Understanding Investment Funds

Transcription

1 Understanding Investment Funds

2 This page intentionally left blank

3 Understanding Investment Funds Insights from Performance and Risk Analysis Edited by Virginie Terraza University of Luxembourg and Hery Razafitombo University of Lorraine, France

4 Editorial matter, selection and introduction Virginie Terraza, Hery Razafitombo 2013 Remaining chapters respective authors 2013 Softcover reprint of the hardcover 1st edition All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6 10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act First published 2013 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number , of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin s Press LLC, 175 Fifth Avenue, New York, NY Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave and Macmillan are registered trademarks in the United States, the United Kingdom, Europe and other countries ISBN ISBN (ebook) DOI / This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress CPI Antony Rowe, Chippenham and Eastbourne

5 Contents List of Figures List of Tables Acknowledgments Notes on Contributors Introduction 1 Virginie Terraza and Hery Razafitombo Part I New Performance Measure Methodologies 1 Is There a Link between Past Performance and Fund Failure? 9 Philippe Cogneau, Laurent Bodson, and Georges Hübner 1.1 Introduction Data and variable construction Data Selection of relevant performance measures The link between fund performance and subsequent disappearance Global results Specific aspects of predictability Conclusion 23 2 The Fund Synthetic Index: An Alternative Benchmark for Mutual Funds 37 Virginie Terraza and Hery Razafitombo 2.1 Introduction Methodology and data Fund synthetic index construction Data Results A statistical comparison Value at Risk analysis Correlation issues Conclusion 51 viii ix xi xii v

6 vi Contents 3 Hedge Funds Risk-adjusted Performance Evaluation: A Fuzzy Set Theory-Based Approach 57 Alfred M. Mbairadjim, Jules Sadefo Kamdem, and Michel Terraza 3.1 Introduction Theoretical background and definitions Possibilistic performance and application Definitions Empirical study Concluding remarks 67 Part II Advanced Risk Analyses and Modeling 4 Hedge Funds Risk Measurement in the Presence of Persistence Phenomena 75 Mohamed A. Limam, Rachida Hennani, and Michel Terraza 4.1 Introduction Value at Risk framework VaR and financial markets The benchmark model: RiskMetrics The ARFIMA-FIAPARCH models Model performances Kupiec s test (1995) The Manganelli and Engle s test (2004) Data and return dynamics analysis Unit root and stationarity tests Testing for persistence Double long memory estimation and prevision The ARFIMA-FIGARCH/FIAPARCH models Backtesting tests Conclusion Conditioned Higher-Moment Portfolio: Optimization Using Optimal Control 106 Marc Boissaux and Jang Schiltz 5.1 Introduction Context The conditioned portfolio problem Portfolio optimization involving higher moments of returns Data Results Mean-kurtosis (MK) optimization Mean-variance-kurtosis (MVK) optimization 118

7 Contents vii Mean-variance-skewness-kurtosis (MVSK) optimization Conclusion The Hazard-Adjusted Portfolio: A New Capital Allocation Scheme from an Extreme-Risk Management Perspective 129 Falk Laube and Virginie Terraza 6.1 Introduction Inferring univariate and multivariate market stability Filtering for market stability Pure bivariate MSM BiMSM crisis indicators Multifractal effects in the fund universe Univariate multifractal effects Multivariate multifractal effects Multifractality between pairs of assets Dynamics of co-movements the structural decomposition of correlations Detecting crises based on multifractal dynamics of volatility processes The structure of the hazard-adjusted portfolio Portfolio structure Optimal weighting of pre-allocated assets Simulation of the hazard-adjusted fund of hedge funds Setup of the FoHF simulation Portfolio performance Concluding remarks Mutual Fund Rating: A Symbolic Data Approach 161 Virginie Terraza and Carole Toque 7.1 Introduction Methodology Symbolic objects and clustering on interval variables Quality indexes for selecting the best partition Financial symbolic data table Application and results Conclusion 172 Index 176

