DTWEXBGS Explained: Meaning, Calculation, and 26-Currency Weighting

DTWEXBGS aggregates the dollar against 26 partner currencies weighted by bilateral trade flows and refreshed annually. The formula is written as weighted geometric averages, but it reads through trade flows.

This piece does not unpack what DTWEXBGS signals: that belongs to the cluster’s main article. It does one thing: expose the calculation mechanics, because without them, reading DTWEXBGS as a signal rests on a black box.

Per the Federal Reserve’s official documentation (Statistical Release H.10, January 2019 revision), DTWEXBGS, for Daily Trade Weighted Exchange Rate Broad Goods and Services, aggregates bilateral dollar moves against 26 partner currencies. Weights are refreshed annually, which separates DTWEXBGS from any fixed-composition construction. The detail is worked through in our analysis of dollar shortage episodes. The index baseline (100) is set at January 2006, the month the official series begins; it is computed every business day. The relevance of DTWEXBGS for analyzing the dollar in the global monetary system depends, first, on these methodological choices, which deserve to be made explicit.

Origin and the January 2019 methodological revision

DTWEXBGS in its current form took effect on February 4, 2019, as described in the Fed’s FEDS Note of January 15, 2019. Before that, the Federal Reserve published a family of separate indices: DTWEXB for the Broad index, DTWEXM for Major currencies, DTWEXO for Other Important Trading Partners. The fragmentation reflected an earlier distinction between developed-economy currencies and emerging-partner currencies, a distinction that became increasingly contested as the actual composition of U.S. foreign trade evolved.

The January 2019 revision unified the architecture. Three indices now coexist: DTWEXBGS for the full broad perimeter (26 currencies), DTWEXAFEGS for Advanced Foreign Economies (a subset of developed-economy currencies), DTWEXEMEGS for Emerging Market Economies (a subset of emerging-market currencies). All three use the same weighting methodology and the same January 2006 = 100 baseline, which makes the sub-indices directly comparable. The Fed continues to publish detailed documentation in Statistical Release H.10, publicly accessible and revised annually.

The revision mainly changed the weights. They used to rest on goods trade alone, blending 50% import weight, 25% export weight and 25% third-market competition; they are now each partner’s simple share of total bilateral trade, services included. Vietnam entered the basket and Venezuela left it. The calculation frequency did not change: the old broad index (DTWEXB) had been daily since 1995, and the Fed kept publishing it until December 31, 2019. DTWEXBGS remains a public statistic of the Federal Reserve, whereas the DXY is a commercial product of Intercontinental Exchange.

The weighted geometric formula

DTWEXBGS is not a simple arithmetic average of the 26 exchange rates. The Fed adopts a weighted geometric mean, the standard convention in exchange-rate index theory, because it avoids the biases associated with adding bilateral variations. Concretely, the index at time t is expressed as the product of bilateral exchange rates at t divided by their product at t-1, with each ratio raised to the power of the corresponding currency’s trade weight.

The practical consequence: a +5% move of the dollar against all currencies weighted equally produces exactly +5% on the index. Conversely, asymmetric moves combine according to weighted product roots, not according to a simple algebraic sum. This property explains why DTWEXBGS can remain relatively stable even when the dollar rises sharply against part of the basket and falls against another, a frequent phenomenon in regional divergence phases. A companion piece: our sub-pillar on the systemic dollar.

The geometric formula carries an analytical cost: it masks individual contributions. An observed DTWEXBGS variation does not, on first reading, indicate which currency drove the move. Disaggregation requires an independent pair-by-pair calculation. For macro analysis, that is an accepted limit; for currency-risk analysis of a portfolio, it is a reason never to use DTWEXBGS as a hedging proxy.

The 26 currencies in the 2026 basket

The 2026 DTWEXBGS composition, per the H.10 trade-weights table updated by the Fed on February 2, 2026, is as follows in descending order of weight: euro (21.0%), Mexican peso (14.8%), Canadian dollar (12.8%), Chinese yuan (10.9%), British pound (5.2%), Japanese yen (5.2%), South Korean won (3.6%), Indian rupee (3.4%), Taiwan dollar (3.0%), Swiss franc (3.0%), Vietnamese dong (2.6%), Singapore dollar (2.2%), Brazilian real (1.9%), Malaysian ringgit (1.4%), Australian dollar (1.4%), Thai baht (1.4%), Hong Kong dollar (1.1%), Israeli shekel (0.9%), Chilean peso (0.7%), Indonesian rupiah (0.7%), Colombian peso (0.7%), Swedish krona (0.6%), Philippine peso (0.6%), Saudi riyal (0.5%), Argentine peso (0.4%), Russian ruble (0.1%).

