The Design and Effect of Tariff Retaliation: Evidence from the European Union

Board of Governors of the Federal Reserve System International Finance Discussion Papers ISSN 1073-2500 (Print) ISSN 2767-4509 (Online) Number 1436 March 2026 The Design and Effect of Tariff Retaliation: Evidence from the European Union Ece Fisgin, Johannes Fleck, Keith Richards Please cite this paper as: Fisgin, Ece, Johannes Fleck, Keith Richards (2026). “The Design and Effect of Tariff Retaliation: Evidence from the European Union,” International Finance Discussion Papers 1436. Washington: Board of Governors of the Federal Reserve System, https://doi.org/10.17016/IFDP.2026.1436. NOTE: International Finance Discussion Papers (IFDPs) are preliminary materials circulated to stimulate discussion and critical comment. The analysis and conclusions set forth are those of the authors and do not indicate concurrence by other members of the research staff or the Board of Governors. References in publications to the International Finance Discussion Papers Series (other than acknowledgement) should be cleared with the author(s) to protect the tentative character of these papers. Recent IFDPs are available on the Web at www.federalreserve.gov/pubs/ifdp/. This paper can be downloaded without charge from the Social Science Research Network electronic library at www.ssrn.com.

The Design and Effect of Tariff Retaliation: Evidence from the European Union* EceFisgin† JohannesFleck‡ KeithRichards§ March16,2026 Abstract WeshowthattheEU’s2018retaliationagainstUSsteelandaluminumtariffstargetedgoods with low US import dependence and high substitutability. For the majority of tariffed goods, the US share of EU imports declined notably and remained below pre-2018 levels even after the retaliatory tariffs were lifted, reflecting asymmetric effects of tariffs on tradediversion. Moreover, althoughtheretaliatorytariffswereinstantlyandfullypassed throughtoEUimporters,theretaliationdidnotleadtodomesticpricepressuresaswefind noevidenceforinflationaryeffectsonconsumerandproducerprices. Keywords: Trade Wars, Retaliatory Tariffs, Tariff Pass-Through, Inflation, US-EU Trade Relations JELClassification: E31,F13,F14,F42,F62 *WethankseminarparticipantsattheFederalReserveBoardforcomments.WearealsothankfultoHarunAlp, JustinPierce,MarianoSomaleandFrancoisdeSoyresforhelpfulconversationsandespeciallytoAaronFlaaenfor detailedcomments. TheviewsexpressedhereinarethoseoftheauthorsandnotnecessarilythoseoftheBoardof GovernorsoftheFederalReserveSystem. †BoardofGovernorsoftheFederalReserveSystem;Ece.Fisgin@frb.gov ‡BoardofGovernorsoftheFederalReserveSystem;Johannes.Fleck@frb.gov §BoardofGovernorsoftheFederalReserveSystem;Keith.P.Richards@frb.gov

1 Introduction The recent sharp US tariff hikes represent a historic departure from the decades-long trend of global trade liberalization. In response to such tariff increases, it is natural to ask what the effectsofforeignretaliationmightbe. Whiletherewereretaliatorythreatsfromothereconomies earlyonduringthecurrentandongoingepisode,theylargelyeased,partlyinresponsetobilateraltradedeals. Butthesetrucesarefragileasthedealsareoftenincompleteandcommitments uncertain, most notably regarding foreign investment pledges. More generally, if there was a breakdown of the global trading order due to ever-stronger protectionist policies, it would likelyresultinsubstantialretaliation. Despite this rising relevance of tariff retaliation, there is only scant evidence regarding its impactontheretaliatingeconomies. Indeed, thefocalpointoftherecentliteratureontheeffects of tariffs is their impact on the US economy. This evidence is not informative about potential retaliation and its effects on the retaliator. Moreover, the design of retaliatory tariffs can take many shapes. Therefore, understanding the effects of retaliation on the retaliator necessarily requiresexaminingparticularepisodesofretaliation. Our paper closes this gap in the literature by assessing the trade destruction and domestic price effects of the European Union’s (EU) 2018 tariff retaliation package. This package was compliantwithWorldTradeOrganization(WTO)provisionsandrepresenteda“rebalancing” response to US steel and aluminum tariffs. Between June 2018 and January 2022, it applied 10 and 25 percent ad-valorem duties on a $3.3bn list of US imports ranging from steel and aluminum to consumer goods like bourbon, motorcycles, and apparel. With this selection of products, the EU sought to minimize domestic economic disruptions while maximizing pressureontheUStorescindthesteelandaluminumtariffs. Standardtradetheorysuggeststhattheincidenceoftariffs,whethertheyarebornebyexporters through lower pre-tariff prices or by importers through higher post-tariff prices, depends on therelative,product-specificelasticitiesofsupplyanddemand. Hence,iftheEU’sdemandfor thetariffedUSimportswasinelastic,orifEUimporterscouldnoteasilyfindsubstitutes,either from domestic or third-country suppliers, the retaliatory tariffs would have materialized as a taxonEuropeanbusinessesandconsumers,resultingininflationarypressures. Weassessthese mechanismsfromanempiricalperspectivebystudyingtheevolutionofimportquantitiesand pricesoftariffedandnon-tariffedproductsinadifference-in-differenceframework. Our key findings are as follows: First, the retaliatory tariffs had a lasting destructive effect on 2

USimportstotheEU.Theimportvalueofthetariffedproductsfelltoabouthalffromthesecond month after the tariffs were implemented. Moreover, tariffed US imports did not recover fullyaftertheretaliatorytariffsweresuspendedbutremainedpersistentlybelowthepre-tariff levels. Our estimate of the import demand elasticity to tariff suspension is just slightly larger than half of that to tariff implementation, demonstrating that both had asymmetric effects. Thus, asliftingthetariffsfailedtoreversetheearlierdeclineinUSimports, theretaliationleft permanentscarsinUS-EUtradeflows. Second, US exporters did not lower their pre-tariff prices to absorb the costs of the retaliatory tariffs. Instead,theywererapidlyandcomprehensivelypassed-throughintothe“at-the-dock” importpricesofEuropeanimporters. ThisfindingisconsistentwithAmiti,Redding,andWeinstein(2019),Fajgelbaum,Goldberg,Kennedy,andKhandelwal(2020)andCavallo,Gopinath, Neiman, and Tang (2021) for US tariffs and emphasizes that, even if tariffs target comparably smallimportvaluesandalargevarietyofproducts,theyarestillassociatedwithhighdomestic pricepass-throughintheshortrun. However, despite their immediate passthrough, we find that the retaliatory tariffs did not lead to inflationary pressures in the EU. We arrive at this conclusion by regressing changes in category-specific tariff rates on measures of excess inflation over consecutive month windowsasotherresearchers,forexampleMintonandSomale(2025)havedonetotracktheprice effectsofUStariffs. Wedososeparatelyforproducerandconsumerprices. We rationalize our findings on trade and price effects by highlighting several features of the retaliationpackage. First,itwasmoderateinsize,applyingtoonly1.5percentofallUSgoods imports to the EU, and to be rolled out in two stages. As the second stage never went into effect,thetariffedimportvolumewaslessthanhalfoftheUSsteelandaluminumtariffs. Based on 2017 import figures and abstracting from any substitutions or import price reductions, the cumulative tariff cost would have amounted to approximately $2.9bn during the 3.5 years in whichtheretaliationwasactive–comparedtoabout$7trnintotalEUimports. Thus,thetariff costshadlittleefficacytonoticeablydriveupaggregatedomesticprices. Second, the retaliation package was asymmetric and distributed potential tariff costs over a varietyofproductsasitdidnotexclusivelyapplytosteelandaluminumimports. Mostofthese categorieshadalowUSshareofEUimports,reflectinglittleimportdependenceandenabling EU importers to quickly and efficiently source tariffed US imports from cheaper alternative suppliers. 3

