The Free Trade Agreement: How It Functions and What It Does

A free trade agreement (hereinafter: FTA) is a treaty between two or more parties that reduces or removes barriers to commerce between them while leaving each party free to set its own trade policy towards everyone else. Several hundred such agreements are now in force, and the great majority of world commerce takes place between parties that have concluded at least one. They have become the principal instrument through which trade barriers are lowered.

Despite the name, most of these treaties are neither about free trade in any complete sense nor confined to trade. They eliminate tariffs on a defined and often incomplete list of goods, according to schedules that may run for decades, and modern examples devote most of their text to matters other than tariffs: services, investment, procurement, standards, intellectual property and the machinery for settling disputes. This article explains the mechanics and then examines what the evidence shows about their economic effects.

A Brief Historical Note

The legal foundation was laid in 1947, when the General Agreement on Tariffs and Trade (hereinafter: GATT) established the principle that a concession granted to one member must be extended to all. Article XXIV of that agreement carved out an exception permitting customs unions and free trade areas, on the condition that they substantially cover all trade between the parties and that they not raise barriers against outsiders. A further exception, the Enabling Clause of 1979, permitted preferential arrangements among developing economies on easier terms.

For four decades, the exception was used sparingly. From the early 1990s, the number of agreements rose steeply, and the pace has not slackened since. Explanations for the acceleration vary, but the difficulty of concluding negotiations at the global level, particularly after the Doha round began in 2001, is generally counted among them.

How Free Trade Agreements Function

The core of any such treaty is a tariff schedule. Duties are not usually removed at once: they are staged over a period, with sensitive lines given the longest phase-ins and some, most often in agriculture, excluded outright or confined to limited quotas. Agriculture is the most common exclusion across agreements from every region, which is one reason the coverage requirement of GATT Article XXIV has been applied loosely in practice.

The second element follows from the definition. Because each party keeps its own external tariff, goods could otherwise enter through whichever member charges least and be forwarded duty-free to the others, a problem known as trade deflection. To prevent this, every FTA needs rules of origin, which determine whether a good counts as originating within the area. These rules take several forms: a requirement that a minimum share of value be added locally, a requirement that the good change tariff classification during processing or a requirement that a specified process be performed. The economist Anne Krueger showed that this requirement marks the fundamental difference from a customs union, which adopts a common external tariff and therefore needs no such rules; the Southern Common Market, known as Mercosur, established among South American states in 1991, is a customs union in design for this reason, whereas the European Free Trade Association, founded in 1960, is not.

Rules of origin are not a technicality. They impose documentation and accounting costs, and where those costs approach the tariff saving, firms simply pay the duty instead. Utilisation rates, the share of eligible trade that actually claims preference, are accordingly well below complete in most agreements, and lowest for small firms and for goods whose ordinary tariff is already low. Rules of origin can also be written restrictively enough to function as protection in their own right, by obliging producers to source inputs within the area rather than from the cheapest supplier.

Beyond tariffs, contemporary agreements contain provisions of a different character. Services are liberalised under commitments modelled on the General Agreement on Trade in Services. Chapters on investment set out protections for foreign investors. Chapters on technical standards, and on sanitary and phytosanitary measures, address the regulatory requirements that now obstruct trade more than duties do, sometimes through mutual recognition of conformity assessment. Others cover government procurement, competition policy, intellectual property, electronic commerce, labour and environmental standards. Agreements are commonly classified as shallow or deep according to how far they extend into these areas.

On the side of the enforcement of FTAs, there are variations. Most agreements provide for consultation followed by arbitration between the parties, with authorised retaliation as the ultimate remedy. Many, though a declining number, also provide for investor-state dispute settlement, allowing a foreign investor to bring a claim directly against a host state before an arbitral tribunal. This mechanism has attracted sustained criticism and has been narrowed or omitted in several recent treaties. The reason for this is that economically and/or more powerful states can muscle smaller trade partners into unfair conditions resulting from dispute settlement. For example, investors from the United States of America (hereinafter: USA) used such provisions to exploit advantages through settling made-up cases in North American courts, which ruled in favour of the North American corporations. Enforcement was then delegated to the state, which exerted diplomatic pressure on, especially South American, states.

The Theory of Economic Impact

The analytical starting point remains the distinction drawn by Jacob Viner in 1950 between trade creation and trade diversion. Trade creation occurs when the removal of a tariff shifts purchases from a higher-cost domestic producer to a lower-cost partner, which raises welfare. Trade diversion occurs when it shifts purchases from a lower-cost outside supplier to a higher-cost partner that now enjoys preferential access, which lowers welfare and also transfers tariff revenue away from the importing state. The decisive insight is that a preferential agreement is not unambiguously beneficial even in the standard framework: its effect depends on which of these predominates, and that depends on the height of the initial tariffs, the size of the partners and how similar their production structures are.

Later analysis has added several channels. Larger markets allow producers to exploit economies of scale. Exposure to competition disciplines domestic firms and can raise productivity. Consumers gain access to a greater variety of goods. Agreements may attract investment by enlarging the market that an investment serves, and by making policy more predictable, since commitments in a treaty are harder to reverse than domestic legislation. Deep provisions matter particularly for production networks that cross several borders, because a component may pass through numerous customs procedures before final assembly.

The distributional consequences are equally part of the theory. Standard trade models predict that liberalisation benefits the factors of production a state holds in abundance and harms the scarce ones, so that aggregate gains coexist with concentrated losses. Where displaced workers cannot move readily to other industries or places, those losses persist rather than dissipate, and the adjustment costs fall on identifiable communities while the gains are spread thinly across consumers.