This weight structure prompts three observations. First, two emerging currencies, peso and yuan, alone account for 25.7% of the basket. Both are entirely absent from the DXY. Second, the five main emerging currencies (peso, yuan, won, rupee, real) add up to 34.6% of the basket, and the 19 emerging currencies together weigh 50.8%, more than the seven advanced ones (49.2%). DTWEXBGS therefore assigns more weight to emerging economies than to the euro taken individually. Third, the basket includes currencies of heterogeneous status: fully convertible, freely floating currencies (euro, yen, pound); currencies managed with discretionary interventions (yuan, real); and currencies long subject to exchange controls (Argentine peso). This heterogeneity follows from the Fed’s inclusion criterion, the share of U.S. bilateral trade (at least 0.5%): the index measures trade flows, not the institutional quality of FX markets. There is a limit, though: Venezuela was dropped in 2019 because its official rate no longer reflected the market. Gold detached from the greenback and its real behaviour sets this observation on a secular scale.

Annual rebalancing and its cyclical effect

The Fed updates DTWEXBGS weights roughly once a year from annual trade data published by the Bureau of Economic Analysis. The methodology is explicit: each partner country’s share of total U.S. exports plus imports, goods and services combined, excluding oil, gold and military equipment. The schedule is not fixed: the weights in use from March 24, 2025 to January 30, 2026 were replaced on February 2, 2026, and the 2024, 2025 and 2026 columns of the table differ by only a few hundredths of a point.

This revision produces a cyclical effect sometimes misunderstood by long-series analyses using the index. When U.S. trade pivots geographically (for instance, the China-to-Mexico rebalancing observed since 2018) the index gradually incorporates the shift. The yuan weight rose from 12.3% in 2006 to 16.0% in 2017, then fell back to 10.9% in 2026. The peso weight, conversely, has moved from 11.3% to 14.8% and has exceeded the yuan since 2022. Over the same period, smaller weights have clearly strengthened: the Indian rupee went from 1.5% to 3.4%, the Vietnamese dong from 0.3% to 2.6%.

Operational consequence: the official series starts in January 2006, and the Fed separately provides an estimate back to 1973 based on estimated services weights before 2006. In both cases each year carries its own weights, not a single weight applied retrospectively. This makes the series mathematically consistent with historical trade realities, but renders any raw level comparison between 1980 and 2026 partially biased by basket recomposition.

Broad, Advanced Foreign Economies, Emerging Market Economies: three sub-families

The DTWEXBGS family actually comprises three indices that the Fed has published in parallel since February 2019. DTWEXBGS itself covers all 26 currencies. DTWEXAFEGS isolates Advanced Foreign Economies, seven currencies: euro, Canadian dollar, yen, pound, Swiss franc, Australian dollar and Swedish krona (49.2% of the 2026 basket). DTWEXEMEGS isolates Emerging Market Economies, the other 19 currencies, from the Mexican peso to the ruble (50.8%).

This decomposition serves a precise analytical purpose: it answers the question “is the dollar rising against developed or emerging currencies?” In March 2020, for example, DTWEXBGS rose 4.1% from end-February to end-March, but this increase decomposes into +1.7% against AFE and +6.6% against EME, signaling specific emerging-market stress, invisible in a broad-only reading. The reverse also happens: in February 2023, the broad index’s 2.0% rise came more from advanced currencies (+2.5%) than from emerging ones (+1.4%). The practice is to read the three indices jointly and use their divergence as a cycle-configuration indicator. The difference with the narrower DXY resides precisely in this granularity: the DXY has no AFE/EME equivalent because its basket contains no emerging currencies. Readers consulting the DTWEXBGS dataset on Eco3min will also find the corresponding AFE and EME series in their downloadable form.

Why methodology changes the reading

Three methodological traits of DTWEXBGS condition what the index can reveal. The trade-flow perimeter limits its scope to dollar pressure on goods and services flows, not on financial flows, which represent the majority of FX market volumes. The geometric weighting masks individual contributions, requiring pair-by-pair disaggregation to identify the drivers of any move. Annual rebalancing introduces a commercial memory that distinguishes DTWEXBGS from any frozen index, but complicates level comparisons across long series.

These three traits are not flaws. They are explicit methodological choices that define what is asked of the index: measure dollar strength against actual trade partners, with dynamic weights and substantial emerging-market coverage. Any interpretation of DTWEXBGS that ignores these three traits drifts toward confusion.

Last updated — 7 October 2026