Notes — Howard Marks on Avoiding Disaster, Risk Posture, and the AI Bubble

Source-grounded literature notes. Own words. Citations by section. Howard Marks, interviewed by William Green for Richer, Wiser, Happier (RWH063). Occasion: 35th anniversary of Marks’s investment memos; a free digital compilation of 45 best memos had just been published.

Four questions [Adler frame]

Q1 — What is it about? A comprehensive statement of Howard Marks‘s defensive investment philosophy built around one founding insight — that avoiding losers matters more than accumulating winners — plus a set of practical doctrines that follow from it: risk-posture calibration, the rarity of warranted macro calls, the virtues of patience and counterintuition, and the dangers of bureaucratic investment management. The interview also includes a situational assessment: the AI investment environment compared to the 1998–2000 internet bubble, and Marks’s analytical reasons for avoiding gold and Bitcoin.

Q2 — How is it argued? Memoir and principle are interwoven. Each doctrine is introduced through a biographical moment: the 1990 Minneapolis dinner generates the founding motto; the 2008 Lehman deployment shows why committees fail; the 1969 Nifty 50 debacle shows that admirable quality is no safeguard against an excessive price. Marks uses counter-factuals consistently: what if I had made 5,000 macro calls rather than five? What if Buffett’s coin-toss predictor applied a Cray supercomputer to a fair coin? The argument is not statistical proof but pattern recognition from a 56-year sample, presented with explicit uncertainty about its generalisability.

Q3 — Is it true? The asymmetry-of-loss argument is mathematically sound and consistent with modern portfolio theory. The Kelly criterion and Nassim Taleb’s writings independently confirm that left-tail management compounds differently than right-tail maximisation. The market-efficiency argument — that edge requires operating where information is unequally distributed — is standard EMH doctrine applied to market selection rather than stock selection; the insight is conventional but correctly stated. The five macro calls are fewer than they might seem over 50 years, and Marks himself notes they were 80/20 propositions at best. The AI-as-internet-bubble comparison is well-reasoned but structurally uncertain: the claim depends on whether AI monetisation will prove as clear as e-commerce did. [?] The gold returns comparison (7.7% vs 12.7% annually from end-2010) is time-period-dependent and should not be taken as a structural conclusion about gold’s long-run performance.

Q4 — What of it? The actionable takeaway for any investor is not defensive timidity but conscious posture: knowing whether you are playing to accumulate winners or to exclude losers, and making that choice explicitly rather than by default. The second takeaway is the rarity standard: macro calls are warranted perhaps five times in a fifty-year career. Every analyst instinct to call a top or bottom should be tested against that baseline. The third takeaway is structural: Marks’s willingness to pre-raise capital so that deployment during a crisis requires no fresh client conviction is a lesson in institutional design, not only investment philosophy.

Glossary

Fewer losers or more winners: A Marks memo title and a practical choice every investor must make. Superior returns can come from having more of the things that go up or fewer of the things that go down; most investors do not have the equipment to do both. The aggressive player maximises winners; the defensive player minimises losers; the choice determines your strategy, your hiring, and your risk architecture.

Risk posture: Marks’s term for an investor’s calibrated position on a 0-to-100 scale from no risk to maximum possible risk. Each investor has a natural baseline determined by age, wealth, income, dependants, aspiration, and intestinal fortitude. The posture question is then whether current market conditions justify being above or below that baseline. Introduced in the Calibrating memo.

Taking the temperature: Marks’s phrase for assessing investor behaviour rather than predicting market outcomes. When exuberance is rampant, prices have already moved to dangerous levels; when depression is universal, prices are already cheap. The method produces observations, not forecasts. Marks used it explicitly five times in fifty years.

Idiosyncratic insight: David Swenson’s phrase for the kind of investment position that is by definition unpopular — because if it were popular, the opportunity would already be priced away. Idiosyncratic insight is incompatible with committee approval, because a committee whose members mostly hold the consensus view cannot rationally endorse a position that contradicts the consensus.

Intrinsic value: The discounted present value of future cash flows. For Marks, any asset whose value cannot be derived from cash flows — gold, Bitcoin, paintings — cannot be analysed in the standard value-investing framework; it can only be bought on belief, historical pattern, or speculation. Not a criticism of such assets, but a statement about the limits of his own analytical toolkit.

Less efficient market: A market where information is not equally distributed and where hard work and skill can therefore produce superior returns. Marks’s entry into high-yield bonds in 1978 illustrates the point: public pension funds would not touch them for reputational and political reasons, so prices contained bargains available to those willing to look. The coin-toss metaphor: no amount of computation can beat a fair coin, because there is no information advantage to exploit.

