Politics, policy, and pandemic control: a global perspective

Figures & data

Figure 1. Visualization of the non-linear, time-dependent trajectory of deaths per 100,000 people generated by the GAM estimator. The y-axis represents the standardized dependent variable (deaths per 100,000 people, logged).

Table 1. Random effects panel regression models of total COVID-19 deaths per 100,000 people as a function of per capita GDP, age structure, government responsiveness, and political reach.⁶

Dependent Variable:	(1)	(2)	(3)
Total deaths per 100,000 (logged)	GAM	GAM	RE Panel
GDPPC (logged)	0.286***	0.080***	0.0418
	(0.019)	(0.016)	(0.0590)
pop 80+	0.191***	0.280***	0.327***
	(0.020)	(0.016)	(0.0625)
RPR		−0.035***	−0.0924**
		(0.011)	(0.0424)
responsiveness (t-2)		−0.497***	−0.678***
		(0.013)	(0.0157)
Constant	0.017	0.009	−0.0178
	(0.014)	(0.011)	(0.0422)
Observations	2,355	2,355	2,355
Number of countries	116	116	116
R-squared	0.566	0.732	0.672

Standard errors in parentheses

*** p<0.01, ** p<0.05

Figure 2. Time series of deaths per 100,000 people overlaid with predicted values. The y-axis is on a logarithmic scale.

Figure 3. Predicted marginal effects of elderly populations on mortality rates.

Figure 4. Predicted marginal effects of policy responsiveness on COVID-19 death rates over time.

Figure 5. Mean reported COVID-19 deaths for low and high levels of RPR.

Figure 6. Overlaid time series plots of policy responsiveness and reported COVID-19 deaths per 100,000 people.

Figure 7. Comparison of death tolls from COVID-19 in Greece and Italy.

Figure 8. Comparison of death tolls from COVID-19 in Canada and the United States.

Table A1. Data descriptions and sources.

Variables	Descriptions	Source
Total Deaths per 100,000	Covid-19 total deaths per 100,000 population	Institute for Health Metrics and Evaluation (IHME)
GDP per Capita	Gross domestic product per capita	World Bank Human Development Indicators
Population age 80 +	The percentage of population over 80 and older	World Bank Human Development Indicators
Responsiveness	An index of government responses	The Oxford COVID-19 Government Response Tracker and IHME
RPR	Relative Political Reach	TransResearch Consortium
RPE	Relative Political Extraction	TransResearch Consortium
Week	Weeks of Covid-19 deaths	Institute for Health Metrics and Evaluation (IHME)

Table A2. Slope Difference Tests: Pairwise Comparisons of Marginal effects of Responsiveness on Total COVID-19 deaths per 100,000 people.

Responsiveness at: VS Responsiveness at:		Contrast Simple Slope	Bonferroni-adjusted P-value	95% Confidence Interval
High RPR & Low RPE	High RPR & High RPE	0.74 (0.305)	0.088	−0.60 1.55
Low RPR & High RPE	High RPR & High RPE	0.87 (0.296)	0.019	0.09 1.65
Low RPR & Low RPE	High RPR & High RPE	−0.29 (0.299)	1.000	−1.08 0.50
Low RPR & High RPE	High RPR & Low RPE	0.13 (0.345)	1.000	−0.78 1.04
Low RPR & Low RPE	High RPR & Low RPE	−1.03 (0.346)	0.017	−1.95 −0.12
Low RPR & Low RPE	Low RPR & High RPE	−1.16 (0.319)	0.002	−2.00 −0.32

Standard errors in parentheses

Table A3. Average Marginal Effects of Responsiveness on Total COVID-19 deaths per 100,000.

Responsiveness at:	Simple Slopes	Delta- method Standard Errors	P-value	95% Confidence Interval
High RPR & High RPE	−5.03	0.207	0.000	−5.43 −4.62
High RPR & Low RPE	−4.28	0.260	0.000	4.79 −3.77
Low RPR & High RPE	−4.15	0.239	0.000	−4.62 −3.69
Low RPR & Low RPE	−5.32	0.233	0.000	−5.77 −4.86

Table A4. Summary statistics.

variable	N	mean	std. dev.	min	max	skewness	kurtosis
total deaths per 100k	2462	7.75	14.85	0.00	89.43	2.83	11.16
GDP per capita	2462	18028.69	20995.33	379.37	92119.52	1.41	4.20
population 80+	2462	2.53	2.13	0.12	8.42	0.69	2.34
responsiveness	2387	439.85	1656.40	1.27	30567.48	9.73	127.73
RPR	2462	0.96	0.13	0.58	1.25	−0.42	2.96
RPE	2462	0.99	0.31	0.26	1.76	0.11	2.59
week	2462	11.10	6.41	1.00	27.00	0.19	1.99