The Effect of Swiss Free Trade Agreements on Agricultural Trade

Fiankor, Dela‐DemDoe; Ritzel, Christian; Irek, Judith; Mack, Gabriele Article — Published Version The Effect of Swiss Free Trade Agreements on Agricultural Trade Kyklos Provided in Cooperation with: John Wiley & Sons Suggested Citation: Fiankor, Dela‐DemDoe; Ritzel, Christian; Irek, Judith; Mack, Gabriele (2025) : The Effect of Swiss Free Trade Agreements on Agricultural Trade, Kyklos, ISSN 1467-6435, Wiley, Hoboken, NJ, Vol. 78, Iss. 4, pp. 1333-1357, https://doi.org/10.1111/kykl.12472 This Version is available at: https://hdl.handle.net/10419/330146 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Kyklos, 2025; 78:1333–1357 https://doi.org/10.1111/kykl.12472 1333 Kyklos ORIGINAL ARTICLE OPEN ACCESS The Effect of Swiss Free Trade Agreements on Agricultural Trade Dela-DemDoeFiankor1,2 | ChristianRitzel2 | JudithIrek2 | GabrieleMack2 1Department of Agricultural Economics and Rural Development, University of Goettingen, Göttingen, Germany | 2Economic Modelling and Policy Analysis Group, Agroscope, Ettenhausen,Switzerland Correspondence: DelaDem Doe Fiankor ([email protected]) Received: 2 May 2025 | Revised: 2 May 2025 | Accepted: 16 May 2025 Keywords: agricultural trade| free trade agreements| gravity models| Switzerland ABSTRACT As a country highly dependent on imports, Switzerland has many free trade agreements (FTAs) that liberalise trade barriers. We assess how these agreements affect Swiss agricultural imports at different margins of trade adjustment. We estimate reducedform gravity models using agricultural trade data for 202 partner countries from 2004 to 2022. We find that Swiss FTAs increase agricultural import values by 8.75%, decrease import prices by 3%, increase the probability of imports by 2% and reduce market exit rates by 1%. These effects are heterogeneous across products, sectors and agreements. Regarding import values and quantities, the positive effects of FTAs are mainly observed for raw products (including vegetables, fruits and nuts, coffee, tea and spices). However, the estimated effects are negative for processed products. Regarding import prices, the effects are positive whenever they are statistically significant. We also find that the number of competing agreements to which a Swiss trade partner is exposed only marginally affects Swiss imports. We extend our analysis to agricultural exports and find that FTAs increase Swiss export values by 47%, quantities by 53% and prices by 3% but do not affect export probabilities or export market exit rates. Thus, although Swiss FTAs generally boost trade on average, policymaking should consider the heterogeneities of the estimated FTA effects regarding products, agreements and time when using FTA estimates for counterfactual analysis and negotiations. JEL Classification: F14, Q17, Q18 1 | Introduction Economists disagree on many things, but the superiority of free trade over protection is not controversial (Rodrik2018). A free trade agreement (FTA) allows countries to reduce barriers to imports and exports on a bilateral basis, allowing consumers to benefit from greater product variety at lower prices. This is particularly relevant for agriculture where trade barriers are traditionally higher relative to other sectors. For example, in 2015, global average tariffs were 5% for nonagricultural products and 11% for agriculture (Niu etal.2018), highlighting the substantial potential gains from liberalizing trade in agriculture. Existing studies on FTAs, however, focus mainly on big countries, such as the European Union and the United States, and assess their economywide effects, leaving a knowledge gap on how FTAs affect the agricultural sector in smaller countries.1 We address this gap using the case of Switzerland—a small, open economy in which imports account for approximately 50% of domestic consumption (Ritzel etal.2024). As of 2024, Swiss trade policy rests upon three main pillars: (i) World Trade Organization (WTO) membership, (ii) association agreements with the European Union (EU) and membership of the European Free Trade Association (EFTA) and (iii) bilateral agreements with other countries. WTO membership means that all Swiss imports are subject to MostFavoredNation (MFN) tariffs. If MFN tariffs are positive but imports originate from a country that has an FTA with Switzerland This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2025 The Author(s). Kyklos published by John Wiley & Sons Ltd. 1334 Kyklos, 2025 (under either of the other two pillars), the goods benefit from lower or even zero tariffs.2 Given the trade cost reductions that come with trade liberalisation, we expect FTAs to increase bilateral trade. The same is true for FTAs that liberalise nontariff measures (NTMs) and administrative procedures. What remains an empirical question is the magnitude of the trade effect and whether the effects vary by product, agreement or over time. Furthermore, the fact that Switzerland has an FTA with its partners does not preclude the partners from signing FTAs with other countries. These thirdcountry agreements could offer a comparable or even higher level of liberalisation to Switzerland's trade partners and divert potential exports destined for Switzerland to alternative destinations. Assessing whether and to what extent these competing thirdcountry agreements affect Swiss imports is necessary to provide a holistic picture of the trade effects of Swiss FTAs. Based on this premise, the economic question underlying our work is how important FTAs are for Swiss agricultural trade. Our use of Switzerland's agricultural sector as a case study is based on the stark contrast between the levels of protection in the agricultural and nonagricultural sectors. Switzerland's tariff pattern reveals high rates of MFN tariffs on agricultural imports compared with low rates on industrial goods. These high tariffs serve a politically motivated protective role for the agricultural sector, limiting opportunities for substantial concessions in reciprocal trade negotiations for domestically sensitive agricultural products. Conversely, the industrial sector faces minimal tariffs, which continue to decrease. For instance, in January 2024, Switzerland implemented a significant trade reform by autonomously eliminating all tariffs on industrial imports, irrespective of origin (Zimmermann 2023). More broadly, the higher protection levels in the agricultural sector visàvis the nonagricultural sector suggest a larger trade increase in agriculture following an FTA. This expectation is consistent with Grant and Lambert(2008), who find that trade agreements increase agricultural trade by an average of 72%, compared with a 27% increase in nonagricultural trade. In Switzerland, where the disparity in protection between sectors is even more pronounced and there exists little flexibility in agricultural concessions, the impact of FTAs on Swiss trade remains an empirical question. Our empirical assessment uses data on agricultural imports and FTAs in force between 2004 and 2022. We define two margins of import adjustments: the intensive margin (measured by import values, import quantities and import prices) and the extensive margin (measured by the probability of imports and market exit). We then estimate a reducedform gravity model that regresses two FTA indicators—the presence of an FTA (dummy variable) and the number of thirdcountry FTAs—on these margins. Our empirical findings show that, on average, Swiss FTAs increase import values by 8.75%, decrease import prices by 3%, increase the probability of imports by two percentage points and reduce market exit rates by one percentage point. On the effect of thirdcountry FTAs, we observe very marginal effects. For instance, an additional thirdcountry FTA decreases Swiss import quantities by about 0.3%, decreases import prices by 0.1% and changes import probability and market exit rates by 0.1 percentage points. Thus, although these effects are statistically significant, the magnitudes are too small to have sizeable negative impacts on Swiss imports. To provide deeper insights into our main findings, we assess the heterogeneity of the average FTA effects across various dimensions. Swiss FTAs, aiming to achieve targeted liberalisation that aligns with Swiss agricultural policy objectives, distinguish between basic (raw) and processed agricultural products. Assessing the heterogeneity of the trade effect across this product classification, we find that FTAs increase the import values and quantities for raw products but decrease them for processed products. Similar heterogeneity is observed across different HS2digit product sectors. Some FTAs increase imports, others decrease imports and others have no effect on imports. This pattern of heterogeneity is consistent across other import margins. To capture the dynamic effects of FTAs, we incorporate lags and leads of the FTA variable. We find no evidence of anticipation effects but find that the trade effects phase in up to 2 years after implementation. For completeness, we extend our analyses to Swiss exports, even though agricultural exports make only a small share of total Swiss trade. We find that Swiss FTAs increase Swiss export values by 47%, quantities by 53% and prices by 3%; however, they do not affect the extensive margins of export. The magnitudes of the exportside effect that we estimate are larger than the import side effects. Given the relatively lower levels of existing Swiss agricultural exports visàvis imports, the larger exportside effect of an FTA is not surprising. That Swiss FTAs increase export prices is consistent with the fact that Swiss exports are of a higher average quality and command a price premium. However, it is also consistent with the idea that the cost savings from lower tariffs may not be fully passed through to domestic consumers but are instead partially appropriated by foreign suppliers. Our work makes two key contributions to the literature. Existing studies on the effects of Swiss FTAs on trade