Video summary

Is The Market Dangerously Expensive Now? | Kevin Muir

Main summary

Key takeaways

Finance

Markets Macro: “Too Expensive” Equities, Lost-Decade Risk, and AI/Capex Caution

Core thesis: markets are priced for perfection

Kevin Muir argues equities are priced for perfection and that risk is higher than during the COVID lows, implying lower forward returns and possible “lost decade” outcomes.

Equity Risk Premium framing

He uses the equity risk premium concept by comparing:

  • S&P 500 earnings yield vs

  • 10-year Treasury yield

Examples mentioned:

  • March 2020: S&P 500 earnings yield ~6% vs 10-year ~1% → a very favorable setup for long-run equity outperformance.
  • Current setup (as stated): 10-year ~4.5% vs earnings yield ~3.5% → equity risk premium is “almost completely the reverse,” implying equities are priced too expensively.

Valuation extremes as evidence

He cites valuation indicators to support the “expensive” claim:

  • Buffett Indicator: market cap vs GDP at record highs.
  • CAPE / Shiller CAPE: at/near levels associated with the dot-com bubble.

Explicit recommendation: get defensive

His recommendation is that investors should be thinking defense—specifically:

  • Reduce equity exposure if you can tolerate possible underperformance for about a year rather than adding more risk assets.

“Lost decade” probability: “probably what you should assume”

Muir’s view is that a lost decade is “probably what you should assume,” and he calls the probability “huge.”

Drivers cited

  1. Valuation implies weak long-run returns Even if the next year can still be strong, starting valuations reduce expected long-run payoff.

  2. Concentration risk in the S&P 500 He warns that the “top 10” are a much larger share than before, referencing volatility in MAG 7-type concentration and historical ~50% corrections in these stocks.

Simple math example (as stated):

  - If the **top 10 = 40%** of the S&P 500
  - and those names suffer **50% corrections**
  - then that concentration can imply roughly **~-20%** on the index mechanically (from those holdings alone).

  1. Semiconductors are “super cyclical” He suggests markets may underestimate how quickly earnings/profits can roll over. He also cautions against assuming “low P/E” automatically equals undervaluation, because Wall Street may be expecting earnings normalization eventually.

    • His suspicion: rollover could be faster than the market assumes.

Other referenced estimates

  • John Hussman: forecasts negative average market returns over the next decade (contextually referenced).
  • Goldman Sachs (framed as “cheerleaders”): projecting low single digits / near-zero returns.

How a “lost decade” may look (and what could still go right)

The hosts caution that a lost decade doesn’t have to mean “flat and dead.” It may include large drawdowns.

Counter-case: sideways grind

A discussed possibility:

  • Returns could resemble a sideways grind for years.
  • Example mechanism: valuation compression (P/E falling) e.g., U.S. forward P/E moving from about ~22–23x toward ~18–16x.

Historical analogies

  • 1968–1982 Dow Jones: brutal period; inflation hurt real returns; long sideways behavior.
  • Tech and housing bubbles: opportunities can exist near bottoms, but drawdowns can be extreme.

AI/Capex bubble risk: “token mirage,” overbuilding, and real constraints

Muir’s AI argument emphasizes overconfidence + overcapacity.

Overbuilt capex and adoption skepticism

He claims AI is driving a major capex buildout, and that compute/software adoption may be overstated using short-term “utilization” style signals.

Depreciation and upgrade risk

He argues data centers/hardware may be short-lived relative to past infrastructure booms (like fiber/railroads). Replacement costs matter:

  • replacement/upgrade hardware/chips might cost about two-thirds of the initial buildout
  • cooling improves quickly, so older cooling equipment may become obsolete

Magnitude/statistics mentioned

  • A “study” claim that semiconductors/memory stocks added:
    • ~$4T market cap in 2025
    • ~$8T YTD in 2026
  • Annualized rate cited: ~$19T
  • Compared to U.S. GDP ~ $33T

“Token mirage” mechanism

He describes a possible mismatch between usage metrics and real business value:

  • Companies allegedly pushed engineers to use AI via leaderboards/performance reviews.
  • That can create demand/usage that doesn’t equal economic value.

Illustrative example mentioned:

  • An AI bill framed as ~$500 million of AI credits, used to contrast credit/usage economics with “consumer-style unlimited” intuition.

