Nomura constructs a composite G10 FX positioning index using five indicators, with the latest data still pointing to room for further unwinding of long USD positions

Overview

Addressing the lack of a unified positioning measure in the FX market, the report combines five indicators spanning futures, options, and CTAs into a composite positioning index for six G10 currencies. Nomura believes this method is more stable and more representative of the overall market than reliance on a single indicator. The latest results show that although long USD positions have declined, other currencies have not broadly reached extreme long positioning, and USD positioning may still shift further toward net short.

Core views

The report begins by noting that assessing FX positioning is both an art and a science. Monitoring positioning serves two main purposes: first, identifying asymmetric drawdown risks after positions become excessively crowded; and second, determining whether a currency has established a trend while positioning remains light, leaving room for the trend to continue. The problem is that any single indicator covers only one segment of the market and may overstate or understate overall positioning. Using EUR/USD since 2017 as an example, COT asset manager data in early 2018 and late 2020 would have suggested that the speculative market was long EUR, while leveraged fund data from the same source indicated the opposite; the reverse divergence emerged in late 2023. Similar contradictions exist in other G10 currencies, and individual indicators frequently diverge significantly from exchange-rate movements. To reduce such bias, Nomura constructs a composite positioning indicator for EUR, GBP, AUD, NZD, JPY, and CAD. The indicator contains three broad categories and five specific components: the number of net long contracts held by asset managers and leveraged funds in IMM non-commercial futures; three-month risk-reversal pricing and the DTCC daily ratio of call-option to put-option trading volume; and an estimate from the CTA FX positioning tracker developed by Nomura's quantitative strategy team in Japan. Both risk reversals and the call/put volume ratio use 20-day averages to reduce short-term volatility. Data for all currencies and indicators cover the period from 2016 to the report date, although CTA data for NZD and CAD begin only in 2020. Each component is first converted into a z-score using a one-year rolling window, measuring whether current positioning is relatively long or short versus its average over the preceding year rather than measuring absolute positioning. This treatment also helps address the long-term one-sided bias of certain indicators. For example, IMM asset manager positioning in EUR has never been net short in absolute terms since 2016, while AUD risk reversals have exhibited a persistent bias toward AUD weakness. Nomura compared different lookback periods: a three-month window tends to generate noise and overstate or understate positioning, while a three-year window produces too few changes and signals. It therefore selected one year as a balance between signal and noise. In terms of weights, the three broad categories—IMM futures, the options market, and CTAs—each account for 1/3. Within IMM, asset managers and leveraged funds are equally weighted, while within options, risk reversals and the call/put trading volume ratio are equally weighted. The final weights are therefore 1/6 for asset managers, 1/6 for leveraged funds, 1/6 for risk reversals, 1/6 for the call/put trading volume ratio, and 1/3 for CTAs. Asset managers and leveraged funds are separated because the former typically hold larger contract volumes, while the latter adjust more flexibly; if aggregate non-commercial positioning were used directly, leveraged fund signals could be overwhelmed. Although price-based and volume-based options data usually move in the same direction, including both can smooth anomalous volatility in individual markets and capture different timing and magnitudes of change. The composite indicator generally tracks FX movements more closely than individual indicators, but Nomura does not expect it to move fully in sync with spot exchange rates: the indicator does not cover all capital flows, and z-scores mean-revert, whereas exchange rates themselves can sustain trends and adjust to new equilibrium levels. The report therefore compares composite positioning with each currency's performance against USD over the preceding three months and finds that, over the full sample period, the overall fit across all currencies is better than that of any individual indicator. This does not mean positioning must replicate past performance; rather, when the composite indicator diverges from an established trend, the divergence warrants more attention than one based on an individual indicator. The report further tests whether positioning can indicate exchange-rate performance over the subsequent three months. Three months was selected because the best-fit horizons for the six currencies mostly fall between 8 and 16 weeks. To avoid excessive currency-by-currency optimization and preserve comparability, a uniform three-month window is used. The test sample begins in 2017, except for NZD and CAD, which begin in 2020 because of their shorter CTA histories. The report examines two “extreme positioning” thresholds, ±1 and ±1.5 standard deviations, and emphasizes that the research is not intended to establish a systematic trading model. At the ±1 standard deviation threshold, the composite indicator has relatively weak explanatory power. In the order of “composite indicator, asset managers, leveraged funds, risk reversals, call/put trading volume, CTA,” the explanatory-power values in the table are: 0.05, 0.03, 0.08, 0.01, 0.03, and 0.01 for EUR; 0.05, 0.05, 0.03, 0.02, 0.00, and 0.00 for GBP; 0.00, 0.00, 0.02, 0.08, 0.03, and 0.00 for AUD; 0.20, 0.03, 0.04, 0.08, 0.04, and 0.09 for NZD; 0.10, 0.02, 0.05, 0.11, 0.09, and 0.05 for JPY; and 0.05, 0.08, 0.02, 0.10, 0.00, and 0.03 for CAD. After the threshold is raised to ±1.5 standard deviations, the composite indicator shows a clearer reversal relationship for currencies other than EUR: currencies with extreme net long positions tend to weaken against USD over the subsequent three months, while currencies with extreme net short positions tend to strengthen. No single component consistently exhibits this pattern. In the same order, the explanatory-power values are: 0.04, 0.02, 0.13, 0.06, 0.02, and 0.00 for EUR; 0.22, 0.11, 0.02, 0.04, 0.00, and 0.00 for GBP; 0.13, 0.03, 0.04, 0.11, 0.07, and 0.00 for AUD; 0.66, 0.07, 0.04, 0.11, 0.09, and 0.06 for NZD; 0.32, 0.03, 0.06, 0.15, 0.20, and 0.04 for JPY; and 0.43, 0.16, 0.01, 0.13, 0.01, and 0.00 for CAD. EUR is the exception: even at the higher threshold, the regression coefficient between positioning and future direction remains slightly positive rather than negative as required by the reversal pattern, and most indicators also have low explanatory power. Nomura tested other combinations and weights but found no approach that consistently improved signals across all currencies. It therefore chose a consistent and comparable framework rather than overfitting each currency separately. Regarding the latest data, long USD positions have declined markedly, while short positions in other G10 currencies have narrowed or even shifted to long over the past month, consistent with the dollar's underperformance since late June. However, composite long positioning in other currencies has rarely reached extreme levels. Based solely on positioning data, there is still room for further unwinding of long USD positions, and the market may shift further toward a more explicit bearish USD stance. The charts also show that long AUD positioning remains in place but has declined. GBP is the currency closest to the crowded range, with a composite positioning z-score slightly above 1, driven mainly by bullish options trading volume. However, the report's historical tests show that one standard deviation is insufficient to constitute a significant near-term reversal signal. Nomura continues to hold a short GBP/NZD trade. The trade has performed well recently, and despite the rise in GBP positioning indicators, the report still believes the cross can continue declining toward 2.25. If long GBP positioning becomes more crowded, it could reinforce this trend. Nomura also holds long EUR/USD and EUR/JPY views. The composite data show that relative short positioning in EUR is the widest among the six currencies, driven mainly by negative readings from the IMM indicators. The report observes that underlying capital flows into the euro area are improving and expects the European Central Bank to raise rates in September, with further upside risks thereafter. This could reduce the negative carry from holding EUR and encourage the market to continue reducing EUR shorts. For JPY, speculative short positions have declined markedly following intervention by Japan's Ministry of Finance and the US Treasury. Because the current weakness in JPY does not appear to be driven by substantial speculative selling, the incentive for further intervention in the near term may be limited.