patterns primarily focus on the aggregate economy.3 For instance, Bergstrand and Baier (2010) show that the Swiss–Mexico FTA of 2001 increased bilateral trade by approximately 37% after just 4 years in place. Nussbaumer's (2017) analysis of 20 Swiss FTAs using data on exports and imports from 1993 to 2014 provides descriptive evidence that points towards a general positive trade effect of FTAs, but the empirical estimates are inconclusive. According to Imhof (2021), Swiss FTAs have no effect on import quality and variety but decreases qualityadjusted prices. We contribute to this stream of findings by assessing the impact of FTAs specifically on agriculture, given the high levels of protection that typically characterise this sector. In this regard, our work is similar to Kohler(2016), who examines the effect of complete liberalisation in cheese between Switzerland and the EU on the Swiss cheese trade. Although the results in Kohler(2016) are positive, they paint a fuzzy picture and do not rule out the possibility that the FTA effect is null. There is also the work by Copenhagen Economics(2016), whose primary focus on EU FTAs offers an assessment of the effects of SwissEU FTAs in the agricultural sector. While relevant, this work is limited to Swiss FTAs with the EU. Our work thus differs from those of Kohler(2016) and Copenhagen Economics(2016) on two fronts: We focus on all agricultural products and consider all Swiss FTAs. Furthermore, Swiss FTAs often distinguish between basic agricultural products and processed agricultural 1335 products, a distinction that has not been incorporated into any ex post assessments. Our work fills this gap. Our second contribution extends beyond the direct trade effects of Swiss FTAs on Swiss imports to consider the broader network of trade relationships involving Switzerland's partners. Many of Switzerland's trade partners maintain bilateral agreements with third countries outside Switzerland. For example, while Switzerland has an FTA with the EU, the EU also holds FTAs with countries such as the Mediterranean basin, Canada, Mexico, Singapore and Chile. Whether these thirdcountry agreements enhance or divert trade away from Switzerland is an empirical question that remains underexplored in the existing literature. The increasing overlap of trade agreements presents both challenges and opportunities. Overlapping agreements can raise trade costs due to the complexity of managing multiple trade rules and regulatory standards. Conversely, countries connected through several FTAs may experience stronger integration and regulatory harmonisation, potentially reducing trade costs. In this context, our study contributes to a growing body of literature examining the interaction between overlapping FTAs and their effects on agricultural trade (e.g., Jafari etal.2023). Our analysis and findings hold important implications for policymaking, particularly in the agricultural sector. Historically, agriculture has been treated as a special sector, often exempt from certain provisions in trade agreements. However, recent trends suggest a shift towards integrating agriculture into broader trade frameworks. A report by the Organisation for Economic Cooperation and Development (ThompsonLipponen and Greenville2019) indicates that the number of trade agreements excluding agriculture has stagnated. Only a few agreements now exclude agriculture entirely, with an increasing tendency to address agricultural trade within the general provisions of agreements rather than in dedicated chapters. Given these developments, our study is timely in assessing the effectiveness of these provisions for agriculture. Furthermore, our attempt to provide evidence for the case of a highly tradedependent economy, such as Switzerland, is important, as there may be crucial policy implications for future agreements. Moreover, our ex post analyses offer a basis for comparison with ex ante simulations conducted by government agencies, such as the Swiss Federal Office of Agriculture (FOAG). This comparison can help the FOAG evaluate whether the anticipated benefits of FTAs have been realised and identify unintended consequences or areas for policy improvement. As agriculture continues to converge with general trade policy, such evidence is critical for refining strategies to support the sector effectively. The structure of the paper is as follows. Section2 provides the conceptual and theoretical background that frames our analyses and aids in interpreting the empirical findings. Section3 discusses the empirical framework employed in the study. In Section 4, we present the data and highlight the key stylised facts relevant to our analysis. We present and discuss the empirical findings in Section5. In Section6, we extend our analysis of Swiss imports to Swiss exports. Finally, Section7 concludes the paper and offers policy implications based on our findings. 2 | Conceptual and Theoretical Considerations In this section, we present the conceptual basis for our analyses. This provides structure for our work, guides our a priori expectations and helps us to discuss our empirical findings. We then present a concise theoretical overview of the gravity model, which serves as the basis for our empirical analyses. 2.1 | Conceptual Background: The Economics of Trade Agreements Standard microeconomic theory predicts that trade agreements generate termsoftrade gains for member countries. To illustrate this, we provide a simplified framework for analysing these effects in a small open economy within a partial equilibrium setting (see also Plummer etal.2011). SectionA.1 in the appendix offers a comprehensive discussion of the microeconomic foundations and mechanisms underlying trade agreements, including their theoretical underpinnings and the key factors that drive their effects. The small country assumption is appropriate in this context, as Switzerland's international market influence is relatively modest, accounting for just 1.67% of global merchandise imports and 2.96% of global imports of commercial services, which together represent 1.9% of total global merchandise and commercial services imports (Zimmermann 2023). FigureA1 depicts the domestic market for a specific good in a country preparing to join an FTA. In the end, two main predictions emerge from this framework and set the basis for the rest of our work: We expect the presence of an FTA to (i) increase import quantities and (ii) lower import prices. In the next subsection, we explain how we intend to test this expectation empirically. In this paper, we focus on the direct trade creation effects of FTAs. We limit the theoretical exposition to tariff reductions, as these remain a central feature of FTAs. However, it is important to note that recent FTAs have become deeper and more comprehensive, encompassing not only tariff cuts but also the liberalisation of NTMs and administrative procedures. These broader provisions, although crucial, are outside the scope of our analysis. Another observation beyond the scope of the current paper is the effect of trade diversion, which occurs when imports previously sourced from the more efficient outsider are displaced by imports from the less efficient but now cheaper FTA partner country. The theoretical prediction that FTAs increase trade carries important welfare implications for different economic agents in the home country. As a result of lower import prices, producer welfare declines because domestic producers receive lower prices for their goods. However, the reduction in domestic prices benefits consumers, increasing their surpluses and available product varieties and making them better off. The government also loses some tariff revenue, and the net welfare effect depends on efficiency gains in other sectors of the economy. Although these nondirect effects are relevant, they are not the focus of this study. Additionally, as we focus on FTAs, which are reciprocal by definition, we exclude unilateral trade preferences granted under the Generalised System of Preferences. On reciprocal versus unilateral trade liberalisation in the Swiss context, Zimmermann (2023) offers a broad discussion, while Ritzel and Kohler(2017) provide an analysis specific to the agricultural sector. 1336 Kyklos, 2025 2.2 | Theoretical Framework Our starting point is the structural gravity equation. Gravity equations are expenditure functions that indicate how consumers allocate their spending across countries when faced with trade cost constraints. It remains the workhorse model for ex post analysis of both the partial and general equilibrium effects of trade agreements (Larch and Yotov2024). In its basic form, the model predicts that bigger countries trade more with each other and that trade decreases with bilateral distance. For a model that was disconnected from economic theory until the 21st century, several theoretical models now yield predictions that are close to gravity. For our case, we adopt the productspecific version of the ArmingtonCES specification, as in Anderson and Van Wincoop(2003), as follows4: where Xodpt denotes exports of product p from origin (i.e., exporter) country o to destination (i.e., importer) country d in year t. Edpt is the import demand of p in d, which is usually proxied by gross domestic product (GDP). Yopt is the level of domestic production in o of p . Ypt is aggregate world production of p. The righthand side of Equation(1) is a product of two ratios. The first ratio is the predicted trade flow under free trade, and the second ratio in brackets captures exogenous bilateral trade costs. The trade cost term consists of three components: (i) the numerator, 𝜏odpt , is the bilateral trade cost between o and d for product p ; (ii) the denominator contains two structural terms, Πopt and 𝜆dpt , that measure the ease of market access for o and d ; (iii) 𝜎pt is the elasticity of substitution parameter. Our interest lies in 𝜏odpt , as it allows us to show how FTAs modify predicted costless trade. We model 𝜏odpt as the following loglinear function of observed trade frictions, including FTAs, NTMs, bilateral tariffs and a vector Ωod of timeinvariant traditional gravity covariates (including bilateral distance, and dummies for sharing a common language, and sharing a common border): 3 | Empirical Application In this section, we specify our econometric models and describe how we estimate the average and heterogeneous effects of FTAs on Swiss agricultural imports. 