Last Wednesday the July inflation report landed, and within about ninety seconds the September rate hike probability dropped to 42 percent.
Almost every headline translated that the same way. The Fed probably won't hike. Odds are against it. The market has spoken.
Every one of those readings is a misuse of the number. And this afternoon, when the Federal Reserve releases the minutes of its July meeting at two o'clock Eastern, you're going to see the same probability quoted again, revised, and misread all over again. So let's fix what that number actually is, because getting this right changes how you build a portfolio.
Start with where it comes from. That 42 percent isn't a survey. Nobody polled economists. It's derived from fed funds futures contracts, which are instruments that settle based on the average effective federal funds rate over a given month. Traders buy and sell them. The price of the contract implies an expected average rate for that month. Compare that implied rate to the current target range, and you can back out how much of a 25 basis point move the market has priced.
So when you see 42 percent, what's really being said is this: the fed funds futures market is currently priced at a level consistent with roughly 42 percent of a quarter point hike being delivered in September. Right now the broader rates market is carrying about 9 basis points of tightening for September and roughly 23 basis points across the rest of the year.
That is a price. It is not a prediction.
The difference matters more than it sounds. A prediction is somebody's best guess about what will happen. A price is where buyers and sellers clear, and clearing prices contain things that have nothing to do with what anyone expects.
Consider who's actually in that market. A large share of the volume comes from institutions hedging, not speculating. A regional bank with a book of floating rate loans isn't expressing a view when it buys protection against higher rates. It's buying insurance. It will pay above fair value for that insurance because the alternative is a hole in its balance sheet. That hedging demand pushes the implied probability around in ways that have nothing to do with anyone's forecast.
Then there's the asymmetry problem. Suppose a hike would cost the market a great deal of pain and a hold would be a non-event. Even if the true likelihood of a hike is 30 percent, participants will pay up to be protected against it, and the implied probability drifts above the true one. Market-implied probabilities are risk-weighted, and risk weighting is not the same thing as probability weighting.
And then there's the part nobody wants to say out loud. Futures markets are excellent at pricing the near term and consistently mediocre at anything beyond a few months. Go back and look at what fed funds futures were pricing for the end of this year, back in January. Then compare it to where we are. The curve has been repriced repeatedly, in both directions, by data nobody had yet.
This year has been an unusually clean demonstration of that. In the space of about seven months the rates market has gone from pricing a continued easing path, to pricing nothing at all, to pricing hikes. Energy shocks reset the inflation picture. A negative monthly CPI print in June pulled expectations one direction, then a positive July print with wages running below inflation pulled them back. None of that was in any curve at the start of the year, because none of it had happened yet. The futures market didn't fail. It simply did what it always does, which is price the information available at that moment and reprice the second the information changes. Treating a repricing machine as a forecasting machine is the error.
None of this means the number is useless. It's extremely useful, as long as you use it for what it is: a real-time readout of what the market has already absorbed. That's its actual job. If the probability sits at 42 and a hike arrives, the surprise gets priced violently, because more than half the market wasn't positioned for it. If it sits at 90 and a hike arrives, almost nothing happens, because everyone already owns it.
That's the insight worth keeping. The probability doesn't tell you what will happen. It tells you how much the market will move if it does.
Which is exactly why 42 is the most dangerous number on the board. At 90 or at 10, one outcome is dominant and the other is cheap to hedge. At 42, both branches are live and both branches are expensive. Anything you build that only works in one of them is fragile by construction.
That brings us to this afternoon.
FOMC minutes are the most misunderstood document the Fed publishes. People treat them as a summary of a decision. They aren't. The decision was announced three weeks ago and the market moved on within hours. The minutes are the record of an argument.
That's the value. The statement tells you where the committee landed. The minutes tell you how spread out the committee was when it got there. And the distribution of views is a far better leading indicator of the next move than the decision itself, because policy changes when the center of gravity shifts, and the center of gravity shifts inside these documents before it shows up in a vote.
There's a specific skill to reading them, and it's mostly about counting.
The Fed uses a deliberate vocabulary of quantifiers, and the people who trade this stuff treat the words as numbers. "A couple" means two. "A few" means roughly two or three. "Several" means somewhere around three to five. "Many" implies a substantial bloc. "Most" means a clear majority. "All" is unanimity. These aren't loose adjectives. They're calibrated, they're consistent across meetings, and the drafting committee chooses them carefully.
So the exercise isn't reading the minutes. It's diffing them. Pull up the July minutes this afternoon alongside the June set and compare the quantifiers on the same recurring themes. If a sentence about upside inflation risk moved from "a few participants" to "several participants," that's a bloc growing. If a sentence about labor market softening moved the other direction, that's a bloc shrinking. Nothing in the headline coverage will capture that, and it's the entire reason the document exists.