Intestinal fortitude: Marks’s preferred phrase for genuine emotional tolerance of volatility — not claimed tolerance but real staying power when prices are falling. It is one of the inputs to risk-posture calibration and the hardest to self-assess accurately because investors tend to overestimate it in bull markets.

Lottery ticket mentality: Buying a laggard stock because it is cheap relative to the market leaders in a bubble, on the theory that it offers a cheap option on the same transformational trend. Marks considers this a mistake: a low probability of success should be accepted as a likely failure, not reframed as an attractively asymmetric bet.

Section notes

The 1990 dinner: ‘if you can avoid the losers…’

David Van Benschoten managed General Mills’s pension fund for 14 years. The equity portfolio was never above the 27th percentile or below the 47th in any single year — solidly second quartile throughout. The 14-year cumulative result: fourth percentile overall. The mathematics are not paradoxical but are counterintuitive: most investors shoot for star performance and occasionally suffer a catastrophic loss; a large loss takes years to recover from, and the compounding time lost cannot be retrieved. Steady second-quartile performance never suffers that penalty, so it compounds past competitors who swing for the fences.

Marks’s first memo (1990): ‘in equities, if you can avoid the losers and losing years, the winners will take care of themselves.’ This became Oaktree’s motto. Graham and Dodd’s 1940 Security Analysis made the same point for fixed income: bond investing is a negative art — if 90 out of 100 bonds will pay, the only thing that matters is avoiding the 10 that default. You improve performance by what you exclude, not what you select.

The connection to defensive tennis: Charlie Ellis’s 1975 paper ‘The Loser’s Game’ observes that professional tennis players must hit winners — their opponents are skilled enough to punish any soft return. Amateur players, who lack that control, should simply get the ball over the net. Investing resembles amateur tennis, not professional: there is too much randomness and uncertainty to swing for the fences without risking being ‘carried out.‘

Fewer losers or more winners: the conscious choice

From the Fewer Losers or More Winners memo. The skillful aggressive investor accumulates winners. The skillful defensive investor excludes losers. Very few can do both because the skills and the biases run in opposite directions. The key insight is that the choice must be conscious: a portfolio that never explicitly answers the question ‘am I playing offence or defence?’ cannot have a winning strategy, because you cannot execute consistently on a posture you have never identified.

The extension to portfolio construction: Are you maximising growth of capital or preservation of capital? Both goals cannot be maximised simultaneously. Every allocation decision implicitly answers this question; the question is whether the investor answers it deliberately.

Risk posture calibration

From the Calibrating memo (approximately 2018). The inputs to a personal baseline posture: age and career stage; wealth relative to income and needs; number of dependants; level of aspiration; proximity to retirement; intestinal fortitude. Once the baseline is established, the active question is: given current conditions, should I be above or below my baseline?

Marks does not advocate active trading. He estimates he has moved materially around his baseline five times in fifty years. The point is not to respond to every market fluctuation but to recognise the rare compulsive cases — when the evidence for defensiveness or aggressiveness is overwhelming — and to act on them deliberately rather than by drift.

Nick Sleep’s advice to William Green (‘don’t fiddle’) is instructive but slightly too idealistic in Marks’s view: if the market moves to extremes, a degree of recalibration is warranted. The Fidelity study (possibly apocryphal) — that the best-performing accounts belonged to investors who were dead — makes the correct point about overtrading, but should not be read as an argument against all adjustment.

Taking the temperature: five calls in fifty years

The five calls (all ‘observations of investor behaviour, not predictions of outcomes’):

January 2000 — the first day of the year, a memo on internet.com excess.
2004–2007 — subprime and mortgage-backed securities.
September 2008 — deploying $7 billion in distressed debt post-Lehman at $450M per week for 15 weeks.
2012 — [referenced without detail in transcript].
2020 — [referenced without detail in transcript].

The method: Buffett’s principle (‘the less prudence with which others conduct their affairs, the greater the prudence with which we must conduct our own’) applied to observable behaviour. When exuberance is rampant and risk is invisible, prices have already moved to dangerous levels. When depression is universal, prices are already cheap enough to reward aggression.

Marks emphasises trepidation: he was never certain on any of these calls. He estimates each was at best 80/20 in his favour. If he had tried to make 5,000 calls over 56 years (one every four days), his record would have been 50/50. The rarity of the calls is the discipline.

Idiosyncratic insight versus committees

Swenson: ‘Active management strategies demand uninstitutional behaviour from institutions, creating a paradox that few can unravel.’

Marks loathes committees because idiosyncratic insight — by definition — is what most people do not see. If most market participants are doing A, a committee majority will endorse A. The very insight that most makes money is the one most likely to fail a majority vote. His 2005 memo Dare to Be Great is a rant against bureaucracy.

