Trade policy changes fast. The figures in this post are as of early November 2026, so check for updates before you quote any of them in an exam.

Protectionism is the topic where my H2 Economics students have the most material and the least method. Everyone has read about tariffs since 2025. Very few can turn that into AO4 marks, because knowing what happened is not the same as knowing which economic mechanism it shows.

This post sets out the method I teach: build a tariff analysis that reaches the welfare effects, then evaluate protectionism in a way that holds up in context. The US tariffs of 2025 and 2026 supply the evidence, but the method is the point. Trade policy will have changed again by the time you sit the paper. The chains will not.

Two things from the SEAB 9570 syllabus shape everything below.

Evaluation sits in about 60% of the marks in both papers. Questions testing AO1 + AO2 + AO3 + AO4 make up about 60% of the marks in Paper 1 and in Paper 2. AO4 is evaluation.

The tariff diagram is compulsory. The syllabus says diagrammatic analysis of the effects of tariffs is required. It is the only diagram Theme 3.3 names as required.

As in my fiscal policy evaluation guide, I work through chains, not checklists: follow the economic logic to the variable the question asks about, then test the chain where it is weakest. For the general idea of evaluation, see my post on evaluation across IGCSE, IB and H2.

Three syllabus points to settle first

These come from the Additional Information column of Theme 3.3.1.

1. Comparative advantage is required. The numerical version is not.

You need comparative advantage as an opportunity cost argument: a country should specialise in the good it produces at a lower opportunity cost, this differs from absolute advantage, and specialisation and exchange can raise output and consumption possibilities for both parties. The syllabus states that numerical illustration is not required, so you do not need two-country output tables and opportunity cost ratios. Spend that time questioning the claim instead, which is where the marks are.

2. Institutional and legal knowledge is not required.

This matters right now. On 20 February 2026 the US Supreme Court ruled 6 to 3, in Learning Resources, Inc. v. Trump, that the International Emergency Economic Powers Act does not authorise the President to impose tariffs. It is a significant event. It is also not what you should write about. Court decisions and statutory authorities are institutional and legal knowledge, outside the assessment.

What the episode is useful for is economic: a clear illustration of policy uncertainty as a cost of protectionism, and of how quickly a trade regime can change. Use it that way, in one sentence, and move on. Write the economics, not the news.

3. The tariff diagram is required. Trade agreement diagrams are not.

Draw the tariff diagram properly. Do not invent one for a free trade agreement.

Part 1: Building the tariff chain

The diagram, done properly

Most students can draw the tariff diagram. Far fewer use it: a diagram your paragraph never refers to is decoration. The construction, assuming a small importing country that takes the world price as given:

  • Domestic supply (S) and domestic demand (D) curves for the good.
  • World price Pw, drawn as a horizontal line below the domestic equilibrium price.
  • At Pw, domestic producers supply Q1, domestic consumers demand Q4, and imports fill the gap from Q1 to Q4.
  • Impose a tariff of t. The domestic price becomes Pw + t.
  • At Pw + t, domestic supply rises to Q2, domestic demand falls to Q3, and imports shrink to the gap from Q2 to Q3.
The effects of a tariff on a small importing countryPriceQuantity0PwPw + tDomestic SDomestic DQ1Q2Q3Q4ABCDA: gain in producer surplusB: production distortionC: tariff revenueD: consumption distortionB + D = deadweight loss
Figure 1. A tariff of t raises the domestic price from Pw to Pw + t. Consumer surplus falls by A + B + C + D. Domestic producers gain A and the government gains C in tariff revenue, so the net welfare loss to the importing country is B + D.

The welfare effects read straight off the areas, and this is the part to get exactly right:

EffectAreaDirection
Consumer surplusA + B + C + DFalls
Producer surplusARises
Government revenueC, which is t multiplied by the imports from Q2 to Q3Rises
Production distortionB, the triangle between Q1 and Q2Deadweight loss
Consumption distortionD, the triangle between Q3 and Q4Deadweight loss

The loss of consumer surplus is larger than the producer surplus gain plus the government revenue. The difference is the two deadweight loss triangles, B and D, and that is the net welfare loss to the importing country.

