To The Moon

The rapid selloff in the bond market is shaking the confidence of macro investors, and what unsettles them even more is that a trading anchor used for years is ceasing to work.

According to a latest "Global Reflections" report from Louis Miller, Managing Director at the Goldman Sachs GSX trading desk, he visited multiple macro clients over the past week and received three core pieces of feedback: the speed of rising rates is more worrying than the terminal level; oil and rates, once inseparable assets, have quietly decoupled; and no one can predict the path of de-escalation in the Middle East. Combined, these three points mean macro investors must not only endure violent bond market swings but have also simultaneously lost their most important forward-looking indicator.

At the same time, internal Goldman Sachs data shows that the best-performing systematic strategy in 2026 is precisely large-scale short volatility—that is, betting that the market will not experience violent swings. This stands in sharp contradiction to clients' deep concerns about "speed risk" and plants potential fragility in the market.

Speed is the real risk

Miller pointed out that clients' core concern is not the 5.25% yield level itself, but the speed at which that level was reached. Since the end of August, the 10-year Treasury yield has risen by about 68 basis points cumulatively, and Goldman Sachs rates strategist Rikin Shah characterized the past month's move as a two-standard-deviation event.

Data from the Goldman Sachs MarketStrats team shows that the 2-year yield has risen 64 basis points since August 4 and 149 basis points cumulatively since the end of February; the 3-month implied volatility on 10-year dollar interest rate swaps has risen 28 basis points, or 2.3 standard deviations, sitting at the 100th percentile of the past year.

Looking at specific figures, the 10-year Treasury yield was 4.77% on September 3 and had risen to about 5.32% by last Thursday; on the same day, the 30-year Treasury auction priced at a yield of 5.618%, the highest since August 2000; the 10-year Treasury auction yield was 5.30%, also a new high since 2000. Yields ultimately ended the week at 5.25%, with the 2-year closing at 4.80%.

Goldman Sachs rates trader April Li further noted that the recent 20 basis point rise is different in nature from earlier moves: previously it was bear flattening, while the latest leg is bear steepening. With 100 basis points of hikes already priced in, upside room at the short end is near its limit, selling pressure is shifting to the long end, and term premium continues to accumulate under the combined effects of fiscal pressure, sustained supply increases, inflation uncertainty, and a shrinking marginal buyer base.

Li also emphasized that the yield level alone is not enough to attract real duration demand back. Whether the 10-year rises to 5.5% or the 30-year rises to 6%, large-scale buying is difficult to achieve in an environment with such high volatility. One piece of evidence: the Treasury's latest bond buyback operation attracted $46 billion in bids, while the average for previous operations was about $10 billion, which Li interpreted as a signal that investors are using buybacks to sell inactive bonds.

The decoupling of oil and bonds

For most of the Hormuz crisis, the bond market had a clear playbook: follow Brent crude. Rising oil prices meant rising inflation, which meant rising rate hike expectations, which pushed yields higher. However, Miller said this relationship has failed, leaving clients deeply uneasy.

This week's market action provided a complete chain of evidence. On Monday, oil prices and yields briefly rose together, then oil gave back gains while Treasuries continued to fall; on Thursday, news that the United States would not strike Iran before the midterm elections briefly boosted bonds, but the relief rally in crude was fleeting. For the full week, oil rose about $2.5 while the 10-year Treasury yield was roughly flat.

April Li gave the most concise description of the new regime: the sensitivity of Treasuries to oil prices has become asymmetric—when crude falls, bonds barely benefit; when crude rises, bonds still come under pressure. In other words, bad news for oil is still bad news for bonds, but good news for oil is no longer good news for bonds.

Goldman Sachs London rates strategist Mark Wilson holds a different view. In a Friday report, he noted that long-end yields remain closely correlated with 12-month forward crude prices, and that stabilizing energy prices would help ease the bond selloff and drive a year-end equity rally. The two views are not entirely contradictory: spot Brent crude includes a $22 per barrel risk premium estimated by the Goldman Sachs commodities team, the second-highest monthly reading in history, making it a headline instrument dominated by geopolitics; the back end of the forward curve is the true inflation pricing tool.

For macro investors, the practical impact is direct: the best hedge for being long oil used to be the trade itself—a ceasefire would push oil lower while duration assets would rebound accordingly. If de-escalation can no longer save bonds, then those clients who told Miller they "cannot predict where the situation is heading" have also lost their reason to wait for de-escalation.

Equity market divergence and the AI island

Miller's weekly charts show that the equity market's response to the above pressures is already clearly visible. Year to date, the S&P 500 is up about 14%, but the S&P excluding AI stocks (SPXXAI) is up only about 7%, and the gap between the two briefly widened to a record 10 percentage points over the past five weeks.

Miller's conclusion is that the market excluding AI has already digested the dual pressures of rates and oil, current positioning is light, and he is inclined to begin adding non-AI exposure, including consumer experiences, healthcare, and call options on SPXXAI (November 103% strike calls cost 0.75%). The appeal of healthcare lies in its correlation with AI, oil, and the 10-year yield all being close to zero. Goldman Sachs' Jack Singer provided a historical benchmark: since 1974, the median S&P 500 return in the month before midterm elections has been -3.5%, and +6.5% in the four months after.

The feast and hidden risks of short volatility

The strategy performance leaderboard appended at the end of Miller's report reveals a deeper contradiction. Of the 21 systematic strategies listed, 10 are some form of short volatility or skew harvesting, and another 5 are dispersion strategies. The top five by Sharpe ratio are all short volatility strategies: daily equity skew harvesting (4.34), copper volatility arbitrage (3.55, year-to-date return +30.9%), weekly S&P variance arbitrage (3.15), weekly WTI volatility arbitrage (2.69), and VIX volatility arbitrage (2.68).

However, significant fragility has accumulated behind this feast. The Goldman Sachs STS team noted that over the past year, the 1-month realized volatility of the S&P 500 has consistently been below 13%, while the VIX average has been 18. Asset managers' net short positions in VIX futures have hit a record, surpassing the July 2024 peak, and the volatility of volatility has fallen to near the August low. Equity volatility trader Oriol Lozano cited data showing that dealer gamma positioning is estimated at a record, and the assets of the five largest S&P oversubscribed ETFs surpassed $65 billion in early October.

The Goldman Sachs US rates trading desk noted that continued risk reduction by short volatility accounts is itself keeping rate volatility elevated, and record-high rate volatility in turn deepens the bond selloff by squeezing arbitrage trades—a self-reinforcing mechanism: the strategy works until volatility rises; and volatility rises precisely because the strategy is being unwound.

The current setup is this: rate volatility is at the top of its one-year range, equity volatility is at the bottom, and the divergence between the two is at a record. While clients tell Miller about speed risk, the most profitable strategies on the leaderboard are precisely those that sell protection against speed on a massive scale in every asset class. One hedge recommended by Goldman Sachs this week is a VIX November 20/30 call spread, costing 0.81 with a maximum payout ratio of 12 to 1—when the institution publishing the performance leaderboard simultaneously recommends buying volatility squeeze trades, that in itself is a signal worth watching.