← Back to list
Colossus (Invest Like the Best / Business Breakdowns)Podcast20 Dec 2022Source: joincolossus.comHost: Patrick O'Shaughnessy

Michael Mauboussin - Sharpening Investor & Executive Toolkits - [Invest Like the Best, EP.308]

In plain words

This interview covers Michael Mauboussin's view that traditional metrics like profit margins can mislead because they ignore intangible investments (e.g., R&D, brand). He argues that high profits at firms like Google and Microsoft come from heavy intangible spending, not monopoly power. He notes that surging real interest rates have hurt asset prices but boosted expected returns, creating capital-allocation opportunities. Key holdings: Snowflake (its ROIC jumps from -416% to 3% after adjusting for intangibles, fitting an early-stage firm), Amazon (overcapacity post-COVID, management adjusting), and OpenAI (AI performance surprises, but value capture is unclear).

AI SummaryAI-generated · may contain errors · verify against the original

Michael Mauboussin (Head of Research at Counterpoint Global) delved into three core topics during the program: market share, return on invested capital (ROIC), and capital allocation. He pointed out that early-stage breakthrough companies in low-concentration markets present attractive investment op

~11 min full read · 7 sections
Deep Analysis

At a Glance

Michael Mauboussin (Head of Research at Counterpoint Global) and Patrick O'Shaughnessy engaged in an in-depth discussion on three core topics: market concentration, return on invested capital (ROIC), and capital allocation. The central thesis is that these seemingly basic metrics, when adjusted for intangible assets, reveal a completely different picture—the high profit margins of "superstar firms" stem not from market monopoly, but from their heavy investment in intangible assets.


Market Concentration ≠ High Profitability: The Key Lies in the "Value Stick"

Mauboussin argues that there is no stable causal relationship between market concentration and industry profitability. Citing academic literature, he notes that highly concentrated industries are not necessarily more profitable than those with low concentration; rather, there is a weaker conditional link between a company's market share and its profitability.

  • Two Sources of Market Concentration: One is M&A-driven consolidation (e.g., the U.S. defense industry over the past 25 years, or the airline industry over the last 20 years); the other is "winner-takes-all" dynamics (e.g., Google in search, Microsoft in word processing). However, early stages often involve significant shakeouts—in the late 1990s, the search market had a very low HHI, with Ask Jeeves, Excite, and Google all competing, and Google's eventual victory was not foreseeable at the time.
  • Investor Tools: ① Market share stability—Bruce Greenwald's method: take the average absolute change in market share over two five-year periods; a high value indicates intense competition and weak moats. ② Industry entry/exit data—early stages see many entries and few exits, while mature stages see more exits than entries, with the number of competitors and market shares stabilizing.
  • Minimum Efficient Scale (MES) and Total Addressable Market (TAM) determine industry structure: a large market can accommodate multiple scaled players (e.g., 5–6 global auto companies with revenues exceeding $100 billion); in a small market, a single company reaching MES can become dominant.
  • Value Stick Framework (Brandenburger & Stuart, 1990s; Oberholzer-Gee, Better, Simpler Strategy): From top to bottom, the sequence is "willingness to pay → price → cost → willingness to sell." Companies should focus on increasing willingness to pay (rather than directly raising prices) and reducing willingness to sell (so that employees/suppliers remain satisfied at lower prices). Mauboussin emphasizes: "Don't worry about raising prices; think about how to increase willingness to pay."
  • Case Study: Walmart's partnership with P&G—Walmart shared sales data to help P&G optimize production and reduce inventory, enabling P&G to accept lower supply prices, creating a win-win outcome.

After Intangible Asset Adjustment, the ROIC and Markup Landscape Is Fundamentally Transformed

Mauboussin argues that traditional ROIC calculations severely underestimate the true capital investment of technology companies. After capitalizing intangible assets, extreme values are pulled back into a reasonable range. Meanwhile, the sharp rise in markup since 1980 nearly disappears once intangible investment is accounted for.

