Should we be scared of China's surpluses?

In the early twentieth century, xenophobic fears of the “Yellow Peril” — Asian immigration to the west — led to the US Immigration Act of 1917, barring immigration from China and other Asian nations.

In the early twenty-first, instead of sending people, the ever-cunning Chinese have discovered an even more sinister thing to send to the shores of California and undermine Western power: cheap cars.

Or, at least, that’s what an increasingly loud and influential contingent would have you believe: “China is making trade impossible”, “Europe has no choice but to intervene on trade”, and “China’s growth target is a global problem” have all been published in the FT in the last five months. They just ran a four-part series on the topic. On the basis of a Goldman Sachs analysis, the Wall Street Journal’s Greg Ip tweeted that China’s “unique status as a ‘beggar thy neighbor’ partner should motivate other nations into forming a trading system without it.”

Economic substacker Noah Smith and Nobel prize winner turned economic substacker Paul Krugman have joined the chorus. Smith writes that “if Europe fails to fight the Second China Shock — if it embraces the comfortable euthanasia of deindustrialization — it will be militarily weaker in the face of the Russian threat, its international financial position will deteriorate, and it may be economically poorer as a result.” Krugman tells us that “one doesn’t have to be a crude, Trumpian-style trade protectionist to understand that it’s a major problem for the global economy when an economy the size of China’s…runs persistent, extremely large trade surpluses.”

Recent reports from the Bank of England and one commissioned by the French G-7 Presidency corroborate these concerns with an added institutional imprimatur.

These doomsayers are responding to what has been dubbed “China Shock 2.0.” China Shock 1.0 was the rise in Chinese exports of low value added goods that occurred in the early 2000s. Shock 2.0 is when China begins exporting the high value added goods that constitute what remains of the industrial bases in the US and Europe. We can see the broad pattern in the time-series of China’s current account balance as a fraction of world GDP:¹

As the chart indicates, balance of payments aficionado Brad Setser thinks the IMF’s data undercounts China’s current account surplus, and its 2025 value was more like 0.75-0.91% of world GDP. Either way, the overall pattern is clear: China shock 1.0 peaked in 2007 at a 0.65% of global GDP surplus, the surplus came all the way down to <0.1% in 2018, and has since risen back up — either to a new all-time high (per Setser) or close to its previous peak.

Given all the panican about shock 2.0, one might assume that leading academic work on shock 1.0 has delivered a negative verdict: that allowing China to run such large surpluses harmed the US and other global economies. Yet, as far as I can tell, that is unambiguously not the case.

Four leading papers that estimate the effect on the US of shock 1.0 — Caliendo et al. 2019, Dix-Carneiro et al. 2023, Rodríguez-Clare et al. 2026, and Kim et al. 2026 — all find that, in aggregate, the US benefitted from China shock 1.0.

How much should we trust the ivory tower on these matters? Michael Pettis recently tweeted a critique of “most academic trade models.” Pettis says we shouldn’t trust models which “assume that we live and trade in the world of Econ 101, i.e. an idealized world of free and balanced trade, in which countries maximize global output by specializing in their areas of comparative advantage, and in which countries do not intervene in their external accounts.”

However, all four of the aforecited papers match the fact that US-China trade is unbalanced, and the recently released Kim et al. 2026 explicitly models Chinese currency manipulation — a form of intervention in the external account. The papers don’t push a blind Panglossian vision of China Shock 1.0 either: they all show that China shock 1.0 contributed to the decline in US manufacturing employment, and some of them find that certain states or communities were net losers from the shock. Nonetheless, the aggregate impact of the shock remains positive.

The reason is straightforward: China shock 1.0 resulted in a wave of new and cheaper goods for people to consume. Contra DJT, children don’t like settling for two dolls instead of 30. Jaravel and Sager find that when a product is 1pp more imported from China, its price declines 1.9%. Across shock 1.0, they estimate that for each job displaced by trade with China, consumer surplus increases by $477,555 — a worthwhile trade-off given that average job earnings were $33,000.

