Howard Marks

Co-founder of Oaktree Capital Management (1995), one of the world's largest alternative investment firms specializing in credit and distressed debt. Author of The Most Important Thing (2011). Has written client memos since 1990 — over 100 memos across 35 years that constitute one of the most coherent bodies of investment philosophy in existence.

Core Philosophy

Marks calls his approach the "I don't know" school of investing: radical intellectual humility combined with rigorous attention to price, cycle position, and risk. The macro future is unknowable; what is knowable is where valuations stand and where we are in the market cycle.

His organizing principle: defense first. "If we avoid the losers, the winners will take care of themselves."

Two epistemic camps drive everything else: the "I Know" school (macro forecasters, confident strategists) vs. the "I Don't Know" school (Marks' own). The first school acts on predictions; the second calibrates exposure to where the cycle stands without needing to know when it turns.

Key Frameworks

Thinking & Decision-Making

Second-level thinking — Marks' most-cited contribution. First-level: "This is good — buy it." Second-level: "This is good, everyone knows it, the price reflects it — at this price, sell." Outperformance requires being both non-consensus and correct.

Anti-forecasting — Macro forecasts are systematically unreliable. Non-consensus forecasts must be both different and correct and acted upon to add value — a rare combination. Prefer observation to prediction. "The future is inherently uncertain; forecasters are systematically overconfident."

Two-layer forecasting failure — Marks' 2016 election memos sharpen the anti-forecasting argument: investors must predict both the event and the market reaction to the event. In November 2016, consensus was wrong on both layers: Trump won unexpectedly, and stocks rose instead of falling.

Decision quality vs outcome — Marks' 2020 memo "You Bet!" connects investing to games of chance: the goal is not certainty, but good probabilistic process under incomplete information. Outcomes are noisy feedback, not proof of decision quality.

Staged action under uncertainty — In "Nobody Knows II" (2020), Marks applies his process to COVID: when facts are limited and psychology is moving fast, avoid false precision. If price has improved relative to value, buy some; if uncertainty remains high, reserve capital to buy more later.

Epistemic humility - Marks' 2020-2022 memos turn "I don't know" into a broader discipline: separate facts from informed extrapolations and guesses, identify whether a topic is knowable, and express confidence only in proportion to evidence. This is especially clear in "Knowledge of the Future," "Uncertainty," "Uncertainty II," and "The Illusion of Knowledge."

Forecasting as assumption stacks - "The Illusion of Knowledge" and "What Really Matters?" argue that macro forecasts are chains of conditional assumptions. If an investor predicts A, then B, then C, then D, then security E, every link must be right. The more links, the less useful the forecast.

Active management as error detection — Marks argues that active investing is justified only when market mistakes exist and the investor can identify them. Distressed debt is his prime example: prior underwriting errors, excessive leverage, and panic selling create mispricings for disciplined buyers.

Passive investing reflexivity — Marks does not reject indexing, but argues passive investing depends on active investors doing price discovery. If passive flows become large enough, they can distort the very prices they rely on, especially in cap-weighted indices, smart-beta products, and ETF structures.

Computers vs judgment — Marks' 2018 view on passive, quant, and AI is that machines can improve investing where the task is countable, rules-based, data-rich, and short-term. The remaining human edge is qualitative judgment: management, incentives, long-term business quality, bankruptcy negotiation, private assets, and situations where the crucial variables are not in the dataset.

Market mind - In "Conversation at Panmure House" (2022), Marks restates that markets are psychological systems, not mechanical systems. Economics is not physics because people have feelings; prices can move from "flawless" to "hopeless" through mood changes as much as through fundamental changes.

"Should" ≠ "Will" — A recurring logical trap. Even when a thesis is correct ("home prices should fall"), the market may not correct on any predictable timeline. Marks navigated the 2004–2007 bubble acutely aware that "early" is indistinguishable from "wrong" in the short run. Being right eventually is not the same as being right profitably.

Economic Reality vs. Political Reality — In 2016, Marks extended his second-order thinking into policy and elections. Economic reality means scarcity, trade-offs, incentives, productivity, and second-order consequences. Political reality means voters and candidates are often rewarded for simplifying those trade-offs into costless promises. Brexit and the 2016 U.S. election become his live cases.