Ourfindingscontributeanewperspectivetoagrowingbodyofliteratureonthe2018-2020and theongoing2025tradewars. MostofthesestudiesfocusontheimpactofUStariffsondomestic prices, product varieties and production relocation choices, such as Amiti, Redding, and Weinstein(2019),Cavallo,Gopinath,Neiman,andTang(2021),Flaaen,Hortac¸su,andTintelnot (2020),andFlaaenandPierce(2024). Somealsoassesstheeffectsofforeignretaliationonprices and quantities of US exporters, especially for the agricultural sector (Fajgelbaum, Goldberg, Kennedy, and Khandelwal (2020), Grant, Arita, Emlinger, Johansson, and Xie (2021) Morgan, Arita, Beckman, Ahsan, Russell, Jarrell, and Kenner (2022)). Our work is a counterpoint to thesestudiesasitprovidesresultsonthetradedisruptiveanddomesticinflationaryeffectson theretaliatingeconomy. Thereiscurrentlyonlylittleandnarrowevidenceregardingthe“self-harm”effectofretaliatory tariffs. For China, Chor and Li (2024) and Li, Balistreri, and Zhang (2020), provide estimates on how retaliation against US tariffs in 2018/2019 reduced activity and increased production costs. Forthe2018EU’sretaliation,thefewavailablestudies,suchasFetzerandSchwarz(2021) and Braml (2020) focus on the political motives behind the design of the retaliation package.1 AstheretaliationpackagewhichtheEUcompiledin2025subsumesandextendsmanydesign elements of the 2018 package, our findings are also informative about the economic effects of futureUS-EUretaliationepisodes.2 Theremainderofthepaperisorganizedasfollows. Section2describestheinstitutionalbackground of the EU’s 2018 retaliation and the design of the EU’s tariff retaliation package. Section3investigatestheeffectoftheretaliationonUSimportsharesandquantities,includingthe asymmetric effects. Section 4 details and applies our empirical strategy for identifying tariff pass-throughandpriceeffectsonEUproducerandconsumerprices. Section5concludes. 1Theypointoutthatcertaintariffedproducts,suchaswhiskeyormotorcycles,wereproducedinareaswhich predominantlyvotedforTrumpinthe2016electionornationaliconssuretoreceiveamplepublicattention.Bown, Jung,andLu(2018)andAndersson(2019)provideselectstylizeddescriptionsofthe2018retaliationpackage. 2FleckandPradhan(2026)computeandcomparethedistributionalimpactonAmericanandEuropeanhouseholds of the US 2025 tariffs on EU imports and the proposed but suspended EU retaliation. Gnocato, Gunnella, Montes-Galdon,Schuler,andStamato(2025)assessthedomesticmacroeconomiceffectsofhypotheticaltariffretaliationschemesfocusingontheintermediateversusfinalgoodscomposition.SeeKhan,Zulfiqar,andChunjie(2025) forarecentsurveyoftheliteratureontheeffectsoftradewars,includingtheuseoftariffretaliation. 4

2 The EU’s 2018 Tariff Retaliation 2.1 Timeline On March 8, 2018, President Trump signed executive orders which invoked Section 232 of the 1962 Trade Expansion Act to impose a tariff rate of 25 percent on imports of steel and a 10 percenttariffrateonimportsofaluminum. WhileimportsfromtheEUwereinitiallyexempt, € theywereaddedfromJune1,2018andtargetedabout 6.4bn($7.5bnatthetime)EUimports. The EU responded by filing a complaint with the WTO and announced two sets of retaliatory tariffs, referred to as Annex 1 and 2. Together, these tariffs represented WTO compliant rebalancing measures and amounted to the same value the US tariffs were targeting and mirrored theUStariffrates. TheAnnex1tariffstookeffectfromJune22,2018andincludedtariffsof10 and25percenton86selectimportsfromtheUS,totalingabout € 2.8bn($3.3bn).3 These tariffs remained in place until January 1, 2022 when they were suspended following a jointEU-USstatementinOctober2021inwhichbothsidescommittedtoremovethesteeland aluminum tariffs as well as the retaliation. For our analysis, we focus on these product-level tariffstostudythedesignandeffectoftariffretaliation.4 InAppendixA,wediscusstheother twootherepisodesinwhichtheEUretaliatedagainstUSimportswhilethe2018Annex1tariffs were in place. They only took effect from May 2020 and targeted distinct goods, providing us ample opportunity to study the short-term effects of the 2018 retaliation without concern aboutconfoundingeffects,includingfromextrememovementsinpricesduetotheonsetofthe Covid-19pandemic. 2.2 Design We now present several key features of the Annex 1 tariff design. These are critical to understand our findings on the trade diversion and inflationary effects we present later. First, 3SeeEuropeanCommission(2018a)andEuropeanCommission(2018b). The10percenttariffrateappliedonly to playing cards. All other products were tariffed at a rate of 25 percent. The EU published the tariffs using its Combined Nomenclature (CN) eight-digit codes, listing 182 products. However, as trade data are generally classified according to the six-digit codes of the Harmonized System (HS), we aggregate the CN codes into HS codesthroughoutthispaper. Thisaggregationdoesnotomitinformationwhichisrelevantforouranalysis. For illustration, consider HS code 100630 which refers to “Rice; Semi-milled or wholly milled rice, whether or not polished or glazed.” The CN Annex 1 tariffs have 16 subcategories for this product which provide only minor additionaldetails,suchasgrainshapes,length-widthratios,semivs.whollymilledetc. 4SeeEuropeanCommission(2021a)andEuropeanCommission(2021b). TheAnnex2tariffswerescheduledto complementtheAnnex1tariffsfromJune2021unlesstheUStariffshadbeensuspendedbythen. TheAnnex2 tariffsappliedtoanadditional158products, accordingtotheCNclassification, withtariffratesof10, 25and50 percent,respectively,andtargetedanother€3.6bn($4.2bn)inUSimports,bringingthetotalto€6.4bn($7.5bn).Yet, astheUSandEUhadenterednegotiationsaimingtoremovemutualtradebarriersinearly2021,theEUpostponed theimplementationoftheAnnex2tariffsonMay28,2021,andsuspendedthemindefinitelyonJanuary1,2022, togetherwiththeAnnex1tariffs. 5

