The Legendary Debate

The Investment Council Debates Equifax (EFX)

Warren Buffett: leans back and folds his hands "Here's what I've been wrestling with on Equifax. In 2017, this company suffered the worst data breach in corporate history — 147 million Americans had their most sensitive information exposed — and yet here we sit, nine years later, and the business is bigger, more profitable, and more entrenched than ever. No customers left permanently. No competitor gained meaningful share. That tells you something extraordinary about the nature of this oligopoly. So my question for the room is this: is Equifax a business that succeeds in spite of management, because the structural position is so dominant that even catastrophic failure can't dislodge it? And if that's true — if the moat is really that wide — then why does the financial performance over the past decade look so... ordinary?"

Charlie Munger: "You've asked exactly the right question, and the answer is uncomfortable for anyone who wants this to be a simple story. Equifax survived the breach for the same reason the railroads survived a century of mismanagement — the infrastructure is irreplaceable, and the customers have no alternative. You cannot build 1.4 billion credit records from scratch. You cannot replicate The Work Number's 209 million active payroll records by wishing them into existence. These are accumulated data assets that took decades to assemble, and they generate economic value every time someone applies for a mortgage, a credit card, or a job. But here's where I apply inversion: what could kill this business? Three things keep me up at night. First, a second data breach — and this time, the political response might not be a fine but structural regulation, maybe utility-style rate-setting for credit bureaus. Second, the CFPB has issued three Civil Investigative Demands targeting The Work Number specifically. That's the crown jewel asset, and the government is looking at it with suspicion. Third — and this is the subtle one — management has spent $5.8 billion on acquisitions and $3 billion on cloud transformation over the past decade, and the returns on that capital are worse today than before they started spending. Show me the incentive, I'll show you the outcome — and I worry management's incentive is to build an empire rather than optimize returns."

Dev Kantesaria: "Charlie, I want to address your CFPB concern head-on because I think it's where the market is making its biggest mistake. I've studied regulated toll booth businesses for eighteen years — Visa faced DOJ antitrust, Moody's was hauled before Congress after 2008, MSCI faces constant fee pressure from index funds. In every case, the regulatory scrutiny felt existential at the time and proved manageable in practice. The reason is structural: regulators understand that breaking critical infrastructure creates more harm than the monopoly pricing they're trying to constrain. The Work Number isn't just a profit center for Equifax — it's the verification engine that enables the mortgage closing process to happen in days instead of weeks. If the CFPB restricts access or mandates pricing changes, the immediate effect is that mortgage closings slow down, government benefit verification becomes less accurate, and employer compliance burden increases. Regulators tend to discover these second-order consequences before they act, not after. Now, to Warren's core question about whether this business succeeds in spite of management — I'd reframe it. This is a toll booth business going through a capital investment cycle. The cloud transformation consumed $3 billion in capex during 2020-2023, which is why the returns look depressed. But capex has already declined from $625 million to $481 million, and the platform is enabling new product launches at a 15% vitality index — 500 basis points above target. The investment is beginning to pay off. That's not management incompetence; that's a long-cycle infrastructure rebuild."

Mohnish Pabrai: "Dev, I respect the toll booth framework, but I want to challenge something you're glossing over. You're comparing Equifax's regulatory risk to Visa and Moody's, but those companies were earning extraordinary returns when regulators came knocking — Visa at 40%+ ROIC, Moody's at 50%+. The regulatory threat was to compress outstanding economics toward merely good. Equifax is earning 8% ROIC — barely above the cost of capital. If the CFPB imposes pricing constraints on The Work Number's verification fees, you're not compressing great returns toward good; you're compressing mediocre returns toward value-destructive. That's a fundamentally different risk profile. And let me raise another point that everyone in this room seems to be dancing around: the 2017 breach didn't just cost $1.4 billion in settlements. It created a permanent cost overhang in security, compliance, and regulatory infrastructure that shows up in the margin compression from 26% operating margins pre-breach to 18% today. That's not a temporary drag — that's the new structural reality of operating a business that holds the most sensitive consumer data in America."

