DXY Dollar Index Forecast 2026: Live Read, Fed and Yields

The dxy dollar index forecast hinges on the Fed rate path, real US yields and the global risk tone. DXY Dollar Index trades at 98.74 as of 21 August 2026, three source 98.759 to 98.74 to 98.76, with the session running 98.69 to 98.88 against a previous close of 98.89. The desk reads the near term bias as broken and retracing: the 99.50 to 100.00 band that held for a fortnight is gone and the index sits at a three month low, down about 0.9 per cent on the week, though Thursday closed higher at 98.89 after trading to 98.56 so the last two sessions are a test and a bounce rather than a straight line. The tell is that United States yields rose overnight, the ten year to 4.7041 per cent and the thirty year to 5.2508 per cent, and the dollar was sold anyway, because the Treasury not the Fed is now driving the long end after its buyback expansion, leaving the 99.22 confluence of the 200 day and the 0.618 retracement of 97.628 to 101.801 as the overhead level that would repair the structure and 98.00 as the first congestion beneath, and frames direction by structure and the macro path rather than a single target. Live levels and scenarios are below.

Today’s DXY Dollar Index Desk Note · 23 August 2026
Last checked: 23 August 2026, 09:00 BST London morning · Sources: spot and performance cross-checked against TradingEconomics, FXStreet and the Monday-open indicative prints. Joint US and Japan intervention confirmed from Japan’s finance ministry and US Treasury statements. Iran and Oman talks per Al Jazeera and CNBC, 3 August.

Ken’s Take. The single most important thing on this page today is a relationship, not a level. Overnight US yields went UP and the dollar was sold anyway. The 10 year rose about 4.5 basis points to 4.7041 per cent, the 30 year added about 1.4 to 5.2508 per cent, and the index is 98.759, 98.74 and 98.76 across three sources, two ticks apart, down about 0.9 per cent on the week at a three month low. Higher yields are normally rocket fuel for a currency, and they are fuel only when they rise for the right reason. When they rise because investors want paying more to hold the risk, rather than because the economy is strengthening, the currency gets nothing. That is the distinction this desk has been putting on this page for a week, and this week it has become visible in the tape rather than remaining an argument. One precision point that cuts against the simple story and belongs here anyway: Thursday actually CLOSED HIGHER at 98.89 after trading down to 98.56. The last two sessions are a test of the downside and a bounce, not a straight line.

The confirmed leg is official. Japan and the United States have announced a joint yen-buying operation, stated out loud by both sides. That is a policy bid that can be repeated. The unconfirmed leg is the Iran de-escalation, which is real on the strikes and unsettled on the substance. One of these can be repeated on command. The other can reverse on a sentence. Size the two differently.

What changed. Bessent went further, and the market went the other way. On Thursday the Treasury Secretary said the government will increase the size of the buyback, possibly beyond 4 billion dollars per issue, declined to name a figure, and said the quiet part out loud: that part of it is signalling, to show that we believe the yields don’t reflect the underlying fundamentals. He also pledged a fiscal consolidation effort alongside budget director Russell Vought. Yields rose regardless. That is the second consecutive session in which an official intervention aimed at the long end has failed to hold it, and the dollar has now been sold through both of them. A currency that ignores its own central bank is one story. A currency that also ignores its own debt manager is a different and larger one.

Main driver. This desk owes a correction before going further, because it goes to the mechanism. Yesterday this page said the quarterly liquidity support envelope was unchanged at 38 billion dollars and that long end operations rise from two to four a quarter. Neither claim is in the Treasury release. Read at source, the release of 19 August says the maximum size per operation in the 10 to 20 and 20 to 30 year nominal coupon sectors goes from 2 billion dollars to at least 4 billion, effective 9 September and running through 4 November, and that an updated tentative schedule will be released later. The Treasury has therefore not said the envelope is unchanged and has not published a frequency. The 38 billion figure is real, but it belongs to the schedule published at the early August refunding, two weeks before the change. The honest formulation is that the per operation maximum in two long end buckets has at least doubled and the new total is not yet known. That correction matters because the whole bull case for the dollar rested on the buyback being a fixed pot reshuffled, which is debt management, rather than a genuine expansion, which is closer to something else. The market appears to have decided it does not much care which, and that is the point. Goldman Sachs strategist Vitali Meschoulam called the problem increasingly fiscal rather than technical and warned that yield suppression becomes progressively less effective the more the market focuses on sovereign financing. Carol Kong of Commonwealth Bank of Australia called it another example of the US government using unconventional tools and flagged the risk of encouraging more dollar hedging. Two independent desks, the same conclusion, and neither of them is treating this as a fix. US debt has topped 40 trillion dollars.

