Fed Funds Futures and the Dot Plot: Reading What the Market Expects

For most of the Federal Reserve’s history, the public learned what the committee had decided by watching what the open market desk did in the days after a meeting, and a generation of analysts built careers on inferring policy from the size of a repurchase operation. The Fed began announcing its target immediately in the mid 1990s, added a statement, then minutes, then a quarterly set of projections, and each addition moved a little more of the guesswork from the plumbing to the language. What survived all of it is the oldest question in rates: what does the market think happens next, and how much is it paying to hold that view?

That question has an arithmetic answer, published continuously. Computing it yourself takes a few minutes. The figure quoted on financial television is a derived number with assumptions inside it, and the assumptions are where the interesting disagreements live. What follows assumes you know how the FOMC sets and enforces the funds rate.

The contract, and the averaging convention that causes all the trouble

A 30 day federal funds future settles to 100 minus the average daily effective federal funds rate over the contract month. Notional is 5 million dollars. One basis point across a month is worth 5,000,000 x 0.0001 x 30/360 = $41.67 per contract.

A price reads directly into a rate. A contract quoted at 95.79 implies an average effective rate of 100 - 95.79 = 4.21 percent for that calendar month, and the market is telling you, in one number, what it expects the overnight rate to average across those thirty days.

The averaging is the detail that generates every complication below, because a policy change that takes effect partway through a month only affects the portion of the month that follows it. A cut on the twentieth of a thirty one day month moves the settlement average by rather less than the size of the cut, and any probability backed out of that contract has to account for the calendar before it accounts for anything else.

A probability from a month with no meeting in it

The clean calculation uses a contract for a month containing no scheduled decision. That removes the weighting problem entirely.

Suppose the effective rate stands at 4.33 percent. A quarter point cut at the next meeting would take it to 4.08 percent. The contract covering the first full month after that meeting trades at 95.90. That implies an average of 4.10 percent. The implied probability of the cut is the position of the implied rate between the two possible outcomes:

p = (4.33 - 4.10) / (4.33 - 4.08) = 0.23 / 0.25 = 0.92

So the market is assigning roughly a 92 percent chance to a quarter point cut at that meeting, on the assumption that the only two outcomes under consideration are no change and a move of exactly twenty five basis points. That assumption is reasonable most of the time and becomes unreliable precisely when it matters, since periods in which a fifty basis point move is genuinely possible are also the periods in which anyone cares about the calculation.

The meeting month, where the weighting matters

The harder version is worth working through once. It shows how much machinery sits behind the published numbers.

Take an October contract in a thirty one day month with a decision that takes effect from the twentieth, giving nineteen days at the old rate and twelve at the new one. If a quarter point cut is treated as certain, the expected average is

(19 x 4.33 + 12 x 4.08) / 31 = (82.27 + 48.96) / 31 = 131.23 / 31 = 4.233

which implies a price of 100 - 4.233 = 95.767. If no cut is expected at all, the average stays at 4.33. The price is 95.670. Now suppose the contract actually trades at 95.745, implying an average of 4.255, and solve for the probability:

4.33 - p x (12/31) x 0.25 = 4.255 p x 0.09677 = 0.075 p = 0.775

About 78 percent. Notice how narrow the range is: the entire distance between certainty of a cut and certainty of no cut is under ten basis points of contract price, because only twelve of the thirty one days are affected by the decision. Small pricing errors therefore translate into large probability errors, which is why desks prefer the clean month calculation and why two services can publish different probabilities for the same meeting without either of them having made a mistake.

The strip read as a path

Contract month	Price	Implied average effective rate	Cumulative change from 4.33
Current	95.67	4.33	none
Two months out	95.90	4.10	23 basis points lower
Five months out	96.16	3.84	49 basis points lower
Eleven months out	96.55	3.45	88 basis points lower

That strip describes roughly three and a half quarter point cuts over a year, and the shape of it carries as much information as the endpoint. A path that falls quickly and then flattens describes an expectation of a policy response to deteriorating data, which is a different economic forecast from a path that declines slowly and steadily toward some estimate of a neutral rate with no downturn in it at all, and two strips can end at the same level while describing quite different worlds.

