Every “year two of a presidential term” since 1986, rebuilt from actual month-end closes — combined with the election result and the latest macro data into a judgment for year-end and the 12 months after the midterms.
中文版 →If you read one screen, read this one.
QQQ closed Nov 5 at $647.77, down 6.4% on the day — since 2020, only three days in March 2020 were worse, and it topped the worst day of the April 2025 tariff shock. The year-to-date gain shrank to +5.4%, 14.7% below the Oct 6 high, and the Nasdaq closed in correction territory. Chips led the way down: AMD −17.7%, Micron −17.2%, Nvidia −9.3%, and SMH is 30% below its Oct 5 high; in the afternoon the selling spread to the whole market — the S&P 500 −4.4%, the Dow −3.5%, VIX 28.
An exit poll found about 76% of voters oppose new data centers where they live; Arizona Rep. Yassamin Ansari called for a pause on advanced AI development on the new Congress's “day 1.” Reuters reported about 600,000 Nvidia H200 chips stuck in Hong Kong awaiting Beijing's approval, and Qualcomm's handset revenue fell 26% year over year. The market's read: the AI buildout will become “slower and more expensive.”
The base and bear cases now carry 40% each. Base: choppy trading at lower levels, ending the year around +7% (≈$657). Bear: a replay of December 2018, ending around −1% (≈$608). Bull 20% (≈$719). Cushions are appearing too: December hike odds fell from about 81% this morning to about 69%, and the 2-year yield fell 11bp.
The next line is $620 (the mid-April closing area) — a close below it makes the bear case the main scenario; reclaiming the 200-day moving average (about $680.7) brings the bear probability back to 35%. Several analysts called today's selling a capitulation; but 2018 showed that a panic can be followed by another leg down, until the Fed turns.
According to Yahoo Finance, all four major networks said Democrats' control of the House was assured, and Democrats also took a Senate majority; the new Congress is sworn in in January 2027, splitting control between the White House and Congress. The market's reaction wasn't a simple up or down: first a rotation, then a broad de-risking — the Dow had its best day of the year the day after the election, and the next day the chip crash spread to the whole market. Three channels:
The market's read: with the president's policy room narrower and Congress likely to tighten war funding and oversight, military escalation with Iran looks less likely. That directly lowered inflation expectations and Treasury yields.
Large fiscal expansion becomes harder to pass, which is seen as good for bonds; tariffs are largely a presidential lever, so the election's direct effect on them is limited. A Senate flip also makes Fed appointments harder to steer.
Data-center policy may reach the 2027 agenda; Texas's governor-elect has pledged a construction moratorium, an exit poll found about 76% of voters oppose new data centers where they live, and Arizona Rep. Yassamin Ansari called for a pause on advanced AI development on “day 1” (a social-media post, not a bill). UBS estimates about 30% of expected 2028 capacity additions are at risk.
Nine indicators, each scored as a tailwind or headwind for the Nasdaq-100: 2 tailwinds, 4 headwinds, 3 neutral or swing (the Oct 25 version had 0 tailwinds and 5 headwinds). The macro picture is improving; the pressure is concentrated in AI hardware. Gray notes show changes from the Oct 25 version.
December hike odds fell from about 81% on the morning of Nov 5 to about 69% (CME FedWatch, 12:15 p.m. ET) — the stock rout has started to pull hike bets down, but a hike is still more likely than not. Held on Oct 28 in a 10–2 vote, with two members favoring a hike; Warsh offered no forward guidance.
Near its lowest since Sep 10, about 50bp below the 5.33% intraday high on Oct 14; during the Nov 5 rout the 10-year fell 5bp and the 2-year 11bp. The earlier decline came from falling oil and inflation expectations; today added safe-haven buying.
September core PCE rose 0.3% month over month and 3.1% year over year, with headline at 3.6% — all in line; Q3 unit labor costs fell 0.7%, easing wage pressure. Lower oil will show up in October–November headline inflation, but core is still well above 2%. October CPI is due Nov 10.
