*	*
Spain	*		*			*
Korea, South		*	*
Portugal	*	*				*
Chile	*			*

*15th percentile of vaccination rate was used to identify the level of critical risk.

Table 12

Risk Factor Assessment for top 10-percent income countries.

Country	Polity 2	GE	LEB	URB	GINI	AGE64	VACCINE+
United States		*			*
Switzerland	*	*	*
Netherlands	*	*	*	*		*
Austria	*	*	*
Germany	*	*				*
Norway	*	*	*
Luxembourg	*	*	*	*
Belgium			*	*		*
Sweden	*	*	*	*		*
Finland	*	*	*	*		*
Ireland	*	*	*
Denmark	*	*	*	*		*
Singapore		*	*	*	*

*15th percentile of vaccination rate was used to identify the level of critical risk.

As seen in Table 9, the vaccination rate in several low-middle income countries is below the 15th percentile. Among these are Haiti, Nigeria, Tanzania, Senegal, Zambia, Ghana, Kenya, Djibouti, Benin, Algeria, Kyrgyzstan, Mauritania, and Zimbabwe. Only Cabo Verde and Mongolia in this group have the highest value on the democracy index. Consequently, it is likely that higher levels of democracy contributed to the COVID-19 mortality rate in these two countries. The income inequality in Zambia, Benin, and Nicaragua is greater than the 85th percentile, whereas their urbanization rates are below the 85th percentile.

Several upper-middle income countries' vaccination rates fall below the 15th percentile, as shown in Table 10. These countries consist of Gabon, Libya, Namibia, Iraq, and Jamaica. In this country group, urbanization is expected to increase the COVID-19 mortality rate in Gabon, Lebanon, Jordan, Brazil, and Argentina, among others. Brazil has greater rates of urbanization and income inequality among these countries. The highest democracy index scores are in Costa Rica and Mauritius, whereas the percentage of population aged 64 and older in Bulgaria and Romania is higher than the 85th percentile.

All high-income and top 10% high-income countries have vaccination rates above the 15th percentile. Many countries with high incomes have the highest democracy Index score. These countries are Trinidad and Tobago, Slovakia, Cyprus, Poland, Slovenia, Hungary, Greece, Lithuania, France, Italy, Uruguay, Japan, Canada, Spain, Portugal, and Chile, as shown in Table 11. On the other hand, many countries in this group have a higher proportion of population aged 64 and older. These are the countries Croatia, Slovenia, Czech Republic, Hungary, Estonia, Greece, Lithuania, Latvia, France, Italy, Japan, Spain, and Portugal. Japan has the oldest populations among these countries, with approximately 29% of its population aged 64 or older. It was discovered that Oman, Israel, Saudi Arabia, Bahrain, the United Kingdom, Kuwait, Uruguay, Japan, and Chile have urbanization rates above the 85th percentile. Notables are the high urbanization rates and income inequality in Bahrain and Japan. Nine countries, including Estonia, Israel, Latvia, the United Kingdom, France, Japan, Canada, South Korea, and Portugal, have governance effectiveness scores above the 85th percentile. Countries with high life expectancy at birth and governance effectiveness include Israel, the United Kingdom, France, Japan, Canada, and South Korea.

The only top 10-percent income countries with democracy index scores below the 85th percentile, as shown in Table 12, are the United States, Belgium, and Singapore. However, life expectancy at birth is lower than 85th percentile in the United States and Germany. Belgium, on the other hand, is the only country with a governance effectiveness level that is below the 85th percentile. Six countries, including the Netherlands, Germany, Belgium, Sweden, Finland, and Denmark, have a population proportion of 64 and older that exceeds the 85th percentile. Netherland, Luxembourg, Belgium, Sweden, Finland, Denmark, and Singapore are among the countries with urbanization rates in this country group that are higher than the 85th percentile. Both the United States and Singapore have levels of income inequality above the 85th percentile, but only Singapore has an urbanization rate above the 85th percentile.

Conclusions and implications

This study indicated that there are mediated pathways from economic performance to the change in the COVID-19 mortality rate, such as through urbanization, elderly population group, income inequality, the vaccination rate, advancements in healthcare systems and the quality of political institutions. A few major findings are noteworthy. First, the estimated results indicated that the COVID-19 mortality rate is related to per-head GDP in a reverse U-shaped relationship or known as the Kuznets curve. At lower levels of per-head GDP, the COVID-19 mortality rate is anticipated to be lower due to economic activities mainly occurring on a domestic scale with more short-distance travel. However, because of economic and population growth, urban sprawl induces a higher demand for long-distance travel. This, in turn, increases the disease transmission rates. At higher per-head GDP levels, people and the government are more capable of investing in disease abatement measures, fostering a decline in disease transmission rates. The rise and fall of COVID-19 mortality rates with per-head GDP as indicated in this study can thus be regarded as a relationship following the explanation of the Kuznets theory. Second, the results also revealed that a higher per-head GDP is required to reverse the rising trend of the mortality rate under higher income inequality conditions.

