Brad Setser

What's more, it is a bit of a myth than tariff differentials  and non tariff barriers explain the US trade deficit.

Consider two sectors: pharma and chips 

1/

Tariffs globally in both sectors are zero.  There are no meaningful differences in tariff rates.   And not many non-tariff barriers (FDA approval is needed for the sale of a drug in the US, and many countries piggy back off that ... but no one would say this is unnecessary)

2/

The US of course now runs a massive (0.5 pp of GDP) trade deficit in pharmaceuticals.  But that is because US firms can avoid paying the headline US tax rate on their most profitable drugs by moving the IP and production abroad

4/

And US drug companies typically also make their patent protected drugs for the global market abroad -- so the standard US trade negotiating demand (longer periods of patent protection and the like) would have no impact on goods exports/ the bilateral goods balance

5/

Chips (semiconductors) have some similarities -- though the current global pattern of production is more tied to past industrial policies (and TSMC's successful model of being an independent cutting edge fab) than to tax avoidance 

6/

There are of course tax advantages to exporting chip designs (NVDIA, Qualcomm, Apple, others) rather than actual chips (Intel) but they aren't huge ...  

But there aren't major tariffs or non -tariff barriers (outside of China, which isn't keen on paying for legacy chip IP)

7/

However tariff free trade can still be distorted (or shaped) by government policies -- Chris Miller's Chip Ways documents how Singapore pulled in foreign chip manufactures with tax advances, and how Taiwan and Korea subsidized their own national champions

8/

And the scale of China's current chip subsidies is really mind blowing, as it (for understandable reasons) a national priority and Xi isn't afraid of writing big checks for technological catch up

9/

Point being that bringing tariffs down to zero won't change the global game in either pharma or chips.   Tariffs are already zero.  Changing the pattern of trade requires thinking more broadly ...

10/

For pharma, I obviously think the right approach is to get rid of the pro offshoring provisions of the Tax Cuts and Jobs Act (George Callas has some interesting ideas that might be palatable to Republicans)

11/

And for chips, the US really did need an industrial policy of its own if it wanted to stay at the technological cutting edge (hence the bipartisan CHIPS act).  A stronger taiwan dollar wouldn't hurt either; it would make TSMC more keen to invest abroad

12/

But the narrative that every trade problem (including for that matter limited US auto exports) can be traced to differentials in tariffs that panelize the US isn't really true.

/end

The January US trade data showed signs of front running in consumer goods (and gold of course) but not in autos.

That changed in February.  The US is on track for a blowout q1 trade deficit ...

1/ 

The FT's reporting confirms all suspicions --  and March will match February in all probability as the incentive to front run April 2 is clear

2/

https://www.ft.com/content/d391b660-2aed-44bf-b020-ed2369d46c88

And since it now seems (based on Bloomberg's latest reporting) that April 2 ("Liberation day") will only see the rollout of reciprocal tariffs not the sectoral tariffs (232s) that have been teased (including by Trump) 

3/

https://www.bloomberg.com/news/articles/2025-03-22/trump-plans-his-tariff-liberation-day-with-more-targeted-push?fromMostRead=true

The fact that Trump is more hawkish on tariffs than most of his advisors only adds to the incentive to ship as much as possible to the US now -- as it is hard to imagine that the sectoral tariffs won't eventually take effect in some form

4/ 

As an aside, targeting all countries with trade barriers/ and or a trade surplus with the US will produce some strange results -- "Only countries that don’t have tariffs on the US, and with whom the US has a trade surplus, will not be tariffed"

5/

Korea and Canada don't have significant trade barriers to US industrial exports, as both have FTAs with the US. But Korea runs a large surplus with the US and it is hard to see how Trump doesn't tariff Canada even tho Canada runs a deficit with the US if oil is excluded 

6/

the US generally runs surpluses with both Australia and Brazil, as both use some of the proceeds from the large iron exports to china to buy US goods ...

7/

The US has a free trade agreement with Australia, so it presumably won't be on the "reciprocal" list now being drawn up by Trump's advisors -- but Brazil has large tariffs and presumably will be on the list ...

8/

And if the tariffs are significant, trade will evolve dynamically and countries that don't now have a surplus and avoid tariffs may soon have a surplus (am thinking of say the UK ... which is an obvious workaround for tariffs on the EU )

9/

I understand the desire not to impose too big a shock -- which presumably is why parts of the White House (the NEC I assume) are signalling that not all the tariffs will be additive and that some of the "IEEPA" tariffs could be replaced by reciprocal tariffs ...

10/

But the Bloomberg story also suggests a prolonged period of uncertainty, as the sectoral tariffs (232s) won't be part of the announcement, &amp; presumably the China 301 review will also continue.  Among other things, that implies more frontrunning

11/11

https://www.bloomberg.com/news/articles/2025-03-22/trump-plans-his-tariff-liberation-day-with-more-targeted-push?fromMostRead=true&sref=pKfngeVk

My regular reminder that foreign governors haven't been the marginal buyer of Treasuries for the last decade, and certainly not the last 3 years ...