Closed at $116.72 on Oct 23; fell 8% in a day after the election and rebounded about 2.5% on Nov 5. Markets are betting on an end to the Iran war — a bet, with no formal agreement reported yet. WTI is still about 50% above where it began the year, and volatility will be high.
Initial jobless claims were 189K, below the 196K expected; Q3 productivity grew 3.6% and GDP 3.4% annualized. ADP added 86K jobs in October; payrolls are due Nov 6, with consensus at +65K (September was just +29K). The mix of strong growth and weak hiring hasn't changed.
Michigan's October final, a five-month low. Gasoline futures are at their lowest since mid-March, which should ease household budgets; the November preliminary reading is due Nov 6.
Google Cloud +103%, Azure +48%, AWS +39%: demand is real and being monetized. But Alphabet's quarterly free cash flow was −$9.1B and Amazon's trailing-12-month −$42.4B, and Reuters reported about 600,000 Nvidia H200 chips stuck in Hong Kong awaiting Beijing's approval. The market is rewarding restraint and punishing cash burn.
Neoclouds remain under pressure: on Nov 5 CoreWeave fell 9.8%, IREN 17%, and Nebius and Applied Digital 15% each; CoreWeave reports Nov 10. The post-election policy risk is escalating: an exit poll found about 76% of voters oppose new data centers where they live, and UBS estimates about 30% of expected 2028 capacity additions are at risk.
Vanda's data show retail investors net sold about $358 million on the day, the biggest selling day since March 2020 — a “total capitulation by retail investors”; VIX posted its biggest one-day jump in over two years. The selling spread from AI hardware to the whole market, but Treasuries rallied and hike bets fell — so far this looks like de-risking, not a credit crisis. UBS says “a lot is already priced into Semis.”
What the intraday version worried about became fact by the close. QQQ closed at $647.77, down 6.4% on the day — below the $650 line written down in advance, and its worst day since March 2020. In the afternoon the selling spread from AI hardware to the whole market: the S&P 500 fell 4.4% and the Dow 3.5%, and Apple and Microsoft closed lower too. There was no single new headline; the market was repricing the election — in the words of an eToro analyst, “the AI infrastructure buildout will become slower and more expensive.”
Why raise the bear probability to 40%: the rule was written in advance — a close below $650 raises it; and the path increasingly resembles 2018: QQQ rose 3.1% the day after that election, then fell 18.3% to the Dec 24 low and ended the year 12.1% below that post-election close. On the same ratio, this year would end around $608. And a Dec 9 Fed hike is still more likely than not (about 69%).
Why not turn more bearish: the macro data isn't bad (jobless claims 189K, productivity +3.6%, unit labor costs −0.7%), and rates and hike bets are moving in a helpful direction — December hike odds fell from about 81% to about 69% in a single day, and the 2-year yield fell 11bp. Retail “capitulation” selling and the biggest one-day VIX jump in over two years led one fund manager to call this the “capitulation phase” for data-center capex beneficiaries. Measured from end-October, QQQ rose in the 6 and 12 months after all 6 previous midterms.
So: bull 20%, base 40%, bear 40%. The next line is $620 (the mid-April closing area) — a close below it makes the bear case the main scenario; reclaiming the 200-day moving average (about $680.7) brings the bear probability back to 35%. After that, watch payrolls on Nov 6, CPI and CoreWeave on Nov 10, Nvidia on Nov 17 and the Fed on Dec 9.
Nvidia beats and raises, H200 shipments to China are approved, and the political talk stays talk; CPI cools and hike odds keep falling, and the Fed hikes in December but signals it's the last. Chips rebound from capitulation-style selling, and QQQ reclaims its 200-day average, returning to around $720.
Chips stop falling but can't lead, and platforms hold up; the Fed hikes on Dec 9 as priced; Brent stays between $85 and $100. QQQ swings in a $610–680 range and ends the year slightly above today.