Why the production distortion triangle matters more than students think

Triangle B represents resources drawn into producing the good at home at a higher marginal cost than the world price. That is comparative advantage again, in welfare terms: the country now makes something it is relatively inefficient at, using resources that had a more valuable alternative use. The cost of protection is not only the higher price consumers pay. It is the output given up elsewhere because factors of production were pulled into a protected sector. Saying that, rather than just naming the triangle, separates an analysed diagram from a described one.

The assumption to notice, and to test

The small-country diagram assumes the world price Pw does not change, so the whole tariff is added on top and the full burden falls on domestic buyers. For once there is good evidence on whether that holds.

Economists at the Federal Reserve Bank of New York found that from January to August 2025, about 94% of the tariff burden fell on US importers, easing to about 86% by November 2025. Independent work reached similar conclusions: research by Gita Gopinath and Brent Neiman found pass-through close to complete, and the Kiel Institute's January 2026 study, "America's Own Goal", using shipment-level data, reached the same verdict.

Two lessons follow, and both are examinable.

First, the standard diagram is broadly right. Near-complete pass-through is what the small-country model predicts, and it is roughly what happened, even though the United States is a very large importer.

Second, the drift from 94% to 86% is the large-country case appearing at the margin. Where an importer is large enough that its reduced demand pushes foreign exporters to cut prices to keep market share, part of the burden shifts abroad and its terms of trade improve. Stating that with a number attached is far stronger than saying "some of the burden may fall on foreign producers". To show it on paper, draw a version of the diagram in which Pw itself falls after the tariff, so the domestic price rises by less than t.

From microeconomic welfare to macroeconomic effects

Tariff questions often sit in Section B, so the chain has to continue past the welfare triangles.

On the price level. Tariffs raise the price of imports and of import-competing goods. Where imports are intermediate inputs, they raise firms' costs directly. This is cost-push inflation from a supply-side shock, not demand-pull. Research at the Federal Reserve Bank of Dallas found tariffs added close to a full percentage point to US inflation.

On aggregate demand and output. The effect is ambiguous, and saying so with reasons beats picking a side. Fewer imports raise net exports and AD. But higher input costs reduce short-run aggregate supply, higher prices erode real income and consumption, and retaliation cuts exports. Which dominates is an empirical question.

On the balance of trade. This is the intended target, and the evidence is striking. The Tax Foundation's assessment is that the tariffs have not meaningfully altered the trade balance: a policy can change prices and trade patterns a great deal while failing at the objective it was introduced for.

On employment. Protected sectors may keep or add jobs. Downstream sectors that use the protected good as an input lose them. See Chain 3.

Part 2: The seven evaluation chains

The syllabus requires "benefits and costs of protectionism" at 3.3.1d. Both halves. The test from my fiscal policy guide applies: would this sentence survive unchanged in an essay about a different country or a different tariff? If yes, it is not yet evaluation. In a 25-mark essay you develop two or three chains properly, not seven.

Chain 1: Incidence, or who actually pays

The chain: tariff imposed → does the world price fall? → pass-through rate → who bears the burden.

The test: the claim that tariffs are paid by foreign exporters can be checked, and it was. Roughly 86% to 94% of the burden fell on US importers, and through them on firms and consumers. Whether the remainder shifts abroad depends on the importing country's market power and on the price elasticity of foreign supply.

The extension: tariffs are regressive. Lower-income households spend a larger share of income on goods, which is what tariffs tax. The Tax Foundation estimated the 2025 tariffs at roughly $1,000 per US household. Linking that to the income distribution material in Theme 3.2 is the kind of cross-topic link that lifts an answer.

Chain 2: Net welfare, or do the gains cover the losses?

