Nikos S. Thomaidis

Active Portfolio Management under a Downside Risk Framework: Comparison of a Hybrid Nature --- Inspired Scheme

Hybrid intelligent systems are becoming more and more popular in solving nondeterministic polynomial-time --- hard optimization problems. Lately, the focus is on nature --- inspired intelligent algorithms, whose main advantage is the exploitation of unique features of natural systems. One type of complex optimization problems is the active portfolio management, where the incorporation of complex, realistic constraints makes it difficult for traditional numerical methods to deal with it. In this paper we perform a computational study of a hybrid Ant Colony Optimization algorithm. The application is a specific formulation of the problem. Our main aim in this paper is to introduce a new framework of study in the field of active portfolio management, where the main interest lies in minimizing the risk of the portfolio return falling below the benchmark. Secondary, we provide some preliminary results regarding the use of a new hybrid nature --- inspired scheme in solving this type of problem.

A Soft Computing Approach to Enhanced Indexation

We propose an integrated and interactive procedure for designing an enhanced indexation strategy with predetermined investment goals and risk constraints. It is based on a combination of soft computing techniques for dealing with practical and computation aspects of this problem. We deviate from the main trend in enhanced indexation by considering a) restrictions on the total number of tradable assets and b) non-standard investment objectives, focusing e.g. on the probability that the enhanced strategy under-performs the market. Fuzzy set theory is used to handle the subjectivity of investment targets, allowing a smooth variation in the degree of fulfilment with respect to the value of performance indicators. To deal with the inherent complexity of the resulting cardinality-constraint formulations, we apply three nature-inspired optimisation techniques: simulated annealing, genetic algorithms and particle swarm optimisation. Optimal portfolios derived from “soft” optimisers are then benchmarked against the American Dow Jones Industrial Average (DJIA) index and two other simpler heuristics for detecting good asset combinations: a Monte Carlo combinatorial optimisation method and an asset selection technique based on the capitalisation and the beta coefficients of index member stocks.

A COMPARISON OF NEURAL NETWORK MODEL SELECTION STRATEGIES FOR THE PRICING OF S&P 500 STOCK INDEX OPTIONS

This paper compares a number of neural network model selection approaches on the basis of pricing S&P 500 stock index options. For the choice of the optimal architecture of the neural network, we experiment with a “top-down” pruning technique as well as two “bottom-up” strategies that start with simple models and gradually complicate the architecture if data indicate so. We adopt methods that base model selection on statistical hypothesis testing and information criteria and we compare their performance to a simple heuristic pruning technique. In the first set of experiments, neural network models are employed to fit the entire options surface and in the second they are used as parts of a hybrid intelligence scheme that combines a neural network model with theoretical option-pricing hints.

An intelligent statistical arbitrage trading system

This paper proposes an intelligent combination of neural network theory and financial statistical models for the detection of arbitrage opportunities in a group of stocks. The proposed intelligent methodology is based on a class of neural network-GARCH autoregressive models for the effective handling of the dynamics related to the statistical mispricing between relative stock prices. The performance of the proposed intelligent trading system is properly measured with the aid of profit & loss diagrams.

On the performance and convergence properties of hybrid intelligent schemes: application on portfolio optimization domain

Hybrid intelligent algorithms, especially those who combine natureinspired techniques, are well known for their searching abilities in complex problem domains and their performance. One of their main characteristic is that they manage to escape getting trapped in local optima. In this study, two hybrid intelligent schemes are compared both in terms of performance and convergence ability in a complex financial problem. Particularly, both algorithms use a type of genetic algorithm for asset selection and they differ on the technique applied for weight optimization: the first hybrid uses a numerical function optimization method, while the second one uses a continuous ant colony optimization algorithm. Results indicate that there is great potential in combining characteristics of natureinspired algorithms in order to solve NP-hard optimization problems.

