Trade & Sourcing
Nearly One-Fifth of Structural Steel’s Country Mix Changed. Its Concentration Score Barely Moved.
Between complete 2024 and 2025 boards, fabricated structural steel has 18.67% country-share turnover while its HHI moves only from 1,021.99 to 1,014.45.
Editorial evidence cutoff: September 9, 2026. Published September 29, 2026. Observation periods are stated throughout; older figures are retrospective evidence.
Country concentration is often used to judge progress on sourcing diversification. For a September 2026 review, the completed 2024 and 2025 records expose a limitation in that practice: an almost unchanged score can conceal substantial change in the underlying distribution. This retrospective uses only the full country boards available by September 9, rather than claiming a new movement occurred that month.
Fabricated structural steel’s country concentration score barely changed between 2024 and 2025. Its Herfindahl-Hirschman Index moved from 1,021.99 to 1,014.45, a decline of only 7.54 points. A dashboard built around that one measure could reasonably look quiet.
The country mix was not quiet. The distance between the two complete origin distributions was 18.67%, measured as half the sum of the absolute changes in country shares. Nearly one-fifth of the value-share distribution was reallocated in this statistical sense, while the concentration score hardly moved.
The finding exposes a limitation in what concentration can tell a purchasing team. It describes how unevenly a market’s shares are distributed. It does not preserve the identities of the countries holding those shares.
A STABLE SCORE CAN HIDE A DIFFERENT MAP
Imagine two origin countries exchanging equal shares. A measure based only on the size distribution can remain unchanged even though the identities behind those shares have switched. Structural steel presents a real, less tidy version of that problem: large changes in some country shares coexist with a small net change in concentration.
The analysis uses the full 2024 and 2025 country boards for 57 tracked manufacturing families from the saved USITC DataWeb snapshot. Structural steel has 98 reporting origins in the first year and 91 in the second. Their customs values sum exactly to the family totals in both years.
That complete denominator is essential. A top-supplier table can make a remaining country appear more important simply by omitting the tail. Here, no residual share is silently redistributed among the largest sources. Each country’s weight is its share of the entire observed family value.
TWO MEASURES ASK DIFFERENT QUESTIONS
HHI is the sum of squared shares, multiplied by 10,000 when shares are expressed as fractions. Squaring gives larger sources more influence. The Justice Department’s explanation describes the arithmetic for firms; applying the same formula to origin countries produces a country-concentration statistic, not a judgment about competition between companies.
Turnover uses a different operation. For each country, subtract its earlier share from its later share, take the absolute value, sum the changes and divide by two. The division avoids counting the same redistribution twice, as a loss somewhere and a gain elsewhere.
For structural steel, that measure is 18.67%. It should not be described as proof that 18.67% of orders changed suppliers. The underlying observations are nominal customs values. Prices and product composition can change country shares even when a buyer keeps the same supplier. The statistic measures the distance between two distributions, with that distinction preserved.
SIMILAR TURNOVER, OPPOSITE CONCENTRATION RESULTS
Semiconductors have almost the same turnover, 18.76%, but their HHI rises from 1,108.90 to 1,248.17. The largest reported origin changes from Malaysia to Taiwan. Here, substantial movement in country shares accompanies a more uneven distribution.
Steel long products show slightly greater turnover, 19.90%, alongside an HHI decline of 450.17 points. Canada remains the largest origin. A leading country can keep its position while the rest of the distribution changes enough to make the overall sourcing pattern less concentrated.
Structural steel sits between these cases: substantial turnover with almost no net concentration movement. Looking across all 57 families prevents the reader from assuming that turnover necessarily means diversification or concentration. The direction depends on which shares rose, which fell and their starting sizes.
REMOVE THE TINY ORIGINS. THE RESULT REMAINS.
Country counts can be noisy at the small end. A tiny shipment can add an origin to the board without materially changing the distribution. To test whether the structural-steel result comes from that tail, the calculation was repeated after retaining only countries with at least 0.1% of value in either year and then renormalizing their shares.
Thirty-nine countries remain, covering 99.27% of 2024 value and 99.18% of 2025 value. Turnover is still 18.56%. HHI still barely changes, declining by 5.73 points in the restricted distribution. Removing the small origins does not make the apparent contradiction disappear.
The same check leaves semiconductor turnover at 18.77%. Steel long-product turnover remains 19.77%. These sensitivity results support the comparison without pretending that the 0.1% threshold is uniquely correct. Both the full-board and restricted calculations should be available beside the chart.
CONCENTRATION IS NOT A DISRUPTION FORECAST
A country-level HHI does not identify how many independent mills, factories or corporate groups supply the goods. Multiple suppliers can share an origin; one supplier can operate across several origins. The trade board cannot resolve that ownership and capacity structure.
Nor does a lower HHI automatically establish lower disruption probability. Transport routes, common upstream inputs, energy exposure and correlated weather risks can connect origins that appear separate in customs data. Merger-review thresholds for firms should not be pasted onto these country scores as supply-risk cutoffs.
The investigation’s narrower conclusion is useful enough: a concentration number can remain stable while the composition behind it changes substantially. That is a measurement finding, not an assertion that structural steel became safer, riskier or cheaper in 2025.
KEEP THE SCORE. ADD THE COUNTRY CHANGES.
A better sourcing dashboard would show the concentration level, its change and the largest origin-share gains and losses together. Turnover provides a compact signal that the country map deserves inspection, even when HHI itself has barely moved.
For an actual purchasing portfolio, the next layer is item-level evidence. Were specifications unchanged? Did a distributor change origin? Are two nominally independent sources dependent on the same upstream mill? These questions require purchase records and supplier information that the national country table does not contain.
The existing structural-steel story follows China’s share decline and Mexico’s leadership over a longer period. This complete two-year comparison asks a different question: how much can change behind a stable summary score? The answer is enough that a nearly unchanged number should not end the review.
Source window: calendar 2024–2025, complete country boards only. HHI and turnover are recomputed from unrounded customs-value shares. Values refer to imports for consumption and do not establish physical quantities or causal supplier relocation.
Sources and evidence
Evidence period: Calendar 2024 versus 2025. The frozen evidence record lists the source files and verified hashes available September 9, 2026. Source revision: 5e4fb7726c3d40060c0151c086c903baae856cab. Later live-data updates do not alter the historical evidence in this article.
justice.gov/atr/herfindahl-hirschman-index
census.gov/foreign-trade/guide/sec2.html
Published 2026-09-29.