The chain: consumer surplus loss → producer surplus gain plus government revenue → deadweight loss triangles → net effect.

The test: this is the diagram's own conclusion, so the evaluative work is explaining why a government might impose a tariff anyway. One answer is distributional. The producer gain is concentrated in an identifiable industry; the consumer loss is spread thinly across millions of households. A policy can reduce welfare in aggregate and still be politically rational. That is a standard public choice point, not a political one.

Chain 3: Downstream effects, the strongest and least-used argument

The chain: tariff on an input → downstream firms' costs rise → downstream output and employment fall → net effect across the whole economy.

This is the most under-used argument in the topic, and it has hard numbers behind it. A United States International Trade Commission study of the Section 232 steel and aluminium tariffs estimated that they raised production in the protected industries by about $2.8 billion a year in total ($1.5 billion in steel and $1.3 billion in aluminium), while production in downstream industries facing higher input costs fell by about $3.4 billion a year.

The protection worked for the protected sector. The economy still finished worse off, because more industries use steel than make it. Use it as a question: how many workers are employed in industries that make this good, compared with industries that buy it?

Singapore in 2026 shows the same mechanism from a different cause. After the disruption to energy supplies through the Strait of Hormuz, MTI's first-quarter 2026 survey reported that refineries and petrochemical crackers had cut operating rates and several downstream chemical firms had declared force majeure. Input cost shocks travel downstream, whatever their cause.

Chain 4: Retaliation, because the effect is not one-sided

The chain: tariff imposed → trading partner retaliates → export industries contract → domestic employment and output effects reverse.

The test: retaliation turns a policy aimed at imports into a policy that also damages exports. The US Department of Agriculture estimated that retaliatory tariffs cut US agricultural exports by more than $27 billion from mid-2018 to the end of 2019.

The qualification that makes this evaluative rather than descriptive: retaliation is not symmetric. It depends on the retaliating country's market power, its dependence on the imposing country's market, and its alternatives. China's response in 2025 combined tariff retaliation with other measures such as export controls, which is a different response from matching tariff for tariff. A small economy with little market power has very little capacity to retaliate at all, which is central to Chain 7.

Chain 5: Effectiveness, or did it achieve the stated aim?

The chain: stated objective → mechanism → observed outcome.

This asks whether the policy worked on its own terms, not whether it is theoretically justified. One stated objective was to reduce the trade deficit and bring manufacturing back. The Tax Foundation's assessment is that the tariffs have not meaningfully altered the trade balance. The reasons are worth stating:

  • Trade diversion rather than reduction. Production and shipping can reroute through third countries. Bilateral deficits can fall while the overall deficit does not.
  • The trade balance reflects saving and investment. The current account balance reflects the gap between national saving and investment. A tariff changes the composition and pricing of trade without addressing that gap. Made correctly, this is a real differentiator.
  • Effective rates differ from headline rates. Gopinath and Neiman found that the tariff rates actually paid in 2025 were well below the statutory rates, because of exemptions, shipping lags and gaps in enforcement.

That last point generalises: announced policy and implemented policy are different things, and evaluating the announcement rather than the outcome is a common error.

Chain 6: Uncertainty, a cost separate from the tariff itself

The chain: policy volatility → firms cannot forecast landed costs → investment and hiring deferred → output loss separate from the tariff rate.

This is the distinctive feature of the 2025 and 2026 episode, and it is not in older textbooks. Consider the sequence facing a firm trading with the United States. US tariffs on Chinese goods escalated to 145% in April 2025, were cut sharply under a truce in May 2025, and the truce was extended in August and again in November. The IEEPA tariffs were then struck down by the Supreme Court in February 2026 and immediately replaced by a temporary 10% global tariff for 150 days, which was in turn replaced on 24 July 2026 by new Section 301 tariffs on imports from 60 economies.

