Two Clocks
Late 2026 against 1973 and 2000: a cycle map and scenario stress test for 2027–2031.
15, November, 2026
FACTS Research, version 1.0. Data as of 14 November 2026. This paper maps position and stress-tests paths. It is not a forecast and it is not investment advice.
Summary
Late 2026 runs on two clocks at once. One is an energy clock: a war-driven oil shock has landed on inflation that was already above target, and the Federal Reserve has started raising rates again. The other is a valuation clock: US equities trade at a cyclically adjusted price-to-earnings ratio last seen in 1999–2000, carried by an AI investment cycle whose returns have not yet shown up in most companies' accounts.
The closest historical match for the first clock is late 1973. The closest match for the second is March 2000. Neither episode had both. Neither had a federal debt above $40 trillion or a central bank hiking into an equity market at these valuations.
That combination is the finding. In 2001 the Fed cut rates eleven times. In 1975 policy turned loose and inflation came back. Late 2026 has the valuation exposure of 2000 and the inflation constraint of 1974, so the rescue that followed 2000 is not available on the same terms.
- Position. If equity valuations reset, the reset is unlikely to be met with 2001-style monetary relief, because inflation and long-term yields constrain it.
- The swing variables. Oil, and whether the bond market absorbs an equity reset without a rise in long yields.
- The reference path. A drawn-out repricing over 2027–2028 rather than a single crash, followed by either broad AI deployment or a populist, stagflationary turn.
- What would prove us wrong. Brent settling in a $75–$85 range with inflation expectations falling, or AI investment returns becoming visible across ordinary companies, would weaken the case for constraint. Section 9 lists the markers.
1. How to read this paper
We do not forecast. The system is reflexive: its participants act on their readings of it, and the fastest part of it, confidence, can run ahead of the slower parts for long stretches. What can be done honestly is to locate position, show the reasoning, and name the evidence that would change the reading.
The method has three steps.
- Read position on three axes. Capital and claims: valuations, debt, credit. Conversion: the stage of the technology cycle and whether investment is turning into output. Dispositions: confidence, trust and political mood, read from surveys and elections rather than from commentary.
- Screen history for matching shapes. We compare late 2026 against candidate episodes on each axis and keep the two closest, plus one modifier.
- Run the matched trajectories through today's constraints. What followed each match is a guide to dynamics, not to dates. Where today's constraints differ, we say how the path changes.
Every figure in this paper traces to a source we have read. Where only a secondary report was available, we say so. Figures we could not verify were removed rather than softened.
2. Where we are: the institutional record
Monetary policy has turned. The Federal Reserve, chaired by Kevin Warsh, raised its target range by a quarter point to 3.75%–4.00% at its September meeting. On 28 October it held that range by a 10–2 vote. Governor Lisa Cook and Cleveland Fed President Beth Hammack dissented in favour of another increase [1, 2]. Futures markets put the odds of a hike at the 9 December meeting at roughly 85% after the October inflation report [3].
Inflation is moving the wrong way. US consumer prices rose 3.7% in the year to September. October's monthly rise was 0.4%, above expectations, with core prices up 0.3% [3, 4].
Long rates are near 5%. The 10-year Treasury yield touched about 5.3% in mid-October and was back near 5% after the October inflation report on 10 November [5, 6].
The debt is large and expensive. Gross federal debt passed $40 trillion on 18 August 2026, five months after crossing $39 trillion [7]. Interest on the debt now costs more than a trillion dollars a year [8].
Global growth is holding, and so is inflation. The IMF's October World Economic Outlook projects global growth of 3.1% in 2026 and 3.4% in 2027, with global inflation rising to 4.7% in 2026 before easing to 4.1% in 2027 [9]. Its forecasts assume oil and gas shipments through the Strait of Hormuz resume by March and that the AI investment boom moderates [10]. Inflation is projected higher in emerging and developing economies (5.7%) than in advanced ones (3.1%) [11].
