Multi Factor Scores

In the previous articles, we discussed how we created single factor scores for value, momentum and quality factors. In this article, we will extend the idea to create multi factor scores which are a combination of 2 or more factors.

The idea behind multi factor scores is to see how a stock scores on more than one factor and base your investment decisions on that. One way to do this is to combine multiple factors to arrive at a single score which we will discuss in this article. The other way is to classify stocks based on their single factor scores into different “boxes”. We will discuss this in our style classification methodology called StyleBoxes.

Creating multi factor scores is based on our standard score-aggregate-score methodology. When combining 2 or more factors, add their respective score to get an aggregate score, then rescore the stocks based on aggregate scores to get the final multi factor score between 0 and 100.

We have 4 multi factor scores on sharpely

• QV score – single score combining Quality and Value scores
• QM score - single score combining Quality and Momentum scores
• VM score – single scores combining Value and Momentum scores
• QVM Score - single score combining Quality, Value and Momentum scores.

A potential drawback of using multi factor scores

While multi factor scores can be very useful in visualizing a stock’s factor tilt across multiple factors, it can, at times, be misleading.

This will generally happen when a stock scores very high on one factor and average on other factor or factors. For example, if a stock scores 100 on value but 65 on momentum, VM score of the stock will be above 70 (our internal cutoff for deciding if the score is “good” or not). 

But the stock only has average momentum and there could be other stocks with “good” value (not 100 but 90) and “good” momentum. To overcome that drawback, we have StyleBox classification. But it comes with its own set of limitations.

Why no “growth” factor?

By now, readers would have realised the conspicuous absence of “growth” factor and hence a “growth” score.

While we do not have a dedicated growth score, profitability growth is used in quality score and earnings momentum is used in momentum score.

As per empirical evidence, historical earnings growth, as a standalone factor does not add much value. As such, we do not have a separate factor score for growth. Rather, we extract the “valuable” aspects of growth for quality and momentum score calculation.