Big Data Offers New Insights into Financial Markets

Creative Investment Professionals Can Gain an Edge

Stock-Market

The age of big data is revolutionizing all business fields, from marketing to investment management. It is also rapidly changing the practice of empirical research.

This new research approach featured prominently in a symposium on “Information and Asset Prices” that I recently organized with Professor Joseph Chen. The symposium focused on how information is processed by financial markets, and it brought top global scholars in the field to UC Davis.

The symposium highlighted the trend among finance researchers toward relying on non-standard, hand-collected, datasets and on new empirical methodologies, such as linguistic analysis.

The study of financial markets has been dominated by the belief in the weak form of market efficiency, which was proposed by the 2013 Nobel Prize winner Eugene Fama. This is the hypothesis that asset prices reflect all publicly available information.

However, research presented in the symposium showed not only that public information is not being efficiently incorporated in asset prices but also that investors do not even necessarily consider all relevant information when making portfolio decisions.

In fact, when investors do analyze public information, their inferences may diverge, and their judgments may be clouded by their moods, referred to as sentiment in the finance literature. Evidence from the symposium clearly supported the view of Fama’s 2013 Nobel Prize co-winner, Robert Shiller, who views markets as irrational and prone to sentiment.

Trading on Proprietary Information

The papers presented by Dmitry Livdan of UC Berkeley and Ohad Kadan of Washington University were closely related in that they showed that institutions profitably trade on information before it enters the public domain (these papers were co-written with Terrence Hendershott and Norman Schurhoff, and Roni Michaely and Pamela C. Moulton).

This research left open the question whether institutions obtain their information advantage through legal or illegal channels. After all, we know from the popular press that many hedge funds also employ big-data techniques, for example, when predicting earnings surprises of big retailers by collecting data on the number of customers in their local stores, when using information on Internet searches for company’s products to predict demand, and so on.

On the other hand, the very public Securities and Exchange Commission cases against Raj Rajarantham of the Galleon hedge fund and Steven Cohen of SAC Capital showed that hedge funds sometimes obtain their information illegally from corporate insiders.

Investors Largely Ignore Corporate Financial Reports

Tim Loughran and Bill McDonald of the University of Notre Dame analyzed the pattern of searches for corporate financial disclosures, such as annual or quarterly reports, mandated by the SEC.

Looking at Internet search patterns allows researchers to directly observe what information investors consider when making portfolio decisions. To their surprise, they found that very few investors download corporate financial reports directly, seemingly preferring to rely on second-hand information sources, such as financial analysis. Much important information may get lost in the process. Even more surprisingly, the number of downloads declines dramatically with the passage of time.

This implies that investors do not consider long-run trends, or compare corporate strategic plans outlined in the reports with the future realizations, when deciding on investment allocations.

There may be money left on the table for the investors who have the patience to study corporate financial disclosures with full attention.

In his work on bubbles, Robert Shiller has argued that the media helps perpetuate bubbles by amplifying investor sentiment. It does so by reporting on assets that have recently increased in price and by writing stories on investors who have made money by investing in such assets, be it real estate or the market index.

Boom and Bust Patterns

Cindy K. Soo of University of Michigan calculated the sentiment of local media coverage of local housing markets. She then used these measures to test the role of sentiment in the run-up and crash of housing prices that instigated the great financial crisis of 2008. Just as Shiller hypothesized, she found that the sentiment from local media was able to forecast the boom and bust pattern of house prices at a two year lead, both in aggregate and across cities.

How is sentiment formed? Diego Garcia of University of North Carolina conducted a meticulous study of this question looking at what influences the sentiment of stories about the broad stock market. He concluded that the tone of market coverage is strongly influenced by negative events in the days prior, but not as much by positive events. He suggested that this asymmetry stems from the demand of readers, who are very concerned about limiting their losses (the so-called loss-averse preferences).

Rose-colored Glasses?

What if investors look at the same information, but disagree on how to interpret it? Previous research showed that prices will be overpriced relative to the average opinion because short-sale constraints and the straight-up unwillingness to hold short positions prevent pessimistic investors from expressing their opinions.

Dong Lou of London School of Economics and his co-authors Byoung-Hyoun Hwang and Chengxi Yin explained why the overpricing is not there for portfolios of assets, be it closed-end funds or corporate conglomerates. These portfolios often trade at a discount relative to their components because investor disagreements about the component assets may offset each other (i.e., investor 1 may be optimistic about stock A but pessimistic about stock B, and investor 2 may be pessimistic about stock A but optimistic about stock B; thus, the value of a closed-end fund that holds stocks A and B will truly reflect the average valuation of both stocks A and B).

The six papers were discussed by Lu Zheng of UC Irvine, Paul Irvine of Texas Christian University, S.P. Kothari, of MIT, David Solmon of University of Southern California, Chris Parsons of UC San Diego, and Karl Diether of Brigham Young University, respectively. The discussants, all top experts in the field of empirical asset pricing, offered constructive criticisms of the papers and their views on how to move the field forward.

Big Data Can Mean Big Advantage

The age of big data offers new creative opportunities across the board. For finance researchers, big data allows the field to settle old debates and discover new puzzles. For investment practitioners, the big data approach offers limitless possibilities to gain informational advantage over the competition by cleverly analyzing public data sources and even utilizing the Freedom of Information Act to obtain the data that is hidden but belongs in the public domain.

The more creative money managers are in using the data, the less competition they will face from like-minded traders, and the more stable and long-lasting their profits will be.

When it comes to the field of investing, there is little glory in blending in with the crowds and a huge payoff to thinking differently and creatively. The key to success is asking interesting questions, knowing how to obtain relevant data, and proficiency with various data analysis techniques.

These skills will be in high demand for many years to come.



Leave a Reply