Wednesday, July 30, 2025

The big divergence in confidence

 


What is surprising is the large difference between the confidence of institutional investors versus individual investors, according to the International Center of Finance at Yale University survey. There is some lag with the data. The most recent data comes from May, yet it shows that institutions are very bullish while individuals are moving in the opposite direction, albeit recent data has demonstrated improvement. While this data shows unusual behavior in the last year, the gap is surprising. Nevertheless, this survey data with information on valuation suggests that the market can go higher - expectations are not discounting higher valuations


The power of combining value and momentum factors

 


One of the more interesting combinations of factor exposures is value and momentum. Value and momentum have both delivered strong excess returns, and blending them together also has strong benefits, even though the value and momentum strategies are negatively correlated. 

An interesting new paper looks at the combination by trying to explain the excess returns through a pricing kernel that uses nine latent variables from the combined value and momentum cross-section, see "Value and Momentum Leftovers". The combination has proven to be much better than investing in each factor by itself. However, there is an issue of what the pricing kernel should be to help combine these two factors because there is a significant alpha after trying to price the combination. Additionally, the idea that you can just blend the value and momentum as a 50/50 combination may not be effective. 

We will not go into the details of this "leftover" argument other than to say that the use of latent variables can help price these two key factors and can lead to better factor combination outcomes. The blend of value and momentum can be improved through well-defined pricing techniques. 




Monday, July 28, 2025

Thematic investing - How does that work?

 


One of the most significant trends in equity investing is the development of thematic portfolios; however, there has been limited testing of this concept. Thematic investing is presented in a model framework in the new paper, "Thematic Investing: A Risk-based Perspective". 

In the paper, the authors cite several definitions: “Themes are structural trends expected to significantly impact economies and redefine business models.” .... “Thematic investing can be seen as an additional dimension in portfolios.” and  “.... themes may pertain to macroeconomic or structural trends that transcend the traditional business cycle.” 

More formally, their model looks at the residuals for a set of assets. Cross-sectionally, these residuals should be uncorrelated after accounting for their main factors. In the case of a theme, there should be positive pairwise correlation across the set of assets identified in a theme. There is a relationship that cannot be accounted for by other known factors. Surprisingly, the pairwise correlations are relatively low within a theme. A theme is a set of assets that are identified by a narrative with a common trend. It is trend-following associated with a narrative.

Many identified themes do not prove to be correct, so there is an art to forming themes and not a science.




Sunday, July 27, 2025

Choose your modeling technique wisely

 



One of the key challenges for any quantitative modeler is selecting the appropriate modeling technique. This involves not selecting the features or predictors to be included in the model, but rather the statistical techniques that will generate the best predictions. Not all modeling approaches are the same, so knowing when to use the best techniques is an important skill. The paper "From OLS to neural nets: Comparing regression models in systematic trading" is not the final word on this question. Still, it does a good job of making comparisons across techniques. It does not address all the machine learning techniques, yet it does compare many of the simple techniques that most researchers may employ. 

The tables from the paper are nice to have handy for anyone starting out when thinking about modeling.