Equity Markets Positioning Model
Positioning Eases but Vulnerabilities Remain
The technology-led selloff weakened positioning across major equity markets. Nasdaq remains vulnerable because long positions are offside, European positioning continues to deteriorate, and KOSPI retains the largest regional deleveraging risk.
Citi’s Take
The recent technology-led selloff triggered a deterioration in positioning across most major equity markets, with long unwinds, new shorts and rising unrealised losses driving flows. The sharpest adjustment occurred in Nasdaq, where positioning reset lower but remains vulnerable given all longs are currently in loss. European investors continued to reduce risk through profit-taking and new shorts, pushing DAX into bearish territory and increasing squeeze potential should sentiment improve. In Asia, bearish flows were widespread, although KOSPI remains the market most exposed to further deleveraging. Here, positioning remains elevated despite the recent index decline. Looking ahead, upcoming earnings will be key in determining whether positioning stabilises or unwinds further.
United States
US positioning deteriorated sharply as the recent AI and technology selloff triggered widespread de-risking, with flows overwhelmingly bearish across large caps. S&P positioning eased primarily through long unwinds, while Nasdaq positioning saw a more aggressive combination of long liquidation and new short flows, leaving positioning at a one-month low. The key positioning risk remains concentrated in Nasdaq, predominantly for longs that are now entirely in loss while positioning levels remain elevated.
Europe
European positioning weakened despite relatively resilient headline index performance. EuroStoxx bullish levels declined, and DAX positioning turned outright bearish. The week's activity was characterised by profit-taking and fresh short selling, reflecting investor concerns around technology and semiconductor exposure. Although positioning momentum remains negative, the increasingly crowded short base creates squeeze risks should sentiment stabilise or macro data surprise positively. In contrast, FTSE positioning improved on short covering and a rise in risk flows over the week.
Asia
Asia positioning experienced the broadest regional deterioration as new short flows dominated activity across all indices tracked in the region. China A50, Hang Seng, Nikkei and KOSPI all saw bearish repositioning following the technology-led selloff, with positioning across China indices moving back towards neutral. The most significant risk remains in Korea, where positioning continues to rank higher despite recent losses. Elevated positioning and substantial unrealised losses leave KOSPI vulnerable to forced deleveraging and further downside pressure.
United States
Large-cap positioning reset as technology and AI-related stocks sold off, while small-cap positioning remained extended.
Another challenging week for US equity markets saw a pronounced selloff in technology and AI-related stocks. Escalating geopolitical tensions between the US and Iran also contributed to a broad selloff in equities, with the growth-sensitive Nasdaq witnessing the largest decline for the week.
Corporate earnings, banks in particular, provided a strong backdrop for an improved earnings season, although they did little to offset losses across the broader AI-exposed segment of the market.
Weekly flows were almost exclusively bearish for US large caps. S&P positioning at +0.8 changed largely through long unwinds, whereas new short positioning dominated flows for Nasdaq at +1.5, leaving bullish positioning at the lowest levels seen over the month. The decline was most pronounced for Nasdaq, which fell from +2.5 to +1.5 after the technology rout.
Russell 2000 positioning remained unchanged, with recent offsetting activity effectively nullifying any significant changes for the week. Positioning levels on Russell 2000 have been at extremes in recent weeks and remain so, while P&L build-up has stayed contained amid substantial return variability.
ETF positioning remained bullish and robust for the S&P 500 at +3.9 and Nasdaq at +2.1, suggesting that moderate and longer-term holders were not capitulating.
The S&P 500 P&L profile reveals mounting stress beneath the surface. While net P&L appears manageable, the long side remains more vulnerable to further unwinds, with two-thirds of long positioning offside. Nasdaq presents an even greater risk. Current P&L is extended at -2, with all longs currently in loss. The average long entry is near 30,000 while current levels are near 28,800, leaving an average positioning loss of about 4 percent.
With $26 billion in long notional completely offside and positioning still elevated at +1.5, Nasdaq downside risk could prompt further upheaval should technology earnings disappoint. The critical test arrives on July 22 and 23 with Alphabet and Tesla earnings. Guidance cuts or margin pressure could trigger another round of adjustment.
The small-cap trade benefited from the broadening participation theme, but the extreme positioning of +5 seen for Russell 2000 means that an economic disappointment could trigger rapid unwinding. The 69 percent offside ratio on longs suggests that many investors were late in chasing the rotation.
Europe
DAX moved into bearish territory while EuroStoxx positioning faded and FTSE positioning improved.
