Risk management is integral to Brandywine’s investment programs and has been since the development of Brandywine’s portfolio allocation model in the late 1980s.
The original concept for Brandywine’s portfolio allocation model was triggered by the realization that the single most important consideration of any model should be to focus on producing the most predictable performance – meaning that future performance should mimic past performance, whether actual or tested, as closely as possible. (This does not mean that Brandywine will produce predictable performance, only that predictable performance should be the goal of the portfolio allocation model. In fact, PAST PERFORMANCE IS NOT INDICATIVE OF FUTURE PERFORMANCE AND THERE IS THE RISK OF LOSS WHEN INVESTING WITH BRANDYWINE.)
In contrast, the portfolio allocation models of that time, and which remain in dominant use today, are designed to produce “optimal” performance on historical data. Known as “mean-variance” or “Markowitz” modeling, these models produce a perfect curve-fit on historical data but fail to provide any predictability that future performance will match that past performance. (This is not a failure of design. In fact, by focusing on optimizing past returns rather than future predictability, they were designed to fail in just this way!)
What Brandywine discovered more than 25 years ago was that in developing the model to produce predictable results, a natural artifact was that it also minimized portfolio risk. That is because the primary characteristic that produces the most predictable performance – creating a balanced, diversified portfolio – also reduces risk by decreasing the exposure to any single return driver or market.
Because Brandywine’s portfolio allocation model is the unique intellectual property of Brandywine and a significant contributor to our performance differentiation, its specific structure is not revealed. But we can reveal that there are a number of key elements necessary to create portfolio balance, including:
- A clear understanding of the return drivers powering each trading strategy. It is not sufficient to merely focus on historical performance, but to understand the basis for each strategy’s returns.
- Event risk associated with each trading strategy.
- Correlation and assessment of future correlations among all trading strategies.
- Event risk associated with each market.
- Correlation and assessment of future correlations among all markets.
- Time-in-market for each strategy/market combination and time-in-market relationships among all strategy/market combinations.
The end result of Brandywine’s portfolio allocation model is a portfolio designed to be balanced across all strategies and markets, such that they will each provide an equitable contribution to the portfolio’s risk and return over time.