Foreign investors’ growing holdings of dollar-denominated bonds have created a large demand to hedge dollar risk. Using outstanding FX forward and swap positions for seven major dollar currency pairs, this column shows that changes in fund hedging are closely tied to exchange rate movements. Dealer banks transmit derivative demand to the spot market, while investors’ tendency to hedge less after the dollar appreciates can amplify currency movements.

Exchange rates are among the hardest asset prices to explain. Since Meese and Rogoff (1983), economists have documented how poorly standard macroeconomic fundamentals account for currency movements – a finding often called the exchange-rate disconnect puzzle. More recent work has emphasised international capital flows. Yet a large part of cross-border bond investment is hedged in derivatives, and this second leg of the transaction is often left out of the analysis.

That omission has become increasingly important. Over the past decade, foreign investors have built large positions in US dollar bonds, while US investors’ holdings of foreign-currency bonds have grown much less. The resulting imbalance creates substantial net demand from non-US investors to sell dollars forward. In a new paper, we explain why variation in these hedging positions helps explain movements in both forward and spot exchange rates (Bräuer and Hau 2026).

From bond investment to the spot exchange rate

Consider a European fund that buys a dollar-denominated US bond. The bond purchase gives the fund a long dollar position. To protect the portfolio against a dollar depreciation, the fund can sell dollars forward and buy euros forward. A primary dealer bank typically takes the other side of this contract.

Dealer banks do not usually keep the resulting currency exposure. They offset it through covered interest parity arbitrage, combining spot-market transactions with borrowing and lending in the two currencies. To supply the fund with a dollar short position, the dealer sells dollars in the spot market. An increase in hedging demand therefore creates additional spot dollar sales and puts downward pressure on the dollar. A fall in hedging demand has the opposite effect.

This hedging channel works alongside the familiar capital-flow channel. The initial purchase of a US bond creates spot demand for dollars, while the hedge partly reverses that demand. When the bond position is fully hedged, only the unhedged component of the capital inflow supports the dollar. The effect is especially relevant because the dollar’s dominant role in global bond markets makes the two sides asymmetric: hedging by foreign holders of US bonds is not offset by an equally large amount of hedging by US holders of foreign bonds.

Figure 1 The hedging channel links forward and spot FX markets

Notes: Non-US funds demand dollar short positions in the forward market. Primary dealer banks supply those positions and offset their exposure through covered interest parity arbitrage, transmitting the demand shock into spot dollar sales. Source: Bräuer and Hau (2026).

Measuring hedging pressure

We use daily data from CLS, the world’s largest multi-currency cash settlement system. CLS reports outstanding FX forward and swap positions by currency, maturity, and counterparty type. The sample runs from September 2012 to March 2022 and covers seven major dollar pairs: the euro, pound sterling, yen, Swiss franc, Canadian dollar, Australian dollar, and New Zealand dollar. The data represent around 20% of global outstanding forwards and swaps and about 28% of their daily trading volume.

Our measure of hedging pressure is the difference between funds’ outstanding dollar short and dollar long positions in derivatives, scaled by average market-wide open interest. A positive value means that funds are net sellers of dollars forward. Across the seven currencies, average hedging pressure is 12.4%, and it rises over the sample period in every currency. Funds are also the largest price-taking participant group in these markets; for EUR/USD, they are a counterparty in 63% of all outstanding forward positions.

A strong link between hedging and the dollar

The headline result is visible in Figure 2. Annual changes in aggregate hedging pressure and annual changes in an equal-weighted dollar index have a correlation of -0.70. Periods in which funds increase their currency hedging through net dollar short positions are periods in which the dollar depreciates; periods of reduced hedging coincide with dollar appreciation.

Panel regressions across the seven currencies confirm that the relation is not limited to the aggregate time series. A one-standard-deviation increase in hedging pressure is associated with a 0.72% dollar depreciation – roughly one-third of the dollar’s monthly variation. Hedging pressure alone explains about 9% of monthly exchange-rate movements across the seven currency pairs and roughly one-fifth of movements in the aggregate dollar basket.

