Crush Without the Move: When Earnings Sell Vol Harder Than Spot

22 September 2026 · SharpeIQ Research Desk · data: SharpeIQ

Abstract

Across 66 September 2026 earnings prints with both equity reaction and 30-day ATM IV change on file, the cross-section is weakly negatively related: Pearson r = −0.20, Spearman ρ = −0.10, OLS slope −0.39 IV points per 1% equity move (R² = 0.04). Spot direction alone barely explains post-event IV. The economically interesting mass sits elsewhere — names such as ORCL (−1.7% / −23.0pt) and PL (−1.3% / −45.6pt) crushed hard with near-flat spot, while TTAN (−30.0% / +70.2pt) and NAVN (−21.8% / +55.1pt) expanded IV into large drawdowns. That split is the signal: event premium can overprice a gap that never arrives, or underprice residual uncertainty after a gap that does. Not investment advice.

Key Takeaways · 15-second read

• 86% of the sample (57 / 66) printed a negative 30d ATM Δ — crush is the modal outcome, not the exception.
• Mean IV Δ was −11.8pt (median −14.4pt) against a mean absolute equity move of 7.9% — vol was sold harder than spot, on average.
• Classic crush-without-the-move cluster: ORCL, PL, SAIL, AGX, UNFI, PVH, ADBE (|spot| < 3%, IV Δ ≤ −15pt).
• Opposite quadrant — large drop plus IV expansion: TTAN, NAVN, BRZE (and milder COO, CASY).
• Rally + crush: SNOW (+16.6% / −26.7pt), GTLB, DELL, SIG — resolution without a leftover fear bid.
• Linear fit is almost useless (R² ≈ 0.04); trade the quadrant, not the slope.

1 · The sample

Source: SharpeIQ Research Earnings reactions table (Company, Period, Move, 30d ATM, Filed), window covering prints filed roughly 1–16 September 2026. One name (KMTS) had equity move but no 30d ATM reading and is dropped from the regression; n = 66.

The 30d ATM series on individual earnings pages is the constant-maturity change across the event — not the front-week expiry crush. That matters when comparing magnitudes: a −23pt move in 30d ATM is a continuum reprice, not merely the shedding of a binary event strip.

Field definitions (Data Ops / production, Sep 2026):

• Move — New York–session close-to-close % return straddling the 8-K: after-close filings use that day’s close → next session close; pre-market filings use prior close → filing-day close. Intraday filings use the same window but are flagged contaminated. Formula (post_close / pre_close − 1) × 100 on split/div-adjusted closes (price_history / Yahoo auto_adjust=True). Premarket gap sits inside that path.
• 30d ATM — Change in IBKR OPTION_IMPLIED_VOLATILITY on the underlying (iv_history.atm_iv): a single ~30-day constant-maturity ATM level, in volatility points (iv_after − iv_before). iv_before = last history date ≤ pre-event as-of; iv_after = first history date strictly after that as-of. Stored as decimal; UI shows points (e.g. ORCL −23.0pt).
• Not interchangeable with front-week: TTAN’s after-close row had front IV ~257 with a large implied move, while 30d ATM rose ~67 → ~137 (+70pt) on this table — front crush elsewhere can coexist with 30d expansion here.

Flagship colour from the ORCL page: Q1 FY2027 revenues $19.3bn (+30%), cloud $11.6bn (+62%), IaaS +121%, RPO $664bn — and the stock still closed −1.7% with −23.0pt of 30d ATM. Planet Labs (PL) printed a similar geometry: revenue beat colour, equity ≈ −1.3%, 30d ATM −45.6pt.

2 · Regression: IV Δ on equity move

Simple OLS of IV_Δ_pts on equity_move_pct across the 66 complete observations:

Interpretation is deliberately modest. A 10% rally predicts only about 4pt more crush than a flat print under the fit — and the scatter around that line is the story. Rank correlation is even weaker than Pearson: extremes in either axis are not ordered the same way.

Largest absolute residuals (actual IV Δ − fitted): WLTH (+15.0% / +66.5pt), TTAN (−30.0% / +70.2pt), NAVN (−21.8% / +55.1pt) on the high side; CRDO (−20.0% / −49.4pt), PL (−1.3% / −45.6pt), MDB (−13.5% / −40.3pt) on the low side. Those six names define the interesting quadrants better than the slope does.

3 · Four quadrants that matter

A · Near-zero spot, large IV crush (classic event-premium crush)

Pre-event IV had priced a distribution of gap outcomes that did not materialise. Once the binary resolved, ATM mean-reverted toward the ex-event continuum. Implied move ≫ realised spot path — overpriced short-horizon premium relative to the path taken, not proof the market “knew” the number.

B · Large spot drop + IV expansion (uncertainty re-anchored)

Opposite regime: the realised path exceeded or re-anchored the pre-event implied distribution, and residual uncertainty rose. Demand for protection / convexity after the catalyst — the event did not exhaust uncertainty.