The point is not the legal detail. It is that a firm making a ten-year investment decision cannot price any of this. Uncertainty depresses investment regardless of the tariff level, because the value of waiting rises when possible outcomes are wide apart. A tariff that is announced, removed and reinstated may do more damage than a stable tariff at the same average rate.

Chain 7: The small open economy, and why Singapore's position is different

The chain: small economy → no market power → cannot retaliate effectively → cannot influence the world price → exposure without instruments.

This is the material most likely to be directly relevant in a Singapore paper.

Singapore faced the 10% baseline tariff despite a free trade agreement with the United States in force since 2004, and despite US goods data showing a US surplus with Singapore. According to US Census Bureau figures cited by the Congressional Research Service and MTI, the US goods trade surplus with Singapore was $3.6 billion in 2025, up from $1.9 billion in 2024. In March 2026 Singapore was also among the economies named in new Section 301 investigations into structural excess capacity.

That is analytically useful: a bilateral deficit cannot explain a tariff when the official goods data run the other way.

The exposure is real and measurable. In the third quarter of 2025, Singapore's non-oil domestic exports slipped 3.3% overall while shipments to the US fell 30.7%. Section 232 sectoral tariffs on automobiles at 25% and on steel, aluminium and copper at 50% also applied to Singapore.

The policy response is instructive, because it was not retaliation, which would raise domestic prices while being too small to change US behaviour. The response ran through domestic adjustment and support, including the Singapore Economic Resilience Taskforce, the Business Adaptation Grant and the GRIT graduate traineeship programme. In his 2025 National Day Rally, PM Lawrence Wong warned that more trade barriers were coming and that small, open economies would feel the pressure.

This chain links directly to the syllabus point at 3.3.1c on the effects of trade on governments, in terms of macroeconomic objectives and policy choice. Singapore's constrained policy choice is the point.

Part 3: When is protection defensible?

An essay that only demolishes protectionism is incomplete. Engage seriously with the case for it before judging.

Infant industry. A new industry may have a latent comparative advantage it cannot yet realise, because incumbents already enjoy economies of scale and learning. Temporary protection may let it reach efficient scale. The evaluation: the government must pick an industry that will become competitive, and the protection must really be temporary. Both are demanding, and protection tends to persist because the protected group organises to defend it.

Strategic and national security. A country may reasonably decline to depend entirely on foreign supply for defence goods, essential medicines or advanced semiconductors. This is an argument about resilience rather than efficiency. The question is not whether it costs something, since it does, but whether the insurance is worth the premium. The 2026 Hormuz disruption is a reminder that concentrated supply chains carry real tail risk.

Anti-dumping. If a foreign producer sells below cost to drive out rivals and raise prices later, a tariff may correct a distortion. The evaluation: predatory pricing is hard to tell apart from a genuine cost advantage, and anti-dumping measures can shelter firms from ordinary competition.

Structural adjustment. Where import competition displaces workers faster than they can retrain, temporary protection may reduce structural unemployment. The evaluation: that compares protection with doing nothing. The stronger comparison is with retraining and adjustment support, which address the transition without the deadweight loss. Singapore's use of SkillsFuture and the Business Adaptation Grant is an example of the second route.

Negotiating leverage. Tariffs may be used to extract concessions rather than to be kept. The evaluation: this depends on credibility and on the other side's alternatives, and it carries the uncertainty costs of Chain 6 either way.

Part 4: Building the judgement

AO4 asks candidates to "synthesise economic arguments to arrive at well-reasoned judgements and decisions." The same three requirements apply as in any evaluation: a stated criterion, a ranking with a reason, and a condition under which the judgement would change.

A weak conclusion:

In conclusion, protectionism has both benefits and costs. It can protect domestic jobs and infant industries, but it also raises prices for consumers and may cause retaliation. Therefore free trade is generally better but tariffs may sometimes be justified.