As demonstrated by this study, the decline in the COVID-19 mortality rate is contingent on improvements in healthcare services, governance effectiveness, and a higher vaccination rate. Clearly, economic growth alone is insufficient to bring about this outcome. Only by increasing the demand for increased public health safety and making resources available for investment in healthcare services does economic growth facilitate the reduction of the COVID-19 mortalities. Therefore, it may be advantageous in the long run to prioritize efforts to improve the health care system and the quality of political agencies, particularly in low-income countries where the level of healthcare service and the quality of political institutions are generally low and direct policy action is possible. In addition, the results demonstrated that a higher level of democracy would increase COVID-19 mortality rates. Considering these findings, it is also prudent to conduct a review of the standard operating procedures for dealing with international affairs to reduce the risk of global disease transmission.

On the other hand, as shown by the findings, increases in income inequality, particularly under higher levels of urbanization, as well as having a higher percentage of the population that is over the age of 64 can both contribute to an increase in the COVID-19 mortality rate. As a result, there is an immediate need to modernize informal settlements and provide adequate services to accommodate the numerous requirements of senior citizens and communities that are economically disadvantaged. Particularly important is the provision of financial assistance, as well as food, water, and other essentials.

Numerous lower income countries in the dataset used in this study have a per-head GDP below the turning point of the reverse U-curve for the COVID-19 mortality rate. This implies that these countries have yet to reach the stage where further income growth will result in a decline in the COVID-19 mortality rate. However, we cannot presume that these countries will strictly follow a similar pathway. This is largely due to other factors included in the analysis that also influence the COVID-19 mortality rate. Considering this, these factors’ effects ought to be considered when enacting infectious disease policies to minimize the impact of COVID-19 specifically on the vulnerable communities.

It is advised that raising vaccination rates, governance effectiveness, and healthcare system quality in low-, lower-middle-, and upper-middle-income countries would help to lower the COVID-19 mortality rate in light of the risk factor assessment's findings. Greater income inequality exists in many upper-middle-income countries, which is predicted to lead to a higher COVID-19 mortality rate as urbanization rates increase. To prevent COVID-19 from spreading in heavily populated areas, strict measures must be put in place. This recommendation holds true for high- and top-10% income countries, as these countries have a high urbanization rate on average. Since the elderly make up a larger part of the population in these high-income countries, enough medical resources should be set aside for them.

According to earlier studies, the COVID-19 pandemic triggered a global economic recession due to an increase in unemployment, volatility in the stock market, a decline in the efficiency of the energy system, a decrease in tourism expenditures, and lowering oil prices. Contrarily, numerous previous studies have also indicated that economic performance has a significant impact on infectious disease mortality, including the COVID-19 mortality. All these studies unmistakably indicate a possibly significant correlation—possibly in both directions—between economic performance and COVID-19 mortality. Studying the relationship as a one-way relationship would be problematic if the relationship between COVID-19 mortality and economic performance is, in fact, bidirectional. The structural parameter estimations from these studies will be unreliable and inconsistent due to the endogenous association between economic performance and COVID-19 mortality. As a result, future research could investigate the bidirectional causality between COVID-19 mortality and economic performance.

Several limitations of the present study should be acknowledged, which also call for future research. First, the life expectancy at birth, utilized as a proxy for the improvements in healthcare systems, does not directly measure the COVID-19 mortalities. A rise in life expectancy at birth does not totally reflect improvements in healthcare systems. However, notwithstanding this limitation, this variable was used mainly due to the high availability of data and was often used in previous studies as a proxy for the improvements in healthcare systems [85,99]. Second, the results of this research may not be used to generalize the relationship between the mortality rates of other infectious diseases and economic performance. This is because infectious diseases vary significantly in many aspects, including the spread mechanism and mortality risks. Therefore, further work is required to examine the impact of economic performance on other infectious diseases like Ebola, HIV, and influenza. Third, the quadratic function of per-head GDP was employed in the present study to examine the Kuznets hypothesis for the COVID-19 mortality rate. The downside of the quadratic function is that it assumes a symmetrical pattern at both sides of the turning point, indicating that the COVID-19 mortality rates will decrease at a similar rate to which they previously grew. Therefore, it is recommended in future studies to fit the Kuznets curve for the COVID-19 mortality rate using regression methods that are not limited by a particular functional pattern, such as the spline regression analysis technique10 .

Author contribution statement

The author confirm contributions of the manuscript as follows: research conceptualization and design: Law and Ng; data collection: Poi and Ng; data analysis and interpretation: Law and Poi; draft manuscript: Law, Ng, and Poi. All authors revised the manuscript and approved the final version.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.