Maya Chen: Nathan, I have a number for you — tell me if this lands the way it landed for me.
Dr. Nathan Hayes: Go ahead.
Maya Chen: The initial subprime mortgage losses in 2008 — under five hundred billion dollars. The global household wealth destroyed — sixteen trillion. I've been sitting with that ratio all week and I can't... I can't make it feel real. Like, the thing that broke the world was smaller than a routine bad month in the stock market.
Dr. Nathan Hayes: Now, the ratio landing emotionally is actually important here — because that's what makes people reach for the word 'greed' and stop there. But the thirty-two-fold difference isn't moral, it's mechanical.
Maya Chen: Wait — thirty-two-fold, you're doing the math in real time.
Dr. Nathan Hayes: Sixteen trillion over five hundred billion, yes, roughly. And I want to be careful — I'd want to separate out how much of that sixteen trillion is pure leverage unwind versus secondary wealth effects, people feeling poorer and spending less. Those are different phenomena mechanically. But even the conservative slice is extraordinary.
Maya Chen: Hmm — so even the careful version of the number is still the story. And the story is leverage. Which is... I mean, everyone uses the word. Borrowed money, right? You borrow to control more than you own. But I think what today is actually about is — why does something that sounds so straightforward become so hard to see when it's winding up?
Dr. Nathan Hayes: Because of procyclical leverage. During the expansion, leverage feels rational — prices are rising, collateral looks solid, lenders loosen standards. The very success of the bet encourages more borrowing. And then collateral values fall, margin calls arrive, and the loop runs exactly backwards at speed.
Maya Chen: The trap was built while everything looked fine. That's what we're going to trace.
Dr. Nathan Hayes: Let me give you the kitchen-table version first, because the math is actually simple — shockingly so. You put down ten thousand dollars, borrow ninety thousand, buy a hundred-thousand-dollar asset. It rises ten percent, you've doubled your money. It falls ten percent — you're wiped out entirely. The borrowed money means every percentage move is counted against the whole position, not just what you own.
Maya Chen: That's — yeah. You own a fraction but you're exposed to everything.
Dr. Nathan Hayes: Exactly that. And now here's where it stops being intuitive — that math is identical whether you're talking about a mortgage, a margin account, a repo agreement, a futures contract, a leveraged ETF. The asset class changes. The time scale changes. The arithmetic does not.
Maya Chen: Wait — identical how? Like, I can see a mortgage and a margin account being similar, but an ETF?
Dr. Nathan Hayes: So — this is actually my favorite example and also, I think, the most quietly alarming one. A two-times leveraged ETF hits its stated daily return target. Perfectly. Every single day it does exactly what it says it does. And over a multi-year period with a volatile index that's ultimately flat, it can still systematically underperform two times the index return. That's volatility decay — what some people call the constant leverage trap.
Maya Chen: Hold on. Is that — is that fraud?
Dr. Nathan Hayes: No. And that's precisely the problem. The product is not fraudulent. The math is correct. The customer's intuition about what 'two-times leverage over five years' means — that's wrong. Daily rebalancing in volatile markets compounds against you in ways that don't match how people think about multiplication.
Maya Chen: So the product works and still misleads most people who buy it.
Dr. Nathan Hayes: Structurally, yes. And this — this is where Thurner, Farmer, and Geanakoplos become important. Their 2009 model showed formally that leverage alone, even with completely stable underlying fundamentals, generates fat-tailed price distributions and clustered volatility. The instrument doesn't matter. The fragility is in the leverage.
Maya Chen: Fat tails meaning — extreme moves happen way more often than you'd expect.
Dr. Nathan Hayes: More than a normal distribution would ever predict, right. And at ten-times leverage, a five-percent price decline is enough to trigger a margin call — a demand to post more funds or liquidate. You don't need a crash. You need a bad Tuesday.
Maya Chen: That's the part that sort of — I mean, when you hear it that way, it's not the catastrophe that breaks you. It's the ordinary bad day that the structure wasn't built to survive.
