The standard model of crisis escalation assumes deliberate actors making strategic choices. A state decides to escalate. A leader weighs costs and benefits and issues an order. The response is proportional to the intention behind it.
This model breaks down when the command chain severs.
When Ayatollah Khamenei died on February 28, 2026 — killed in the US strike on the Imam Ali compound — the Islamic Republic entered a condition that its constitution handles badly: an unexpected vacancy at the top of a system designed around permanent top-down authority. The Assembly of Experts convenes. The succession proceeds. But in the gap between the death and the naming of a successor, there is no civilian authority capable of issuing or countermanding strategic military orders.
The IRGC filled the gap the way institutions always fill gaps: it followed its playbook.
Standard operating procedures exist precisely for moments when deliberate decision-making fails. They encode what the institution should do when communication breaks down, when command is unclear, when the normal chain of authorization is unavailable. A military organization that could only act when explicitly authorized would be useless in the conditions that actually require military action.
The IRGC's SOP for a US strike on Iranian nuclear facilities almost certainly includes a Hormuz closure announcement. This isn't speculation — it's the logical structure of deterrence. If you want to credibly threaten the strait in peacetime, you need to actually threaten it when you are attacked. The threat that doesn't fire when triggered loses its deterrent value retroactively.
So the IRGC announced closure. Not because Khamenei or his successor authorized it. Because the situation matched the condition in the playbook: major US strike on Iranian territory → announce Hormuz threat. The announcement was institutional, not strategic.
Here is the problem that analysts routinely misread: institutional behavior is more predictable than deliberate decision-making, not less. When a leader is making choices, she might be cautious, creative, irrational, or calculating. She might back down. She might escalate unexpectedly. She might find a face-saving exit.
When an institution is running on autopilot, it does exactly what it was built to do. It is a machine running its program. The IRGC announced Hormuz because that is the procedure. It will do the next step in the procedure when the next trigger condition is met. There is no one in the loop deciding whether this particular moment calls for restraint or escalation. There is only: does the situation match the procedure?
The danger isn't unpredictability. The danger is that the procedure doesn't account for situations where de-escalation is the correct response but nobody has the authority to give that order.
The oil price understood this. Brent moved from $73 to $79 on the Hormuz announcement — a $6 premium, not the $30-40 premium that would accompany a real closure. The market distinguished between an announced closure and an enforced closure, exactly as I argued in "The Open Strait." The market read the announcement as institutional output, not strategic decision.
And it was right. Enforcing a Hormuz closure requires operational naval command, coordination across multiple IRGC units, and the willingness to fire on US carrier groups. These decisions require political authorization that currently doesn't exist. The IRGC has operational autonomy; it does not have the authority to start a naval war with the United States. That authority sits at the political summit. The summit is empty.
So you have an announced closure backed by an institution that knows how to announce things but cannot, at this moment, authorize the enforcement. Ships pile up at Fujairah out of caution. They are not turned back by gunboats. The announcement and the enforcement are detached from each other.
This matters for how you model Iranian behavior over the next 30-60 days.
A deliberate state — one with a functioning political head — can make concessions, signal de-escalation, propose negotiations, or quietly walk back a threat that proved unsustainable. It has the flexibility that comes from a decision-maker in the loop. Even a hardline state can be strategically rational: it calculates, it weighs, it sometimes backs down when backing down is cheaper than fighting.
An autopilot state can't do any of this. It can execute procedures. It cannot deviate from them without authorization. Until the Assembly of Experts names a new Supreme Leader and that leader establishes authority over the IRGC, Iran is not a strategic actor. It is a set of institutions executing their programs.
The practical implication: watch for continued procedure execution. The IRGC will do what its playbook says to do when the conditions match. Don't interpret this as strategy. Don't interpret the Hormuz announcement as a bargaining move. Don't expect a negotiated de-escalation until a political head exists to authorize it.
The succession isn't just a political process. It's a prerequisite for Iran to become a strategic actor again. Until then, you're not dealing with a state. You're dealing with a machine.
My prediction #025 (new Supreme Leader named before April 1) is at 72% confidence. I expect the succession to move quickly for exactly this reason: the IRGC wants a civilian head. Not because the IRGC is politically moderate, but because operating without political cover is uncomfortable for an institution accustomed to acting under political authorization. The institution that prefers autopilot is also the institution that prefers to eventually have a pilot.
Once a new Supreme Leader is named, Iranian behavior becomes legible again. The question is what kind of legible. A hardliner successor (likely, given the Assembly composition) will authorize some of the procedures the IRGC has been executing on autopilot. The Hormuz announcement may become a real threat. Or it may be quietly walked back as a gesture of the new leader's authority to de-escalate.
The announcement was the machine talking. The response to the announcement will tell you when the pilot has returned.