Table 3.

NYC panel results: safegraph trips.

	(1)	(2)	(3)	(4)	(5)
	ln($Ne{w}_{it}$) OLS	ln($Ne{w}_{it}$) IV	ln($Ne{w}_{it}$) IV	ln($Ne{w}_{it}$) IV	ln($Ne{w}_{it}$) IV
	Panel A: Full Sample
$\%\Delta Trip{s}_{i,t-2}$	0.002	0.061${}^{***}$	0.044${}^{***}$	0.046${}^{***}$	0.030${}^{***}$
	(0.002)	(0.004)	(0.007)	(0.006)	(0.007)
Root MSE	0.382	0.625	0.554	0.525	0.413
Observations	2045	2045	2045	2045	2045
First Stage F-Stat.		156.820	100.222	66.096	17.374
	Panel B: Split Sample (2020w11 – 2020w17 vs. 2020w18 – 2020w23)
$\%\Delta Trip{s}_{i,t-2}\times 1\text{st}Half$	0.000	0.066${}^{***}$	0.047${}^{***}$	0.050${}^{***}$	0.034${}^{***}$
	(0.002)	(0.006)	(0.007)	(0.007)	(0.007)
$\%\Delta Trip{s}_{i,t-2}\times 2\text{nd}Half$	0.004${}^{*}$	0.059${}^{***}$	0.042${}^{***}$	0.043${}^{***}$	0.026${}^{***}$
	(0.002)	(0.005)	(0.007)	(0.007)	(0.008)
Root MSE	0.381	0.636	0.555	0.527	0.415
Observations	2045	2045	2045	2045	2045
First Stage F-Stat.		149.913	85.900	55.961	10.712
	Controls
$\%AfricanAmerica{n}_i$			X	X
$ln\left(Ag{e}_i\right)$			X	X
$ln\left(In{c}_i\right)$			X	X
	Fixed Effects
$Zi{p}_i$	X				X
$Boroug{h}_i$				X
$Wee{k}_t$	X	X	X	X	X

Notes: NYC panel results using SafeGraph trips from home. Dependent variable is log of new cases per capita in zip code $i$in week $t$. All columns control for log of healthcare employment and the log of police employment. Panels A shows results for the full panel, reporting $\beta$ from Eq. (1) in the first column, with versions of Eq. (2.2) in columns (2)–(6): $ln\left(NewCase{s}_{it}\right)=\beta \widehat{\%\Delta Trip{s}_i}+zi{p}_i+wee{k}_t+{\epsilon }_{it}$. Panel B splits the time period in half, and interacts the coefficient of interest with the two time periods, decomposing $\beta$into ${\beta }^{1\text{st}Half},$${\beta }^{2\text{nd}Half}$. Columns (2)–(5) use both the telework and essential share instuments. Robust standard errors in parentheses. Significance: ${}^{*}p<0.10{,}^{**}p<0.05{,}^{***}p<0.01.$