Full Citation

Title: Stochastic Modeling and Forecasting of Health Changes in the U.S. Population

Citation Type: Miscellaneous

Publication Year: 2013

ISBN:

ISSN:

DOI:

NSFID:

PMCID:

PMID:

Abstract: This paper provides a methodology to generate stochastic forecasts of future age- and gender-specific self-assessed health. We first apply the framework of Lee and Carter (1992), which decomposes self-assessed health into a time effect and an age effect. We then extend the Lee-Carter model by further decomposing the time effect into observed macro-economic quantities (GDP and unemployment rate) and an unobserved latent time factor. Using a vector auto regression model (VAR) to forecast the observed and unobserved time effects, this paper forecasts future health and quantifies the forecasting uncertainty. Changes in the U.S. populations self-assessed health for both males and females are estimated and forecasted based on this approach. The estimation results show that trends in health can be largely captured by trends in the observed macroeconomic quantities. A back testing analysis suggests that the Lee-Carter model with macro-economic quantities significantly improves the forecasting accuracy for future health development compared with the original Lee-Carter model. An extensive sensitivity analysis adds robustness to our results. As an application of health forecasting, this paper estimates and predicts life expectancy and healthy life expectancy, and quantifies the degree of uncertainty in their predictions.

User Submitted?: No

Authors: De Waegenaere, Anja; Yang, Ying; Melenberg, Bertrand

Publisher: Tilburg University

Data Collections: IPUMS Health Surveys - NHIS

Topics: Methodology and Data Collection, Other

Countries: