Table 6.

Direct and indirect policy effects without national case/death variables.

Cases
	Direct	Indirect	Total	PI$\to$Y Coef.	Average	Difference
masks for employees	−0.090${}^{\ast \ast \ast }$	0.043	−0.047	−0.083${}^{\ast \ast }$	−0.065	0.036${}^{\ast \ast }$
	(0.031)	(0.028)	(0.043)	(0.038)	(0.040)	(0.015)
closed K-12 schools	−0.074	−0.374${}^{\ast \ast \ast }$	−0.448${}^{\ast \ast \ast }$	−0.226${}^{\ast \ast \ast }$	−0.337${}^{\ast \ast \ast }$	−0.223${}^{\ast \ast \ast }$
	(0.080)	(0.095)	(0.116)	(0.086)	(0.098)	(0.055)
stay at home	−0.063	−0.034	−0.096${}^{\ast }$	−0.127${}^{\ast \ast }$	−0.112${}^{\ast \ast }$	0.031${}^{\ast \ast }$
	(0.049)	(0.029)	(0.055)	(0.058)	(0.056)	(0.015)
business closure policies	0.051	−0.153${}^{\ast \ast \ast }$	−0.101	−0.076	−0.089	−0.025
	(0.062)	(0.045)	(0.068)	(0.067)	(0.067)	(0.021)
${\sum }_j{\mathrm{Policy}}_j$	−0.176	−0.517${}^{\ast \ast \ast }$	−0.693${}^{\ast \ast \ast }$	−0.512${}^{\ast \ast \ast }$	−0.603${}^{\ast \ast \ast }$	−0.181${}^{\ast \ast \ast }$
	(0.129)	(0.145)	(0.182)	(0.147)	(0.162)	(0.061)
$\Delta log\Delta C_{it}$	0.015	0.012	0.027	0.040	0.033	−0.013${}^{\ast \ast }$
	(0.026)	(0.009)	(0.025)	(0.025)	(0.025)	(0.007)
$log\Delta C_{it}$	−0.105${}^{\ast \ast \ast }$	−0.045${}^{\ast \ast \ast }$	−0.150${}^{\ast \ast \ast }$	−0.137${}^{\ast \ast \ast }$	−0.144${}^{\ast \ast \ast }$	−0.013${}^{\ast }$
	(0.018)	(0.015)	(0.025)	(0.021)	(0.023)	(0.008)
Deaths
	Direct	Indirect	Total	PI$\to$Y Coef.	Average	Difference
masks for employees	−0.146${}^{\ast \ast \ast }$	0.035${}^{\ast \ast \ast }$	−0.111${}^{\ast \ast }$	−0.134${}^{\ast \ast \ast }$	−0.122${}^{\ast \ast }$	0.022
	(0.049)	(0.000)	(0.049)	(0.051)	(0.049)	(0.021)
closed K-12 schools	−0.232${}^{\ast \ast }$	−0.420${}^{\ast \ast \ast }$	−0.653${}^{\ast \ast \ast }$	−0.610${}^{\ast \ast \ast }$	−0.632${}^{\ast \ast \ast }$	−0.042${}^{\ast }$
	(0.095)	(0.084)	(0.111)	(0.109)	(0.109)	(0.023)
stay at home	−0.066	−0.008	−0.073	−0.082	−0.078	0.008
	(0.068)	(0.031)	(0.067)	(0.067)	(0.067)	(0.016)
business closure policies	0.098${}^{\ast \ast \ast }$	−0.163${}^{\ast \ast }$	−0.065	−0.059	−0.062	−0.006
	(0.000)	(0.065)	(0.065)	(0.086)	(0.061)	(0.089)
${\sum }_j{\mathrm{Policy}}_j$	−0.346${}^{\ast \ast \ast }$	−0.557${}^{\ast \ast \ast }$	−0.902${}^{\ast \ast \ast }$	−0.885${}^{\ast \ast \ast }$	−0.894${}^{\ast \ast \ast }$	−0.018
	(0.122)	(0.137)	(0.156)	(0.157)	(0.147)	(0.106)
$\Delta log\Delta D_{it}$	0.016	−0.040${}^{\ast \ast \ast }$	−0.024	−0.001	−0.012	−0.023${}^{\ast \ast \ast }$
	(0.034)	(0.011)	(0.030)	(0.032)	(0.031)	(0.005)
$log\Delta D_{it}$	−0.055${}^{\ast \ast }$	−0.015	−0.070${}^{\ast \ast }$	−0.078${}^{\ast \ast \ast }$	−0.074${}^{\ast \ast \ast }$	0.009
	(0.024)	(0.009)	(0.028)	(0.026)	(0.027)	(0.005)

Direct effects capture the effect of policy on case growth holding behavior, information, and confounders constant. Direct effects are given by in Eq. (BPI→Y). Indirect effects capture how policy changes behavior and behavior shift case growth. They are given by from (BPI→Y) times from (PI→B). The total effect is . Column “PI$\to$Y Coefficients” shows the coefficient estimates from (PI→Y). Columns “Difference” are the differences between the estimates from (PI→Y) and the combination of (BPI→Y) and (PI→B) while column “Average” are their averages. Standard errors are computed by bootstrap and clustered on state.