Akademik Veri Yönetim Sistemi
Tezin Türü: Doktora
Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Uygulamalı Matematik Enstitüsü, Aktüerya Bilimleri Anabilim Dalı, Türkiye
Tezin Onay Tarihi: 2015
Öğrenci: BÜŞRA ZEYNEP TEMOÇİN
Danışman: SEVTAP AYŞE KESTEL
Özet:In this thesis, various portfolio insurance strategies are designed and proposed for portfolio management of defined-contribution type pension plans. These type of plans consist of consecutive and defined premium payments which are invested in financial markets and lead to a benefit that will be collected at the retirement. Since the beneficiary faces all of the financial risk throughout the plan, a capital protection mechanism is needed in such retirement systems. The main contribution of the present research is to formulate this problem using different portfolio insurance methods with the aim of providing a minimum guarantee on the portfolio value under the assumption of stochastic floor processes. More specifically, various versions of Constant Proportion Portfolio Insurance (CPPI) method with distinctive floor processes are developed in different markets and with certain trading constraints. Modifying the classical dynamics of CPPI for the pension fund framework, the portfolio efficiencies of these newly introduced strategies are analyzed in continuous- and discrete-time trading markets. In a market with continuous-time trading, two distinctive CPPI strategies are introduced. With the aim of eliminating the discontinuities resulting from the contribution payments, a replication strategy is carried out and a continuous-time environment is achieved. Through a detailed sensitivity analysis and terminal wealth distributions illustrated by Monte Carlo simulations, portfolio performances are studied. To ensure that there is no bias in the comparison, optimal CPPI-multiplier for each guarantee framework is obtained via using a classical stochastic control approach. Showing that critical risks (such as cash-lock risk and gap-risk) can arise once one steps outside of continuous-time environment, the need of modeling in a more realistic market with discrete-time trading is addressed. Considering the problem in a discrete-time trading setting, additional path-dependent CPPI strategies are proposed imposing certain trading constraints. In these strategies, the floor processes are designed to vary randomly based on the performance of the portfolio with the aim of capturing cash-lock and gap risks. In addition to sensitivity analysis conducted, risk analysis is carried out via computing the local risk measures; cash-lock probability, expected shortfall (ESF) and shortfall probability. Portfolio performances of these proposed strategies are then discussed through calibration of risk measures.