Active portfolio management from a fuzzy multi-objective programming perspective

We consider the problem of structuring a portfolio that outperforms a benchmark index, assuming restrictions on the total number of tradable assets. We experiment with non-standard formulations of active portfolio management, outside the mean-variance framework, incorporating approximate (fuzzy) investment targets and portfolio constraints. To deal with the inherent computational difficulties of cardinality-constrained active allocation problems, we apply three nature-inspired optimisation procedures: simulated annealing, genetic algorithms and particle swarm optimisation. Optimal portfolios derived from these methods are benchmarked against the Dow Jones Industrial Average index and two simpler heuristics for detecting good asset combinations, based on Monte-Carlo simulation and fundamental analysis.

Improving neural network based option price forecasting

As is widely known, the popular Black & Scholes model for option pricing suffers from systematic biases, as it relies on several highly questionable assumptions. In this paper we study the ability of neural networks (MLPs) in pricing call options on the SP in particular we investigate the effect of the hidden neurons in the in- and out-of-sample pricing. We modify the Black & Scholes model given the price of an option based on the no-arbitrage value of a forward contract, written on the same underlying asset, and we derive a modified formula that can be used for our purpose. Instead of using the standard backpropagation training algorithm we replace it with the Levenberg-Marquardt approach. By modifying the objective function of the neural network, we focus the learning process on more interesting areas of the implied volatility surface. The results from this transformation are encouraging.

Stochastic Convergence Analysis of Metaheuristic Optimisation Techniques

Commonly used metaheuristic optimisation techniques imbed stochastic elements into the selection of the initial population or/and into the solution-search strategy. Introducing randomness is often a means of escaping from local optima when searching for the global solution. However, depending on the ruggedness of the optimisation landscape and the complexity of the problem at hand, this practice leads to a dispersion of the reported solutions. Instead of relying on the best solution found in a set of runs, as is typical in many optimisation exercises, it is essential to get an indication of the expected dispersion of results by estimating the probability of converging to a “good” solution after a certain number of generations. We apply a range of statistical techniques for estimating the success probability and the convergence rate of popular evolutionary optimisation heuristics in the context of portfolio management. We show how this information can be utilised by a researcher to obtain a deeper understanding of algorithmic behaviour and to evaluate the relative performance of competitive optimisation schemes.

On Detecting the Optimal Structure of a Neural Network Under Strong Statistical Features in Errors

The purpose of this article is to investigate the empirical performance of various statistical techniques for detecting the optimal structure of a neural network (NN) regression model. We are particularly concerned with the specification of the NN architecture when the error component is characterized by special statistical properties, such as heteroskedasticity and non-normality. We consider the sequential testing procedure based on standard Lagrange multiplier (LM) tests for neglected nonlinearity and also examine three modifications of this test that are robust to heteroskedasticity. By means of Monte Carlo simulations, we investigate the ability of these procedures to detect the right structure of the NN under different types of heteroskedasticity and noise distributions. Simulation results show that robustified LM tests allow the researcher to control the complexity of the NN without having to explicitly model all statistical aspects of the data-generating process, something which is not generally feasible with the standard LM test. The combination of robust regression-based testing with bootstrapping and generalized autoregressive conditional heteroskedasticity modelling techniques increases the efficiency of the statistical sequential procedure in eliciting the optimal NN architecture.

The impact of cross‐listings on the UK and the German stock markets

Purpose - The purpose of this paper is to examine volatility transmissions between portfolios of cross-listed equities and exchange rate differences and also the volatility persistence for home, foreign equities, and exchange rate differences in the UK and German markets. Design/methodology/approach - A primary focus of this paper is to see if there is an impact first on the volatility persistence for foreign equities that are listed in the UK and German markets, second on the respective home portfolios of cross-listed equities, and third on the exchange rate differences. In addition, whether there are any bilateral spillovers between the following equity portfolios: foreign cross-listed equities, home cross-listed equities, and also local or global exchange rate differences are investigated. Findings - The paper finds that the volatility persistence is more prominent than error persistence from cross-listed equities, foreign or home, and the exchange rate differences. Furthermore, the transmission mechanism indicates a bilateral integration process in some of the cases that were examined. Based on these results, it is concluded that in the UK market the foreign cross-listings affect less the domestic equities compared to the German market. Originality/value - This paper examines the interdependence of portfolios of home and foreign equities for cross-listings that belong to the same stock exchange with two exchange rates, a local and a global one in order to provide more evidence in this area of literature.