David Tepper: "Mohnish, you're right about the margin compression, but you're wrong about the conclusion. You're looking at this as a permanent impairment. I see a classic post-investment-cycle setup — and I've made my career buying exactly these situations. Think about what's happening here: Equifax spent the equivalent of two years' worth of operating cash flow rebuilding its entire technology infrastructure from the ground up. During that period, every metric that value investors watch — ROIC, FCF, margins — looked terrible. The market priced it accordingly. But the spending cycle is ending. Capex peaked at $625 million and is already down to $481 million. Free cash flow was negative $2 billion in 2021, and it's $1.06 billion today. That's a $3 billion swing in cash generation in four years. And here's what the bears are missing: CEO Begor said on the earnings call that they grew revenue 7% even with the mortgage market down 7% and the hiring market down 2%. Strip out the cyclical headwinds, and the underlying business is growing at high single digits in a weak environment. When those end markets normalize — and they will, because Americans buy houses and employers hire workers — you get a revenue tailwind on top of a margin expansion story on top of a declining capex trend. That's three separate levers all pulling in the same direction."

Robert Vinall: "David, your framing is compelling, but I want to probe the reinvestment economics more carefully, because that's where the bull and bear cases ultimately diverge. The question isn't whether capex is declining — it clearly is. The question is whether the new cloud platform generates better unit economics than the old infrastructure. If Equifax spent $3 billion to build a platform that enables faster product development, lower marginal delivery costs, and higher incremental margins, then the investment was brilliant and ROIC will recover toward 12-13% as revenue grows on the new infrastructure. But if the platform merely replaces aging technology at roughly the same economic return — just preventing deterioration rather than enabling improvement — then management spent $3 billion to run in place, and the 8% ROIC is the new normal. The early evidence is mixed. On the positive side, the vitality index hit 15% — meaning new products contributed $900 million in revenue last year, which is genuinely impressive product innovation velocity. On the negative side, operating margins have flatlined at 17.7-18.5% for three straight years despite the cloud migration being largely complete. If the cloud were generating operating leverage, margins should be expanding. They aren't — not yet."

Pulak Prasad: "Robert, you've identified the pivot point, and I want to add an evolutionary perspective. The 2017 breach was the most important event in Equifax's history — not because of the financial cost, but because of what it revealed about the organism's survival capacity. Every credit bureau's competitive moat is ultimately trust — the trust of consumers that their data is secure, the trust of lenders that the data is accurate, and the trust of regulators that the industry operates responsibly. Equifax shattered all three forms of trust simultaneously. And yet the organism survived. Not because management responded brilliantly — the initial response was widely criticized — but because the ecosystem couldn't function without it. Lenders didn't stop pulling Equifax credit reports. Employers didn't stop contributing to The Work Number. Regulators didn't shut the bureau down. This is what I call a keystone species — an organism so deeply embedded in its ecosystem that removing it would collapse the entire structure. Kodak had 80% film market share and vanished because digital cameras were a genuine alternative. Equifax has a 33% credit bureau market share and survived a catastrophic breach because there is no alternative — you can't build a replacement ecosystem from scratch. But here's my concern: management has layered financial leverage onto this resilient franchise. Negative working capital of $962 million, $6.15 billion in total debt, and capital returns exceeding free cash flow in FY2025. The organism is durable, but the financial structure is optimized for favorable conditions. And I'm not sure conditions will stay favorable."

