Ben Okonkwo: Here's a number that genuinely unsettled me when I saw it. U.S. margin debt hit one-point-oh-two trillion dollars in July 2025. FINRA's figure. That's a record.
Jonathan Ingles: And nobody's panicking. Which is either reassuring or it's exactly the problem.
Ben Okonkwo: Right, and — okay, so the core idea here is actually simple. Leverage means borrowed money amplifies whatever happens to your position. A five-to-one ratio: one dollar of your equity controls five dollars of assets. Gain ten percent on the assets, your equity doubles. Lose ten percent — your equity is gone.
Jonathan Ingles: That's the clean version. The version nobody teaches you is what happens when someone else decides to sell.
Ben Okonkwo: Which is exactly the Archegos story. Hwang's fund — technically solvent, the math actually worked — until one lender flinched and ten banks were in forced liquidation simultaneously. The positions weren't the problem. The confidence was.
Jonathan Ingles: Leverage isn't a math problem. It's a trust problem wearing math's clothes.
Ben Okonkwo: Now, the math is worth spelling out though, because — okay. Take a hundred dollars of your own equity, borrow four hundred at five percent, you're controlling a five-hundred-dollar position. Ten percent market gain, that's fifty dollars profit. On your hundred. That's a fifty percent return.
Jonathan Ingles: And the same math runs in reverse.
Ben Okonkwo: Exactly — ten percent loss, fifty dollars gone, fifty percent of your actual capital. The leverage ratio just scales both directions symmetrically. That's the textbook. But wait — actually, that's where it starts to get misleading. Because the interest doesn't stop.
Jonathan Ingles: The borrowing cost is running the whole time. Every day you're levered, you're paying for it.
Ben Okonkwo: Which means the symmetry is — it's actually broken from the start. Gains are being shaved by the financing cost. Losses are being compounded on top of a shrinking equity base. And then the margin call mechanism... that's where I mean, that's the real trapdoor. The broker isn't waiting for your math to resolve itself.
Jonathan Ingles: They liquidate at the worst moment. Which locks in the loss.
Ben Okonkwo: And risk of ruin doesn't scale linearly with the leverage ratio — it accelerates. Exponentially. Because every loss reduces the equity base, which means the same dollar loss next time is a bigger percentage hit. The textbook is technically correct and practically a trap.
Ben Okonkwo: Now, Long-Term Capital Management is the case that should have ended this. Nobel Prize-winning economists — rigorous models, genuine edge in fixed-income arbitrage — and September 1998, Russia defaults, and their highly leveraged positions unravel simultaneously. Fourteen banks. Three-point-six billion dollars. William McDonough at the Federal Reserve Bank of New York essentially convening a bailout room.
Jonathan Ingles: And the Fed didn't deploy its own funds. That's the part that gets buried. McDonough facilitated — he got the banks in the room — but the money was private. Which tells you exactly who was on the hook.
Ben Okonkwo: Right. And then Archegos, 2021 — same failure mode, completely different instrument. Total return swaps. Hwang gets all the economic exposure to a stock — price moves, dividends — without the shares ever showing up on any ownership record. Each prime broker sees one slice. None of them sees twenty-times leverage spread across ten of them.
Jonathan Ingles: The opacity wasn't a side effect. It was the mechanism.
Ben Okonkwo: Exactly — and that's where Kelly becomes interesting, actually — wait, not interesting, it's where Kelly becomes inadequate. The Kelly Criterion tells you the optimal fraction of capital to bet, maximizes your geometric growth rate. Fractional Kelly cuts that stake, trades some growth for much lower risk of ruin. The math is clean. But MacLean, Thorp, and Ziemba — they catalogued this systematically — the whole framework assumes you know your probability distribution. Stable, known, repeatable.
Jonathan Ingles: Which markets are not.
Ben Okonkwo: Non-stationary by definition. LTCM's models were built on historical correlations that broke the moment the crisis hit. So the question isn't whether they overbid Kelly — it's whether Kelly was ever solving the right problem. And that same blindness is what made LBOs look sophisticated while retail margin looked reckless, even though the institutional failures were the ones that needed fourteen-bank rescues.
Jonathan Ingles: And Archegos proved it, structurally. Total return swaps make the leverage invisible to regulators until the crisis — that opacity isn't a gap in oversight, it's the product. The instrument was designed that way.
Ben Okonkwo: Which is the part I can't quite resolve. Because — okay, even if you mandate full disclosure, even if every prime broker sees Hwang's complete position across all ten counterparties... I mean, does that actually stop them from lending? LTCM had sophisticated counterparties who understood the risk.
Jonathan Ingles: They lent anyway.
Ben Okonkwo: They lent anyway. So the transparency argument assumes rational restraint. And the historical record — LTCM, Archegos — doesn't really support that assumption.
Jonathan Ingles: Frankly, that's the darker read. Opacity is the excuse. Remove the excuse, you still have the incentive. The prime broker profits from the leverage whether or not they're holding the loss when it unwinds.
Ben Okonkwo: Right. And now margin debt's at a trillion dollars, and — hm. I keep turning over whether the next version even uses total return swaps. Or something we genuinely haven't categorized yet.
