Hedge Fund Replication From Replication To Forecasting

19

Click here to load reader

description

Case Study Presentation at the Hedge Fund Replication and Alternative Beta conference on 27th – 29th November 2007 Ritz-Carlton Hotel, Hong Kong, China

Transcript of Hedge Fund Replication From Replication To Forecasting

Page 1: Hedge Fund Replication From Replication To Forecasting

s

1

December 07

From Replication to Forecasting –Creating a new and active hedge fund benchmark

Hedge Fund Replication & Alternative Beta28th – 29th November 2007Ritz-Carlton Hotel, Hong Kong

Page 2: Hedge Fund Replication From Replication To Forecasting

December 07

2

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

This presentation and the analysis herein contains proprietary information and is not to be copied, reproduced, used, or divulged to any person in whole or in part without proper written authorization from an officer or director of Siemens AG. This information is the property of Siemens AG and is subject to completion and amendment. The content of the presentation should not be interpreted as legal, tax, or investment advice. This document has been prepared by Siemens for discussion purposes only, based upon unaudited financial data. Siemens does not make any representation that the strategy will or is likely to achieve performance comparable to that shown. This document is not an offer to sell or a solicitation for the sale of a security nor shall there be any sale of security in any jurisdiction where such offer, solicitation, or sale would be unlawful. An investment in any of the products may involve a high degree of risk, including the risk of complete loss of an investment, and may only be made pursuant to final offering documents. Past performance of Siemens and / or any of its respective affiliates, employees, members, or principals is not indicative of future results and is no guarantee targeted performance will be achieved. Siemens is under no obligation to release to the public any revised financial data that reflect anticipated or unanticipated events or circumstances. This presentation does not claim to be all-inclusive or to contain all of the information that any particular party may desire. No representation or guarantee is made regarding the accuracy or completeness of any of the information contained herein. Any person in possession of this presentation agrees that all of the information contained herein is of a confidential nature. Furthermore, the same person will treat the information in a confidential manner and will not directly or indirectly, disclose, or permit agents or affiliates to disclose, any of such information without the prior written consent of Siemens.

BY ACCEPTING THIS DOCUMENT YOU ACKNOWLEDGE THAT ALL OF THE INFORMATION HEREIN SHALL BE KEPT STRICTLY CONFIDENTIAL BY YOU.

The views and opinions expressed in this presentation are those of the authors only, and do not necessarily represent the views and opinions of Siemens AG, or any of its employees. The authors make no representations or warranty, either expressed or implied, as to the accuracy or completeness of the information contained in this presentation, nor are they recommending that this presentationserves as the basis for any investment decision. This presentation is prepared for the Hedge Fund Replication & Alternative Beta 2007, 27th November – 29th November 2007, Ritz-Carlton, Hong Kong only. Research support from Fin4Cast is gratefully acknowledged.

Prof. Georg Dorfleitner*, Maria Crepaz**, Klaus Gams**, Dr. Martin Kuehrer** and Dr. Miroslav Mitev**

* Professor of Finance, Department of Finance, University of Regensburg, Germany.

** Siemens AG Österreich, Siemens IT Solutions and Services, Program and System Engineering, Fin4Cast, Gudrunstrasse 11, 1100 Vienna, Austria, Phone: +43 (0) 517 07 46360, Fax: +43 (0) 517 07 56256, email: [email protected].

The corresponding paper “From Replication to Forecasting – Creating a new and active hedge fund benchmark” is available upon request.

Disclaimer

Page 3: Hedge Fund Replication From Replication To Forecasting

December 07

3

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

From indexation to replication – what‘s next? Creating a newand active hedge fund benchmark

Replication of hedge fund returns – does it really work?

The magic behind – how to replicate?

Limits of hedge fund replication – good to know.

Synthetic replication – presenting the results

Conclusion & research outlook

From Replication to Forecasting – Creating a new and active hedge fund benchmark

Page 4: Hedge Fund Replication From Replication To Forecasting

December 07

4

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Replication of hedge fund returns – does it really work?

