On the Lee-Carter model as applied to mortality tables of Baltic states and the Philippines

• Main Authors: Dela Cruz, Kristine Mae M., Dela Cruz, Sara Jane C.
• Format: text
• Language: English
• Published: Animo Repository 2009
• Subjects: Mortality > Philippines > Statistics; Philippines > Statistical, Vital; Statistics and Probability
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/5222
• Institution: De La Salle University

Mortality tables are used in planning for social security and health care systems and in computing premiums of life insurance policies. Thus, modeling these mortality tables is essential in the field of Actuarial Science. There are several models available for modeling mortality tables. One of the most popular is the Lee-Carter (LC) model. In fact, several countries applied the LC model in estimating and projecting their life tables.The final outputs of this paper are the projected mortality tables of Baltic states namely Estonia, Latvia and Lithuania for year 2025. This was realized by firstly using a singular value decomposition (SVD) to obtain parameter estimates for the LC model using SAS program. Then, ARIMA models were used to project the time series component of the LC model. This projected parameter was used to project mortality tables for the Baltic states. Furthermore, chi-square goodness-of-fit test was performed to validate the results for the Baltic states.The researchers attempted to project mortality tables for the Philippines. However, due to insufficient data the Simple Linear Regression Model (SLRM) that should be used to project parameter kt was not significant.Also, the researchers checked the forecasting ability of the ARIMA time series modeling under the LC model using two methods. First, data for the Baltic states from years 1959-2004 was used to come up with lifetable estimates for year 2005.This estimated life table was included in the life tables for years 1959-2005 and was used to forecast a life table for 2006. Similar procedure was performed until coming up with a forecasted life table for year 2007. The second method was done by directly projecting a life table for year 2007 using data of the Baltic states for years 1959-2006. The results of these two methods were compared against each other.