But…but…what about Autor, Dorn, and Hanson? Didn’t ADH deliver the final blow to the myth of free trade? There are two things to say here: 1) ADH has come under serious replication scrutiny, and a series of papers since then make it far from clear that the China shock had as negative an impact on specific communities as ADH claimed 2) There is a crucial difference between our best estimates of the net effects of the China shock and our best estimates of whether some were harmed by the China shock.

I don’t mean to be glib about the real economic pain that specific communities experienced as a result of China Shock 1.0. However, if our best current estimates are that the net effect of the shock on US welfare was positive, I also don’t think we can blindly write articles that suggest rising Chinese surpluses are some sort of existential threat to the rest of the world.

Instead, if we are worried that China Shock 2.0 is going to be a net negative outcome for the US, or Europe, or whomever else, we need to make arguments that specifically address either why existing estimates of China shock 1.0’s effects are wrong, or why China shock 2.0 will be different.

In my mind, there are five broad classes of argument one could make. I’m going to go through them each, in ascending order of how plausible I find them (most plausible last).

1. Beggar thy neighbor

2. Security risks

3. Inter-generational trade-off

4. Financial risks

5. Growth risks

To spoil the results: there is good evidence that the US should have some sort of policy response to China Shock 2.0 beyond laissez-faire, but the evidence that any of these five forces are strong enough to suggest the net impact of China Shock 2.0 will be negative (for the US or Europe) is lacking. Furthermore, the policy responses I believe are appropriate would likely still result in a world where the US is a large current account debtor and China runs a large current account surplus. So while there are elements of China Shock 2.0 to take seriously, surpluses themselves are not something to be scared of.

Beggar thy neighbor

Michael Pettis frequently claims that, by running large surpluses, China is forcing “the demand-suppressing cost of their policies onto their trading partners.” The idea here is relatively straightforward: by disincentivizing consumption within China, China’s policies are reducing domestic demand, which, ceteris paribus, reduces global demand.

The problem with this logic should be fairly obvious: ceteris is not in fact paribus. It assumes other countries passively hold their own demand fixed in response to suppressed Chinese demand. But if that were the case, we should expect to see excess unemployment in the rest of the world in response to rising Chinese surpluses.

The empirical record decisively rejects this prediction: both US and EU unemployment was falling during China Shock 1.0 (2000-08), and post-2021 we’ve seen falling unemployment in the EU and stable full-employment in the US.

Monetary policy in other countries adjusts, preventing any shortfall of demand. The more sophisticated version of this argument recognizes that it only bites if other countries are constrained by the zero lower bound (ZLB). In that case, monetary policy cannot lower interest rates to offset a fall in demand elsewhere in the world.

But reports of the death of domestic demand levers at the ZLB have been greatly exaggerated. The monetary authority can raise the inflation target, implement (NGDP or price) level targeting, use forward guidance, and open up the spigots on QE. Furthermore, fiscal policy remains an adequate tool for boosting demand.

One modulation of the beggar thy neighbor story Pettis has articulated recently is that rather than forcing unemployment on the rest of the world, suppressed demand in China forces either an increase in household or governmental debt in the current account debtor countries. This mechanism has some bite, but the concerns it raises naturally fit in the “inter-generational trade-off” and “financial risks” sections below.

Security risks

In my opinion the most straightforward reason — though not the most compelling reason — to worry about China’s surpluses is the implication for national security relevant supply chains. If there is real risk of conflict with China in the future, increasingly relying on China for inputs of goods critical to national security is not a good idea.

But even if we accept the premise — that there is real risk of conflict with China in the future — the remedies needed to ensure national security are unlikely to change the fact that China runs large surpluses while the US runs large deficits. Ergo, the national security threat vector is not really a reason to be scared of surpluses.

To be concrete: the natural solution to whatever national security threats we face is to directly subsidize US (or allied) production of those goods deemed necessary for sovereignty over military production. Since aggregate trade deficits are driven by macroeconomic saving and investment patterns, subsidizing a few security-relevant industries is overwhelmingly unlikely to change the global imbalance status quo.