Analysis framework

Nomura first uses historical examples to demonstrate divergences among individual positioning indicators and between those indicators and exchange-rate movements, then standardizes and combines five futures, options, and CTA data series using assigned weights. The report subsequently compares the composite indicator's relationships with performance over the preceding three months and returns over the subsequent three months, tests ±1 and ±1.5 standard deviation thresholds and different weights, and finally applies the unified framework to the latest USD, GBP, EUR, and JPY positioning and related trading views.

Methodology notes

• Quantitative/Factor/Portfolio TheoryMulti-factor model

  Weighted combination of five positioning indicators

  The report combines IMM asset managers, IMM leveraged funds, three-month risk reversals, the call/put option trading volume ratio, and CTA estimates into a single indicator to reduce the dominance of any one market segment over the overall assessment.

• Quantitative/Factor/Portfolio Theory

  One-year rolling z-score standardization and 20-day smoothing

  Each positioning data series is standardized using its mean and volatility over the preceding year to measure the extent to which it is long or short relative to recent history. Options price and volume data use 20-day averages to reduce short-term noise.

• Quantitative/Factor/Portfolio Theory

  Extreme positioning thresholds and subsequent three-month return tests

  The report defines different degrees of extreme positioning using ±1 and ±1.5 standard deviations and compares their explanatory power for exchange-rate returns over the subsequent three months. A uniform three-month horizon is used because the best-fit periods mostly range from 8 to 16 weeks and to avoid excessive currency-by-currency optimization.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• USD

  Long USD positions have already declined markedly, but the report believes there is still room for further unwinding and a shift toward more explicit net short positioning.

  Weaknesses

  It has underperformed since late June, while short positions in other G10 currencies have narrowed or shifted to long over the past month.

  Comparison

  Long positions in other currencies have generally not yet reached extreme levels.

  Risks

  This assessment is based solely on positioning data, and the composite indicator does not cover all types of capital flows.

• EUR, EUR/USD, EUR/JPY

  Nomura holds long EUR/USD and EUR/JPY views, believing that the relatively broad short EUR positioning has further room to be covered.

  Strengths

  Underlying capital flows into the euro area are improving, and an expected ECB rate hike in September may reduce the negative carry of EUR positions.

