22 May 2012 in European Financial Management
1 August 2016 in Computers & Operations Research
In accordance with Basel Capital Accords, the Capital Requirements (CR) for market risk exposure of banks is a nonlinear function of Value-at-Risk (VaR). Importantly, the CR is calculated based on a bank’s actual portfolio, i.e. the portfolio represented by its current holdings. To tackle mean-VaR portfolio optimization within the actual portfolio framework (APF), we propose a novel mean-VaR optimization method where VaR is estimated using a univariate Generalized AutoRegressive Conditional Heteroscedasticity (GARCH) volatility model. The optimization was performed by employing a Nondominated Sorting Genetic Algorithm (NSGA-II). On a sample of 40 large US stocks, our procedure provided superior mean-VaR trade-offs compared to those obtained from applying more customary mean-multivariate GARCH and historical VaR models. The results hold true in both low and high volatility samples
1 March 2017 in European Journal of Operational Research
1 January 2014 in Computer Science and Information Systems
In this paper we solve the problem of static portfolio allocation based on historical Value at Risk (VaR) by using genetic algorithm (GA). VaR is a predominantly used measure of risk of extreme quantiles in modern finance. For estimation of historical static portfolio VaR, calculation of time series of portfolio returns is required. To avoid daily recalculations of proportion of capital invested in portfolio assets, we introduce a novel set of weight parameters based on proportion of shares. Optimal portfolio allocation in the VaR context is computationally very complex since VaR is not a coherent risk metric while number of local optima increases exponentially with the number of securities. We presented two different single-objective and a multiobjective technique for generating mean-VaR efficient frontiers. Results document good risk/reward characteristics of solution portfolios while there is a trade-off between the ability to control diversity of solutions and computation time.
1 January 2010 in Economic Annals
The objective of this paper is to estimate Serbian benchmark spot curves using the Svensson parametric model. The main challenges that we tackle are: sparse data, different currency denominations of short and longer term maturities, and infrequent transactions in the short-term market segment vs daily traded medium and long-term market segment. We find that the model is flexible enough to account for most of the data variability. The model parameters are interpreted in economic terms.