One pointed example is rare earths. Per this report, 72% of US rare earth imports from 2019-22 were from China. Concerning! However, in 2025 the US imported a total $165 million of rare earths. If we assume the 72% fraction held up (which it probably did not since the bilateral deficit has shrunk), that is $119 million of rare earth imports from China. Which accounts for a whopping 0.04% of total US imports from China in 2025.

All imports of ships, aviation, and ammunition — three sectors singled out by a recent DoD initiative — regardless of whether they come from allies or adversaries and whether they are for military or commercial use, account for about 1.1% of total US imports.²

Coming it at from a maximally conservative perspective, the DoD requested $170bn worth of procurement in FY 2024. Even if we assume all of those goods rely on some critical inputs from China, $170bn is only 51% of 2024 imports from China (and <5% of all Chinese exports). Given that a very large chunk of those goods value surely does not come from China (unfortunately we don’t have exact stats on this), I think at most we should expect that fully de-risking our military supply chain from China would reduce the bilateral surplus by 25%, and barely reduce China’s overall trade surplus. That’s not nothing, but it would still leave us in a world of large Chinese surpluses and US deficits.

The point is, even if you think addressing supply chain risks is a top national security priority, it’s not a decisive factor in analyzing whether surpluses themselves are scary.

Inter-generational trade-off

Noah Smith puts the inter-generational trade-off point as follows:

A trade deficit is not a gift; it is a loan. That loan must be paid back. The Europeans who take out the loan may be insufficiently long-term in their thinking, effectively borrowing against their own futures to consume in the present. It’s up to government to look out for the prosperity of future generations.

A nice paper by Christopher Phelan that came out a couple months ago formalizes this logic in a simple model: if country A suppresses consumption and runs trade surpluses with country B, that is likely good for country B residents in the short-run, but unambiguously bad for country B residents in the far-off future.

The neatness of this accounting-identity based argument is enticing, but I don’t think it holds up to more careful scrutiny.

The central issue is just how far-off this far-off future is. To riff on Keynes, in the long-run, either we’re all dead, or we’re all likely to be much richer than we are today. The US debt introduces a generational trade-off — someone is likely to have pay for it eventually — but we understand that this trade-off is worthwhile if future generations are going to be better-off than current ones.

So, assuming protectionist hackles aren’t raised more than they already have been, how long should we expect China’s surpluses and US/EU deficits to persist? Mostly, we don’t really know. But aside from a blip in 1991, the US has run current account deficits continuously since 1982, and China has run surpluses since IMF data begins in 1997.

The paper that takes this question most seriously is Auclert et al. “Demographics, Wealth, and Global Imbalances in the Twenty-First Century.” In their projections, the US net foreign asset position (which you can think of as the cumulative current account) continues worsening until after 2050, and then only mildly reverses course for the rest of the century. This pattern is mirrored in China, where the NFA surplus increases until past 2050, and then barely declines afterwards. Demographics are the driving force: the more rapid aging of the Chinese population leads to net asset accumulation relative to the US.

Another perspective comes from Itskhoki and Mukhin’s “Global Imbalances: A Progress Report.” They note that, in principle, it is possible the US never has to run a trade surplus to offset accumulated trade deficits. This is possible if the US persistently maintains “exorbitant privilege”: the foreign assets the US owns abroad earn higher returns than the US assets owned by foreigners.

Given the opposite pattern held from 2007-24, the more plausible case to the authors (and to me) is that a combination of some exorbitant privilege and some improvement in trade balances eventually contributes to improving the NFA position. In their words: “Because the net foreign asset position represents only a small share of aggregate national wealth and can easily be dominated by valuation effects from movements in equity and housing prices, such an adjustment may be extremely gradual and unfold over several decades.”

My takeaway from this literature is that the “loan” that is the US trade deficit likely does not “need” to be paid off for decades at the least, and this trade-off of present consumption for future consumption is benign.