Nvidia's guidance disappoints, or the megacaps slow 2027 capex to repair cash flow; data-center pauses spread and neocloud credit stress spreads; the Fed hikes even as hiring cools. On the 2018 ratio the year ends around $608, and the low along the way could be deeper (about $566 on the ratio to the Dec 24, 2018 low).
Judgments go wrong, so here's what would change this one, written down in advance. If any of these trigger, the scenario probabilities get reshuffled. The next version will score each one.
| Indicator | Now | Bullish signal | Bearish signal |
|---|---|---|---|
| QQQ close | $647.77 (below $650) | Reclaims the 200-day moving average (about $680.7) → bear back to 35% | Closes below $620 → bear becomes the main scenario |
| Nvidia earnings (Nov 17, after the close) | — | Revenue and guidance beat; H200 shipments to China approved | Guidance misses, H200s stay stuck, or customers delay orders |
| Data-center & AI policy | Exit poll: about 76% oppose new local data centers; a lawmaker calls for an AI pause | Stays at the state level and in rhetoric; major projects proceed | Congress introduces pause or restriction bills, or more states follow |
| AI credit (CoreWeave reports Nov 10) | −20% since Oct 23 | Smooth refinancing; customer contracts stable | Financing stalls, or stress spreads to investment-grade credit |
| Brent crude | ≈$97 | Falls below $85 and stays there | Back above $110 (talks collapse or conflict escalates) |
| 10-year Treasury yield | 4.83% | Below 4.75%, with inflation expectations falling too | Back above 5.0% |
| December hike & core inflation (October CPI, Nov 10) | Hike odds ≈69% (about 81% this morning) | Core CPI m/m ≤ 0.2%, and hike odds fall below 50% | Core CPI m/m ≥ 0.4%, or the Fed signals several more hikes in 2027 |
The election is behind us. Using end-October — the last month-end before the vote — as the starting point, here is how QQQ did after the 6 midterms from 2002 to 2022: up in every 6- and 12-month window, but December itself was weak.
| Year | End-Oct distance from 12-month high | Nov | Dec | Next 6 months | Next 12 months |
|---|---|---|---|---|---|
| 2002 | −38.1% | +12.9% | −12.1% | +11.8% | +43.3% |
| 2006 | 0.0% | +3.4% | −2.0% | +7.9% | +29.2% |
| 2010 | 0.0% | −0.2% | +4.6% | +13.2% | +11.1% |
| 2014 | 0.0% | +4.5% | −2.6% | +6.1% | +11.8% |
| 2018 | −9.0% | −0.3% | −8.9% | +11.6% | +16.1% |
| 2022 | −30.1% | +5.5% | −9.2% | +16.0% | +26.2% |
| 2026 | −6.7% | −6.2%* | ? | ? | ? |
In the 3 years that ended October more than 3% below their month-end high (2002, 2018, 2022), the next 12 months averaged +28.5%; the 3 years at their high (2006, 2010, 2014) averaged +17.4%. 2026 ended October 6.7% below its month-end high, which puts it in the first group. On the political configuration, a second-term president facing an opposition-controlled Congress after the midterms happened in 2006 and 2014; the next 12 months returned +29.2% and +11.8%. With only 6 observations, treat this as a reference, not a law — and these are 12-month statistics, not December ones: after a −8.9% December in 2018, the next 12 months were still +16.1%.
QQQ month-end closes (2002–2022), starting at end-October; “distance from high” uses the trailing 12 month-end closes. * Through the Nov 5 close. The Nov–Dec statistic also includes 1998 (NDX, +31.1%); 1994 lacks a verifiable October value and is excluded.
Three lenses, three answers.
Still the smallest Jan–Oct path error (sum of squared errors 476; next is 2014 at 685), and July–October is nearly identical (just 24). The post-election path rhymes too: QQQ rose 3.1% the day after the 2018 election, then fell 18.3% to the Dec 24 low (including −9.4% in the four sessions starting with the Dec 19 hike); this year it rose 1.0% the day after the election and fell 6.4% the next day. The difference: core inflation was about 2% then, so the Fed could pivot within a month; now it's 3.1%.