European equity markets navigated a complex week, caught between spillover weakness from US and Asian technology stocks and improving regional inflation dynamics. The STOXX Europe 600 managed a marginal gain of 0.07 percent, masking significant internal divergence. The UK's FTSE 100 climbed 0.98 percent, benefiting from low technology exposure and energy-sector strength, while Germany's DAX fell 0.94 percent.
Despite index levels holding steady, EuroStoxx positioning declined from +1.2 to +0.8 and DAX positioning fell to -0.4, flipping to bearish levels. Weekly activity was driven by profit-taking unwinds and new short positioning. In contrast, FTSE 100 saw a positioning uplift from risk flows and short-covering activity.
DAX remains the only developed-market index tracked where shorts exceed long positioning. The bearish switch reflects its greater technology exposure, which leaves it more vulnerable to a further semiconductor selloff. The DAX short book has reached a three-year high, creating potential squeeze risk if a positive rebound emerges.
FTSE shorts are currently all in loss. Any potential rebound from covering could be limited by the smaller size of the short book.
European Banks saw modest profit-taking despite strong second-quarter earnings across the sector. Positioning is one-sided bullish at +1.4, though it is not extended. Long profits remain relatively small, leaving limited positioning risks.
Asia and Emerging Markets
Bearish flows spread across the region while KOSPI retained the largest loss-sensitive positioning risk.
Asian equity markets bore the brunt of the recent technology selloff, with technology-oriented indices leading declines. Citi’s equity strategy team remained broadly supportive of the region, upgrading China and Taiwan while turning neutral on Korea in view of volatile trading conditions.
Positioning flows were firmly bearish across all indices in the region, with new shorts dominating weekly activity. The largest decline came from China A50, where positioning activity was solely attributed to short reallocations. The deterioration for Hang Seng was less extensive, but positioning on both indices was trending back towards neutral despite beginning from opposite ends.
KOSPI positioning retained its bullish +1.7 tilt despite the sharp decline, while positioning momentum did not change. Nikkei weekly flows followed a similar trend, with new shorts and long unwinds leading positioning lower.
The rapid decline across indices has led to a build-up of P&L risks, notably across markets where positioning is still elevated. KOSPI remains the stand-out. All of the $5 billion long book is in loss, loss levels are substantial, and existing positioning remains near the 91st percentile. Because losses are already extensive, a continued decline across semiconductor stocks leaves the market vulnerable to forced covering that could add further downside.
Similar P&L risks are present in Nikkei and China A50, though they are less extensive. Each retains a long portfolio that is completely offside and could face similar challenges if the selloff continues.
How Positioning Changed
The following charts compare weekly changes, three-year positioning histories, current rankings, exposure, P&L and breakeven levels across the markets in the model.
Equity Markets Positioning Charts
Each chart combines a current position snapshot with market data, cumulative flows, normalised positioning and normalised P&L.
United States
S&P 500
Nasdaq 100
Russell 2000
Europe
EuroStoxx 50
DAX
FTSE 100
European Banks
Asia Pacific
S&P ASX 200
Nikkei 225
KOSPI 200
Hang Seng
China A50
Broad international indices
MSCI EAFE
MSCI EM
The Equity Markets Positioning Model
The aim of the Equity Markets Positioning Model, or EMP, is to provide information on current equity market positioning and how this might influence subsequent price action. It allows readers to visualise investor exposures in equity futures across multiple markets.
- Quantify current market positioning and profitability.
- Identify extreme market positions.
- Categorise periods in which price action will be dominated by positioning.
The model calculates a headline positioning score between -5 and +5 and provides a visualisation of the current and past market.
How the model works
At a micro scale, price action is dominated by the size, level and speed at which investors choose to execute trades with financial intermediaries. Buy and sell orders can drive prices higher or lower, and the EMP attempts to quantify these flows in order to understand market direction and motivation.
This is an ambitious objective because no single market participant holds a complete view of all market flows. The analysis aggregates futures trading to extrapolate positioning across the whole market, including cash trading. Based on this information, the model estimates two related measures.
- Market positioning
- The outstanding market positioning, including whether there has been net buying or selling in recent history and the price and size at which trades were executed.
- Market P&L
- The unrealised profit or loss within the market, given the current outstanding positioning, and the point at which this may motivate investors to manage risk.
The analysis has limits. Incomplete trade visibility, opaque instruments and unknown investor trading strategies create uncertainty in the true picture of high-level equity market positioning.
The EMP model calculates historical equity market positioning over the last three months and the associated unrealised P&L for the main markets around the world. The assets and contracts used in the EMP are presented in Figure 30, followed by the calculation methodology.
