Nowcasting: Comparing Sector Returns

Ajit Agrawal, Suchita Sridhara*, Neeraj Sudhakar

In a previous blog, we highlighted established a relationship between hit rate and forward returns, and made an argument for the hit rate as a measure of AKAnomics’ confidence. In this analysis, we go further to ask the question whether different sub-sectors within the US Industrials & Materials sectors have a very different outcome in relationship to our signals.

Information theory would suggest that sub-sectors which have higher informational asymmetry should have higher forward returns. To normalize for this, we take the ratio between the forward paper returns generated by AKAnomics historical signals per weekly beat/miss event, and that generated by the perfect signal (with no errors). Let’s call this ratio the Return Ratio. We then posit that we should see a linear relationship between sector hit-rate, and the corresponding Return Ratio, as a higher hit rate should imply more accuracy and a larger capture of potential returns available through a perfect signal.

The graph below broadly confirms such a relationship for the 7 sub-sectors we track in our universe of 120 companies, where the sector hit rates vary between a narrow window of 63%-69%. It, however, does have two outliers - Building/Construction, and Transportation. Building/Construction industries tend to have higher informational asymmetry due to a lack of cohesive set of industry-level data to guide investors to the diversity of business drivers, which could potentially explain better (outlier) AKAnomics outcomes. However, better AKAnomics outcomes for Transportation (where high frequency public data is widely available) points to the likely value generated from systematic harnessing of available industry data that most investors are unable to undertake without significant investments.

As before, for our experiment, we identified each event where AKAnomics estimate pointed to a revenue beat/miss for a company at any week. Beat/miss events imply significant deviation of our estimate from consensus (0.75 or more historical standard error). For each such event, we looked at forward 1-week return of that company relative to the XLI (Industrials) Index – (positive for revenue beats, and negative of the returns for revenue misses), and measured the average return per event by sub-sector.

For our experiment, we identified each event where AKAnomics estimate pointed to a revenue beat/miss for a company at any week. Beat/miss events imply significant deviation of our estimate from consensus (0.75 or more historical standard error). For each such event, we looked at forward 1-week return of that company relative to XLI (Industrials) Index – (positive for revenue beats, and negative of the returns for revenue misses), and added up the returns since 2015. For the Perfect Signal analysis, we repeated this hypothetical experiment with the actual revenues known in advance as a measure of sub-sector informational asymmetry.

To round up, we share the average return/event for each of the sub-sectors both for AKAnomics signals, as well as the perfect signals. This graph suggests that there is a large potential for improving returns for the Distributors sub-sector.

*Suchita Sridhara is a Summer Intern at AKAnomics Inc, and we thank her for her contributions. Suchita is an undergrad at Wesleyan University.