8 Figures 1.1 ROC curve for performance horizon h = 4 years and prediction period of 3 months (T = 0.25) The evolution of benchmarks during the whole period Correlation circles Possibility distributions of the hedge funds returns resulting from probability possibility conversion Hedge fund daily NAVs Graphical analysis of the six studied funds: log-returns, density, ACF and PACF functions Ex-ante and ex-post mean absolute returns kurtosis for all 21 points on the discretized efficient frontier MK ex-post returns time path for highest attainable expected return point Expected (ex-ante) and observed (ex-post) utility values for both unconditioned and conditioned optimizers with the given set of MVK utility functions The multifractal structure for equities and corporate bonds The correlations of switching behavior across individual variance regimes Correlations of identical volatility frequencies between different pairs of assets The conditional correlation decomposition of return returns and variance regime switches The evolution of joint crisis probabilities (JCP) over time for some select cases The Markowitz unrestricted efficient frontier FoHF simulation results for the crisis-resilient, minimum-risk approach FoHF simulation results for the crisis-resilient, maximumreturn approach Benchmarking the pure Markowitz against the extreme-risk filtered FoHF Superimpose VSTAR for the 3-partition with three clusters Superimpose VSTAR for the 5-partition with two clusters 172 viii

9 Tables 1.1 Summary statistics of the fund returns Fund delistings per year, currency and motive Aggregate number of performance estimates Selected performance measures per horizon Somers s D statistic for the global sample Somers s D statistic per type of disappearance Coefficients of the discriminant function Somers s D statistic for 13 popular performance measures Somers s D statistic for subsets of funds Database Data properties of indexes Cornish Fisher Value at Risk Cornish Fisher GARCH Value at Risk Correlation matrix List and tickers of hedge funds used in the empirical study Summary statistics of the hedge monthly returns Comparison between probabilistic and possibilisic mean and variances Funds ranking with possibilistic versus probabilistic approaches, based on Sharpe ratio Funds ranking with possibilistic versus probabilistic approaches, based on information ratio Names of selected funds Descriptive statistics BDS test Unit root and stationarity tests LO s R/S test Estimation of the fractional integration parameter ARFIMA-FIGARCH/FIAPARCH estimation results under normal distribution ARFIMA-FIGARCH/FIAPARCH estimation results under student and skewed student distributions In-the-sample results of Kupiec s test for the ARFIMA-FIGARCH model In-the-sample results of Manganelli and Engle s test for the ARFIMA-FIGARCH model In-the-sample results of Kupiec s test for the ARFIMA-FIAPARCH model 96 ix

10 x List of Tables 4.12 In-the-sample results of Manganelli and Engle s test for the ARFIMA-FIAPARCH model Out-of-the-sample results of Kupiec s test for the ARFIMA-FIGARCH model Out-of-the-sample results of Manganelli and Engle s test for the ARFIMA-FIGARCH model Out-of-the-sample results of Kupiec s test for the ARFIMA-FIAPARCH model Out-of-the-sample results of Manganelli and Engle s test for the ARFIMA-FIAPARCH model Summary statistics Mean metrics of portfolio returns based on mean-kurtosis (MK) optimization Polynomial coefficients for the 9 MVK utility functions used Mean metrics of portfolio returns based on mean-variance-kurtosis (MVK) optimization Polynomial coefficients for the 9 MVSK utility functions used Mean metrics of portfolio returns based on mean-variance-kurtosis (MVK) optimization Details of the time series used for construction of the hazard-adjusted portfolio Comparison of MSM models for different numbers of frequencies k Lengths of the different volatility cycles for all markets A structural representation of the construction process for the hazard-adjusted portfolio Details of the simulation setup for the hazard-adjusted FoHF The general properties of the hazard-adjusted FoHF and the cost structure Comparing extreme-risk filtered PF and the pure constrained Markowitz A data table with interval variables An extract of the financial symbolic data table The three number-of-clusters criteria for k = The stability measures for the 3-partition The stability measures for the 5-partition 170

11 Acknowledgments The contents of this book are based on the workshop on investment funds held in Luxembourg in March We would like to thank the Center for Research in Economic analysis (CREA) at the University of Luxembourg for their helpful assistance as the local organizer. This publication could not have been achieved without the financial support of the Fonds National de la Recherche (FNR) in Luxembourg, which provided the funds for our research project titled Banques, marchés financiers et législations: interdépendances et mesure comparée des performances from the program Vivre. We are hence grateful to the FNR for their support and contribution. Finally, thanks are also due to all who participated in this workshop. The ideas and rich experiences that they shared are presented in this book. xi