Financial risks

China’s surpluses are often described as “unsustainable.” The underlying logic here is basically the same as the inter-generational trade-off argument, just with a more ominous implied accelerated timeline.

The worry is that persistent imbalances place strains on the financial system: continued export-growth in China alongside currency manipulation creates pressure for the yuan to appreciate; continued trade deficits in the US lead to rising debt and equity liabilities, sowing the seeds of a sharp correction.

This was one of the main narratives surrounding China Shock 1.0: in 2005, Maurice Obstfeld and Ken Rogoff warned us of “the growing U.S. current account deficit and the potentially sharp exchange rate movements any future adjustment toward current account balance might imply.” Nouriel Roubini and Brad Setser wrote in 2004 that “the tensions created by this system [of global imbalances] are large, large enough to crack the system in the next three to four years.”

Yet, instead of a sharp correction, since 2005 we’ve seen the negative US NFA position about 4x in size (as a share of GDP)³ and the real dollar exchange rate appreciate by 14%. With hindsight, we have a better understanding of why demographic factors can lead to persistent and stable global imbalances.

A goal post-shift worth considering is that though the main concern expressed at the time was a sharp dollar depreciation, the great financial crisis (GFC) of 2008 was another manifestation of the financial risks created by global imbalances. The idea being that mid-2000s foreign inflows into the US are what pushed down interest rates, propped up the housing boom, and created the conditions for its collapse.

There is a substantial amount of work supporting this hypothesis; but as with all things macroeconomic, an equally substantial literature either directly disputing the global imbalances story or emphasizing other drivers of the housing boom. Not to mention, the Scott Sumner view — worth taking seriously — which disputes the importance of the housing bust in contributing to the recession altogether.

I’m not going to litigate this debate here — so you’ll have to take my word on it, do the reading yourself, or ask your local chatbot oracle — but in short: I think global imbalance driven foreign inflows were a factor, but far from the dominant factor, in explaining the GFC. Regulatory mistakes, monetary policy mistakes, inadequate housing supply, and irrational expectations all deserve a larger share of the blame.

Which leaves me with the view that the “financial risks” reason to be worried about global imbalances is not to be dismissed, but not a dramatic threat either.

Growth risks

A longstanding reason for abandoning laissez-faire and meddling in trade policy is the idea that there may be increasing returns in the production of traded goods. Think of the growth of China’s solar energy and electric car industries, shipbuilding and other industrial sectors in South Korea, or even the Lancanshire cotton industry, to name a few. There is some sort of compounding advantage that builds from developing the production of highly complex manufactured goods, which leads to greater technological progress and economic growth. Meanwhile, net importing countries shrink their manufacturing sector, foregoing those increasing returns, and slowing subsequent economic growth.

The connection between increasing returns in the tradable sector and global imbalances has been demonstrated sharply in two recent papers: “The Global Financial Resource Curse” and “Industrial Policies, Global Imbalances and Technological Hegemony.”

The basic story goes as follows: when a country disincentivizes consumption, in addition to running a current account surplus, that country will have relatively more activity in its tradable sector — since tradables can be exported to other countries where consumption is not suppressed. If there are increasing returns to scale in tradable goods, the country will experience subsequently faster growth as a result of their concentration in tradables production. The inverse applies to the debtor countries: capital inflows push up their consumption of both tradables and non-tradables, but since they are importing more of the tradables from the surplus country, production re-allocates to non-tradables. Subsequent growth is then lower than it would be without the capital inflows, due to less ability to capitalize on increasing returns in the tradable sector.

China produces ever-cheaper and more capable electric vehicles, the US produces ever-pricier and sloppier salad bowls.

Industrial policies amplify this distortion, by further increasing activity in the tradable sector of the country implementing the policy and further reducing it in the country idly placing bulk orders on Temu.

