The Four Filters Between a Currency Depreciation and an Actual Export Gain

The Federal Reserve's nominal broad dollar index averaged 120.60 in July 2026, against 128.84 in January 2025 — a decline of roughly 6.4 percent over eighteen months on an index where January 2006 equals 100. The yen has gone the other way, averaging 162.33 per dollar in July 2026 against 158.68 in March. Two currencies, two directions, and in both cases the same reflex appears: a cheaper currency makes exports cheaper, cheaper exports sell better, so the trade balance improves.

The reflex is not wrong. It is conditional, and the conditions are specific enough to check. The International Monetary Fund's own work — the most cited defence of the textbook channel — estimates that a 10 percent real effective depreciation is associated with a rise in real net exports averaging 1.5 percent of GDP. That is a real effect and it should not be waved away. But the same estimate carries a cross-country range of 0.5 to 3.1 percent of GDP, a spread of more than six to one. When the dispersion around an average is that wide, the average is the least interesting number in the table. The interesting question is what sorts an economy toward the bottom of the range rather than the top.

Four structural filters do most of that sorting, and each is observable in published data before the trade figures print.

The Surface Reading and What It Quietly Assumes

The standard chain of reasoning has four links. The currency falls. Foreign-currency prices of the country's exports fall with it. Foreign buyers respond to the lower price by buying more. Volumes rise, and the trade balance improves. Each link is an empirical claim, and each has a measured elasticity attached to it.

The IMF's estimates for the first two links are the ones worth memorising. Long-run pass-through from the exchange rate into foreign-currency export prices runs at an elasticity of about 0.552 — a 10 percent depreciation lowers export prices in foreign currency by roughly 5.5 percent, not by 10. The corresponding import-price elasticity is about 0.605, so a 10 percent depreciation raises domestic-currency import prices by around 6.1 percent. The third link, price to volume, is weaker still: a 10 percent rise in export prices reduces export volumes by roughly 3 percent in the long run, an elasticity near 0.32. Multiply the links together and a 10 percent depreciation delivers something on the order of a 1.8 percent export volume gain through the pure price channel, most of it arriving in the first year but not all of it.

That is the mechanism working as designed: a real but modest effect, and the product of three fractions, none of which is one. Anything pushing any of the three toward zero collapses the chain. The four filters below are the things that do so.

Filter One: The Invoice Is Not Written in the Currency That Moved

The textbook chain implicitly assumes exports are priced in the exporter's own currency, so that a domestic depreciation mechanically lowers the foreign price. Very little of world trade works that way. Federal Reserve research covering annual averages from 1999 to 2019 finds that the dollar accounted for 96 percent of export invoicing in the Americas, 79 percent in the rest of the world, and 74 percent in the Asia-Pacific region. Europe is the single exception, where the euro dominates at 66 percent. Outside Europe, the overwhelming majority of trade invoices are denominated in a currency that belongs to neither the buyer nor the seller.

When an exporter in a third country prices in dollars, a depreciation of that exporter's own currency does not lower the price the foreign buyer sees. The dollar price is sticky in the short run; what changes is domestic-currency revenue, which rises. The margin widens. The buyer's incentive to order more does not change, because from the buyer's side nothing happened.

This is testable. Research on dominant-currency pricing using data covering 91 percent of world trade finds non-commodity terms of trade essentially uncorrelated with exchange rates, and the dollar exchange rate quantitatively dominating the bilateral rate in both pass-through and trade-elasticity regressions — with the effect increasing in the share of imports invoiced in dollars. The bilateral rate, the one most commentary watches, is close to the wrong variable.

The same body of work produces the most direct number in this whole discussion: a 1 percent appreciation of the US dollar against all other currencies predicts a 0.6 percent decline within a year in the volume of total trade between countries in the rest of the world, controlling for the global business cycle. Trade that has nothing to do with the United States contracts when the dollar rises. That is not a competitiveness effect in any conventional sense.

Filter Two: The Export Basket Is Full of Imports

The second filter is arithmetic. If inputs are purchased abroad, a depreciation raises their domestic-currency cost at the same moment it raises the domestic-currency value of the sale. The net gain is the difference between two moves in the same direction, not a windfall.