What the Evidence Shows

The empirical workhorse is the gravity model, which explains bilateral trade by the economic size of the parties and the frictions between them, and which permits the effect of an agreement to be estimated against what would otherwise have been expected.

Estimating that effect is harder than it appears, because parties choose to conclude agreements and tend to do so where trade was already growing. Once this self-selection is accounted for, the measured effects are substantial: work by Scott Baier and Jeffrey Bergstrand found that an FTA raises bilateral trade between its members by roughly one hundred per cent over about a decade, an effect considerably larger than earlier estimates that ignored the problem.

On the balance between creation and diversion, the weight of evidence indicates that creation generally predominates, and that diversion, while real and demonstrable in particular sectors, is usually the smaller effect. This partly reflects the low level of most tariffs before agreements are signed, which limits the scope for diverting trade.

Effects on income and productivity are more modest and harder to identify than effects on trade flows, and estimates vary widely with method. Deep agreements are associated with larger increases in the trade in intermediate goods, consistent with the argument that regulatory provisions matter most for production networks. Effects on investment are positive in most studies, but sensitive to specification.

The consequences for non-members have been examined closely, and the results cut both ways. Some studies find that members subsequently lowered their tariffs on outsiders as well, so that preferential liberalisation acted as a building block towards broader opening. Others, including work by Nuno Limão, find the opposite in specific cases: where preferences were used to obtain concessions on other matters, the party granting them became less willing to reduce its external tariffs, since doing so would erode the value of what it had traded away.

Examples Across Regions

The variety of these arrangements is considerable, and generalisation from any single one is unwise.

The African Continental Free Trade Area, whose agreement entered into force on 30 May 2019 with trading formally commencing on 1 January 2021, covers almost the whole of the African Union and aims at a continental market; intra-African trade has historically been a small share of the continent’s total, and raising it is among the principal objectives. In the Asia-Pacific, the Regional Comprehensive Economic Partnership, in force since 1 January 2022, links fifteen economies, including Japan, China, the Republic of Korea, Australia, New Zealand and the members of the Association of Southeast Asian Nations (hereinafter. ASEAN); it is comparatively shallow in its regulatory provisions but harmonises rules of origin across a large production network. The Comprehensive and Progressive Agreement for Trans-Pacific Partnership, in force since 2018, is, by contrast, a deep agreement spanning parties in Asia, Oceania and the Americas.

In North America, the agreement among the USA, Mexico and Canada, in force since 1 July 2020, replaced the earlier North American Free Trade Agreement of 1994 and tightened rules of origin in the automotive sector while adding labour provisions. In South Asia, the South Asian Free Trade Area, operative since 2006, illustrates how extensive lists of excluded products can limit an agreement’s practical reach. Among European states, the European Union (hereinafter: EU) is not an FTA but a customs union within a single market, though it has concluded FTAs with partners including Japan, Canada and the Republic of Korea. It constitutes several levels above a classical FTA, as harmonisation of standards and even fiscal policy directives govern economic conduct within the EU. As a result, the EU was the leading economic entity by gross domestic product for a short time in the late 2000s.

Continuing Debates

Three lines of criticism recur in the literature. Jagdish Bhagwati argued that a proliferation of overlapping agreements, each with its own tariff schedule and origin rules, produces a tangle he termed the spaghetti bowl, in which the applicable treatment of a good depends on where it was made and which treaty is invoked, raising costs and favouring firms able to navigate the complexity.

Richard Baldwin offered a partly contrary account, arguing that agreements spread by a domino effect, because exclusion from one becomes costly and induces outsiders to seek membership or their own arrangements, and that regional agreements might eventually be consolidated into broader ones.

A third line questions what these treaties are principally for. Dani Rodrik has argued that as tariffs have fallen, the content of agreements has shifted towards regulatory harmonisation, investor protections and intellectual property rules and that provisions of this kind do not carry the same presumption of mutual benefit that classical tariff reduction does, since some may serve the interests of particular industries rather than raise aggregate welfare. On this view, the case for any given agreement has to be made on its specific contents rather than inferred from the general case for open trade.

Conclusion

Free trade agreements function through staged, incomplete tariff schedules, rules of origin necessitated by the absence of a common external tariff and, increasingly, provisions that extend well beyond the border into domestic regulation. Their measured effect on trade between members is large and reasonably well established; their effects on income, productivity and investment are positive in most estimates but smaller and less certain; and their consequences for non-members and for the distribution of gains within member societies remain genuinely contested. The general case for open trade does not settle the question of whether any particular agreement is beneficial, because that depends on what the treaty actually contains.

From a devletist perspective, FTAs are viewed very positively, as they reduce transaction costs and enable goods and products to move more freely around the world. This has normative implications that facilitate and accelerate genuine knowledge production. As tariffs are lifted, goods and services that arise from talented people’s production in one state where demand for that specific good or service is weak can find suitable markets elsewhere through FTAs without additional entry barriers to the market. That way, the probability of people being able to engage in genuine knowledge production in a way that also funds their living increases, which has amplified development effects on a society.

What is a necessary precondition for FTAs, however, is that the frameworks must always be balanced and fair. Of course, states need to agree to an FTA in the end, but political and economic pressures are often misused by stronger states to exploit weaker states, which hampers long-term growth and development. As FTAs are concluded in an international setting, where supranational oversight is not given, these structural imbalances will persist for as long as there is no global shift to the devletist mode of governance; not because it intriduces a supra-national organ, but because states themselves are normatively organised, allowing fair and mature economic diplomacy.