Three other places carry weight. The staff economic projection section, which is written by Board economists rather than the policymakers, and which occasionally disagrees with the committee in ways that get resolved later in the committee's favor. The risk management paragraph, usually near the end, where the committee explains what it's most afraid of getting wrong. And any explanation attached to a dissent, which is the closest thing to a public argument the institution permits.
Read those four things and you'll know more about the September meeting than someone who watched an hour of coverage.
Now, the practical part, because none of this is worth much if it doesn't change what you do.
The mistake I see most often is people building a portfolio around a single expected path. They decide the Fed is going to hike, or hold, or cut, and then they position for it. When you do that at a 42 percent probability, you are essentially making a coin flip bet with your household balance sheet and calling it strategy.
The alternative is to build for both branches, and it's less complicated than it sounds.
The hike branch hurts you in exactly one place: anything you owe at a floating rate. Credit card balances, home equity lines, variable rate business debt, margin. Those repriced immediately and painfully. So the hike branch is handled not by prediction but by elimination. Whatever floating rate debt you can retire or convert to fixed before September, do it now, while the decision is cheap and reversible. That's not a bet on a hike. It's removing your exposure to the question.
The hold branch is where your cash lives. If rates stay where they are, cash keeps paying reasonably well and there's no urgency. If they eventually fall, today's yields on money market funds and short Treasuries were the good old days. So the hold branch is handled by matching duration to obligation rather than to forecast. Money you need in six months stays short. Money you don't need for five years shouldn't be sitting in overnight instruments waiting for clarity that isn't coming.
Notice that neither of those requires you to know what the Fed will do. That's the design goal. A system that produces good decisions under both outcomes beats a forecast that's right sixty percent of the time, because the forty percent is where people get wiped out.
The exercise that makes this concrete takes about twenty minutes. Draw two columns. Label the left one "rates go up 25 basis points" and the right one "rates stay where they are." Then walk every line of your balance sheet through both columns and write down what changes.
Your card balance in the hike column costs you more, immediately, because card APRs are tied to prime and prime moves with the Fed within a billing cycle. Your home equity line does the same. Your money market yield goes up a little in the hike column and stays put in the hold column, which is pleasant in both. Your bond fund loses a bit of price in the hike column and gains income over time. Your fixed rate mortgage does absolutely nothing in either column, which is the entire reason people fixed it.
Now look at the two columns side by side and find the lines where one column is genuinely painful. Those are the only places a decision is required. Everything else is noise you can safely ignore for the next month, and the relief of knowing which is which is worth the twenty minutes by itself.
Most people discover they have exactly one or two exposed lines. Fix those and the September meeting becomes something you read about rather than something you brace for.
The same logic applies to your equity allocation, and this is where mechanics beat conviction. If your target is 70 percent equities and markets have pushed you to 78, you're carrying more risk than you chose, and you're carrying it into a meeting where both outcomes are live. Rebalancing bands solve this without requiring an opinion. Set a threshold, five percentage points is a common choice, and when a position drifts past it, you trim back to target. No forecast required. No meeting to watch.
This is the sort of thing that works far better automated than remembered. M1 Finance is built around this specific behavior: you define target allocations once, and every contribution routes to whatever is underweight, so the rebalancing happens as a byproduct of funding rather than as a decision you have to make while a Fed announcement is scrolling across your screen. Taking the decision out of the moment is most of the value.
At two o'clock this afternoon the minutes will drop and the probability will move. It might go to 50, it might go to 30. Either way, the number you see tomorrow is telling you the same thing it's telling you today: how much room there is for a surprise. That's it. That's the whole message.
Treat it as information about market positioning rather than as a forecast, and you'll stop being whipsawed by every revision.
I built a short reference guide for this. It covers how implied probabilities are derived in plain language, the Fed's quantifier ladder with what each term maps to numerically, the four sections of the minutes worth reading and what to look for in each, and a two-column worksheet for stress testing your own balance sheet against the hike branch and the hold branch. It's the checklist I'd hand someone who wanted to stop reacting to headlines.
Reply to this email with the word ODDS and I'll send you The Probability Translator.
Own the system,
Taylor Voss
Money Systems Lab
DISCLOSURE: This newsletter is educational content and does not constitute financial, investment, tax, or legal advice. Money Systems Lab is not a registered investment advisor. All figures cited are drawn from public sources as of the date of publication and are subject to change. This issue contains affiliate links, which means Money Systems Lab may earn a commission at no additional cost to you if you open an account through them. We only reference tools we would use ourselves. Always do your own research and consult a qualified professional before making financial decisions.