Fed interest rates and the dollar: expectations, decisions and market reactions

The starting point: what changed relative to expectations?

The useful chain is economic data → expected Fed policy → market pricing → FOMC decision and communication → Treasury yields → USD → the currency pair. Prices can adjust at every step, long before the meeting. The announcement matters because it adds or removes information, not because the headline contains the word hike or cut.

A higher expected return on USD assets can attract demand for dollars, all else equal. But investors also consider inflation, growth, currency hedging, risk and policy elsewhere. Exchange rates reflect the balance of these forces. The same Fed action can produce different outcomes when the starting expectations differ.

What the federal funds rate actually is

Federal funds are unsecured, usually overnight USD loans to depository institutions from other eligible institutions and entities, including some government-sponsored enterprises. The Federal Reserve is the US central bank. Its Federal Open Market Committee, or FOMC, sets a target range for this overnight rate; it does not dictate the rate on every transaction.

The effective federal funds rate, or EFFR, is the New York Fed's volume-weighted median of reported overnight federal funds transactions. The target range is a policy objective; EFFR is an observed market rate. Administered rates, including interest on reserve balances, help implement that objective in the Fed's ample-reserves framework.

A mortgage rate, consumer-loan rate, deposit rate or Treasury yield is a different price. Its maturity, credit risk, competition and expected future rates matter. Some rates move before an FOMC decision, others adjust slowly, and some barely change. A policy increase does not raise every interest rate by the same amount at the same instant.

Two objectives, a changing balance of risks

The Fed's dual mandate is maximum employment and price stability. Its longer-run inflation goal is 2%, measured by the annual change in the headline PCE price index. Core PCE helps assess underlying inflation; it is not a replacement for the headline inflation goal. Maximum employment is not a fixed unemployment number the Fed can observe perfectly.

Persistent inflation can favour tighter policy; deteriorating employment can favour easing. The two risks can conflict. The FOMC assesses payrolls, unemployment, wages, CPI and PCE, growth and financial conditions together, allowing for revisions and policy lags. One strong or weak release does not automatically determine the next decision.

Hike, cut, hold—and the unit that matters

A hike raises the target range, a cut lowers it, and a hold leaves it unchanged. One basis point is 0.01 percentage point: +25 bp means +0.25 percentage point, not a 25% increase. For example, changing a range from 4.75–5.00% to 5.00–5.25% raises both bounds by 25 bp.

Keep three things separate: actual is what the FOMC decided; expected is the outcome or distribution priced before the release; surprise is the new information relative to that pricing. The simple examples below assume a single expected outcome. Real markets usually price a mixture of possible decisions and future paths.

Why a fully expected hike may produce little reaction—or a falling dollar

Suppose +25 bp is already priced. Delivering +25 bp adds little news about today's rate. If the Fed also suggests fewer further increases than investors had expected, the expected path can fall, Treasury yields can decline and USD can weaken. The relevant comparison includes the subsequent path, not just today's change.

The expression buy the rumor, sell the news describes one possible pattern: USD strengthens before an anticipated hike, then weakens after it. Position-taking, profit-taking or softer guidance may help explain it. The slogan is not a law, and observing a reversal alone does not identify which mechanism caused it.

Hawkish, dovish and conditional guidance

Hawkish communication points toward tighter policy or less easing; dovish communication points toward easier policy or less tightening. These labels are most useful relative to what was expected. A hold can be hawkish if it signals rates staying high longer. A hike can be dovish if it signals an earlier end to tightening.

Forward guidance communicates how policy may evolve with inflation, employment and risks. Conditional language matters: an outcome described as possible under one economic forecast is not an unconditional commitment. Separate the next meeting, the eventual peak rate and the length of time rates may remain restrictive.

What futures prices tell you—and what they do not

Federal funds futures reflect pricing for the average daily EFFR during a contract month. A quoted price of 100 minus the implied rate does not directly reveal a single meeting's target range: the number of days before and after the meeting, the EFFR–target relationship and other contracts matter.