A strong conclusion:

Judged by aggregate welfare in the imposing country rather than by the interests of the protected sector, broad tariffs on industrial inputs are difficult to defend. The decisive evidence is not the deadweight loss triangles, which are theoretical, but the downstream effect: the United States International Trade Commission found that a production gain of about $2.8 billion a year in the industries protected by steel and aluminium tariffs came with a production loss of about $3.4 billion a year in industries that use those metals as inputs. Since more sectors buy a widely used intermediate good than produce it, the employment argument for protection tends to fail on its own terms in exactly the cases where it is most often made. This judgement would not extend to a narrowly defined strategic sector where the objective is supply resilience rather than efficiency, because there the relevant test is whether the insurance premium is proportionate to the tail risk, and the deadweight loss is the premium rather than the failure.

The second version is not much longer. It names its criterion, identifies which evidence is decisive and why, and marks the boundary of its own claim.

Part 5: The evidence bank

Every figure here is as of early November 2026. This is the part of the post most likely to date.

ChainEvidenceFigures
IncidenceFederal Reserve Bank of New YorkAbout 94% of the burden on US importers, January to August 2025, easing to about 86% by November 2025
Incidence (distribution)Tax FoundationAbout $1,000 per US household in 2025; tariffs regressive in incidence
Price levelFederal Reserve Bank of Dallas researchClose to 1 percentage point added to US inflation
DownstreamUSITC study, steel and aluminiumAbout $2.8bn a year more output in protected industries; about $3.4bn a year less in downstream industries
RetaliationUS Department of AgricultureMore than $27bn in lost agricultural exports, mid-2018 to end-2019
EffectivenessTax FoundationTrade balance not meaningfully altered
Effective vs headlineGopinath and NeimanRates actually paid well below statutory rates
UncertaintySequence, 2025 to 2026145% on Chinese goods (April 2025), truce (May 2025), extensions (August, November), IEEPA tariffs struck down (20 February 2026), 150-day 10% replacement, Section 301 tariffs on 60 economies (24 July 2026)
Singapore exposureEnterprise Singapore; Section 232Q3 2025 shipments to US down 30.7%; Section 232 at 25% on autos, 50% on steel, aluminium and copper
Singapore anomalyUS Census data via CRS and MTIUS goods trade surplus with Singapore of $3.6bn in 2025 (up from $1.9bn), yet the 10% baseline tariff applied despite an FTA in force since 2004
Singapore forecastingMTI2025 growth forecast: 1% to 3%, cut to 0% to 2% (April), raised to 1.5% to 2.5% (August), raised to about 4% (November); actual outturn 5.0%

On that last row. MTI cut its 2025 forecast to as low as zero in April as the tariffs landed, then raised it twice as front-loading, AI-related demand and de-escalation proved stronger than expected. The outturn, 5.0%, was above the top of every range published during the year. The lesson is not that MTI forecasts badly. It is that the effect of a trade shock on a small open economy is hard to predict, because firms, trading partners and governments all adapt at once. If a question asks you to assess likely impacts, that uncertainty is part of an honest answer.

Part 6: Applying the method in each paper

Paper 2 (essays): 75 marks, 60%, 2 hours 30 minutes

Six questions, answer three: at least one from Section A (mainly microeconomics), at least one from Section B (mainly macroeconomics), and the third from either. Each is worth 25 marks, split into a 10-mark part (a) and a 15-mark part (b), which gives you about 50 minutes per essay including planning.

Protectionism sits across both sections. The welfare analysis is microeconomic; the effects on growth, inflation, employment and the trade balance are macroeconomic. Read which one is being asked before you decide whether the tariff diagram or AD/AS leads. Since about 60% of Paper 2 marks sit in questions that include AO4, an essay that explains all five areas at length and then adds two sentences of evaluation has misallocated its time.

Paper 1 (case studies): 60 marks, 40%, 2 hours 30 minutes

Two compulsory case studies, 30 marks each, each worth 20% of the total. About 12 marks per case study go to data response questions and about 18 to higher-order questions. Each case study presents two to three pages of data followed by six or seven part-questions.

