Oct 7, 2018
Today, we're going to discuss the concept of a Factor Mimicking Portfolio (FMP). These portfolios are constructed to isolate the exposure of a single factor within a given model, so that they have unit exposure to that factor and no exposure to any other factor in the model. When we describe the returns of a specific factor, we're essentially looking at the weighted average returns of these long/short FMPs, because they are built to solely exhibit the characteristics of that factor. Thus, a 1% move for a factor in the market would equal a 1% move for the underlying FMP.
We can use Omega Point's portfolio decomposition tools to learn about the Size factor's composition and return characteristics by analyzing its FMP. As a reminder, Size factor differentiates between large and small stocks and is the natural logarithm of the total issuer market cap, averaged over the last month. Please note that the Size FMP is not only weighting large cap names, but identifying the large cap companies that also have the least exposure to all other factors.
This is what the net style exposures look like for the Size FMP on 10/3/18:
Since the beginning of July, Size factor has staged a rally, up 3.6% on a cumulative basis during that time.
Diving into the performance contribution of the underlying FMP, we can see what stocks have been driving the price movement for this factor. The top 10 names that drove about 30% of the total performance are below, along with their average weight in the FMP and the % return over this period.
Size factor composition - 10/3
We can also utilize the Size FMP to get a better sense of which industries are most represented in the factor. While the FMP is inherently built to be net neutral across all factors other than size, we can look at the gross weights for each industry.
As of 10/3/2018, the top 10 industries below represent almost 50% of the overall gross weight in the Size factor:
We'll be taking a look at the FMPs for other known factors such as Value and Momentum in coming weeks. If you'd like to discuss FMPs, or would like to better understand how we measure the relationships between factors, please don't hesitate to reach out.