3.1 | Econometric Specification To assess the average effect of Swiss FTAs and the number of competing FTAs that Swiss trade partners have with other thirdcountries on different margins of Swiss agricultural imports, we estimate the following generic reducedform gravity equation: where o is the origin country (i.e., the country of production), p is the HS6digit product, and t is time measured in years. Xopt is the outcome variable, which varies depending on the import margin under consideration. FTAot is a dummy variable that takes the value 1 if there exists a FTA between Switzerland and o in year t , and 0 otherwise. 𝛽1 captures the effect of the presence of an FTA between country o and Switzerland in year t on agricultural imports, holding constant other factors that might influence trade. Using an FTA dummy, we capture the average effect of FTAs on agricultural imports, abstracting from the complexities of specific agricultural concessions or productlevel commitments. This allows us to estimate trade effects without requiring detailed productspecific data. The FTA dummy implicitly reflects the reduction in trade costs, capturing the combined effect of all tradefacilitating measures under an FTA, including, where relevant, tariff preferences, quota arrangements and reductions in nontariff barriers.5 𝛽2 captures the effect of thirdcountry FTAs that do not involve Switzerland. This accounts for such FTAs as those between the EU and South Korea, the EU and Türkiye, among others. ThirdCountryFTAd≠CHE ot is the number of other FTAs owned by country o excluding Switzerland. GDPot is the timevarying gross domestic product of the origin country. NTMopt captures the number of originand productspecific NTMs imposed on imports. Tariffopt is the applied ad valorem (bilateral) tariffs charged on imports of product p from country o in year t . 𝜆pt and Πop are product–time and origin–product fixed effects that control for the multilateral resistance terms that are typical of structural gravity models. Another important distortionary trade policy tool frequently used in Switzerland is the tariff rate quota (TRQ) system (Hillen2019). TRQs allow a predetermined quantity of a product to be imported at lower tariffs (inquota duty) while imposing higher tariffs on imports exceeding this quota (outofquota duty). They are often applied during specific periods within the year, particularly during domestic supply seasons, to protect local producers. Due to the annual nature of our dataset, however, we are unable to account for the intrayear variation in TRQs. Nevertheless, the inclusion of product–year fixed effects in our estimations accounts for their impact, as TRQs are applied on a productspecific basis. 𝜖opt is the error term. Our estimation equation is a loglinearised form of Equation (1) that embeds Equation (2). However, there are a few issues that are worth highlighting, given that at first glance, Equation(3) does not look exactly like the theoretical specification in Equation(1). In our setup, Switzerland is the only importing country, so the destination index d is redundant and is dropped from the empirical specification for simplicity. For this same reason, the inclusion of origin product fixed effects Πop absorbs all the timeinvariant traditional gravity variables contained in the vector Ωod in Equation(2). Because d is redundant, the dimensions of the countrypair variables included in vector Ωod reduce to Ωo , which is further embedded in Πop . Nonetheless, bilateral fixed effects—in our case Πop —are better measures of bilateral trade costs than the standard set of timeinvariant traditional gravity variables (Egger and Nigai2015; Agnosteva etal.2019; Fiankor etal.2021). The multilateral resistance terms Πopt and 𝜆dpt in Equation(1) reduce to Πop and 𝜆pt in the empirical specification. 𝜆dpt simplifies to 𝜆pt because d is redundant, but we resort to Πop in the empirical estimation because allowing the origin (1) X odpt = Y opt E dpt Y pt (𝜏odpt Π opt 𝜆 dpt ) 1−𝜎pt (2) 𝜏 odpt =FTA𝛽1 odtThird Country FTAd≠CHE,𝛽2 odt NTM3 odptTariff𝛽4 odptexp 7 ∑ n=5 𝛽nΩ od (3) Xopt =𝛽0+𝛽1FTAot +𝛽2Third Country FTA d≠CHE ot +𝛽3logGDP ot +𝛽 4 NTM opt +𝛽 5 log ( 1+Tariff opt) +𝜆 pt +Π op +𝜖 opt 1337 country fixed effects to vary over time (as in Πopt ) would result in perfect collinearity with our variables of interest, FTAot and ThirdCountryFTAd ≠ CHE ot . 3.2 | Defining Different Measures of Xopt In this study, we are interested in how FTAs affect different margins of import adjustments. This is important, as different margins of trade may adjust differently when faced with trade costs. We define five different margins of imports. The first three margins come directly from our theoretical framework in Figure A1, in which we illustrate how tariffs are predicted to affect import quantities and prices. We refer to these margins as the intensive margin of import adjustment and define them as follows: 1. The value of imports in CHF of product p from country o in year t , that is, Import valueopt . 2. The quantity of imports in kilograms of product p from country o in year t , that is, Import quantityopt . The entry into force of an FTA reduces trade costs for partners involved in the trading relationship. The exporters in the foreign country must no longer bear the costs of tariffs and other NTMs that were liberalised as part of the FTA. In return, this may reduce the prices of imports, as producers and other actors along the value chain no longer need to bear the extra costs of production and trade. To test this prediction, we define an import price margin: 3. The price—measured as unit values in CHF/kg—of imports of product p from country o in year t , that is, Import priceopt . The three outcome variables we consider focus on absolute trade values or quantities. Thus, our estimates provide insight into the size of the change in the value or quantity of Swiss imports in response to an FTA. However, it is possible that the expansion of trade may manifest not only as increased values or quantities of existing products or importers but also in other ways. For instance, new exporters may enter the Swiss import market. The reduction in trade costs as part of the FTA should also reduce the number of exporters that exit the Swiss market. These trade measures are often referred to as extensive margins. We define these margins as follows: 4. The probability of imports of product p from country o in year t , that is, Pr(Vopt > 0) . 5. The probability that imports of product p from country o cease in year t , that is, Pr( Exit opt >0 ) . 3.3 | Estimation Procedure Depending on the outcome variable, we estimate Equation(3) using different estimators. On the effect of FTAs on import values and import quantities, we use the Poisson pseudomaximum likelihood (PPML) estimator. The PPML estimator's loglinear objective function allows us to specify the estimation equation in its multiplicative form without logtransforming the dependent variable and is consistent under heteroscedasticity (Silva and Tenreyro2006). Because import prices are never zero, we estimate the effect of FTAs on import prices using ordinary least squares (OLS). Regarding the effect of FTAs on the probability of trade and market exit, we estimate a linear probability model (LPM). We employ the LPM for practical reasons, as it allows for a straightforward interpretation of the coefficients as marginal effects and facilitates the inclusion of highdimensional fixed effects without encountering the incidental parameter problem typical of many other nonlinear models. That notwithstanding, we also estimate both probit and logit models to ensure that the model choice does not drive our findings. 3.4 | Identification Strategy Endogeneity has been a major obstacle in gravity models. The sources of the problem are very clear, often arising from reverse causality and/or omitted factors that simultaneously affect trade and the probability of signing an agreement.6 Due to its intuitive appeal and easy implementation, the leading method to handle endogeneity of FTAs is that of Baier and Bergstrand(2007), who, consistent with the approach to control unobserved timeinvariant heterogeneity with panel data by Wooldridge(2010), propose the use of bilateral fixed effects, thus controlling for most of the unobserved correlation between the endogenous FTAs and the error term in gravity models (Larch and Yotov2024). In our onecountry case, the originproduct fixed effects, Πop , capture all bilateral variations. As such, threats to identification due to endogeneity are addressed using standard approaches in the literature. Nevertheless, we interpret our findings as associations rather than causal estimates. This is because in our single importing country setting, we cannot entirely rule out the additional effect of other origin–time specific effects, including climate change and extreme weather events. Our variable of interest is identified by the country and time variation in the agreements that entered into force during the study period. 4 | Data Our empirical analyses depend on data from two main sources: data on Swiss FTAs and data on Swiss bilateral trade, as detailed below. 