The replication of hedge fund returns aims:

to deliver similar month-to-month returns to a particular hedge fund style

to replicate the statistical properties of a particular hedge fund index

to separate the hedge fund alpha of a particular hedge fund style from the traditional and the alternative beta

to lower cost and provide greater transparency and liquidity

to provide benchmarks for investments in hedge funds

to provide liquid underlings for structured products

to eliminate single-manager risk and style drift

to provide access for a larger number of investors

Page 5: Hedge Fund Replication From Replication To Forecasting

December 07

5

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Replication of hedge fund returns – does it really work?

Strong evidence from the recent academic research that a large portion of hedge fund returns can be synthetically replicated through a dynamic long/short portfolio of tradable liquid futures:

Mechanical rule-based trading

(Fung and Hsieh, 1997) use look-back straddles to replicate a trend following strategy mechanically

Products: Merrill Lynch Equity Volatility Arbitrage Index, Merrill Lynch FX Clone, Deutsche Bank Currency Return Index, and Bear Stearns “Mast” (Fixed Income) Index

Multi-factor modeling

(Schneeweis et al, 2003) introduce futures and options as observable factors and replicate the return process of various hedge fund strategies

(Jaeger and Wagner, 2005) estimate factor models to model the underlying hedge fund risk premiums using a broad set of risk factors and (non-linear) rule-based strategies

(Hasandhodzic and Lo, 2006) estimate linear factor models to replicate individual hedge funds using six common factors corresponding to liquid exchange traded instruments

(Fung et al, 2006) estimate a seven-factor model for fund of funds using one traditional and six alternative factors

(Gams, Kuehrer and Mitev, 2007) introduce an integrated and dynamic two stage multi-factor approach to replicate the month-to-month returns of HFR Hedge Fund Index.

Products: Goldman’s Absolute Return Tracker index (GS-ART), Merrill Lynch Factor Index, JPMorgan Alternative Beta Index (ABI), Deutsche Bank Absolute Return Beta Index. Partners Group’s Alternative Beta

Copula-based algorithm

(Kat and Palaro, 2006) use a copula-based approach to design trading strategies that generate returns with predefined statistical properties similar to those of hedge funds or hedge fund indices

Page 6: Hedge Fund Replication From Replication To Forecasting

December 07

6

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Replication of hedge fund returns – does it really work?

HEDGE FUNDS

• Absolute Returns

• Non-Directional

Returns

• Diversification

Benefits

• Distinctive Risk

Profile

DIRECT INVESTMENTREPLICATION APPROACHES

FEASABILITY

• Capital

Requirements

• Long Holding

Periods

• Management and

Incentive Fees

• Legal Requirements

• Transparency of

Risks involved

• Disclosed Hedge

Funds

• Funds of Funds

• Mechanical rule-

based Trading

• Multi - Factor

Modeling

• Copula Approach

• Hedge Fund

Returns are

Available

• Opportunities on

International

Markets

• Determine Return

Driving Factors

• “Reverse

Engineering”

Page 7: Hedge Fund Replication From Replication To Forecasting

December 07

7

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Liquid Futures Factor Pre-Selection Replication Methodology

• RWOLS (Restricted and Weighted Ordinary Least Squared)

• RWLAD (Restricted and WeightedLeast Absolute Deviation)

Factor Selection - Search Algorithms (Best Descriptive

Models)

• Heuristic Search Algorithm

• Greedy Forward Search

• Cross validation

• Fast Stepwise Local Search

Dynamic Portfolio Construction

• Multimodel Inference –

Weighted Average

Approach

• Dynamic Leverage Factor

Dynamic Selection of the best Replication Strategies

Measuring the Results

• Average Strategy

• R-squared Selection Strategy

• Tracking Error Selection Strategy

• Absolute Deviations Selection Strategy

• Statistical Properties

• Replication Accuracy

• Stable Portfolio Development

• Distribution Features

The magic behind – how to replicate?fin4cast two stages integrated multi-factor Hedge Fund Replication Approach

• 45 Traditional Factors and 3 Spreads

13 Commodities9 Stock Indices1 vola Index

6 Bond Indices11 Currencies5 Money Markets

• 6 Alternative FactorsMechanical Trading

Rules (MTRs)

