APPLICATION OF MULTIVARIATE LAMBDA DISTRIBUTION WITHIN THE PORTFOLIO SELECTION MODEL

MÁRIO PČOLÁR

https://doi.org/10.53465/EDAMBA.2021.9788022549301.383-389

 

Abstract: The article deals with the application of the resampling procedure using multivariate lambda distribution within the process of optimization of portfolio selection models. The aim of the resampling procedure is to achieve portfolios that provide better quality results on out-of-sample data compared to the traditional optimization-based approach using estimates from historical data. In this paper, we deal with the application of the resampling procedure on daily data of 30 assets within the model of portfolio selection in the space of expected return and CVaR (Conditional Value at Risk). We are dealing with the application of two approaches, an approach based on the assumption of normal distribution of data using multivariate normal distribution for data generation and a procedure using data generation from multivariate generalized lambda distribution.

Keywords: CVaR optimization, multivariate distribution, resampling

JEL classification: G11, G17

Fulltext: PDF

Online publication date: 12 May 2022

 

To cite this article (APA style): 

Pčolár, M. (2022). Application of Multivariate Lambda Distribution within the Portfolio Selection Model. Proceedings from the EDAMBA 2021 conference, 383 – 389. https://doi.org/10.53465/EDAMBA.2021.9788022549301.383-389