The use of big data in investment management has been revolutionary, but harnessing its potential is the next big challenge for the active asset management industry. Big data is the residue of information that we all leave behind as we buy things, sell things, browse the internet, use our smartphones and generally live our lives. Alpha, or excess returns, is not in the data itself, it is in how data is processed and creatively interrogated. Big data without data science lacks any power in uncovering insights that can deliver real alpha.
It all depends how the data is used
Despite the buzz that big data is generating within the research-driven world of finance and investment, achieving an integrated approach to using it is far from easy. Alternative data can be opaque or even misleading. Amid the hype, many will forget that big data is of little use without the combined insights of experienced data scientists and investment professionals.
So-called ‘big data’ is proliferating. At the same time, advances in cloud computing, machine learning and artificial intelligence allow extraction of coherent, strategic insights from these digital residues. Combined, as data science, they have the potential to be a richly-enhanced source of information about our world. It is information that is deeper and more detailed than we have ever had before, and yet also broader and more comprehensive.
In simple terms, data science potentially brings unique insights into the inner workings of a company under diligence. Most investors wait for quarterly earnings updates or an occasional meeting with an executive. Data science now enables tracking in near real-time the sales of a company and its competitors, the morale of its employees and their view of the CEO, how a new product launch is being received in a given market, research and development budgets, as well as numerous other key performance indicators.
These insights, sometimes unknown even to the company itself, give an advantage over the competition as they allow for a differentiated view of the company’s earnings power, or at the very least minimise impairment risk. Data science often provides insight into the operation of a business at a deeper level than is communicated more broadly, and in public reports.
Examples of big data opportunities
For example, credit card data is a popular form of alternative data but often it is used to infer total top line revenue in a quarter. This can contain significant bias and errors. Data science provides methods to detect and correct bias. In addition, statistical methods can be used to reveal demographics and psychographics in the data. Classifications of the data panel members by gender, age, urban density, income band, and buying preferences, etc. can be revealing about the cohorts and composition of the customer base and growth areas of a business. When evaluating a company it is important to know if the growth is from more customers, or from more loyalty spend by the same customers.
Online activity is another source of data science insight. Searching for something that a customer intends to buy is a leading indicator of the actual purchase. This leading indicator is particularly useful for higher cost items that are often considered for weeks before a purchase. Online transactions also provide insight into the competitive environment that the company of interest operates in, as well as a measure of the advertising spend in various channels.
Job listings provide a wealth of insight into the way that companies are growing, but this involves the use of data science methods to analyze the text of the job description. The descriptions show the areas and geographies where the company is growing, and within the text the company often signals the products from other business that they prefer to use.
Six categories of opportunities
Overall, many of the data science research hypotheses fall into one of the following six analytical categories:
Three years ago, we added a data science capability to its team of nearly 650 investment professionals. While many in the firm have come to quickly embrace data science, the recently-launched Global Equities Data-science Integrated strategy (or ‘GEDI’ for short) is the first to fully integrate data science with fundamental research and ESG engagement.
Of course, different strategists working to different time horizons will make different buy and sell decisions, but whatever they do, they will now be informed, to a greater or lesser extent, by the additional depth and breadth that these data science insights bring.
State Super’s senior investment manager, Andrew Huang, said of these new data-based techniques:
“To us, Neuberger Berman appears to be meaningfully ahead of the curve in building and implementing this approach. We believe that supporting fundamental research with a solid data-science and ESG discipline will be a growing advantage over time”.
Data science is not a replacement for traditional investment research, but a complement that brings a fresh and sometimes counter-intuitive perspective. It is not a technology support function for investment professionals but an extension of what they already do. For the same reasons, simply hiring a team of data scientists and setting them to work is not necessarily going to enhance an investment manager’s search for alpha. Finding a common language with which to integrate that team into the existing research flow that investment teams generate is critical.
Ultimately, we believe those who engage with big data seriously and ethically will find it transformative.
Michael Recce is Chief Data Scientist at Neuberger Berman, a sponsor of Firstlinks. This material is provided for information purposes only and nothing herein constitutes investment, legal, accounting or tax advice, or a recommendation to buy, sell or hold a security. It does not consider the circumstances of any investor.
For more articles and papers by Neuberger Berman, please click here.