The scale has changed materially over fifty years. Foreign value added embedded in gross exports — backward participation, in global value chain terminology — rose from roughly 15 percent in the 1970s to about 25 percent by 2013, though the increase has since decelerated. A quarter of the average export, by value, was bought from somewhere else. For assembly-specialised economies the figure runs well above that.

A 10 percent depreciation lowers foreign-currency export prices by about 5.5 percent, but if a quarter of the input bill has just become 6.1 percent more expensive in domestic currency, much of that price cut comes out of margin rather than from a genuine cost advantage. Exporters facing that squeeze tend not to cut foreign prices at all. Filters one and two therefore align: both point toward stable foreign-currency prices and unchanged volumes.

Filter Three: Capacity Has to Exist Before Volume Can Move

Suppose the first two filters clear. Prices genuinely fall in the buyer's currency, foreign demand genuinely rises, and orders come in. The volume response still requires that somebody can make more of the thing. That is a physical constraint, and it is measured monthly.

United States total industry capacity utilisation ran at 76.3 percent in July 2026, which the Federal Reserve's own release describes as 3.1 percentage points below its long-run 1972-2025 average of 79.4 percent. Manufacturing specifically ran at 76.0 percent against a long-run average of 78.2 percent, a gap of 2.2 percentage points. Readings below the long-run average mean headroom exists, the condition under which a depreciation transmits most cleanly. The IMF finding is explicit on this point: the export response to depreciation is strengthened when economic slack is present and dampened during banking crises.

The inverse case is the one that catches people out. An economy running at or above its own long-run utilisation average cannot convert an order book into shipments on the timescale over which the currency move is being discussed. New capacity takes quarters to years, and the currency can reverse before the plant opens. The honest form of the question is not "did the currency fall" but "did it fall while there was slack" — and the second version has a numerical answer before the trade data arrives.

Filter Four: Depreciation Arrives as a Credit Event Before It Arrives as a Price

The fourth filter is least represented in the standard framing, and it runs opposite to the other three. Exporting is credit-intensive: goods are produced before they are paid for, shipments cross borders under trade finance, and firms in long production chains carry large working capital balances. Much of that credit is dollar-denominated regardless of where the firm sits.

Bank for International Settlements research examining 4.6 million observations of export shipments finds that following dollar appreciation, exporters relying on dollar-funded bank credit suffer a greater decline in credit and a corresponding slowdown in exports. Banks with heavy reliance on dollar wholesale funding cut dollar credit supply more sharply than less-exposed banks when the dollar strengthens. Exporters with higher working capital needs and longer production chains contract most severely. The effect holds even controlling for non-credit factors.

This is a competitiveness story turned inside out. The trade channel says a weaker home currency against the dollar helps exporters; the financial channel says a stronger dollar tightens the credit those same exporters need to make the goods. Both fire simultaneously and point opposite ways. Which dominates depends on funding structure — visible in bank balance sheet data long before it is visible in customs data.

What the Market Tends to Miss

Three specific misreadings recur, and each has a cheap diagnostic.

Values are mistaken for volumes. Export values in domestic currency rise almost mechanically after a depreciation, because the same foreign-currency receipts translate into more local currency. A headline of record exports in domestic-currency terms is compatible with flat or falling shipments. Volume and value indices can diverge for years, and only one is evidence about competitiveness.

The import side moves faster than the export side. The import-price elasticity is higher than the export-price elasticity — 0.605 against 0.552 in the IMF estimates — and the pain lands on domestic buyers immediately, while any export gain accumulates over quarters. The trade balance can deteriorate before it improves, a pattern old enough to have its own name, and the timing gap is where policy credibility is usually lost.

The broad dollar index is read as a bilateral relative price when it functions as a global financial variable. As of end-2022 the dollar comprised 58 percent of disclosed global official reserves, down from 71 percent in 2000; it was on one side of roughly 88 percent of global foreign exchange transactions in April 2022; and around 70 percent of foreign-currency debt has been dollar-denominated since 2010. A variable sitting inside reserve management, funding markets and debt issuance at once does not behave like a simple relative price.