  Weaknesses

  Current composite positioning remains the most bearish among the covered currencies, mainly due to negative IMM readings.

  Comparison

  Relative short positioning in EUR is the widest among the six covered currencies.

  Risks

  EUR is the only currency in the historical tests that does not exhibit a clear reversal pattern following extreme positioning.

• GBP/NZD

  Nomura maintains its short GBP/NZD position and believes the cross can continue declining toward 2.25.

  Strengths

  The trade has performed well recently, and if long GBP positioning becomes more crowded, it could reinforce the cross's decline.

  Weaknesses

  GBP's composite positioning z-score is already slightly above 1, driven mainly by bullish options trading volume.

  Comparison

  GBP is currently the only one of the six currencies beginning to approach the crowded range.

  Risks

  The one-standard-deviation threshold itself has not provided a significant near-term reversal signal.

• JPY

  Following intervention by Japan's Ministry of Finance and the US Treasury, speculative JPY shorts have declined markedly; Nomura also holds a long EUR/JPY view.

  Strengths

  Speculative net short positioning has narrowed from previous levels.

  Weaknesses

  JPY remains weak, but the report believes its weakness is not clearly driven by large-scale speculative selling.

  Comparison

  Compared with the period before the intervention, speculative investors are no longer as bearish on JPY.

  Risks

  The incentive for near-term intervention may be limited because speculative selling has weakened.

Key data

• Currencies covered6EUR, GBP, AUD, NZD, JPY, and CAD, all analyzed in terms of performance relative to USD.
• Underlying data period2016 to the report dateAll indicators and currencies cover this period, but CTA data for NZD and CAD begin in 2020.
• Standardization window1 yearThe three-month window produces excessive noise, while the three-year window generates too few signals; the report selects one year to balance signal and noise.
• Final component weightsAM 1/6, LF 1/6, RR 1/6, C/P 1/6, CTA 1/3The three broad categories—IMM, options, and CTA—each account for 1/3.
• Options data treatment3-month RR, 20-day averageRisk reversals are used as the price signal, while the DTCC call/put trading volume ratio is used as the volume signal; both are smoothed using 20-day averages.
• Return test windowSubsequent 3 monthsThe best-fit period for each currency is generally 8 to 16 weeks, and the report uniformly uses three months to maintain comparability.
• Extreme positioning thresholds±1 and ±1.5 standard deviationsThe ±1 standard deviation threshold has limited predictive power, while at ±1.5 standard deviations the composite indicator provides a clearer signal of reversals for most currencies.
• Composite indicator explanatory power at ±1 standard deviationEUR 0.05, GBP 0.05, AUD 0.00, NZD 0.20, JPY 0.10, CAD 0.05The report considers this threshold to have only very limited predictive ability overall.
• Composite indicator explanatory power at ±1.5 standard deviationsEUR 0.04, GBP 0.22, AUD 0.13, NZD 0.66, JPY 0.32, CAD 0.43Except for EUR, the composite indicator generally outperforms individual indicators in reversal tests following extreme positioning.
• Latest GBP composite positioningz-score slightly above 1Driven mainly by bullish options trading volume, but it has not yet reached the 1.5-standard-deviation level that provides a more meaningful reversal indication.
• GBP/NZD strategy target2.25Nomura maintains its short GBP/NZD position and believes further crowding in long GBP positions could reinforce the downward trend.

Impact & implications

The report believes the composite positioning index is better suited as a unified reference for assessing market crowding and identifying anomalous divergences than as a standalone systematic trading signal. Long positions in other G10 currencies have not yet broadly reached extreme levels, implying that there is still room for long USD positions to unwind and for the market to shift toward a more explicit bearish USD stance. Meanwhile, GBP is approaching but has not yet reached an effective crowding threshold, EUR shorts remain substantial amid improving capital flows and interest-rate differentials, and the decline in speculative JPY shorts may weaken the incentive for near-term intervention.

Risks

• The historical predictive power of the ±1 standard deviation threshold is very limited and cannot be used to conclude that a reversal is imminent.
• EUR does not exhibit the same extreme-positioning reversal pattern as the other covered currencies, and most indicators have low explanatory power.
• The composite indicator does not cover all capital flows, z-scores mean-revert, and exchange rates can sustain trends or shift toward new equilibrium levels.
• CTA data and related test samples for NZD and CAD begin only in 2020, resulting in shorter histories than for the other currencies.
• Different component combinations and weights did not consistently improve signals across all currencies, and the indicator is not a systematic trading model.

What to watch

• Monitor whether long USD positions continue to decline and whether the market shifts toward more explicit bearish USD positioning.
• Monitor whether GBP's composite positioning rises from slightly above 1 toward or beyond 1.5 standard deviations.
• Monitor underlying capital flows into the euro area, the ECB's September rate hike, and their effects on EUR's negative carry and short covering.
• Monitor changes in speculative JPY shorts and whether they alter the incentive for near-term intervention by Japan and the United States.