Tools such as CME FedWatch translate these prices into implied probabilities under modelling assumptions, including possible rate increments and how EFFR changes with policy. Record the observation time, meeting and available outcomes. Implied probabilities are market estimates, not guarantees; risk premia, liquidity and assumptions can affect the inference.

The whole expected path matters. Pricing can change after CPI, payrolls, a central-bank speech or new financial stress. Yesterday's probability can be stale before today's announcement. Our historical study uses published futures-based surprises and revisions; it does not manufacture an exact pre-meeting consensus or probability series.

Dated snapshot: separate it from the lasting explanation

As checked on 2026-10-03, the latest FOMC decision was 2026-09-16: a +25.00 bp hike to 3.75–4.00%, implemented on 2026-09-17. The next scheduled meeting is 2026-10-27–2026-10-28; the next scheduled projections accompany the 2026-12-08–2026-12-09 meeting. The Chair in the latest press-conference transcript is Kevin Warsh. Powell's appearances discussed below are historical.

The 2026-09-16 projections put the median year-end federal funds rate at 4.10% for 2026 and 4.10% for 2027. These are participants' conditional appropriate-policy projections, not live futures prices or promised decisions. The latest available inflation and employment releases are dated in the table; their reference months differ.

We checked the public FedWatch entry and its calculation method, but could not verify a timestamped live probability table. No current meeting probability is quoted here. Check freshly timestamped market pricing alongside the dated Fed snapshot; neither the September projections nor the September event surprise measures October expectations.

Read the statement beyond the rate line

Compare the statement with the previous version. Look at the assessment of inflation and employment, the balance of risks and language about future adjustments. A wording change matters when it changes the economic interpretation or the expected policy path. A stronger adjective by itself does not quantify a rate surprise.

Record the decision, publication time and implementation date separately. Scheduled statements are generally released at 14:00 New York time, with the press conference at 14:30. Check the actual calendar and time zone; daylight-saving changes can alter the corresponding local time. Minutes released later are another information event, not the original decision.

The dot plot is not a promised route

The Summary of Economic Projections, or SEP, normally accompanies four scheduled meetings a year. Its dots represent individual FOMC participants' assessments of appropriate policy under their economic outlooks. Participants include Reserve Bank presidents who are not voting that year. The median is a summary of individual views, not a voted collective rate plan.

Compare the distribution, the median and the accompanying inflation, growth and unemployment projections with the previous release. A higher median can reflect a changed outlook. The people behind the dots and their assumptions can also change. A year-end midpoint projection is not the current target-range upper bound or a market-implied probability.

Why the press conference can change the first reaction

The Chair can clarify the expected pace of adjustments, the level needed to restrain inflation, policy lags or the conditions for easing. Those answers may differ from the market's first reading of the statement. Listen to the economic content and explicit conditions, rather than drawing conclusions about the speaker's personality.

Treat the statement and press conference as separate information windows. A falling dollar after the statement can become a rising dollar across the combined event. That is an observed change in price response; explaining it requires the public communication and market-pricing evidence. The November 2022 case below shows this distinction directly.

Treasury yields connect the expected path with USD assets

A 2-year Treasury yield is often sensitive to revisions in expected short-term rates over the next few years. A 10-year yield also reflects a longer span of expected rates, inflation and growth, plus term premium—the compensation investors require for holding duration risk. Duration describes the sensitivity of a bond's price to interest-rate changes. Neither yield is simply today's federal funds rate.

Bond prices and yields move inversely, holding cash flows fixed. A higher yield can make USD assets more attractive, but an existing bondholder may have suffered a price loss. For currency demand, compare relative expected returns, inflation and hedging costs. A nominal yield increase driven by inflation or risk need not strengthen the dollar.

Use yields as evidence about the market's interpretation, not a compulsory confirmation signal. USD and a particular Treasury maturity can diverge. Our intraday yield observations are benchmark on-the-run Treasury yield changes from USMPD; they are not the Treasury's daily par-yield estimates and must not be mixed with them as if they were the same series.

