Table 2. Simulation results under EMH,...

Table 2.

Simulation results under EMH, three-month forward rate. RE implies MA(11) errors; general implies ARMA(1,11) errors

Period	Case	Estimator	MSE	Bias	Variance	Coverage	Length
11/83–01/23	RE	OLS	1.6898	1.0541	1.5036	0.92	4.7404
		FGLS	0.7098	0.6761	0.7261	0.95	3.3387
		Durbin	1.3941	0.9358	1.4148	0.95	4.6647
	General	OLS	18.5744	3.4972	15.3260	0.91	15.0322
		FGLS	0.6019	0.6283	0.6332	0.96	3.1202
		Durbin	1.3964	0.9370	1.4193	0.95	4.6720
01/89–04/21	RE	OLS	1.8895	1.0869	1.6558	0.91	4.9454
		FGLS	0.9995	0.7817	0.9834	0.94	3.8838
		Durbin	1.9843	1.1502	2.0973	0.96	5.6801
	General	OLS	20.6180	3.5933	16.2122	0.88	15.3295
		FGLS	0.8825	0.7451	0.9027	0.95	3.7256
		Durbin	1.9782	1.1467	2.0995	0.96	5.6832

Period	Case	Estimator	MSE	Bias	Variance	Coverage	Length
11/83–01/23	RE	OLS	1.6898	1.0541	1.5036	0.92	4.7404
		FGLS	0.7098	0.6761	0.7261	0.95	3.3387
		Durbin	1.3941	0.9358	1.4148	0.95	4.6647
	General	OLS	18.5744	3.4972	15.3260	0.91	15.0322
		FGLS	0.6019	0.6283	0.6332	0.96	3.1202
		Durbin	1.3964	0.9370	1.4193	0.95	4.6720
01/89–04/21	RE	OLS	1.8895	1.0869	1.6558	0.91	4.9454
		FGLS	0.9995	0.7817	0.9834	0.94	3.8838
		Durbin	1.9843	1.1502	2.0973	0.96	5.6801
	General	OLS	20.6180	3.5933	16.2122	0.88	15.3295
		FGLS	0.8825	0.7451	0.9027	0.95	3.7256
		Durbin	1.9782	1.1467	2.0995	0.96	5.6832

Note(s):

We use US-UK observed data $x_{t} = f_{t, k} - s_{t}$ ⁠, $T_{1} = 2, 050$ ⁠, $T_{2} = 1, 600$ ⁠. For OLS we use HAC standard errors as described in the text

[ViewLarge]

Sharing Unavailable