4.1 | FTAs Our primary data source on Swiss FTAs is the State Secretary of Economic Affairs (SECO 2023). In addition to the EFTA Convention and the FTA with the European Union, Switzerland currently has a network of 33 FTAs with 43 partners. Figure1 illustrates the network of partner countries with which Switzerland has FTAs. In contrast to FTAs concluded jointly as the EFTA bloc, agricultural concessions are often granted in separate bilateral agricultural agreements between Switzerland and its trading partners. For instance, the agreement concerning trade in agricultural products between Albania and Switzerland was concluded following the FTA between Albania and the EFTA countries. These agreements are designed to address the specificities of agricultural trade, which often involves more complex regulatory and tariff structures than trade in industrial goods. These can take the form of TRQs, rebates or price 1338 Kyklos, 2025 compensation mechanisms. Unlike FTAs for industrial goods, which generally ensure the full elimination of tariffs, agricultural agreements feature more nuanced concessions. Tariffs on agricultural products are significantly higher than those on industrial goods. According to the WTO, the latest ad valorem equivalents of the tradeweighted average MFN applied rates for 2021 are 24.8% for agricultural products compared with only 0.7% for nonagricultural imports (Zimmermann2023). As these agriculturespecific agreements do not involve the same level of liberalisation, their trade effects may also be limited in comparison to industrial FTAs. Figure2 depicts the years in which the agreements entered into force. It also illustrates the variations that we exploit in our empirical analysis. According to the figure, different countries signed the agreements with Switzerland at different times, allowing our identification strategy to exploit this time and country variation in the entry into force of the agreements. Aside from the EFTA Convention and the agreements with the EU, which date far back to the 1960s and 1970s, the oldest agreement is the Swiss–Türkiye FTA, which has since been modernised, with the updated agreement becoming active in October 2021. FTA negotiations are currently underway with Kosovo, India, Vietnam, Malaysia and the MERCOSUR, while negotiations with the Russia–Belarus–Kazakhstan Customs Union have been suspended. To account for thirdcountry bilateral agreements that are outside the control of Switzerland, we use data from the regional trade agreement database maintained by Egger and Larch(2008) and count the number of FTAs these countries are signatories to in a year that do not include Switzerland. FIGURE 1 | Swiss FTAs in 2022. Note: The map shows which countries have a free trade agreement with Switzerland in 2022. The bilateral FTAs include those signed bilaterally with Switzerland and those signed together as part of the EFTA. Source: The Swiss FTA Monitor (SECO2023). [Colour figure can be viewed at wileyonlinelibrary.com] FIGURE 2 | Swiss free trade partners in 2022 and years of entry into force of the agreement. Note: For clarity of presentation, we exclude the EFTA Convention which came into force in 1960 and the FTA with the European Community members in 1973. SACU stands for the South African Customs Union and includes South Africa, Botswana, Eswatini, Lesotho and Namibia. CAS represents the Central American States of Costa Rica, Guatemala, Honduras and Panama. GCC represents the Gulf Cooperation Council members: Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and the United Arab Emirates. Source: SECO(2023). [Colour figure can be viewed at wileyonlinelibrary.com] 1339 4.2 | Agricultural Trade Data Our analysis focuses on the agricultural sector, defined according to the Swiss Federal Office for Agriculture to include HS01–H24 (excluding fish and fish products, HS03), 290543, 290544, 3301, 3501–3505, 380910, 382360, 4101–4103, 4301, 5001–5003, 5101–5103, 5201–5203, 5301 and 5302. We analyse Swiss customs trade data (SwissImpex2023) at the level of the partner country and HS6digit products over time. It includes data on import quantities in kilograms (kg) and import values in Swiss Francs (CHF). A preliminary glance at the data confirms that most Swiss trade occurs with FTA partners, with this trend increasing over time (Figure3). Furthermore, as shown in TableA2 in the appendix, the majority of Swiss bilateral trade is with EU members. However, FTAs with nonEU countries also play a significant role in Swiss trade policy. In aggregate, approximately 84% of Swiss trade occurs with FTA partners, while only about 16% of Swiss trade occurs with countries that do not have an FTA with Switzerland. FiguresA2 and A3 in the appendix map the geographic distribution of Swiss trade flows, highlighting diverse trading partners. European Union member states dominate Swiss imports, with Germany, Italy and the Netherlands as key suppliers. Outside Europe, the United States, China and Brazil are notable trade partners, while imports from developing regions such as Africa and South America focus on primary products, with Morocco, Colombia, Côte d'Ivoire and Ghana making significant contributions. Switzerland's exports, meanwhile, are concentrated in European markets, particularly Germany, Austria and the United Kingdom, with the United States and Japan being major nonEuropean partners. Switzerland also exports to emerging markets, such as China, India and Brazil. The composition of traded products is equally important. Swiss imports are dominated by primary agricultural goods, with high shares in fruits and nuts (HS08) and vegetables (HS07), reflecting dependence on foreign supplies. Cereals (HS10) and oil seeds (HS12) also have significant import shares, with minimal exports. Other sectors, such as beverages, spirits and vinegar (HS22) and dairy produce (HS04), reflect substantial imports. By contrast, Swiss exports are concentrated on highvalue, processed agricultural products. Beverages, spirits and vinegar (HS22) lead the export flows, followed by dairy products (HS04) and preparations of cereals (HS19). Niche sectors, such as cocoa and cocoa preparations (HS18) and miscellaneous edible preparations (HS21), highlight Switzerland's competitive advantage in highquality, valueadded production. These patterns reveal Switzerland's strategy of importing raw materials while excelling in processed, highvalue exports in niche global markets. Recent advancements in the structural gravity literature emphasise the importance of including intranational trade flows, as they allow the identification of international trade costs relative to domestic trade costs (Yotov etal.2016; Yotov2022). However, due to data limitations, most empirical applications, including ours, rely solely on international trade data. In our case, we lack domestic trade data at the HS6 digit level for Switzerland. Without a domestic trade benchmark, we cannot fully assess whether increased international trade flows under FTAs replace or complement domestic production. This is a key issue in the agricultural sector, where domestic production often meets a share of demand and may respond differently to FTAs than international trade. 4.3 | Auxiliary Data SwissImpex(2023) also provides access to data on specific tariffs in CHF/kg imposed on imports from partner countries over time. Switzerland stands out in its tariff application as one of the few countries that explicitly express tariffs in specific or perunit terms. Given that these tariffs are fixed amounts per unit rather than a percentage of value, their impact depends on the price of the product. As such, perunit tariffs place a heavier burden on lower priced items within a given tariff line. Developing countries, which typically export at lower prices, face higher ad valorem equivalents for the same specific tariff compared with highincome countries. As a result, while specific tariffs may appear nondiscriminatory as MFN measures, they can effectively discriminate against developing countries' exports (Chowdhury2012; Fiankor etal.2024). However, the tariffs are only reported when trade flows are observed. Thus, when we introduce zero trade observations, information on tariffs is missing. To deal with this situation, we resort to the MAcMapHS6 database maintained by the CEPII and the International Trade Center (Guimbard etal.2012). As the MAcMap dataset is available only for every third year between 2007 and 2019, we interpolate using data from previous years whenever we encounter missing data. While this is limiting, there remain substantial challenges with the quality of publicly reported tariff data, especially when multiple countries are concerned. Teti(2023) highlights that standard sources for tariff data suffer from significant measurement errors due to misreporting and the resulting false imputations, which lead to artificial spikes in bilateral time series data and, consequently, cause massive inaccuracies in the measurements. We also include data on NTMs, which are policy measures other than tariffs that affect international trade by affecting quantities, prices or both (UNCTAD2019). As tariffs have been significantly liberalised since the establishment of the WTO, there has been a concurrent rise in standardlike NTMs as tools for market access. Therefore, it is crucial to account for these NTMs FIGURE 3 | Swiss agricultural imports by FTA status of the partners. [Colour figure can be viewed at wileyonlinelibrary.com] 1340 Kyklos, 2025 in our estimations. Given that the proliferation and increasing relevance of NTMs, including those in Switzerland, are driven by sanitary and phytosanitary (SPS) and technical barriers to trade (TBT) measures (Irek2022; Fiankor2023b), we account for NTMs using the aggregate productlevel number of SPS and TBT measures imposed by Switzerland on imports from an origin country each year. The data on NTMs are accessed from the WTO's comprehensive data on NTM notifications via the Trade Analysis and Information System (UNCTAD2019). Data on GDP are accessed from the World Bank World Development Indicators. Our final estimation sample covers imports from 202 countries (see TableA3), 730 HS6digit products, over 19 years (i.e., 2004– 2022). Summary statistics on all the variables included in the estimation are presented in TableA4 in the appendix. 5 | Results and Discussion We present and discuss the results of our analysis in this section. We first present the average effects before assessing whether and to what extent they are heterogeneous along the three dimensions, and end by assessing dynamic effects. 