The 30-second read: the FOMC gives you two things that move rates: the statement (what they did and how they framed it) and, four times a year, the SEP with the dot plot (where each official thinks policy is headed). Read the statement by diffing it word-for-word against the last one: every changed phrase is deliberate. Read the dot plot off the median dot, not the range. The median is the signal, the tails are noise. Then reconcile both against what was already priced. The trade is almost never the level; it's the surprise versus market pricing, and the gap between the statement and Powell's press conference.
The FOMC statement is a controlled document. It changes by a handful of words per meeting, and each edit is negotiated. Do not read it top to bottom. Pull up the prior statement side by side and diff it. The market reaction lives in the deltas: an upgrade of growth from "moderate" to "solid," a shift on inflation from "remains elevated" to "has eased," the addition or removal of a risk-balance sentence. A single verb change in the characterization of the labor market or inflation is worth more than three paragraphs of boilerplate.
The load-bearing sentences, in order: the economic assessment (first paragraph), the forward-guidance clause ("the Committee anticipates..." or "in considering the extent and timing of additional adjustments..."), and the balance-sheet/runoff paragraph. Watch specifically for the guidance verb tense: moving from "will be appropriate" to "may be appropriate," or dropping a tightening bias entirely, is the pivot signal. When the target range for the fed funds rate itself is the surprise (a hike or cut the market only half-priced), the front end repin is immediate; when the range is as expected, the whole trade is in the language and the dots.
The dot plot is the SEP's chart of appropriate policy: each of the 19 participants (7 Board governors + 12 Reserve Bank presidents, whether or not they vote that year) places one dot for the year-end fed funds level for the current year, the next two or three years, and the longer run. Anonymous, no names attached. The number that matters is the median: the 10th dot when you rank all 19. That's what desks quote as "the dots showed two cuts this year." The mean, the mode, and the full range are decoration; the median is the committee's central tendency and what gets priced.
Read the dot plot as a delta, exactly like the statement. The question is never "how many cuts do the dots show" in isolation. It's "how did the median move versus the last SEP, and how does it sit versus the market?" A median that shifts from three cuts to two is a hawkish revision even if the economy looks fine. Then check the dispersion: a tightly clustered set of dots signals conviction; a bimodal split (a cluster high, a cluster low) tells you the committee is genuinely divided and the median is fragile. One or two participants moving can flip it next quarter. Also read the longer-run dot, the committee's estimate of the neutral rate; a drift higher there lifts the whole term-premium conversation and the back end.
The dot plot is one page of the Summary of Economic Projections. The projection tables around it are what justify the dots. Read them together or you'll misread both. Four rows carry the weight: real GDP growth, the unemployment rate, headline PCE inflation, and core PCE inflation. The Fed forecasts in PCE, not CPI, so anchor your read there and cross-check against the live core PCE and core CPI prints you're already tracking. The classic tell is a stagflationary revision: core PCE nudged up and GDP nudged down in the same SEP means the committee is stuck, and the dots usually can't ease as much as the market wants.
Pair each projection with its dot. If the median dot moves hawkish while the unemployment-rate forecast is revised higher, that's a hawkish-on-inflation, dovish-on-labor tension: the reaction function is prioritizing prices, and the curve should flatten. If they cut the growth forecast and the dots still show fewer cuts, they're telling you the neutral rate is higher than the market assumes. The projections are the "why"; the dots are the "what." Trade the pair, not either alone.
The statement names the voters and records dissents. A unanimous vote on a surprise move is a strong signal; a 9-2 with two dissents in opposite directions tells you the committee is straddling and the next meeting is live. Dissents from Board governors carry more weight than a regional-president dissent because governors are the standing core of the Fed. Track the direction: a hawkish dissent (wanted less easing / more hiking) caps how dovish you read the statement, and vice versa. Because the dots are anonymous you can't map a dissent to a dot, but you can infer the tails: a known hawk dissenting usually owns one of the high dots.
The FOMC almost never surprises on the decision itself; SOFR futures and OIS have the meeting priced to a fraction of a basis point going in. Your edge is the gap between the statement-plus-dots and that pricing. Before the release, know the strip: how many cuts/hikes are implied for this year and next, and where the terminal rate sits. When the dots print, the instant question is whether the median is above or below the market's implied path. Dots above market pricing = hawkish surprise, front end sells off, 2s10s flattens; dots below = dovish, front end rallies, curve steepens. The magnitude of the repricing is proportional to the gap you measured, not to the absolute dot level.
Then hold fire for the press conference. The statement and SEP drop at 2:00pm ET; Powell speaks at 2:30. The single most common intraday reversal in rates is a statement that reads one way and a presser that reframes it: a hawkish dot plot walked back by a dovish "we're not on a preset course," or a benign statement turned hawkish by a live-Q&A admission that cuts are further off. Two separate moves, two separate trades. Watch the first reaction into 2:00, then re-underwrite it against the tone at 2:30 on the live rates board before you size up.
Turn this into a repeatable sequence so you're not reading prose under fire: (1) diff the statement against prior: flag every changed word; (2) on SEP meetings, read the median dot and its move versus last quarter; (3) check dot dispersion and the longer-run neutral dot; (4) scan the four projection rows for a stagflationary or dovish-growth revision; (5) note the vote and any dissents; (6) measure all of it against the pre-meeting SOFR/OIS strip; (7) trade the surprise, then re-underwrite at 2:30. Only four meetings a year carry a SEP. Mark them on the calendar and check the FOMC hub for the schedule; the other four are statement-and-presser only, which shifts more of the signal into the language and Powell's tone.