Dr. Nathan Hayes: And the person inside that margin call doesn't see the mechanism. They see a number on a screen and a deadline. Which is probably the next thing we need to actually look at.
Maya Chen: And that number on the screen — that's sort of where I want to pick up, because what I keep getting stuck on is how that bad Tuesday becomes everybody's bad Tuesday. Like, the person inside the margin call didn't make anyone else's position worse. Except... they did.
Dr. Nathan Hayes: That's the liquidity spiral. And it's — the mechanism is almost embarrassingly simple once you see it. Prices fall, collateral is worth less, lender demands more collateral or liquidation, the forced sale pushes prices lower, which makes the next investor's collateral worth less.
Maya Chen: It feeds itself.
Dr. Nathan Hayes: Endogenously, yes. The selling isn't driven by new bad news about the asset. It's driven by the structure of the lending. And the — now, here's the part people miss — the fire sale harms everyone holding the same asset class. You didn't borrow. You didn't get a margin call. But your neighbor's forced liquidation just reset the comparable price for your position.
Maya Chen: Which is — wait, that's literally a homeowner in 2007. Watching a foreclosure sale close down the street and suddenly their own home's value has dropped. They didn't touch their mortgage. Nothing about their situation changed. But the price floor moved under them.
Dr. Nathan Hayes: Exactly that scenario. The forced sale creates a negative externality for every other leveraged position in the same asset class. It's not transmitted through sentiment — it's transmitted through comparable valuations and collateral rules.
Maya Chen: And this — I mean, this isn't new, right? I keep thinking, we've watched this before. 1929 — I always assumed that was just, I don't know, euphoria, speculation, people being reckless. But the research on it is actually more specific than that.
Dr. Nathan Hayes: Much more specific. Borowiecki, Dzieliński, and Tepper traced what actually happened in 1929, and it's — so, the margin requirements tightened in the first nine months of 1929. Before the crash. That tightening, combined with September price declines, made enough investors collateral-constrained that the system was already at the edge. October and November weren't the cause. They were the cascade that a pre-stressed structure produced.
Maya Chen: So the crash was the liquidity spiral running, not the thing that started it.
Dr. Nathan Hayes: Right. And Long-Term Capital Management in 1998 is the same architecture, just — derivatives instead of margin accounts, global instead of domestic. LTCM was using derivatives leverage so extensively that when their positions began unwinding, it moved across global markets simultaneously. The Federal Reserve had to broker a coordinated bailout. One fund.
Maya Chen: One fund required a Fed-brokered intervention. That's — the scale of the contagion from a single node is what I find hard to sit with.
Dr. Nathan Hayes: And Schularick and Taylor documented that this isn't exceptional — 14 countries, 1870 to 2008, the same procyclical pattern repeats across every regulatory regime they looked at. Credit booms reliably precede financial crises. 138 years of it. Which means — actually, the piece that makes all of this considerably harder is what Geanakoplos's framework says about *why* the spiral is so hard to interrupt. But that's where we need to go next.
Maya Chen: Yeah — because if the transmission mechanism isn't interest rates at all, then the main tool we reach for in a crisis might be pointed at the wrong thing entirely.
Dr. Nathan Hayes: That's exactly the inversion Geanakoplos is making. And it's — so, the conventional model assumes price declines trigger the cascade. Watch the asset price, adjust the interest rate, done. But what Geanakoplos's leverage cycle framework actually says is: the trigger is a shift in collateral terms. Haircuts. Lenders wake up on a Wednesday and demand more collateral per dollar borrowed — and the underlying asset hasn't moved at all.
Maya Chen: Wait — the asset price didn't change?
Dr. Nathan Hayes: Doesn't have to. The haircut widens — that's the lender's discretionary term — and suddenly the borrower is undercollateralized by definition. Forced to deleverage. Before any fundamental repricing has occurred.
Maya Chen: So the Fed is watching asset prices, adjusting interest rates — and the actual pressure point is... a term in a lending agreement that nobody's monitoring in real time.