5.1 | Baseline Findings We present the average effect of Swiss FTAs on imports in Table 1, with each column depicting one of the five import margins. In column (1), we find that, on average, the presence of an FTA leads to an 8.5% increase in import values. In terms of magnitude, this coefficient translates into an effect size of 8.75%.7 In column (2), we find no statistically significant effect of FTAs on import quantities. In column (3), we find a negative effect of FTAs on import prices; specifically, FTAs decrease import prices by 3.4%. At the extensive margin, we find that FTAs increase the probability of trade by two percentage points and decrease the probability of market exit rates by one percentage point.8 That we do not observe a statistically significant change in import quantities is inconsistent with the theoretical framework in FigureA1. However, the finding that FTAs increase import values and import probabilities and lower import prices and market exit rates confirms our a priori expectations. These findings are also consistent with the existing empirical literature. A recent metaanalysis of the effects of trade agreements on agricultural trade based on 61 empirical studies and 1961 effect sizes (Afesorgbor etal.2024) find that trade agreements generally have a positive and significant effect on agricultural and food trade. The fact that FTAs do not lead to an increase in import quantities suggests that the negative price effect outweighs the quantity effect. This phenomenon is consistent with the idea that the cost savings from lower tariffs may not be fully passed through to domestic consumers but may be partially captured by foreign suppliers. Additionally, the reduction in trade costs may incentivise the entry of higher quality goods, which are priced higher, increasing import values without a proportional rise in quantities. On the effects of thirdcountry agreements, we find that an extra agreement signed by a partner country that excludes TABLE 1 | The effect of Swiss FTAs on different margins of Swiss agricultural imports. Outcome variable Intensive margin Extensive margin Import values Import quantity Import prices Import probability Import market exit (1) (2) (3) (4) (5) FTAot 0.085*** −0.020 −0.032*** 0.021*** −0.010*** (0.029) (0.037) (0.010) (0.002) (0.003) Third Country FTAd≠CHE ot 0.000 −0.003* −0.001** 0.001*** −0.001*** (0.002) (0.002) (0.000) (0.000) (0.000) logGDPot 0.458*** 0.319*** 0.154*** 0.034*** −0.018*** (0.031) (0.034) (0.010) (0.002) (0.003) NTMopt −0.061*** −0.039*** −0.000 −0.002*** 0.002*** (0.006) (0.007) (0.002) (0.000) (0.000) log( 1+Tariff opt) −0.000 0.000 0.000 0.000 0.000 (0.000) (0.000) (0.000) (0.000) (0.000) Product–time FE Yes Yes Yes Yes Yes Origin–product FE Yes Yes Yes Yes Yes Observations 587,108 587,108 206,194 587,108 484,345 Estimator PPML PPML OLS LPM LPM Note: ***, ** and * denote significance at 1%, 5% and 10%, respectively. Intercepts are included but are not reported. Cluster–robust standard errors are in parentheses. The differences in the number of observations across columns are due to differences in estimators. Columns (1), (2), (4) and (5) account for zero trade observations, which are dropped in column (3). The number of observations in column (5) differs because countries exporting to a product destination market every year are excluded from the exit analysis. 1347 similar patterns. For example, Fiankor(2023a) shows that a Swiss firm exported the same HS8digit product, ‘hard cheese’ (HS 0406 9099), to 18 countries, with freeonboard (FOB) prices ranging from 10.70 CHF/kg in Peru to 16.00 CHF/kg in South Korea. While such price differences may arise from exporters arbitrarily adjusting markups, they may also reflect quality variations, such as more durable packaging for highercost markets. Unlike raw agricultural products, where quality differentiation is limited, Swiss agrifood exports are largely processed products where quality sorting is common. This suggests that Swiss exporters may tailor product quality across destinations. At the extensive margin, we find no statistically significant effect of FTAs. As to whether the effects we find are heterogeneous across basic and processed products, we show in TableA5 that this is not the case for exports. In relation to the effects we estimate for imports, the exportside effects are larger in economic magnitude. What explains the asymmetry in the size of the trade effects for exports and imports? Although our estimates cannot provide direct answers, we can offer plausible reasons based on the policy environment. First, it is important to note that these average effects are conditional on the value of existing imports and exports between trade partners at the inception of the agreement. Second, the concessions granted by Switzerland's trade partners are often more substantial, as these partners typically have fewer defensive positions in agriculture. In contrast, Swiss agricultural policy is highly protectionist, with significant tariffs and NTMs limiting the scope of liberalisation on imports. As a result, the relative gains from FTAs on imports may be smaller, given Switzerland's constrained concessions. Third, the nature of the traded products themselves plays a key role. Swiss agricultural exports, such as cheese and other highvalue processed goods such as coffee and chocolate, are often niche products with strong international demand. FTAs enhance market access, leading to disproportionately large gains in export value and quantity. By contrast, strong protections for sensitive domestic products limit the potential for significant import increases. Lastly, NTMs further contribute to this asymmetry. Whereas FTAs reduce tariffs, NTMs—such as TRQs, quality standards and certification requirements—remain particularly restrictive for agricultural imports into Switzerland (Fiankor etal.2025; Fiankor and Shingal2025). These constraints can dampen import growth despite tariff reductions. Conversely, Swiss exports may adapt more readily to the partner country's standards, resulting in greater export increases. 7 | Conclusions The WTO has been making little progress in multilateral trade liberalisation for years. As a result, since the Doha round, we have observed a rise in the number of bilateral FTAs. Switzerland has kept pace with this trend, signing numerous FTAs. In 2024, Switzerland had in place a network of 33 FTAs with 43 partners. Among the primary goals of these agreements is to facilitate trade among member countries, allowing consumers to benefit from lower prices and increased product variety. The aim of this paper is to assess whether these objectives are achieved in practice. Specifically, we assess the effect of Swiss FTAs on different margins of agricultural imports over the period between 2004 and 2022. Furthermore, because partner countries often sign additional FTAs with other countries, we also assess how the network of FTAs Swiss partners are involved in influences their exports to Switzerland. Empirically, we situate our analysis within a gravity framework and estimate a reducedform gravity model. Our findings show that Swiss FTAs increase imports, decrease import prices and reduce market exit rates. These findings are, however, heterogeneous along different dimensions. Swiss FTAs increase the import values and quantities of raw products but decrease the imports of processed products. We find further heterogeneous effects across HS2digit product sectors and for individual agreements. Thus, while the average effects of Swiss FTAs on imports and product prices are in line with our theoretical priors and the available empirical evidence, the heterogeneities we find also highlight the importance of examining different sectors and agreements and support our empirical choice of going beyond just the average effects. Nevertheless, these heterogeneities also suggest that in some cases, the findings are inconsistent with theoretical priors. For instance, in some cases, we find that FTAs decrease imports. Our empirical findings are not without limitations. The existence of the agreement only solves the trade barrier issue but does not reflect the quality of domestic institutions and traderelated infrastructure or local shocks (e.g., climate change and extreme weather events, political instability and economic crisis) in the product–origin country. As long as these factors remain countryand timespecific, they cannot be captured by our model specifications. In this case, our FTA effects may be biased, as the FTA variable picks up other confounding factors that drive trade. Recent reviews of the regional trade agreement literature, such as those by Larch and Yotov(2024) and the metaanalysis by Afesorgbor etal.(2024), show that although trade agreements generally enhance trade, in cases of individual agreements or products, the empirical findings do not always align with the theoretical predictions. As such, even if Swiss FTAs generally achieve the intended trade effects for which they were signed, policymakers should keep these associated heterogeneities in mind when using average FTA estimates for counterfactual analysis and/or trade negotiations. Acknowledgements This project benefited from extensive feedback from members of the Trade Relations Unit of the Swiss Federal Office for Agriculture (FOAG). We are especially grateful for the insightful comments provided by Axel Tonini and Yvan Decreux. We also thank the handling editor, Christoph A. Schaltegger, and two anonymous referees for their valuable suggestions. Open Access funding enabled and organized by Projekt DEAL. Data Availability Statement The data that support the findings of this study are openly available from publicly available data sources that are all mentioned in the manuscript. 