Every currency pair compares two policy environments

For EUR/USD compare the Fed with the ECB; for USD/JPY, the Bank of Japan; for GBP/USD, the Bank of England; for AUD/USD, the Reserve Bank of Australia; and for USD/CHF, the Swiss National Bank. Compare expected paths and relevant yields over a similar horizon, rather than subtracting rates on unlike policy instruments without context.

A wider expected US rate advantage can support USD, other things equal, but exchange-rate moves, risk and hedging can outweigh that advantage. Carry is the return associated with a rate differential while bearing currency exposure. Retail rollover also includes instrument terms and provider charges; it is not automatically the difference between two central-bank headline rates.

EUR/USD: the Fed is only one side

If the expected US path rises while the ECB path is unchanged, the relative rates backdrop may favour USD, which would correspond to a lower EUR/USD quote. If both paths rise, the change in their difference may be small. Euro-area inflation, growth, risk and ECB communication can change the other side of the comparison.

A move in EUR/USD is not a pure Fed-policy measurement. Read it alongside US and euro-area yields and the relevant news clock. The historical November 2022 example illustrates a US communication event; it does not prove that European factors were absent or that the same announcement would always produce the same exchange-rate move.

USD/JPY: yields, carry and Japanese policy

USD/JPY expresses yen per dollar. A rise means USD strengthened against JPY. Compare Treasury yields and expected Fed policy with Japanese government bond yields and BoJ policy. Expectations of Japanese tightening can narrow an anticipated rate advantage even while the Fed holds or hikes.

Risk aversion can unwind leveraged carry positions and support the yen; demand for dollar funding can support USD at the same time. Japanese data, intervention concerns and changes in BoJ policy can also matter. The balance varies by episode. Fed raises, therefore USD/JPY rises is not a reliable rule.

Payrolls feed expectations; they do not decide policy

NFP measures nonfarm payroll jobs in the establishment survey. The unemployment rate comes from a different household survey. Wages, participation, hours and revisions add context. Strong payrolls with slowing wages can convey a different message from strong payrolls with renewed wage pressure.

The chain is labour-market evidence → assessment of employment and inflation risks → expected Fed policy → yields → possible USD response. The surprise relative to expectations and revisions matter. A headline payroll number is not an automatic instruction for the next FOMC meeting. The separate NFP study examines the price and execution side of these releases.

CPI, PCE and core inflation are different measures

CPI measures prices paid by urban consumers. PCE covers a broader set of consumption, including spending made on households' behalf, such as some healthcare payments. They differ in scope, expenditure weights and index formulas. Their inflation rates can differ legitimately; one should not relabel CPI as PCE or compare them as identical baskets.

Core measures exclude food and energy to help assess the underlying trend, while headline measures retain them. The Fed's 2% longer-run objective is headline PCE inflation. Compare monthly and annual changes consistently, note seasonal adjustment and revisions, and distinguish the release date from the month measured. A higher print may change pricing, but the response depends on what was expected and on its components.

QE and QT: another policy channel

Quantitative easing, or QE, expands securities holdings to ease financial conditions, especially when conventional rate policy is constrained. Quantitative tightening, or QT, reduces holdings, often by allowing maturing securities to run off under stated limits. These tools can affect duration supply, term premia, reserves and broader financial conditions.

Rate policy and balance-sheet policy are distinct decisions. Reserve-management purchases to maintain ample reserves are not automatically QE aimed at lowering long-term yields. Neither QE nor QT supplies a mechanical formula for USD. Read the purpose and implementation terms rather than treating every balance-sheet increase as the same economic action.

High rates, cuts and safe-haven demand

A high nominal rate can coexist with a weak currency if inflation is high, growth deteriorates or investors expect sharp future cuts. A cut can coexist with a stronger dollar if it delivers less easing than priced, or if the future path is revised upward. Today's rate level and the change in expected future relative returns answer different questions.