Dr. Nathan Hayes: Geanakoplos's literal argument — and this is taught in MIT's Financial Crises course, Alp Simsek formalizes exactly this transmission — is that the Fed would, quote, 'do better to attend to the economy-wide leverage and leave the interest rate alone.' That's not a fringe position. It's just institutionally unresolved.
Maya Chen: Institutionally unresolved meaning — we know this and still don't act on it?
Dr. Nathan Hayes: Meaning the Financial Stability Board, the ECB, the Basel Committee — they have macroprudential leverage ratio tools, capital adequacy requirements, yes. But the structural incentive to maximize leverage hasn't changed. And that's — I mean, this is where DeAngelo and Stulz come in, because their argument is that high leverage isn't irrational for banks. It's not even a risk preference. Providing liquid claims against illiquid assets — that's banking. That business model requires high leverage to function.
Maya Chen: The systemic risk is baked into what a bank actually is.
Dr. Nathan Hayes: Structurally, yes. Which is why 2008 ran through repo markets and prime broker financing specifically — lightly regulated channels, almost no haircut oversight — and haircut tightening there forced deleveraging before asset prices had even fully fallen. The spiral was running in the plumbing before it showed up on the screens.
Maya Chen: That's — that's the part I find genuinely hard to sit with. Because if the transmission is in the funding terms, in those repo agreements, then the thing that actually needed watching was... sort of invisible to the standard instruments.
Dr. Nathan Hayes: Right — and Simsek's framework formalizes exactly why. The transmission runs through funding terms, not price signals. So a central bank adjusting the federal funds rate is, in a leverage crisis, moving the wrong dial.
Maya Chen: Wrong dial. That's a devastating way to put it, actually — not because the rate doesn't matter, but because by the time the rate matters, the haircut has already done the damage.
Dr. Nathan Hayes: Precisely. The leverage cycle — the boom where collateral terms loosen, lending expands, everyone's position looks healthy — that's when the fragility is accumulating. And the policy response is calibrated for the bust, not for the quiet accumulation.
Maya Chen: Which means the moment you'd actually need to intervene is the moment when nothing looks wrong. When the numbers are good and the terms are loose and everyone's comfortable. That's when the trap is being set.
Dr. Nathan Hayes: And the uncomfortable corollary to that is — the early warning signal isn't the asset price moving. It's haircut creep. Margin tightening in repo markets, in prime broker agreements, in unregulated lending channels that the IMF tracks in its Global Financial Stability Reports precisely because no one else has eyes on them in real time. That's the canary. And it doesn't make the news.
Maya Chen: Which means by the time it's visible — it's already running. Central banks can't actually stop a margin call. They can try to prevent the shock that triggers it, but once the haircut widens and the collateral math shifts... the call goes out.
Dr. Nathan Hayes: That's — yes. Precisely that. The Fed doesn't have a margin call lever. It has an interest rate and emergency liquidity tools, and those matter, but they reach a different part of the machine. The collateral regime — who sets haircuts, when, under what rules — that's still not treated as a direct policy instrument. Not consistently. The Basel framework, macroprudential ratios — the tools exist on paper. The structural incentive to maximize leverage hasn't moved.
Maya Chen: So the open question just... sits there. Whether regulators start treating the collateral regime as the actual lever — not a side note, not a macroprudential afterthought — but the thing you watch the way you watch the federal funds rate. And I don't know the answer to that. I don't think we do.
Dr. Nathan Hayes: Neither do I. And I find that — mm, I don't want to say alarming, but it's not comfortable. Geanakoplos made this argument clearly enough that it's in MIT's curriculum. The IMF publishes it. And the next cascade is, in some unregulated lending channel right now, accumulating quietly while the asset prices look fine.
Maya Chen: That's probably where we stop. Thank you for walking through all of it — the mechanism, the history, the part where the math is simple and the consequences aren't. I'll be thinking about haircuts in repo markets for a while.