1348 Kyklos, 2025 Endnotes 1 The terms big country and small country here is used without prejudice to the economic size of the countries. The small country case references a situation where a country's imports constitute a very small share of the world market and, therefore, do not influence world market prices. In this context, an existing study that also examines the agricultural sector in small economies is Copenhagen Economics (2016). However, our study differs in focus: while Copenhagen Economics(2016) analyses trade relationships between the EU common market (a large economy) and its partners, we examine trade relationships between a small country and its trade partners. 2 There are also trade programs such as the Generalised System of Preferences (GSP) designed to promote economic growth in developing countries by giving them preferential access to the markets of developed countries. Under the GSP, selected goods from eligible developing countries can enter the importing country at reduced or zero tariff rates. The GSP grants developing countries nonreciprocal, preferential market access to developed countries through reduced or zero tariffs, unlike FTAs, which are reciprocal arrangements with mutual obligations. Our focus here is on reciprocal arrangements. 3 Much of the literature assessing the effects of trade agreements focuses on multiple countries (e.g., Baier and Bergstrand2007; Baier etal. 2019; Sun and Reed2010; Jean and Bureau 2016). However, a smaller subset of studies examines the impacts of trade agreements on specific countries, including Japan (Yamanouchi 2019; Ando etal.2022), Canada (McDougall2020), India (Jagdambe and Kannan2020), and the United States (Ajewole etal.2022). Our work contributes to this second stream of literature by providing a focused analysis on Switzerland. 4 Two main assumptions underlie the model. First, goods are differentiated by country of origin (i.e., the Armington assumption) such that two goods of the same kind coming from different countries are imperfect substitutes, for example, German, and Italian cheese are distinct goods in the composite group cheese. Thus, the reason Swiss consumers purchase foreign goods is that they are different from the ones produced at home. Other motivations may exist for purchasing foreign goods, for example, in a Ricardian world, foreign goods will be purchased because they are produced more efficiently abroad than at home. Second, consumer preferences are identical and homothetic across countries and captured by a constant elasticity of substitution (CES) utility function. Given that the formal derivation of the gravity equation is now standard in the literature (see, e.g., Anderson and Van Wincoop2003; Yotov etal.2016), we do not reproduce the derivation. 5 This choice is motivated by challenges in obtaining detailed data on productlevel preferential margins across multiple countries. Nevertheless, using an FTA dummy enables us to consider the broader context of FTAs, which often involve not only tariff preferences but also quota arrangements and reductions in other nontariff and quota barriers. This approach is standard in the trade literature (Baier and Bergstrand2007; Baier etal.2019; Egger and Larch2008; Egger etal.2022) and offers a practical way to estimate trade effects without requiring detailed data on productspecific tariff reductions or concessions, which are often difficult to compile across multiple agreements. The limitation, however, is that our model abstracts from the complexity of individual concessions within FTAs, and our effects reflect the cumulative impact of these individual concessions. 6 Addressing this concern using instrumental variable techniques is challenging because very often what determines the probability to sign a trade agreement also affects the volume of trade flows. The interested reader should refer to Larch and Yotov(2024) for a discussion of these issues. 7 The trade effect of an FTA can be calculated as [exp(β1) − 1] × 100. 8 We also estimate the effect of FTAs on the extensive margins using logit and probit models. The results presented in Table A5 of the Appendix are in line with our main findings in terms of direction, magnitude, and statistical significance. Thus, the choice of estimator does not influence our results. 9 Basic products are defined to include products of HS sections 01–14, excluding Section4.3, headings 0402–0406 and 0408, and subheading 0801.32, plus headings 1801, 1802, 2401, 5001, 5101 to 5103, 5201, 5202, 5301 and 5302. Everything else is considered a processed product. This definition was provided by the Swiss Federal Office of Agriculture (FOAG) based on the official definitions adopted by the Swiss Secretariat for Economic Affairs (SECO). 10 For cases in which the agreements are signed within a bloc such as SACU or the EU, we assess the effects at the countrylevel. For instance, for the effect of the EUSwitzerland association agreement, we estimate different effects for Croatia and Romania that joined the EU over the study period. Note that we are unable to estimate unique effects for the founding members of the EU as there is no variation in the FTA dummy for them over the study period. For members of the Gulf Cooperation Council, we also estimate countryspecific effects for Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and the United Arab Emirates. Thus, in essence, the variation we exploit here is more at the country level that at the agreement level. References Afesorgbor, S. K., D.- D. D. Fiankor, and B. A. Demena. 2024. “Do Regional Trade Agreements Affect AgriFood Trade? Evidence From a MetaAnalysis.” Applied Economic Perspectives and Policy 46, no. 2: 737–759. Agnosteva, D. E., J. E. Anderson, and Y. V. Yotov. 2019. “IntraNational Trade Costs: Assaying Regional Frictions.” European Economic Review 112: 32–50. Ajewole, K., J. Beckman, A. Gerval, W. Johnson, S. Morgan, and E. Sabala. 2022. “Do Free Trade Agreements Benefit Developing Countries? An Examination of US Agreements.” Technical Report, Economic Information Bulletin Number 240. Economic Research Service, United States Department of Agriculture. Anderson, J. E., and E. Van Wincoop. 2003. “Gravity With Gravitas: A Solution to the Border Puzzle.” American Economic Review 93, no. 1: 170–192. Ando, M., S. Urata, and K. Yamanouchi. 2022. “Do Japan's Free Trade Agreements Increase Its International Trade?” Journal of Economic Integration 37, no. 1: 1–29. Baier, S. L., and J. H. Bergstrand. 2007. “Do Free Trade Agreements Actually Increase Members' International Trade?” Journal of International Economics 71, no. 1: 72–95. Baier, S. L., Y. V. Yotov, and T. Zylkin. 2019. “On the Widely Differing Effects of Free Trade Agreements: Lessons From Twenty Years of Trade Integration.” Journal of International Economics 116: 206–226. Bergstrand, J. H., and S. L. Baier. 2010. “An Evaluation of Swiss Free Trade Agreements Using Matching Econometrics.” Aussenwirtschaft 65, no. 3: 239–250. Chowdhury, S. 2012. “The Discriminatory Nature of Specific Tariffs.” World Bank Economic Review 26, no. 1: 147–163. Copenhagen Economics. 2016. “Impacts of EU Trade Agreements on the Agricultural Sector.” Technical Report, Report for DG-Agri, European Commission, Luxembourg: Publications Office of the European Union. Egger, P., and M. Larch. 2008. “Interdependent Preferential Trade Agreement Memberships: An Empirical Analysis.” Journal of International Economics 76, no. 2: 384–399. Egger, P. H., M. Larch, and Y. V. Yotov. 2022. “Gravity Estimations With Interval Data: Revisiting the Impact of Free Trade Agreements.” Economica 89, no. 353: 44–61. 1349 Egger, P. H., and S. Nigai. 2015. “Structural Gravity With Dummies Only: Constrained ANOVAType Estimation of Gravity Models.” Journal of International Economics 97, no. 1: 86–99. Fiankor, D.- D. D. 2023a. “Distance to Destination and Export Price Variation Within AgriFood Firms.” European Review of Agricultural Economics 50, no. 2: 563–590. Fiankor, D.- D. D. 2023b. “Estimating Ad Valorem Equivalents of NonTariff Measures in Swiss Agriculture.” Agroscope Science 156: 1–35. Fiankor, D.- D. D., D. Curzi, and A. Olper. 2021. “Trade, Price and Quality Upgrading Effects of AgriFood Standards.” European Review of Agricultural Economics 48, no. 4: 835–877. Fiankor, D.- D. D., B. Dalheimer, D. Curzi, O. Hoffmeister, and B. Brümmer. 2024. “Does It Matter How We Ship the Good Apples Out? On Specific Tariffs, Transport Modes, and Agricultural Export Prices.” Agricultural Economics 55, no. 3: 498–514. Fiankor, D.- D. D., B. Dalheimer, and G. Mack. 2025. “Pesticide Regulatory Heterogeneity, Foreign Sourcing, and Global Agricultural Value Chains.” American Journal of Agricultural Economics 107, no. 2: 611–634. Fiankor, D.- D. D., and A. Shingal. 2025. “Pesticide Regulatory Homogeneity and Firms' Import Decisions: Evidence From EUSwiss AgriFood Trade.” Journal of Agricultural Economics. https:// doi. org/ 10. 1111/ 14779552. 12623 . Grant, J. H., and D. M. Lambert. 2008. “Do Regional Trade Agreements Increase Members' Agricultural Trade?” American Journal of Agricultural Economics 90, no. 3: 765–782. Guimbard, H., S. Jean, M. Mimouni, and X. Pichot. 2012. “Macmaphs6 2007, an Exhaustive and Consistent Measure of Applied Protection in 2007.” International Economics 130: 99–121. Harrigan, J., X. Ma, and V. Shlychkov. 2015. “Export Prices of US Firms.” Journal of International Economics 97, no. 1: 100–111. Hillen, J. 2019. “Market Integration and Market Efficiency Under Seasonal Tariff Rate Quotas.” Journal of Agricultural Economics 70, no. 3: 859–873. Imhof, P. 2021. “Switzerland's System of Free Trade Agreements: Assessing the Impact on Imported Goods.” Aussenwirtschaft 71, no. 1: 35–71. Irek, J. 2022. “Characterizing Swiss NTM Trade Policy for AgriFood Products: From Technical Barriers to Sustainability Standards.” Agroscope Science 148: 1–27. Jafari, Y., H. Engemann, and A. Zimmermann. 2023. “Food Trade and Regional Trade Agreements–a Network Perspective.” Food Policy 119: 102516. Jagdambe, S., and E. Kannan. 2020. “Effects of ASEANIndia Free Trade Agreement on Agricultural Trade: The Gravity Model Approach.” World Development Perspectives 19: 100212. Jean, S., and J.