The result is robust across daily, weekly, monthly, and quarterly frequencies and to controls for yield differentials, the Treasury basis, the VIX, and currency and time effects. The relationship is almost identical for spot and three-month forward rates. In this sample, changes in hedging positions have more explanatory power than changes in bilateral bond holdings themselves.

Figure 2 Hedging pressure and the US dollar move in opposite directions

Notes: The blue line is the annual change in aggregate hedging pressure from funds; the green dashed line is the annual change in the equal-weighted dollar spot rate across seven currencies. A higher spot rate denotes dollar appreciation, while higher hedging pressure denotes more net short selling of dollars. The correlation is -0.70. Source: CLS, Bloomberg, and Bräuer and Hau (2026).

Estimating the elasticity of hedging demand

The strong negative relationship in Figure 2 can run in both directions. More hedging can weaken the dollar because dealer banks sell dollars in the spot market. But funds may also change their hedges after the dollar moves. Distinguishing between these directions is central to our contribution. It allows us to estimate not simply whether hedging and exchange rates move together, but how funds’ demand for dollar protection responds to changes in the dollar. In our model, this feedback determines whether hedging dampens or amplifies exchange rate movements.

To isolate the response of funds, we use changes in dealer-bank capital as a supply-side source of variation. Better-capitalised banks can intermediate more FX trades and supply more dollar-short positions, but changes in their capital should not directly alter funds’ desire for currency protection. Using daily data for 640 banks, we isolate bank-specific capital changes from common shocks, persistent differences across banks, quarter-end effects, and days with widespread extreme movements. Our preferred instrument is constructed in the spirit of Gabaix and Koijen (2024) and compares capital changes between large and small banks. So-called ‘granular supply side shocks’ allow us to isolate how the fund demand for hedging reacts to changes in the dollar price.

We find that funds hedge less after the dollar strengthens. A 1% dollar appreciation reduces demand for net dollar-short positions by about 0.19%, equivalent to approximately $15.6 billion in 2022. In economic terms, the estimated hedging-demand elasticity is -0.19. Put simply, investors buy less protection against a dollar decline after the dollar has risen. This response is important because it can reinforce, rather than offset, the initial currency movement.

Why might funds behave this way? After the dollar appreciates, unhedged foreign investments outperform hedged investments, which may make investors less willing to hedge. Expectations that recent dollar strength will continue, or momentum-based portfolio decisions, could produce the same response. We do not separately identify these motives, but each is consistent with less demand for dollar hedges at a higher dollar price.

Why a negative demand elasticity amplifies currency movements

The negative response of hedging demand to the dollar creates a reinforcing feedback loop. When the dollar appreciates, funds hedge less, so dealers make fewer spot dollar sales, reinforcing the appreciation. When the dollar depreciates, funds hedge more, prompting additional dollar sales by dealer banks and reinforcing the depreciation. In our model, a negative demand elasticity in the forward market can therefore account for both the strong inverse relationship between hedging pressure and the dollar as well as a greater exchange-rate volatility. Derivatives are therefore not merely passive instruments used to insure exposures created elsewhere: through dealers’ spot-market transactions, changes in hedging demand can alter the equilibrium exchange rate and make the dollar rate more sensitive to shocks.

Implications and caveats

The findings suggest that models and policy assessments of exchange rates should track derivative positions alongside cash-market capital flows. They also reveal a less familiar consequence of the dollar’s international role. Foreign demand for dollar bonds supports the currency through the capital-flow channel, but the associated demand for currency protection can push in the opposite direction. Large dollar bond issuance therefore creates a correspondingly large hedging channel.

These findings should be interpreted with some qualifications. Although CLS provides unusually detailed data on global FX settlements, it does not capture the entire market, and our analysis is limited to a ten-year period. The causal interpretation of the instrumental-variable estimates also rests on the assumption that bank-specific capital changes influence funds’ hedging exclusively by altering dealers’ capacity to intermediate FX transactions. Nevertheless, the consistency of the evidence across currencies, frequencies, controls, and alternative instruments points to a clear conclusion: time variation in currency hedging is an important part of exchange-rate determination, and the response of hedging demand to past currency movements can itself be destabilizing. This means that we are starting to understand currency disconnect much better and can locate at least one of its sources.

Source : VOXeu