When conditions are not only unusual but evolving at a furious pace, as they have been since the novel coronavirus was identified early this year, clarity can seem an impossibility.
During such times, official data is often a mix of slow and noisy observations ― making it difficult for investors to develop a holistic and clear picture of the state of play. At the same time, many companies have been withdrawing guidance amid heightened uncertainty, making analysis at the industry and company level even more challenging.
Against this backdrop, alternative data and analytical tools take on new importance in identifying emerging economic and market trends that traditional sources can be slow to efficiently isolate.
What is alternative data?
Alternative (or big) data is physical, unstructured (text) or non-financial data generated by the technologies of our everyday lives ― smartphones, GPS and smart home devices, to name just a few. When aggregated and analyzed in the right way, alternative data can provide valuable insights into country, industry and company prospects.
The BlackRock Systematic Active Equity team has been using various forms of alternative data for more than a decade. Some data sources have proved particularly valuable this year, especially when analyzed with the right context in mind. Here we look at three examples and how they have been additive in assessing the macroeconomic and investing environment during the global coronavirus pandemic.
1. Foot traffic
Foot traffic around shopping facilities was one of the original big data metrics embraced by systematic investors early on, though its relevance has gradually diminished as a significant share of shopping has moved online.
Online buying is even more prevalent amid coronavirus-driven closures, yet foot traffic patterns have reasserted their importance in new ways. Differences in social-distancing policies have led to wide dispersion in activity across countries. We find that comparing foot traffic activity at various points of interest has provided a faster read on economic activity across countries and industries. Our readings at the end of May, for example, showed movement in China was more than two-thirds back to normal since re-openings began in March and April.
2. Natural language processing
Natural language processing (NLP) of text has proved especially useful in gleaning insights from analysts, many of whom were relatively slow to update their earnings estimates for the first quarter of 2020. This is understandable, and similar to what we saw during the global financial crisis of 2008-2009, as point-in-time numerical forecasts are difficult in a world with tremendous room for error. While an analyst may take some time to update a numerical forecast, analyzing the text of their reports helps paint a more complete investing picture in the absence of a definitive arithmetic stance.
NLP also has been helpful is getting an early read on fiscal policy, allowing us to parse analyst language for a sense of how policy is moving across countries. We then lean into those with easing tendencies, such as the U.S. The chart below shows the magnitude of change in fiscal policy direction as determined from sentiment measures gleaned from thousands of analyst reports.
3. Job postings
Hiring trends can reveal important information about both industry- and company-level growth prospects. During the COVID-19 crisis, we’ve extended our coverage of job postings to emerging markets where data scarcity would normally make it less informative. But given the velocity of change today, even lower-breadth measurements have proved useful in identifying potential winners and losers. We also have been able to identify firms seeking skillsets that support work-from-home and accelerating digitalization in general. First movers may have opportunity to gain market share versus competitors.
From big data to meaningful insights
We have seen this year ― and over our long history using alternative data ― that different insights tend to add value at different times. Some are more effective around earnings releases, while others pick up on slow-moving trends that play out over several months. We have experimented with timing but found that the most effective approach is having diverse and differentiated insights to inform our investment decision-making.
Ultimately, all alternative data sources have flaws and the goal is to get better, not perfect. Our experience has taught us that having multiple datasets to draw from can reveal inconsistencies and anomalies just as well as potential trends. All of this together helps us to develop a picture of the backdrop before it is evident in the official data. This is the type of information edge than can make a difference when the world and the opportunities and risks in it are changing fast.
Jeff Shen, PhD, is Co-CIO of Active Equities and Co-Head of Systematic Active Equity (SAE) at BlackRock. He is a regular contributor to The Blog.