- C. Bureau. 2016. “Do Regional Trade Agreements Really Boost Trade? Evidence From Agricultural Products.” Review of World Economics 152: 477–499. Kohl, T. 2014. “Do We Really Know That Trade Agreements Increase Trade?” Review of World Economics 150: 443–469. Kohler, A. 2015. “Determinanten der Schweizer Agrarexporte–Eine Anwendung des Ökonomischen Gravitationsmodells.” Journal of SocioEconomics 8: 21–38. Kohler, A. 2016. “Contra Facta–Die Auswirkungen des Schweizer Käsefreihandels mit der EU.” Agroscope Science 10: 1–12. Larch, M., and Y. V. Yotov. 2024. “Estimating the Effects of Trade Agreements: Lessons From 60 Years of Methods and Data.” World Economy 47, no. 5: 1771–1799. Manova, K., and Z. Zhang. 2012. “Export Prices Across Firms and Destinations.” Quarterly Journal of Economics 127, no. 1: 379–436. Martin, J. 2012. “Markups, Quality, and Transport Costs.” European Economic Review 56, no. 4: 777–791. McDougall, B. 2020. “The Impacts of Free Trade Agreements on the Intensive and Extensive Margins of Canadian AgriFood Trade.” Master's Thesis, University of Guelph. Niu, Z., C. Liu, S. Gunessee, and C. Milner. 2018. “NonTariff and Overall Protection: Evidence Across Countries and Over Time.” Review of World Economics 154: 675–703. Nussbaumer, T. 2017. “A Study Case on the Caveats in the Measurement of FTAs Effect on Trade: Switzerland's Free Trade Agreements.” Aussenwirtschaft 68, no. 1: 139–167. Plummer, M. G., D. Cheong, and S. Hamanaka. 2011. Methodology for Impact Assessment of Free Trade Agreements. Asian Development Bank. Ritzel, C., and A. Kohler. 2017. “Protectionism, How Stupid Is This? The Causal Effect of Free Trade for the World's Poorest Countries: Evidence From a QuasiExperiment in Switzerland.” Journal of Policy Modeling 39, no. 6: 1007–1018. Ritzel, C., A. Möhring, and A. von Ow. 2024. “Vulnerability Assessment of Food Imports—Conceptual Framework and Empirical Application to the Case of Switzerland.” Heliyon 10, no. 5: e27058. Rodrik, D. 2018. “What Do Trade Agreements Really Do?” Journal of Economic Perspectives 32, no. 2: 73–90. SECO. 2023. “Free Trade Agreements.” Swiss State Secretariat for Economic Affairs. Retrieved December 20, 2024, from https:// www. seco. admin. ch/ seco/ en/ home/ Ausse nwir t scha f t spol itik_ Wir ts chaft l iche _ Zu sa m men a r b eit/ W irts ch a f t s be zi ehu ng en/ F r ei h a ndel s abko mmen. html. Silva, J. S., and S. Tenreyro. 2006. “The Log of Gravity.” Review of Economics and Statistics 88, no. 4: 641–658. Sun, L., and M. R. Reed. 2010. “Impacts of Free Trade Agreements on Agricultural Trade Creation and Trade Diversion.” American Journal of Agricultural Economics 92, no. 5: 1351–1363. SwissImpex. 2023. “SwissImpex Database.” Technical Report, Swiss Foreign Trade Data. Federal Office for Customs and Border Security (FOCBS). https:// www. gate. ezv. admin. ch/ swiss impex/ priva te/ expert. xhtml . Teti, F. 2023. “Missing Tariffs.” Technical Report, Collaborative Research Center Transregio 190 Discussion Paper No. 458. ThompsonLipponen, C., and J. Greenville. 2019. “The Evolution of the Treatment of Agriculture in Preferential Trade Agreements.” UNCTAD. 2019. International Classification of NonTariff Measures. UN. Wooldridge, J. M. 2010. Econometric Analysis of Cross Section and Panel Data. MIT Press. Yamanouchi, K. 2019. “Heterogeneous Impacts of Free Trade Agreements: The Case of Japan.” Asian Economic Papers 18, no. 2: 1–20. Yotov, Y. V. 2022. “On the Role of Domestic Trade Flows for Estimating the Gravity Model of Trade.” Contemporary Economic Policy 40, no. 3: 526–540. Yotov, Y. V., R. Piermartini, and M. Larch. 2016. An Advanced Guide to Trade Policy Analysis: The Structural Gravity Model. WTO iLibrary. Zimmermann, T. A. 2023. “A Case of Unilateral Trade Liberalization: The Autonomous Abolition of Industrial Tariffs by Switzerland in 2024.” Aussenwirtschaft 73, no. 1: 113–169. 1350 Kyklos, 2025 Appendix A A.1 | Tables TABLE A1 | HS2 product sectors and their import and export shares. HS2 product sector Import share (%) Export share (%) HS01: Animals, live 0.01 0.11 HS02: Meat 3.52 0.90 HS04: Dairy produce 3.91 6.43 HS05: Animal products, nes 0.76 3.89 HS06: Trees and other plants 3.78 0.08 HS07: Vegetables 8.66 0.22 HS08: Fruits and nuts 10.68 0.23 HS09: Coffee, tea, mate, spices 3.60 2.45 HS10: Cereals 5.99 0.13 HS11: Products of milling industry 2.95 0.32 HS12: Oil seeds 5.10 0.21 HS13: Lac; natural gums, resins 0.27 0.17 HS14: Vegetable plaiting materials 0.19 0.11 HS15: Animal, vegetable fats & oils 5.40 0.82 HS16: Preparations: meat, fish 0.52 0.04 HS17: Sugars & sugar confectionery 4.16 1.20 HS18: Cocoa & cocoa preparations 2.15 4.99 HS19: Preparations: cereals 1.05 6.69 HS20: Preparations: vegetables, fruits 6.99 4.28 HS21: Misc. edible preparations 3.40 6.39 HS22: Beverages, spirits, vinegar 13.80 51.88 HS23: Residues of food industry 8.61 4.59 HS24: Tobacco 1.11 1.49 HS29: Organic chemicals 0.34 0.01 HS33: Essential oils and resinoids 0.34 0.90 HS35: Albuminoidal substances 1.54 0.69 HS38: Misc. chemical products 0.77 0.03 (Continues) TABLE A2 | Swiss agricultural trade relationships with FTA and nonFTA partners in 2022. Partner Imports Exports Trade Share of trade (%) EFTA 137 85 222 0.78 EU 13,102 5409 18,511 65.33 FTA 2194 2737 4931 17.40 No FTA 2146 2526 4672 16.49 Total 17,579 10,757 28,336 100.00 Note: Trade is the sum of imports and exports. Imports, exports and trade values are in million CHF. Data used for the calculations come from SwissImpex. The ‘No FTA’ group is derived as the residual difference between the total reported trade flows and the trade values that fall within the three FTA groups. Furthermore, given that unilateral trade preferences are not FTAs, it is possible that the ‘No FTA’ group includes imports from developing and least developed countries that enjoy nonreciprocal preferential exports to Switzerland under the GSP scheme. HS2 product sector Import share (%) Export share (%) HS41: Raw hides and skins 0.00 0.68 HS43: Fur skins and artificial fur 0.00 0.00 HS50: Silk 0.00 0.00 HS51: Wool 0.03 0.03 HS52: Cotton 0.35 0.06 HS53: Other vegetable textile fibres 0.03 0.00 TABLE A1 | (Continued) 1351 TABLE A3 | List of countries included in the study. Aruba, Afghanistan, Angola, Albania, Andorra, Argentina, Armenia, American Samoa, Antigua and Barbuda, Australia, Austria, Azerbaijan, Burundi, Belgium, Benin, Burkina Faso, Bangladesh, Bulgaria, Bahrain, Bahamas, Bosnia and Herzegovina, Belarus, Belize, Bermuda, Bolivia, Brazil, Barbados, Brunei, Bhutan, Botswana, Central African Republic, Canada, Chile, China, Cote d'Ivoire, Cameroon, Democratic Republic of the Congo, The Republic of the Congo, Colombia, Comoros, Cape Verde, Costa Rica, Cuba, Curacao, Cayman Islands, Cyprus, Czech Republic, Germany, Djibouti, Dominica, Denmark, Dominican Republic, Ecuador, Egypt, Eritrea, Spain, Estonia, Ethiopia, Finland, France, Faroe Islands, Micronesia, Gabon, Georgia, Ghana, Gambia, Equatorial Guinea, Greece, Grenada, Greenland, Guatemala, Guyana, Hong Kong, Honduras, Croatia, Haiti, Hungary, Indonesia, India, Ireland, Iran, Iraq, Iceland, Israel, Italy, Jamaica, Jordan, Japan, Kazakhstan, Kenya, Kyrgyzstan, Cambodia, Saint Kitts and Nevis, South Korea, Kuwait, Laos, Lebanon, Liberia, Libya, American Samoa, Sri Lanka, Lesotho, Lithuania, Luxembourg, Latvia, Macao, Morocco, Moldova, Madagascar, Maldives, Mexico, Marshall Islands, North Macedonia, Mali, Malta, Myanmar, Montenegro, Mongolia, Northern Mariana Islands, Mozambique, Mauritania, Mauritius, Malawi, Malaysia, Namibia, New Caledonia, Niger, Nigeria, Nicaragua, Netherlands, Norway, Nepal, Nauru, New Zealand, Oman, Pakistan, Panama, Peru, Philippines, Papua New Guinea, Poland, Portugal, Paraguay, Palestine, French Polynesia, Qatar, Romania, Russian Federation, Rwanda, Saudi Arabia, Sudan, Senegal, Singapore, Solomon Islands, Sierra Leone, Slovenia, San Marino, Somalia, Serbia, South Sudan, Sao Tome and Principe, Suriname, Slovakia, Slovenia, Sweden, Swaziland, Seychelles, Syria, Turks and Caicos Islands, Chad, Togo, Thailand, Tajikistan, Turkmenistan, TimorLeste, Tonga, Trinidad and Tobago, Tunisia, Türkiye, Tuvalu, Tanzania, Uganda, Ukraine, Uruguay, United Kingdom, United States, United Araba Emirates, Uzbekistan, St. Vincent and the Grenadines, Venezuela, US Virgin Islands, Viet Nam, Vanuatu, Yemen, South Africa, Zambia, Zimbabwe. TABLE A4 | Summary statistics of variables included in the estimation. Variable Mean SD Min Max NUnit Import valueopt 2,805,434 3,371,736 0360,572,139 669,864 CHF Export valuedpt 298,200 5,135,876 0831,598,983 490,637 CHF Import quantityopt 76,122 619,352 030,022,336 669,864 Kg Export quantitydpt 91,542 6,132,433 01,578,214,294 490,637 Kg Import priceopt 37 685 0207,386 235,830 CHF/kg Export pricedpt 100 1566 0419,885 136,268 CHF/kg GDPot 1,095,511 2,808,794 223 25,439,700 656,877 Million USD NTMopt 12 12 052 669,864 Tariffopt 523 1821 022,430 669,864 CHF/kg FTAot 0.532 0.499 0 1 669,864 Third Country FTAot 22.848 20.267 066 666,881 1352 Kyklos, 2025 TABLE A6 | The effect of FTAs on different margins of Swiss agricultural exports across basic and processed product types. Outcome variable Intensive margin Extensive margin Import values Import volume Import prices Import probability Import market exit (1) (2) (3) (4) (5) FTAot 0.083 0.246 0.047 0.006 −0.005 (0.143) (0.158) (0.033) (0.006) (0.006) FTAot × Processedp 0.166 0.118 −0.023 −0.007 0.006 (0.153) (0.165) (0.035) (0.007) (0.007) Third Country FTAd≠CHE ot −0.010*** −0.008*** −0.002*** 0.001*** −0.000*** (0.002) (0.003) (0.001) (0.000) (0.000) logGDPot 0.695*** 0.208** 0.024 0.043*** −0.045*** (0.062) (0.097) (0.015) (0.003) (0.004) NTMopt 0.002 −0.005 0.001* 0.002*** −0.002*** (0.002) (0.003) (0.001) (0.000) (0.000) log( 1+Tariff opt) −0.000 −0.000 −0.000 0.000*** −0.000** (0.000) (0.000) (0.000) (0.000) (0.000) Product–time FE Yes Yes Yes Yes Yes Origin–product FE Yes Yes Yes Yes Yes Observations 362,303 362,303 115,616 362,303 306,582 Estimator PPML PPML OLS LPM LPM Note: ***, ** and * denote significance at 1%, 5% and 10%, respectively. Intercepts are included but are not reported. Cluster–robust standard errors are in parentheses. TABLE A5 | The effect of FTAs on the extensive margins of Swiss agricultural exports: alternative estimators. Outcome variable Import probability Import market exit (1) (2) (3) (4) FTAot 0.086*** 0.160*** −0.048*** −0.093*** (0.006) (0.011) (0.006) (0.011) Third Country FTAot 0.005*** 0.009*** −0.006*** −0.010*** (0.000) (0.000) (0.000) (0.000) logGDPot 0.154*** 0.278*** −0.080*** −0.147*** (0.001) (0.003) (0.001) (0.003) NTMopt −0.006*** −0.011*** 0.007*** 0.013*** (0.000) (0.000) (0.000) (0.000) log(1 + Tariffopt) 0.000 0.000 0.000 0.000 (0.000) (0.000) (0.000) (0.000) Product–time FE Yes Yes Yes Yes Origin–product FE Yes Yes Yes Yes Observations 587,108 587,108 484,345 484,345 Estimator Probit Logit Probit Logit Note: ***, ** and * denote significance at 1%, 5% and 10%, respectively. Intercepts are included but are not reported. Cluster–robust standard errors are in parentheses. 