  • Three components of ROIC: ① NOPAT (unlevered, cash-free cash earnings) ÷ invested capital (cumulative investment on the asset side); ② The distribution is roughly normal in the middle with fat tails on both ends (a significant number of both low-ROIC and high-ROIC companies); ③ Through a simplified DuPont decomposition (NOPAT margin × capital turnover), companies can be classified as "cost leaders" (low margin, high turnover) or "differentiators" (high margin, low turnover).
  • Intangible asset adjustment: 30% of non-R&D SG&A plus all R&D are treated as investments, with industry-specific asset lives assigned. In 1985, intangible investment was only 0.7 times capital expenditure; by 2021, it had risen to 1.6 times. After adjustment: Snowflake's ROIC shifts from -416% to 3% (consistent with an early-stage company); the overall distribution's tails revert toward the mean; companies migrate from the upper-left corner (low margin, high turnover) to the lower-right corner (high margin, low turnover).
  • The markup debate: The traditional calculation (sales ÷ cost of goods sold × 0.85) shows that markup was low and stable from 1955 to 1980, then rose in a hockey-stick pattern after 1980. However, after incorporating intangible investment, the upward trend is significantly moderated, and the "superstar firm" effect within industries disappears—high-markup companies are precisely those that invest the most in intangible assets. Mauboussin cautions: "So-called superstar firms are, in essence, the companies with the largest intangible investments."
  • Absolute ROIC vs. change: Buying a high-ROIC portfolio and a low-ROIC portfolio yields the same Sharpe ratio—the market has already priced it in. The key lies in identifying the persistence of ROIC (transitions from high to low or low to high). Only a few companies can sustain top-quartile ROIC over the long term; if identified early, the returns are significant.

Capital Allocation: Internal Cash Dominates, M&A Leads, Intangible Investment Rises to Second

Mauboussin finds that the largest source of capital for U.S. corporations is internal cash flow, the largest use is M&A, but intangible investment (R&D and certain expenses within SG&A) has surpassed capital expenditure to become the second-largest spending item. Common traits of superior capital allocators include "zero-based thinking" and a "willingness to act."

  • Total spending in 2021 was approximately $7.1 trillion, with sources: internal cash as the primary source, debt as a supplement, and net equity reduction (buybacks and dividends exceeding issuance). Buybacks totaled about $1.1 trillion, with stock-based compensation at approximately -$250 billion.
  • Spending ranking: M&A (largest) → Intangible investment (SG&A + R&D) → Capital expenditure (roughly on par with buybacks in recent years) → Dividends → Divestitures.
  • CFO behavioral biases (John Graham's 30-year survey): ① Using discount rates far above the cost of capital; ② Underestimating uncertainty about the future; ③ Viewing dividends as "sacred" (not to be cut) and buybacks as "residual" (done only with spare cash); ④ Growth rates in M&A and buybacks are highly volatile (pro-cyclical), while dividend and capital expenditure growth remain stable.
  • Characteristics of superior capital allocators: ① Zero-based resource allocation — starting from a blank slate each year and asking, "How much resource does this business need to maximize its potential?" Academic research shows 98% of companies operate below the optimal adjustment level, exhibiting significant inertia. ② Strategic rather than project-based thinking — a good project may belong to a bad strategy (e.g., an individual optimization project within an excessive number of distribution centers). ③ Willingness to act — e.g., Bill Stiritz of Ralston Purina (a case from The Outsiders), who continuously assessed the timing for buying or selling each asset.
  • Human factors: Mauboussin shares a classroom anecdote — a CFO knew they should sell a non-core business but found it difficult to raise the issue because the CEO was a neighbor and the coach of their child's baseball team. "These are real, human factors that influence capital allocation decisions."

Current Market: Surging Real Rates Create Capital Allocation Opportunities, AI Is the Biggest Source of Disruption

Mauboussin argues that over the past 12 months, real rates have risen from -100bp to +120bp (a shift of over 200bp), putting pressure on all asset classes while simultaneously creating mispriced capital allocation opportunities. Artificial intelligence is the most noteworthy disruptive innovation at present.

  • Changes in Expected Returns: According to Aswath Damodaran's data, as of January 1, 2022, the equity risk premium stood at 5.75%, with 10-year inflation expectations around 2.5% and real returns at approximately 3.25% (half the historical average). Currently, this figure has risen to over 8%. High-yield spreads have widened by 150bp, and BBB spreads by 70bp. "Although the adjustment in asset prices is painful, expected returns have improved significantly."
  • Post-COVID Aftermath: Companies such as Amazon invested heavily in capacity to meet surging demand, only to see demand reverse later, leaving management teams struggling to adjust—this in itself represents a capital allocation opportunity.
  • Disruptive Nature of AI: The performance improvement from GPT-2 to GPT-3 exceeded expectations. Two key questions arise: ① The impact on the labor market (history suggests new technologies do not necessarily lead to unemployment); ② Who ultimately creates and captures value—"this is not clear from the outset."

Mentioned Positions

Position Analyst View Key Data
Google Historical case (winner-takes-all in search market) Search market share rose from extremely low in the late 1990s to 85-90%
Microsoft Historical case (dominance in word processing market) Traditional ROIC 49%, adjusted for intangible assets down to 34%
Snowflake Early-stage company case Traditional ROIC -416%, adjusted for intangible assets 3%
Walmart Positive value stick case Data sharing with P&G reduced supplier willingness to sell
Procter & Gamble Positive value stick case Leveraged Walmart data to optimize production and reduce inventory
Amazon Capital allocation case Heavy investment in capacity during COVID, followed by demand reversal
OpenAI Current focus GPT-2 → GPT-3 performance improvement exceeded expectations

Judgments Worth Remembering

1. "There is no stable causal relationship between market concentration and profitability" (Mauboussin) — Highly concentrated industries are not necessarily more profitable; only company-level market share and profitability have a conditional correlation. Investors should focus on market share stability and entry/exit data, rather than simply looking at concentration.

2. "Don't worry about raising prices; think about how to increase willingness to pay" (Mauboussin, citing Oberholzer-Gee) — The core of the value stick framework: raise the maximum price consumers are willing to pay, rather than directly increasing prices. Network effects and complementary goods (e.g., charging stations for electric vehicles) are classic paths to boosting willingness to pay.

3. "So-called superstar companies are essentially those with the largest investments in intangible assets" (Mauboussin) — The surge in markups after 1980 almost disappears once intangible investments are accounted for. High-markup firms are precisely those that spend the most on SG&A and R&D, not monopolistic price-setters.

4. "Snowflake's ROIC went from -416% to 3% — fully consistent with early-stage company characteristics" (Mauboussin) — Capitalizing intangible assets brings extreme values back to a reasonable range, revealing the true economic picture. Traditional accounting severely underestimates the capital investment of technology companies.

5. "98% of companies operate below the optimal resource adjustment level" (Mauboussin, citing academic research) — Corporate capital allocation exhibits enormous inertia. Asking "how much resources does this business need" from scratch each year is a core habit of excellent capital allocators.

6. "Dividends are seen as sacred, buybacks as residual" (Mauboussin, citing a survey by John Graham) — Although theoretically equivalent, CFOs have vastly different mental accounts for the two. Dividends cannot be cut, while buybacks are only done when there is spare cash.

7. "No asset class can escape a shift in real interest rates from -100bp to +120bp" (Mauboussin) — The over 200bp swing in real rates over the past 12 months has pressured all assets, but expected returns have risen significantly, creating mispricing opportunities for capital allocation.

8. "AI performance improvements have exceeded expectations, but who ultimately creates and captures value remains unclear" (Mauboussin) — The leap from GPT-2 to GPT-3 was astonishing, but two key questions remain unresolved: the impact on employment (history suggests it may not lead to job losses), and the attribution of value (which is not clear from the outset).