What the desk is watching. 99.222, and it is still overhead. That is the 61.8 per cent retracement of the move from the 97.628 war era low to the 101.801 high, and an independent reading puts the 200 day moving average at 99.181, four ticks away and rising gently. Two independent methods on the same price is the strongest single piece of technical evidence on this page. Beneath price the level that has actually been doing work is 98.613, which is the 76.4 per cent retracement of that identical swing. Thursday traded to 98.56, through it, and closed back at 98.89, which is a failed break rather than a break. Today’s low is 98.69, above it. The reason this desk is confident in that number is worth stating: an independent technical desk publishing on 20 August derives its own primary support at 98.60 as the 76.4 per cent of the SAME 97.628 to 101.801 swing this page has used since 14 August, and names 97.63 where this page names 97.628. Neither desk cites the other. That is corroboration, not an echo.

Key levels. Resistance begins at 98.88, the session high, then 99.00 which an independent map names as congestion, then 99.181 the 200 day average, then 99.222 the 61.8 per cent retracement of 97.628 to 101.801, which is the level that would repair the structure and where two independent methods agree to four ticks, then 99.42 which was the trigger and the underside of the old range, then 99.50 as the floor of Ken’s box which is now its ceiling, then 99.715 the 50 per cent retracement, then 100.00 as the top of that box, then 100.207 the 38.2 per cent and 100.35 which is the 15 July weekly low. Support sits at 98.69 the session low, then 98.613 which is the 76.4 per cent retracement and was breached intraday on Thursday and reclaimed, then 98.56 which is Thursday’s low, then 98.521 the 78.6 per cent, then 98.00 which an independent map names as congestion, then 97.628 which is the war era low and where that same independent map names 97.63 as the April to May lows, then 97.00. The 52 week range is 95.55 to 101.80. No moving average other than the 99.181 reading beside the 99.222 retracement is published on this page today, because that one is corroborated by a second independent method and the rest of the vendor set is not.

What would invalidate the view. A daily close back above 99.222 reclaims the 200 day average and the 61.8 per cent retracement together and turns this week into a stop run rather than a structural break, and it remains the cleanest invalidation on the board because two independent forms of resistance would have to fail at once. Short of that, a close above 99.50 puts price back inside the broken box and deserves to be treated as a failed breakdown. On the downside, a daily close beneath 98.521 opens 98.00 as the first named congestion and then 97.628. It has to be a close and it has to be on Ken’s own chart. The honest caveat is unchanged from yesterday and it cuts against the bearish case: the 30 year is STILL about 10 basis points higher on the month, which is wider than it was yesterday, not narrower. A spike was capped. A trend was not reversed. If the long end resumes rising into a Treasury that has now shown its hand twice, the second leg of this dollar move is a faster animal than the first.

Data and event risk. The Treasury announcement, Bessent’s Thursday remarks and the 28 to 29 July minutes are all done and in the price. Flash purchasing managers indices land today: the euro area at 09:00 London, where the July composite was 52.0 after 50.0 in June and the strongest reading in eight months, the United Kingdom at 09:30, and the United States this afternoon, where the July composite was 54.5 after 51.9 and the strongest since October 2025 with input cost inflation the fastest since November 2022. That last combination, decent activity with accelerating input costs, is the live risk to the short dollar trade and it is under reported. The Kansas City Federal Reserve’s symposium runs from Thursday 27 to Saturday 29 August at Jackson Lake Lodge on the theme of financial innovation and its implications for payments and policy, with roughly 120 central bankers and academics from more than 70 countries. Chair Kevin Warsh delivers the keynote on the morning of Friday 28 August, his first in that role since taking office on 22 May 2026, and it lands 19 days before the 16 September decision. It carries unusual information value because this chair has removed forward guidance from Federal Reserve communication. The target range is 3.50 to 3.75 per cent after five consecutive holds and the July vote was 9 to 3 with all three dissents FOR a hike. On pricing, this page publishes a September figure today where it published none yesterday, and the reason is stated in the pack: two dated readings of the same tool four days apart now agree at about 30 to 32 per cent, where yesterday vendor readings ran from 30 to 65 per cent. October and December readings remain cumulative and are still not published here.