C · Large spot rally + IV crush

Good news and resolution: spot reprices higher while the event strip is sold. Contrast WLTH (+15.0% / +66.5pt) — a rally that widened IV, closer in spirit to the expansion regime than to clean crush.

D · Large drop + still crushing (premium overkill even into pain)

CRDO (−20.0% / −49.4pt), MDB (−13.5% / −40.3pt), PATH (−16.6% / −32.3pt), ASAN (−12.7% / −30.0pt), AEO (−14.0% / −26.4pt) — spot was ugly, yet 30d ATM still shed tens of points. Event premium had been rich enough that crush survived a double-digit gap lower. That is not “vol ignored the move”; it is pre-event IV that was even richer than the realised path required.

4 · Charts: crush leaders and IV Δ ordered by spot

Figure 1 ranks the largest absolute crushers by 30d ATM Δ (most negative first). Figure 2 is the scatter substitute: the same IV Δ metric ordered left-to-right by equity move (worst drawdowns on the left, strongest rallies on the right). If spot drove IV, the bars would trend smoothly down the page; they do not — TTAN / NAVN / BRZE spike positive on the left, PL / ORCL crush near zero, and the right tail is mixed crush vs WLTH expansion.

5 · Flagship: ORCL as crush-without-the-move

Oracle is the cleanest large-cap illustration in the window. Equity −1.7% close-to-close around the 10 September release; 30d ATM −23.0pt. The fundamental tape was anything but quiet — cloud acceleration and a $664bn RPO print — yet the spot path was muted relative to what short-dated premium typically prices into an ORCL event week. Live widgets below for post-print context:

6 · Interpretation (Strategy Trader / SharpeIQ desk framing)

Desk framing — not a trade call:

1. Steep negative 30d ATM IV Δ with ~0 spot move = classic event-premium crush. Pre-event IV priced gap outcomes that did not materialise; ATM mean-reverts toward the ex-event continuum. Implied move ≫ realised — overpriced short-horizon premium relative to the realised spot path, not evidence the market “knew” the number.

2. Large spot move with IV expansion = opposite regime. The realised path exceeds or re-anchors the pre-event implied distribution; residual uncertainty rises. Demand for protection and convexity after the catalyst — the event did not exhaust uncertainty (TTAN, NAVN, BRZE).

3. Operational caveats before sizing either read: (a) liquidity and quote quality — thin ATM or wings inflate IV Δ; (b) stale / pre-open versus closing marks bias both axes; (c) mega-gap, limit prints and corporate actions dominate as outliers; always check DTE mix so IV Δ is tenor-comparable (front binary strip versus 30d ATM continuum — this note uses the latter).

7 · What would falsify the reading

• If a refreshed sample with strict liquidity filters (minimum OI / bid-ask) erased the near-zero-spot crush cluster, the “overpriced event strip” claim would be an artifact of thin names.
• If front-week Δ and 30d ATM Δ systematically disagreed on quadrant membership for the same tickers, tenor mix — not economics — would be driving the table.
• If subsequent sessions showed ORCL / PL-style crushers re-expanding 30d ATM without a new catalyst, the “resolution” label would have been premature.

References

1. SharpeIQ Research — Earnings reactions table (Sep 2026 sample): https://www.sharpeiq.com/research
2. SharpeIQ — ORCL Q1 FY2027 earnings reaction (−1.7% / −23.0pt 30d ATM): https://www.sharpeiq.com/research/earnings/orcl-q1-fy2027-earnings
3. SharpeIQ — PL Q2 FY2027 earnings reaction (−1.3% / −45.6pt 30d ATM): https://www.sharpeiq.com/research/earnings/pl-q2-fy2027-earnings
4. SharpeIQ Research — Risk Resolution and Volatility Compression: The SPCX Post-IPO Implied-Vol Collapse (event-driven crush precedent): https://www.sharpeiq.com/research/risk-resolution-and-volatility-compression-the-spcx-post-ipo-implied-vol-collapse
5. SharpeIQ Research — Reading Implied Volatility: https://www.sharpeiq.com/research/reading-implied-volatility
6. SharpeIQ Research — When 100% Vol Is Cheap (IV vs realised framing): https://www.sharpeiq.com/research/when-100-vol-is-cheap-mu-sandisk-in-the-ai-memory-supercycle

Methodology footnote

Move is the NY-session close-to-close return straddling the 8-K (split/div-adjusted); 30d ATM is the change in IBKR ~30-day constant-maturity ATM IV across that same window, in volatility points — not front-week event IV. Validated against production (earnings_reaction.py / event_move_store.py / iv_term.py) by Data Ops, 22 Sep 2026. Sample drops null IV (KMTS). Intraday filings and unrevalidated liquidity filters remain open caveats; OLS here is tenor-comparable only for this 30d CM ATM Δ series.

Educational research only. Not investment advice. Options involve substantial risk of loss. Figures above are transcribed from SharpeIQ tables and linked earnings pages; regression statistics computed on the parsed complete sample (n = 66).