Michael Mauboussin

A share price is not an answer. It is the starting constraint in an argument about what a business must deliver.

Begin With What the Price Demands

Michael Mauboussin is Head of Consilient Research at Counterpoint Global, Morgan Stanley Investment Management, and an adjunct professor of finance at Columbia Business School. Columbia's current faculty profile says he has taught there since 1993. Those roles make him a researcher and educator working near investment practice. They do not identify him as the sole decision-maker for a public portfolio with an attributable return record.

His distinctive method is expectations investing, developed with Alfred Rappaport. Conventional valuation often starts with an analyst's forecast, discounts the forecasted cash, and compares the result with price. Their method reverses the order. Start with the observable price, use a cash-flow model to infer the operating path that would justify it, and then judge whether the market will have to revise that path.

The inversion prevents a common category error. A company can have excellent products, rapid growth, and capable management while its shares offer a poor return because the price already requires even better results. A troubled company can be an attractive investment if the price assumes a deterioration more severe than the likely outcome. Quality and attractiveness are related only through expectations and price.

The Model Exposes an Operating Burden

The publisher's description of Expectations Investing and the authors' chapter guide set out the sequence. An analyst estimates price-implied expectations, examines historical performance and competitive strategy, isolates the value trigger that matters most, runs scenarios, and converts possible revisions into expected value and a buy, sell, or hold decision.

Sales, costs, and investment are the operating triggers. Volume, price and product mix, operating leverage, scale economies, cost efficiency, and investment efficiency connect those triggers to cash flow. The model also needs an opportunity cost of capital and a view of continuing value. Instead of hiding those judgments inside a multiple, the process asks how much growth is required, what margin must be reached, how much capital that growth consumes, and how long returns can remain above the cost of capital.

This is not a discovery of one true market forecast. Different combinations of growth, margin, reinvestment, duration, and discount rate can reconcile with the same price. Price also reflects risk preferences, taxes, fund flows, mandates, and liquidity. The output is better described as a set of operating hurdles chosen by the analyst. Its value lies in making those choices visible and testable.

Domino's Shows the Full Sequence

The revised book uses Domino's Pizza as its main implementation case. According to the official chapter summaries, chapter 5 estimates price-implied expectations, chapter 6 combines the company's history with the expectations infrastructure and competitive analysis to isolate a key trigger, and chapter 7 translates scenarios into an investment decision. The case replaced Gateway from the original edition.

That progression is more important than a retrospective claim that Domino's became a successful stock. A useful case begins before the outcome: establish what the quotation requires, identify the business variable with the greatest valuation sensitivity, specify alternative operating states, and decide whether the probability-weighted value leaves an adequate margin over or under price.

The public chapter guide does not disclose every input, valuation date, transaction, position size, or realized return. It therefore supports Domino's as a documented teaching application, not as Mauboussin's personal investment. The distinction keeps an analytical example from becoming an invented performance anecdote.

Competition Determines How Long Returns Last

Price-implied growth is incomplete without the cost of producing it. Mauboussin and Dan Callahan's Measuring the Moat defines sustainable value creation through three connected quantities: the spread between return on invested capital and the cost of capital, the amount a company can reinvest at that spread, and the period during which the opportunity persists.

A high historical return attracts competitors. Markets mature, customer preferences change, successful practices spread, and regulation can alter who is allowed to act. The 2026 competitive-advantage-period report argues that simple earnings or cash-flow multiples obscure duration, while discounted-cash-flow models often bury it in an ungrounded terminal value. The corrective is not to declare that a company has a moat. It is to explain the choices, customer economics, industry structure, and capital requirements that could keep excess returns from fading.

Accounting adds friction. Reported invested capital may omit internally created intangible assets, while acquisitions place similar assets on the balance sheet. Leases, inflation, write-downs, and changing business mix can distort comparisons. Return on invested capital is a recorded result that prompts causal investigation; it is not a self-executing measure of competitive advantage.

Base Rates Discipline the Inside View

A company-specific thesis is an inside view. Base rates supply an outside view: what happened to companies that began in a comparable position? In Bayes and Base Rates, Mauboussin and Callahan examine ambitious artificial-intelligence forecasts using nearly 18,900 observations for US public companies with $2 billion to $5 billion of starting sales from 1950 through 2024.

The historical distribution does not prohibit an exceptional outcome. It establishes how exceptional the forecast is before company-specific evidence is considered. New contracts, economics, capacity, or competitive developments should update that prior. The report also warns that base rates can change as the world changes.

Reference-class choice is therefore part of the judgment. A semiconductor producer, software platform, and data-center builder can all benefit from artificial intelligence while facing different reinvestment, rivalry, and scale constraints. A broad class may hide the relevant economics; a narrow class may contain too little evidence or be selected to flatter the forecast. The analyst should preserve the class definition and test alternatives.

Process Does Not Substitute for Performance

Mauboussin's work on skill and luck explains why one favorable outcome cannot validate a process. A noisy result contains the decision, unpredictable events, competitor actions, and error. Conversely, a sound probabilistic decision can lose. Repeated, dated forecasts and feedback are more informative than a celebrated winner, though a process that persistently fails still needs repair.

No complete audited return series attributable to Mauboussin appears in the cited official sources. Counterpoint Global is an investment organization, but its research papers do not show whether a portfolio manager used a particular analysis, what position was taken, how it was sized, when it traded, or what it earned. The research is also coauthored: expectations investing belongs to Rappaport and Mauboussin, while the Morgan Stanley reports cited here belong to Mauboussin and Callahan.

Morgan Stanley's disclosures describe the reports as informational and educational, not recommendations, and say the views may not represent all investment personnel or products. That institutional boundary is material. Research can clarify a decision; a portfolio also requires mandate authority, execution, liquidity, risk limits, a benchmark, and clients able to tolerate divergence while a thesis is tested.

An Assumption Ledger, Not a Valuation Oracle

The method transfers as a disciplined sequence. Translate price into an operating burden. Identify the variable carrying the most value. Compare the inside story with business records, competition, and a named reference class. Build scenarios rather than a single point forecast. Record what would revise each assumption. Separate the quality of the decision from the luck in one result.

Its limit is equally useful. A reverse model cannot prove what every market participant believes, a historical distribution cannot settle a structural change, and a precise expected value remains sensitive to probabilities chosen by the analyst. Mauboussin's contribution is not certainty. It is a way to make an investment disagreement specific enough to test: what must happen, what could prevent it, and what the current price already assumes.

Inside CompanyGraph

The deep-value configuration is observable: companies priced well below book value while current assets exceed current liabilities by a wide margin and the equity ratio sits high for the industry.