Beat and fall: why good results can still send a stock lower

What a beat is measured against

A result is a beat or a miss only relative to a forecast. In most earnings coverage that forecast is an analyst consensus: an average of estimates compiled by a data vendor. Those figures are licensed products, which is why they do not appear here.

Two forecasts are public and checkable. The first is the company's own guidance, published in an earlier Item 2.02 8-K, which the new release usually restates next to the update. The second is harder to see: the share price itself. The close on the day before a release is the price at which buyers and sellers last agreed, knowing that results were due. Whatever they collectively expected is already in that number. A release then moves the price by how far it differs from that expectation, not by how good it looks in isolation. A quarter can be strong against last year, strong against guidance and still fall short of what the price had assumed. Nobody publishes that assumption, so the reaction is the only trace of it.

Three companies, one morning

All three 8-Ks list Item 2.02 and were accepted on EDGAR before the 09:30 ET open on 20 October: GE Aerospace at 06:30:57, 3M at 06:35:38 and General Motors at 07:30:18 (EDGAR filing index pages). The first session to react was therefore 20 October itself, and the comparison is with the 19 October close.

GE Aerospace: 311.28 to 299.87, so 299.87 ÷ 311.28 − 1 = −3.67%. General Motors: 82.75 to 77.09, so −6.84%. 3M: 162.01 to 171.37, so +5.78%. The S&P 500 went from 7,690.73 to 7,712.57, or +0.28%. Subtracting the index move gives a rough market-adjusted figure: −3.95, −7.12 and +5.49 percentage points (our arithmetic). That subtraction treats each stock as moving one-for-one with the index, which is a simplification, but on a day when the index moved this little it changes the picture very little.

Guidance and the quarter still to come

A full-year range issued in October mostly describes one remaining quarter, so the useful step is to subtract what has already been reported. GE Aerospace's release gives adjusted EPS of $6.13 for the first nine months of 2026. Against the new full-year range, that leaves $8.00 − $6.13 = $1.87 to $8.10 − $6.13 = $1.97 for the fourth quarter. The third quarter itself came in at $2.25. For the year-earlier comparison, the same release gives 2025 adjusted EPS of $6.37 and nine-month 2025 adjusted EPS of $4.80, so the fourth quarter of 2025 was $1.57. The growth this works out to for the fourth quarter of 2026 is 1.87 ÷ 1.57 − 1 = 19% to 1.97 ÷ 1.57 − 1 = 25%, against 36% reported for the third quarter (our arithmetic; GE notes its per-share amounts are computed independently, so they may not add exactly). The release raised the guide; its own figures also leave room for only slower growth in the final quarter than in the one just reported.

General Motors shows the same mechanism with a range that narrowed rather than rose. The EBIT-adjusted midpoint was $15.0bn before and after: (14.0 + 16.0) ÷ 2 and (14.5 + 15.5) ÷ 2. The release gives EBIT-adjusted of $12,302m for the first nine months, which leaves $2.2bn to $3.2bn for the fourth quarter, against $4,106m reported for the third (our arithmetic). Fourth quarters can differ from third quarters for seasonal and model-year reasons, so this is a reading of GM's own numbers, not a forecast. The adjusted EPS midpoint did rise, from $13.00 to $13.25, and GM raised its adjusted automotive free cash flow range to $11.5bn–$12.5bn from $9.5bn–$11.5bn.

GE had also raised its guide in July: its 16 July 2026 8-K under Item 2.02 moved adjusted EPS to $7.65–$7.85 from $7.10–$7.40. A second raise in a row is still a raise, but by October the price had had three months to absorb the first.

The quality of the result

Headline adjusted figures are only one line of a release. The rest of the tables can point the other way, and the three releases each show this.

GE Aerospace reported adjusted EPS of $2.25, up 36%, and operating profit margin of 22.5%, up 220bp. On a GAAP basis, profit margin fell 80bp to 19.9%. Cash from operating activities was $2,203m against $2,556m, and free cash flow $2,042m against $2,362m, both down 14%. In its largest segment, Commercial Engines & Services, revenue rose 21% but margin slipped 10bp to 27.9%, which the release attributes to mix and inflation.

General Motors reported EPS-diluted-adjusted of $3.70, up 32.2%, with North America EBIT-adjusted up 43.4% to $3,592m. Outside North America the picture was weaker: GM International EBIT-adjusted fell to $28m from $226m, China equity income to $32m from $80m, and GM Financial EBT-adjusted to $668m from $804m.

3M shows the reverse case, where the GAAP headline flatters. GAAP EPS of $2.63 was up 70%, but the reconciliation includes a $0.60 per-share gain from the higher value of its Solventum stake. Adjusted EPS was $2.50, up 12%. The guidance raise also mixes two things: 3M now treats amortisation of acquired intangibles as a special item, which the release says lifts third-quarter adjusted EPS by about $0.04. The prior range of $8.80–$8.95 and the new $9.10–$9.15 are on different bases, and the release does not give a full-year figure for the change, so the $0.25 rise at the midpoint (8.875 to 9.125) is not all operating improvement.

What one day can and cannot show

Each of these releases contains good news and less good news, and the price moved in different directions for companies with broadly similar headlines. The filings do not say which line drove the move, and neither does the price. On the same morning there was a conference call for each company, analysts' notes, and trading for reasons no filing records; none of that is in the Item 2.02 8-K. A fall after a raise can reflect the next quarter, cash flow, a weak segment, a raise smaller than the price assumed, or selling unrelated to the results. Attributing it to one cause requires evidence beyond the close.

The same pattern appeared a week earlier. Johnson & Johnson raised its 2026 adjusted EPS range on 13 October 2026 and closed 5.73% lower that session (263.73 to 248.61, Yahoo Finance daily closes), while the S&P 500 was little changed at −0.02%. We walk through that filing's figures in /articles/how-to-read-an-earnings-release-and-its-8-k.

Two cautions about the arithmetic itself. A one-day close-to-close change includes everything that happened between the two closes, not only the release. And an index adjustment of the kind above removes the broad market only roughly; sector moves and each stock's own sensitivity to the market are not taken out.

How to check the numbers yourself

Every figure above can be rebuilt from two public sources. For the filings, open the company's submissions file at data.sec.gov (GE Aerospace is CIK 0000040545, General Motors 0001467858, 3M 0000066740) or its EDGAR filings page, and look for 8-Ks whose items include 2.02. The filing index shows the accepted time, which tells you which session was the first to react. In Exhibit 99.1, the guidance table usually shows prior and updated ranges side by side; the earlier Item 2.02 8-K confirms the prior figures. Subtract the nine-month figures from the full-year range to see what the company's range leaves for the remaining quarter.

For prices, take the daily close on the session before the release and the first session after it. A release before the open reacts the same day; one after the close reacts the next day. Do the same for the S&P 500 to see what the market did. Our earnings calendar at /earnings lists upcoming report dates, and /markets shows the index closes.