Oil spiked, then fell fast. Brent began 2026 near $61 a barrel [12]. After the war with Iran began in late February, it peaked several times. In late October it briefly topped $117 [13]. It then lost 16% in a single week to 9 November, reaching its lowest level since late August, as Gulf exports recovered and JPMorgan estimated a global surplus of about 1.4 million barrels a day in October [14]. Refined products remained tight [15].
Valuations sit where they sat in 1999. The S&P 500's cyclically adjusted price-to-earnings ratio stood at 41.07 in October 2026 on Multpl's monthly series, against a peak of 44.19 in December 1999 and a long-run mean near 17 [16]. It has stayed above 40 since May 2026, a level held for a sustained period only once before [17].
Credit stress has moved into private markets. Fitch's US private credit default rate, tracking about 1,300 borrowers, reached a record 6.3% for the twelve months to August, and held there for the twelve months to the third quarter [18, 19].
The AI build-out is running ahead of its cash. Goldman Sachs estimates global AI investment will exceed $1 trillion in 2026. Bank of America forecasts combined free-cash-flow deficits at eight hyperscalers and neoclouds of $128 billion in 2026 and $335 billion in 2027, narrowing to $77 billion in 2028 [20]. On the demand side, MIT's NANDA study found that only 5% of organisations it examined had taken embedded or task-specific generative AI tools to a successful implementation, defined as marked and sustained impact on productivity or profit within about six months. The authors call the findings preliminary, and the sample is small: 52 interviews, 153 survey responses and a review of more than 300 public initiatives [21].
| Domain | Reading | As of | Source |
|---|---|---|---|
| Fed funds target | 3.75%–4.00%, held 10–2 | 28 Oct 2026 | [1] |
| US CPI | 3.7% y/y (Sep); +0.4% m/m (Oct) | Nov 2026 | [3, 4] |
| 10-year Treasury | about 5.3% peak (mid-Oct); near 5% (10 Nov) | Nov 2026 | [5, 6] |
| Gross federal debt | above $40 trillion | 18 Aug 2026 | [7] |
| Global growth / inflation (IMF) | 3.1% / 4.7% (2026) | 13 Oct 2026 | [9] |
| Brent crude | above $117 peak; -16% in week to 9 Nov | Nov 2026 | [13, 14] |
| Shiller CAPE | 41.07 | Oct 2026 | [16] |
| Private credit defaults (Fitch) | 6.3% TTM, record | Q3 2026 | [19] |
3. The dispositional field
Institutional data describe an economy that is growing. Household data describe one that is under strain.
Sentiment is near its record low. The University of Michigan's consumer sentiment index hit an all-time low of 44.8 in May 2026. October's final reading of 48.0 was the second lowest on record, and buying conditions for durable goods hit a record low that month [22, 23]. The preliminary November reading rose to 51.0, still about 10% below February, just before the war began [24]. Year-ahead inflation expectations eased to 4.3% from 4.7%, and five-to-ten-year expectations to 3.4% from 3.5% [24]. One detail matters more than the headline: in November, sentiment among the largest stockholders weakened as concern about an AI bubble entered their answers [25].
Necessities are hard to afford. In a Pew survey fielded from 20 July to 9 August 2026, 44% of US adults said they had trouble paying for at least one of four basic necessities in the past year: medical care, daily transport, rent or mortgage, or food. Among lower-income adults the figure was 71% [26].
Trust is near its floor. Gallup's June 2026 survey put average confidence in 14 core US institutions at 27%, one point above the record low set in 2023. Twelve of the fourteen were at or near their lowest readings. Confidence in large technology companies fell to a new low of 20%, from 32% when Gallup first asked in 2020, and 41% said they had little or no confidence in them [27, 28].
The electorate moved before the economy did. On 3 November, Democrats won control of both chambers of Congress. They flipped seven Senate seats, including Kansas, Iowa, Texas, Ohio, Maine, North Carolina and South Carolina, to hold at least 54 [29, 30]. The Associated Press tally on 8 November gave them at least 254 House seats and 28 governorships, with two governors' races uncalled [31]. Texas elected its first Democratic governor since 1990, in a race where the fight over data centres was a major issue [32]. The President has said he will veto Democrat-led bills [33]. Divided government begins in January 2027.
