What it measures
The Hesitation Index reads the language of caution. When confidence drops, dispersion widens, and the room collectively talks more about layoffs and recession while talking less about buybacks, the index rises. Above 65, you are in a measurable hesitation regime — the corpus is communicating defensive posture. This is the leading-edge soft indicator that precedes harder macro confirmation by 30–90 days. Named for The Hesitation Current — the slow, multi-quarter slowdown that doesn't quite become a recession but reshapes capital deployment across the economy.
Why these components
Layoff mentions catch labor-market caution directly. Recession mentions catch forward-looking macro caution. Confidence dispersion catches the regime in which executives are increasingly disagreeing about what comes next — the moment a coherent story breaks down and individual companies start hedging their narrative. The inverse of buyback mentions catches the capital-allocation tell: when boards stop talking about returning capital and start talking about reserving it, that is the strongest single signal that they are bracing for something.
How to read it
Above 65 — the corpus is in measurable hesitation. Between 35 and 65 — normal range. Below 35 — confidence regime; companies are deploying, not defending. The strongest read happens during early earnings season for a given quarter: a fast rise in The Hesitation Index across the first 20% of S&P 500 reports tends to land before the broader macro data catches up by 4–8 weeks.
What it does not say
The Hesitation Index reads language, not action. Companies can talk defensive and still grow, and vice versa. The number is most useful read against actual capex disclosures and the FRED labor data — when the language and the actions diverge, the index identifies the gap rather than resolving it.
History note
History recomputed 2026-09-21 on the rebuilt five-year Discourse corpus under the canonical discourse.* definition: readings before 2026-09-19 are weekly as-of recomputes that replace the history originally computed on the pre-rebuild corpus.