Risk Driver Analysis

IA’s Factor-Based models are equally useful for funds that only have access to a manager’s returns and for funds that have full position transparency (including the hedge funds themselves).

 
Our groundbreaking returns-based technique identifies meaningful drivers of hedge fund risks. The process has 4 steps:

1. Choose the list of candidate factors.  We offer hundreds of possible factors to choose from in creating the 'starting set' that the system will consider as candidate factors. 
2. Quantitative: systematically build the basket of driving factors.  The system determines which combination of factors best reproduces the returns of the fund or holdings.  This can be done either with the fund's realized returns (if there is no transparency), or with the returns of the current portfolio holdings (if full transparency is available).
3. Qualitative: incorporate manager discretion.  The risk manager can choose to override the system's choices of risk drivers by requiring the system to include or exclude any of the selected factors.  This gives complete discretion to the risk manager, who can even choose to build several baskets of risk drivers for different parts of the portfolio or for use in different simulations.
4. Learn from history.  Once the baskets of risk drivers are established, we put our expertise to work by examining the histories of the factors to identify periods of interest and atypical returns and volatilities.  This analysis then guides our clients when constructing "nightmare" scenarios.

How it works:
The goal of the quantitative analysis is to distill the historical returns (of either the fund or its positions) into a set of meaningful risk “drivers” using a series of mathematical functions. The returns are modeled as:

The returns (R_fund) can either be the realized returns of the actual fund or they can be the rolled-up summary returns of the currently held positions.  This means that the same analysis is equally useful regardless of your degree of transparency.

The results speak for themselves.

The factors or drivers can be individual time series returns, power series, lagged returns, non-linear returns and other more complex factors. In our considerable experience, a fund is typically best modeled with four to seven distinct factors, and the relationship between these factors is often non-linear.

The key to this extraordinary method is to analyze the different parts of the portfolio on equal footing, synthesizing transparency for managers who do not
offer it themselves.

The bottom line: A returns-based analysis that works, measuring the risks of your portfolio even with mixed levels of transparency. No other risk management method brings you this level of intelligent transparency about your risk exposure.