We consider the identification of a Markov process when only {Wt} is observed. In structural dynamic models, Wt includes the choice variables and observed state variables of an optimizing agent, while denotes time-varying serially correlated unobserved state variables (or agent-specific unobserved heterogeneity). In the non-stationary case, we show that the Markov law of motion is identified from five periods of data Wt+1,Wt,Wt−1,Wt−2,Wt−3. In the stationary case, only four observations Wt+1,Wt,Wt−1,Wt−2 are required. Identification of is a crucial input in methodologies for estimating Markovian dynamic models based on the “conditional-choice-probability (CCP)” approach pioneered by Hotz and Miller.
Authors
Johns Hopkins University
Matthew Shum
Journal article details
- DOI
- 10.1016/j.jeconom.2012.05.023
- Publisher
- Elsevier
- Issue
- Volume 171, Issue 1, November 2012
Suggested citation
Hu, Y and Shum, M. (2012). 'Nonparametric identification of dynamic models with unobserved state variables' 171(1/2012)
More from IFS
Understand this issue
What are the challenges in getting debt on a falling path?
28 June 2024
Election Special: Your questions answered
27 June 2024
What is the two-child limit in benefits?
27 June 2024
Policy analysis
How would the parties’ tax and spending plans affect Scotland and Wales?
28 June 2024
Free breakfast clubs in schools: what Labour’s plans would mean for pupils and families
25 June 2024
What are the parties’ plans for benefits and taxes?
24 June 2024
Academic research
Fiscal Studies: Volume 45, Issue 2
28 June 2024
Income inequality in Ireland, 1987–2019
28 June 2024
Inequality in Denmark, 1987–2021
28 June 2024