Pricing Power to the People: A Profile of 'Pricing Recovering' Calls
We examine 35,541 earnings calls from a corpus of 165,182 spanning 1990 to 2026 in which the model answered YES to the battery item 'Pricing Recovering'. These calls skew candid and specific: candor 7.01 vs 6.86, specificity 7.76 vs 7.56, and confidence 7.34 vs 7.21, with lower evasion (2.53 vs 2.70) and stress (2.31 vs 2.43). They raise guidance more often (25.6% vs 21.1%) and lower it less (8.4% vs 11.6%). Yet the language lift is modest: 'Volume About to Step Up' appears at 1.28x, while rehearsed-sounding and hype-flavored phrases run under baseline. Post-call 5-day returns show a median of -0.0837 vs -0.0716 baseline, with 37.6% beating vs 39.5%.
- 'Pricing Recovering' YES calls account for 35,541 of 165,182 calls, a 21.5% share (95% CI 21.3%-21.7%).
- Guidance was raised on 25.6% of these calls versus 21.1% of the base, and lowered on 8.4% versus 11.6%.
- The phrase 'Volume About to Step Up' appears at a 1.28x lift, while 'Scale-Dependent Advantage Claims' (0.59x) and 'Calls That Read Rehearsed' (0.68x) are underrepresented.
- The returns sample of 5,152 calls shows a median 5-day return of -0.0837 versus -0.0716 for the 22,449-call base, with 37.6% beating versus 39.5%.
1Introduction
When management says pricing is recovering, they are making one of the most consequential claims on an earnings call: that the company has regained the ability to push prices through, which shows up directly in margins and guidance. Analysts parse these statements closely, but rarely in aggregate. Do calls flagged for pricing recovery actually sound different — more candid, more specific, less evasive? Do they carry stronger guidance actions, and different language? And how do subsequent returns compare? This study profiles all calls in the corpus where the model answered YES to the 'Pricing Recovering' item.
2Data & methodology
The corpus comprises 165,182 earnings-call transcripts published between 1990 and 2026, each scored independently by a large language model on an identical 37-field battery: seven categorical business verdicts, eight 0–9 behavioral meters, and twenty yes/no judgments. The study group is defined as calls where the model answered YES to the battery item "Pricing Recovering" (n = 35,541; 21.5% of the reference set, 95% Wilson interval 21.3%–21.7%). Baseline figures use all scored calls. Market outcomes join a fixed sample of 22,449 calls with twelve-month total returns in excess of SPY, measured from the first close after each call; this sample skews toward liquid U.S. names and is reported as descriptive history only.
3Results
The profile deltas are directionally 'healthier': candor +0.15, specificity +0.21, confidence +0.13, with evasion -0.16 and stress -0.12. Guidance actions align: 25.6% raised versus 21.1% base, 8.4% lowered versus 11.6%, and only 1.5% withdrawn versus 2.7%. Language lifts are mild; the top overrepresented phrase, 'Volume About to Step Up', sits at 1.28x with wide uncertainty (0.28-0.37 base rates), while hype-tinged phrases like 'A Tiny Fraction of the Market' (0.64x) and 'Calls That Read Rehearsed' (0.68x) appear less often. Returns tell a flatter story: median -0.0837 versus -0.0716, and 37.6% beats versus 39.5%.
| Meter | Study group | Baseline | Δ |
|---|---|---|---|
| Candor | 7.01 | 6.86 | +0.15 |
| Evasion | 2.53 | 2.70 | -0.16 |
| Specificity | 7.76 | 7.56 | +0.21 |
| Stress | 2.31 | 2.43 | -0.12 |
| Promotion | 4.93 | 5.05 | -0.12 |
| Confidence | 7.34 | 7.21 | +0.13 |
| Action | Study group | Baseline |
|---|---|---|
| Raised | 25.6% | 21.1% |
| Maintained | 51.0% | 48.8% |
| Lowered | 8.4% | 11.6% |
| Withdrawn | 1.5% | 2.7% |
| Signal | Lift | In group | Baseline |
|---|---|---|---|
| Volume About to Step Up | 1.28× | 36.5% | 28.5% |
| Scale-Dependent Advantage Claims | 0.59× | 6.5% | 11.1% |
| A Tiny Fraction of the Market | 0.64× | 19.3% | 30.0% |
| Calls That Read Rehearsed | 0.68× | 26.4% | 38.7% |
| Statistic | Study group | Returns sample |
|---|---|---|
| Median excess return | -8.4% | -7.2% |
| Interquartile range | -25.9% to +9.8% | — |
| Share beating SPY | 37.6% (95% CI 36%–39%) | 39.5% |
| Observations | 5,152 | 22,449 |
| Ticker | Quarter | Call date | Call grade |
|---|---|---|---|
| CNC | Q2 2025 | 2025-07-25 | F |
| FFIC | Q2 2025 | 2025-07-25 | B+ |
| OMF | Q2 2025 | 2025-07-25 | A |
| GBCI | Q2 2025 | 2025-07-25 | A |
| UVE | Q2 2025 | 2025-07-25 | C+ |
| FLG | Q2 2025 | 2025-07-25 | B |
| TNET | Q2 2025 | 2025-07-25 | C+ |
| SSB | Q2 2025 | 2025-07-25 | B+ |
4Discussion
A careful reader should conclude that calls where the model says pricing is recovering tend to feature slightly more candid, specific, confident language and somewhat stronger guidance actions than baseline. They should not conclude that the model's YES means pricing genuinely recovered, that these calls perform better or worse as investments, or that the language profile causes anything. The returns gap is small and measured over one fixed window; with 37.6% versus 39.5% beat rates, the differences are descriptive, not actionable, and we make no predictive claims.
5Limitations
The YES/NO battery fields are AI-read judgments and noisy at the individual call level, so the 21.5% share carries real misclassification error. The returns subsample covers 22,449 base calls (5,152 flagged) and is skewed toward liquid names, limiting generalizability. Our own forward tests falsified directional prediction from these signals. Additionally, LLMs partially remember famous stocks' histories, which can contaminate any backtest by leaking hindsight into the labels themselves. See the full methodology, including the C1 pattern’s forward-test failure and the LLM-memorization finding.