Table 2.

Monte Carlo simulation study results by method and scenario

Scenario	Root mean squared error
	A1	A2	B1	B2	C1	C2	D1	D2
Parallel trends	Holds	Holds	Fails	Fails	Fails	Fails	Fails	Fails
Homogenous treatment effects	Yes	No	Yes	No	Yes	No	Yes	No
Time‐invariant covariates and treated units in convex hull	No	No	No	No	Yes	Yes	No	No
Group‐specific shock postintervention	No	No	No	No	No	No	Yes	Yes
18 pre‐treatment periods
Difference‐in‐differences	0.01	0.01	43.74	43.74	10.78	10.78	74.25	74.25
Synthetic controls	0.09	0.09	4.46	4.46	0.06	0.06	14.88	14.88
Interactive fixed effects	0.01	0.36	0.06	1.88	0.02	0.11	4.37	6.43
Generalized synthetic controls	0.01	0.01	0.07	0.07	0.02	0.02	4.28	4.28
12 pre‐treatment periods
Difference‐in‐differences	0.01	0.01	23.19	23.19	5.65	5.65	46.38	46.38
Synthetic controls	0.11	0.11	3.97	3.97	0.07	0.07	13.47	13.47
Interactive fixed effects	0.02	0.32	0.14	0.71	0.04	0.14	4.04	1.70
Generalized synthetic controls	0.02	0.02	0.11	0.11	0.03	0.03	4.30	4.30
9 pre‐treatment periods
Difference‐in‐differences	0.01	0.01	15.34	15.34	3.78	3.78	34.93	34.93
Synthetic controls	0.13	0.13	3.80	3.80	0.10	0.10	12.98	12.98
Interactive fixed effects	0.02	0.34	0.17	0.98	0.05	0.17	4.19	1.52
Generalized synthetic controls	0.02	0.02	0.18	0.18	0.05	0.05	4.58	4.58
6 pre‐treatment periods
Difference‐in‐differences	0.02	0.02	9.05	9.05	2.24	2.24	24.98	24.98
Synthetic controls	0.16	0.16	3.66	3.66	0.13	0.13	12.97	12.97
Interactive fixed effects	0.03	0.33	0.19	1.33	0.09	0.19	4.27	1.39
Generalized synthetic controls	0.03	0.03	0.34	0.34	0.10	0.10	5.00	5.00