Analysts' vocabulary 'herd-like' in run-up to market bubble


GET-RICH-quick schemes, cash advances and exuberant investors are 10-a-penny during stock market bubbles, but verbs and nouns may be in shorter supply.

Computer scientists at University College Dublin have found that when the stock market is overheating, the language used by financial analysts and reporters becomes increasingly similar.

After examining almost 18,000 articles published by the Financial Times, the New York Timesand the BBC between 2006 and 2010, they discovered that the language used by financial commentators converges in a “herd-like” fashion in the run-up to an investment bubble. Immediately after the bubble bursts, the language used becomes more varied again.

The research was completed by Aaron Gerow as part of his Masters degree at UCD and Prof Mark Keane, who is chair of computer science at the college.

Their paper, Mining the Web for the ‘Voice of the Herd’ to Track Stock Market Bubbles, will be presented at the International Joint Conference on Artificial Intelligence in Barcelona today.

“Our analysis shows that trends in the use of words by financial journalists correlate closely with changes in the leading stock indices – the Dow Jones Index, the Nikkei-225 and the FTSE-100,” said Prof Keane.

As wallets loosened, vocabularies tightened. “Our study shows that reporters converge on the same language – ‘stocks rose again’, ’scaled new heights’ or ‘soared’ – as their commentaries became more uniformly positive in the lead up to the 2007 crash.”

Although it is possible journalists simply have wider vocabularies when it comes to the use of “negative” verbs, the researchers claim a shrinking pool of nouns used during bubble periods suggests “classic herd-like behaviour” is taking place, with commentators fixating on a handful of rapidly rising stocks.

“It’s the dotcom effect where everyone is talking about the three companies that are doing well and they’re not talking about the 400 companies that are not doing well,” said Prof Keane.

Almost 13,300 of the articles analysed were from The Financial Times,about 2,400 from the New York Timesand 2,000 from the BBC’s online news. Prof Keane said the appropriation of words and phrases from sources such as news wires should be constant in bull and bear markets, so this could not account for language converging during a bubble.