There is a crucial distinction between static and dynamic effects here: the debtor country importing all their tradables experiences immediate gains — they get to consume stuff more cheaply. This abstracts from job loss in the sectors facing import competition, but as the earlier literature I referenced makes clear, our consensus estimates are that these static gains are positive. However, the static gains could be more than offset by the dynamic losses of slower subsequent growth, due to the re-allocation of production to the non-tradable sector, which lacks increasing returns.

Before preaching protectionism, there are two questions to resolve: 1) on net, do we expect there to be dynamic losses from China Shock 2.0? 2) if so, should we expect those dynamic losses to outweigh the static gains?

The nice little fable we just told about how China shock 2.0 could lead to a shrinking tradable sector and dynamic losses in the US/EU is very much one angle on the issue. It’s the angle most of the press coverage of the topic takes, but it’s worth emphasizing that it is the opposite of the conventional academic wisdom on the topic.

In the conventional — theoretical — account, supported by a variety of different underlying models, greater openness to trade, including import competition, leads to dynamic gains, not losses. Firms get to import better intermediate inputs, heightened competition increases their incentives to innovate, and knowledge of new technologies flows across borders.

The empirical literature has generally supported these mechanisms, albeit with some important qualifications. The 2019 paper “The Impact of Trade Liberalization on Firm Productivity and Innovation” provides a review of the evidence. They summarize the results as follows: “in emerging countries, trade liberalization appears to spur productivity and innovation. In developed countries, export opportunities and access to imported intermediates tend to encourage innovation, but the evidence on import competition is mixed, especially for firms in the United States.”

I’m only aware of a few full-fledged quantitative models which attempt to answer these questions. Two recent “state-of-the-art” papers find pervasive dynamic gains.

Most pertinent to the China Shock 2.0 debate is “Globalization and the Ladder of Development: Pushed to the Top or Held at the Bottom?” The paper doesn’t focus on the China shock specifically, but globalization writ large, from 1964 to 2014. The relevance of the paper comes from its mechanism: different export products can be thought of as sitting on a ladder of ascending complexity. When a country imports from the world, their own production shifts away from products which face the most foreign competition. Either down the ladder of complexity if more complex goods face greater competition, or vise-versa.

Contra to the other literature, the paper does find robust evidence of dynamic losses from globalization. However, those dynamic losses are largest in low-income countries (primarily sub-Saharan Africa), and more importantly yet, the static gains outweighed the dynamic losses for all but 8 of the 120 countries in their sample.⁴ In the US, there were static gains and no dynamic losses (or gains) whatsoever.

It’s only one paper, and its sample overlaps with shock 1.0, when China was competing more in low value-added goods — thus competing more with the low-income countries which experienced larger dynamic losses. So I don’t take it as strong evidence that we shouldn’t be worried about dynamic losses outweighing static gains from shock 2.0. But I think we should be clear-eyed about the fact that even if we think China Shock 2.0 will lead to dynamic losses, we have no good reason to believe those dynamic losses will offset the static gains.

Furthermore, I return again to the question of whether this whole discussion motivates fear of surpluses qua surpluses. Taking returns to scale in the tradables sector seriously might very well imply some combination of targeted export-subsidies and consumption-suppressing policies are a good idea, but the underlying demographic destinies mentioned beforehand lead me to expect China will run large surpluses and the US large deficits even if the optimal “response” is concocted.

Conclusion

There are real things to be concerned about from China Shock 2.0: Leibniz taught us that natura non facit saltus — nature makes no leaps. Two centuries of economic growth have taught us that development makes no Pareto improvements. But it’s all too easy to focus on the negatives and forget about the value that comes from giving people more opportunities to buy things for cheap. Especially in a time when affordability is the political issue du jour. We can take seriously China Shock 2.0 without castigating it as a villain.

1. I used IMF data for China’s current account in $ and world GDP in $.

2. Calculated by summing 2024 imports of aircraft/spacecraft, arms and ammunition, and ships, boats, and other floating structures, divided by total 2024 imports.

3. Based on figure 1 in Atkeson et al.

4. Belize, Bermuda, Central African Republic, Fiji, Gambia, Guyana, Rwanda, and Togo