The 2008 deployment proves the point: Marks and Bruce Karsh spent $7 billion in one quarter. A committee could not have approved that. The structural solution was pre-raising committed capital so that deployment during the crisis required no fresh client convincing. The institutional design (committed fund + two principals with authority) made idiosyncratic action possible.

Marks on marriage and institutions: ‘marriage is a wonderful institution for people who like living in institutions — and I don’t.’ He regards institutional constraints as anti-competitive.

Market efficiency and the less efficient market

The counterfactual: if you were given 15 PhDs and a Cray supercomputer and asked to predict coin tosses at the start of football games, you would fail. A fair coin cannot be beaten because there is no information to exploit. An efficient market (where all participants have the same information) is analogous. Marks’s lesson from high-yield bonds (1978): public pension funds would not touch them for reputational and political reasons. That unpalatable status was itself the opportunity. ‘Oh great — an asset class I can buy that nobody else will buy at any price. Maybe it’s full of bargains.’ The bargains were partly structural: when only risk-tolerant specialists will transact, prices reflect that exclusion.

The knowledge-advantage requirement: you cannot claim superiority from intelligence (everyone is smart) or from pedigree (everyone went to good schools). You need a specific edge — either superior analysis of the same data, access to data that others lack, or a willingness to operate where others will not.

The AI bubble comparison

Marks identifies the closest historical parallel as the 1998–2000 internet bubble, not the Nifty 50 or the subprime mortgage market. Both involve a genuinely transformational technology that ‘fired the imagination.’ The Nifty 50 was about established great companies — no technological crises. Subprime was a financial invention, not a technological one; no one thought subprime mortgages would change the business of housing.

The distinction between the internet and AI: in 1999–2000, the vision of how the internet would change the world was fairly clear, and it mostly came true. E-commerce became the dominant force that was imagined. Today, the vision of how AI will produce profits — not merely productivity — is less clear.

Buffett’s 2000 annual meeting point: ‘There’s no doubt that the internet will produce a great increase in productivity. It’s not clear that it’ll have a positive impact on profitability.’ Marks applies the same question to AI: if providers compete on price, or if their customers pass savings to consumers, AI-generated productivity may not accrue to investors.

The three mistakes to avoid in euphoric AI investment: (1) assuming today’s leaders will be tomorrow’s leaders; (2) assuming that laggards are therefore safer bets (‘lottery ticket mentality’ — a low-probability-of-success bet should be accepted as probably unsuccessful); (3) failing to distinguish binary bets (pure-play startups with no revenues) from incumbent technology companies that will get moderate AI benefit if it works, and survive if it does not.

Marks’s general principle: bubbles coalesce around new things because imagination has no prior data to discipline it. Trees can be imagined to grow through the sky when no one has seen one reach the sky before.

Gold and Bitcoin

Marks’s value-investor constraint: he cannot invest analytically in an asset without cash flows. Without cash flows, there is no intrinsic value to compare against price. Gold can be a store of value, a hedge, a speculation, or a religion — but it cannot be analysed as cheap or dear in the standard sense.

The returns comparison (end-2010 to early 2025): gold approximately 7.7% annually; S&P 500 approximately 12.7% annually. Not catastrophic, but not an obvious substitute for equities. The comparison is time-period-sensitive.

Oil at $147 in July 2007 and $35 in January 2009: ‘nothing changed about oil.’ Commodities without cash flows have prices that are what the market will bear, not what analysis will justify.

Charlie Munger

Munger was exceptionally well-read and developed what he called a ‘lattice work of mental models’ — a toolkit for pattern recognition across domains. The practical effect: when a new situation arose, Munger did not reason from scratch but recognised a pattern and applied the appropriate tool. This is the investing analogue of a doctor who can say ‘I’ve seen this before.’

Munger’s major contribution to Buffett: convincing him to abandon cigar-butt investing (buying okay companies at great prices, then extracting the residual value) in favour of buying great companies at okay prices. That revision is the intellectual foundation of Berkshire’s long-run record.

Munger’s other contributions: brutal directness, no patience for committees or bureaucracy, self-deprecating clarity. On his late-life Alibaba purchase: ‘I just bought it yesterday, so it’s bound to go down 50% immediately.’ (It did.)

Living a balanced life

Marks’s idol: Christopher Morley — ‘there is only one success: to be able to live your life in your own way.’ He has played tennis and backgammon, spent time with family, and specifically designed his role at Oaktree (setting investment philosophy, meeting clients, writing memos) around his temperament. He skips to work in the morning.

The counsel: once you have enough money — more than enough — giving up enjoyment to acquire more is not a rational trade. The trite saying (‘nobody on their deathbed said I wish I worked more’) is trite because it is true. The harder version: figure out at 22 what will make you happy at 70, knowing that you will change and that you may have an inaccurate vision of yourself.