Nowcasting Correlations with Quant & Fundamental Factors

Ajit Agrawal, Rhea Pandit**, Neeraj Sudhakar

Contrary to macro Nowcasting which has been quite popular with economists, equity/company Nowcasting concept is quite new to the investing world. AKAnomics has been a pioneer in this field, and clients ask how our “quantamental” signals relate to other known quant and fundamental factors. Here we show that our signals have low correlation with the known quant factors of momentum and mean reversion, and Fama-French (F-F) fundamental factors of size and book value*. This implies that the AKAnomics signals can be significantly additive to quantitative portfolios employing the above factors.

To conduct our experiment, we used the daily F-F factors for size (SMB – Small minus Big market cap) and book value (HML – High minus Low book-to-market value), as well as the quant factors of momentum and mean reversion (Long-Term and Short-Term) obtained from Prof Ken French’s data library. We compared the accumulated returns for the weeks for these series against the returns of a sector-neutral paper-portfolio constructed from AKAnomics weekly signals (extra market exposure hedged with XLI index among the group of ~100 US Industrials names). We note limited correlation of 18%, -1%, 0.3%, -9%, 4% respectively.

We should have expected this low correlation, since AKAnomics factors are quite different from these other factors. Our portfolio of companies includes companies of differing sizes and book values. Moreover, our signals are based not on the company’s past market performance, but on whether the regional industries each company is exposed to are getting better or worse based on our close-to-real-time understanding of the global economy.

The below chart shows the cumulative returns of these fundamental and quant factors and AKAnomics signals. While this is not an apples-to-apples chart since AKAnomics signals use only 24% capital on average (vs 100% for other factors), it shows the cumulative returns for each of the different factors/signals we considered, and shows how the AKAnomics historical signal return patterns differ significantly from other factors.

*https://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html

**Rhea Pandit is a Summer Intern at AKAnomics Inc, and we thank her for her contributions. Rhea is an undergrad at Williams College.