Where the idea comes from

Howard Marks, co-founder of Oaktree Capital Management, made the term popular in his client memos and in his 2011 book The Most Important Thing. His point is simple. To earn better results than the market, you need a view that differs from the consensus and turns out to be right. A correct view that everyone shares is already in the price.

The idea is older. John Maynard Keynes compared markets to a newspaper beauty contest, where readers win by picking the faces others will pick. The task is not to find the best answer, but to anticipate the average opinion and judge when it is wrong.

First level vs second level

SituationFirst-level thoughtSecond-level questions
Strong companyGreat business, so the stock goes upWhat growth does the price assume? Can it deliver more?
Earnings reportResults will beat, so the stock risesHow big a beat is priced in? What move do options imply?
Cheap stockLow P/E, so it is a bargainWhy is it cheap? What would make the market change its mind?
Announced mergerThe deal will close, so the spread is free moneyWhat odds does the spread imply? What could break it?
Bad newsResults missed, so the stock fallsWas the miss expected? Is the selling overdone?

The second-level column always starts by measuring the market's expectation. That is the step most investors skip.

How to measure what the market expects

Expectations are not a mood. They can usually be read from prices and published numbers.

Share price. A reverse DCF works backwards from the current price to the growth and margins it implies. If the price needs 20% annual growth for a decade, a forecast of 12% is a disagreement, even if 12% sounds strong.

Consensus estimates. Analyst estimates show what results are expected for the next quarter and year. Watch the direction too: estimate revisions move the bar before the report.

Option prices. The straddle price before an event implies the size of the expected move. Estimating an earnings move from options shows the arithmetic.

Spreads. In a merger, the gap between the share price and the deal price implies the odds of closing. A tight spread means the market sees little risk.

Valuation multiples. A multiple far above or below peers is a compressed statement of expected growth and risk.

Forming a variant view

Once the expectation is measured, the question becomes whether you see something different. That is variant perception: a view that differs from consensus, with a reason.

A usable variant view has three parts.

  1. The expectation. What the price or consensus implies, stated as a number.
  2. Your difference. What you expect instead, and the evidence behind it.
  3. The test. A dated event, such as an earnings report, a vote or a court ruling, that will show which view was right.

Without the third part, a disagreement can last for years. This is why event-driven investors care about catalysts. An event forces the market to update, and it limits how long you have to wait. What a stock catalyst is covers the types.

A worked example

A first-level reading says Company B has strong momentum and will beat. A second-level reading notes that the rise in estimates and price means a beat is already expected. A result of $1.05 might be in line with what holders really want. The interesting question is whether results could be far above or below what the rising expectations assume, and whether anything in the company's own disclosures supports either view.

If the honest answer is "I expect a modest beat, like everyone else," the second-level conclusion is that you have no edge. That is a useful answer.

Second-level thinking in special situations

Special situations are where the habit pays most, because the market's expectation is often written in a single number you can read directly.

SituationWhat the price impliesThe second-level question
Announced cash mergerThe spread implies the odds of closing and the time it takesIs the regulatory or financing risk larger or smaller than the spread assumes?
Stock below net cashThe market expects the cash to be spent badly or burnedIs there a catalyst, such as a buyback, a sale or a liquidation, that returns it first?
Spin-offForced selling by index funds and parent holders pushes the price downIs the selling over, and does the new company's value differ from what the selling implied?
Binary drug readoutThe option straddle implies the size of the moveAre the odds of success better or worse than the move suggests?
Bankruptcy equityOld shares trade on hope of a recoveryDoes the plan leave anything for equity at all?

In each case the work starts with what the price says and ends with a specific reason to disagree. Net cash per share and probability-weighted valuation for binary events show how to turn the reasoning into numbers.

A second advantage of special situations is that the event is dated. A merger closes or breaks, a vote passes or fails, a court rules. The market's expectation and your own are tested on a known day, which keeps the second-level view honest. When the event passes and the outcome matches the consensus, the lesson is that the price was right, and that is worth recording too.

Where this goes wrong

Second-level thinking is a discipline, not a formula, and it fails in recognizable ways.

  • Contrarian by reflex. Disagreeing with the crowd is not the same as being right. Most of the time the consensus is reasonable.
  • Guessing the expectation. Saying "the market is too pessimistic" without measuring what the price implies is a first-level thought dressed up.
  • Infinite regress. "Everyone knows that everyone knows" can run forever. Stop at the measurable expectation and the evidence that would change it.
  • No catalyst. A correct variant view without an event can stay unrewarded long enough to cost more than it earns.
  • Overconfidence. A different view needs stronger evidence than the consensus, because the consensus already aggregates many informed people. Confirmation bias makes it easy to collect only the evidence that agrees.

Building it into a routine

The habit sticks when it is part of the write-up. Before any position, write down the expectation in numbers, your difference, the evidence and the date it will be tested. A dated investment thesis does this in one page. Afterwards, compare the outcome with what you wrote.

Measuring expectations needs data: estimates, option prices and filings. The guide to stock research tools for individual investors compares the platforms that provide them. Dombey's published research states the market's expectation and the variant view before each catalyst; see the research archive.

The bottom line

Second-level thinking asks what the price already assumes before asking whether a company is good. The market's expectation can usually be measured from the share price, estimates, option prices and spreads. A view has an edge only when it differs from that measured expectation, rests on evidence and faces a dated test. When your view matches the consensus, the disciplined answer is that the price is probably right.