TY - JOUR
T1 - Markov-switching models with endogenous explanatory variables II
T2 - A two-step MLE procedure
AU - Kim, Chang Jin
N1 - Funding Information:
I thank the anonymous referees, associate editor, and Cheng Hsiao for thoughtful comments on the previous version of the paper. I acknowledge financial support from Korea Research Foundation Grant (KRF-2007-327-B00086) and from the Bryan C. Cressey Professorship at the University of Washington.
PY - 2009/1
Y1 - 2009/1
N2 - This paper proposes a two-step maximum likelihood estimation (MLE) procedure to deal with the problem of endogeneity in Markov-switching regression models. A joint estimation procedure provides us with an asymptotically most efficient estimator, but it is not always feasible, due to the 'curse of dimensionality' in the matrix of transition probabilities. A two-step estimation procedure, which ignores potential correlation between the latent state variables, suffers less from the 'curse of dimensionality', and it provides a reasonable alternative to the joint estimation procedure. In addition, our Monte Carlo experiments show that the two-step estimation procedure can be more efficient than the joint estimation procedure in finite samples, when there is zero or low correlation between the latent state variables.
AB - This paper proposes a two-step maximum likelihood estimation (MLE) procedure to deal with the problem of endogeneity in Markov-switching regression models. A joint estimation procedure provides us with an asymptotically most efficient estimator, but it is not always feasible, due to the 'curse of dimensionality' in the matrix of transition probabilities. A two-step estimation procedure, which ignores potential correlation between the latent state variables, suffers less from the 'curse of dimensionality', and it provides a reasonable alternative to the joint estimation procedure. In addition, our Monte Carlo experiments show that the two-step estimation procedure can be more efficient than the joint estimation procedure in finite samples, when there is zero or low correlation between the latent state variables.
KW - Control function approach
KW - Curse of dimensionality
KW - Endogeneity
KW - Markov switching
KW - Smoothed probability
KW - Two-step estimation procedure
UR - http://www.scopus.com/inward/record.url?scp=58349109655&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=58349109655&partnerID=8YFLogxK
U2 - 10.1016/j.jeconom.2008.09.023
DO - 10.1016/j.jeconom.2008.09.023
M3 - Article
AN - SCOPUS:58349109655
SN - 0304-4076
VL - 148
SP - 46
EP - 55
JO - Journal of Econometrics
JF - Journal of Econometrics
IS - 1
ER -