This is the gap the paper turns on. Markets priced AI optimism to near-record valuations while households recorded sentiment near the worst in the survey's history. In November the two began to converge from the top: the people with the most equity began to worry about it.
4. Shape on three axes
Capital and claims. Valuations are at the level of 1999–2000. Long rates are near 5%, so the earnings yield on equities sits below the yield on a risk-free government bond. Debt is above $40 trillion and interest exceeds a trillion dollars a year. Default stress is concentrated in private credit, where marks are least visible. The record that distant investors read, prices and reported values, is running well ahead of the conversion it claims against.
Conversion and the technology cycle. In Carlota Perez's account of technological revolutions, each surge moves from an installation period, driven by financial capital and ending in a frenzy, through a turning point of crash and institutional reform, into a deployment period led by production [34]. AI infrastructure in late 2026 shows the marks of late installation: investment above $1 trillion a year, financed increasingly with debt and vendor arrangements, against enterprise returns that most firms cannot yet show. The labour signal is early but visible. Stanford researchers found employment for workers aged 22–25 in the most AI-exposed occupations fell 13% relative to less exposed peers after late 2022 [35]. Physical limits, power above all, have made data centres a local political issue.
Cohesion. Peter Turchin's structural-demographic theory links long waves of political instability to popular immiseration, an oversupply of credentialed aspirants competing for too few elite positions, and fiscal strain on the state [36]. The theory is contested and its empirical tests are mixed. We use it as a lens rather than a law. Its markers are present: affordability strain across nearly half of adults, institutional trust near its floor, and a midterm swing large enough to end one-party control of Washington.
5. Historical screening
No single episode matches every axis. Four candidates were screened.
| Candidate | Valuation and capex | Energy and inflation | Fiscal and credit | Politics and mood | Verdict |
|---|---|---|---|---|---|
| 1893 (Panic of 1893, after the McKinley Tariff) | Railway overbuild on shaky finance; valuations far below today | Opposite: gold-standard deflation | Run on Treasury gold reserves | Very close: tariff-driven living costs, agrarian and labour populism, an 1894 midterm rout | Modifier. Captures the tariff-populist realignment |
| 1929 | Speculative boom, but valuations below today's | Mismatch: low inflation under gold | Mismatch: low federal debt | Mismatch: euphoria at the peak, not pessimism | Rejected |
| Late 1973 – early 1974 | No frontier technology boom | Close: Middle East oil shock on top of inflation already above 7% | Post-Bretton Woods deficits, but debt far lower than today | Close: collapsing confidence, distrust of policy | Match 1. Energy, inflation and mood |
| March 2000 | Close: CAPE above 40, debt-financed telecom build-out ahead of demand | Mismatch: low inflation, no energy shock | Opposite: a federal surplus | Mismatch: high consumer confidence at the peak | Match 2. Valuation and the capex-return gap |
Late 1973 is the closest energy and inflation match. The October 1973 embargo hit an economy where inflation was already above 7% after years of loose fiscal and monetary policy. The recession that followed ran from November 1973 to March 1975, sixteen months, the deepest since the 1930s [37]. Late 2026 has a war-driven oil shock, above-target inflation, a Fed tightening and households near record pessimism.
March 2000 is the closest valuation match. The Nasdaq peaked at 5,048.62 on 10 March 2000 and the S&P 500 at 1,527.46 on 24 March, with the Shiller CAPE having peaked at 44.19 the previous December [16, 38]. The telecom build-out of the late 1990s laid capacity, much of it debt-financed, well ahead of demand. The Fed raised rates to 6.50% by May 2000 [38].
1893 is the modifier. Neither 1973 nor 2000 combined steep tariffs with mass affordability strain and political realignment. 1893 did. Railway overbuilding coincided with the 1890 McKinley Tariff and working-class distress, and in 1894 the governing Democrats lost more than 100 House seats [39].
6. What followed the two matches
After late 1973
The recession ended in March 1975. Unemployment kept rising after output turned and peaked at 9.0% in May 1975 [37]. The S&P 500 lost close to half its value between January 1973 and October 1974.
