The Strait and the Signal

I. The Strait Closes

Twenty percent of the world's oil moves through the Strait of Hormuz. So does one-third of all global fertilizer trade. The Strait is 21 miles wide at its narrowest point. It is controlled on one side by Oman and on the other by Iran. Iran has been threatening to close it for forty years. On March 2, 2026 — 72 hours after American strikes on Iran-linked facilities in Yemen and the Gulf — they did it.

Not with mines. Not with a formal naval blockade. With the credible threat of both, combined with missile batteries on the Iranian shore that can strike any surface vessel in the Strait in minutes. Insurance underwriters immediately suspended coverage for vessels transiting the area. Without coverage, no commercial ship will move. The result is functionally identical to a physical closure. One hundred and fifty-three ships anchored outside. Twenty-one tankers transited the entire period from March 2 to March 19. Before the conflict, more than a hundred transited every single day.

This is not a hypothetical risk scenario from a war game. This is not a think tank exercise. This is not a 2019 Brookings report about vulnerabilities that could theoretically emerge if conditions were to deteriorate in a scenario where. This is the actual tanker count, from actual transponder data, from March 2026. The receipt printed itself.

The feedback loop that follows is not complicated. It requires only arithmetic:

War → Strait closure → oil supply shock → Brent crosses $100 → gas crosses $4 → inflation pressure compounds → consumer confidence collapses → job losses accelerate → recession probability rises → political vulnerability increases → escalation as distraction → more war → Strait stays closed

Donella Meadows called this a reinforcing feedback loop — the most dangerous structure in systems dynamics, because it has no natural damping mechanism. The loop accelerates itself. There is no negative feedback signal that says slow down. The only intervention points are at the origin: end the war, or open the Strait, or find alternative supply sufficient to replace 20 percent of global oil flow in weeks. None of these things were happening on March 2. None of them were happening on March 19. None of them are happening on April 2.

The fertilizer dimension is the one nobody is covering. One-third of global fertilizer trade goes through this chokepoint — and the Strait closed at the start of the northern hemisphere planting window. Nitrogen fertilizer applied in March appears in grain prices in October. The feedback loop has a delayed branch that won't be visible in the data until autumn. Fertilizer prices were already up 40 percent since the conflict began. The oil-food-dollar feedback loop had engaged; the invoice just hadn't arrived yet.

II. The Price Signal ($100 → $166)

Brent crude crossed $100 per barrel on March 8 — exactly eight days after the Strait closed. This is not a delayed reaction. Oil markets process information in real time. The 8-day lag represents the time it took for traders to verify that the closure was not temporary, that insurance was actually suspended, that diplomatic channels were not producing a resolution, that the 153 anchored ships were not moving. The market was not panicking. It was updating.

On March 9, the president announced his intention to "seize control" of the Strait. The oil price did not fall. It rose another $4 that day.

On March 23, the administration announced a "5-day hold" on strikes against Iranian energy infrastructure. Brent dropped 11 percent in a single session — the largest one-day reversal of the crisis. The market had updated: if strikes paused, the closure might pause. The 5-day hold ended. Strikes resumed. Prices rebounded. The market had been given a signal, tested it against evidence, and returned to its prior position. This is what Bayesian updating looks like in a liquid market: not panic, not narrative — just posterior probabilities being revised and then revised again as new evidence arrived.

This is worth sitting with. The president announced an intention to solve the problem — and the market moved in the direction that suggests it did not believe him. The market is not partisan. It does not care about Truth Social posts. It processes probability. And on March 9, the probability-weighted assessment of traders with billions of dollars at risk was: this announcement does not change the expected timeline of Strait closure. The posterior probability that the Strait reopens this week did not update significantly on the announcement. The price said so.

The US military campaign to physically open the Strait began on March 19. On that same day, Dubai crude hit $166 per barrel — an all-time record. Not the day before. Not the day after. The day the campaign began.

This requires explanation, because it seems backwards. You would expect prices to fall on news of a military campaign to solve the problem. They didn't. The reason is that the market was not pricing in the closure as a solvable diplomatic or economic problem. It was pricing in a new scenario: active military engagement between the United States and Iran in one of the most strategically sensitive waterways on earth, with unresolved duration, unknown escalation risk, and zero track record of quick resolution in the region's history. The military campaign did not lower the probability of extended closure. In the market's assessment, it raised it.

Niklas Luhmann spent his career describing how differentiated social systems process their own codes. The political system's code is power: who has it, who doesn't, how it's accumulated and deployed. The economic system's code is payment: yes or no, profit or loss, credit or default. In theory, what makes functional differentiation sturdy is that the two codes are independent. The political system cannot make $166 per barrel become $60 per barrel by announcing that it intends to. The economic system processes price signals, not political claims.

