fed

Tuesday's settle had the odds of a 28 October hike at 15.5%, down from 18.8% the session before, and there's still no cut priced at any meeting out to December 2027. At those odds I think next week's FOMC is likely a hold, and for $XAU and $BTC the bigger variable going into it is what the long end of the Treasury curve does.
Why the long end
Gold and Bitcoin pay no cash flow, so you discount them at the long real yield, and when that rises they've got nothing to offset it with. I read Bitcoin as the purest duration asset in the market, with gold close behind it. You could see it on 1 September and again on 15-16 September, when the assets with no cash flow were the first to give as yields climbed.
What's priced for the Fed
The Fed hiked 25bp to 3.75-4.00% on 16 September, 12-0, and the dots put the median at 4.1% for both end-2026 and end-2027, with cuts only from 2028. Tuesday's settle has 1.04 hikes by December, 1.50 by January, 2.56 by April 2027 and 3.19 by October 2027. The front got a little less hawkish on the day while the back of the strip added, 3.11 to 3.19 by next October.
Warsh said on 16 October that the data had broadly conformed with what the Fed expected when it hiked. I read that as taking the urgency off the 28th while leaving October and December both open, which fits a front-month number in the teens. Every voter who's spoken since the hike leans to a hike or a hold... Hammack thinks conditions aren't restrictive and wants to move quicker, and Waller thinks higher yields are already doing some of the work. There's no new dot plot at this meeting either, that comes on 8-9 December.
Where the yield move came from
The ACM model had the 10-year term premium at +80.0bp on 19 October, the 99th percentile since 2015. Kim-Wright had it at +98.6bp on 16 October, the 100th percentile since 2015, so both models have it at the top of their post-2015 range.
ACM splits the move differently depending on how far back you look. Over the past year the fitted 10-year rose 105.8bp and term premium was about a quarter of it, so most of that was the market pricing a higher Fed path. Over the past month it rose 15.3bp and term premium was 110% of the move, which means the expected path actually fell and the premium did more than all the work. Over the last week the 10-year fell 10.4bp and 10.3bp of that was term premium coming back out.
So the recent swings, both ways, have come from the extra yield investors want for holding long bonds, and that responds to supply and to who's buying. At his September press conference Warsh listed government supply, weak savings, defence spending and capex among the reasons for higher long yields (hyperscalers competing for capital included), and on 16 October he named the AI build as one of the things pushing inflation up. I think the AI financing story that $QQQ trades is as much a bond-market story as an equity one, and Warsh has pointed at it twice now.
The 20-year kink
The 20-year now yields more than the 30-year, and today's $13bn 20-year auction is the long end's test this week. Bond volatility is off its peak but still higher than for most of the year, and my liquidity gauge reads worse than it did in September. I'd want both lower before calling liquidity back.
The case the other way
The data under the headlines is softer than the inflation prints. September PPI rose 1.3% on the month, but producer energy was +10.8% and core goods only +0.3%, so the oil move hasn't reached core goods yet. The retail sales control group rose only 0.1%, and the Beige Book had just 2 of 12 districts growing. Canada looks the same, with September CPI up to 3.1% on oil while the Bank of Canada's core measures averaged 1.85%, their lowest in six years.
There's a fair counter to the term-premium read too. Back in August, US nominal income was growing about 4.5% a year against a 30-year near 5.3%, and on that framing the long end looked like a buy. Bonds did rally on 15 October during the AI-debt selloff, with the long end gaining more than the front, and I read that flattening as growth worry.
On gold itself, the monthly downtrend runs out in October without having reached its target, and I count that as a bullish variable. There's no weekly signal yet.
What's dated
Thursday has jobless claims, the leading index and a 5-year TIPS auction, and Friday the flash PMIs. The FOMC meets 27-28 October with Warsh's press conference, the BoJ decides on the 30th after Ueda said more hikes are coming with the pace tied to the yen, and the October monthly bar closes on the 31st, the same day as the Mt. Gox deadline. The US midterms are on 3 November.
At 15.5% the decision itself probably won't move much, so I think the press conference and how the 2027 strip trades around it could count for more with gold and Bitcoin than the rate call. A BoJ hike without a stated pace hasn't threatened the yen carry trade so far. One that comes with a pace might be a different story.
I'll be watching the result of today's 20-year auction and whether the ACM term premium keeps climbing into the 28th.
#Fed #FOMC #Gold #Bonds #Macro