relativetotheEU’stotalimportsfromtheworldandevenrelativetototalimportsfromtheUS, the share of imports affected by the retaliatory tariffs was very small. Expressed in values of 2017tradedata,theyonlyappliedto1.5percentofthevalueofallgoodsimportedfromtheUS andto0.2percentoftheEU’stotalgoodsimports. Thus,theretaliatorytariffshadanegligible effectontheEU’sweightedaveragetariffrateandoverallimportcosts.5 Second, the EU’s retaliatory tariffs were asymmetric to the US steel and aluminum tariffs as they focused on a broader set of products. They included a number of particular final goods aimingtocreatepoliticalpressureintheUS,asdiscussedby,forexample,FetzerandSchwarz (2021),whilelimitingdomesticcostsbyavoidingkeyinputsindownstreamindustries. Toprovide more details on the composition of the tariff list and the pre-tariff import trends of the productcodes,westudytradedatafrom2017,theyearbeforetheimplementationoftheretaliatory tariffs. Figure 1 shows that only 31.5 percent value of tariffed US imports belonged to productcategorieswhichincludealuminumandsteel. Meanwhile,finalgoodscategorieslike “Beverages, spirits, and vinegar”, “Ships and boats”, “Vehicles”, and “Cereals” contributed about 37.2 percent of the value and included high visibility imports such as Harley Davidson motorbikes and Bourbon whiskey. Moreover, comparing the number of individual products tariffedwithinsectorsalsoemphasizesthatthesetoftariffedfinalgoodshadadisproportionallysmallernumberofproductsrelativetoitsoverallimportvalue. Inotherwords,theretaliatorytariffsweredesignedtotargetafewhigh-valueproductsotherthaninputsfromsteeland aluminumsectorsandwerehighlydiversifiedwithinthesteelandaluminumcategory. Third, the value of final or processed products targeted by the retaliatory tariffs was larger thanthatofanyraworintermediateproducts. Toillustratethisdimensionofthetariffdesign, we classify the HSproduct codes into Broad Economic Categories(BEC) by mapping product codes through concordances provided by the World Integrated Trade Solution (WITS). By incorporating a System of National Accounts (SNA) end-use dimensions, such as intermediate consumption,capitalgoods,andfinalconsumption,BECcategoriesallowustoidentifywhich stagesofproductiontheEUtargetedtoavoidspilloversindownstreamindustries. Figure2showsthat48percentoftariffedimportsareconsideredintermediategoods. Thus,the shareappliedtoproductstypicallyforfinalconsumptionwasmarginallylaterthanthoseused as inputs in production. This breakdown contrasts sharply with the US tariffs, which applied tointermediateproducts(steelandaluminum)exclusively,andreinstatesthatthedesignofthe retaliatory tariffs was deliberately chosen to minimize supply chain disruptions for European 5AccordingtoWorldBank(2024),thisratewas4.42percentin2017,4.36in2018and4.70in2019. 6

Figure 1: Breakdown of the EU’s 2018 Annex 1 tariffs on US imports using 2017 trade values at the eight-digit CN level. Product classifications are aggregated to the two-digit HS2 chapters. “Apparel & other textile articles” includes HS2 chapters 61, 62, and 63. “Steel (material/mill, fabricated products)”includesHS2chapters72and73(whichincludeproductswithironcontent). Fabricatedrefersto more refined and final products, while material/mill refers to products at the initial state of production. NumbersinparenthesisaretotalnumberofproductsgroupedundereachHS2chapterateight-digitCN productclassification. Source: UNComtrade. companies and to curtail any negative effects on domestic employment and activity. Instead, the design aimed to incentivize European consumers to substitute final goods imported from theUSbygoodssuppliedbyEUandthird-partyproducers. Fourth,theretaliatorytariffstargetedproductcategoriesinwhichtheEUhadminimalimport dependence on the US. To illustrate this point in more detail, Figure 3 plots the relationship betweentheUSimportshareoftariffedproductsgroupedunderHS2chapters,andthetariffed product share of imports from all partners within total imports of two-digit HS2 chapters. To definetheformermoreformallyletT denotethesetoftariffedproductswithinchapterh,and h M be the 2017 import value of tariffed product l from partner p. We define total imports of lp 7

Figure 2: Breakdown of the EU’s 2018 Annex 1 tariffs on U.S. imports using 2017 trade values as classifiedunderBroadEconomicCategories(BEC).Source: UNComtrade. tariffed products in chapter h as MT = ∑ ∑ M . The US import share for chapter h is h l∈T h p l,p computedastheshareofthosetariffedimportssourcedfromtheUS. ∑ M sUS = l∈T h l,US ×100 (1) h MT h To measure how large the tariffed subset of products is relative to the chapter as a whole, let MAll denote total 2017 imports in HS2 chapter h. The within-chapter import coverage of h tariffedproductcodesisthen, MT c = h ×100 (2) h MAll h Byplottingthesetwoshares,wecapturenotonlytheUSfootprintinthetariffedsubsetofproductsasatradingpartnerfortheEU,butalsohowsignificanttheseproductsareincomparison to the entire basket of goods under an HS2 chapter As Figure 3 illustrates, there is a strong negative relationship between the two measures. For example, only two HS2 chapters with tariffed products had more than 20 percent of all EU imports from the U.S. However, in both 8

of them, the within-chapter value of total imports from all partners was just above 10 percent andalmost0percentwithintheirrespectiveHS2chapters. Conversely,inthosetwocategories with high import coverage of tariffed products, only less than 10 percent of tariffed imports camefromtheUS.Finally,fortheoverallmajorityofcategories,bothmeasureswerebelow20 percent,meaningtheUSwasnotadominantsourceofimportsforthesetariffedproducts,and additionallythattheproductsthemselveswhengroupedunderHS2chaptersdidnotrepresent alargevalueshare. Figure3: ComparisonoftheEU’s2018Annex1tariffsonimportsusing2017tradevaluesbypartner shareandbytotalimportcoverage. Tariffedproductcodesattheeight-digitCNlevelhavebeengrouped underHS2chapters. Source: UNComtrade. Finally,wecomputetheimport-partnerconcentrationfortariffedproductsusingtheHerfindahl- Hirshman Index (HHI) calculated at the HS6 subheading level from 2017 import values of tariffed products. This provides a simple proxy for substitutability of imports within a product group: when imports are concentrated in a small number of partners, replacing a tariffed supplierislikelytobemoredifficultintheshortrun. ForeachHS6subheadingi,defines as hp partner p’sshareofEUimportswithinthetariffedproductsoftheHS6code. Theconcentration indexis HHI = ∑ s2 0 < HHI ≤ 1 (3) i hp i p Highervaluesinthisindexindicatethatimportsaresourcedfromfewerpartners,makingthem harder to substitute if the dominant supplier becomes more expensive, while lower values indicate a more diversified supplier base. Notably, partner counts can be high even when concentrationishigh;severalcategoriesmayhavemanypartnerswhileexhibitingahighHHI. i 9

