AKAnomics: Equity Nowcasting Innovation

Samarth Nagaraj*, Rhea Pandit**, Suchita Sridhara***

As interns at AKAnomics this summer, we’ve gotten a look into Nowcasting by exploring how AKAnomics signals are constructed and how they are used by AKAnomics’ clients. We wanted to summarize our learnings for those new to Nowcasting and the AKAnomics’ Nowcasting process.

What is Nowcasting?

Each of us is familiar with forecasting. However, Nowcasting may be a less familiar concept. The main difference between forecasting and nowcasting lies in the amount of information used to make predictions. Forecasting techniques can predict future events using historical or other relevant data. However, the longer the prediction horizon (the time between an estimate and its predicted event), the lower is the accuracy of the estimate. Nowcasting closes the gap on predicting shorter horizon events with higher precision, by incorporating new sources of information. A concept introduced in meteorology, Nowcasting was brought into macroeconomic predictions over the last two decades.

What does AKAnomics do?

AKAnomics further brings the concept of Nowcasting in finance from the world of macroeconomic predictions to estimating company revenues using available macroeconomic, industry, and company data. AKAnomics analyzes how different industries are changing with each new piece of macro/industry data around the globe. It further examines what changing industry growth implies for the revenues of different companies, and which companies are likely to beat or miss consensus revenue expectations. Consensus figures are based on the expectations of Wall Street analysts, and often mirror the guidance provided by each company on their quarterly performance. AKAnomics has found that being able to identify the disconnect with consensus can significantly help in assessing forward returns.

How are the signals constructed?

AKAnomics company signals represent the likelihood of a company’s revenues being disconnected from consensus expectations. The estimates are currently updated weekly but can be updated in real-time. AKAnomics tracks a company’s revenue, specifically the portion that is sensitive to macroeconomic conditions, namely organic revenues.  Historical organic growth for each company must be built from filings, while total revenue growth is easily available. AKAnomics uses the Nowcasting technique to construct a running view of each regional industry from the macro/industry data, and further matches this information with that of each company’s business segments to derive its signals. It is hence a data-supported fundamental signal creation approach, sometimes referred to as a “quantamental” signal.

How do we measure the efficacy of estimates?

AKAnomics prefers to use “hit rate” as its measure of the efficacy of its signals. Hit rate is a measure of how frequently AKAnomics estimates fall on the same side of consensus as the company reported revenues. Historical data suggests an average weekly hit rate of 66% over the years (higher for quarterly hit rates), and across the 100+ companies it tracks. Additionally, historical data shows a linear relationship between a company’s hit rates and forward returns, suggesting that improving hit rates could enhance stock-picking.

What criteria leads to good signals?

Most investors will consider a signal “good” if it is good in estimating the fundamental indicator it is trying to predict (i.e., revenues), as well as it has good forward-return characteristics. The farther the AKAnomics estimate is from consensus, the more likely it is to rightly identify a revenue beat or miss relative to consensus, and the more likely to offer favorable forward-returns characteristics. The choice of the threshold to determine when a signal is far enough away from consensus to be considered valuable depends on several factors - the threshold must be low enough to yield a substantial number of signals of interest to portfolio managers, and high enough to produce a decent hit rate. Historical data suggests a strong relationship between AKAnomics signals and forward returns, and a choice of threshold that yields ~45% of the estimates as “good” signals.

How to use the signals in portfolios?

AKAnomics helps clients use data-driven decisions either at systematic hedge funds or discretionary hedge funds. AKAnomics’ uniqueness can be understood by examining the relationship between AKAnomics’ ‘quantamental’ signals and other common factors used by fundamental and quantitative hedge fund portfolio managers. Analysis shows low correlations between historical AKAnomics signals and fundamental factors like size and book-to-market value, as well as quantitative factors like momentum and mean reversion. Real-time granular industry-level analysis of macroeconomic data helps AKAnomics differentiate its signals while aiding hedge funds in building new systematic portfolio strategies, or risk-managing their existing strategies.

Does the economic cycle matter?

AKAnomics’ estimates of average revenue growth across tracked companies closely follow both actual revenue growth for these companies, as well as Global Industrial Production growth. AKAnomics’ metrics of revenue surprise dispersion can also serve as an indicator of stock-picking regimes, with dispersion showing a linear relationship to forwards returns. Thus, AKAnomics’ signals are not only uncorrelated to other fundamental and quantitative factors used by clients, but they can be concurrent indicators of the economic cycle as well.

* Samarth is an undergrad at George Washington University

** Rhea is an undergrad at Williams College.

*** Suchita is an undergrad at Wesleyan University.

The authors are all Summer 2024 Interns at AKAnomics Inc.