The policy response is the lesson. Policy loosened while inflation expectations were still unanchored and deficits persisted. Real short-term rates stayed negative. Inflation came back, reaching double digits by 1979, and it took Paul Volcker's rate increases, taking the federal funds rate close to 20%, to break it [40]. A sharp shock became a decade-long problem because relief came before credibility.
After March 2000
There was no single crash day. The decline rolled on for thirty-one months. The Fed cut rates eleven times in 2001, from 6.50% to 1.75%, and the Nasdaq still fell 78% to 1,114.11 by 9 October 2002, with the S&P 500 down 49% [38]. Owning the eventual winners did not protect investors: Amazon fell more than 90% from its 1999 high. The Nasdaq did not regain its March 2000 close until 2015.
The recession itself lasted eight months, from March to November 2001. The aftermath lasted longer. The bust exposed accounting fraud and hidden leverage at Enron (December 2001) and WorldCom (July 2002), and produced the Sarbanes-Oxley Act in July 2002. The infrastructure survived and, cheaper, carried the platform companies of the following decade. The easy money that cushioned 2001 also fed the housing and credit bubble that followed.
What is different in 2026
| Modifier | 1973–74 and 2000 | Late 2026 | Effect on the path |
|---|---|---|---|
| Fiscal room | 1970s debt far lower; 2000 ran a surplus | Debt above $40 trillion; interest above $1 trillion a year | Limits 2001-style fiscal support; puts a floor under long yields |
| Monetary room | In 2001 the Fed cut to 1.75% with inflation below 3% | The Fed is raising rates into 3.7% inflation | Early cuts risk a 1970s-style second wave of inflation |
| Energy | Supply tightened again into 1979 | Gulf exports recovering; Brent fell 16% in a week | Cheaper oil could ease inflation in 2027 if the war stays quiet |
| Politics | 2000 had free trade; 1974's midterm followed a bear market | Tariffs in place; the midterm swing came with stocks near highs | Divided government shields Fed independence but sharpens oversight of AI and private credit |
| Labour | 2000 hit programmers; 1974 hit industrial workers | Early pressure on entry-level white-collar work | Feeds demands for structural change |
7. Scenario stress tests, 2027–2031
These are conditional paths, not predictions. Each states the conditions under which it would unfold. The reference path is the one most consistent with today's position, not one we expect on a timetable.
| Scenario | Macro and inflation | Markets and AI | Credit, fiscal and labour | Politics |
|---|---|---|---|---|
| A. Rolling reset (reference path) | Growth below trend; inflation sticky above 3% before easing | Valuations compress over several quarters; AI capex growth slows below current expectations | Private credit defaults rise through the 2027–2029 maturities; long yields stay near current levels; white-collar hiring freezes | Congressional oversight of AI infrastructure and private credit; state limits on data centres |
| B. Orderly deployment | A brief 2027 slowdown, then recovery; Brent in a $75–$85 range pulls inflation toward target | Rotation from AI hardware toward software and grid investment; returns broaden | Long yields ease; enterprise AI returns become visible | Price relief lowers acute hardship; policy turns to workforce transition and grid upgrades |
| C. Fiscal-credibility break | Recession with inflation at 5% or more as oil spikes again | A deep equity and AI-asset decline; forced selling by leveraged funds | Long yields rise sharply; credit and emerging-market currency stress coincide | An 1890s-scale political rupture: price controls, tariff walls, demands for system overhaul |
A. Rolling reset. If inflation stays in a 3%–4.5% band, the Fed holds policy tight and the gap between a 5% bond yield and a much lower equity earnings yield closes the slow way, through falling prices rather than falling rates. A Democratic Congress investigates data-centre power costs, private credit opacity and pricing. Hyperscaler spending growth slows, exposing vendor-finance arrangements and writing down specialised leases. Middle-market borrowers meet the 2027–2029 refinancing wall at high rates, and private credit defaults rise past August 2026's record. As in 2001–2003, a disclosure reform analogous to Sarbanes-Oxley could follow for non-bank credit and AI infrastructure. By 2029–2031, lower valuations and cheaper compute let deployment lift productivity outside technology, with long yields still capping growth.