• Reuters• Thomson Financial• Bloomberg

Sub-pool 154 Factors

Sub-pool 227 Factors

Sub-pool 322 Factors

Page 8: Hedge Fund Replication From Replication To Forecasting

December 07

8

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Limits of hedge fund replication – good to know

Replication of hedge fund indices (average return of hedge funds), but NOT a single hedge fund

Quality of replication vary considerably among different hedge fund styles, i.e. Global Macro, Long/Short Equity or Market Neutral

Replication results very among different providers of hedge fund indices, i.e. HFR or CS/Tremont

Quality of replication suffers from:

the lack of liquid instruments to replicate specific risk premia, i.e. emerging market and M&A

the time lag to adjust the model’s coefficients with respect to “external shocks” and regime switches

the time lag of the data availability, i.e. 15th of each month

the low frequency of the available data, i.e. monthy returns

the short history, i.e. just 167 data points since January 1994 for CS/Tremont Hedge Fund Composite Index

Page 9: Hedge Fund Replication From Replication To Forecasting

December 07

9

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Synthetic replication – presenting the results

A negative compound alpha -1.71% for the observed period!

Page 10: Hedge Fund Replication From Replication To Forecasting

December 07

10

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Synthetic replication – presenting the results

Page 11: Hedge Fund Replication From Replication To Forecasting

December 07

11

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Synthetic Replication – presenting the results

Page 12: Hedge Fund Replication From Replication To Forecasting

December 07

12

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

From indexation to replication – what‘s next? Creating a new and active hedge fund benchmark!

Building of forecast models to predict the direction of the monthly returns of the CS/Tremont Composite

Combining the results of the forecast models with the results of the replication models by adjusting the model’s coefficients:

if the return forecast is positive the coefficients stay the same as for the replication model

if the return forecast is negative the coefficients are multiplied by -1

The objective is:

☺ to create a new and active hedge fund benchmark

☺ to out-perform the average of the hedge funds by generating positive returns during periods of negative returns of CS/Tremont Composite

Page 13: Hedge Fund Replication From Replication To Forecasting

December 07

13

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

From indexation to replication – what‘s next? Creating a new and active hedge fund benchmark!

Page 14: Hedge Fund Replication From Replication To Forecasting

December 07

14

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

From indexation to replication – what‘s next? Creating a new and active hedge fund benchmark!

☺ The new active hedge fund benchmark out-performed the CS/Tremont Composite Index by 13.34% during the observed period!

Page 15: Hedge Fund Replication From Replication To Forecasting

December 07

15

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Conclusion

Does replication of hedge funds returns really work?

Strong evidence from the recent academic research supports the motion

Replication approaches – mechanical rule-based, multi-factor and copula

How to replicate?

Replication of CS/Tremont Hedge Fund Composite Index using fin4cast two stage multi-factor integrated Hedge Fund Replication Approach

Good to know:

Replication works for average hedge fund returns, but not for a single hedge funds

Replication quality varies among differnt hedge fund styles and index providers

Replication lags behind due to time lag of the data availability and adjustment of the model‘s coefficients

What are the results?

☺ Our results give strong evidence that the synthetic hedge fund portfolio is able to replicate the statistical properties of the monthly returns of the CS/Tremont Hedge Fund Index with respect to the month-to-month return and the standard deviation

☺ Our findings show that the compound alpha of the CS/Tremont Index compared to the cost and interest rate adjusted returns of the synthetic portfolio is negative

New idea about an active hedge fund benchmark was born:

Combination of return replication and return forecast!

Page 16: Hedge Fund Replication From Replication To Forecasting

December 07

16

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Research outlook

What’s next?

Extensive research for the creation of new and active hedge fundbenchmarks for different hedge fund styles

Building of qualitative mathematical forecast models for different hedge fund indices

Intensive live-testing of new and active hedge fund benchmarks

Page 17: Hedge Fund Replication From Replication To Forecasting

December 07

17

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Universe of liquid futures used for the replication

Page 18: Hedge Fund Replication From Replication To Forecasting

December 07

18

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

ReferencesAckermann, C., McEnally, R. and Ravenscraft, D. (1998); The performance of hedge funds: risk, return and incentives

An, H. and Gu, L. (1985); On the selection of regression variables; Acta Mathematicae Applicatae Sinica, Vol. 2, No. 1 (pp. 27-36)

An, H. and Gu, L. (1989); Fast stepwise procedures of selection of variables by using AIC and BIC criteria; Acta Mathematicae Applicatae Sinica, Vol. 5, No. 1 (pp. 60-67)

Burnham, K. and Anderson, D.R. (1998); Model selection and inference: a practical information-theoretic approach, Springer Verlag

Crepaz, M., (2007): Replication of Hedge Fund Returns, Diploma Thesis, Vienna University of Economics and Business Administration.

