Quant Memo
Core

Measuring Narrative Attention and Momentum

How quant researchers quantify how much the market is "talking about" a stock or theme, and whether rising attention itself predicts returns independent of the underlying news.

Markets don't just react to facts, they react to how much attention a story is getting — a mediocre earnings beat can move a stock sharply if it's suddenly the topic everyone is discussing, while genuinely important news can be ignored if nobody's looking. Narrative attention measures try to capture this directly, counting things like news article volume, social media mentions, or search interest for a company or theme over time, separately from whether the coverage is positive or negative.

Narrative momentum extends this by tracking whether attention itself is accelerating or decaying — a stock with rapidly rising mention counts week over week is behaving differently from one with steady, flat coverage, even if both currently have the same sentiment score. Researchers have found that spikes in raw attention, regardless of tone, tend to precede short-term price moves and elevated volatility, because attention is a rough proxy for how many new investors are becoming aware of a name.

Worked example. A stock's average weekly news mention count is 50. Over two weeks it jumps to 200 and then 400, a rising trend regardless of whether the articles are bullish or bearish. A narrative-momentum signal flags this doubling pace as informative on its own, distinct from any sentiment score computed on the same articles.

Narrative attention counts how much a name is being talked about; narrative momentum tracks whether that attention is accelerating — both carry information about coming price moves that is separate from, and can outweigh, whether the coverage itself is positive or negative.

Related concepts

Further reading

  • Da, Engelberg & Gao, 'In Search of Attention', Journal of Finance
ShareTwitterLinkedIn