Türk emeklilik sistemi için parametrik reform alternatiflerinin genetik algoritma yardımıyla saptanması

ABSTRACTA GENETIC ALGORITHM-BASED SEARCHFOR PARAMETRIC REFORM ALTERNATIVESFOR THE TURKISH PENSION SYSTEM: 2005-2060Alparslan, Mustafa ArtunM.A., Department of EconomicsSupervisor: Assoc. Prof. Serdar SayanAugust 2005In this thesis, we search for parametric reform alternatives so as to achieve abalance in the long term accumulated difference between pension expenditures andrevenues of the SSI, the largest pension fund in Turkey, from 2005 to 2060. Theprojected worker-retiree composition of the pension system is allowed to change alongwith proposed policy changes in contribution and replacement rates. These changes inpopulation structure are modeled by estimated work and pension income elasticities oflabor supply values so that the resulting changes in the incomes of the worker could nowaffect the retirement decision. Possible policy alternatives are then found by a geneticalgorithm developed for this purpose. Finally, the results obtained from the program arecompared with previous and current reform proposals.Keywords: Pension Reform, Genetic Algorithms