Why USITC Rejected the Tomato Market-Change Evidence
USITC found that purchase-share shifts and scanner data did not show a meaningful change since 2019 in competition between U.S. and Mexican tomatoes. Publication 5762 shows how to build a stronger Section 751(b) record.
Primary lensTrade remedies
Sub-topicInjury determination
Evidence base8 records used
Use caseTrade-remedy exposure
USITC's full views on fresh tomatoes from Mexico explain where a familiar consumer-preference argument failed. The purchase mix changed. The record did not show that competition between U.S. and Mexican tomatoes changed meaningfully after 2019.
As USITC applied Section 751(b), the parties seeking revocation had to show either that a circumstance relevant to the original investigation had changed in a meaningful way, or that a new circumstance had arisen that might affect whether revocation would lead to continued or recurring material injury within a reasonably foreseeable time. Product differentiation without the right time comparison could not do that work.
The questionnaire record pointed the other way. Most responding producers, importers, and purchasers said cross-country interchangeability had not changed from 2019 through 2025. The Commission found that neither market-segmentation report established clear dividing lines between tomato categories. It separately found that the econometric study lacked transparency, left important endogeneity concerns unresolved, and did not provide data sufficient for replication.
USITC found no changed circumstance on the full record. It stopped before analyzing likely import volume, price effects, or industry impact after revocation. Counsel preparing the next request should first identify the original finding alleged to have changed, then build the data and economic work around that comparison.
Here, 2019 refers to the Commission's December 2019 affirmative threat determination and the conditions examined in that investigation. It is not the issuance date of the current order. The U.S. Department of Commerce Fresh Tomatoes from Mexico order, 90 FR 33363, took effect July 14, 2025, after termination of the 2019 Suspension Agreement.
Institution did not decide the merits
The Commission instituted the review because the submissions justified further inquiry. Institution did not establish that any asserted market change had occurred.
The vote itself makes the distinction concrete. Commissioner Kearns voted to institute, Chair Karpel voted not to institute, and Commissioner Johanson did not participate. The statutory rule for an evenly divided Commission produced an affirmative institution decision. After the record was developed, Kearns and Karpel both found no changed circumstance sufficient to warrant revocation.
19 U.S.C. 1675 changed-circumstances review places the burden of persuasion on the party seeking revocation. USITC described institution and revocation as different thresholds. Enough information to investigate is not enough evidence to terminate an order. Its discussion of the legislative history also starts from the existing order, rather than requiring the domestic industry to justify the order again.
The USITC Publication 5762 detailed views identified two ways to meet the threshold. The requester could show that a circumstance relevant to the original investigation changed meaningfully. Or it could show that a new circumstance arose that may affect whether revocation would lead to continued or recurring material injury within a reasonably foreseeable time.
The Commission found no changed circumstance and stopped before the likely-volume, price-effect, and impact analysis. A future requester must establish the change itself before trying to prove what revocation would do.
The headline shift became smaller after the categories were reconciled
Round tomatoes lost 5.6 percentage points of total purchase share during the review period, while Roma and plum tomatoes gained 5.4 points. On their own, the aggregate figures suggested movement from mature-green product toward vine-ripe and specialty product.
The category bridge showed what the headline left out. The share of mature-green rounds fell 2.6 points, while the share of vine-ripe rounds fell 3.0 points. Both maturity groups contributed to the decline in rounds. Within Roma and plum tomatoes, the mature-green share was stable and vine-ripe product gained 5.4 points. The Commission therefore treated 2.6 points as the maximum purchase-share movement that could support the claimed shift from mature-green rounds to vine-ripe Roma and plum tomatoes. This was a ceiling derived from category shares, not a measured flow of individual buyers.
Aggregate signal
Reconciled record
Evidentiary consequence
Rounds down 5.6 points
Mature-green rounds down 2.6 and vine-ripe rounds down 3.0
The decline was not confined to mature-green product
Roma and plum up 5.4 points
The gain came from vine-ripe Roma and plum while mature-green Roma and plum stayed stable
Category growth did not identify which origin lost a sale
Varieties differed in taste, quality, and use
Producers in both countries sold the major varieties through overlapping channels
Product differentiation did not establish a break in competition between origins
Those figures put a limit on what the product-mix trend could establish. They did not show whether U.S. and Mexican tomatoes competed differently than they had in 2019.
The questionnaire record did not show a change from 2019
USITC asked firms about that comparison. Fourteen of 15 responding U.S. producers, 77 of 81 importers, and 21 of 23 purchasers reported that interchangeability between U.S. and Mexican tomatoes had not changed from 2019 through 2025. Thirteen of 16 producers, 77 of 81 importers, and 19 of 22 purchasers reported no change in interchangeability by growing method.
These counts are not market-share estimates, and they do not make every tomato substitutable for every use. The Commission acknowledged limits to substitution among types. Their value came from the comparison over time. The responses addressed the relationship the statute required the Commission to examine, while the purchase-share numbers described the product mix.