1353 TABLE A7 | Estimates for specific agreements (complete table of results). Import values Import volume Import prices Import probability Import market exit (1) (2) (3) (4) (5) FTAot −0.410*** −0.065 −0.062* −0.073*** 0.091*** (0.088) (0.104) (0.035) (0.011) (0.015) FTAot × Albania 1.090*** 1.464*** 0.224* 0.088*** −0.097*** (0.000) (0.337) (0.041) (0.000) (0.000) FTAot × U.A.E. −0.365 −0.331 −0.148* 0.092*** −0.088*** (0.315) (0.227) (0.085) (0.016) (0.019) FTAot × Bulgaria 0.457** 0.091 0.389*** 0.088*** −0.071*** (0.190) (0.333) (0.091) (0.016) (0.020) FTAot × Bahrain 1.965*** 1.787*** −0.313 0.023 −0.013 (0.639) (0.658) (0.392) (0.027) (0.029) FTAot × Bosnia 1.906*** 1.546*** 0.048 0.032* −0.039* (0.175) (0.207) (0.059) (0.017) (0.021) FTAot × Botswana 2.679** 1.419 −0.623 0.023 0.001 (1.335) (1.061) (0.762) (0.047) (0.056) FTAot × Canada 0.343** −0.320** −0.000 0.069*** −0.078*** (0.144) (0.157) (0.057) (0.014) (0.018) FTAot × Chile 0.210 −0.166 −0.005 0.109*** (0.240) (0.221) (0.105) (0.022) FTAot × China 0.366*** −0.361*** 0.056 0.194*** −0.267*** (0.105) (0.130) (0.042) (0.013) (0.017) FTAot × Colombia 0.606*** −0.268* 0.066 0.109*** −0.125*** (0.111) (0.157) (0.056) (0.015) (0.018) FTAot × Costa Rica 0.636*** 0.416*** −0.146*** 0.016 −0.019 (0.129) (0.155) (0.056) (0.016) (0.020) FTAot × Ecuador 0.678*** 0.158 0.051 0.059*** −0.060** (0.177) (0.175) (0.064) (0.021) (0.023) FTAot × Egypt 0.207 −0.595** −0.019 0.080*** −0.099*** (0.158) (0.255) (0.065) (0.015) (0.019) FTAot × Georgia 1.193*** 0.840*** 0.266** 0.138*** −0.128*** (0.324) (0.301) (0.121) (0.021) (0.023) FTAot × Guatemala 0.361*** 0.055 −0.099 0.077*** −0.090*** (0.106) (0.173) (0.070) (0.018) (0.022) FTAot × Hong Kong 0.528* 0.387 −0.139* 0.099*** −0.103*** (0.278) (0.245) (0.081) (0.015) (0.019) FTAot × Honduras 0.307*** 0.294* 0.137 0.059*** −0.062*** (0.117) (0.152) (0.098) (0.020) (0.023) FTAot × Croatia 0.676*** 0.176 0.217*** 0.054*** −0.021 (0.161) (0.232) (0.058) (0.015) (0.019) (Continues) 1354 Kyklos, 2025 Import values Import volume Import prices Import probability Import market exit (1) (2) (3) (4) (5) FTAot × Indonesia 0.256* −0.057 0.155** 0.144*** −0.164*** (0.135) (0.209) (0.071) (0.025) (0.028) FTAot × Japan 0.803*** 0.348** −0.034 0.032** −0.051*** (0.146) (0.138) (0.050) (0.014) (0.019) FTAot × Korea 1.209** 1.122** −0.098 0.136*** −0.130*** (0.470) (0.567) (0.076) (0.017) (0.023) FTAot × Kuwait 1.615** 2.595*** −0.477** 0.043* −0.022 (0.784) (0.812) (0.227) (0.022) (0.025) FTAot × Lebanon 0.485** 0.311 0.270*** 0.105*** −0.124*** (0.214) (0.210) (0.069) (0.018) (0.024) FTAot × Lesotho 6.035*** 9.174*** 0.015 0.018 (1.146) (0.732) (0.054) (0.064) FTAot × Montenegro 2.342*** 0.685** −0.122 0.180*** −0.165*** (0.418) (0.345) (0.127) (0.023) (0.026) FTAot × Namibia −0.294 0.550 0.148 0.022 0.005 (0.430) (0.470) (0.128) (0.022) (0.028) FTAot × Oman −1.695** −3.063*** −0.622*** 0.015 −0.001 (0.666) (0.714) (0.236) (0.024) (0.027) FTAot × Panama 1.154*** 0.322 −0.064 0.053*** −0.062*** (0.290) (0.217) (0.104) (0.020) (0.023) FTAot × Peru 1.030*** 0.804*** 0.001 0.199*** −0.221*** (0.147) (0.161) (0.053) (0.014) (0.018) FTAot × Philippines 0.023 −0.829*** 0.132*** 0.051*** −0.069*** (0.186) (0.283) (0.049) (0.016) (0.019) FTAot × Qatar 1.153 0.462 0.289 −0.016 0.027 (1.112) (0.931) (0.601) (0.027) (0.029) FTAot × Romania −0.472 −0.806** 0.102 0.105*** −0.087*** (0.358) (0.355) (0.095) (0.015) (0.019) FTAot × Saudi Arabia 1.277*** 0.132 −0.294*** 0.049*** −0.044** (0.487) (0.838) (0.104) (0.016) (0.019) FTAot × Serbia 1.464*** 1.113*** 0.052 0.160*** −0.122*** (0.159) (0.181) (0.057) (0.015) (0.019) FTAot × Swaziland 2.530*** 1.904*** 0.002 0.075** −0.056 (0.475) (0.634) (0.131) (0.030) (0.035) FTAot × Tunisia 0.990*** 1.042*** −0.146 0.092*** −0.080*** (0.195) (0.256) (0.100) (0.020) (0.027) FTAot × Ukraine 0.355 0.065 0.140* 0.150*** −0.136*** (0.242) (0.250) (0.078) (0.015) (0.019) (Continues) TABLE A7 | (Continued) 1355 TABLE A8 | The effect of Swiss FTAs on different margins of Swiss agricultural imports: Relaxing stringency of fixed effects. Import values Import quantities Import prices Import probability Import market exit (1) (2) (3) (4) (5) FTAot −0.011 0.098*** 0.047*** 0.004*** −0.002 (0.040) (0.030) (0.007) (0.002) (0.002) Third Country FTAot 0.004*** 0.009*** 0.004*** 0.001*** −0.001*** (0.001) (0.001) (0.000) (0.000) (0.000) logGDPot 0.489*** 0.401*** 0.036*** 0.073*** −0.054*** (0.009) (0.007) (0.001) (0.000) (0.000) NTMopt 0.208*** 0.130*** −0.006*** 0.010*** −0.008*** (0.012) (0.010) (0.000) (0.000) (0.000) log(1 + Tariffopt) −0.000*** −0.000*** 0.000 0.000*** −0.000** (0.000) (0.000) (0.000) (0.000) (0.000) Distanceo −0.288*** −0.235*** 0.168*** −0.038*** 0.025*** (0.022) (0.017) (0.004) (0.001) (0.001) Bordero 1.764*** 1.546*** −0.127*** 0.291*** −0.256*** (0.046) (0.043) (0.009) (0.002) (0.004) Languageo −0.590*** −0.241*** 0.078*** 0.028*** −0.019*** (0.042) (0.039) (0.007) (0.002) (0.002) Product–time FE Yes Yes Yes Yes Yes Observations 607,700 607,700 213,422 607,700 488,393 Estimator PPML PPML OLS LPM LPM Note: ***, ** and * denote significance at 1%, 5% and 10%, respectively. Intercepts are included but are not reported. Cluster–robust standard errors are in parentheses. Import values Import volume Import prices Import probability Import market exit (1) (2) (3) (4) (5) Third Country FTAot 0.001 −0.001 −0.002*** 0.001*** −0.002*** (0.002) (0.002) (0.000) (0.000) (0.000) logGDPot 0.514*** 0.437*** 0.144*** 0.025*** −0.004 (0.037) (0.039) (0.010) (0.002) (0.003) NTMopt −0.064*** −0.045*** −0.000 −0.002*** 0.002*** (0.006) (0.008) (0.002) (0.000) (0.000) log(1 + Tariffopt) −0.000 0.000 0.000 −0.000 0.000 (0.000) (0.000) (0.000) (0.000) (0.000) Product–time FE Yes Yes Yes Yes Yes Origin–product FE Yes Yes Yes Yes Yes Observations 587,108 587,108 206,194 587,108 484,345 Estimator PPML PPML OLS LPM LPM Note: ***, ** and * denote significance at 1%, 5% and 10%, respectively. Intercepts are included but are not reported. Cluster–robust standard errors are in parentheses. TABLE A7 | (Continued) 1356 Kyklos, 2025 A.2 | The Economics of Trade Agreements Standard microeconomic theory predicts that trade agreements generate termsoftrade gains for member countries. To illustrate this, we provide a simplified framework for analysing these effects in a small open economy within a partial equilibrium setting (see also Plummer etal.2011). The small country assumption is appropriate in this context, as Switzerland's international market influence is relatively modest, accounting for just 1.67% of global merchandise imports and 2.96% of global imports of commercial services, which together represent 1.9% of total global merchandise and commercial services imports (Zimmermann2023). FigureA1 depicts the domestic market for a specific good in a country preparing to join an FTA. We refer to this country as the ‘home’ country, other signatories to the FTA as partner countries and nonmember countries of the FTA as outsiders. Before the FTA enters into force, the home country imposes a mostfavourednation tariff ( tMFN ) on all imports, irrespective of their origin. We express tariffs in specific terms as a fixed monetary amount per unit of imports. At this stage, the home country collects tariff revenue equivalent to the product of the tariff rate and the volume of imports (i.e., tMFN × [S0 − D0] ). Additionally, we assume that the outsider is the most efficient producer of the good and offers the lowest price among the three. Before the FTA, domestic producers supply S0 units of the good, while domestic consumers demand D0 units. The excess demand, D0 − S0 , is met through imports from the outsider, who supplies the product at the lowest price. In this preFTA scenario, domestic consumers in the home country pay a price of pHome = pOutsider + tMFN per unit of the good, assuming that the product is homogeneous or perfectly substitutable. After signing the FTA, the removal of tariffs on imports from the FTA partner reduces the price of these imports to pPartner , making them cheaper than imports from the outsider. This price reduction leads to increased consumption, with domestic demand rising to D1 . As a direct consequence, imports will increase from D0 − S0 to D1 − S1 , with all imports now sourced from the FTA partner rather than the outsider. The lower domestic price also results in a reduction in local production, with domestic producers supplying only S1 . The trade creation effect of the FTA is represented by two components. First, the reduction in domestic production, S1 − S0 , is replaced by more efficient imports from the partner country. Second, the increase in consumption, D1−D0 , is also satisfied by additional imports. Overall, trade creation is captured by the change in total imports due to the FTA: [ S1 − D1 ] − [ S0 − D0 ]. A.3 | Figures FIGURE A1 | The economic effects of trade agreements on imports in a small open economy. FIGURE A2 | Import sources. [Colour figure can be viewed at wileyonlinelibrary.com]