Skip to main content

Sensitivity to the Dollar (DXY Beta)

Updated Jun 24, 2026 at 2:35pm

Research Draft High 1,165 words

Dollar sensitivity ("DXY beta") measures how much a stock, sector, or index moves in response to swings in the U.S. dollar — most commonly proxied by the ICE U.S. Dollar Index (DXY). It is the regression coefficient of an asset's returns on dollar returns. The core tension is that this single number bundles together several distinct economic channels — currency translation of foreign earnings, competitiveness of exports, commodity pricing, and risk-on/risk-off capital flows — that don't always pull in the same direction. A measured beta is therefore a useful summary, but a fragile one: its sign and magnitude shift with regime, and a stable-looking beta can mask offsetting forces.

How it's calculated / formed

The standard estimate is a linear (or CAPM-style) regression of asset excess returns on the contemporaneous return of the dollar index:

R_asset − R_f = α + β_$ · R_DXY + ε

where β_$ is the dollar beta. Multi-factor versions add the market and other factors so the dollar coefficient is the residual sensitivity after stripping out broad equity beta (per the standard factor-model form Rp − Rf = α + β₁F₁ + … + βₙFₙ + ε; Meketa, Factor Exposure Analysis).

Practical conventions:

  • Index used. DXY is a geometric weighted average of the dollar against six currencies — EUR 57.6%, JPY 13.6%, GBP 11.9%, CAD 9.1%, SEK 4.2%, CHF 3.6% — with fixed weights, published by ICE (Wikipedia; Financer.com). It is heavily a EUR/USD trade and contains no emerging-market or Asian (ex-Japan) currencies. The Fed's broad trade-weighted dollar index is the more economically complete measure and is what most academic "dollar factor" work uses (Bruno, Shim & Shin, BIS WP No. 1000).
  • Window. Rolling windows are typical because the relationship is non-stationary; asset-pricing work often uses long (e.g. five-year) windows, while trading desks use shorter rolling betas to capture regime shifts — at the cost of noisier estimates.
  • Returns. Log returns; daily, weekly, or monthly. Higher frequency raises noise; lower frequency loses regime detail.

How to read it

A negative β_$ (the common case for U.S. large-caps, commodities, and EM) means the asset falls when the dollar rises. A positive β_$ means it rises with the dollar — typical of domestically focused, import-reliant, or safe-haven-substitute exposures. Magnitude matters: a beta of −0.5 implies roughly a half-percent move against a one-percent dollar move, all else equal. Because the channels differ, read the driver alongside the number, not the number alone.

How it's used in practice

DXY beta is mainly used for positioning tilts and risk attribution, not standalone signals:

  • Sector and style rotation. Energy and materials tend to carry strongly negative dollar betas (commodities are dollar-priced) and historically underperform in clear DXY uptrends, leading when the dollar rolls over. A sustained dollar breakout often favors domestic-revenue exposures — small-caps, regional banks, utilities — that earn almost entirely in dollars (Hartford Funds; Bitget Wiki). These are tendencies, not laws.
  • Earnings-translation overlay. The S&P 500 earns roughly 28–30% of revenue abroad by FactSet's geographic-revenue methodology (S&P's own methodology puts it higher, ~40%), and Information Technology is the highest-exposure sector at roughly 55–59% (FactSet GeoRev; Visual Capitalist). A widely cited practitioner rule of thumb holds that each ~10% move in the dollar shifts aggregate S&P 500 EPS by roughly 2–4% in the opposite direction (J.P. Morgan cites ~2% per 10%; Fed and sell-side estimates run toward ~3–4%), larger for international-heavy sectors — useful for framing currency as an earnings headwind/tailwind, but an approximation that varies by source and period, not a precise constant.
  • EM and global allocation. A stronger dollar tightens global dollar funding and is associated with lower EM equity returns; the broad dollar acts as a global risk factor more powerfully than bilateral rates (Bruno, Shim & Shin, BIS WP No. 1000).
  • Risk attribution. Funds decompose portfolio return into factor exposures (including a dollar factor) to know how much P&L is really a hidden FX bet.

Standing & evidence

The dollar's role as a global equity factor is well supported. The BIS work (Bruno, Shim & Shin, 2022) shows the broad dollar index helps explain cross-country stock returns, a stronger dollar correlates with lower returns, and EM indices with higher dollar beta earn higher average returns over time — i.e. dollar beta behaves as a priced risk factor: bearing dollar risk is compensated. That paper does not quantify the premium's magnitude, and the result is clearest for emerging markets.

For the U.S. broad market, the relationship is real but unstable. The dollar's inverse link to the S&P 500 is regime-dependent and not reliable as a timing tool over short horizons. The dollar-smile theory (coined by Stephen Jen at Morgan Stanley, ~2001) captures why: the dollar strengthens in both global-fear (risk-off, flight to safety) and U.S.-outperformance regimes, while weakening in stable risk-on growth — so the same DXY rise can coincide with falling or rising equities depending on which arm of the smile you are on (HeyGoTrade; BIS quarterly review). The classic dollar–oil and dollar–gold negative correlations are likewise real on average but have broken down for extended stretches (e.g. CME notes an "evolving" gold–dollar relationship). Treat measured base rates as conditional, not structural.

Strengths & limitations

Strengths. A compact, quantifiable summary of FX exposure; genuinely useful for risk attribution and for explaining why a portfolio's returns diverge from its sector thesis. Strongest and best-evidenced as a cross-sectional sorting variable for global/EM equities.

Limitations.

  • Non-stationarity. Betas drift and flip sign with regime; a backward-looking number can mislead at exactly the turning points that matter.
  • Channel confusion. One coefficient nets translation, competitiveness, commodity, and risk-flow effects that can offset — a near-zero beta may hide large opposing exposures.
  • Endogeneity / common driver. The dollar and equities often move together in response to a third force (rates, risk sentiment), so beta reflects co-movement, not a clean causal lever.
  • Index choice. DXY's heavy EUR weight and lack of EM currencies make it a poor proxy for the FX exposure that actually drives many firms; the broad trade-weighted index is more faithful.
  • Hedging. Many multinationals hedge translation/transaction risk, so realized stock sensitivity is smaller than revenue-geography suggests.

The single most common misuse: treating a recently estimated DXY beta as a stable, tradable timing signal for U.S. equities. The dollar-smile shows the sign itself is regime-conditional — using last year's beta to predict next month's move is the canonical error.

Sources

Dispute flags: the U.S.-equity ↔ dollar relationship is regime-dependent and contested as a timing tool; the ~2–4%-EPS-per-10%-dollar figure is a practitioner rule of thumb that varies by source (J.P. Morgan ~2%, others ~3–4%), not a precisely measured constant; the priced-factor result is strongest for EM, not U.S. large-caps.