Warren Buffett: "Let's turn to the numbers, because I want to test whether the financial history supports the qualitative story we've just debated. The headline that jumps off the page for me is this: Equifax roughly doubled its revenue over the past decade, from $3.1 billion to $6.1 billion. That's a respectable 7.6% compound annual growth rate. But GAAP EPS went from $4.13 to $5.52. That's 34% total — a 3.3% annual compounding rate over nearly a decade. So for every dollar of new revenue the business generated, only about seventeen cents reached the bottom line. That gap between revenue growth and per-share earnings growth is the single most important number in this entire analysis, and it tells a story of massive capital consumption with disappointing translation to shareholder value."

Charlie Munger: "Warren's seventeen-cents-on-the-dollar figure is devastating, and it demands an explanation. Where did the other eighty-three cents go? The answer is a combination of acquisition goodwill amortization, cloud transformation depreciation, breach remediation costs, and higher ongoing security and compliance expense. The $5.8 billion in acquisitions shows up on the balance sheet as $6.75 billion in goodwill — which means Equifax paid substantial premiums for these businesses — and the amortization on those deals flows through GAAP earnings as a persistent drag. This is why adjusted EPS of $7.65 diverges so dramatically from GAAP EPS of $5.52 — a 39% gap, which is among the widest in the S&P 500. Management would have you focus on the adjusted number. I would have you focus on the gap. Serial acquirers always want you to ignore amortization because it makes their deals look accretive. But amortization exists because they paid for those assets, with real shareholder capital, at premiums to tangible book value. If you buy a company for $1 billion and it has $200 million in identifiable intangibles, that $200 million amortizes through earnings for the next fifteen years. It's not 'non-recurring' — it's the ongoing cost of the acquisition strategy."

Dev Kantesaria: "Charlie, I hear the accounting critique, but I think it conflates two different capital allocation decisions that deserve separate evaluation. The cloud transformation was a mandatory infrastructure investment — Equifax was running on technology so antiquated that it enabled the 2017 breach. The choice wasn't 'should we spend $3 billion?' It was 'should we spend $3 billion or face existential risk?' Judged on those terms, the investment was rational. The acquisitions are a different story and deserve more scrutiny. But here's what I keep coming back to: The Work Number. This is a data asset covering 209 million active payroll records from employers contributing voluntarily. That flywheel — more employers contribute, making the database more valuable, attracting more verification buyers, which incentivizes more employers — is the Equifax compounding engine. And it's strengthening: they added 20 million records last year alone, an 11% increase. Government verification revenue is growing low double digits. The vitality index is 500 basis points above target. If you strip out the legacy credit bureau and focus on Workforce Solutions — 51.5% EBITDA margins, 8% revenue growth, accelerating to 10% in Q4 — you're looking at a genuine compounding machine that happens to be attached to a decent-but-unexceptional credit bureau business."

David Tepper: "Dev's right to isolate Workforce Solutions, and I want to sharpen the point because this is where the catalyst math gets interesting. Mortgage revenue is approximately 20% of total Equifax revenue per management's own commentary. It was down 7% in FY2025. Think about what that means: Equifax grew total revenue 7% despite a headwind from its most cyclical and highest-margin segment. If mortgage origination volumes simply normalize — not boom, just normalize — that 20% revenue segment recovers with near-total operating leverage because the fixed infrastructure costs are already covered. A 15-20% mortgage recovery could add $100-150 million in operating income. Meanwhile, D&A is running at $720 million — up from $269 million in 2016 — and the heaviest cloud investments capitalized in 2020-2022 will begin fully depreciating over the next three years. Even if capex remains at $450-500 million, the D&A will mechanically decline as old assets roll off faster than new ones are added. That's the margin expansion catalyst that I think the market is completely ignoring."

Mohnish Pabrai: "David, I want to push back on the D&A rolloff narrative because I think it's doing more work in the bull thesis than the evidence supports. D&A has gone from $200 million in 2015 to $269 million in 2016 to $490 million in 2021 to $681 million in 2024 to $720 million currently. It has risen every single year for a decade. At what point do we stop projecting a decline and start acknowledging that this might be a permanently elevated cost of operating a modern cloud-native data business with 400+ AI patents and ongoing security infrastructure requirements? Capex is still running at $481 million — yes, it's down from the peak, but it's nearly double the $200-250 million pre-transformation run rate. If capex stabilizes at $400-450 million and assets depreciate over 5-7 years, D&A doesn't decline — it stabilizes at roughly current levels. The margin expansion bull case requires D&A to drop $200-300 million. Show me the math for why that happens when capex remains this elevated."