However,thisindicatesthatoneorafewsuppliersmakeupforamajorityofimports. Figure4showsthetop5andbottom5tariffedproductsrankedbytheirimport-partnerconcentration. Notably, only 34 products out of the 180 that were tariffed and traded in 2017 had an HHIabove0.5,meaningthatthevastmajorityofproductswerenotatriskofsupplychaindisruptionsduetodependenceonafewpartners,includingtheUS.Thefigurealsoillustratesthat final use products tend to rank higher in HHI while products in steel and iron, or aluminum i subheadings,generallyrangearound0.1or0.2. Figure 4: Import-partner concentration for the EU’s 2018 Annex 1 tariffs products using 2017 trade values at eight-digit CN level products grouped by six-digit HS6 subheadings. Top 5 and bottom 5 productsrankedbytheirHerfindahl-HirschmanIndex(HHI)areshown. Numbersinparenthesesdenote the total number of import partners for all tariffed products combined under each HS6 subheading. Source: UNComtrade... Figure 4 further decomposes HHI into the US contribution, s2 , indicated with blue shadi (h,US) ing. Crucially,highconcentrationdoesnotautomaticallyimplyUSdependence. Inthesubset ofUS-tradedtariffedproductswithan HHI above0.5,onlysixaremajoritysourcedfromthe i US. These are concentrated in a small set of salient goods, most notably (Bourbon) whiskies, cranberry juice, ground nuts, and playing cards, which are in line with the types of products that have the capacity to generate political pressure but have little downstream industrial exposure. By contrast, some highly concentrated categories like orange juice have very low US importsharesdespitehigh HHI,indicatingconcentrationisfocusedonnon-U.S.suppliers. i In summary, the retaliatory tariffs were designed to result in minor disruptions to domestic producers as the tariffs balanced final and intermediate products. This design is consistent with the broader insight that tariffing final goods is generally less distortionary than tariffing intermediate inputs, because input tariffs propagate through supply chains and raise produc- 10

tion costs as argued by, for example Antra`s, Fort, Gutie´rrez, and Tintelnot (2024). Moreover, the structure of EU import sourcing suggests the measures were crafted to keep substitution costslow;mosttargetedproductgroupsweresuppliedbymultiplepartnersandexhibitedlow importconcentration,implyingthatEuropeanimporterscouldreallocatepurchasesawayfrom the US without major disruption. Where concentration was high, reliance on a dominant US supplierwaslimitedtoasmallsubsetofpoliticallysalientgoods,reinforcingtheinterpretation thattheretaliationbalanceddomesticcostcontainmentwithexternalpoliticalleverage. 3 Effect on Import Shares, Quantities and Prices The four panels of Figure 5 show the evolution of EU annual imports between the years 2010 and2024foreachoftheproductcategoriessubjecttothe2018retaliatorytariffs. Theblackbars representthevalueoftotalimports(fromallpartners,includingtheUS)incurrentUS-Dollars (leftaxis). ThebluelinerepresentstheimportsfromtheUStotheEUasashareoftotalcategory importvalues(rightaxis). Ineachpanel,theverticalblacklineindicatestheintroductionofthe EU’sretaliatorytariffsinJune2018. For all categories except Textiles and Apparel, we find that US import shares to the EU fell notably and persistently after 2017. While import values from all partners decreased during the first two years following the tariff implementation—also due to the onset of the Covid19pandemic in 2020—they eventually returned to trend-growth. However, the US share did not recover to its pre-tariff level but remained below, even after the suspension of the retaliatory tariffsonJanuary1,2022. Inotherwords,theretaliatorytariffsresultedinlastingtradediversionasEuropeanimporterssubstitutedawayfromtariffedUSimportsandsourcedthemfrom non-U.S.suppliers.6 This finding is consistent with other papers studying the effect of foreign tariff retaliation in 2018 on US exports. For example, Carter and Steinbach (2020) find that foreign tariffs on US agricultural products significantly reduced exports, already in the short-run, and resulted in a re-configuration of trade relationships as retaliators started sourcing tariffed products from countriesnotaffectedbyretaliatorytariffs. Ourresultsshowthatthismechanismalsoapplied to other export sectors and that the US import shares have not recovered even six years after theapplicationoftheEU’sretaliatorytariffsandaboutfouryearsaftertheirsuspension. 6FigureC1inAppendixCshowsthechangeintotalimportvaluesofalltariffedgoodsatthemonthlyfrequency forthe30monthsbeforeandthefirsttwelvemonthsafterthetariffswereimposed. Itshowsthat,alreadyinthe firstyear,year-over-yeargrowthinimportvalueshadfallentoabouthalfofthatseenbeforeJune2018. 11

Figure5: TotalimportvaluesfromallpartnersandUSimportshareoftheEU’s2018Annex1tariffed productsusingtradevaluesatsix-digitHS6subheadings,concordedacrossHS2007toHS2022using WITSconcordancetables. Source: UNComtrade. ToinvestigatehowtheretaliatorytariffsaffectedtheEU’simportprices,wenowusemonthly data on import quantities and prices (before tariffs) at the CN product level to construct and inspect unit values of tariffed and non-tariffed goods. To this end, we multiply goods subject to tariff retaliation with the applicable rates, aggregate imports into tariffed and non-tariffed goods and normalize their twelve-month price changes to zero for the month before the tariff wasimplemented. Figure6showsbothtimeseries. Priortotheimplementationofthetariffs,untreatedpricesbarelyfluctuate,indicatingthatsubstantial movements are likely related to tariffs. Moreover, there is no evidence of a difference inpricetrendsbetweentariffedandnon-tariffedgoodspre-2018. However,afterthetariffimplementation, the twelve-month change in tariff-inclusive import prices for the tariffed goods climbedroughly20to30percentagepointsrelativetothepre-tariffmonth. Thisrangeiscommensuratewiththeactualtariffrateappliedtoalmostalltariffedgoods(25percent)indicating that the retaliatory tariffs were rapidly and fully passed through to European importers. Finally, the series shows no evidence for price increase spillovers from tariffed to non-tariffed 12