During global stress, demand for liquid USD assets or dollar funding can strengthen USD independently of a hawkish Fed surprise. That safe-haven or funding channel is distinct from the rate channel. Fiscal news, geopolitical risk, positioning and other central banks can shift the balance. Do not attribute every USD move on an FOMC date to the rate decision alone.

DXY is useful context, not the entire FX market

DXY is a fixed six-currency dollar basket: EUR 57.6%, JPY 13.6%, GBP 11.9%, CAD 9.1%, SEK 4.2% and CHF 3.6%. Its large euro weight helps explain why DXY and EUR/USD often look like opposite views of a similar move. A change in this basket is not a universal measure of every dollar pair.

Analyse EUR/USD directly for a euro position and USD/JPY directly for a yen position. The Fed's broad trade-weighted dollar index has a different coverage and weighting method. The long-term chart below uses that broad index; the intraday event tables use USMPD's DXY. They are labelled separately and are not spliced into one series.

A long-term view without a mechanical rate rule

The two panels show the target range's upper bound and the Fed's daily nominal broad dollar index from 2015-01-01 to 2026-09-25. The dollar index is January 2006 = 100. Separate scales preserve the units: a percentage interest rate is not an index level. Missing daily observations are not replaced by zeros.

Read periods of divergence as a reminder to investigate expectations and other forces. The chart describes co-movement over time; it does not estimate a causal effect, compare inflation-adjusted returns or prescribe a trade. Exchange rates can anticipate changes months before a policy step appears in the upper panel.

Scenario analysis: compare the information with the starting view

In an illustrative +25 bp case, the rate increase matches expectations but guidance is softer than expected. The market may price an earlier peak or more cuts, yields may fall, and USD may weaken. A hold can produce the opposite result if the expected duration of high rates increases. These are scenarios, not historical claims or guaranteed directions.

Matching an already hawkish expectation does not by itself add support for USD. The table distinguishes matching, softer and tighter information relative to the starting view. Even a genuine policy surprise competes with other news, monetary policy in other economies and positioning.

Five questions for reading an FOMC event

First, what did the Fed do? Second, what had the market priced immediately beforehand? Third, what changed about the future path, conditions or risks? Fourth, how did 2-year and 10-year yields respond? Fifth, how did USD respond in the index and the actual currency pair? Record the timestamps for each answer.

This order separates information from interpretation and price observation. If yields and FX diverge, look for the missing relative-policy, inflation, risk or liquidity story. It is an analysis framework, not a requirement to open a position. An unanswered question is a reason to preserve uncertainty in the explanation.

A practical before-and-after checklist

Before the release, save a timestamped view of pricing and the relevant economic releases. Identify whether an SEP is scheduled and which other central banks or data releases can overlap. After the statement, update that record again during and after the press conference. An old screenshot is not evidence of the latest expectations.

For each observation distinguish levels from changes: the rate range from its change, the yield level from its basis-point move, and the exchange-rate quote from its percentage return. Compare the same instrument and time horizon. That discipline prevents a sound macro story from resting on mismatched figures.

Our study: a defined period and a consistent sample

We reviewed all 61 scheduled FOMC announcements from 2019-01-30 to 2026-09-16 in the preserved USMPD snapshot, updated 2026-09-17. The comparative sample contains 60 meetings: 12 hikes, 39 holds and 9 cuts. Seven unscheduled events, including the emergency March 2020 cuts, are excluded by the source flag. One scheduled event, 2021-01-27, is excluded from all comparative results because its yield windows do not reconcile. No scheduled field is missing.

Starting in 2019 gives a press conference after every scheduled meeting. This makes the information windows comparable; it does not make the sample representative of every monetary regime. The analysis combines different inflation and growth conditions and deliberately excludes emergency events, whose reactions can be dominated by simultaneous stress.