Robert Vinall: "Mohnish raises a fair point about D&A, and I want to approach it from the reinvestment perspective. The real question for me is whether the new cloud platform changes the incremental return on capital. The blended ROIC of 8% reflects the accumulated weight of $10.6 billion in invested capital. But what matters going forward is the return on the next dollar invested. If Equifax can now launch new data products at a 15% vitality rate on relatively modest incremental investment — because the cloud platform is built and the marginal cost of new product development has dropped — then the incremental ROIC could be substantially higher than the blended figure. That would show up as gradual improvement in overall ROIC even if D&A doesn't decline dramatically. The test is simple: over the next eight quarters, does operating income grow faster than revenue? If yes, the operating leverage thesis is real. If operating income merely grows in line with revenue, the cloud investment was a treadmill, not a springboard."

Pulak Prasad: "I want to address Guy Spier's 'away from desk' question, because I think it cuts to the heart of whether Equifax is a long-term holding or a trading position. If I owned Equifax and couldn't check the price for twelve months, would I sleep well? Honestly — no. Not because the business would collapse, but because this is a company with $6.15 billion in debt, negative working capital, $1 billion in short-term maturities, CFPB investigations targeting its crown jewel asset, and a management team that returned more capital to shareholders last year than the business generated in free cash flow. Any one of these is manageable. But the combination means this is a business where the downside scenarios — a recession coinciding with refinancing needs, or a regulatory action coinciding with a housing downturn — could create genuine stress. Compare this to FICO, which sits upstream of Equifax with 47% operating margins and $9 million in annual capex. FICO is a toll booth I'd happily own without checking the price. Equifax is a toll booth with a lot of maintenance debt."

Warren Buffett: "All right, the stock's at $169.50. Let's talk about what that price buys you and whether it makes sense. Dev, you've been the most constructive — what's your case for buying here?"

Dev Kantesaria: "At $169.50, you're paying approximately 14.7x EV/EBITDA and getting a 5.1% free cash flow yield on a toll booth data oligopoly at the trough of both a capital investment cycle and a mortgage origination cycle. My back-of-envelope math is straightforward: FCF per share of $8.84 today, growing at 10-12% annually as capex normalizes and Workforce Solutions compounds. Within two years, the FCF yield on a $169.50 cost basis exceeds 6.3%, well above the risk-free rate. For a business embedded in the essential infrastructure of American credit and employment decisions, that's a price I'm willing to pay. I own Equifax as a core holding because the toll booth is structural. Every mortgage, every credit card, every tenant screening, every employment verification — the data flows through Equifax or one of two competitors. That's inevitability. I'm buying."

David Tepper: "I'm also buying at current prices, and my reasoning is catalyst-driven. This business is sitting at the intersection of three normalization trends: capex declining, mortgage recovering, and D&A approaching its peak. If Q2 or Q3 2026 operating margins cross above 20%, the market will reprice this stock to 18-20x EBITDA — that's $220-$250. Even my downside scenario — mortgage weakens further, margins stagnate, EPS drops to $4.80 — puts the stock at $120 at 25x trough earnings. So I'm risking maybe 29% to make 30-50%. The expected value favors buying."

Robert Vinall: "I share Dev and David's conviction on the business quality, but I want a margin of safety the current price doesn't provide. My fair value is approximately $180 based on a 10-year DCF with 10% FCF growth and a 9.5% discount rate. I'd start buying at $150 — a 17% discount that gives me cushion against the ROIC stagnation scenario Mohnish described. At $150, the FCF yield approaches 6%, and even if margins don't expand materially, the compounding of revenue growth and modest buybacks should deliver a satisfactory return."