The Scenario That Undoes This

The argument above is a set of conditions, not a law. Treating it as a general rule would be the same error in the opposite direction.

Large, sustained depreciations do work. The IMF's evidence is that after depreciations averaging 25 percent in real effective terms, export volumes rise about 10 percent over five years. That is a substantial response and it is measured, not theoretical. The filters described here bite hardest on moves of 3 to 8 percent over a few quarters — the size of move that dominates commentary and is mostly noise. A 25 percent multi-year realignment is a different object and should be treated as one.

Commodity exporters are a genuine exception. The finding that terms of trade are uncorrelated with exchange rates was specifically about non-commodity trade. Commodities are priced in dollars on global markets, so a producer country whose currency falls against the dollar sees domestic-currency revenue rise directly, and the cost base is largely domestic. The transmission there is close to the textbook version.

Services exports respond faster than goods. Tourism, education and locally delivered services are priced in the destination currency by construction, carry little imported input content, and often have spare capacity within a season rather than within years. Three of the four filters are weak or absent, which is why services balances frequently move on depreciation while goods balances do not.

Europe is invoiced differently. With the euro at 66 percent of regional export invoicing, intra-European trade is much closer to producer-currency pricing than the global pattern suggests. The dominant-currency argument is strongest outside Europe and weakest within it.

Slack changes the answer. An economy with utilisation several percentage points below its long-run average, no banking stress, and a domestically sourced export basket can produce a textbook response. That combination is uncommon but it is not hypothetical, and the components are all published.

What to Watch Next Week

Five releases speak to the filters directly, each on a fixed cadence.

• Federal Reserve H.10, weekly. The broad dollar index is the single most informative variable for the fourth filter. The reference points worth holding: 128.84 at the January 2025 monthly peak, 117.91 at the February 2026 monthly low, 120.60 in July 2026.
• Federal Reserve G.17, mid-month. Capacity utilisation against the stated long-run average of 79.4 percent for total industry and 78.2 percent for manufacturing. The gap, not the level, is the signal.
• National customs releases, monthly. The specific thing to extract is the volume index separately from the value index. Where an agency publishes only values, the series cannot answer this question.
• Cross-border banking statistics, quarterly. Dollar-denominated claims on non-bank borrowers outside the United States are the closest published proxy for the credit channel in the BIS research.
• Producer and export price indices, monthly. Pass-through is testable in near real time by comparing the export price index in foreign-currency terms against the exchange rate move over the same window. A ratio well below 0.55 indicates exporters are holding price and taking margin.

Concrete Framework — How to Verify It Yourself

A five-step check, in order. Any step can terminate the analysis.

1. Establish the size and persistence of the move. Measure in real effective terms, not against the dollar alone, and over a window of at least four quarters. Anything under roughly 10 percent sustained is unlikely to clear the filters. Compare against the 25 percent benchmark associated with a measured 10 percent volume response over five years.
2. Identify the invoicing currency of the relevant export sector. If it is the dollar and the country is outside Europe, assume the regional base rate — 74 to 96 percent depending on region — and expect foreign-currency prices to be sticky. The bilateral exchange rate is then largely uninformative about volumes.
3. Estimate imported input content. Take 25 percent as the global reference for foreign value added in gross exports and adjust upward for assembly-heavy sectors, downward for resource extraction and services. Subtract the input-cost effect from the headline price advantage before treating any of it as a competitiveness gain.
4. Check capacity against its own long-run average. Not against 100 percent, and not against another country. The relevant comparison is the same series' own historical mean, which statistical agencies publish alongside the current reading.
5. Check the direction of the credit channel. If the depreciation coincides with broad dollar strength, the financial channel is working against the trade channel for any exporter with dollar-denominated working capital. Where the two channels conflict, the funding structure of the export sector decides which one prints in the data.

The output is a probability judgement, not a forecast. A move clearing all five steps has a reasonable chance of showing up in volumes within a year. A move failing step two or step three most likely shows up in exporter margins and domestic-currency export values, and never reaches volumes at all. That distinction is answerable from published data before the trade balance prints.

This article is macroeconomic analysis, not investment or financial advice. Figures cited are drawn from published statistical releases as of August 2026 and are subject to revision.