Second-Level Thinking

Concept Analysis

Definition & Origins

Second-level thinking is the foundational cognitive framework in Marks' entire body of work — the answer to the question what does it take to generate superior investment results? It appears in the opening pages of The Most Important Thing and is revisited throughout the 35-year memo corpus.

The terminology is Marks' own coinage, though the underlying idea draws on Keynes' beauty contest analogy, Graham's Mr. Market, and basic game theory. Marks observed that most investors think at what he calls the "first level" — forming a view about the world and expressing it in a portfolio. The problem is that all other investors are also forming first-level views, and those views are already embedded in market prices. If your view of the world matches the consensus, even if correct, it is already in the price — and you earn approximately the consensus return.

Superior returns require disagreeing with the consensus and being right. That is the high bar of second-level thinking: not just forming an independent view, but forming an independent view that turns out to be more accurate than what the market has priced.

The coinage itself has a precise birthday. In 2009, Warren Buffett wrote to Marks promising a jacket blurb if he would finally write his book. When Columbia Business School Publishing asked for a sample chapter, Marks sat down — without, by his own account, previously having given the topic any organized thought — and knocked out a chapter on something he labeled "second-level thinking." It became the first chapter of The Most Important Thing, and the memo "It's Not Easy" (2015) later reprinted that chapter for the memo readership, conceding that in two decades of memos he had never covered the subject explicitly.

Core Ideas

First-level vs. second-level — the fundamental distinction. First-level thinking: The company's earnings will grow. The stock will rise. Buy. Second-level thinking: Everyone expects earnings to grow. That expectation is priced in at a 30x P/E. What if growth is merely 'good' rather than 'excellent'? What is the downside if earnings disappoint the consensus by only 10%? Is the asymmetry of outcomes favorable at this price? The first-level thinker asks "what will happen?" The second-level thinker asks what is the consensus expecting, how does my estimate differ, and how is the probability-weighted outcome distribution affected by the price I'm paying?

Consensus thinking produces consensus returns. This is the logical core of the argument. Markets aggregate the views of all participants through the pricing mechanism. When one investor's view is identical to the average of all investors, their portfolio will produce the average return. To generate returns above average, you must depart from consensus — which means sometimes being early, sometimes appearing wrong, always enduring the discomfort of holding a position that isn't immediately validated by price movement.

Being contrarian is not enough — you must also be right. Marks is careful to distinguish second-level thinking from reflexive contrarianism. Doing the opposite of the crowd is as mindless as following the crowd, because sometimes the crowd is right. The second-level thinker asks "where is the consensus wrong?" — which sometimes leads to a contrarian conclusion and sometimes leads to agreement with consensus at a different price or on different terms.

The difficulty sets the magnitude of the opportunity. Because second-level thinking is difficult — requiring independent analysis, tolerance for feeling wrong, deep understanding of what the consensus view actually is — few investors consistently practice it. This difficulty is precisely what makes superior returns possible for those who do. If second-level thinking were easy, prices would already reflect it.

The question that matters most is not "what is this worth?" but what does the market think it's worth, and is that estimate wrong? Intrinsic value analysis is necessary but not sufficient. You must also understand how your estimate compares to the consensus estimate embedded in the price. A correct intrinsic value estimate that matches the consensus produces no alpha. A correct intrinsic value estimate that differs significantly from consensus is the source of investment edge.

The second-level thinker asks a battery of questions the first-level thinker never raises. What is the range of likely future outcomes? Which outcome do I think will occur, and with what probability? What does the consensus think, and how does my expectation differ? How does the current price comport with the consensus view of the future — and with mine? What will happen to the price if the consensus turns out to be right, and what if I am right? The first-level thinker, Marks notes, is ignorant of the very existence of a second level at which to think; the second-level thinker understands the market as a marketplace in which prices reflect and depend on the expectations of the other participants.

Practical Application

In credit markets: The second-level insight that gave Oaktree its foundational edge: in the late 1980s, the consensus view was that high yield bonds were inherently speculative and dangerous. Marks and his team concluded that the consensus systematically overestimated default rates and underestimated recovery rates, producing prices that implied higher risk than actually existed. This non-consensus view — if correct — would generate superior returns. It was correct.

At market cycle extremes: Second-level thinking is most powerful and most difficult at market extremes. In 2007-2008, the consensus through most of the credit bubble was that structured credit products (CDOs, SIVs) were safe. The second-level question: What assumptions must be true for these instruments to be safe? Are those assumptions actually true? What happens to correlations between mortgage defaults in a nationwide housing decline? The investors who asked these questions — and came to different conclusions than the consensus — generated extraordinary returns when the crisis arrived.

In distressed debt: When a company files for bankruptcy, the instinctive first-level response is "this is broken, avoid it." The second-level question: What does the market price imply about recovery rates and timeframes? Is that implied recovery realistic given the asset coverage, business economics, and restructuring probabilities? If recovery is higher than the market implies, what return does the current price offer?

In equity markets: Marks' first job supplied his formative case study. When he showed up for work in First National City Bank's investment research department in 1968, the bank was investing heavily in the Nifty Fifty — the stocks of America's best, fastest-growing companies — on the official dictum that the price you paid didn't matter much for companies where nothing could go wrong. At 80 to 90 times earnings, the first-level view ("great company, buy the stock") was universal, and being universal, it was fully in the price. When the market ran into trouble in the early 1970s, many of these stocks lost the vast majority of their value, and once-dominant names like Kodak, Polaroid and Xerox eventually went bankrupt or required turnarounds. The lesson that seeded second-level thinking: the merits were real but too obvious; the ideas became too popular; and the prices became dangerously high.

Common Misconceptions

Misconception 1: Being different is sufficient. Holding an unconventional view is not valuable in itself. The value is in holding an unconventional view that turns out to be correct. The non-consensus view must be right often enough, by a large enough margin, to compensate for the periods when it is wrong.

Misconception 2: Second-level thinking is more complex thinking. Sometimes second-level thinking leads to simpler conclusions than first-level. If everyone is building a complex model to value a security, the second-level question might be: Why is everyone so confident in complex models for an inherently uncertain business? Is the complexity providing false precision? Simplicity can be the contrarian insight.

Misconception 3: You need a second-level view on everything. Most markets, most of the time, are reasonably efficient — the consensus view is approximately correct, and no superior information is available. The opportunity to add value through second-level thinking exists in specific niches: complex, stigmatized, or structurally constrained markets where the consensus is more likely to be wrong for systematic reasons.

Misconception 4: A forecast must be correct to be profitable. Marks inverts this too. The easiest correct forecast — extrapolating a trend and seeing it continue — is broadly shared, so it is discounted in the price in advance and earns nothing unusual when it comes true. Conversely, a forecast can be wrong in its particulars and still be highly profitable if it is merely less wrong than the consensus: when a trend everyone else extrapolated fails to continue, even an imprecise prediction of the deviation pays off. What matters is not the accuracy of your forecast in the abstract, but its accuracy relative to the one embedded in the price.

Howard Marks' Own Words

Thought Evolution

Related Concepts

Key Memos

Mentioned In

Source: Chian.io — Howard Marks Knowledge Base