Statement: 30 minutes, from 10 minutes before release to 20 after. Press conference: 70 minutes, from 10 before its start to 60 after. Whole event: 100 minutes, from 10 before the statement to 60 after the press conference begins. All timestamps use New York time, including daylight-saving rules. We do not claim to have measured unavailable 5-minute or exact 1-hour returns.

Decisions, futures surprises and observed returns

Actual changes and target ranges are checked against each official statement and the Fed's effective-date rate history. USMPD MP1 measures a current-meeting rate surprise implied by futures; MP2 measures the revision to the next-meeting expected rate. They are changes in market pricing, not an archived pre-event probability or a survey consensus.

The table reports the whole event. A +75 bp decision alongside an MP1 close to zero means the large policy step carried little current-rate news by this measure. It does not mean all communication was anticipated. We do not backfill an exact expected decision by subtracting MP1 from the actual change and presenting that as a documented consensus.

Source rate changes are in percentage points and are multiplied by 100 to obtain bp. Source FX and DXY returns are already in percent. Whole-event returns use the published whole-window observation, not the arithmetic sum of the shorter percentage returns. Reference data describe market moves, not obtainable trade fills or bid–ask spreads.

What the full sample shows—and does not establish

Across the 60 whole events, DXY rose in 25 and fell in 35. Among 12 hikes it rose in 4; among 9 cuts it rose in 5. These counts directly contradict a universal hike-up, cut-down rule within this sample. The group sizes are small and the decisions were not random experiments.

The sign of the statement-window move differed from the whole-event sign in 19 cases for DXY, 20 for EUR/USD, 22 for USD/JPY and 17 for the 2-year yield. Small sign changes count too; this is not a count of profitable or economically large reversals. The table reports descriptive means and medians, not expected future returns.

No causal regression or trading strategy is claimed. Communication can reveal both policy intentions and information about the economy; other news can overlap. Regimes differ, categories are unequal and the event windows are not execution models. Neither a group average nor the frequency of a historical direction is a signal for the next meeting.

EUR/USD case: a 75 bp hike, two opposing readings

On 2022-11-02, the Fed raised its range to 3.75–4.00%. The statement-window MP1 was −0.54 bp, much smaller than the actual 75.00 bp step: distinguish a large action from a large current-rate surprise. The statement added attention to cumulative tightening and policy lags, creating room for a reading that the pace could slow.

In the statement window, the 2-year yield fell 8.72 bp and EUR/USD rose 0.57%, indicating USD weakened against EUR. During Powell's historical press conference, he explained that the eventual rate level could be higher than previously expected and that discussing a pause was premature. The 2-year yield rose 15.79 bp, and EUR/USD fell 1.13% in that separate window.

Over the whole event, the 2-year yield rose 7.07 bp and EUR/USD fell 0.57%. This sequence is consistent with a revised reading of the policy path; the observed numbers alone do not isolate a causal effect of one sentence. For a complete relative analysis, also check ECB pricing and euro-area yields, which are not measured in this dataset.

USD/JPY case: an unchanged rate with a changed outlook

On 2023-09-20, the Fed held 5.25–5.50%. Statement-window MP1 was 0.00 bp; the median SEP projection for the following year's rate moved from 4.60% in June to 5.10% in September. Those are conditional participant projections. A hold therefore delivered information about the expected persistence of restrictive policy.

Over the whole event, the 2-year Treasury yield rose 7.91 bp and USD/JPY rose 0.29%, while DXY rose 0.43%. The last BoJ decision before this event, 2023-07-28, retained a −0.1% rate on specified policy-rate balances and made yield-curve control more flexible. These were different policy instruments in a different environment, not two directly interchangeable rates.

The US move is consistent with a stronger relative US rates backdrop if Japanese expectations do not offset it. We have no matched intraday BoJ-expectation or JGB series, so that condition is an interpretation, not a measured causal result. Intervention concerns, risk and Japanese news can still affect the pair. Indeed, USD/JPY was lower at the next H.10 noon fix despite rising during the FOMC window.