Educational only, not financial advice. The note above is the desk’s read on the date shown. Trading is leveraged and most retail accounts lose money.

Work with the desk

One to one mentorship with Ken. How a desk actually builds a view, sizes it and survives being wrong. Applications are open. Apply here

The KenMacro AI powered gold system. Automated gold execution with a published reference record. The numbers, the risk data and exactly how copying works. See the record

New to gold EAs? What one actually is, why backtests lie, and how to spot martingale and grid before it costs you. The full review framework

Some links are affiliate links and the desk may earn a commission at no extra cost to you. Past performance is not a guide to future results.

Latest desk read · updated 21 August 2026
98.74-0.14%three source 98.759 TradingEconomics to 98.74 Investing.com to 98.76 on the wire  |  session 98.69 to 98.88, previous close 98.89  |  London pre open read

The single most important thing on this page today is a relationship, not a level. Overnight United States yields went up and the dollar was sold anyway. The ten year rose about four and a half basis points to 4.7041 per cent and the thirty year added about one and a half to 5.2508 per cent, and the index is 98.759, 98.74 and 98.76 across three sources, two ticks apart, down about 0.9 per cent on the week at a three month low. Higher yields normally support a currency, and only when they rise because the economy is strengthening. When they rise because investors want paying more to hold the risk, the currency gets nothing, and that is the distinction this page has carried for a week. One precision point that cuts against the simple story and belongs here anyway: Thursday actually closed higher at 98.89 after trading down to 98.56, so the last two sessions are a test of the downside and a bounce, not a straight line. The 99.50 to 100.00 box that this desk carried for a fortnight is gone, and what matters is that it broke on a session when the Federal Reserve was signalling it wanted to hike.

Macro backdrop

What moved it was the Treasury, not the Federal Reserve, and on Thursday the Secretary went further. Scott Bessent said the government will increase the size of the buyback, possibly beyond four billion dollars per issue, declined to name a figure, and said the quiet part out loud, that part of it is signalling to show that we believe the yields do not reflect the underlying fundamentals. Yields rose regardless, the second consecutive session in which an official intervention aimed at the long end failed to hold it, and the dollar has now been sold through both. This desk owes a correction that goes to the mechanism: yesterday it said the quarterly liquidity support envelope was unchanged at 38 billion dollars and that long end operations rise from two to four a quarter, and neither claim is in the Treasury release of 19 August, which says the maximum size per operation in the ten to twenty and twenty to thirty year sectors goes from two billion dollars to at least four billion, effective 9 September through 4 November, with an updated schedule to follow. So the envelope is not stated as unchanged and the frequency has not been announced, and the honest formulation is that the per operation maximum in two buckets has at least doubled while the new total is unknown. Goldman Sachs strategist Vitali Meschoulam called the problem increasingly fiscal rather than technical, and Carol Kong of Commonwealth Bank of Australia called it another example of the United States using unconventional tools and flagged more dollar hedging. Two independent desks, the same conclusion, and United States debt has topped 40 trillion dollars. The scheduled events that resolve it are the flash purchasing managers indices today and Chair Warsh’s first Jackson Hole keynote on 28 August.

Resistance
98.88this morning’s session high
99.00congestion on an independent map
99.181the 200 day moving average, rising gently
99.222the 0.618 retracement of 97.628 to 101.801, four ticks from the 200 day, the level that repairs the structure
99.42the old trigger and the underside of the broken range
99.50the floor of Ken’s box, now its ceiling
100.00the top of that box
100.35the 15 July weekly low
Support
98.69this morning’s session low
98.613the 0.764 retracement of 97.628 to 101.801, breached intraday on Thursday and reclaimed
98.56Thursday’s low
98.521the 0.786 retracement
98.00congestion on an independent map
97.628the war era low, where the same independent map names 97.63
97.00the next level beneath

Levels read from the desk’s own chart and cross checked against TradingEconomics, Investing.com and a market wire on 21 August 2026, with the United States curve from TradingEconomics and the prior session ranges from Investing.com daily historical tables. The 99.222 and 98.613 retracements of 97.628 to 101.801 and the 99.181 200 day are third party technical work, reproduced from their own anchors before publication and labelled as such, and an independent desk derives 98.60 as the 0.764 of the same swing. Ken’s own call is the 99.50 to 100.00 range, now broken. Levels, not trade ideas. Two sided risk, always.

What actually moves the dollar index

The dollar index, DXY, measures the US dollar against a basket of six major currencies, and the euro alone is more than half of that basket. So the index is really a rate and risk story. It firms when the market expects the Federal Reserve to hold rates higher for longer relative to the other central banks, and when global risk appetite sours and capital runs to the dollar as the world’s reserve haven. It softens when the Fed is expected to ease faster than its peers, or when risk appetite is strong and money leaves the dollar for higher beta assets.

Two forces sit underneath that. The first is real US yields, the Treasury yield after inflation, which set how much you are paid to hold dollars. Rising real yields pull capital in and lift the index; falling real yields do the opposite. The second is relative growth and policy: the dollar is always priced against someone else, so a weak euro area or a dovish ECB can lift DXY even when nothing has changed in the US. The desk reads the index as the Fed path minus the rest, adjusted for where global risk sits.

Why the dollar is the centre of gravity

Almost every other market is quoted against the dollar or funded in it, so the index is the single most important macro read on the board. A firmer dollar is a headwind for gold, for the dollar priced commodities and for the FX majors; a softer dollar tends to lift them. That is why the desk reads DXY first and the individual pairs second: much of what looks like a move in EUR/USD, GBP/USD or gold is really a dollar move wearing a different label.

How the desk reads the chart

DXY trends in long, persistent moves punctuated by sharp corrective pullbacks, so the level map matters more than any single candle. The near term bias stays firm while the index holds its nearest defended support and the run of higher lows is intact, and it turns cautious when a support that held several times finally breaks on a closing basis. The named levels in the live block come from a six lens confluence scan, prior day and weekly extremes, round numbers, moving averages, pivots and the volume point of control, not from a single indicator.

None of this is a trade instruction. It is a map of where the market has shown it cares, so you can frame your own plan and your own risk.

Frequently asked questions

What is the DXY forecast right now?

The live read at the top of this page is refreshed every session with the current dollar index level, the day’s move and the nearest support and resistance from the desk’s level scan. Direction hinges on the Fed path and real US yields: a higher for longer Fed and rising real yields support the index, while a dovish turn or a strong risk on tape tend to soften it.

What drives the US dollar index?

Three things. The Fed’s rate path relative to the other major central banks, especially the ECB, since the euro is over half the basket. Real US yields, the Treasury yield after inflation, which set the reward for holding dollars. And global risk appetite, because the dollar is the world’s reserve haven and catches a bid when risk sours.

Why does a strong dollar move gold and the FX majors?

Because they are priced against the dollar. When DXY strengthens, gold and the dollar priced commodities cost more in every other currency and tend to soften, and the FX majors fall by definition. Much of what looks like a gold or EUR/USD move is really a dollar move, which is why the desk reads the index first.

What are the key levels for the dollar index?

The named levels card above lists the current support and resistance zones from a six lens confluence scan, each tagged with why it matters, for example a prior weekly high or a defended round number like 100.00. The levels refresh every session as structure shifts, so the page always shows the live map.

Is the dollar a buy at current levels?

The desk does not publish buy or sell instructions on this page. It publishes the structure: where the index sits relative to defended support and overhead resistance, and what would shift the near term bias. Use that map to frame your own plan and risk. Trading carries risk and most retail accounts lose money.

Trade the dollar from the macro down, not the candle up

The dollar index turns on the Fed path, real yields and risk sentiment. The free KenMacro framework is the same risk-first, macro-aware approach behind this desk view, built so you can apply it to your own trading.

Get the free framework →

More from the desk: US dollar outlook  ·  how the dollar moves gold  ·  world interest rates