Econometric models serve the decision makers in formulating policies in social, economic and political fields. Econometricians make use of the data and/or pre-policy issues to arrive at the modelling results. One significant aspect of the data and the ad hoc policy is that the observations/documentation is not entirely reliable and complete, often the case in the transportation systems (travel demand, safety, infrastructure) modelling, political decision making, health policies, etc. Fundamentally they are partial observability problems. The key econometric issues related to partial observability and identification due to endogeneity in transportation systems is addressed. The purpose is to bring to the forefront the identification problem in transportation systems stemming from partial observability. Inferences from recent work in pavement condition and traffic safety models are used. Given the highly empirical nature of transportation contexts, sources of endogeneity and forms of exogeneity and consequent impacts on identification and parameter estimation are discussed. The modelling of latent processes gives rise to other potential forms of partial observability. These potential forms of partial observability like endogeneity, exogeneity, causality and state dependence should not be overlooked, considering the dynamic nature of infrastructure modelling. Specification tests help in identifying forms of endogeneity and suggestions to care of identification issues are made. Linear and non-linear frameworks have to be treated differently in instrumental regressions. Extreme caution in modelling and identifying the processes can help estimation problems like state dependence and heterogeneity.
Endogeneity and identification in transportation systems: Econometric relationships to partial observability
Endogenität und Identifikation in Transportsystemen: Ökonometrische Beziehung zur partiellen Beobachtbarkeit
2004
6 Seiten, 18 Quellen
Aufsatz (Konferenz)
Englisch
Controllability and Observability Analysis for Intelligent Transportation Systems
Springer Verlag | 2018
|Preference endogeneity in discrete choice models
Online Contents | 2014
|Endogeneity in Residential Location Choice Models
Transportation Research Record | 2006
|