goods as import prices of tariffed goods remained permanently higher after the tariff implementation. Figure 6: “Untreated” includes all goods and countries that do not face retaliatory tariffs. June 2018 is set as month 0 as tariffs were effective on June 22, 2018, i.e. after the middle of the month. Data for the change in import prices is the geometric mean of the 12-month tariff-inclusive unit value relatives, weightedbythelogarithmicmeanoftheirimportsharesbyvalue. Intheleftpanel,pricesarenormalized so that so a zero is a price change that equals its value before the tariffs were implemented. In the right panel, import values are normalized to one in month zero so the import values are relative to imports inthelastmonthbeforethetariffswereimplemented. Petroleumproducts(HSchapters2709-2715)are removed. Source: UNComtrade. To formally quantify the effects of the EU’s retaliation on imports and prices, we now study twelve-month changes, expressed in logs, in tariffed imports and their prices between June 2017andJune2019,i.e.aboutoneyearbeforeandelevenmonthsafterthetariffsbecameactive (on June 22, 2018). In our regressions, we use the EU’s statutory retaliatory tariffs as an independent(exogenous)variableandincludeproduct-levelandmonthlytimecontrolstoaccount for changes in other variables which can lead to changes in trade quantities and values, such asmovementsintheexchangerate.7 PanelAofTable1presentsresultsfortheperiodofJune2017toJune2019,capturingtheimport demand elasticity of US tariffed imports in quantities and values in the wake of the Annex I tariffs. Panel B spans the period of December 2020 to December 2022, studying the effects on import demand elasticity following the suspension of Annex I tariffs on January 1, 2022. The first column of Panel A shows the result of regressing the change in foreign exporter prices 7OureconometricapproachissimilartoAmiti,Redding,andWeinstein(2019)butconsidersalongertimehorizonafterthetariffimplementation. Notethat,unliketherecentUStariffs,theEU’s2018retaliatorytariffsdidnot featureinitialincompleteenforcementandnumerousexemptions. See,forinstance,GopinathandNeiman(2026). Thus,thereisnomeaningfuldifferencebetweenstatutory(announced)andeffectivetariffrates. 13

on product specific tariff rates. The estimated impact of tariffs on unit values of 0.336, which is statistically insignificant, and suggests that tariff retaliation had no discernible impact on pricesreceivedbyforeignexporters. Thus,itvalidatesthefindinginFigure6whichpointedto substantialtariffpassthroughtopricesbeingpaidbyEUimporters. This finding on the passthrough of the EU’s tariff retaliation costs in 2018 is consistent with evidence on US tariff passthrough in 2018 and 2025, as documented by, for example, Amiti, Redding, and Weinstein (2019) and Gopinath and Neiman (2026). Thus, a key finding of our analysis is that there is no difference in the pass-through of tariffs applying to a large share of totalimports–asfortheUS“reciprocaltariffs“–andtariffstargetingasmallshareandproducts withlowimportconcentration–asfortheEU’sretaliation. The remaining columns of Panel A show that the introduction of EU retaliatory tariffs in June 2018ledtosharpdeclinesinUSimports. Incolumn2,weregresschangesintariffsonchanges in import quantities to obtain an estimate of the import demand elasticity. The coefficient is precisely estimated and suggests that a one percent increase in the tariff rate reduced import quantitiesby2.9percent. Incolumn3,werunthesameregressionbutapplyaninversehyperbolicsince(IHS)transformationtoimportquantitiesofzerosinsteadofdroppingthem. While our estimate of 4.8 percent is slightly below the analogous estimate reported by Amiti, Redding, and Weinstein (2019) for the US tariffs against China in 2018, it still falls into the typical rangeoftradeelasticityestimates. In columns 4 and 5, we replace import quantities with import values. The resulting estimates suggest that import values fell by about 3.6 or 5.2 percent, respectively in response to a one percentincreaseinthetariffrate. Again,theIHSestimateissomewhatbelowthatreportedby Amiti, Redding, and Weinstein (2019) but still indicates that the EU retaliatory tariffs reduced affectedUSimportsbysizableamountscomparedtothenon-affectedimports. Finally,Table1shedsmorelightontheasymmetrybetweenthetrade-destroyingeffectsoftariff imposition and the trade-restoring effects of tariff suspension. While the results in Panel A show that EU importers substituted away from US goods as the EU’s tariffs became active, PanelBestimateshowtheyrespondedtothetariffsuspensiononJanuary1,2022. Allestimates are notably smaller than their corresponding Panel A estimates, providing clear evidence of onlypartialrecoveryoftariffedUSimports. Thissuggeststhattariffremovalfailedtoreverse the full extent of the initial trade destruction, suggesting that a substantial share of the disrupted trade relationships did not renew once the tariffs were lifted, likely reflecting the sunk costsinvolvedinsupply-chainreconfigurationandconsistentwithscarringeffectsoftariffs. 14

Table1: EffectofRetaliatoryTariffsonUSImports (1) (2) (3) (4) (5) Logchangeforeign Logchange IHSchange Logchange IHSchange exporterprices importquantities importquantities importvalues importvalues PanelA ∆ln(1+Tariff ) 0.336 −2.916*** −4.758*** −3.558*** −5.241*** it (0.202) (0.374) (0.811) (0.457) (0.864) Observations 98249 118586 144278 153772 177433 R2 0.074 0.111 0.100 0.090 0.083 PanelB −∆ln(1+Tariff ) 1.418*** 1.464* 1.964*** 2.971*** it (0.348) (0.815) (0.383) (0.781) Observations 115699 148348 154575 184272 R2 0.120 0.213 0.095 0.217 Source:UNComtrade. Note:DataareattheCN2018productlevelwithamonthlyfrequency.PanelAistheperiodbetweenJune2017 toJune2019,andPanelBistheperiodbetweenDecember2020toDecember2022.InPanelA,∆ln(1+Tariffit) reflects tariff imposition; in Panel B, −∆ln(1+Tariffit) reflects tariff suspension. EU Legislation 2018/886 AnnexItariffswereimplementedonJune22,2018,andsuspendedonJanuary1,2022. IHSistheinverseof hyperbolicsinetohandlezero-valuetradedataintort-12. Columns1-3excludeunitvalueratiosoftand t−12thataregreaterthan3orlessthan1/3.StandarderrorsinparenthesesandareclusteredattheCN2018 eight-digitlevel,noted*** p<0.01,** p<0.05,* p<0.1. 4 Effect on EU Domestic Prices 4.1 ProducerPrices In the previous section, we show evidence that the EU’s retaliatory tariffs did not result in lowerpriceschargedbyUSexporters. Inotherwords,Europeanbusinessandconsumersbore thebruntofthetariffcosts. Yet,wealsofoundthatthetariffedimportsweresmallandthatimportsfromtheUSinthetariffedcategoriesdeclinedafterthetariffswereimplemented. Tosee which of these offsetting forces had a stronger impact, we now conduct a formal passthrough analysiswhichmapstheuseoftariffedgoodsineachsectortoasectorspecificEuropeanprice index. ThisanalysiswillprovidequantitativeinsightsonifandhowtheEU’stariffretaliation affected domestic prices. Since the retaliation had a roughly even split for tariffed goods between intermediate and final use, as detailed in Figure 2, our analysis estimates price effects bornebybothconsumersandproducersseparately. Lookingfirsttoproducers,weexaminethepassthroughoftariffsintotheEUProducer’sPrice Index(PPI)whichisclassifiedaccordingtotheNACERev. 2systemandavailableataquarterly frequency. Using data from the 2017 edition of the European Union’s Full International and GlobalAccountsforResearchinInput-OutputAnalysis(FIGARO)tables,wecanobservehow 15

