. 2020 Oct 21;127:103292. doi: 10.1016/j.jue.2020.103292

Table 2.

Multiple city results.

	(1)	(2)	(3)	(4)	(5)	(6)
	ln( $D e a t h s_{i})$ NYC	ln( $C a s e s_{i}$ ) NYC	ln( $C a s e s_{i}$ ) Atlanta	ln( $C a s e s_{i}$ ) Boston	ln( $C a s e s_{i}$ ) Chicago	ln( $C a s e s_{i}$ ) Philadelphia
	Panel A: OLS
$% Δ T r i p s_{i}$	0.019 $^{* * *}$	0.023 $^{* * *}$	0.003	0.051 $^{* * *}$	0.019 $^{* * *}$	0.012 $^{* * *}$
	(0.006)	(0.004)	(0.019)	(0.011)	(0.003)	(0.003)
r-sq.	0.125	0.224	0.018	0.521	0.171	0.251
obs.	159	159	22	19	206	42
	Panel B: OLS With Demographics
$% Δ T r i p s_{i}$	0.004	0.010 $^{* * *}$	0.001	0.034 $^{*}$	0.005	0.008
	(0.005)	(0.003)	(0.013)	(0.016)	(0.003)	(0.005)
R-Sq.	0.484	0.437	0.567	0.569	0.514	0.465
Obs.	159	159	22	19	206	42
	Panel C: IV With Demographics
$\hat{% Δ T r i p s_{i}}$	0.029 $^{* *}$	0.068 $^{* * *}$	0.010	0.066 $^{*}$	0.010	0.016 $^{*}$
	(0.015)	(0.020)	(0.029)	(0.034)	(0.012)	(0.009)
Root MSE	0.420	0.487	0.537	0.422	0.482	0.222
Obs.	159	159	22	19	206	42
F-Stat.	32.170	22.199	4.040	5.648	38.891	5.091
	Controls for Panels B & C
$% A f A m_{i}$	X	X	X	X	X	X
$l n (A g e_{i})$	X	X	X	X	X	X
$l n (I n c_{i})$	X	X	X	X	X	X

Notes: The dependent variable is total cases, or total deaths, per capita in zip code $i$ . All columns control for log of healthcare employment. Panels A and B show versions of OLS Eqs. (1) for separate cities’ cases, as well as NYC’s deaths. Panel C shows results from Eq. (1.2), adding additional demographic controls, $X_{i}$ : $l n (T o t a l C a s e s_{i}) = α + β \hat{% Δ T r i p s_{i}} + Γ X_{i} + ε_{i}$ . Panel C uses both the telework and essential share instuments. Robust standard errors in parentheses. Significance: $^{*} p < 0.10,^{* *} p < 0.05,^{* * *} p < 0.01 .$