He returns to the same framework in 2019 with tariffs, the Trump-China trade conflict, wealth taxes, anti-capitalist rhetoric, rent control, and expanded government promises. The common question is always: what cost or second-order consequence is being hidden by the political slogan?

"This time it's different" detection — Marks treats broad acceptance of optimistic exception claims as a late-cycle warning. In 2019 the claims included no recession, permanent QE prosperity, harmless deficits/debt, lower-for-longer rates, profitless companies at huge valuations, and growth stocks beating value forever.

Unconventionality 2×2 Matrix — The only way to generate superior returns is to hold non-consensus views that turn out to be correct. This requires the willingness to be wrong and look foolish. A 2×2 of (consensus/non-consensus) × (right/wrong) shows that consensus positions produce average results at best; non-consensus-wrong positions destroy capital; only non-consensus-correct delivers alpha. "Dare to Be Great" (2006) codifies this.

Active bets are double-edged - "I Beg to Differ" (2022) sharpens the same point: every attempt to outperform requires departing from consensus, and every departure creates the possibility of underperformance. "Try to be right" always comes paired with "risk being wrong."

Long-term compounding vs market timing - In "Selling Out" (2022), Marks argues that selling because something is up or down is usually psychology masquerading as analysis. The valid selling question is relative selection: has the thesis weakened, or is there a better risk-adjusted use of the capital?

Reasonable expectations - In The Most Important Thing Illuminated, Marks adds a chapter arguing that unrealistic return expectations are themselves a risk source. If the promised result is too high, too smooth, or too dependable relative to alternatives, skepticism should activate.

Cycle & Risk

The Pendulum — Markets oscillate between euphoria and depression. Cyclicality is "one of the few constants." "Trees don't grow to the sky, and few things go to zero." The investor's job is to observe the pendulum's position, not predict when it turns.

Confidence Cycle — Confidence is self-reinforcing in economies and markets. When people believe the future will be good, they spend, invest, lend, hire, and buy assets; those actions help make conditions better. But excessive confidence lowers skepticism and demanded risk premiums, creating danger. Low confidence is unpleasant but can create safer prices.

Liquidity Risk — Liquidity is usually available when investors do not need it and absent when they do. Marks emphasizes that funds, ETFs, and market structures cannot be more liquid than their underlying assets during stress.

The Credit Cycle — More volatile than the underlying economy. "Look around at every crisis and you'll find a lender." Credit excess enables every boom; credit contraction causes every bust. The cycle operates via a Virtuous/Vicious Circle:

Virtuous (up): Rising prices → paper gains → more credit available → more buying → higher prices
Vicious (down): Falling prices → margin calls → forced selling → more supply → lower prices → more calls

Mark-to-market accounting acts as an accelerant: unrealized losses trigger real-world covenant breaches and capital requirements.

See credit-cycle.

Credit investing as negative art - Marks distinguishes credit from equity: equity investors can be paid for finding huge winners; credit investors are mostly paid for avoiding losers, because upside is capped and downside can be severe. The 2025 First Brands/Tricolor/private-credit scares reinforce that good credit research is mosaic-building and early error detection.

Perversity of Risk — Risk actually increases as markets rise, because rising prices compress future returns and attract more leverage. Conversely, after a crash, forward risk has diminished even as perceived risk peaks. "The protection of capital is incompatible with maximizing returns." Marks formalized this against the CFA/MPT definition of risk as volatility, arguing that the real risk is permanent capital loss — and it hides precisely when people feel safest.

Bubble detection - Marks evaluates bubbles by behavior rather than labels. The recurring signs are "no price too high" thinking, FOMO, lottery-ticket math, valuation indifference, circular financing, debt-financed buildouts, and newness that makes history feel irrelevant. In 2025 he applies this to AI while allowing that some "inflection bubbles" can accelerate real progress and still destroy investor capital. See bubble-detection.

Debt vs equity in technological revolutions - In the final 2025 AI memo, Marks argues that speculative winner-take-most races are better suited to equity than debt. Equity can survive many losers if it owns the winner; debt has capped upside and can still be impaired when the losers fail.

Low-rate world - Marks' post-2020 framework for why asset prices can be high, prospective returns low, and relative valuations still internally coherent. Near-zero rates pull down the whole capital market line, force return-seeking, reopen credit markets, and make future cash flows look more valuable. The vulnerability is higher inflation or higher rates.

