Factor Modeling

Although quantitative methods for fund analysis have been around for decades, they tend not to work very well on hedge funds. They need lots of data, and they require funds to behave consistently over time.

Hedge funds, on the other hand, often have short histories. They report returns infrequently and they source their risks from numerous factors. They usually have highly dynamic and non-linear risk exposures. All of this makes empirical analysis very difficult.

Our algorithms overcome these obstacles by augmenting traditional statistical techniques with advances in machine learning. They often produce substantial improvements over conventional approaches.

Why should you care? Our methods can help you evaluate more funds, faster. They allow you to expand your investable universe and design more efficient portfolios. They also facilitate analyses that have previously been either difficult or impossible to perform, such as:

  • Identifying sources of risk and return
  • Generating near real-time, intra-month performance projections
  • Simulating returns prior to a fund’s inception
  • Predicting future returns, correlations, and volatilities
  • Determining the value of any particular fund to its portfolio
  • Optimizing portfolios based on factor exposures

Composite Hedge Fund Indexes

Hedge funds are difficult to benchmark. Managers typically report to only one or two databases, if any. Although several indexes may exist for any given strategy, each measures only a small subset of the total hedge fund universe.

The Eqira Composite Hedge Fund Indexes are indexes of indexes that are designed to reduce the bias inherent in provider indexes. By weighting each component index by its sensitivity to a common factor, we produce indexes that much more accurately represent true strategy performance.

We provide composite indexes for 30 hedge fund strategies, most of which have history dating back to the early 1990s.

Intra-Month Return Projections

It is exceptionally difficult to get reliable estimates of hedge fund performance in real-time. Funds and indexes provide returns infrequently, often with considerable lag, and the few daily indexes that do exist exhibit enormous tracking error. Eqira’s Factor-Based Projections tackle this problem head on. We provide daily, model-generated return estimates for 30 hedge fund strategy indexes, as well as custom assets and portfolios of our clients’ choosing.

We derive our projections using our database of Market Factors.  Doing so allows us to produce forecasts that are significantly more accurate at projecting actual hedge fund performance than investable index returns. Furthermore, since we aren’t subject to the same limitations as investable hedge fund indexes we can generate projections for many strategies they don’t cover.

Market Factors

Identify the risks worth taking and those worth leaving behind using our collection of proprietary factor-mimicking time series, which systematically capture the underlying forces moving markets. Our database tracks the returns to thousands of traditional and alternative investment strategies across most major asset classes.

Through Eqira’s factor-based lens, you’ll see assets for what they truly are: bundles of risk exposures. By decomposing assets into their component parts, we make it easier to understand security behavior, identify relationships, and draw useful inferences. Our Market Factors allow us to discover hidden characteristics of even the least transparent hedge fund strategies.

Factor-Based Optimization

Mean-variance optimization traditionally relies on three inputs: estimates of return, risk, and correlation. However, knowing that two securities are correlated is much different from knowing why two securities are correlated.

We’ve built an optimizer that takes a much more granular approach to portfolio construction. It optimizes portfolio exposures based on each security’s underlying factor structure, which allows us to penalize securities for their exposure to undesirable risk factors. Doing so allows us to produce portfolios with low correlations to traditional and alternative risk factors, while still maximizing expected performance in consideration of client-specific risk tolerances.

© 2017 Eqira LLC. All rights reserved.