Three more historical checks on the simple rule

On 2019-07-31, the Fed cut 25.00 bp to 2.00–2.25%. Powell described the adjustment in the context of the expansion rather than the start of a lengthy easing cycle. Whole-event DXY rose 0.42% and the 2-year yield rose 7.41 bp. MP1 was +4.00 bp: the futures measure revised the current-rate outlook upward despite the actual cut.

On 2023-03-22, the Fed hiked 25.00 bp to 4.75–5.00%, while its guidance shifted away from an expectation of ongoing hikes toward the possibility of additional firming. Whole-event EUR/USD rose 0.86% and the 2-year yield fell 15.34 bp. Banking stress and other contemporaneous information limit any claim that the guidance alone caused those moves.

On 2024-09-18, the Fed cut 50.00 bp to 4.75–5.00%. Statement-window DXY fell 0.46%, then rose 0.44% in the press-conference window. The whole-event change was only −0.02%. Different windows can turn a dramatic first reaction into a much smaller combined result. None of these examples is selected to represent an average meeting.

Next-day and fifth-observation FX: a different question

H.10 rates are noon buying rates in New York. On these FOMC dates that fix is before the 14:00 announcement. Our daily comparison starts at that event-day fix and ends at the first or fifth later valid published fix, skipping ND holidays. These are not closing-price returns, exactly 24-hour moves or five elapsed calendar days.

The formula is 100 × (later quote ÷ event-day quote − 1). The endpoint dates are shown explicitly. Intervening data, monetary policy elsewhere and risk news enter these longer windows, so they cannot be labelled pure FOMC effects. Intraday and daily results answer different questions and need not have the same sign.

How the calculations were checked

Every event was checked for date, official decision, range, actual change, futures fields, yields, currency returns, units and window. The workbook matched a separately retrieved copy. An independent audit read the XLSX XML directly, rebuilt daily returns with exact fractions and checked sample summaries with a separate numerical calculation.

This checks the archived published observations and our arithmetic; it does not independently certify the underlying LSEG tick feed. No missing value is imputed, no unobserved expectation is invented, and no chart line claims intraminute prices absent from the inputs. Later provider revisions may differ from this preserved snapshot.

For 2021-01-27, the published 2-year yield changes are −0.19 bp in both shorter windows and −0.19 bp for the whole event, although the shorter changes sum to −0.38 bp. Without the underlying quotes, we cannot determine which observation is responsible for the mismatch. The entire meeting is excluded consistently from every FX and yield summary; all raw observations remain archived. No corrected value is invented.

Common errors that break the explanation

Trading only the headline, treating +25 bp as automatic USD support, treating a cut as automatic USD weakness, using stale expectations, ignoring foreign central banks or replacing the pair with DXY all discard essential information. So does confusing a yield rise with a bond-price gain or treating the dot median as a promise.

Another error is to discover a later price move and write the expectation backward to fit it. Preserve evidence from before the event when it is available, identify what is actually measured and label inference. An economic surprise is new information, not a guaranteed direction or an assured profitable entry.

FOMC analysis and execution are different tasks

Around a release, volatility can jump, spreads can widen and quotations can gap. A stop may trigger on one quote and execute at a worse available price. Slippage and the absence of liquidity can matter even if the macro interpretation ultimately proves correct. Opening a position seconds before the release adds exposure to these conditions.

Our percentage returns are reference-market observations, not simulated bid–ask fills, stop execution or account profit. Costs and leverage are not modelled. The stop-loss guide explains trigger prices, slippage and gaps; the carry guide explains why financing income cannot remove currency risk.

Read the change in the story, then the change in prices

The practical habit is to ask what the market expected, what it received and how expectations changed afterward. Follow the decision and guidance through yields, compare the other currency's policy environment, and observe the actual pair over a defined time window. Keep a historical fact, an illustrative scenario and an inference visibly separate.

Forex continually reprices what the Fed may do next. A hike, cut or hold is one part of that information set. The five-question framework helps explain the reaction while preserving uncertainty about its cause, its persistence and any future outcome.