Table 2.

Transition probabilities and goodness of fit statistics for MS models: (1998:Q1–2019:Q2)

StatisticsMSM1MSM2MSM3MSM4
Sigma0.01618280.01599530.01481960.0143143
Linearity Test98.412 (0.0000)a97.249 (0.0000)107.05 (0.0000)111.23 (0.0000)
p_{0|0}0.9663820.9688640.9678700.966282
p_{1|0}0.0336180.0311360.0321300.033718
p_{0|1}0.1872630.1754880.1733560.178535
p_{1|1}0.812740.824510.826640.82146
log-likelihood213.071302214.391491220.073541222.595543
AIC*T−408.142603−406.782982−414.147083−415.191086
AIC−4.80167768−4.78568215−4.87231862−4.88460101
ConvergenceStrong convergence
by SQPFb using analytical derivatives
Strong convergence
by SQPF using analytical derivatives
Strong convergence
by SQPF using analytical derivatives
Strong convergence
by SQPF using analytical derivatives
Observations85858585

Notes:

a

The values in parentheses are p. values.

b

SQPF (feasible sequential quadratic programming) follows Ox function MaxSQPF, as is indicated in Lawrence and Tits (2001) (Oxmetrix-PCGive, 2014)

or Create an Account

Close Modal
Close Modal