Natural language processing (NLP) has been used by hedge funds, asset managers and other traders to study how much a term or word shows up in everything from news stories to social media, and how that affects a particular market (or other area of study). Commodity traders have been using it to determine weather patterns to trade crops for several years.
A new study by Indexica, an NLP alternative data firm, which uses predictive indexing intelligence to “find leading signals” in news patterns, studied those related to President Donald Trump.
“While it’s clear that Trump’s tweets and actions move markets within seconds, what hasn’t been clear until now is whether the broader conversation stemming from him and his actions moves markets systematically,” Indexica noted in its research.
In its study, the company found four key ways Trump affected markets:
- He drives volatility. The company used its BuzzSentiment metric, which “measures the ratio of news that is related to Trump rather than other topics,” combined with its Attitude measurement to predict Trump’s impact on market volatility.
“When Trump is written about more often and in a more negative tone, equities tend to become more volatile,” Indexica said in a statement, adding that “some may argue that the BuzzSentiment of Trump is highly volatile because the person the metric is measuring is highly volatile, and markets follow.” However, it notes that BuzzSentiment is a short-term metric, and that “markets tend to calm down as the metric reverts towards its moving average.”
Zak Selbert, Indexica CEO, told ThinkAdvisor that “Trump comes into play with volatility,” although the firm doesn’t provide Trump-centric information to clients as it does weather pattern information for trading coffee, for example.
- He generally drives equities higher. The company states that when its metric Anger is associated with Trump, the market tends to rise. This may be counterintuitive because “most anger associated with Trump is uncorrelated to corporate performance.”
However, another metric, Action Share, which “quantifies the degree to which news language about Trump is tilted toward verbs (actual action) vs. adjectives (lots of talk),” is correlated with rising markets, the company states. The exceptions to this are trade war and immigration actions, which actually can decrease equity market performance, Selbert said.
- He polarizes the market. No surprise here, but Indexica says that, using its Connectivity metric for thematic/exposure scoring, it found that companies highly linked to the term “trade war” underperform those less connected to the term.
Indexica acknowledges that’s not a surprise considering Trump’s polarization of the U.S. population, but states that “what is surprising is the degree to which the Connectivity metric is predictive of market performance. Trump tends to ratchet up his rhetoric in quantifiable peaks, and by identifying early levels of a predictable Trump tonality trend, predictive signals can be gleaned.”
- Nothing Trump does persists. Many of Indexica’s factors exhibit patterns of “broad, historically persistent drivers of return,” the company states, including “Futurity, Severity and Opportunity, which each measure trends using linguistic techniques.” It gives the example that top companies, as measured by Futurity, outperform the lower companies “by a multiple of three over long periods of time. But no factor related to Trump is persistent. The ways in which he drives market returns are short-term, chaotic and ever-changing,” the company states.
Indexica has no ax to grind; Selbert said separately in a statement that his firm’s job “is to agnostically identify and measure what drives markets. Though Trump might disagree, news articles narrate what’s happening in the world and are an ideal source for quantifying events, trends and opinions.”
— Related on ThinkAdvisor: