Skip to main content
. 2021 Jun 20;61(4):799–825. doi: 10.1111/jors.12536

Table 2.

ML estimated spatial panel data models

Spatial panel models without endogenous spillovers Spatial panel model with endo‐genous spillovers
Containment measure SEM SEDM SDM
School closure −0.0436** (0.0099) −0.0526** (0.0108) −0.0487** (0.0108)
Establishment closure −0.0252*** (0.0131) −0.0213 (0.0147) −0.0216 (0.0147)
Shopping mall closure −0.0233* (0.0110) −0.0177 (0.0119) −0.0163 (0.0119)
Restaurant closure −0.0234** (0.0087) −0.0212* (0.0096) −0.0206* (0.0096)
Contact restrictions −0.1193** (0.0075) −0.1269** (0.0082) −0.1261** (0.0081)
Face mask duty −0.1223* (0.0555) −0.1233* (0.0555) −0.1255* (0.0551)
Spatial autoregressive parameter 0.0803** (0.0093) 0.0787** (0.0093) 0.0802** (0.0089)
Region‐specific time trends Yes Yes Yes
Regional FE Yes Yes Yes
Time FE Yes Yes Yes
LogLik 7877.9 7895.7 7901.5
Local CD 12.126 12.118 12.118
Local BSCLM 61.760 61.760 61.760

Note: Standard errors are given in parentheses.

Abbreviations: HC, heteroscedasticity consistent; Local BSCLM, Breusch and Pagan's bias‐corrected scaled LM test for spatial autocorrelation; Local CD, Pesaran's CD test for spatial autocorrelation; LogLik, logarithmic likelihood; ML, maximum likelihood; SDEM, spatial error Durbin model; SDM, spatial Durbin model; SEM, spatial error model.

*5% significance level, **1% significance level, ***10% significance level.