Unobserved heterogeneity is one of the main concerns for applied economists, this is particularly so when modelling health and health related behaviours. This thesis illustrates four studies on modelling unobserved heterogeneity using some recent developments in latent class analysis. Chapter 2 examines two sources of individual unobserved heterogeneity when subjective indicators are used to measure health status: variations in unobservable true health and differences in self-reporting behaviour for a given level of “true health”. These two sources are separately identified using both objective (biomarkers) and subjective health indicators. Chapter 3 examines the so called positive correlation test. This test rejects the null of absence of private information in a given insurance market when individuals with greater coverage experience more of the insured risk. This test is shown to lead to puzzling results where there exists multiple sources of private information (multidimensional heterogeneity). An alternative strategy proposed uses a finite number of heterogeneous types and extends the standard adverse and favourable selection definitions into local and global ones. We implement a finite mixture model to identify the unobserved types and test the multidimensionality of private information. We apply these approaches to the US long-term care and Medigap insurance markets.
Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:557171 |
Date | January 2010 |
Creators | Li Donni, Paolo |
Contributors | Jones, Andrew ; Rice, Nigel |
Publisher | University of York |
Source Sets | Ethos UK |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Source | http://etheses.whiterose.ac.uk/1162/ |
Page generated in 0.0149 seconds