Goldman Sachs: Long-Short Crowding Nears Five-Year Extremes, Quarter-End Pension Selling Pressure ~$30 Billion
nashnova research
Goldman's John Flood warns that U.S. equity long and short crowding factors are both near five-year extremes, momentum exposure sits at the 98th percentile, and roughly $30 billion in quarter-end pension rebalancing selling hits next week — if the trend reverses, deleveraging will accelerate fast.
How rare is $30 billion of quarter-end selling?
Roughly $30 billion in U.S. equities will be sold next week as pensions rebalance — mechanically trimming stocks that have outgrown their target allocation.
This means → it is not a panic sell; it is passive, rules-driven trimming. But the size ranks at the 89th percentile of all buy/sell estimates over the past three years, and at the 95th percentile going back to January 2000.
In plain terms = pensions are not calling a top. Stocks simply rose past their target weight and must be cut back. The problem: this time, the cut is larger than almost any in recent history.
Crowding and momentum at extremes — what does that mean?
Both long-crowding and short-crowding factor exposures are near the most extreme levels in five years. Mid-term momentum factor exposure sits at the 98th percentile of its five-year lookback.
This means → capital is heavily concentrated in the same trend trades. Everyone is on the same side of the boat. If direction reverses, the stampede for the exit will be simultaneous — deleveraging pressure amplifies fast.
Flood's conclusion is blunt: prepare for sustained volatility.
Are hedge-fund positions extreme?
U.S. long-short gross leverage is 207.3%, at the 4th percentile over one year (near the lowest). Net leverage is 54.5%, at the 74th percentile.
Overall positioning is not extreme, but a structural split has emerged: IT-sector exposure has surged to a near-five-year high, while Mag 7 gross and net exposure have both fallen to one-year lows, driven by short selling since June.
This reflects a rotation: hedge funds are pulling back from the mega-caps and spreading into the broader tech sector. Flood sees the Mag 7 drawdown as a potential long entry opportunity.
Record equity supply — how did the market absorb it?
Alphabet completed a $40 billion equity raise — the largest primary-market deal in U.S. corporate history. Just six trading days later, SpaceX's $75 billion IPO broke that record. Combined supply of over $115 billion was absorbed smoothly in under two weeks.
Goldman's trading desk saw no meaningful liquidation selling from asset managers or sovereign wealth funds. Mutual funds hold roughly $170 billion in cash, in line with the historical average.
On June 18, total U.S. exchange volume hit 33 billion shares, an all-time single-day record. This means → liquidity is deep enough for now; retail buying is a key pillar, and Flood expects this tailwind to last through year-end.
Record semiconductor positioning — is sector broadening a good sign?
Semiconductors were the most net-bought sub-industry globally in 2025 and topped the list again in H1 2026. Net allocation has more than doubled year-to-date to an all-time high, led by Asian chipmakers.
Since June, U.S. single-stock flows have broadened — 8 of 11 sectors saw net buying, led by financials, industrials, and consumer discretionary. IT and energy were the heaviest net sellers.
In plain terms = money is no longer pouring into semiconductors alone; it is spreading across industries. Flood sees sustained broadening into H2 as a positive signal — it means the market is no longer standing on one leg.
What is the biggest risk for the second half?
Q1 S&P 500 EPS grew 18% year-over-year (excluding one-offs); median stock earnings growth hit 14%, one of the strongest quarters in a decade. AI-infrastructure stocks are expected to contribute roughly half of the S&P 500's full-year earnings growth.
But new Fed Chair Warsh's first FOMC meeting was more hawkish than expected — of 18 participants who submitted rate projections, half expect one or more hikes in the remainder of 2026. The median core-PCE inflation forecast for Q4 2027 rose to 2.5%.
Goldman's economists maintain their base case: no hike remains the most likely outcome, because roughly half of those projecting hikes are non-voting regional Fed presidents. Goldman also cut its U.S. recession probability from 25% to 15% and raised its H2 annualized GDP growth forecast to 2%.
This means → fundamentals are solid for now, but whether extreme positioning can unwind in an orderly way atop those fundamentals is the defining test for markets in the second half.
市场有风险，内容仅供研究参考，不构成投资建议。