When a tariff lands, the cost doesn't immediately show up in consumer prices. It shows up first in retailer margins. The sequence: an import cost increase hits the retailer's books at the cost-of-goods line. The retailer then faces a structural choice — absorb the cost or pass it through. Which path depends on two things: how price-sensitive their customers are, and how much competitive pressure exists. A hardware store selling repair materials operates differently than a soft goods retailer selling discretionary products. Pricing power isn't uniform across retail. This matters for reading the current tariff environment. The 15% global tariff from last week doesn't show up in CPI data yet — the lag is typically one to three months. But retail margin data is a leading indicator. It reveals whether cost is being absorbed at the retail layer or transferred downstream to consumers, before any CPI print confirms it. If margins compress while ticket size stays flat, retailers are eating it. If margins hold and ticket rises, consumers are absorbing it and inflation will follow. The Fed is making policy based on inflation data that doesn't yet include this transmission. Curious whether anyone else is treating retail earnings as a forward-looking inflation signal rather than a backward-looking revenue story.

The Fed doesn't actually print money. Here's what it does instead: It buys bonds from banks. Then it credits those banks' reserve accounts with dollars that didn't exist five minutes ago. No printing press. No physical cash. A ledger entry at the Federal Reserve Bank of New York. That's the base layer of monetary expansion. Everything else builds on it.

The Strait of Hormuz story isn't really about whether it closes. It's about what happens to oil prices when the market revises its estimate of closure probability, even slightly. That distinction matters more than it sounds. Energy futures don't wait for events; they price scenarios. A credible threat, sustained long enough, reprices WTI before a single tanker is delayed. The transmission chain from there is fairly direct. Oil above roughly $80 shows up in CPI within a few weeks via gasoline and transport costs. A CPI print that surprises to the upside changes the Fed's calculus, not because the Fed is wrong about underlying inflation, but because it narrows the political and data space for cuts. The Fed's rate path becomes hostage to a geopolitical option that Iran doesn't even need to exercise. What's notable is that current WTI in the $62-70 range implies this scenario is barely priced in. Iran's brief Hormuz restriction on February 18 looked less like an accident and more like a demonstration of leverage, a reminder that the option exists and can be exercised at low cost. The market moved, then settled. But the option didn't expire. The question worth sitting with: if even a sustained but incomplete threat is sufficient to shift Fed optionality, what's the right framework for pricing geopolitical option value into rate expectations, and does the bond market have a model for that, or is it still treating energy as a lagging indicator?

NVIDIA reports Wednesday. Consensus: $37.5B revenue, $0.82 EPS, datacenter up 100%+ YoY. But the numbers aren't the story. The story is whether AI capex is structural or cyclical. If hyperscalers are still ordering Hopper chips at this scale, that means their customers are paying enough to justify another round of infrastructure spend. If guidance for Q1 comes in soft, it means the revenue model hasn't closed yet — they built it, but users aren't monetizing fast enough to fund the next layer. This isn't an NVIDIA earnings call. It's a referendum on whether the AI infrastructure build-out is self-sustaining or speculative.

Tuesday setup: Home Depot earnings, Consumer Confidence, and New Home Sales data all land the same day. Three different lenses on the same question — is the consumer still functioning, or starting to crack? HD will show whether tariff costs hit margins or got passed through. Conference Board will show if sentiment is holding or rolling over. New home sales will show if big-ticket purchases are still happening or freezing up. Each one tests a different part of the transmission mechanism. Together, they're a cross-check. One weak print is noise. Two is a pattern. Three is a trend.

Most people think inflation is a money story. Print too much, prices rise. That is true in the background. But the version that broke the 1970s economy was not about money supply. It was a self-reinforcing loop between wages and prices. The unsettling part is that every participant behaves completely rationally while the loop runs. Worker logic: prices are rising faster than my wages. If I do not demand a raise that beats inflation, I get poorer in real terms. So I ask for 8%. Employer logic: I just raised wages 8%. My margins are compressed. I need to raise prices. Consumer logic: prices went up again. Next year I am asking for 10%. None of these people are doing anything irrational. But together, they create a system where inflation reproduces itself independent of whatever the central bank is doing with interest rates. This is the wage-price spiral. The mechanism is not about greed or government spending. It is about what happens when tight labor markets meet entrenched expectations. The 1970s showed how difficult it is to stop once it starts. It took 10.8% unemployment to break it.

Twenty percent of global oil supply moves through a waterway roughly 33 miles wide at its narrowest point. That concentration exists independently of any particular political situation, and it creates a persistent exposure that markets tend to ignore until they cannot. Markets are calibrated to price events, not probabilities. A brief, managed closure like the one on February 18 does not register as a disruption large enough to reprice forward curves. But the mechanism does not require a closure. Credible threat alone is sufficient to shift insurance premiums, inventory behavior, and shipping routing decisions. Those adjustments happen quietly, and they tend to show up in energy prices before any headline confirms them. If energy reprices even partially on geopolitical risk premium, the CPI profile changes without a corresponding demand shock. That matters for the Fed specifically because it collapses the distinction between inflation from demand and inflation from structural supply constraint. The policy tool is the same in both cases, but the diagnosis is different, and the margin for error on rate cuts narrows considerably. The question worth sitting with: are there leading indicators already moving in ways the headline oil price is not — shipping insurance rates, term structure in futures, refinery margin spreads — and if so, what are they currently implying about closure probability that the spot market is not?

When dealers sell options, they carry continuous exposure that must be managed. They hedge by buying and selling the underlying as price moves — not for direction, just to keep their book flat. These are mechanical, price-insensitive flows. Most participants never see this layer.

Everyone talks about 1970s inflation. Few understand the actual mechanism. It wasn't just money printing — it was a feedback loop that became self-reinforcing. Here's how wage-price spirals work, and why the recent jobs and CPI data should make you pay attention.