What Constitutes a Robust Momentum Factor Model?

What Are the Attributes of a Robust Momentum Model? While smart beta strategies aim for market beating performance, it doesn’t always work out that way. A strategy is only as good as the expertise of the team constructing the model as well as how rigorously it was tested against actual market conditions. Momentum Factor Models are not all created equal. Many are rooted in theories dating back over 25 years. Although they are capturing the essence of momentum, they are late by definition in capturing reversals. The resulting effects are late entries in bull trends and late exits when a negative trend reversal occur. The optimal model is built on a strong foundation which can also adapt quickly to change.

A Robust Momentum Factor Model:

Identifies Trends Quickly: The sooner an investor can identify momentum, the longer they can participate in a trend.  Some widely implemented momentum models require 6 to 12 months of validation on a stock before including it.  Trendrating can identify and rate the momentum of a stock within 1 to 2 months.

Multiple Indicators: A legacy model used to power some of the world’s largest momentum ETFs relies on a single indicator (Relative Strength). This model can identify trends quickly but has not fared well in volatile markets. Trendrating is built on a number of complementary indicators, which make it more stable.

Speed of Adjustment: Momentum exists in many cases and goes unnoticed by models that only rebalance quarterly or bi-annually.  Speed of analysis is crucial to promptly fine-tune the index components.  Trendrating adjusts our model components monthly so that it identifies opportunities faster.

Breadth & Depth: A number of asset managers try to build their own momentum models. This is a major undertaking that requires years of time, a substantial investment, but more important the right skills and expertise on how to build and test a model. Trendrating offers a robust model that is systematic, repeatable, and well validated via massive testing without the drawbacks of human interpretation of charts and subjective picking of technical indicators.

Leave a Reply