products—categorized according to the classification of products by activity (CPA)—are used by various NACE Rev. 2 industries as inputs for production. Most importantly, the FIGARO tablesalsoprovideanaccountofwheretheseproductswereoriginallyimportedfrom. WemaptheoriginallistoftariffedCNcodesandtheirassociatedtariffratestothe64CPA2.1 product categories provided in the FIGARO tables. Using correspondence tables provided by the European Commission we are able to link the tariffed goods to CPA products at the sixdigitlevel. AstheproductdataintheFIGAROtablesisaggregatedtotheCPAtwo-digitlevel weaggregatethetariffratesaccordingly. Todothis,wecomputeanaveragetariffrateforeach two-digitcategory,weightedbytherespectiveUSimportshareforthatsix-digitproduct.8 Finally,foreachPPI(NACERev. 2)industry,wecomputeaweightedaveragetariffrateas T¯ = ∑∑ ω τ (4) k ijk ij i j where s ijk ω = (5) ijk ∑ ∑ s i j ijk Intheseequations,s isthevalueofproductifromcountryjusedinsectorkandτ isthetariff ijk ij rate corresponding to product i from country j which is assumed to be zero for all imported goodsotherthanthetariffedonesfromtheUS. Producer price indices are not published by Eurostat for every sector and for many collection beganonlyin2021. Thus,tostudypricechangesintheaftermathofthe2018tariffimplementation,wehavetorestrictoursampleto31industries. Usingthisdata,weestimatethefollowing linear regression of the sector specific tariff rate T¯ on changes in quarterly PPI, π , to assess k k theinflationaryeffectofthetariffs: π = α+βT¯ +ε (6) k k Since the tariffs were effective as of June 22, 2018, we look at changes in producer prices in the third quarter of 2018 and consider three different measures for PPI changes; quarter-overquarter percent change, 4-quarter percent change and excess 4-quarter inflation. The excess inflation measure is adapted from Minton and Somale (2025) and computes the difference be- 8ThisensuresthatthatproductswheretheEUdoesnotimportanygoodsfromtheUSdonotbiastariffrates downward. Also, to ensure that we are not artificially inflating the tariff rate for a given sector by including all tariffsthatwereinplaceasof2018, wedefineourtariffrateastheadditionaldutiesleviedinthe2018EUtariff announcement. Computationally, thisresultsintariffratesof0percentforallproductsimportedfromcountries thatarenottheUnitedStates. 16

tween the 4-quarter inflation in 2018:Q3 and the average 4-quarter inflation from 2000:Q1 to 2018:Q1. Our results displayed in Table 2 indicate no statistically significant relationship between higher 2018 tariff rates and increased producer prices. This finding is consistent across all inflation measures, including when we replicate our analysis but use 2018Q2 and 2018Q4 PPIdataasbenchmarks. Table2: 2018:Q3TariffPassthroughRegressions Dependentvariable: 2018:Q3ProducerPriceIndex (1) (2) (3) Q-o-QChange 4-quarterchange Excess4-quarterinflation TariffRate −902.1 −4946.7 −7157.5 (2304.0) (10230.9) (8661.7) Observations 31 31 31 R2 0.005 0.008 0.023 AdjustedR2 −0.029 −0.026 −0.011 Note: Largecoefficientestimatesaredrivenbyvolatilityincokeandrefinedpetroleumproducts. ∗p<0.1;∗∗p<0.05;∗∗∗p<0.01 Finding that the retaliation did not affect European producer prices is consistent with the design of the retaliation presented earlier. Recall that retaliatory tariffs were small and targeted sectors that were not overly reliant on US imports for production. As such, even if there was completetariffpassthroughasarguedpreviously,therewasnodiscernibletransmissiontoEU domestic prices. Yet, given the limitations on PPI data, we cannot categorically conclude that PPIpriceeffectswerenegligibleandthusturntoinvestigateconsumerpricesnext. 4.2 ConsumerPrices The EU’s Harmonized Index of Consumer Prices (HICP) is categorized according to the Classification of Individual Consumption According to Purpose (COICOP) framework. We use correspondencetablesprovidedbyCaiandVandyck(2020)tomaptheuseof63CPAproducts into36COICOPinflationcategories.9 UsingdataonimportsprovidedinourFIGAROtables,wecomputeaweightedaveragetariff rateforeachconsumptioncategorykgivenby T¯ = ∑ s ωUS τ (7) k ik i i i where s is the share of product i in category k, ωUS is the share of US imports in product i ik i 9SeeCaiandRueda-Cantuche(2018)fordetailsonthegenerationofthiscorrespondence. 17

and τ is the product specific tariff rate. Stated simply, this COICOP specific tariff considers i bothhowmucheachproductfactorsintoagivencategorybutalsohowmuchofthatproduct isimportedfromtheUnitedStates. Wethenusethesetariffratestoestimate πexcess = α+β T¯ +ε (8) k k where, again as in Minton and Somale (2025), πexcess are HICP category specific differences k betweenaverage3-monthinflationfrom2000to2017andthe3-monthinflationratefollowing tariff implementation. Table 3 reports our findings. Columns 1-3 estimate excess inflation at 3 different monthly horizons after implementation. Consistent with our analysis for producer prices, we find no discernible effect of tariffs on consumer prices as all coefficients are not statisticallysignificant. Again,thisfindingisconsistentwithouranalysisofthetariffpackage whichfoundthatitwassmallandextremelytargeted. Itisalsoworthnotingthatweareonly requiredtodrop2outof36HICPcategoriesduetodatalimitations. Table3: ConsumerPricePassthroughRegressions Dependentvariable: ExcessInflation (1) (2) (3) πexcess πexcess πexcess t+1 t+2 t+3 TariffRate −12.5 9.20 19.88 (14.5) (16.24) (23.54) Observations 34 34 34 R2 0.023 0.010 0.022 AdjustedR2 −0.008 −0.021 −0.009 Note: The subscript t+i corresponds to the 3 month excess inflation in the months following tariff implementation,wheret+1denotesSeptember. ∗p<0.1;∗∗p<0.05;∗∗∗p<0.01 5 Conclusion A growing body of literature studies the effect of US tariffs and potential foreign retaliation on the US economy. However, retaliatory tariff strategies can take many shapes, which may notjustreflecttit-for-tattariffincreases,andlittleisknownabouttheeffectsoftariffretaliation on the retaliator. This stands at odds with the increasingly relevant role of retaliation in a less multilateraltradeorder. Ourpaperstudiesthedomesticeffectsoftariffretaliationbyagroupoflargeadvancedeconomies, the EU, that publishes high-quality data on import quantities and prices as well as producer 18

