A COMPARATIVE STUDY FOR MULTI-PERIOD ASSET ALLOCATION OF DEFINED CONTRIBUTION SCHEMES: EVIDENCE FROM TURKEY

Long term asset allocation is more complicated than the usual asset allocation paradigm due to the multi-period nature of the problem. Since analytical models usually oversimplify for the sake of mathematical convenience, a fundamentally different approach is needed. Numerical solutions such as genetic algorithms provide an important alternative to analytical solutions in handling the inherent complexity which manifests itself as discontinuities and nonlinearities of the solution space. Complementing our previous studies with hypothetical and US market data, this study brings further evidence on the comparative performance of numerical solutions for multi-period asset allocation from an emerging market.