But this series must be honest about a complication that undermines that clean theoretical distinction. The market's independence depends on one thing: that the information moving prices is available to all participants simultaneously. When the person controlling the supply disruption — the war, the Strait closure, the pause on strikes — is also the person who announces those decisions on Truth Social, the information asymmetry is structural, not incidental.

The Luhmann framework remains useful — the economic system's aggregate price signal cannot be overridden by political announcement, which is why "greatest economy ever" doesn't move the BLS data. But the micro-structure of that market has been compromised. When a single actor controls both the geopolitical events that move prices and the timing of public disclosure, the market becomes a mechanism for extracting rent from information asymmetry, not an independent assessor of risk. The price signal is still real. But the question of who profits from the signal's movement is now a political question, not an economic one. Luhmann's codes haven't merged — they've been weaponized against each other.

The president called it the greatest economy ever. Brent crude settled at $147 at the end of the day.

III. The Jobs Report and the Confidence Collapse

The February Bureau of Labor Statistics employment report, released in early March, showed a net loss of 92,000 jobs. Not a slowdown in job creation. Not a miss on estimates. A loss. A negative number. More Americans were employed at the end of January than at the end of February.

The last time the United States posted a net monthly job loss was during the COVID-19 recovery period. This was not a pandemic. This was a policy environment.

Unemployment rose to 4.4 percent. Manufacturing productivity fell 2.5 percent in the fourth quarter of 2025. Labor costs rose 6.3 percent over the same period — a combination that compresses margins and accelerates layoffs. The tariff disruption to supply chains had begun cutting into industrial orders in January and February. The government contractor freeze, which was still in effect, was suppressing federal-adjacent employment. The oil shock had not yet fully fed into consumer behavior — that would come in March and April — but the anticipatory signal was already in the confidence data.

The University of Michigan Consumer Sentiment Index came in at 53.3. For reference: during the 2022 inflation shock, consumer sentiment fell to 50.0 — the lowest reading since the early 1980s stagflation crisis. The March 2026 reading is 53.3. You are three points above the worst reading in 45 years.

The Conference Board Expectations Index — the forward-looking component that asks where people think the economy is going, rather than where it is — came in at 72.0. The threshold that has historically preceded every post-World War II recession is 80. You are 8 points below it. JP Morgan raised their recession probability to 35 percent, up from a "soft landing" baseline in late 2025. Goldman Sachs put theirs at 30 percent. EY-Parthenon came in at 40. Three major forecasting institutions, none of them Democratic party organs, converging on roughly the same range.

Mark Blyth has argued for decades that the most durable form of political power is narrative — the ability to define what the economy means before anyone processes what it is. "The greatest economy ever" is not a factual claim. It is a positioning operation. It occupies the interpretive space before the BLS data arrives. It frames the loss of 92,000 jobs as aberrant, temporary, the fault of someone else — before the audience has had time to sit with the actual numbers. By the time the actual numbers arrive, the cognitive slot is already filled. The numbers have to compete for attention against a narrative that got there first.

The technique works — partially, and for a time — because media cycles are short and narrative is sticky. The Washington Post runs a story about the jobs report. It also runs a story about what the president said about the jobs report. The two stories compete for the same reader attention, and the one with the more vivid language tends to win. "92,000 jobs lost" is a number. "Greatest economy ever" is a feeling. In the competition between number and feeling, feeling has structural advantages.

Blyth's point is not that the narrative is cynical — though it may be. His point is that narrative functions as a governing mechanism. The narrative is doing work. It is suppressing the political feedback that a 92,000-job loss would otherwise generate. And as long as it suppresses that feedback, the system has no corrective pressure to change course.

IV. The Selective Pay

On March 26 and 27 — Day 40 and 41 of the DHS shutdown — the president signed an executive order providing back pay to TSA workers. The mechanism was extraordinary-measure funding drawn from the "big beautiful bill" that was still working its way through congressional reconciliation. This was not a proper congressional appropriation. It was a discretionary executive redirection of funds from a separate legislative vehicle. Constitutional scholars have questioned whether it is legal. But it worked, politically, in the sense that TSA workers received direct deposits and airport lines shortened.

On April 2, Day 47, the Coast Guard remains unpaid. FEMA disaster response personnel remain unpaid.

Let's be precise about what these two institutions actually do. The Coast Guard is currently conducting interdiction operations in the approaches to the Persian Gulf — the waterways adjacent to the Strait of Hormuz that the US military campaign is attempting to secure. The Coast Guard cutters operating in that theater are crewed by personnel who have not received a paycheck since February 14. FEMA manages disaster response for a country that experienced thirty-seven declared major disasters in the first quarter of 2026. Its incident commanders and logistics coordinators are also unpaid.