B. Orderly deployment. If the November fall in oil holds and Brent settles near $75–$85, energy and transport costs drain out of inflation without a deep recession. Divided government restrains new spending, and long yields drift lower. The Fed pauses and moves toward neutral without a 1975-style rebound. Inside AI, capital moves from training races toward uses that show measurable returns, and the share of companies reporting real gains rises well above today's low base. Concentration unwinds through earnings catching up, not through an index crash.
C. Fiscal-credibility break. If hostilities resume and close Hormuz again, oil above $115 pushes inflation toward 5% or more and forces the Fed into aggressive increases while Treasury issuance stays heavy. Long yields jump, the equity valuation anchor breaks, and leveraged funds sell what they can. Emerging markets with near-term dollar debts face capital flight. Under acute household stress, the political mood visible in 2026 turns into an 1894–1896-scale rupture over 2028–2030, with pressure for price controls, redistribution and restrictions on AI and offshoring.
8. Feedback loops that decide the path
The wealth-effect and affordability scissors. Spending by upper-income equity holders has carried aggregate demand while 71% of lower-income adults report strain on necessities [26]. An equity decline cuts the top end of spending at the moment living costs are still squeezing the bottom. Both ends of the K-shaped economy weaken together.
The thin-cushion loop. Households under strain have little buffer. Even moderate white-collar layoffs in 2027 would land on balance sheets with no room, accelerating defaults and political anger. Each feeds the other.
The premature-relief trap. With institutional confidence at 27% [27] and inflation above target for several years, political pressure on the Fed to cut during an equity sell-off would feed straight into long-run inflation expectations and push long yields up, cancelling the relief. This is 1975's mistake under 2026's debt load.
9. What would change the reading
These markers distinguish the paths over the next 24 months. Each is observable.
- The 9 December decision and the first months of divided government. A hike to 4.00%–4.25% that markets absorb calmly, and budget deadlines met without protracted standoffs, point away from Scenario C.
- The 10-year yield regime. Sustained trading above 5.5% while growth softens points to Scenario C. A descent below 4.5% led by falling inflation points to B.
- 2027 hyperscaler capex guidance. Downward revisions or cancelled data-centre leases mark the 2000-style turn toward cash discipline (A or C). Guidance held with visible free-cash-flow improvement weakens our reading.
- Private credit defaults and fund gating. Fitch's default rate rising past 6.3% alongside widespread limits on investor redemptions marks the start of a restructuring wave.
- Brent's settling range. $75–$85 supports Scenario B. A renewed move above $110 traps policy in A or C.
- Long-run inflation expectations and young white-collar employment. Michigan's five-to-ten-year expectations breaking above 3.8% would signal 1970s-style unanchoring. A deepening of the employment gap for young workers in AI-exposed jobs would confirm the labour channel.
- State data-centre permitting pauses and congressional oversight. Their enactment measures how fast social grievance is hardening into supply constraints.
If oil settles low, inflation expectations fall and AI returns broaden beyond a few companies, the constraint at the heart of this paper loosens, and Scenario B becomes the reference path. We will revisit each marker as data arrive.
10. Limits
Historical matching maps tensions and feedback, not calendar dates. Late 2026 has already compressed sequences that took years in earlier cycles: the electorate swung before any recession, and oil fell back toward surplus while inflation was still rising. Several 2026 figures rest on secondary reporting of primary data, noted in the sources. Survey data describe moods, which move fast. The scenario ranges are stress tests bounded by historical precedent, not probabilities.
Sources
Grades follow our casebook standard: A primary, B independent secondary.