Mohnish Pabrai: "My number is $135, and let me explain why. At 31x trailing GAAP earnings with an 8% ROIC, I need to see a path to tripling my money to justify the risk. At $169.50, a triple means a $63 billion market cap. For a company earning $5.52 per share, that requires EPS to reach $17-18 at a 30x multiple. That's heroic. At $135, my downside in a recession — EPS drops to $4.00, stock at 25x = $100 — is 26%. My upside if the margin recovery plays out — EPS reaches $8.00, stock at 28x = $224 — is 66%. That's a 2.5:1 ratio. Not quite my preferred 3:1, but adequate for an oligopoly. The point is, I need the price to give me the asymmetry that the business quality doesn't provide at current ROIC levels."

Charlie Munger: "I'd start buying at $142. My approach is inverted: I want a price where even if the worst plausible scenario materializes — CFPB restricts TWN pricing, mortgage stays depressed, D&A remains elevated — I'm still owning a $6 billion revenue oligopoly at a price that doesn't permanently impair my capital. At $142, the EV/EBITDA is roughly 12.5x. That's what I'd pay for an oligopoly with execution uncertainty. The quality of the franchise deserves a premium, but the quality of recent capital allocation does not."

Pulak Prasad: "I'm at $148 — close to Robert and Charlie. The organism survived the 2017 breach, which proves the franchise is durable. But the leveraged balance sheet means I need a purchase price that works even if the deleveraging doesn't happen as fast as the bulls hope. At $148, I'm paying roughly 13.5x EV/EBITDA for a keystone species in the American credit ecosystem. That's a price that acknowledges both the franchise durability and the financial fragility."

Warren Buffett: surveys the room "Let me try to synthesize what we've arrived at, because this has been one of the more nuanced discussions we've had. On the qualitative side, we achieved near-unanimity — and that's significant. Equifax occupies a structural position that no one in this room believes can be replicated. You cannot build 1.4 billion credit records from scratch, and The Work Number's 209 million payroll records represent a data asset with genuine flywheel characteristics that strengthen with every new employer integration. Pulak's 'keystone species' framing captures it perfectly: even a catastrophic breach couldn't dislodge this organism from its ecosystem. The industry structure is getting better for incumbents, not worse — the data accumulates, the switching costs compound, and no new entrant has cracked this oligopoly in half a century."

"Where we part company — and this is where the genuine analytical insight lies — is on whether the past decade's massive capital deployment has improved the business or merely maintained it. Dev and David see a classic post-investment inflection: capex declining from $625 million toward $400 million, FCF swinging from negative $2 billion to positive $1 billion, Workforce Solutions accelerating to 10% growth in Q4 with 51.5% EBITDA margins. They're buying the recovery before the market prices it. The rest of us see the same data and draw a more cautious conclusion: ROIC has declined from 13% to 8% during a period when the business doubled its revenue. Operating margins have flatlined at 18% for three years despite the cloud transformation being largely complete. As Mohnish pointed out, D&A has risen every single year for a decade — the projected decline remains theoretical. Charlie's point about the 39% GAAP-to-adjusted earnings gap reminds us that the acquisition-heavy strategy has a real ongoing cost that adjusted metrics conveniently hide."

"At $169.50, the stock is approximately fairly valued on our blended framework — not cheap enough for five of us, but not unreasonable for two who see the cyclical inflection as imminent. I'd start buying at $150, where the FCF yield improves to nearly 6% and the margin of safety compensates for the ROIC uncertainty. Five of us want a lower price; two would buy today. The disagreement is honest and instructive: it comes down to whether you trust the forward-looking operating leverage story or demand the financial evidence first. I prefer to see the evidence. But I'll be ready when it arrives."