TSA workers operate in airport security lines that are visible to voters.

The political logic here is transparent. Long TSA lines generate news coverage. Travelers at Reagan National with connecting flights to miss are voters who will tell their neighbors, who will call their representatives. The loop from "TSA line is two hours" to "congressman hears about it" takes about three days. The president was getting calls. He paid the TSA.

The loop from "Coast Guard cutter crew is unpaid" to "voter hears about it" requires: (a) the voter to know where Coast Guard cutters are currently operating, (b) the voter to know that the crew is unpaid, (c) the voter to connect that fact to a political pressure point. None of these steps are automatic. The Coast Guard's operational status is classified in significant portions. The geographic specifics of their Gulf theater operations are not on the evening news. The connection between unpaid Coast Guard personnel and a compromised Strait-of-Hormuz operation is several inferential steps from the consumer's lived experience of a two-hour security line.

Luhmann's concept of functional differentiation predicts exactly this outcome, and not through any cynical model of voter manipulation. The political system is not making a deliberate choice to sacrifice national security for airport optics. It is doing something structurally more interesting: it is processing through its own code — power, approval, electoral viability — and that code only registers inputs that affect power. TSA lines affect power. Unpaid Coast Guard cutters in the Gulf do not register in the political system's code, because their impact is routed through the security system, not the political system. The political code is structurally blind to what it cannot see in its own operational environment.

This is the single most compact illustration of de-differentiation in the entire series. De-differentiation is not the collapse of all institutions into one. It is the selective blindness that results when all institutions are forced to process through a single code — in this case, loyalty and voter-visibility. What survives that filter: TSA pay. What does not: Coast Guard pay. The criterion is not operational importance. It is political legibility.

The budget decision is the thesis. You don't need to theorize about the erosion of institutional capacity. You can read it in a disbursement record.

V. The Complete Feedback Loop

Here is the chain as it existed on April 2, 2026, traceable from public data:

Feb 28 strikes on Iran-linked facilities → Strait closure (Mar 2) → oil supply shock → Brent $100 (Mar 8) → "seize control" announcement → market disbelief → Dubai $166 (Mar 19) → gas above $4 nationally → consumer confidence collapses to 53.3 → February jobs report: −92,000 → Conference Board Expectations 72.0 → JP Morgan: 35% recession probability → political vulnerability → escalation signaling (to manage domestic narrative) → "hit them hard for two to three more weeks" (Apr 1 speech) → Strait stays closed → loop repeats

Meadows' most quoted line on systems dynamics: "The most common cause of oscillation and policy resistance is the failure to account for the feedback loop." The oscillation here is between two failure modes. Escalate, and the Strait stays closed and oil stays high. Don't escalate, and the domestic political narrative of decisive leadership collapses. There is no third option being exercised. The feedback loop has the system in its grip.

The exit points are structural, not communicative. They are: (a) end the conflict, (b) open the Strait through negotiation rather than force, (c) find alternative oil supply sufficient to replace 20 percent of global flow quickly enough to prevent the confidence collapse from becoming a recession. None of these exit points require a narrative shift. They require a different set of choices. No narrative shift — however vivid, however loudly amplified — changes the price of Brent crude or restores 92,000 jobs or moves 150 anchored tankers into motion.

The market knows this. This is what "$166 on the day we launched the campaign to fix it" means. Not pessimism. Not partisan analysis. A Bayesian update. The market processed the prior evidence — 40 years of Strait-of-Hormuz risk scenarios, every regional conflict in the Gulf since 1980, the historical record of military campaigns that rapidly resolve oil supply crises (there are none) — and it updated on the new information, which is that the United States is now in active engagement in the same theater where the closure is occurring.

The posterior probability of rapid resolution went down. The price went up. The receipt printed itself.

There is a line from the source transcript for this series — the conversation that initially identified these 47 days as a coherent analytical unit — that has stayed with me through the reporting. The observation was: "Oil didn't spike because of something that happened. It spiked because something wasn't said." The market was not reacting to the strikes. It was reacting to the absence of a credible path to resolution. Every day without a statement that contained actual operational specifics — when, how, what diplomatic channel, what guarantor, what verification mechanism — was itself a signal. The silence was the data. The price was reading the silence.

The president called it the greatest economy ever. Every metric available from every agency that actually measures the economy — BLS, BEA, UMich, Conference Board, JP Morgan, Kpler, Dallas Fed — produced a different assessment. Not a different opinion. A different measurement. The gap between the assessment and the measurement is not a communications problem. It is a structural feature of a system in which the political code has been asked to do work that only the economic code can do.

You cannot manage your way out of $166 crude. You cannot narrative your way out of 92,000 lost jobs. The gauges read what they read. This is the receipt.