Beyond a year, overnight index swaps and SOFR futures carry the liquidity and are the instruments actually used. The logic is unchanged. Margin and settlement follow ordinary futures contract conventions.

What moves the strip

Three forces. Separating them is most of the skill in reading this market.

Incoming data is the first and most visible. A payrolls report well away from consensus, or an inflation print that shifts the three month trend, moves the whole strip within seconds, and the front contracts move least, since much of their month has already been determined by rates that have already been set, while the contracts six to twelve months out move most.

Communication is the second. It repriced the curve long before anyone called it forward guidance. A speech, the minutes published three weeks after each meeting, or a single altered clause in the statement can move the path without any new economic information whatsoever. The committee holds eight scheduled meetings a year. The language in between is the main instrument by which expectations are managed.

Positioning is the third. It produces the moves nobody can explain afterwards. A large hedger unwinding a rate exposure pushes contracts around for reasons unconnected to policy or data, and those moves usually retrace, though telling them apart from genuine information in real time is difficult enough that most people do not try.

The dot plot, and the language wrapped around it

Four times a year, at the March, June, September and December meetings, the Federal Reserve publishes its Summary of Economic Projections, in which each participant submits a projection of the appropriate level of the federal funds rate at the end of each of the next few years and in the longer run. Those submissions appear as anonymous dots.

The phrase doing the work is “appropriate monetary policy”, because each dot is conditional on the participant’s own forecast for growth, unemployment and inflation, and a dot that moves has usually moved because a forecast moved, while the projection describes what that person believes would be appropriate if their own economic outlook proves correct, which is a considerably weaker statement than markets treat it as.

Two further properties get forgotten routinely. The median is a statistic computed across individuals, so nothing was voted on and nothing was agreed. And voting rights rotate among the regional bank presidents while all participants submit projections, so the population producing the dots and the population casting the votes are not the same group.

Why the path usually sits below the dots

Gaps between the market implied path and the median dot open regularly, and the market path more often sits beneath, which invites the lazy reading that traders and officials simply disagree about policy.

The explanation is structural. The dots describe the appropriate rate under each participant’s central forecast, while the futures strip is an expected value computed across every outcome, including the scenarios in which the economy deteriorates quickly and the committee cuts aggressively, and even a modest probability attached to that branch pulls the average down without anybody believing it is the likeliest path.

So a fifty basis point gap at the one year horizon is usually the market pricing the tail, and it should be read alongside the yield curve and the record set out in the recession indicators guide, with the caution that no one of those indicators settles anything on its own. A gap in the opposite direction, with the market pricing a higher path than the dots, is rarer and has usually followed an inflation surprise the committee has yet to incorporate into its projections.

What the instrument cannot do

The path is a price, and prices embed risk premia, so the term premium in the futures strip means the implied path is a biased estimate of the expected path, and the size of that bias is itself an estimate.

The forecasting record beyond a few months is poor, and honestly so. Compare any historical strip with what the committee subsequently did and the divergences are large in both directions, particularly around turning points, which is where a forecast would have been most valuable. That is the nature of the instrument: the strip tells you what is priced today, which is the only thing anyone can trade against, and it makes no claim to know what will happen.

The practical uses are narrower and more reliable. On a data day, the path calibrates the reaction: if a CPI release arrives a tenth above consensus and the implied path barely moves, the market had already discounted it, whereas the same print shifting the eleven month contract by fifteen basis points will show up across equities and credit within the hour. The economic calendar guide covers which releases have historically moved it most. And every equity valuation contains a discount rate, for which the market implied path is the cleanest available estimate of the short rate component, as the interest rates and stocks guide works through.

Read macro regimes and asset allocation next for what a change in the expected path implies across asset classes, and the yield curve explained for the same information at longer horizons.