Hedge Fund Replication by TrueBeta
The first fully independent, fully transparent approach
Methodology Summary                                                                    info@true-beta.com

Introduction

The primary focus of TrueBeta is on creating a composite representing overall hedge fund returns.  We do this with a bottom-up building block approach, based on five underlying sub-strategies (as defined in the HFR hedge fund classification).

The investment objective for the TrueBeta hedge fund replication models is to maximize correlations with the relevant broad (non-investable) HFRI indices while ensuring daily liquidity and a high degree of scalability.


The Truebeta replicators meet these objectives by investing in highly liquid futures instruments or ETFs, creating portfolios that closely match the risk profiles of the five underlying HFRI strategy indices and the HFRI fund weighted composite.

This is accomplished by identifying a maximum of 10 risk factors across asset classes, including currencies and commodities.

Note that it is a specific design objective, while achieving the above investment objectives, that the methodology should be rules based and as simple as possible. We believe the emphasis on a fully rules based approach sets TrueBeta apart from most replication strategies in the market place today.

Apart from the transparency implicit in a rules based methodology, it also enables flexible customization, be it in terms of exposures, or e.g. risk and currency overlays.


Sub Strategies & Composite

The current sub-strategies are

 Equity Hedge
 Event Driven
 Macro
 Relative Value Arbitrage
 Emerging Markets

    These five sub-strategies have been chosen as they (a) are readily replicable using a factor based approach, and (b) they have historically represented approximately 80 % of hedge fund AUM, and 85% of new capital flows.

    The Composite index is created as an equally weighted composite of the five style replicators. The equal weighting has been chosen for its simplicity, as our research indicates that we would not obtain improved results using a more complex weighting scheme.

    Risk Factors

    The Risk Factors are determined applying a set of filters as follows:

    1. A liquidity filter based on the contribution of the strategy to daily volume, and the open interest of the underlying instruments relative to the S&P500 E-mini futures contract (currently max 1% of average daily volume, and at least 5% of the S&P contract)

    2. A statistical filter with constraints for cross correlation, diversification, and variance inflation (to avoid multicollinearity), and finally

    3. An economic sensibility overlay, which is applied if the number of factors exceed  10, following steps (1) and (2)

    The current Risk Factors are:

     S&P 500
     FTSE 100
     Topix
     MSCI Emerging Markets
     US 10 Year Note
     Euro Bunds (10 Year)
     Yen
     Euro
     Gold
     Crude Oil
     (Cash)


    Note that the 10 factors are common to all four 5 sub-strategies.

    Cash is included as a residual investment, meaning that any funds not invested in the risk factors, are invested in a money market instrument.


    Monthly Signals

    The monthly trading signals for the TrueBeta Replicators are generated using a multi-factor regression model with a twelve month rolling window, estimated on a monthly basis (upon the release of the flash update of the monthly HFRI indices on the 5th business day of the month).

    To compensate for the lack of daily data for the HFRI indices, TrueBeta creates a synthetic daily series using an auto regressive ARX model combining the monthly data for the HFRI indices with the daily data for the investable HFRX indices. This significantly improves both reaction time to potential regime shifts, and the stability of the model.

    Dynamic Exposure

    TrueBeta has developed a dynamic leverage model that enables it to explicitly reflect the impact of changes in hedge fund leverage across market cycles. TrueBeta does this by applying a dynamic (variable) multiplier, generated by the model, to the basic factor exposures on a monthly basis. Please go to the Dynamic Exposure section of this web site for further detail.

    Maintenance

    In addition to the monthly signals, the risk factors employed in the regression are reviewed and updated on a semi-annual basis in May and November. The results of the risk factor reviews will be announced on this web site.

    Methodological Enhancements

    TrueBeta conducts on-going research that may result in enhancements to the methodology from time to time. We expect to introduce minor enhancements semi-annually in May and November, and more significant changes (if any) annually in May.