Table E.1.

Triple-difference regression on the number of deaths, wave 1 .

	Number of deaths, wave 1
2020 $=$ 1	0.12952***	0.18605***
	(0.03773)	(0.05675)
2020 $=$ 1 ×Red area $=$ 1	0.51121***	0.75283***
	(0.17458)	(0.26704)
2020 $=$ 1 ×Poor $=$ 1	−0.05089	0.19666*
	(0.03267)	(0.11114)
2020 $=$ 1 ×Red area $=$ 1 ×Poor $=$ 1	0.81664*	2.41010**
	(0.41663)	(1.08131)
Municipality FE	✓	✓
Sample	All	Major urban centers
	municipalities	and their outskirts
Control outcome mean	1.736	2.290
Observations	188688	100686

* $p<$0.1, ** $p<$0.05, *** $p<$0.01. Standard errors in parentheses clustered at the département level.

NOTE: This table reports coefficients from a triple-difference regression on the number of deaths in wave 1 (March–April) over the years 2018–2020. The three dimensions are (i) time, with 2020 being the post period; (ii) the level of infection in the département before the first wave; (iii) the poverty status of the municipality, defined as the bottom quartile of the national distribution of municipal income weighted by the municipal population size. We include municipality fixed-effects to account for time-invariant factors at the municipal level, such as the population measured in 2014. The control outcome mean is the mean number of deaths in March–April in the years 2018–2019 in non-poor municipalities in green zones.