He warns that when policy shifts from “usage” to “outcomes,” markets could re-rate.

Physical constraints that could limit capex

Potential bottlenecks discussed:

  • permitted land constraints (municipal pushback)
  • engineer/contractor talent shortages
  • water availability
  • copper availability

Geopolitical / software substitution risk

  • China may succeed with more efficient software, reducing marginal hardware needs.
  • He references “DeepSeek”-type surprises to suggest new models can be “pretty damn good,” potentially lowering additional compute requirements.

Macro/Fed: more hawkish than markets expect

Muir rejects a “dovish Fed chair” narrative.

Inflation target and Fed behavior

He argues the Fed will prioritize returning to 2% inflation (interpreting speeches in terms of “left digit” to 2 and “right digit” to 0).

Inflation and labor references:

  • PCE inflation ~3.3%
  • Not at 2% since 2021
  • Employment improvement: “three decent months” referenced

Bond market implications

He frames curve dynamics as consistent with a more hawkish stance:

  • markets pricing different outcomes → yield curve flattens
  • long-end dynamics described as fitting the hawkish interpretation

Taylor Rule framework

He uses a Taylor rule lens:

  • compares policy stance implied by:
    • inflation gap vs target
    • unemployment gap vs natural rate
  • if the implied Taylor level is rising, the Fed must be more hawkish—and he argues recent data pushes it higher.

Positioning: defensive equities (New Harbor), hedges not fully implemented

New Harbor partners agree that defense is warranted due to valuation/macro risk.

Allocation / risk posture (as stated)

  • Equity exposure around 50% max for their system
  • They moved from about 48% equity to ~50%
  • Signals described as “whippy” with reversals

Hedging status (explicitly cautious)

  • They say they have not started “mental stops”
  • They have not taken equity off
  • They also say they haven’t added index options or index puts yet, despite a risk-management focus.

Relative-strength examples mentioned

  • Regional banks outperforming (ETFs referenced: IWM and JJR)
  • Biotech breaking out
  • Homebuilders up ~5% on a referenced day
  • REITs improving
  • Market trading described as in the “74/7500 range” (numeric precision unclear from the subtitles)

Oil outlook: bullish on reserves and geopolitics (SPRs)

Muir prefers:

  1. Oil
  2. then oil stocks

Mechanism: strategic petroleum reserves (SPRs)

He attributes the “oil didn’t go to $200” outcome to strategic petroleum reserves, especially:

  • mentioning China’s planning and reserve drawdowns

“Next surprise”: governments refill SPRs

He expects governments will refill SPRs after peace normalization, so oil may not fall as much as the market expects.

  • He argues oil likely won’t bounce quickly back to $40.

Energy equities vs crude

He prefers stocks over crude:

  • mentions buying energy stocks around ~16x earnings
  • notes the market was ~21x at the time
  • references war-era rally to ~20x earnings
  • argues pessimism implies ~12x forward earnings now, making them look cheaper again
  • USO (United States Oil Fund) referenced as an example related to positioning/short interest

Dislocation thesis

In an equity correction:

  • energy may outperform “nobody owns it” areas
  • short covering could contribute

Precious metals: long-term gold bullish, near-term silver volatility

Gold

Muir’s long-term gold thesis is tied to:

  • People’s Bank of China (PBoC) reserve diversification

He acknowledges:

  • Western “return-to-model” effects via the U.S. dollar and real yields but maintains the China-driven long-term focus.

When he would become less bullish:

  • if/when PBoC stops buying gold

He also notes:

  • miners can look cheap relative to bullion
  • preference: miners over the metal

Silver

He describes silver pullbacks as gut-wrenching but not declaring the trade broken.

Key references:

  • ETF: SLV (iShares Silver Trust)
  • Silver futures described as breaking below $60/oz
  • A prior breakout zone around $35
  • Later move to ~$110–115 (context)
  • SLV testing a “breakaway gap” level; a “high 40s” range cited as an invalidation threshold (imprecise in subtitles)

Historical precedent

  • Silver has previously seen about ~50% corrections, with examples like:
    • 1973
    • 2008–2009
  • sometimes followed by new highs

Positioning / risk limits (as stated):

  • bullion allocation target: 5–10%
  • current model: around ~10% in the “mining and bullion sleeve”
  • trimmed gold (and previously silver); mentions trimming about ~2.5% of gold after a technical breakdown

Psychological caution

  • warns of shakeouts: “this is trying to get people out,” implying investors may sell near bottom due to fear.

