|
Cedars Hill Group — March 9, 2026 — Last week opened with the realization of a long-held existential fear by the oil market: the closing of the Strait of Hormuz. Oil futures were up on Sunday night but by Monday the ambiguity of that fear began to settle int... CHG Issue #218: The Week the Map BrokeCedars Hill Group — March 9, 2026 — Last week opened with the realization of a long-held existential fear by the oil market: the closing of the Strait of Hormuz. Oil futures were up on Sunday night but by Monday the ambiguity of that fear began to settle into the market. The waterway itself was not physically closed, but ships could no longer get insurance to traverse the waterway because of the IRGC’s threats. So, Trump being the master of rhetoric that he is said that the straight wasn’t closed, that the Iranian regime was being annihilated, that the US would backstop insurers, and the Navy would escort tankers.
In reality, what happened last week was that the IRGC used a maximalist label to project power by throwing the Western financial markets into disarray, a well-known strategic lever. Tankers paused as insurers responded to the threat and Trump countered with another label to counter that power projection. The military and naval posture on the ground hadn’t materially changed in Iran’s favor, only the verbal posturing. In fact, the frequency of Iranian missile strikes and drone attacks materially slowed over the course of the week. The contrasting kinetic war and war of words gave a strong example of the power of narratives, especially in our “post-truth” world and pluralistic society. We are all too familiar with the cyclical nature of journalism today and how quickly trends and phrases pass through our collective consciousness. The cognitive machinery behind these phenomenon reveals the importance of classification which gives us specificity and comparability. When confronted with new information we want to know specifically what it is and then relate it to something more familiar so we can judge for ourselves the importance of the new thing. But there is a tradeoff between these competing desires for clarity. If you want more of one you must take less of the other. We cannot specifically know what something is but also be able to broadly compare it. So now that the “Strait of Hormuz is closed” was finally realized we need a basis for comparison. Is this as bad as the 2003 Iraq War? We saw the first sinking of an enemy ship by a torpedo since WWII—is this as bad a WWII? Is this like 1973? The specificity of the event leads us grasping for comparisons to base our judgements. The markets leaned towards pricing the risk under the framework from the most recent experiences with conflict. For most of the week the assumption was that this conflict would be short-lived, but by the end of the week the pricing was starting to turn away from that as the reality of war started to set in. Conflicts are not monolithic things that start and end neatly, they often behave like a system with feedback loops, escalation ladders, and distributed actors with varying motives. Wars can start explicitly, through an act of Congress in the US, but they can also evolve from conflicts. It is important to remember that the map is not the territory, in other words, the map (our view of war) is merely an abstraction of the reality on the ground. Lawmakers spent their week debating the definition of “war” with the Democrats trying to catch Trump breaking the law and Republicans mindlessly coming to the defense of their leader. This was just another manifestation of the same epistemic error that the markets were making. The labels are outdated but the system still relies on them. Congress spent the week trying to score political points instead of working to resolve the conflict. The spike in oil prices and the employment report on Friday also brought back the debate around monetary policy and inflation. The system that defines these things is hard coded with phrases like the Philips Curve, NAIRU, The Sahm Rule, words like “transitory”, and so on that are simply abstractions of a very complex reality. Economists, strategists, and policymakers are straining to fit the current reality into these fixed and outdated containers to make their point and validate their theories without recognizing that possibility that the old containers don’t fit today’s reality. This extends to the craze around AI. Since the original release of the LLMs we have seen new labels like “frontier models” and “agentic AI” rise and help people see the application of this new technology but also feed a mania in the markets around it. The large tech companies see AI as the holy grail and are willing to spend any amount of money to be the leader in it. However, amidst this mania we risk mistaking the capability for the category and losing sight of reality on the ground. This tension has also surfaced in the push to “democratize alternatives”, which is another catch phrase that has captured the minds of investors. Alts once meant hedge funds and now they mean private equity, private credit, real estate, infrastructure, farmland, venture capital, and so on. There are more alts today than publicly traded companies, which has necessarily changed the meaning of alternative. We create these categories to aid our understanding, but the categories need to be updated, and they often outlive their usefulness. We tend to get stuck in the status quo—we like the certainty and familiarity. We’ve been using the “New Framework” developed in It’s Always Different This Time to specify the big shifts taking place today and to put them in light of history (comparability). We’ve suggested that the 40-year secular bond bull market is over and that means we are entering a period where interest rates will not fall much and may rise significantly. This has all sorts of downstream effects that we can theorize about but can’t truly know until they happen. We posited that one of these effects would be a reversal of the outperformance of financial assets over physical assets and we’ve received new information that supports that thesis, strangely enough from one of the AI companies. Anthropic released a study last week titled: Labor market impacts of AI: A new measure and early evidence. CNBC was reporting on Friday about this and they characterized the study findings as finding that jobs in the “knowledge economy” were experiencing more displacement from AI than jobs in the physical economy. New labels for a changing future. Paradoxically, it is language itself that holds the promise of freeing us from the constraints of language. We use language to construct abstractions of reality to increase fluency, relatability, and comparability but lose specificity in the process (ex. the term “football” has different meanings depending on your geography). It takes too long to describe something specifically, at the most atomic level, and doing so makes comparability very difficult so we find something that is “good enough” but is fundamentally fragile. Cognitive dissonance, the state of tension that occurs when individuals hold two psychologically inconsistent cognitions (ex. smoking a cigarette while knowing it is bad for you) makes it very difficult to accurately update our prior beliefs and often leads to new consonant, but not necessarily accurate cognitions (ex. smoking helps to lower my anxiety). Yet this is where LLMs offer a solution because they can do all the heavy lifting of language for us which can reduce or even eliminate our need for abstractions. This is what we meant last week when talking about SaaS: ...building durable architecture based on domain primitives that can endure across rapidly changing trends. LLMs enable this by making it possible to encode the language of business within systems which allow customers to escape from the one-size-fits-all SaaS model and get access to durable solutions at scale. Previously, systems forced the business into fixed and arbitrary boxes which needed to be changed every time the business changed. Understanding the power of the LLMs only requires understanding the power of language. We’ve been using language as an abstraction of reality since the first grunt was heard. What we are doing today is far more complex, but not that different. The ability to process language at scale enables us to use far more specific abstractions to describe and compare reality broadly which holds the promise of far more durable systems in the future. Source: https://cedarshillgroup.substack.com/p/chg-issue-218-the-week-the-map-broke?utm_source=post-email-title&publication_id=1104657&post_id=190426898&utm_campaign=email-post-title&isFreemail=true&r=82apz&triedRedirect=true&_src_ref=url.emailprotection.link&utm_medium=email Comments are closed.
|
About UsPickwick serves exceptional companies and leading investment funds across the US, offering investment banking expertise and fund placement advisory to lower middle-market clients. We tightly couple our strategic advisory services with our financial and transaction services and believe that this approach is highly differentiated in the financial advisory marketplace. |
|
