Overview
Goldman Sachs trading desks say algorithm-driven funds are set to redeploy a historically large sum into U.S. equities after a period of sharply reduced exposure. The downgrade in risk-taking followed a market selloff tied to the Iran war, during which Commodity Trading Advisers sold roughly $48 billion of S&P 500 futures over the past month, according to Goldman traders.
Model-driven reversal
Goldman’s trading models signal a material change in CTA behavior. The models show Commodity Trading Advisers turning into net buyers of U.S. equities across both one-week and one-month horizons, independent of whether markets rise or fall. In a scenario where markets remain flat, Goldman’s estimates put potential CTA demand at approximately $45 billion over the next week - a level described by the desk as among the largest such weekly inflows on record. The analysis and figures were reported by Goldman’s Paul Leyzerovich.
Momentum indicators shifting
Alongside CTA positioning, momentum signals used by Goldman’s trading desk have shifted following the recent market rebound. The thresholds that inform short-, medium- and long-term demand have moved into positive territory, implying that model-based demand readings are now oriented to buy rather than sell across those horizons.
Implications for market participants
The anticipated rotation from heavy selling to sizable buying by algorithmic strategies could materially influence flows into equity index futures and broader U.S. equity markets. The change in CTA positioning and the flip in momentum thresholds are central inputs to the desk’s expectation of a strong near-term bid.
Key points
- CTAs sold about $48 billion of S&P 500 futures over the past month, per Goldman traders.
- Goldman’s models project CTAs becoming net buyers across one-week and one-month horizons, with a flat-market scenario implying roughly $45 billion of demand next week.
- Momentum thresholds for short-, medium- and long-term demand have flipped to positive after the market rebound.
Risks and uncertainties
- Model projections may change if market conditions evolve in ways not captured by the current inputs; the analysis is model-driven and therefore subject to model limitations.
- The prior reduction in CTA exposure occurred during a selloff linked to the Iran war; similar geopolitical developments could again alter algorithmic positioning.