Cutting the Future: What Happens When Capex Goes Down on Calls
This study examines 23,150 earnings calls (14.0% of a 165,182-call corpus spanning 1990-2026) where capex was read as decreasing. Compared with the rest of the corpus, these calls show lower confidence (6.71 vs 7.21), less promotion (4.68 vs 5.05), and markedly higher stress (3.06 vs 2.43). Guidance is lowered far more often (22.96% vs 11.56%) and withdrawn more than three times as often (8.51% vs 2.66%). Language patterns skew toward 'Underused Fixed Costs' (1.47x) and away from growth stories like 'Deferred Revenue Growing' (0.62x). Among 2,450 calls with return data, the median return is -8.88% versus -7.16% for the 22,449-call baseline, and only 38.6% beat the baseline's 39.5%.
- Calls with decreasing capex show a stress score of 3.06 versus 2.43 elsewhere, the largest behavioral gap in the profile.
- Guidance is lowered on 22.96% of these calls versus 11.56% of the baseline, and withdrawn on 8.51% versus 2.66%.
- The phrase pattern 'Underused Fixed Costs' appears 1.47 times more often, while 'Deferred Revenue Growing' appears 0.62 times as often.
- The 2,450-call returns sample has a median return of -8.88%, below the -7.16% baseline median.
1Introduction
When a company says it is spending less on capital expenditures, the sentence can mean discipline or distress, and earnings-call language around it is often where the truth leaks out first. Anyone who listens to these calls knows that capex cuts arrive dressed in many tones: some confident, some defensive, some quietly stressed. Because capex is a forward-looking commitment, how management talks about reducing it may reveal how they view their own runway. This study examines 23,150 calls, 14.0% of a 165,182-call corpus from 1990-2026, where capex was read as decreasing, comparing their language, guidance behavior, and post-call outcomes 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 capex was read as decreasing (n = 23,150; 14.0% of the reference set, 95% Wilson interval 13.8%–14.2%). 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 behavioral profile points one direction: confidence runs 6.71 versus 7.21 in the baseline and promotion 4.68 versus 5.05, while stress is elevated at 3.06 versus 2.43. Candor is slightly higher (7.04 vs 6.86), suggesting these managements are not hiding the cut. Guidance behavior confirms the mood: lowered guidance appears on 22.96% of calls versus 11.56% baseline, and withdrawals on 8.51% versus 2.66%. Language lifts reinforce the picture, with 'Underused Fixed Costs' at 1.47x and 'Results Worse Than Direction' at 1.35x, while growth-narrative phrases are underrepresented, including 'A Tiny Fraction of the Market' at 0.72x. Returns skew mildly negative, with a median of -8.88% versus -7.16%.
| Meter | Study group | Baseline | Δ |
|---|---|---|---|
| Candor | 7.04 | 6.86 | +0.18 |
| Evasion | 2.82 | 2.70 | +0.12 |
| Specificity | 7.58 | 7.56 | +0.02 |
| Stress | 3.06 | 2.43 | +0.63 |
| Promotion | 4.68 | 5.05 | -0.37 |
| Confidence | 6.71 | 7.21 | -0.51 |
| Action | Study group | Baseline |
|---|---|---|
| Raised | 12.7% | 21.1% |
| Maintained | 42.2% | 48.8% |
| Lowered | 23.0% | 11.6% |
| Withdrawn | 8.5% | 2.7% |
| Signal | Lift | In group | Baseline |
|---|---|---|---|
| Underused Fixed Costs | 1.47× | 61.1% | 41.6% |
| The Finished-Story Tell | 1.36× | 6.0% | 4.4% |
| Results Worse Than Direction | 1.35× | 69.2% | 51.1% |
| Deferred Revenue Growing | 0.62× | 5.5% | 8.9% |
| A Tiny Fraction of the Market | 0.72× | 21.5% | 30.0% |
| Early Products Growing Fast | 0.74× | 28.6% | 38.5% |
| Statistic | Study group | Returns sample |
|---|---|---|
| Median excess return | -8.9% | -7.2% |
| Interquartile range | -30.3% to +12.8% | — |
| Share beating SPY | 38.6% (95% CI 37%–41%) | 39.5% |
| Observations | 2,450 | 22,449 |
| Ticker | Quarter | Call date | Call grade |
|---|---|---|---|
| AN | Q2 2025 | 2025-07-25 | B |
| CHTR | Q2 2025 | 2025-07-25 | C+ |
| PUBL | Q2 2025 | 2025-07-25 | B+ |
| MTH | Q2 2025 | 2025-07-25 | C |
| PUBL | Q2 2025 | 2025-07-25 | C |
| VWAGY | Q2 2025 | 2025-07-25 | C |
| HLKHF | Q2 2025 | 2025-07-25 | B |
| WZZAF | Q1 2026 | 2025-07-25 | F |
4Discussion
A careful reader should conclude that calls announcing capex decreases carry a noticeably more stressed and less promotional tone, and that their guidance actions skew sharply toward cuts and withdrawals. That is a description of the sample, not a forecast: nothing here shows capex-cutting language causes lower returns, and the median gap of -8.88% versus -7.16% is modest with wide dispersion (Q1 -30.26%, Q3 12.83%). The trend series also swings widely year to year, from 6.73% in 2021 to 26.45% in 2020, so prevalence is regime-dependent. Treat these patterns as context, not signals.
5Limitations
The language fields are read by AI models and are noisy; a 'decreasing capex' label may misread hedged or ambiguous statements. The returns comparison covers 2,450 calls against a 22,449-call baseline skewed toward liquid names, so composition differs. Our own forward tests falsified directional prediction from these features, and LLMs partially remember famous stocks' histories, contaminating any backtest. Correlations here describe the corpus; they do not establish causation, prediction, or a trading edge. See the full methodology, including the C1 pattern’s forward-test failure and the LLM-memorization finding.