Sitemap

Oil, War, and Algorithms: Why Brent Tells Us the U.S. Is Not Winning (and That an Internal Crisis Is Possible)

5 min readMar 26, 2026

--

Press enter or click to view image in full size

How a statistical model confirms that the Iran‑Israel‑U.S. conflict is not a black swan, but the manifestation of a risk that was already there.

The problem: when headlines lie, markets speak

We are living through an open war between the United States and Iran. Official statements speak of “surgical victories” and “technological superiority.” But if one steps away from the headlines and looks at financial markets, the story changes.

  • Brent Crude Oil remains above $100, a sign that the Strait of Hormuz remains de facto closed.
  • Defense stocks (Lockheed Martin, Northrop Grumman) surged before the war, but have stagnated since the bombings began.
  • Gold, the quintessential safe‑haven asset, has fallen 21% since the start of the conflict.

These contradictory signals point to something that official rhetoric hides: the market does not believe in a quick U.S. victory, and large investors are more concerned about an internal fiscal crisis than about geopolitical escalation.

But there is even stronger evidence. A recent ARMA‑GARCH modeling study (attached to this article) demonstrates that oil volatility during the 2026 war was not unpredictable: it behaved exactly as it had during the 2023‑2025 tensions.

That completely changes the narrative. Let us see why.

2. Quantitative validation: Brent as a “seismograph”

The author of the study asked: can an ARMA‑GARCH model calibrated on 2023‑2025 data (the prior tensions) explain Brent volatility during the open crisis of 2026?

The answer is yes, with compelling figures.

The winning model (reference period 2023‑2025)

  • Mean component: ARMA(1,1)
  • Variance component: eGARCH(2,1)
  • Distribution: Skewed Student‑t (sstd)
  • Volatility persistence: β ≈ 0.95

Results when applied to the 2026 crisis

Metric2023‑2025 period (reference)2026 crisisAIC (best fit)–5.2587–5.1987 (only 1.14% higher)MAE (mean absolute error)0.013600.01401 (2.9% higher)Hit rate (accuracy in 95% intervals)66‑100% depending on subperiod93.1%

Conclusion of the study:

“The 2026 conflict did not create a new risk dynamic; it simply activated the one that already existed, adapting its expression to the nature of the crisis regime.”

Translated: the oil market had already internalized the possibility of open war. When it broke out, volatility simply followed the statistical rules that had been in place since 2023.

This has two immediate implications:

  1. Brent is a reliable indicator of the conflict. There is no need to invent anything: price and volatility tell us whether tensions are escalating or easing.
  2. The “black swan” narrative collapses. If the risk was already modeled, then the war was not a surprise; it was a political choice made on a board whose costs were already discounted.

3. What do the stock market instruments tell us?

If we accept that Brent is the main thermometer, we can build a set of financial assets that help us track the conflict in real time and, in turn, validate the hypothesis that the U.S. may be heading toward an internal crisis.

3.1. Oil (the central thermometer)

  • Brent Crude Oil (BZ=F) / ETF BNO
  • WTI Crude Oil (CL=F) / ETF USO

What it measures: closure of the Strait of Hormuz and geopolitical risk premium.

Get Luis Adrian Martinez Perez’s stories in your inbox

Join Medium for free to get updates from this writer.

What to watch:

  • Price > $110 → severe escalation.
  • Implied volatility (OVX) spiking → extreme uncertainty.
  • If the GARCH model begins to break (as happened with COVID‑19, which caused the hit rate to drop to 33%), that would signal a regime change (fiscal default, ground invasion, or internal crisis in the U.S.).

3.2. Defense (the silent distrust)

  • Lockheed Martin (LMT)
  • Northrop Grumman (NOC)
  • RTX Corp (RTX)

What it reveals: defense stocks rose sharply before the war (anticipating it), but have stagnated since the bombings began.

Interpretation: institutional investors do not believe in a quick victory. If U.S. military technology were so superior, these stocks would be at all‑time highs. They are not.

3.3. Gold (the contradictory signal)

  • Gold spot (XAU/USD) / ETF GLD

What it measures: real financial panic.

What is happening: gold has fallen 21% since the start of the war. This is an anomaly. In theory, it should be soaring as a safe haven. Its decline indicates that large funds are selling gold to cover losses in other assets (a “dash for cash” effect), which reveals liquidity stress, but not yet a systemic collapse.

3.4. Treasury Bonds (the true thermometer of an internal crisis)

  • Treasury 10‑Year (TNX) / ETF IEI

What it measures: confidence in the U.S. government’s ability to finance the war without going bankrupt.

What to watch:

  • If the 10‑year yield exceeds 4.5% on a sustained basis, the fiscal alarm is on.
  • If it exceeds 5%, it is a sign that investors are doubting the solvency of the state. In that scenario, the combination of fiscal crisis + political polarization + demoralized military makes the possibility of an internal crisis a concrete risk.

4. The lesson from COVID‑19: when the regime breaks

The study we are discussing does not only analyze the 2026 war. It also compares the behavior of Brent during the COVID‑19 crisis.

CrisisMAE (error)Hit rateRegime change?COVID‑19 (2020)0.0189533.3%Yes (structural break)Iran‑U.S. War (2026)0.0140193.1%No (regime remains)

Interpretation:

  • COVID‑19 was a true black swan. The volatility structure broke down. Models stopped working.
  • The 2026 war, on the other hand, is a manifestation of the same geopolitical risk attractor that already existed.

The implication for our thesis:
As long as the GARCH model continues to perform as it has so far (low errors, high hit rate), the war remains within the bounds of a “conventional” geopolitical conflict.
If the model begins to break — as it did with COVID‑19 — that will be the signal that the system is entering a regime change: default, failed ground invasion, or internal fragmentation of the United States.

5. Conclusion: a predictable conflict that exposes the cracks in the U.S.

The combination of geopolitical analysis, stock market instruments, and quantitative modeling allows us to draw a picture very different from the official headlines:

  • Brent is not just any price. It is a statistically validated seismograph. Its behavior during the war confirms that the risk was already discounted since 2023.
  • Stagnant defense stocks reveal that the market does not believe in a quick U.S. victory.
  • Treasury bonds are the most sensitive variable: if their yields spiral out of control, we will be facing the prelude to a fiscal crisis that, combined with political polarization and military discontent, would make the scenario of an internal crisis real.

The ARMA‑GARCH study gives us a quantitative tool to distinguish between regime changes (like COVID‑19) and events within the same regime (like the current war). As long as the model continues to work, the conflict is “conventional.” When it begins to fail, we will know that something much deeper is happening.

In short, the Iran‑U.S. war is not a black swan. It is the manifestation of a latent tension that already had its statistical rules. And those rules tell us that, even as missiles continue to fall, the real battle for the stability of the United States is being fought in the debt markets, in the confidence of its troops, and in the legitimacy of its institutions.

This article is based on official statements before the U.S. Senate, reports from non‑Western media, open‑source intelligence analysis, and the quantitative study “Brent Crude Oil Volatility Modeling During the 2026 Iran‑USA‑Israel Crisis [https://github.com/ladrianx/articles/blob/440f70a3de0b9a7bf26cd9f2029395efc1198e1c/BrentCrudeOil.pdf],” whose summary has been incorporated for analytical purposes.

--

--

Luis Adrian Martinez Perez
Luis Adrian Martinez Perez

Written by Luis Adrian Martinez Perez

Applied mathematician (UNAM/UAM). ML, data science, AI. University professor. Asset analysis and quantitative trading.