Dorfleitner, G., (2003): Why the return notion matters. International Journal of Theoretical and Applied Finance, Vol. 6, No.1, pp. 73-86, 2003

Fung, W. and Hsieh D. (1997); Empirical Characteristics of Dynamic Trading Strategies: The Case of Hedge Funds; The Review of Financial Studies, No. 2, (pp. 275-302)

Fung, W. and Hsieh D. (1999); A Primer on Hedge Funds, Journal of Empirical Finance, 6, (pp. 309-331)

Fung, W. and Hsieh, D. (2004); Hedge Fund Benchmarks: A Risk Based Approach. Financial Analyst Journal

Fung, W., Hsieh D., Naik, N. and Ramadorai, T. (2006): Hedge Funds: Performance, Risk and Capital Formation.

Gams, K., Kuehrer, M. and Mitev, M. (2006); Hedge Fund Replication using Fin4Cast Technology, Siemens Fin4Cast Working Paper

Gams, K., Kuehrer M. and Mitev, M. (2007): Synthetic Replicaton, The Hedgefund Journal, October 2007

Hasanhodzic, Jasmina and Lo, Andrew W. (2006): Can Hedge-Fund Returns Be Replicated? The Linear Case.

Jaeger, Lars and Wagner, Christian (2005): Factor Modeling and Benchmarking of Hedge Funds: Can passive investments in hedge fund strategies deliver?, Journal of Alternative Investments.

Kat, H. and H. Palaro (2005); Who Needs Hedge Funds? A Copula-Based Approach to Hedge Fund Return Replication, Working Paper 27, Alternative Investment Research Centre, Cass Business School

Kuehrer, M. and Mitev, M. (2007); Forecasting the future return of the oil price, The Hedgefund Journal, May 2007

Schneeweis Thomas, Kazemi Hossein and Karavas Vassilis (2003): Eurex Derivative Products in Alternative Investments: The Case for Hedge Funds.

Page 19: Hedge Fund Replication From Replication To Forecasting

December 07

19

s

From Hedge Fund Replication to Hedge Fund Forecasting

Dr. Miroslav Mitev

Biography

Dr Miroslav Mitev is a managing director and head of quantitative securities research and portfolio management. Dr Mitev is responsible for the development of innovative, systematic long-short investment strategies for institutional investors world wide based on Siemens/fin4cast technology. After joining Siemens in 2001 Dr Mitev successfully formed a qualified team of 25 professionals which is continuously developing the Siemens/fin4cast Technology and building mathematical forecasting models for a variety of financial instruments like currency futures, commodity futures, stock index futures, bond futures, single stocks and hedge fund indices. Dr Mitev is in charge of the Siemens/fin4cast’s research cooperation with various universities and is actively involved in the scientific management of numerous master thesis and dissertations. Dr Mitev is a regular speaker at international conventions on liability driven investing, asset management, hedge funds, portable alpha, advanced quantitative studies, algo-trading and system research. Dr Mitev’s research is published on a regular basis in international journals and presented on international scientific conferences.

Prior to joining Siemens Dr Mitev was at CA IB, the Investment Bank of Bank Austria Group, where he was in charge of the quantitative research of the securities research division.

Dr Mitev received a Master of Economics and Business Administration with main focus on Investment Banking and Capital Markets. Dr Mitev also received a PhD in Economics with main focus on Finance and Econometrics.

Dr. Miroslav MitevSiemens AG ÖsterreichSiemens IT Solutions and Services PSE/fin4castPhone: +43 (0) 51707 46253Fax: +43 (0) 51707 56465Mobile: +43 (0) 676 9050903Email: [email protected]