andconsumerprices. WefindthattheEU’sretaliationagainstUStariffsin2018wasminorin scope and highly diversified. Its design targeted intermediate and final goods in about equal proportions and focused on goods which EU importers could easily obtain from non-US suppliers,leveragingthattheEUonlyincreasedtariffsononeofitsmanytradingpartners. TheEU’sretaliatorytariffshadanasymmetricandlastingeffectontheUSsharesinEUimports in most of the tariffed categories. These shares decreased notably in the wake of the tariffs and did not recover after the retaliatory tariffs they were lifted but left scars in US-EU trade relations. Moreover, even though the tariffs costs were rapidly and fully passed through to European importers, we do not find evidence that they led to upward pressure on domestic producer or consumer prices, reflecting the small size of the retaliation, as well as its careful targetingandtherapidsubstitutiontonon-USsuppliers. 19

References AMITI, M., S. J. REDDING, AND D. E. WEINSTEIN (2019): “The Impact of the 2018 Tariffs on PricesandWelfare,”JournalofEconomicPerspectives,33(4),187–210. ANDERSSON, S. (2019): “The Price, Quantity, and Welfare Impacts on the EU Following IncreasedImportTariffs,”Masterthesis,LundUniversity. ANTRA`S, P., T. C. FORT, A. GUTIE´RREZ, AND F. TINTELNOT (2024): “TradePolicy and Global Sourcing: AnEfficiencyRationaleforTariffEscalation,”JournalofPoliticalEconomyMacroeconomics. BOWN, C. P., E. JUNG, AND Z. LU (2018): “Harley is a tariff trend setter–but not in a good way,”Policybrief,PetersonInstituteforInternationalEconomics. BRAML, M. (2020): “Beggar-thy-Neighbor or Favor thy Industry? An Empirical Review of TransatlanticTariffRetaliation,”ifoWorkingPaperSeries326,ifoInstitute. CAI, M., AND J. M. RUEDA-CANTUCHE(2018): “Bridgingmacroeconomicdatabetweenstatisticalclassifications: thecount-seedRASapproach,”EconomicSystemsResearch,31(3),382–403. CAI, M., AND T. VANDYCK (2020): “Bridgingbetweeneconomy-wideactivityandhouseholdlevelconsumptiondata: MatricesforEuropeancountries,”DatainBrief,30,105395. CARTER, C. A., AND S. STEINBACH (2020): “The Impact of Retaliatory Tariffs on Agricultural andFoodTrade,”WorkingPaper27147,NationalBureauofEconomicResearch. CAVALLO, A., G. GOPINATH, B. NEIMAN, AND J. TANG (2021): “Tariff Pass-Through at the Border and at the Store: Evidence from US Trade Policy,” American Economic Review, 111(8), 2701–33. CHOR, D., AND B. LI (2024): “Illuminating the effects of the US-China tariff war on China’s economy,”JournalofInternationalEconomics,147,103852. EUROPEAN COMMISSION (2018a): “Commission Implementing Regulation (EU) 2018/886 of 20June2018oncertaincommercialpolicymeasuresconcerningcertainproductsoriginating intheUnitedStatesofAmerica,”OfficialJournaloftheEuropeanUnion. (2018b): “EU adopts rebalancing measures in reaction to US steel and aluminium tariffs,”PressRelease. (2021a): “CommissionImplementingRegulation(EU)2021/2083of26November2021 on suspending commercial policy measures concerning certain products originating in the United States of America imposed by Implementing Regulations (EU) 2018/886 and (EU) No2020/502,”OfficialJournaloftheEuropeanUnion. (2021b): “Joint EU-US Statement on a Global Arrangement on Sustainable Steel and Aluminium,”JointStatement/PressRelease. FAJGELBAUM, P. D., P. K. GOLDBERG, P. J. KENNEDY, AND A. K. KHANDELWAL (2020): “The ReturntoProtectionism,”TheQuarterlyJournalofEconomics,135(1),1–55. FETZER, T., AND C. SCHWARZ (2021): “Tariffs and Politics: Evidence from Trump’s Trade Wars,”TheEconomicJournal,131(636),1717–1741. FLAAEN, A., A. HORTAC¸SU, AND F. TINTELNOT (2020): “TheProductionRelocationandPrice Effects of US Trade Policy: The Case of Washing Machines,” American Economic Review, 20

110(7),2103–27. FLAAEN, A., AND J. PIERCE (2024): “Disentangling the Effects of the 2018–2019 Tariffs on a GloballyConnectedU.S.ManufacturingSector,”TheReviewofEconomicsandStatistics. FLECK, J., AND A. PRADHAN (2026): “The Distributional Effects of the US-EU Trade War,” WorkingPaper. GNOCATO, N., V. GUNNELLA, C. MONTES-GALDON, T. SCHULER, AND G. STAMATO (2025): “Tariffsacrossthesupplychain,”Policyinsight,VoxEU.org. GOPINATH, G., AND B. NEIMAN(2026): “TheIncidenceofTariffs: RatesandReality,”Working Paper34620,NationalBureauofEconomicResearch. GRANT, J. H., S. ARITA, C. EMLINGER, R. JOHANSSON, AND C. XIE (2021): “Agricultural exportsandretaliatorytradeactions: Anempiricalassessmentofthe2018/2019tradeconflict,” AppliedEconomicPerspectivesandPolicy,43(2),619–640. KHAN, H., A. ZULFIQAR, AND Q. CHUNJIE (2025): “Economic, Institutional, and Strategic EffectsofTariffWars: SystematicLiteratureReview,”JournalofEconomicSurveys. LI, M., E. J. BALISTRERI, AND W. ZHANG (2020): “The U.S.–China trade war: Tariff data and generalequilibriumanalysis,”JournalofAsianEconomics,69,101216. MINTON, R., AND M. SOMALE (2025): “Detecting Tariff Effects on Consumer Prices in Real Time,”Fedsnotes,BoardofGovernorsoftheFederalReserveSystem. MORGAN, S., S. ARITA, J. BECKMAN, S. AHSAN, D. RUSSELL, P. JARRELL, AND B. KENNER (2022): “The Economic Impacts of Retaliatory Tariffs on U.S. Agriculture,” Economic ResearchReport304,U.S.DepartmentofAgriculture,EconomicResearchService. WORLD BANK (2024): “Tariff rate, applied, weighted mean, all products (percent),” World IntegratedTradeSolution(WITS),Accessed: 2025-11-21. 21