- Federal Reserve, FOMC statement, 28 October 2026 (A). link
- Wealth Management, "Fed Holds Rates, Says More Data Will Help Next Decision", 28 October 2026 (B). link
- Beinsure, "US Treasury yields surge as October CPI inflation exceeds forecasts", November 2026, reporting BLS data (B). link
- New York Times via DNYUZ, "Inflation runs hotter than expected at 3.7% in September", 14 October 2026 (B). link
- Trading Economics, US 10-year yield news, 14–15 October 2026 (B). link
- MarketWatch via Morningstar, "Stocks fall as the 10-year Treasury yield reclaims 5%", 10 November 2026 (B). link
- Fox Business, "US national debt hits $40 trillion milestone for first time ever", 19 August 2026, reporting Treasury data (B). link
- Michigan Public / NPR, "The U.S. debt tops a record-shattering $40 trillion", 19 August 2026 (B). link
- IMF, World Economic Outlook, October 2026, Executive Summary (A). link
- AP via KSAT, "IMF raises its forecast for world growth this year but warns of risks from AI boom and debt", 13 October 2026 (B). link
- Australian Industry Group, "Global forecasts October 2026", summarising the IMF WEO tables (B). link
- Al Majalla, "How the US-Iran war took oil prices on a wild ride" (B). link
- Al Jazeera, "Oil flows are up as Iran war turmoil eases. Why do prices keep rising?", 30 October 2026 (B). link
- The Nation (Thailand), "Brent losses top US$7 a barrel after November 9 settlement", 10 November 2026 (B). link
- Hydrocarbon Engineering, "IEA: the latest developments in oil markets", 26 October 2026 (B). link
- Finnovate, "Shiller CAPE ratio above 40", 9 October 2026, citing Multpl's series built on Robert Shiller's data (B). link
- Angel Investors Network, summarising The Motley Fool on CAPE above 40 since May 2026 (B). link
- Investment Executive, "U.S. private credit defaults rise: Fitch", September 2026 (B). link
- ABF Journal, "Fitch Ratings: U.S. Private Credit Default Rate Remains Elevated in Q3", 30 October 2026 (B). link
- TheStreet, "Bank of America has unexpected verdict on AI stocks after post-midterm Democrat sweep", 4 November 2026 (B). link
- eCommerce Fastlane, "The MIT 95% AI Failure Stat: What the Study Actually Found", summarising Challapally et al., The GenAI Divide, MIT Project NANDA, July 2025 (B). link
- Reuters via GV Wire, "US Consumer Sentiment Near Record Low Ahead of Midterm Elections", 23 October 2026 (B). link
- TheStreet Pro, "Consumer Sentiment Rebounds to a Three-Month High as Inflation Expectations Ease", 6 November 2026 (B). link
- Advisor Perspectives, "Consumer sentiment recovers significantly", 6 November 2026 (B). link
- Finimize, "US consumer mood picked up but AI bubble fears crept in", November 2026 (B). link
- Pew Research Center, "Many Americans Are Struggling to Afford Basic Necessities", 14 October 2026 (A). link
- Gallup, "Confidence in U.S. Institutions Remains Near All-Time Low", July 2026 (A). link
- Daily Caller, "Trust in big tech drops to record low: Gallup poll", 14 July 2026, reporting Gallup data (B). link
- NBC News live updates, 4 November 2026 (B). link
- CBS News, "Where Democrats won big in the 2026 midterms" (B). link
- California Today, "2026 Midterm election results master guide", AP tallies as of 8 November 2026 (B). link
- CBS News, as [30] (B).
- Washington Examiner, "Takeaways from the 2026 midterm elections" (B). link
- Carlota Perez, Technological Revolutions and Financial Capital, Edward Elgar, 2002.
- Erik Brynjolfsson, Bharat Chandar and Ruyu Chen, "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence", Stanford Digital Economy Lab, August 2025.
- Peter Turchin, End Times: Elites, Counter-Elites, and the Path of Political Disintegration, Penguin, 2023.
- NBER business cycle dating; Federal Reserve History, "The Great Inflation".
- Federal Reserve History and market records for the 2000–2002 bear market and the 2001 easing cycle.
- Federal Reserve History, "Banking Panics of the Gilded Age".
- Federal Reserve History, "The Great Inflation".
FACTS Advisory Ltd. This paper locates position and stress-tests paths. It does not forecast and it is not investment, legal or tax advice.