Table 9.

Wind and PM 2.5 IV — Regressions over alternative outcomes: Hospitalizations, use of ICUs, and use of ventilators .

	(1)	(2)	(3)	(4)	(5)
	Same-day	Same-day	Same-day	3-day	7-day
Hospitalizations
Weighted PM 2.5	13.18***	10.15***	16.41***	22.31**	22.27
	(3.495)	(3.672)	(4.902)	(10.08)	(25.15)
F Stat	9.65	9.88	9.98	9.60	9.59
Dep Var Mean	1646.36	1647.19	1647.18	4962.06	1.2e＋04
Pct change mean	0.800	0.616	0.996	0.450	0.191
Controls	✓	✓	✓	✓	✓
County & State-by-week FEs	✓			✓	✓
County and MSA-by-week FEs		✓
County and CBSA-by-week FEs			✓
Observations	70,875	70,823	70,801	70,757	70,587
Intensive care units
Weighted PM 2.5	3.260**	3.192**	4.433***	3.892	9.823
	(1.430)	(1.446)	(1.641)	(3.914)	(7.771)
F Stat	13.09	13.14	13.06	13.05	13.03
Dep Var Mean	639.95	640.08	640.22	1925.66	4507.23
Pct change mean	0.509	0.499	0.692	0.202	0.218
Controls	✓	✓	✓	✓	✓
County & State-by-week FEs	✓			✓	✓
County and MSA-by-week FEs		✓
County and CBSA-by-week FEs			✓
Observations	42,307	42,276	42,266	42,207	42,064
Ventilators
Weighted PM 2.5	−0.0993	0.333	0.0985	0.0935	2.396
	(0.477)	(0.581)	(0.556)	(1.190)	(2.617)
F Stat	18.07	17.40	17.54	18.43	17.84
Dep Var Mean	207.99	207.99	208.00	615.11	1392.01
Pct change mean	−0.048	0.160	0.047	0.015	0.172
Controls	✓	✓	✓	✓	✓
County & State-by-week FEs	✓			✓	✓
County and MSA-by-week FEs		✓
County and CBSA-by-week FEs			✓
Observations	30,614	30,591	30,580	30,561	30,484

${}^{\ast }$$p<0.1$, ${}^{\ast \ast }$$p<0.05$, ${}^{\ast \ast \ast }$$p<0.01$. Standard errors clustered at the county level in parentheses. Controls include state-level and county-level policy adoption, wind speed, minimum and maximum daily temperature, precipitation, prior two-week social distancing behavior, day-of-week, and two lagged wind direction-by-monitor cluster interactions. The table displays regressions over three different outcomes: hospitalizations, use of Intensive Care Units (ICU), and use of ventilators, at the state-day level. All outcomes represent the number of patients using these healthcare resources on a given day. Columns (1) through (3) analyze these outcomes on the same day over three types of geography-by-week fixed effects. Column (4) and column (5) display the relationship between fine particulate matter and summed daily outcomes over three days and seven days, respectively. Regressions employ county population weights. F-statistics of the relevance test are computed assuming first-stage standard errors are not serially correlated.