A second-term Republican president whose party lost both chambers in the midterms, with oil having fallen sharply from its summer high in the months before the vote. QQQ rose 3.4% in November and fell 2.0% in December, then gained 29.2% over the next 12 months. The difference: the Fed had stopped hiking in June 2006; this year it's still hiking.
A rally driven by a long tech cycle that ran into a credit shock (LTCM) in the fall; the Fed cut three times, and the Nasdaq-100 gained 31.1% in November–December. This year's neocloud credit stress has echoes of that, but the Fed is hiking, not cutting — so a 1998-style year-end sprint is the least likely of the three.
Put together: the path looks like 2018 and the political configuration like 2006 — and today's drop tilts the balance toward 2018. The final direction depends on whether AI hardware stabilizes after Nvidia's report, and whether the Fed pivots after a December hike the way it did in early 2019 — with core inflation at 3.1%, a pivot needs more evidence.
Eleven midterm years overlaid on one Jan–Dec axis. 2026, 2018 and 1998 are in color; the rest are gray context lines (hover for each year's values). The 2026 line runs through the Nov 5 close.
Every version's calls, checked against real data.
| Indicator | Thresholds then (bullish / bearish) | Result |
|---|---|---|
| Megacap earnings | Cloud accelerating, capex mainly from cash flow / capex up while cloud slows, or a shift to debt or equity financing | Both touched: cloud growth accelerated sharply (Google Cloud +103%, Azure +48%, AWS +39%), but Alphabet's and Amazon's free cash flow turned negative |
| AI credit | Spreads tighten / stress spreads to big cloud providers or investment-grade | Not triggered: CoreWeave stock −20% since Oct 23, but no reports yet of contagion to investment-grade credit |
| Brent crude | < $100 / > $130 | Bullish triggered: −8% on Nov 4 to about $95 |
| 10-year Treasury | < 4.75% with inflation expectations falling / > 5.25% | Not triggered: peaked at 5.04% intraday on Oct 28; about 4.83% at the Nov 5 close |
| December hike odds | < 50% / signals several 2027 hikes | Not triggered: fell to about 69% on Nov 5, still above 50%; Warsh offered no forward guidance |
| Core inflation (September PCE) | m/m ≤ 0.2% / ≥ 0.4% | Not triggered: 0.3% m/m and 3.1% y/y, in line |
The level call broadly held: the Oct 25 probability-weighted $698 was within 1% of the Nov 4 close of $691.95. The biggest miss was treating the election as just a calendar event. It changed markets directly through two channels: oil — the previous version treated post-election Iran as a risk, and the opposite happened as markets judged escalation less likely; and data-center politics — a policy risk to the AI buildout that wasn't on the list at all, and one of the drivers of QQQ's worst day since March 2020 on Nov 5.
Cumulative monthly change (%, vs. prior year-end close). 2026 Jan–Oct are actual month-end values, November is the Nov 5 close, and December and the full year are the base case (italic).
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | FY |
|---|
≈ estimated monthly path (full-year return from published data); † some months anchored to NDX month-end values, others estimated; * Nov 5 close.
Data: Month-end, daily and intraday prices for QQQ and for individual stocks and ETFs, as well as the NDX and VIX index levels, come from Robinhood (split-adjusted, excluding dividends). Macro, earnings and election items come from the public reports and official releases listed below, as of the Nov 5 close.
On the judgment: The scenarios and probabilities here are reasoning from historical analogs and current macro conditions — not a forecast, not a recommendation, and not investment advice. Section 09 scores the previous four versions; weigh the method accordingly. The author is not a licensed investment adviser; do your own research and consult a professional before making investment decisions. Election results are cited from media projections and reports; this page discusses only market effects and takes no position on candidates, parties or the outcome. Items about Iran are cited from market reports only and imply no position.
Sources: