← All research
Issue

Vol. 4 · Oct 2, 2026 · ACIES CAPITAL · 20 min read

The Bond Market Broke a 2007 Ceiling. The Peso Did Almost Everything We Asked.

Executive Summary

Since September 17, the regime this book has tracked since June 4 changed its engine without changing its name. The Fed leg that drove three straight issues — hike odds rising from a coin flip to 93%, then the hike itself on September 16 — has gone into reverse: New York Fed President John Williams told markets on September 29 that the September move "removed any urgency" to act again, and an annual revision to the PCE data released the next morning retroactively cut July's core reading from 3.3% to 3.0% — the same 3.3% the Fed had hiked against two weeks earlier. October 28 hike odds, close to 70% two weeks ago, have fallen to roughly 35% on the most recent pricing this issue could source. And yet the dollar is stronger than it has been at any point in this book's life, because a second leg has taken over: the 10-year Treasury yield broke its 2007 closing high on September 30, the 30-year sits at a matching multi-decade high, and the driver is fiscal — weak auction demand and rising real yields, not Fed policy. Against that backdrop, and against a fresh, quantified Mexico-specific risk premium tied to a stalled USMCA round, USD/MXN has done nearly everything this book's June 4 thesis asked of it: the pair marked 18.3369 today (+6.19%), its best level since inception, 94% of the distance from entry to the 18.40 target, and more than 70 centavos clear of the 17.60 confirmation line — though that confirmation is not yet formal, since the rule requires a weekly close and the next one falls tomorrow. GLD, the book's rates-and-dollar-agnostic hedge, improved only to −6.88% even as equities rallied and bonds broke records; spot gold at $4,177 sits a bare $27 above the $4,150 stop, the closest this position has come to its exit since inception, eclipsing the cushion reported in Monday's note. Blended portfolio return on the $200,000 deployed (20% of the $1,000,000 fund) stands at −1.74% as of this issue, against −1.64% on September 17. Both benchmarks remain reference points reported unblended: the S&P 500 TR has rallied to +1.50% since inception (level 17,204.71, up from 16,932.47 on September 17), while USD-adjusted CETES has collapsed to −3.86% (nominal 6.3%) — the mechanical result of the same peso depreciation now carrying this book's flagship position toward its target. For the first time since June 4, the book sits ahead of one of its two benchmarks.

Global Macro & FX Regime

Classification: Risk-Off USD Strength — unchanged, and for the second time this book has had to report it, the mechanism underneath the label has changed while the label hasn't. In mid-September the dollar's strength ran through the Fed's policy rate. As of this issue, it runs through the Treasury market's term premium, a distinct and in some ways more durable channel — and this issue reports that handoff plainly, because conflating the two would misstate why this book's flagship trade just had its best two weeks since inception.

The Fed leg reversed, and reversed on data the Fed itself will have to live with having hiked against. New York Fed President John Williams said on September 29 that the September 16 hike "removed any urgency" for a further move, pulling October 28 odds from roughly 70% toward 50% in a single session. The following morning's August PCE report compounded it: core inflation came in at 0.2% month-over-month (forecast 0.3%) and 3.0% year-over-year (forecast 3.3%) — and buried in the release was an annual methodology revision that retroactively cut July's core PCE reading from 3.3% to 3.0%, the same 3.3% print that had helped justify the Fed's own hike two weeks earlier. October 28 odds on the most recent pricing this issue could source stand near 35%, down from roughly 70% two weeks ago. December 9 — the final 2026 meeting, carrying a fresh dot plot — has in the same move become the market's new near-certainty, priced at odds this issue found ranging from 83% to 100% depending on source and day; this book reports both ends of that range rather than picking the cleaner one.

The fiscal-dominance leg, flagged as unresolved throughout September, resolved — against the Treasury, not for it. The 10-year yield closed Tuesday, September 30 at 5.297%, through its 2007 closing high of 5.26%, with an intraday print of 5.304% that was the highest level since May 2002. The 30-year matched it at a multi-decade high the same session. The proximate causes are structural rather than cyclical: a seven-year Treasury auction drew its weakest bid-to-cover ratio in a year, and real (inflation-adjusted) yields climbed from 2.43% to 2.93% even as the headline and core PCE prints both came in cooler than expected. Mortgage rates followed to 7.3%, their highest since November 2023. The pattern this book first flagged in early September — the Treasury's doubled long-bond buyback operation capping yields only temporarily on announcement, not on trend — has now failed outright, in the most visible way available: a 19-year record.

The Hormuz leg eased on physical supply without resolving diplomatically, and gold's muted reaction follows directly from that split. Saudi Arabia's East-West pipeline restart is now complete in substance rather than in promise: Gulf oil exports have recovered to roughly 23.3 million barrels a day, matching the 2025 average, after effectively doubling through September. Brent has fallen to $97.3 and WTI to $89.6, both down from the $109.21 intraweek peak this book reported on September 15 and well off the $105 level this book has used as its crude watch-list trigger. Qatar is running shuttle diplomacy between Washington and Tehran; President Trump said this week the war would end "very soon" while denying reports that sanctions relief is on the table, and also rejected a further Iranian ceasefire proposal days earlier. No durable de-escalation has been confirmed — the diplomatic track remains exactly as frozen as this book described in the September 23 Wednesday note — but the physical oil-supply channel that drove gold's safe-haven bid through August and early September has largely closed. That is this issue's best explanation for why GLD improved only modestly (−7.50% to −6.88%) in a week when a textbook flight-to-safety trade would have expected more: the bond-market turmoil driving this week's dollar strength is a real-yield story, which is a headwind for gold, not a geopolitical one, which would have been a tailwind.

A fourth, Mexico-specific leg is now explicit in the reporting, and it is new to this issue. The USMCA round that Monday's note described as opening in Washington on September 28 did not produce the breakthrough that framing implied; as of this issue, no new round date is confirmed, and Commerzbank — cited directly in Wednesday-cycle FX coverage — attributes roughly one-third of the peso's more-than-4% slide since September 22 to a fresh, Mexico-specific risk premium tied to the stalled talks, with the remaining two-thirds driven by the broader dollar and bond-market move described above. That is the cleanest decomposition of this week's move this book has been able to source, and it matters for the regime call: part of this week's dollar strength against the peso is global (yields, Fed repricing), and part of it is idiosyncratic to Mexico (trade-policy risk), which is a distinction this book's original thesis anticipated but has not, until this issue, been able to size.

What would change the classification. A durable USMCA breakthrough that removes the idiosyncratic third of this week's move, landing alongside continued fiscal-dominance pressure on yields, would be the cleanest test yet of whether the global or the idiosyncratic leg is doing more of the work. A reversal in the 10-year back below its pre-auction range on strong demand at the next long-end auction would argue the "2007 ceiling" break was a liquidity event rather than a trend. Neither has happened as of this issue.

The Trade — USD/MXN: Six Centavos From a Target the Confirmation Line Hasn't Formally Cleared Yet

The seven questions, restated against a week in which this position moved more than it has moved in any prior single-issue window:

1. The price inefficiency. Narrower than at any point since inception, and narrowing fast. The peso was priced for a risk-on world in a risk-off regime; four months later, USD/MXN at 18.3369 is now 94% of the way from this book's 17.2682 entry to its 18.40 target — the inefficiency this book identified in June has, in large part, closed through price action rather than through this book taking profit, because no rule published at entry calls for an exit at the target itself. Entry-to-mark move: 106.87 centavos of a 113.18-centavo entry-to-target distance.

2. The macro driver. Materially reinforced, but by a different mechanism than the thesis's original framing anticipated. The original driver was a stuck-hawkish Fed and a compressing-but-still-positive carry differential against Banxico's 6.50%. Both of those specific conditions have moved against the thesis's original logic — the Fed's hike odds just collapsed, and the carry differential (now roughly 250–275bp with the Fed at 3.75–4.00% and Banxico unchanged at 6.50% since its September 24 hold) has compressed enough that FX coverage this week describes the carry trade as having "lost its edge" entirely. The pair rallied anyway, on the fiscal-dominance leg described above and on the fresh, Mexico-specific USMCA risk premium. The driver changed; the direction didn't.

3. What confirms it. Unchanged in the rule, newly almost moot in substance: a weekly close above 17.60. Today's 18.3369 mark sits 73.69 centavos above that level — the widest margin of this book's life — but the rule requires a weekly close, and the next one falls tomorrow, Friday, October 2. As of this issue, formal confirmation has not yet occurred.

4. What falsifies it. Unchanged: a credible, durable Middle East de-escalation plus a constructive USMCA opening restoring the risk-on carry bid, confirmed by a weekly close below 16.85. Today's mark carries a 148.69-centavo cushion above that stop — more than twenty-seven times the 5.4-centavo near-miss of September 9, the closest approach to a stop either position has made. The position is not remotely close to falsification as of this issue.

5. Entry / target / stop. Entry 17.2682. Target 18.40 — now six centavos away, the closest this book has been to a published target on any open position. Confirmation: weekly close above 17.60 (not yet formally triggered; see above). Stop: weekly close below 16.85. No levels are changed this issue, consistent with this book's standing rule against revising a published thesis.

6. Time horizon. Week 17 of a position originally scoped to an 8–12 week window (July 30–August 27 by that original math). This issue carries the extended-horizon disclosure first applied when the position ran past its original outer edge. The position is now, by price action, closer to resolution than at any point in that extended window; the disclosure stands regardless, because the rule governing exit is the published stop and confirmation levels, not elapsed time.

7. The repricing catalyst. Sequential, updated: Williams' September 29 remarks removing "urgency" from the Fed's hiking path — confirmed, pulled October odds from ~70% toward 50%. The August PCE release and its retroactive July revision (September 30) — confirmed, pulled odds further to ~35%. A weak seven-year Treasury auction and the 10-year's break of its 2007 ceiling (September 30) — confirmed, the single largest driver of this week's dollar strength and, by extension, of this week's mark. The stalled USMCA round and the absence of a new date (ongoing into this issue) — confirmed, and now quantified by outside analysts at roughly a third of the week's peso move. Next: tomorrow's weekly close is the formal test of the 17.60 confirmation line, landing the same day as the September jobs report; December 9's FOMC, now the market's base case for the cycle's next move, is the next scheduled source of a fresh Fed-leg catalyst.

Status as of live API (2026-10-01T18:43:32.331Z UTC): +6.19%, mark 18.3369, $120,000 deployed.

Actionable Event — Tomorrow's Weekly Close, a Jobs Report on the Same Day, and a Confirmation Line That Price Has Already Left Behind

Friday, October 2 is the weekly-close date that governs every published exit and confirmation rule in this book, and it lands the same day as the September employment report — consensus estimates this book could source put the print near +90,000 jobs with unemployment at 4.1%, though the data-revision volatility that reshaped this week's PCE narrative is a live reminder that initial prints in this cycle have not always held. The government shutdown risk flagged in earlier coverage is not a factor here: President Trump signed a continuing resolution on September 2 funding the government through December 11, so Friday's release proceeds on its normal schedule.

For USD/MXN, tomorrow's close is close to a formality. The pair would need to reverse more than 70 centavos in a single session to fall back below the 17.60 confirmation line — a move with no precedent in this book's four months of live marks. Barring that, this book will report the position's confirmation as formally triggered in its next publication, a milestone that, unusually, will arrive after price has already closed to within six centavos of the position's original target rather than before it.

For GLD, tomorrow's close is the live risk this book has flagged with increasing urgency across the last three weeks. Spot gold at $4,177 sits $27 above the $4,150 weekly-close stop — thinner than the roughly $39 cushion reported in Monday's note and the thinnest cushion this position has carried since inception. The move is, as argued above, a real-yield story rather than a geopolitical one: the same bond-market turmoil driving USD/MXN toward its target is the headwind working against gold. If spot tests $4,150 into Friday's close, the exit executes on the rule as published on June 4 — not on a judgment call made this week, and this book's post-mortem commitment (losses reported at the same depth as wins) would apply in full.

No new positions are triggered by Friday's close under any scenario currently in range. The crude watch-list candidate, at $97.3 Brent, is further from its $105 trigger than at any point since the Saudi pipeline strikes in mid-September — a genuine retreat, not a proximity fact worth reporting as a near-miss.

Portfolio Review

Live marks from investacies.com/api/quotes as of 2026-10-01T18:43:32.331Z. All figures authoritative from the live feed.

| Position | Dir | Size | Entry | Mark | P&L | Status |
|---|---|---|---|---|---|---|
| USD/MXN | Long USD | $120,000 | 17.2682 | 18.3369 | +6.19% | Best mark since inception by a wide margin; six centavos from the 18.40 target; confirmation line cleared by 73.69 centavos, pending tomorrow's formal weekly close |
| GLD | Long | $80,000 | 411.27 | 382.96 | −6.88% | Spot gold ~$27 above the $4,150 stop — thinnest cushion since inception, ahead of tomorrow's weekly close |

Portfolio return on deployed capital: −1.74%. At the fund level (20% deployed, 80% cash): −0.35% on the $1,000,000 fund, per the live API. Unblended against both benchmarks since June 4 inception: S&P 500 TR +1.50% (index level 17,204.71), USD-adjusted CETES −3.86% (nominal 6.3%). The book trails the S&P 500 TR by a wide and now-growing margin, as US equities continue a record-setting 2026 even as this book's two positions sit well below water on a blended basis. Against CETES, the comparison has inverted for the first time since inception: the same peso depreciation carrying USD/MXN toward its target has pushed the USD-adjusted CETES benchmark to a steep loss of its own, and this book's −1.74% is, for the first time, ahead of a benchmark it has trailed at every prior mark. That inversion is reported here exactly as it would be reported in reverse — without qualification, and without claiming more credit for it than a currency move common to both figures deserves.

USD/MXN — four publications, one accelerating arc. +0.96% (Wednesday, September 23) to +2.85% (Monday, September 28) to +6.19% (today) — each mark not merely an improvement on the last but a larger one, as the fiscal-dominance and USMCA legs described above layered onto a dollar-strength trade that had already turned positive. No change to entry logic, target, confirmation gate, or stop. The position is now closer to the terms of its own published thesis than at any point since June 4.

GLD — improved on a week that, by the textbook, should have been easiest. A bond-market selloff of historic proportions, a weakening equity-market correlation with gold typical of real-yield-driven stress, and continued (if easing) Hormuz-war uncertainty are the kind of backdrop that should support a safe-haven hedge. GLD improved only from −7.50% to −6.88% against that backdrop, and spot gold's $27 cushion above the $4,150 stop is now this book's thinnest since inception. This is read here, as in the prior two weeks, as evidence that gold is currently being priced more on the real-rate channel (a headwind) than on the geopolitical channel (a tailwind) — and that read has held for three consecutive publications.

Post-Mortem

No position has been closed since inception. Every rule stated in a prior note — including the USD/MXN 16.85 stop tested at 5.4 centavos on September 9 and the GLD stop now sitting a bare $27 from spot — has resolved, or stands live, without having triggered as of this issue. The full record of every rule stated and every outcome is available at investacies.com/portfolio. If either position closes before the next issue, its post-mortem will run at the same depth this book has committed to for wins and losses alike — and given how close GLD now sits to its stop, that commitment is live, not theoretical, heading into tomorrow's close.

Watch List — Preliminary, Not Taken

Nothing below has been entered in TradingView or the live portfolio. Named here before any action, per house rules.

Crude (USO/BNO) — active watch, retreating from its trigger rather than approaching it. Brent at $97.3 and WTI at $89.6 are both well off the $109.21 intraweek peak this book reported on September 15 and further from the $105 sustained-close trigger than at any point since the Saudi pipeline strikes. Gulf exports have recovered to roughly 23.3 million barrels a day — matching the 2025 average — as the East-West pipeline restart completes in substance. Qatar-brokered shuttle diplomacy between Washington and Tehran continues without a confirmed outcome. Still no entry; this issue reports a genuine retreat from the trigger, not a near-miss dressed up as one.

FEMSA (FMX) — long, preliminary; workup remains paused pending a legible USMCA outcome. The round this book expected to narrow the unresolved list further did not produce a confirmed new date; Mexican officials described a prior round as "postponed" with "October possible," and as of this issue no date has been confirmed. Mexico's nearshoring fundamentals continue to strengthen independent of the trade dispute — reporting this book could source this week put 2026 FDI on pace for a fresh annual record — but this workup stays paused on trade-policy risk specifically, not on sector fundamentals, exactly as stated in each prior issue.

Fiscal-dominance / long-duration Treasury pair — active watch, and the call this book has carried for three issues has now resolved, against the Treasury. The 10-year yield's close through its 2007 high and the 30-year's matching multi-decade high are no longer evidence of "not a durable cap" on long rates — the cap has failed outright, in the most visible way a bond market can register it. No position, no workup started; this book's house rules require a position to exist in TradingView before any issue references it as taken, and none has been entered.

Short-BRL/short-ZAR pair — active watch, reinforced and for the first time explicitly sized. Commerzbank's estimate that two-thirds of this week's peso move is attributable to the broader dollar and bond-market story — the same story this book has used to justify watching a wider EM-carry unwind since inception — is the clearest external confirmation this specific watch item has received. No entry.

Week Ahead

1. Friday, October 2 — the weekly close, and the September jobs report, on the same session. GLD's $4,150 stop is the live risk; USD/MXN's 17.60 confirmation is, barring an unprecedented reversal, a formality the market has already priced. Consensus for payrolls sits near +90,000 with unemployment at 4.1%, though this week's PCE revision is a reminder that initial prints in this cycle have not always held on review.

2. October 28 — the FOMC decision that this book spent three weeks treating as a near-certainty and must now treat as a genuine two-sided risk. Odds have fallen from roughly 70% two weeks ago to roughly 35% on the most recent pricing this issue could source, on Williams' September 29 remarks and the August PCE's retroactive downward revision to July's core reading.

3. December 9 — the final FOMC meeting of 2026, carrying a fresh dot plot, and the market's new base case for the cycle's next move. Pricing this issue found ranges from 83% to 100% depending on source and day; this book reports the range rather than picking the figure that reads cleaner.

4. USMCA — no confirmed date for the next round. The stalled talks are now a quantified driver of peso weakness in outside analysis, not just a standing risk this book has flagged since June. This remains the binding constraint on both the USD/MXN friction leg and the paused FMX workup.

5. December 11 — the next government-funding deadline, distant but on the radar. The continuing resolution signed September 2 runs out that day; this book will not treat it as live until it is within the normal two-to-three-week window this book uses to flag calendar risk, but it is noted here as the first mention.

Sight before strike.

Paper portfolio · research only · not financial advice.

Sources

Not investment advice. The Acies paper portfolio is hypothetical and uses no real capital. Past paper performance does not predict future results.