Portfolio rotation theme: “rolling mini bubbles” and where to look

Muir argues markets go through multiple mini-bubbles (examples given):

  • Bitcoin
  • EVs
  • SPACs
  • silver spikes

Advice:

  • don’t focus only on “page 1” hype—look for what’s earlier (“page 17”) before hype returns.

Candidate sectors listed

  • Agriculture
  • Refiners
  • Biotech

He also suggests liquidity/plumbing changes and momentum trading may contribute to the pattern.

Behavioral/risk framing: prospect theory and loss aversion

Hosts emphasize psychology:

  • Prospect theory: losses feel about 2x as painful as equivalent gains feel pleasurable.

Applied warning:

  • investors may chase extra risk (FOMO/greed) to “fix” losses after drawdowns—often at the wrong time.

Frameworks / methodologies referenced

  • Equity Risk Premium
    • using S&P 500 earnings yield vs 10-year yield to infer long-run equity expectations
  • Buffett Indicator
    • market cap / GDP vs historical extremes
  • CAPE / Shiller CAPE
    • valuation extremes and dot-com-like conditions
  • CAPE-like variant
    • described as a market-cap divided by corporate gross value-added style ratio (similar intent to predict subsequent returns)
  • Taylor Rule
    • policy stance implied by inflation gap and unemployment gap vs natural rates
  • New Harbor portfolio risk planning
    • retirement probability modeling under different return sequences
    • comparing assumptions using historical averages vs realized sequences (e.g., 2000–2025)

Key numbers and levels (as stated)

  • March 2020:
    • S&P 500 earnings yield ~6%
    • 10-year ~1%
  • Current context:
    • 10-year ~4.5%
    • earnings yield ~3.5%
  • Valuation signals: Buffett Indicator and CAPE at record/dot-com-like levels (no exact CAPE values given in subtitles)
  • Concentration example:
    • “Top 10” ~40% of S&P 500
    • if top names drop 50% → roughly ~ -20% mechanically
  • Forward P/E commentary (hosts):
    • ~22–23x forward (next 12 months)
    • possibly ~18–16x after normalization
  • AI capex/buildout market cap additions:
    • $4T (2025)
    • $8T YTD (2026)
    • annualized ~$19T vs U.S. GDP ~$33T
  • Fed/macro:
    • PCE ~3.3%
    • not at 2% since 2021
  • Silver/metals:
    • silver futures breaking below $60/oz
    • breakout zone ~$35
    • later move to ~$110–115
    • invalidation line cited as “high 40s” (imprecise)
    • bullion target 5–10%, model ~10%
    • gold trim: ~2.5% of gold
  • Oil:
    • discussion includes $200 not realized; oil argued potentially capped around $40–50 under peace normalization

Disclosures / notes

  • No explicit “not financial advice” line appears in the provided subtitle excerpt.
  • The discussion uses typical advisor-style caution (risk management, avoiding FOMO, allocation emphasis).

Tickers / assets / instruments mentioned

  • S&P 500
  • IWM (Russell 2000 ETF)
  • JJR (ETF mentioned; exact context unclear)
  • USO (United States Oil Fund)
  • SLV (iShares Silver Trust)
  • Concepts without tickers: SPACs, EVs, Bitcoin, Comex/deliveries (commodity market structure references)

Presenters / sources named

  • Adam Tagert (host; Thoughtful Money)
  • Kevin Muir (guest; “macro tourist”)
  • John Hussman (referenced)
  • Goldman Sachs (referenced)
  • Warren Pi / 314 Research (referenced)
  • Michael Sembly / JP Morgan Asset Management (referenced)
  • GMO / Grantham Mayo Van Otterloo & Co. (referenced via “GMO chart”)
  • New Harbor Financial (partners introduced)
    • John Lodra
    • Mike Preston
  • Daniel Kahneman (referenced for prospect theory)
  • Charlie Munger (incentives → outcomes reference)
  • Plaude (sponsor/product referenced)
  • Andy Sheckchman
  • Thoughtful Money / ThoughtfulMoney.com
  • New Harbor Financial / ThoughtfulMoney Endorsed
  • Worsh / Fed chair Warsh (referenced; likely a transcription error)

Original video