Investors turn to artificial intelligence to find profitable trading signals in corporate word salads

0

We could be at the start of a text investing revolution that could shake up the industry

Content of the article

When United States Steel Corp. published America’s first annual report over a century ago, it was a beautifully cut 40 page, and mostly chimney-piece. Its 2020 annual report had 162 wrapped pages, and even that is concise by modern standards.

Publicity

Content of the article

State-owned companies publish tons of financial and business information every quarter, but the annual report is the big blue whale of corporate reporting. These days, the average length is the equivalent of a 240-page novel, according to S&P Global.

It’s popular to lament that quarterly and annual reports are now salads of words, consisting mostly of copious amounts of often unnecessary reporting requirements and standardized legal caveats, then sprinkled with a big dose of PR. Basic accounting figures have not changed significantly in quantity or quality over the past century, cynics complain.

Content of the article

It is true that corporate reports contain verbiage that would make even a journalist blush. But instead of despising these reports, savvy investors should embrace this admittedly far-fetched textual information as a potential gold mine that can finally be mined with modern technology.

Publicity

Content of the article

Historically, investment has relied primarily on numbers, such as stock prices and earnings, income and research expenses. Traditional stockpickers would naturally complement this with plenty of qualitative analysis, such as interviews with a company’s CFO, discussions with industry experts, and annual reports.

Yet the growing volume of corporate reporting means that no one can realistically consume everything. In the United States, the “risk factors” section of annual reports has nearly tripled since 2006 and now averages over 11,000 words, according to a recent report from S&P Global. Yet even in the most subtle changes, valuable signals are hidden, notes Frank Zhao, analyst with S&P’s Market Intelligence team.

Publicity

Content of the article

The tool for gleaning tradable signals from textual noise is known as natural language processing (NLP), an increasingly popular area of ​​artificial intelligence that involves teaching machines how to read and understand signals. subtleties of human language. NLP makes it possible to systematically collect and analyze at breakneck speeds previously redesigned “unstructured” non-digital data sets.

The potential is immense. Kai Wu, a former GMO analyst who now runs Sparkline Capital LP, a startup investment firm in New York City, says many traditional data-driven trading strategies are ‘exploited’ after being analyzed to death for decades and now exploited to oblivion. .

“But once you’ve crossed the Rubicon into the world of unstructured data, the fruit is suddenly much lower,” Wu said in a post in May.

Publicity

Content of the article

Once you’ve crossed the Rubicon into the world of unstructured data, the fruit is suddenly much lower.

Kai wu

Many sophisticated quantitative investors – those who primarily use algorithms to trade systematically, rather than traditional human fund managers – are already grasping this fruit. Last year, a National Bureau of Economic Research article estimated that algorithmic downloads of quarterly and annual reports in the United States exploded to 165 million in 2016, when they accounted for 78% of all downloads, up from about 360,000 in 2003.

Since 2016, the rate has almost certainly increased further. Industry insiders say there is a near arms race today between NLP algorithms scouring company statements and company executives trying to trick them into avoiding certain tricky words and phrases. .

Publicity

Content of the article

Yet in reality NLP is an opportunity for all investment managers, not just quants who try to systematically exploit text signals. For example, Nomura Securities International Inc. analyst Joe Mezrich used an NLP system to browse transcripts of business executives chatting with financial analysts, rating companies based on their apparent adherence to environmental standards, social and corporate governance (ESG). He found that the stocks of the most ESG compliant companies outperformed the broader stock market.

  1. The US Federal Reserve will tighten its policy according to its timetable, not that of the market, writes David Rosenberg.

    David Rosenberg: The Truths Bothering Investors About Economic Recovery

  2. Yield-hungry investors in Canada are in love with bank stocks.

    Why Investors Should Focus On Generating The Best Total Return, Not The Best Return

  3. Whenever a promising industry takes hold in the media, it rarely takes too long before new investment products are developed and marketed to investors.

    FP Answers: What place should cannabis and cryptocurrency ETFs have in my wallet?

  4. The financial planning and advisory industry might well consume half of your returns while you take all the risk - and even that can be bullish.

    Yields aren’t what they used to be, but investor expectations remain stubbornly high

Publicity

Content of the article

Quarterly and annual reports are now generally published in a user-friendly format, but they are just the tip of the iceberg of written information that investors can dig for valuable signals. Transcripts of management calls with analysts or TV interviews with CEOs, newspaper articles, central bank speeches or even social media chats can all be pulled.

The financial industry loves its buzzwords, and anything to do with artificial intelligence is particularly hot these days and should be treated with caution. But we could be at the start of a text investing revolution that could shake up the industry.

© 2021 The Financial Times Ltd

_______________________________________________________________

If you liked this story, sign up to find out more in the FP Investor newsletter.

_______________________________________________________________

Publicity

In-depth reporting on The Logic’s innovation economy, presented in partnership with the Financial Post.

comments

Postmedia is committed to maintaining a lively but civil discussion forum and encourages all readers to share their views on our articles. Comments may take up to an hour of moderation before appearing on the site. We ask that you keep your comments relevant and respectful. We have enabled email notifications. You will now receive an email if you receive a reply to your comment, if there is an update to a comment thread that you follow, or if a user that you follow comments. Visit our Community rules for more information and details on how to adjust your E-mail The settings.


Source link

Leave A Reply

Your email address will not be published.