Supplemental Appendix Supplemental appendix to ”The Design and Effect of Tariff Retaliation: Evidence from the EU” (Fisgin, Fleck and Richards,2026) A Other EU-US Tariff Retaliation Episodes between 2018 and 2022 In addition to the 2018 retaliation episode we study in this paper, the EU also implemented retaliatory tariffs against US imports on two other occasions, in May and November 2020. Importantly, all of these additional tariffs applied to products other than those included in the 2018 episode. While the retaliation in May 2020 was related to the 2018 retaliation in the sense that it was in response to US tariffs on steel and aluminum derivatives, it was much smaller in magnitude and applied to different products which than those targeted in the 2019 retaliation. Figure A1 provides a timeline and illustrates the relative magnitudes of the different retaliation episodes while the next two paragraphs summarizetheimplementationanddesigndetails. FigureA1: EpisodesofEUtariffretaliationagainstUSimportsbetween2018and2022. Importssubjecttoretaliatorytariffsare measuredincurrentbnUSD.CompiledfrominformationprovidedinEuropeanCommission(2018a,b,2020b,a,2021b,c). OnApril7,2020,theEUrespondedtoUStariffsonderivativesteelandaluminumimportsfromtheEUwithregulation European Commission (2020c). This regulation applied tariffs on three US import products in two annexes. The first applied20percentand7percentonlightersandplasticfittingsfromMay8,2020,worthabout$21.7mio. Thesecond applied4.4percentonimportedplasticfittingsfromFebruary8,2023,worthabout$21.3mio. Together,thesetwosets of tariffs amounted to about $43 mio worth of US imports, mirroring the target of the US tariffs, as required by WTO provisions. Likethe2018Annex1tariffs,thefirstsetoftheseadditionaltariffsfirstweresuspendedindefinitelyfrom January1,2022. Thesecondsethadnotyetbeenimplementedandwasequallysuspended. SeeEuropeanCommission (2021b). 1

Second, on November 10, 2020, as part of the WTO dispute regarding US subsidies to Boeing, the EU implemented additional tariff measures, unrelated to the US steel and aluminum tariffs. See European Commission (2020b) and European Commission (2020a). Again, they were published in two annexes and Annex 1 listed tariffs of 15 percent on “Aeroplanes and other aircraft” while Annex 2 applied 25 percent to accessories, agricultural and manufactured products as well as agricultural utility vehicles and leisure items. Together, they were applying to imports worth about $4bn. They were suspended from July 9, 2021, for a period of five years, following a joint US-EU statement of understandingonacooperativeframeworkforlargecivilaircraftdisputes. SeeEuropeanCommission(2021a). B Additional results on the design of the EU’s 2018 tariff retaliation package InFigureB1,weprovideabreakdownofthetariffedproductsbyBECcategorywhichgoesbeyondintermediate,final, andcapital/investmentgoods. Figure B1: Breakdown of the EU’s 2018 Annex 1 tariffs on U.S. imports using 2017 trade values as classified under Broad EconomicCategories(BEC).Datasource: UNComtrade. C Additional results on the effect of the EU’s 2018 tariff retaliation on imports Figure C1 shows the change in total import values of tariffed goods following the imposition of the EU’s retaliation in June 2018. Two months prior to the imposition of the tariffs, there appears to be a spike in imports, potentially indicating tariff “front-running” by EU importers. However, given the historical volatility of the series, the evidence is limited. In contrast, year-over-year growth in total import values in the twelve months following the imposition of retaliatory tariffs was less than half of that seen in the month prior. By contrast, unaffected products and partners remainedrelativelysteady,highlightingtheimpactofretaliationinshiftingimportsawayfromtariffedUSproducts. 2

FigureC1: “Untreated”includesallgoodsandcountriesthatdonotfaceretaliatorytariffs. June2018issetasmonth0astariffs wereeffectiveonJune22,2018,i.e. afterthemiddleofthemonth. Dataforthechangeinimportpricesisthegeometricmeanofthe 12-monthtariff-inclusiveunitvaluerelatives,weightedbythelogarithmicmeanoftheirimportsharesbyvalue. Intheleftpanel, pricesarenormalizedsothatsoazeroisapricechangethatequalsitsvaluebeforethetariffswereimplemented. Intherightpanel, importvaluesarenormalizedtooneinmonthzerosotheimportvaluesarerelativetoimportsinthelastmonthbeforethetariffs wereimplemented. Petroleumproducts(HSchapters2709-2715)areremoved. Source: UNComtrade. References EUROPEAN COMMISSION (2018a): “Commission Implementing Regulation (EU) 2018/886 of 20 June 2018 on certain commercialpolicymeasuresconcerningcertainproductsoriginatingintheUnitedStatesofAmerica,”OfficialJournaloftheEuropeanUnion. (2018b): “EUadoptsrebalancingmeasuresinreactiontoUSsteelandaluminiumtariffs,”PressRelease. (2020a): “BoeingWTOcase: TheEUputsinplacecountermeasuresagainstU.S.exports,”PressRelease. (2020b): “Commission Implementing Regulation (EU) 2020/1646 of 7 November 2020 on commercial policy measures concerning certain products from the United States of America following the adjudication of a trade dispute under the Dispute Settlement Understanding of the World Trade Organization,” Official Journal of the EuropeanUnion. (2020c): “Commission Implementing Regulation (EU) 2020/502 of 6 April 2020 on certain commercial policy measuresconcerningcertainproductsoriginatingintheUnitedStatesofAmerica,”OfficialJournaloftheEuropean Union. (2021a): “CommissionimplementingRegulation(EU)2021/1123of8July2021suspendingcommercialpolicy measuresconcerningcertainproductsfromtheUnitedStatesofAmericaimposedbyImplementingRegulation(EU) 2020/1646 following the adjudication of a trade dispute under the Dispute Settlement Understanding of the World TradeOrganization,”OfficialJournaloftheEuropeanUnion. (2021b): “Commission Implementing Regulation (EU) 2021/2083 of 26 November 2021 on suspending commercialpolicymeasuresconcerningcertainproductsoriginatingintheUnitedStatesofAmericaimposedbyImplementingRegulations(EU)2018/886and(EU)No2020/502,”OfficialJournaloftheEuropeanUnion. (2021c): “Joint EU-US Statement on a Global Arrangement on Sustainable Steel and Aluminium,” Joint Statement/PressRelease. 3