In a previous post , I explained how we think of our investment process in terms of three stages: Where, Why, and How. Today, I’d like to delve more deeply into the How—specifically, the tools we use to dynamically allocate risk over time.
As I explained, we use two proprietary, forward-looking risk models to guide portfolio construction: an Equilibrium risk model, which encompasses a long-term view that can be thought of as a “normal” state of risk, and an Outlook risk model, which is a “distorted” version of our Equilibrium risk model that we believe is consistent with the current risk environment (because a true normal state of risk rarely exists).
The Starting Point: Valuation-Based Allocation
As a starting point to portfolio construction, we use a proprietary portfolio calibration tool called Valuation-Based Allocation (VBA). This tool incorporates our forward-looking views on risk, correlations, and value-to-price profiles for each market and currency in our universe—not only in isolation, but also in relation to every other market and currency in our universe. We call this a “matrix” approach. The result is a portfolio of suggested risk exposures based on expected contributions to both return and risk and the specified risk budget for the portfolio.
The primary reason we use VBA instead of an optimization approach is that VBA allows us to allocate risk appropriately over time. An optimization approach can indicate the best return/risk trade-off at a point in time, but it does not help with how much risk should be taken at any point in time or with how to consistently allocate risk over time.
The resulting suggested exposures in each market and currency are only the first step in portfolio construction—many aspects of the Why stage of our investment process are qualitative and not captured in the VBA process. We embrace the notion that it is impossible to model every single aspect of our analysis. The setting of actual portfolio exposures, therefore, may at times result in significant deviations from the signals suggested by VBA.
We do not target a constant level of risk over time, but dial risk up and down based on the magnitude of fundamental opportunities we observe, as well as our assessment of Why those opportunities exist.
The Importance of Dynamic Risk Allocation
Importantly, we embrace the approach of dynamic risk allocation. That is, we do not target a constant level of risk over time, but dial risk up and down based on the magnitude of fundamental opportunities we observe, as well as our assessment of Why those opportunities exist. All else equal, we want to take more risk when larger fundamental opportunities arise.
In addition to dynamically managing risk at the portfolio level, our team also dynamically allocates risk within the portfolio—among equities, fixed income, and currencies—over time. This stands in stark contrast to many other approaches, which seek to take an equal amount of risk across portfolio components and/or keep the total strategy’s risk level constant through time.
In fact, we rely on four distinct risk budgets (systematic, unsystematic market, currency, and total) as we allocate risk capital over time. If there are more (or larger) fundamental opportunities in currency, for example, then we will dynamically adjust exposures to take more currency risk. Ultimately, all of our strategies have an expectation of an average level of volatility that lies roughly at the midpoint of the strategy’s total risk budget.
We also make explicit decisions about the use of optionality and the convex/concave profile of the strategy at any point in time. There are times when we want to be buyers of insurance and employ a more convex profile. Similarly, there are times when we believe a more concave profile is warranted and are happy to be sellers of insurance to capture the valuable premia the market is providing.
Determining Dynamic Risk Level
In sum, the dynamic risk level of our strategy, and from where that risk is coming, can be thought of as an equation:
Where + Why + Risk = How
If prices are far from fundamental value in the Where stage, and there are no headwinds, for example, to price reverting to value in the Why stage, then we will likely lean toward taking above-average risk in the portfolio.
If there are fewer opportunities, the opportunities are smaller, and/or there are headwinds to those opportunities being realized, then we want to take below-average risk. This approach can be summarized in a simple rubric: