Reading Notes

Joel Greenblatt on Special Situations, the Magic Formula, and Paying Up for Quality

Episode: Richer, Wiser, Happier

Notes — Joel Greenblatt on Special Situations, the Magic Formula, and Paying Up for Quality

Source: Richer, Wiser, Happier (William Green, interview), c. 2021–22. Joel Greenblatt is founder of Gotham Capital (1985), creator of the Magic Formula, co-founder of the Value Investors Club (1999) and Success Academy (2006), former professor of value investing at Columbia Business School.

Four questions [Adler frame]

Q1 — What is it about? A career retrospective covering four decades of investment practice: hyper-concentrated special situations at Gotham Capital (1985–2000, 40% annualised); the evolution toward paying fair value for exceptional businesses (Moody’s as the first, modelled on Buffett’s Coca-Cola purchase); the creation of the Magic Formula as a systematic distillation of those principles; and the application of replicable-model thinking to education reform via Success Academy. Interwoven is Greenblatt’s account of the emotional life of investing — managing outside capital, surviving losses, finding joy in the game.

Q2 — How is it argued? Through case studies and counter-examples: the Host Marriott spin-off (complex capital structure → margin of safety → 40% position justified); the Florida Cypress Gardens sinkhole (risk you cannot anticipate); the Moody’s position (reverse-engineering Buffett; $10 Coke equivalent → $13 Moody’s, still warranted); the Magic Formula backtesting (one clean test, top decile beats second beats third — an ordered relationship). Arguments are inductive and often grounded in a specific trade or conversation.

Q3 — Is it true? The performance claims (40% annualised at Gotham) are well-documented and widely accepted. The Magic Formula back-test (1987–2004, published in The Little Book That Beats the Market) shows the ordered decile relationship but has underperformed in the 2010–2020 period of growth-stock dominance — which Greenblatt acknowledges obliquely when noting his diversified value portfolio ‘hasn’t been a great help’ during the bull market. The framework is intellectually honest: he describes the Magic Formula as ‘one test of something that makes sense,’ not data-mined. The critique that cheap-only (Tobias Carlisle’s approach) may beat Magic Formula is conceded.

Q4 — What of it? The central insight is that complexity is a source of margin of safety. When no analyst is looking at a situation and you have done the work to understand it, you have a structurally asymmetric bet — and that bet justifies concentration, not diversification. As businesses became more efficiently priced (internet-era information diffusion), the edge migrated from finding overlooked situations to owning structurally superior businesses at fair prices. The Magic Formula is the bridge: a systematic way to screen for cheap-and-good without requiring per-security deep dives.

Glossary

Special situation — a corporate event (spin-off, recapitalisation, merger stub, restructuring) that introduces complexity sufficient to drive away most analysts. The lack of attention — not the event itself — creates the margin of safety. Greenblatt’s phrase: ‘almost not investing, almost cheating.’

Unfair bet — a position where you know what something is worth and no one else is looking at it. The asymmetry between your knowledge and the market’s attention is the edge. You can ‘almost not gamble’ because the downside is bounded by your analytical work.

Margin of safety — Ben Graham’s term; the gap between intrinsic value and purchase price, providing buffer against analytical error. In special situations, complexity is the mechanism that creates the gap (other investors leave the field).

Return on tangible capital (ROTC) — Greenblatt’s quality screen in the Magic Formula: operating earnings / (net working capital + net fixed assets). Strips out intangibles. A high ROTC means the physical assets deployed in the business earn at a high rate. Buffett’s actual criterion for great businesses.

Earnings yield — the cheapness screen: operating earnings / enterprise value. Inverse of EV/EBIT. Equivalent to asking ‘what rate of return do I get if I pay today’s price for this year’s earnings?’

Magic Formula — the two-metric composite screen (ROTC + earnings yield) that Greenblatt developed in the early 2000s, backtested from 1987. Rank the market by both, sum the ranks, buy the top decile. One test; not data-mined. The ordered decile relationship (top > second > third) is the key validation.

The no man — Greenblatt’s term for a partner who will not agree to any idea until independently satisfied, regardless of the proposer’s track record. Rob Goldstein functions this way. Prevents the smart-solo-investor failure mode (catastrophic error with no check).

Value Investors Club — the online investment idea community Greenblatt and John Petry founded in 1999. Entry required submitting an investment write-up that would earn an A+ in Greenblatt’s Columbia class. Became the mechanism through which he identified Michael Burry and Norbert Lou.

Section notes

Stage 1: concentrated special situations (Gotham Capital, 1985–2000)

Greenblatt started in risk arbitrage and found the asymmetry intolerable: win $1–2 if the deal closes, lose $10–20 if it does not. He migrated to the ‘perimeter’ of deals — spin-offs given out as part of transactions, complicated pieces of paper thrown away in restructurings — where complexity created distorted prices.

The intellectual move: when you know what something is worth and no one else is looking at it, position size should reflect your confidence in the margin of safety, not diversification convention. The Host Marriott example makes the logic explicit: he paid $4 for (i) $6 in unencumbered assets (parent company had no corporate debt — the debt was asset-specific, like mortgages on individual buildings) plus (ii) an option on a subsidiary encumbered by debt but potentially worth a lot. The capital structure was so complicated that most analysts didn’t follow it. His conclusion: buying something debt-free at a two-thirds discount ‘doesn’t feel that risky to you.’ He put 40% of the fund in it.