Sea change in rates - Marks' 2022-2024 thesis that the four-decade decline in rates was a once-in-a-career tailwind, not a permanent law. The 2022 inflation shock and Fed hikes moved markets from an easy-money regime toward a more normal world where credit yields, underwriting skill, and bargain selection matter more.

Value vs growth investing - "Something of Value" (2021) records Marks updating his own value-investing instincts after conversations with Andrew Marks. Growth is not the opposite of value; growth is part of intrinsic value. The danger is not high multiples in themselves, but "no price too high" thinking.

Supply chain resilience - In "The Pendulum in International Affairs" (2022), Marks applies the pendulum to geopolitics: after decades of cheap global sourcing, Russia/Ukraine and semiconductor shortages revealed the cost of ignoring resilience, security, and dependence.

Calibration, not binary timing — Marks repeatedly rejects "get in/get out" calls. The practical question is where to sit on the aggressive-to-defensive continuum based on current valuations, risk appetite, credit terms, liquidity, and investor psychology. This is cycle observation, not macro prediction.

7 Reasons Risk Is Hard to Manage:

The future is inherently uncertain
Past data reflects survivorship bias
Models assume normal distributions; tails are fat
Risk concentrates in exactly the places models say it's absent
Correlation rises toward 1.0 in crises
Leverage transforms manageable volatility into ruin
Human psychology amplifies rather than dampens cycles

Crises & Contagion

Fire Sale Mechanism — In a liquidity crisis, the causal chain runs: leverage → price decline → margin call → forced sale → more supply → lower prices → more calls. This loop continues until either sellers are exhausted or a buyer with patient capital steps in. The rational response (holding) is unavailable to leveraged players — they must sell regardless of fundamental value.

Financial Innovation as Cyclical Symptom — Every credit boom produces financial innovations that are represented as risk-reducing but actually concentrate and obscure risk. CDOs, CLOs, structured credit in 2003–2007 didn't eliminate risk; they moved it to entities (SIVs, off-balance-sheet vehicles) that were least equipped to bear it. The Subprime Factory illustrates the incentive chain: borrower → originator → securitizer → rating agency → investor — seven parties, none with skin in the game.

3 Mechanisms of Crisis Contagion:

Leveraged forced selling — margin calls force asset liquidation, spreading losses
Collateral chains — falling asset prices reduce collateral value across all borrowers
Confidence collapse — counterparty uncertainty freezes credit markets (the "nobody knows who has the losses" problem)

Bear Market 3 Stages:

A few forward-looking investors recognize things have gotten ahead of fundamentals
Most participants acknowledge the deterioration
Everyone concludes things can only get worse (capitulation/maximum pessimism — the actual buying opportunity)

12 Lessons of Every Credit Crisis (from "Whodunit," 2008 and earlier memos):

Excessive lending is always the root cause
Financial innovation obscures rather than eliminates risk
Rating agencies are paid by issuers — structural conflict of interest
Leverage magnifies both gains and losses, non-linearly
Illiquidity is indistinguishable from insolvency in a panic
Correlations approach 1.0 when it matters most
"Contained" crises rarely are
Moral hazard grows with every bailout
Markets can remain irrational longer than you can remain solvent
The last crisis is always the template — but the next one is always different in form
Human nature (greed, fear, envy) is the constant; instruments change
Cycles will cease only when human emotion and profit-seeking no longer go to extremes

Forest Fire Analogy — Marks uses the ecology of forest fires to explain moral hazard: suppressing every small fire allows fuel to accumulate until a catastrophic fire is inevitable. Similarly, central bank rescues of every credit downturn prevent the clearing mechanism that keeps the system healthy, building conditions for a larger eventual crisis.

Short-termism as Structural Problem — Compensation structures in finance reward recent performance, creating incentives to sacrifice long-run risk management for short-run gains. Fund managers who underperform in a bull market lose assets under management before the cycle turns. The structural result: the whole industry systematically takes on more risk than it should.

Capital Allocation

Price vs. Value — "What matters most is not what you invest in but when and at what price." No asset is so good it can't be overpriced; no asset is so bad it can't be underpriced.

Leverage is dynamite — "Volatility + leverage = dynamite." Documented via LTCM post-mortem ("Genius Isn't Enough," 1998): 25x leverage wiped out on a 4% price move. "Two kinds of people lose money: those who know nothing and those who know everything."

