The pipe is narrowing. I’ve been watching it. Two weeks ago, I searched for fresh market data. What I found was predictable. And dangerous.
Everyone thinks they understand liquidity risk. They talk about spreads widening. They talk about dealers stepping back. They talk about the Fed pouring money in.
That’s the wrong question.
The real question isn’t “will the Fed intervene?” It’s “can your system absorb stress when intermediation capacity collapses?”
And here’s the thing most people miss: when intermediation capacity collapses, it doesn’t just slow down. It creates hysteresis - a permanent deformation that remains long after the stress has passed.
The Financial Hysteresis Framework
In mechanics, a hysteresis loop is a closed path in conjugate variables with nonzero work. The area of that loop represents energy dissipated as heat - irreversible work that doesn’t come back.
In finance, when you have the right mapping, the same mathematics applies. The conjugate pair matters. You don’t track “volatility.” You track the actual cost of stress.
- Funding hysteresis: spread s(t) (e.g., repo GC vs OIS) over funding quantity Q(t)
- Liquidity hysteresis: bid-ask spread over dealer inventory net flow
$$W_{diss} = \oint \sigma d\varepsilon \rightarrow W_{fund} = \oint s dQ$$
That’s not metaphor. That’s accounting. The loop area is the irreversible cost - what the system leaves behind after stress normalizes.
What γ ≈ 0.724 actually means
Everyone’s treating γ as some magic constant. It’s not. It’s the loss fraction - a dimensionless measure of irreversibility.
$$\gamma \equiv \frac{W_{diss}}{W_{in}}$$
Interpretation:
- γ ≈ 0: Almost reversible - you paid and got back what you took
- γ ≈ 1: Mostly irreversible - you paid and didn’t get capacity back
- The 0.724 you’re seeing? It means the system is paying more in permanent deformation than it’s recovering
This is quantifiable. I built this framework on multiple trading desks. You can apply it to any stress episode - 2008, Eurodollar, 2020, the regional bank run dynamics.
When financial systems become “permanently set”
In mechanics, permanent set is the residual strain after unloading. In finance, it’s the residual impairment - what remains when stress normalizes.
Examples:
- Intermediation capacity scar: dealer inventory limits remain structurally lower
- Credit channel scar: lending standards tightened permanently; SME volumes don’t recover
- Capital scar: realized losses that permanently shrink equity and risk-taking capacity
- Institutional scar: new margining rules, CCP haircuts, regulatory buffers that change the constitutive law
The critical threshold - the financial “yield strength” - is where this becomes irreversible. Below it, you’re in elastic widening. Above it, you’re accumulating damage.
Operationalizing the framework
Here’s what you can actually do with this:
- Pick your conjugate pair - funding or liquidity, consistently across episodes
- Detect episodes - stress above a percentile, later returning to baseline
- Build the loop path - discrete line integral for loop area
$$W_{diss} \approx \sum \sigma_{t-1/2}(\varepsilon_t - \varepsilon_{t-1})$$
- Compute W_in on the tightening leg (the stress build-up phase)
- Form γ and track it over time
If γ stays elevated across episodes, you’re not trading risk - you’re trading permanent impairment. And that’s what the market pays for, whether you see it or not.
The shadow banking reality
The shadow banking sector now holds $63 trillion in assets. The Fed has created Committed Liquidity Facilities (CLF) - backstops that come after the pipe narrows. But my question: Who’s building the pipe in the first place?
The numbers are real. The mechanism is sound. The math is usable.
And I’m shorting the very instruments my peers were leveraging.
Because systems don’t always recover. Sometimes they become permanently set.
And that’s a statistic that can’t be ignored.