Trade produces exactly the data the paper uses: trade balances, tariff schedules, exports by destination and price indices. AO2 asks you to "make valid inferences based on the information presented and its limitations." On trade data, the limitations are unusually rich:

  • Bilateral versus overall balances. A fall in a bilateral deficit may reflect rerouting through a third country rather than any change in the overall position.
  • Value versus volume. Export values rise with prices. A rise in export value during a period of inflation may hide a fall in volume.
  • Front-loading. Firms brought shipments forward ahead of announced tariffs in 2025, inflating figures in one period and depressing the next.
  • Headline versus effective rates. Announced tariff rates and the rates actually paid can differ a lot.
  • Whose data. In March 2026 Singapore's trade ministry publicly disputed US figures on the bilateral trade relationship. Two governments produced different numbers for the same trade flows.

One sentence noting one of these earns AO2 credit directly and takes almost no time. If you are earlier in the course and this all feels like a lot, my post on starting H2 Economics from zero is the place to begin.

I teach H2 Economics at Higher Nucleus Learning Studio. Class details are on the Higher Nucleus H2 Economics page.

Sources

  • Singapore Examinations and Assessment Board, 9570 Economics GCE Advanced Level H2 Syllabus (2026), Scheme of Assessment and Theme 3.3.1
  • Ministry of Trade and Industry Singapore, Economic Survey of Singapore, First Quarter 2026, and GDP forecast releases, 2025 to 2026
  • Enterprise Singapore, Review of 3Q 2025 Trade Performance
  • Federal Reserve Bank of New York, Amiti, Flanagan, Heise and Weinstein, tariff incidence analysis, February 2026
  • Gopinath and Neiman, The Incidence of Tariffs: Rates and Reality, NBER Working Paper 34620
  • Kiel Institute for the World Economy, America's Own Goal: Who Pays the Tariffs?, January 2026
  • Federal Reserve Bank of Dallas, research on tariff pass-through to inflation, 2026
  • United States International Trade Commission, Economic Impact of Section 232 and 301 Tariffs on U.S. Industries (2023)
  • Tax Foundation, Trump Tariffs: Tracking the Economic Impact of the Trump Trade War
  • United States Department of Agriculture, Economic Research Service, The Economic Impacts of Retaliatory Tariffs on U.S. Agriculture (2022)
  • Congressional Research Service, U.S.-Singapore Relations (2026)

Questions people ask

How do I evaluate protectionism in H2 Economics?

Draw the tariff diagram, read the welfare effects off it, then test the chain: incidence, net welfare, downstream effects, retaliation, effectiveness, uncertainty and the small open economy. Develop two or three properly, take the case for protection seriously, and end with a judgement that states its criterion.

Do I need to memorise comparative advantage tables?

No. The syllabus requires comparative advantage in terms of opportunity cost and states that numerical illustration is not required. Learn the logic and how to evaluate it instead.

Should I write about the Supreme Court tariff ruling in an essay?

Only as economic evidence, in one sentence, as an example of policy uncertainty raising the cost of protection. Institutional knowledge and legislation are outside the syllabus, so legal detail earns nothing.

Who actually pays a tariff?

Mostly the importing country's firms and consumers. The New York Fed found about 94% of the burden fell on US importers from January to August 2025, easing to about 86% by November. Some shifts abroad when the importer is large enough that exporters cut prices, which is the large-country case.

Why can't Singapore retaliate against US tariffs?

Because retaliation needs market power. Tariffs on US goods would raise prices at home while being far too small to change US policy, so Singapore's response ran through domestic adjustment and support instead.

Is free trade always better than protectionism?

The aggregate welfare analysis favours free trade, but the syllabus requires both benefits and costs. Evaluate infant industry, strategic resilience, anti-dumping, structural adjustment and negotiating leverage on their own terms, then judge which considerations dominate in the context the question gives.

This article is also in Chinese: 阅读中文版.