Show Me the Money (Ask): Calls Where Pricing Read as Raising
We examine earnings calls where pricing language was read as raising, drawing on 165,182 calls from 1990 to 2026, of which 49,574 (30.0%, 95% CI 29.8% to 30.2%) qualify. These calls show higher confidence (7.46 vs 7.21), specificity (7.76 vs 7.56), and candor (6.99 vs 6.86), with lower stress (2.20 vs 2.43) and evasion (2.54 vs 2.70). Guidance was raised on 28.9% of calls versus 21.1% in the base, and guidance was withdrawn less often (1.5% vs 2.7%). The topic Pricing Recovering appears 2.17x more often than expected. Forward returns were not better: median -6.7% versus -7.2% base.
- 30.0% of 165,182 calls (n=49,574) were read as pricing raising, with a 95% CI of 29.8% to 30.2%.
- Confidence scores run 7.46 vs 7.21 in the base, and stress 2.20 vs 2.43.
- Guidance was raised on 28.9% of these calls versus 21.1% of base calls; guidance was withdrawn on 1.5% vs 2.7%.
- The 9,323-call returns sample shows a median forward return of -6.7%, essentially matching the -7.2% base median.
1Introduction
Pricing is the most direct lever a company has, and how management talks about it is a fast read on operating conditions. When speakers discuss price increases, they tend to sound noticeably steadier: more confident, more specific, less stressed. That tonal gap is why pricing-raising calls are often treated as a signal worth tracking. But tonal polish and subsequent outcomes are different things, and it is worth checking whether the apparent strength carries into results. This study profiles 49,574 pricing-raising calls out of 165,182 from 1990 to 2026, comparing their language, guidance behavior, topic mix, and forward returns against the rest of the corpus.
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 pricing was read as raising (n = 49,574; 30.0% of the reference set, 95% Wilson interval 29.8%–30.2%). 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 language profile is consistently stronger: confidence 7.46 vs 7.21, specificity 7.76 vs 7.56, candor 6.99 vs 6.86, while stress (2.20 vs 2.43) and evasion (2.54 vs 2.70) run lower. Guidance actions point the same way: 28.9% raised versus 21.1% base, and only 1.5% withdrawn versus 2.7%. Topic lifts reinforce the picture: Pricing Recovering appears 2.17x more often than expected (46.6% observed vs 21.5% base rate), while Scale-Dependent Advantage Claims (0.42x) and The Question Left Hanging (0.74x) are rarer. The trend peaked at 49.3% of calls in 2022, easing to 27.3% in 2024. Forward returns, however, show no edge: median -6.7% vs -7.2% base.
| Meter | Study group | Baseline | Δ |
|---|---|---|---|
| Candor | 6.99 | 6.86 | +0.13 |
| Evasion | 2.54 | 2.70 | -0.15 |
| Specificity | 7.76 | 7.56 | +0.21 |
| Stress | 2.20 | 2.43 | -0.22 |
| Promotion | 4.98 | 5.05 | -0.07 |
| Confidence | 7.46 | 7.21 | +0.24 |
| Action | Study group | Baseline |
|---|---|---|
| Raised | 28.9% | 21.1% |
| Maintained | 50.5% | 48.8% |
| Lowered | 9.0% | 11.6% |
| Withdrawn | 1.5% | 2.7% |
| Signal | Lift | In group | Baseline |
|---|---|---|---|
| Pricing Recovering | 2.17× | 46.6% | 21.5% |
| Scale-Dependent Advantage Claims | 0.42× | 4.6% | 11.1% |
| The Question Left Hanging | 0.74× | 35.4% | 48.0% |
| Statistic | Study group | Returns sample |
|---|---|---|
| Median excess return | -6.7% | -7.2% |
| Interquartile range | -24.5% to +11.1% | — |
| Share beating SPY | 39.2% (95% CI 38%–40%) | 39.5% |
| Observations | 9,323 | 22,449 |
| Ticker | Quarter | Call date | Call grade |
|---|---|---|---|
| DOC | Q2 2025 | 2025-07-25 | C |
| CNC | Q2 2025 | 2025-07-25 | F |
| BFH | Q2 2025 | 2025-07-25 | B |
| FFIC | Q2 2025 | 2025-07-25 | B+ |
| OMF | Q2 2025 | 2025-07-25 | A |
| MOG.A | Q3 2025 | 2025-07-25 | B+ |
| HMDPF | Q2 2025 | 2025-07-25 | B |
| TNET | Q2 2025 | 2025-07-25 | C+ |
4Discussion
The consistent picture is descriptive: pricing-raising calls come with steadier tone, more raised guidance, and more recovery talk. A careful reader should not conclude that the tone causes outcomes, or that these calls predict better or worse returns. The median forward return of -6.7% sits close to the -7.2% base median, and the beat rate of 39.2% (95% CI 38.2% to 40.2%) is essentially indistinguishable from the base 39.5%. The honest summary is that pricing-raising language co-occurs with stronger-sounding calls but does not translate into a measurable outcome difference in this data.
5Limitations
The language fields are AI-read and noisy, so small deltas like the 0.13 candor gap should be treated cautiously. The returns sample covers 22,449 calls and is skewed toward liquid names, so it may not represent the full corpus. Our own forward tests falsified directional prediction, and no claim of trading edge is made here. Additionally, LLMs partially remember famous stocks' histories, which can contaminate backtests of model-read fields; any apparent historical pattern should be interpreted with that contamination in mind. See the full methodology, including the C1 pattern’s forward-test failure and the LLM-memorization finding.