This week Thailand is hosting the IMF and World Bank annual meetings, and the venue’s chef is ageing tilapia for five days with makrut lime, lemongrass, cinnamon and cloves. About 70–80% of the roughly 2,000 ingredients served will be Thai. The exercise costs little and is aimed at people who allocate capital. Sceptics reasonably ask what it buys. Their usual answer is that attraction counts as power only when it secures costly decisions, the kind a government’s interests would otherwise rule out. This article argues that the test is set too high, that the research record meets a more reasonable version of it, and that soft power remains a working instrument of statecraft, with limits worth stating plainly.
A standard no instrument meets
Power is conventionally understood in probabilistic terms. Sanctions, tariffs and military deployments shift decisions at a cost to the target, and each fails often when the target’s stakes are high. We still classify them as power. Nye’s claim for attraction was equally modest: it can produce preferred outcomes without coercion or payment. Requiring it to overturn a core interest sets a higher bar than the one applied to coercive tools.
Where researchers have tested soft power on costly decisions, they have found effects. Goldsmith and Horiuchi matched survey data from 58 countries to their governments’ decisions in 2003.
Researchers found that foreign public opinion about US foreign policy had a significant and large effect on troop commitments to the Iraq war, after controlling for hard-power factors.
Few choices cost more than committing troops to a contested war. The authors also found that the effect depends on how salient the issue is to mass publics, so the result comes with a limit attached.
Mechanisms that run through people and programmes
Soft power is easiest to observe where it works through identifiable people and programmes. A study in the British Journal of Political Science examined more than 900 leaders from 147 non-Western countries between 1947 and 2001.
The British Journal of Political Science found that leaders who had attended university in a Western democracy were less likely to initiate militarized interstate disputes, after accounting for leader selection and country-level factors. The claim concerns the costliest decision a government takes.
Causal identification from observational data remains contestable, since leaders who study abroad may differ from those who do not in ways the controls miss.
The wider literature adds a caution. Dreher and Yu’s work on the “alma mater effect” found that foreign-educated leaders were on average less friendly towards their host countries but more friendly towards countries sharing the host’s cultural and political goals, and that the pattern weakens as elections approach. On this evidence, attraction moves leaders towards a bloc, and domestic politics limits how far it carries.
Aid supplies a second case. Goldsmith, Horiuchi and Wood used instrumental variables based on the severity of the HIV/AIDS burden to estimate the effect of the US PEPFAR programme on opinion. They found that it substantially improved perceptions of the United States across more than 80 recipient countries. The result is contested. A later conference paper using Pew data from the Trump years reports that the link between aid and opinion disappears once attitudes towards the president are controlled for.
Both findings point to the same condition: aid that is targeted, sustained, effective and visible can raise a donor’s standing, and perceptions of its leader can easily swamp the gain.
That sits awkwardly with the treatment of soft-power instruments as disposable. Shortly before his death, Nye warned that cancelling USAID assistance and silencing Voice of America deprives a country of a major instrument of power.
The cost of cooperation
A more modest claim holds that attraction changes the price of cooperation. It is easy to wave away, but transatlantic trade suggests it matters. In July 2025 the EU agreed to remove tariffs on US industrial goods in return for a 15% cap on American tariffs. The source is an advocacy outlet, the Rosa-Luxemburg-Stiftung’s Brussels office, though its account of the terms matches other reporting. Washington obtained that agreement while its standing was falling, which is a hard-power outcome and should be recorded as one. The same source quotes Trade Commissioner Maroš Šefčovič framing the concession as a matter of security and Ukraine, which weakens any soft-power reading of the deal.
The terms of ratification are more informative. The European Parliament’s approval included a safeguard allowing the Commission to suspend concessions on steel and aluminium if the US has not cut its tariffs on those products below 15% by the end of 2026. Use of the safeguard rests with the Commission and nothing terminates automatically, so its practical weight is uncertain. It does show that the legislature demanded insurance against a partner it did not trust. This is how reputation operates in practice. A trusted partner is dealt with on easier terms, and a distrusted one is still dealt with, at a price paid in conditions and delay.
The reputational trend behind this is large. Pew’s 2026 survey found China viewed more favourably than the US in 25 of 36 countries, reported as the first time in about 20 years of tracking that China had come out ahead.
Canadian favourability towards the US fell from 57% in 2023 to 33%. In additional polling across several middle-income countries, a median of 75% said the US interferes in other nations’ affairs a great deal or a fair amount, against 45% for China. Whether such figures decide any single negotiation is the wrong question. Advantages and penalties of this kind accumulate across many negotiations.
What governments do about attraction
Governments act as though attraction matters, and sometimes act against it. After Seoul agreed to host the US THAAD missile-defence system, China imposed an unofficial boycott on Korean content. Reporting in 2023 found the restrictions still in force, with Beijing denying they exist and analysts expecting them to last until political ties improved.
A government that spends coercive resources on keeping another country’s songs and dramas off its screens treats them as a force worth managing.
This does not show that culture can dissolve the state that governs those who consume it. It shows that states regard shaping what populations want as consequential.
Circularity is not emptiness
Attraction and material power are hard to separate. Brand Finance’s 2026 index identifies perceptions of a strong and stable economy as the single most influential driver among its 35 attributes. The same report argues that hard power exercised by self-described liberal democracies is likely to damage perceptions. The influence therefore runs both ways: economic strength feeds attraction, attraction eases access, and access supports further economic strength. A loop of this kind shows that the categories belong to one system. It does not show that either is empty, and pure hard power cannot be isolated from the loop either. Brand Finance is a consultancy that sells nation-branding services, and its index measures perceptions and not outcomes, so its framing deserves caution. Its direction of travel for the US matches Pew’s independent data.
Where the contest is fought
States keep competing on this ground, and the evidence on how they win is mixed. Ne Zha 2 became the first film to exceed US$1 billion in a single market, and the East Asia Forum analysis stresses that it succeeded through the market and not through party direction. The same analysis reports that Beijing reportedly spent around US$6 billion on state media abroad, in efforts that had “to a large extent backfired”. South Korea’s government-backed Hallyu study put 2025 exports at US$18.98 billion.
Taken together, these cases support a qualified claim: attraction accumulates where states enable cultural production and do not direct it.
The limits
Thailand’s own figures temper any optimism. Foreign arrivals in the first eight months of 2026 were down 3.08%. South China Morning Post reporting says Chinese arrivals fell by nearly 30% in 2025, and Thai tourism authorities attributed the fall to safety headlines, the border war with Cambodia, a stronger baht and Beijing’s incentives for domestic travel. Cultural appeal addressed none of those. Researchers at ISEAS argue that Thai strategy conflates economic output with cultural influence, which is a fair criticism of what governments count as success. The trade outcome above also shows that a state with declining attraction can still obtain agreements. Attraction supplements leverage and does not replace it.
What the sceptical case has not shown is that attraction contributes nothing at the margin. A better test asks what another actor did differently and at what price. A diner who buys Thai rice, an investor who chooses Bangkok over a near-identical rival, a leader schooled abroad who hesitates before initiating a dispute, and a legislator who drafts one more safeguard are all small effects. Most of international politics is made of small effects, and a middle power with little other leverage over a finance ministry has good reason to try to be remembered well.
Essay: Maseera Shaik- Research Assistant (Security, Crisis and Strategy) at The Advanced Study Institute of Asia (ASIA)
Produced by Decypher Team in New Delhi, India