Forced sellers = opportunity — In crises, falling prices drive out buyers (countereconomic behavior) due to margin calls, redemptions, and rating triggers. Patient capital with no forced-selling risk can buy at extraordinary prices. "Major bottoms occur when everyone forgets that the tide also comes in."

Investment Fashion Cycles — Marks repeatedly uses the hemline analogy: asset classes and strategies move in and out of favor. The key question is not whether stocks, bonds, gold, private equity, or any other category is "good," but whether its merits are already over-owned and over-priced. See investment-fashion-cycles.

Quants Critique — Quantitative models fail at exactly the moments they're most needed because they're calibrated on history that excludes the current tail event. "The models said 25-sigma events were once-in-the-universe occurrences; we got several in a week." Models mistake the map for the territory; they embed the assumption that markets are stationary, which they are not.

Metric skepticism — In "Lines in the Sand" (2017), Marks applies the same anti-simplification instinct to private fund reporting. Subscription lines can improve reported IRR by delaying capital calls without improving lifetime dollar gains or capital multiples. No single metric can answer whether a GP did a good job.

Landmark Memos

"The Route to Performance" (1990) — Foundational: avoid losers; defense > offense
"The Pendulum" (1991) — Introduces the cycle metaphor
"The Value of Predictions" (1993, 1996) — Proves forecasters are measurably, systematically wrong
"Risk in Today's Markets" (1994) — Leverage + volatility framework; perversity of risk
"bubble.com" (Jan 2, 2000) — Written at the exact dot-com top; five lessons on bubbles and valuation
"Genius Isn't Enough" (1998) — LTCM post-mortem; seven lessons on overconfidence and leverage
"The Realist's Creed" (2002) — Marks' own synthesis: contrarianism, skepticism, humility, defensive investing
"Risk" (2006) — Formalizes perversity of risk; risk ≠ volatility
"Dare to Be Great" (2006) — Non-consensus × correct 2×2 matrix; prerequisites for outperformance
"Race to the Bottom" (2007) — Documents loosening standards in real time; early GFC warning
"The Tide Goes Out" (2007) — Diagnoses credit excess; virtuous/vicious circle
"Whodunit" (2008) — Post-mortem of subprime factory; 12 lessons of credit crises; 7-party incentive chain
"Nobody Knows" (2008) — Epistemic humility at peak GFC panic; "I don't know" school codified
"The Long View" (2009) — Multi-decade era of increasing willingness and expansiveness
"Touchstones" (2009) — Crisis lessons as reusable mental anchors
"It's Greek to Me" (2010) — Eurozone sovereign credit crisis; shared credit without shared fiscal discipline
"Open and Shut" (2011) — Credit cycle and low-rate forced risk-taking
"What's Behind the Downturn?" (2011) — U.S. downgrade, Europe, and confluence-driven volatility
"What Are the Chances?" (2012) — Active management as search for mistakes; hedging and conviction calibration
"On Uncertain Ground" (2012) — Slow-growth macro, low-return world, and "move forward, but with caution"
"Ditto" (2013) — Risk-attitude cycle and handcuff volunteers under zero-rate policy
"The Role of Confidence" (2013) — Confidence as self-fulfilling economic and market force
"The Lessons of Oil" (2014) — Oil-price collapse as failure of imagination and second-order consequence case study
"Liquidity" (2015) — Liquidity as transient; funds/ETFs cannot exceed the liquidity of underlying assets
"Risk Revisited Again" (2015) — Risk as permanent loss and future probability distribution
"It's Not Easy" (2015) — Second-level thinking, outcome bias, and why obvious opportunities disappear
"Economic Reality" / "Political Reality" (2016) — Scarcity, trade-offs, Brexit, populism, and political incentives that deny economic constraints
"Implications of the Election" (2016) — Trump-era voter anger as a symptom of globalization, automation, cultural change, and institutional failure
"Go Figure!" (2016) — Event prediction and market-reaction prediction both failed around the Trump election
"Expert Opinion" (2017) — Brexit/Trump/Italy polling failures; experts need scored records, not just confident narratives
"Lines in the Sand" (2017) — Subscription lines; IRR optics versus true LP wealth creation