Positioning calculations
The model identifies an objective subset of positions entered into or closed on a given day and classifies these as the relevant marginal long or short positions that determined the price move as the aggressor. The report refers to these positions as longs, shorts and net position, but these terms always refer to marginal positioning. The presumption is that this subset of positions is the most price-sensitive and is more likely to generate new flows, including profit-taking or squeezes, in response to future price moves.
In practice, the model uses public trade-activity data on exchange-traded futures contracts, which is available daily. Net positioning is always zero for futures because the exchange matches buyers and sellers and creates or cancels contracts as necessary. By tracking daily changes in overall open interest and combining that measure with the change in price, the model can classify the marginal positioning change for that day as set out in Figure 31.
The change in open interest is taken to be the relevant marginal subset of positions for that day. Together with the change in price, it determines whether the change is treated as a long or short position.
A worked example
Performing this analysis day by day creates a trade-activity history over a three-month period, which can be converted into a snapshot of cumulative positioning and unrealised P&L. Figure 32 presents these calculations for S&P 500 futures in early March 2020.
During this period, open interest increased and prices fell, suggesting an increase in the short investor base. The model shows numerous days when new short positions were opened and relatively few when long positions were established, in addition to the covering of some long and short positions. At the end of the period, cumulative uncovered positions and P&L equal the sum of the outstanding positions and P&L. In Figure 32, cumulative positions equal -459,041 contracts and unrealised P&L equals $868 million.
This positioning can be visualised as a horizontal bar graph ordered by price. Figure 33 shows that shorts were centred around 2740 to 2750 and were significantly in the money. This contrasts with the large long position at 2696, which was losing money as the market fell. The snapshot can be summarised by cumulative position and unrealised P&L levels.
EMP uses the daily volume-weighted average price in its calculations instead of closing prices because this produces more robust results. The model also chooses to cover trades with the greatest economic impact, meaning that trades with the largest gains or losses are closed first. In every other respect, the methodology is identical to the method presented above.
The reported numbers are normalised to a range from -5 to +5, representing large shorts or losses through to large longs or profits, so that results can be compared across assets and instruments.
What Output Is Generated by EMP
The EMP model provides a snapshot of current positioning. Figures 34 through 36 summarise outstanding exposure and unrealised P&L for a three-month period in both futures and ETFs.
Figure 34 shows typical cross-market positioning in futures. In this historical illustration, investors are neutral to net short EuroStoxx 50 contracts after a decline from a net long position over the previous three months. This contrasts with MSCI EM futures, where investors are sitting on profitable positions.
The model further breaks down market positioning by analysing long and short positions, the average entry price and the position relative to trading over the previous three months.
A complete visualisation of positioning is presented in Figure 36. It orders current positioning by volume-weighted average prices over a three-month period. Bars show the levels at which trades were executed, with red for short positions and blue for long positions. Lighter bars imply older trade activity. When a trade is covered, the bar is hatched. Shading indicates whether it was covered at a loss in yellow or a gain in green. Horizontal red and blue dotted lines show the respective average entry prices of contracts over the preceding three months.
What it means for cash equities
Although futures are traditionally used as a hedge, investors also use them to take an outright market view. This becomes increasingly relevant when futures positioning is extended. Citi explores this topic in Regime Modelling using Futures Positioning, Futures vs. Cash, A Distant but Meaningful Relationship, which examines the interplay between futures positioning and behaviour in underlying cash markets.
Citi Quant Research Team
David Chew1
david.chew@citi.com+44-20-7986-7698Anju Bhandari4
anju.bhandari@citi.com+1-212-816-3812Richard Schlatter4
richard.w.schlatter@citi.com+1-212-816-0591Yue Hin Pong1
yue.hin.pong@citi.com+44-20-7986-3953Alex Saunders4
alexander.saunders@citi.com+1-212-723-10581 Citigroup Global Markets Ltd. 2 Citigroup Global Markets Asia Limited. 3 Citigroup Pty Limited. 4 Citigroup Global Markets Inc..
The Research Analyst Affiliations listed above, other than those identified as employed by Citigroup Global Markets Inc., are not registered or qualified as research analysts with FINRA. Such research analysts may not be associated persons of the member organisation and therefore may not be subject to NYSE FINRA 2241 restrictions on communications with a subject company, public appearances and trading securities held by a research analyst account. Unless indicated in Appendix A-1 of this document or any referenced document, the analysts listed above have not contributed to this document or any referenced document.
The supplied PDF contains 29 file pages although the cover states 38 pages. Printed page 28, Figure 37 and Appendix A-1 are not included in the supplied file. This webpage does not reconstruct the absent material.