File Download
There are no files associated with this item.
Supplementary
-
Citations:
- Appears in Collections:
Conference Paper: Identification Problems in Linear Social Interaction Models: A General Analysis Based on Matrix Spectral Decompositions
Title | Identification Problems in Linear Social Interaction Models: A General Analysis Based on Matrix Spectral Decompositions |
---|---|
Authors | |
Issue Date | 2014 |
Publisher | The Econometric Society. |
Citation | The 2nd China Meeting of the Econometric Society (CMES), Xiamen, China, 25-27 June 2014 How to Cite? |
Abstract | This paper develops a framework for analyzing the identification problems in linear social interaction models with exogenous networks. Manski's (1993) reflection problem, models with group-level unobservables (fixed effects), differencing methods, and Graham's (2008) variance contrast method can be analyzed under the proposed framework. The analysis reveals that parameters are identified if and only if the number of distinct eigenvalues of the adjacency matrix is larger than or equal to the number of parameters, and that network diameter is a key network property related to identification. |
Description | Session CP29: Applied Econometrics |
Persistent Identifier | http://hdl.handle.net/10722/205082 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Kwok, HH | en_US |
dc.date.accessioned | 2014-09-20T01:24:38Z | - |
dc.date.available | 2014-09-20T01:24:38Z | - |
dc.date.issued | 2014 | en_US |
dc.identifier.citation | The 2nd China Meeting of the Econometric Society (CMES), Xiamen, China, 25-27 June 2014 | en_US |
dc.identifier.uri | http://hdl.handle.net/10722/205082 | - |
dc.description | Session CP29: Applied Econometrics | - |
dc.description.abstract | This paper develops a framework for analyzing the identification problems in linear social interaction models with exogenous networks. Manski's (1993) reflection problem, models with group-level unobservables (fixed effects), differencing methods, and Graham's (2008) variance contrast method can be analyzed under the proposed framework. The analysis reveals that parameters are identified if and only if the number of distinct eigenvalues of the adjacency matrix is larger than or equal to the number of parameters, and that network diameter is a key network property related to identification. | en_US |
dc.language | eng | en_US |
dc.publisher | The Econometric Society. | - |
dc.relation.ispartof | China Meeting of the Econometric Society (CMES) | en_US |
dc.title | Identification Problems in Linear Social Interaction Models: A General Analysis Based on Matrix Spectral Decompositions | en_US |
dc.type | Conference_Paper | en_US |
dc.identifier.email | Kwok, HH: kwokhh@hku.hk | en_US |
dc.identifier.authority | Kwok, HH=rp01632 | en_US |
dc.identifier.hkuros | 235210 | en_US |