Domestic producers and importers also supplied vine-ripe and mature-green rounds, Roma and plum tomatoes, and cherry and grape tomatoes across growing environments and package types. Sales from both origins moved mainly through distributors. Some varieties could occupy distinct commercial niches without severing competition between domestic and subject supply.
Some specialty and commodity categories occupied different commercial niches, but suppliers from both origins continued to compete within them.
The econometric study could not fill the gap
The parties seeking revocation offered two reports by economist Timothy Richards. One used retail scanner data, cross-price elasticities, the law of one price, and cluster analysis to examine substitution among tomato categories. The other used several econometric approaches to argue that specialty tomatoes did not compete with commodity tomatoes.
USITC found that neither report showed clear dividing lines among distinct fresh tomato product types. Its concerns about transparency, endogeneity, the observation count, and replication applied to the second report, the econometric study.
Price and quantity are often determined together. When a demand model treats price as if it moved independently, it can mistake a response to demand for a cause of demand. USITC said the demand equations did not instrument for price and the wholesale model did not address the endogeneity of total market volume. It questioned an unconventional variable called `MX Below US` because the study did not provide a clear theoretical reason for including it. The Commission also noted that numerous controls might absorb variation needed to identify the effects of interest, that the observation count was limited, and that the study did not provide data sufficient for replication.
The Commission's criticisms were confined to the two studies before it. For future filings, the practical point is narrower. Causal assumptions and results that cannot be tested carry less weight, even when a study presents its result as statistically certain.
The substitution study covered retail prices from March 2024 through February 2026. That recent data window could describe relationships among tomato categories. For this Section 751(b) inquiry, it did not by itself furnish a matched comparison with the conditions examined in 2019. This is an inference from the study period and the Commission's stated changed-circumstances test, not a separate finding announced in the report.
The model critique addressed one of several alleged changes. USITC separately examined greenhouse investment, U.S. growers' operations in Mexico, Canadian supply, and structural conditions in the United States and Mexico. It found that the alleged changes, separately and together, did not establish changed circumstances. A reproducible market model could still fail if it did not show how competition between domestic and subject supply had changed.
Eight records for a market-boundary filing
For counsel and economists, the weaknesses identified in Publication 5762 translate into eight records that should be available for inspection.
Record
Minimum question it should answer
Failure it prevents
Claim map
Is the theory a meaningful change to an original circumstance or a genuinely new circumstance?
Applying a historical-comparison demand to the wrong statutory theory
Baseline map
How were comparable products, origins, growing methods, packages, and channels classified in 2019 and now?
Calling a continued trend a new circumstance
Category bridge
Which share movement is observed, and which movement is only the maximum allowed by aggregate data?
Converting composition into an unsupported switching claim
Data provenance and access
What extracts, filters, exclusions, transformations, licenses, and BPI controls produced each series?
A result that cannot be traced or lawfully reviewed
Identification and order-effect memo
How does the design address joint price and volume determination, and separate market change from effects caused by the order?
Biased estimates or an order-induced change presented as an external circumstance
Replication archive
Can an authorized reviewer rerun the code, recover the tables, and identify the data version?
A result that cannot be tested within data-access limits
Overlap and response matrix
Which suppliers, products, seasons, and channels overlap by origin, and who answered the questionnaires?
Product difference without origin separation, or unexplained coverage gaps
Legal bridge
Which original finding changed, and how could that change affect volume, price effects, or impact after revocation?
A market fact with no consequence under Section 751(b)
Because the study window crossed July 14, 2025, when the current order took effect, the design should separate external market change from any response to the remedy. Crossing the effective date alone does not invalidate the study.
The response matrix should identify firms' market roles, coverage, affiliate status, and reasons for material nonresponse. Retail modeling and questionnaires measure different things. Counsel should reconcile both to the same historical comparison and the same claim about competition between origins.
Teams often begin a changed-circumstances project with the freshest available data. The tomato decision points to an earlier first step. Counsel should identify which finding from the original determination the new evidence is meant to displace. The data team can then preserve the old and current taxonomies, the economist can connect the two periods, and procurement can recover bids, refusals, substitutions, and supply constraints for the relevant product-origin pairs.
The public report redacts business proprietary information, and its full-year firm data run through 2025. A different result could follow from a later matched dataset, a reproducible design that addresses endogeneity, or evidence that product and channel overlap between origins has disappeared. A court remand or another agency proceeding could also change the posture. Publication 5762 addresses only the record before the Commission.
For a request based on the consumer-preference theory presented here, counsel still has to show when the change occurred and how it altered competition between domestic and subject supply. When the showing relies on modeling, transparency and reproducibility affect the weight USITC gives the result. In this review, that weight was not enough.
From reading to review
Run the numbers on your lane.
The duty calculator runs the current stack for any HTS code and origin. A free account opens full tool output, AD/CVD detail, Chapter 98 processing, and available exports.