The counterpoint: the Florida Cypress Gardens sinkhole. One of his first investments — a merger arbitrage deal where the target fell into a literal sinkhole weeks before close. The lesson is not carelessness but epistemic humility: ‘risk of sinkhole’ is not on any checklist; you cannot anticipate every tail. The response is not to avoid concentration but to ‘live to play another day’ — maintain enough diversification that a sinkhole doesn’t bankrupt you.

After 10 years he returned all outside capital. The motive was partly emotional: managing other people’s money kept him awake in a way managing his own money did not. He had found a business he loved and was making it unpleasant. The calculation: ‘I would have been silly to take a business I love and make it something that keeps me up at night.‘

Stage 2: paying up for quality (c. 2000 onward)

The transition began with Moody’s — the first time Greenblatt willingly paid more than 20× earnings. His method: reverse-engineer Buffett’s Coca-Cola purchase, make adjustments for Moody’s higher capital efficiency (Moody’s requires almost no reinvestment to grow; Coke requires some), and translate back to a ‘Buffett equivalent price.’ He calculated he was paying $13 for what Buffett paid $10 for — and Buffett quadrupled his money. He concluded the premium was warranted given quality.

The key conceptual move: separating return on tangible capital from total returns required for growth. A business earning 60% ROTC on 3 possible locations is worth less than one earning 30% ROTC on 1,000 possible locations. Quality is not just the rate of return; it is the scalability of that return.

His current portfolio includes concentrated positions in Google, Amazon, and Microsoft — ‘businesses like I’ve never seen in my career,’ held because he still does not believe they are fully priced given network effects, internet-scale reach, and compounding dynamics that break the normal ‘law of large numbers’ limits.

The Magic Formula: one test, not data-mined

Greenblatt is explicit about the epistemology: he did not spin the computer thousands of times to find the combination that worked best in back-testing. He chose two intuitive metrics — crude proxies for ‘good’ and ‘cheap’ — ran one test, and it worked. The intellectual honesty matters because data-mined strategies tend not to persist; strategies grounded in first principles may.

The ordered decile result is the key validation beyond just ‘top group beats market’: the 2nd decile beat the 3rd, the 3rd beat the 4th, all the way through — a monotone relationship between the composite rank and forward returns. This is harder to achieve by chance than top-vs-bottom.

He concedes the period caveat: if you had shorted the bottom decile against the top decile in 2000, you would have lost everything before the reversal. ‘Zero doesn’t compound very well.’ The practical lesson is long-only or modest net-long, with enough staying power to survive the period when the strategy underperforms.

On Tobias Carlisle’s cheap-only approach: it also works, is probably more volatile, and his own evidence suggests cheap-only does slightly better in back-tests. He runs both — cheap-only and cheap-plus-quality — and reports that since 2010 his diversified value portfolio ‘has done incredibly well, almost as good as the S&P 500,’ which he calls ‘incredible’ for a deep value strategy during the decade of growth-stock dominance.

Position sizing and the ‘not getting it right’ trap

One of the most quotable frameworks in the episode: ‘Finding one of the best things you’ve ever seen and putting 2% in it — that’s not getting it right, that’s getting it wrong.’ The insight is that position sizing must be commensurate with your degree of genuine understanding and the width of the margin of safety. If you understand the bet and the downside is bounded, not concentrating is an error — you’ve done the work of identifying an edge and then failed to take it.

This depends critically on the premise that you actually understand the bet. Without that understanding, concentration is not courage but recklessness. Knowing what you own is what makes emotional endurance possible: when the price falls from $6 to $5 on something you think is worth $10, you know it’s an opportunity rather than a signal to sell.

Partner dynamics: the no man as circuit-breaker

The Buffett/Munger dynamic — Munger as the ‘abominable no man’ — is Greenblatt’s frame for his own partnership with Rob Goldstein. Goldstein joined in 1989, is a stronger analyst in Greenblatt’s own assessment, and does not defer to Greenblatt’s track record. Every idea must satisfy both independently.

The structural importance: brilliant solo investors sometimes make errors that would be obvious to any peer, because there is no one they will listen to. The longevity of a partnership depends on both partners actually wanting the truth rather than validation. He notes this is psychologically harder than it sounds — it requires treating a wrong call as ‘our mistake’ because both agreed to it.

Value Investors Club and talent identification

The VIC was founded on a specific insight: there is talent in unlikely places (the supermarket deli counter analyst who had independently identified the same complex situation Greenblatt had), and the internet could surface it if you set the right filter. The entry filter — investment write-up good enough to earn an A+ in a Columbia class — is deliberately high. The result over years was a community of practitioners who think similarly.

The talent identification heuristic: he is not looking for someone who is right about the investment he already likes. He is looking for someone who anticipates his questions in sequence — who says the next thing he was about to ask before he asks it. That signal of shared mental models is what made him back Burry and Lou.

Education: replicable model as investing principle

The Success Academy parallel is explicit in the transcript. Greenblatt and Petry applied the same ‘replicable-model, output-measured’ framework to education that Greenblatt applies to investing:

  • Invest in inputs (teacher credentials, class size) → equivalent to buying based on process rather than results
  • Measure outputs (student achievement vs. control) → equivalent to returns

The key constraint they identified: if your model requires the top 1% of teachers, you cannot scale, just as a strategy requiring the top 1% of investment ideas cannot scale. The replicable model must be good enough to make the average excellent.

Current scale: 47 schools, ~23,000 students, predominantly low-income minority children who outperform students from the wealthiest New York school districts. The ‘communist capitalist’ reading of Success Academy — a for-output-performance institution built on market logic deployed in a public service context — is in the background throughout.

Cross-references