12 Notes on Contributors Laurent Bodson is Affiliate Professor of Finance at HEC Management School of the University of Liège, Belgium, and Head of Asset Management at Gambit Financial Solutions. His areas of expertise include portfolio and risk management in non-gaussian frameworks, as both a practitioner and a researcher. Marc Boissaux is a research assistant in Finance at the Luxembourg School of Finance (LSF), University of Luxembourg. His research focuses on a new generic optimal control formulation for the conditioned portfolio optimization problem. He has worked for various financial companies and his research interests include portfolio optimization and algorithmic trading. Philippe Cogneau is a senior consultant in the Banking Sector and Associate Researcher, HEC Management School of the University of Liège, Belgium. The main focus of his research is the performance measurement of funds, but he is also involved in research concerning bootstrapping methods in finance. He lectures on banking products at the High Fr. Ferrer School in Brussels and on derivatives at the University of El Jadida, Morocco. Rachida Hennani is a doctoral candidate in Econometrics at the Laboratoire Montpelliérain d Economie Théorique et Appliquée (LAMETA), University of Montpellier 1, France. Her research focuses on the non-linear econometrics of Value at Risk, including the chaotic approach of financial markets. Georges Hübner is Full Professor of Finance and the Deloitte Chair of Portfolio Management and Performance, and Board member of HEC-Management School, University of Liège, Belgium. He is also Associate Professor of Finance at Maastricht University and an affiliate professor at EDHEC (France - Singapore) and Solvay Brussels School of Economics and Management. He regularly lectures on management at various European institutions such as INSEAD, including preparing seminars for the GARP (Global Association of Risk Professionals) certification. His research focuses on financial risk management, performance evaluation, and portfolio optimization. Jules Sadefo Kamdem is Associate Professor at the University of Montpellier 1 (LAMETA). He received his Ph.D. in mathematical finance from the University of Reims, France. His research focuses on quantitative risk management, financial econometrics, and fuzzy economics and finance. Falk Laube is a full time research associate at R.G. Niederhoffer Capital Management, Inc., New York, USA. He currently conducts research in the areas of risk management, Bayesian econometrics, and signal processing. xii

13 Notes on Contributors xiii Mohamed A. Limam is a researcher in econometrics and a doctoral candidate at the University of Montpellier 1 (LAMETA), France. His research focuses on long memory processes, persistence phenomena, and their effect on the fund industry. Alfred M. Mbaraidjim is a researcher in econometrics and a doctoral candidate at the University of Montpellier I (LAMETA), France. His research areas are computational and applied statistics, operational research, and financial modelling. Hery Razafitombo is Associate Professor of Finance and the co-founder of the Master s in International Finance at the Paul Verlaine University, Metz, France. He is a member of the CEREFIGE research group, the Centre for Economics, Finance and Management Research at the University of Lorraine (Nancy-Metz). His research focuses on the relevance of performance measures for fund analysis and selection. His interests include theory, empirical, and consultative projects with financial institutions and the fund industry in France and Luxembourg. Jang Schiltz has been Associate Professor in Applied Mathematics at the University of Luxembourg since His current research interests are functional data analysis, mathematical finance, and pension systems. Michel Terraza is Professor at the University of Montpellier 1 and founder of the Master s in Finance at that university. He is the head of the Axis Econometrics Department, which includes finance, commodities and agricultural markets, and wage inequality. Virginie Terraza is Associate Professor of Finance, a researcher at CREA (the Centre for Research in Applied Economics and Management), and the director of the Bachelor s of Management programme at the University of Luxembourg. She is also an associate researcher and lecturer at the University of Montpellier 1 (LAMETA), France. Her fields of interest are financial econometrics, risk management, and the fund industry. Carole Toque is a researcher in finance and data mining, R & D consultant, and statistical analyst at Syrokko, a data mining consultancy based in France.