"There They Go Again . . . Again" (2017) — Late-cycle risk appetite, low VIX, FAANG extrapolation, and boom ingredients
"Yet Again?" (2017) — Clarifies caution as calibration, not a market-timing call; revisits Bitcoin as currency/speculation
"Latest Thinking" (2018) — Market pros/cons, Trump tax law, SALT second-order effects, and late-cycle caution
"Investing Without People" (2018) — Passive investing, ETFs, quants, AI, and price discovery with fewer human investors
"The Seven Worst Words in the World" (2018) — "Too much money chasing too few deals"; direct lending and late-cycle debt-quality warnings
"Political Reality Meets Economic Reality" (2019) — Tariffs, capitalism, populist taxation, and second-order policy consequences
"Growing the Pie" (2019) — Inequality, populism, redistribution, and capitalism debate
"This Time It's Different" (2019) — Nine optimistic late-cycle propositions and the risk of unlearning history
"On the Other Hand" (2019) — Fed rate cuts as ambiguous signals; recessions may be postponable rather than avoidable
"Mysterious" (2019) — Negative interest rates and the uncertainty of financial systems built on unfamiliar assumptions
"You Bet!" (2020) — Games, probabilities, Annie Duke, and decision quality under uncertainty
"Nobody Knows II" (2020) — COVID uncertainty; facts/inferences/guesses; staged buying based on price versus value
"Latest Update" (2020) — COVID curve-flattening, closures, job losses, and crisis-policy response begins
"Which Way Now?" / "Calibrating" (2020) - COVID scenario discipline; offense/defense calibration; buying without waiting for the bottom
"Knowledge of the Future" / "Uncertainty" (2020) - Epistemic humility, path dependence, expert selection, and tail risk
"The Anatomy of a Rally" (2020) - Fed/Treasury liquidity, FOMO, and the rapid market rebound
"Time for Thinking" / "Coming into Focus" (2020) - Economy as induced coma, non-cyclical shock, low rates, and rescue side effects
"Something of Value" (2021) - Value vs growth as false dichotomy
"Thinking About Macro" (2021) - Inflation debate and prepare-don't-predict response
"The Winds of Change" (2021) - Technology, work, political structure, generational inequity, China, and Fed activism
"Selling Out" (2022) - Relative selection and long-term compounding versus psychological selling
"The Pendulum in International Affairs" (2022) - Ukraine, energy security, semiconductors, and supply-chain resilience
"Bull Market Rhymes" (2022) - 2020-21 bull psychology and the correction in tech, SPACs, meme stocks, and crypto
"Conversation at Panmure House" (2022) - Market mind, economics vs physics, active/passive reflexivity
"I Beg to Differ" (2022) - Active difference, second-level thinking, contrarianism, and the risk of being wrong
"The Illusion of Knowledge" (2022) - Macro forecasting as unreliable process; stationarity and single-scenario forecasting fail at inflection points
"What Really Matters?" (2022) - Short-term macro, trading mentality, volatility, and one-year performance are usually distractions
"Sea Change" (2022) - End of the 40-year declining-rate tailwind; higher rates make credit newly competitive
"Lessons from Silicon Valley Bank" (2023) - Rising rates expose duration mismatch and easy-money errors
"Fewer Losers, or More Winners?" (2023) - Credit as a negative art: avoid defaults rather than chase uncapped upside
"Easy Money" / "The Impact of Debt" (2024) - Cheap money distorts behavior; leverage creates path-dependent ruin risk
"The Folly of Certainty" (2024) - Political, macro, and market certainty are usually false precision
"On Bubble Watch" / "Is It a Bubble?" (2025) - AI, bubble behavior, debt-financed infrastructure, inflection bubbles, and investor capital destruction
"Cockroaches in the Coal Mine" (2025) - Private credit/fraud scares; good times create bezzle conditions

The Realist's Creed (Marks' Own Summary)

Contrarianism — buy unpopular, sell popular
Skepticism — doubt consensus, especially optimistic
Modest expectations — calibrate to the environment
Humility — know what you don't know
Defensive investing — survival first
Patience — wait for the fat pitch

Sources in This Wiki

the-complete-collection-howard-marks — Complete Oaktree memo collection (1990–2025); primary source for all frameworks above
second-order-thinking — Farnam Street explainer drawing on Marks' second-level thinking