. 2020 Jul 30;220(1):63–85. doi: 10.1016/j.jeconom.2020.07.038

Table 2.

Bias of estimates of $R_{0}$ for DGP (5.3)–(5.4).

	Deaths	Cases	SUR, Efficient $W$	SUR, Diagonal $W$	SUR, Identity $W$
Panel A: True $E_{0} = 2$ and $I_{0} = 0$ , Assumed $E_{0} = 2$ and $I_{0} = 0$

Levels, Cumulative	0.019	−0.001	0.002	−0.001	−0.001
Levels, Daily	0.026	0.000	0.005	0.001	0.002
Logs, Cumulative, Trim	0.016	0.001	−0.001	−0.004	0.006
Logs, Daily, Trim	0.045	0.009	0.003	0.003	0.022
Logs, Cumulative, No Trim	−0.141	0.032	−0.026	−0.028	−0.064
Logs, Daily, No Trim	−1.030	0.032	−0.091	−0.115	−0.515

Panel B: True $E_{0} = 2$ and $I_{0} = 0$ , Assumed $E_{0} = 1$ and $I_{0} = 0$

Levels, Cumulative	0.018	−0.001	0.002	−0.001	−0.001
Levels, Daily	0.025	0.000	0.004	0.000	0.000
Logs, Cumulative, Trim	0.016	0.000	−0.001	−0.004	0.006
Logs, Daily, Trim	0.042	0.009	0.003	0.003	0.021
Logs, Cumulative, No Trim	−0.142	0.031	−0.026	−0.028	−0.064
Logs, Daily, No Trim	−1.030	0.031	−0.092	−0.117	−0.508

Panel Dc: True $E_{0} = 2$ and $I_{0} = 0$ , Assumed $E_{0} = 16$ and $I_{0} = 0$

Levels, Cumulative	0.034	0.011	0.014	0.011	0.011
Levels, Daily	0.046	0.026	0.034	0.043	0.026
Logs, Cumulative, Trim	0.027	0.007	0.007	0.003	0.014
Logs, Daily, Trim	0.102	0.022	0.016	0.016	0.037
Logs, Cumulative, No Trim	−0.132	0.035	−0.023	−0.025	−0.060
Logs, Daily, No Trim	−1.029	0.038	−0.085	−0.110	−0.514

Panel D: True $E_{0} = 2$ and $I_{0} = 0$ , Estimated $E_{0}$ and $I_{0}$

Levels, Cumulative	0.020	0.000	0.009	0.009	0.014
Logs, Cumulative, Trim	0.017	0.001	0.001	0.000	0.007
Logs, Cumulative, No Trim	−0.137	0.033	−0.023	−0.023	−0.062

The table presents the average bias of the estimates of $R_{0}$ across 500 simulation draws. The estimates in the first column are based on the deaths data only. The estimates in the second column are based on the reported cases data only. The estimates in the remaining three columns combine the deaths and reported cases data and use different weighting matrices.