1/ 

Official demand has been extremely flat the last two years -- with Chinese sales offset by other countries purchases.   China is adjusted for Euroclear -- and Chinese sales were stronger in 23 and early 24 than in recent months ...

2/ 

And unlike in 22 and 23, China is letting its Agency book roll off -- so on net it looks to be reducing its US bond holdings in the US data (subject to the caveat that it may be using custodians other than Euroclear more intensively)

3/ 

The marginal buyer, especially of notes, is now likely to be a levered US investor -- or an investor in a European financial center (UK, one of the custodial centers)

4/ 

Now globally there is a missing flow -- I don't think net FDI outflows from China alone are big enough to offset the true Chinese export surplus, so some must be intermediating the offshore dollars of Chinese exporters/ private firms (the oil flow has faded)

5/ 

There is no real mystery why the official flow to Treasuries has faded at the margin (sales of course can still matter, there is a big stock position) -- most countries have concluded that they have enough reserves so they aren't generally adding to their formal holdings

6/

There are some subplots on top of that: China may be diversifying at the margin (&amp; certainly using a broader set of custodians than the FRBNY facility) but the main theme has been strong private demand for Treasuries (given yield pickup and USD expectations)

7/7 

The old exportweltmeister has been dethroned -- and its economy is suffering at the hand of the new exportweltmeister (China).

That is the story told by both a new ECB paper and the FT in an excellent new piece

1/

Put simply, Germany is the most exposed large G-7 economy to the second China shock (Japan has been buffered by an incredibly weak yen). 

2/

The impact of the second China shock is in all the relevant data sets -- &amp; it reflects a clear Chinese policy choice: “As a country, the Chinese have been in the last years much better, more proactive, more consistent in going after the big technologies and conquering them”

3/

Germany's industrial economy has been in a 6 year slump, which corresponds well with China's pivot during the pandemic from an investment driven toan export driven economy

4/

That corresponds with a swing in the euro area's trade with China -- falling exports (absolutely and relative to EA GDP,), a rising bilateral deficit that corresponds with a falling global surplus once Irish tax distortions are stripped out 

5/

The FT highlights the swing in Germany's capital goods balance with China -- an important point.   The euro area's overall auto balance with China has also swung ...

6/

Germany of course was the one major economy that really relied on exports to China from 2008 to 2018 -- so it is simply more exposed to the second China shock.  Exports to China and HK are down a full p. point of German GDP

7/

And German auto exports to China are on a trajectory that takes them to zero ... or at least back to the levels last seen when China was still quite poor

8/

As an aside, auto imports are now down to about 2% of China's auto market (lots of import substitution) while China's exports are least 15% of non-Chinese auto demand globally if the closed US market is excluded (7m v 45m market)

9/

The FT story focuses on what Germany itself needs to do to respond to the new China shock (defense spending alone won't quite cut it, the internal EU market needs to be reinforced with a less "naive" / VW driven trade policy)

10/

The ECB's paper highlights the global dimension to Germany's industrial angst -- by highlighting how China's internal imbalances have fueled its export surplus (in a much clearer way than the IMF has ... )

11/

The ECB paper uses China's goods trade data (memo to the IMF, you cannot just use China's self reported current account numbers any more) to show that export growth and import growth have diverged, bigly

12/

That divergence maps well to the real depreciation of the Chinese yuan by the way -- Germany cannot afford to ignore currency issues any more (incidentally, the ECB paper largely ignored this, which is my only real criticism of a v good piece)

13/

Europe incidentally has not been rewarded for its relatively restraint (compared to the US) on trade protection -- China's imports from the EU have fallen by as much China's imports from the US since the pandemic (v trend)

14

The ECB paper also highlights something that is obvious in the data but that neither the IMF nor the WTO has been able to state clearly, namely that China is the deglobalizer in chief (import growth has massively lagged GDP growth)

15/

And the ECB finds that China's exports are outperforming the domestic economy across the board -- but especially in sectors where the domestic side of the Chinese economy is weak ...  

16/

The ECB's paper more than justifies France's plan to make the return of global imbalances a focus on its G-7 Presidency next year

17/

And the FT's Big Read should raise questions at the IMF about whether or not Germany is really still a big contributor to global imbalances (exports are falling, and domestic spending is set to increase) let alone a bigger contributor than China

18/

And it also should raise questions about the IMF's advice to China over the last 2 or 3 years, which more or less was to export your way out of the property crash more aggressively (monetary easing &amp; long-term fiscal consolidation = weaker CNY).  More later

19/19

A few reactions to Andrew Batson, and indirectly to Nick Lardy as well --

First, just b/c China has increased its social expenditures since 2010 doesn't mean it couldn't increase them more ...

1/

As Lardy recognizes (along with many Chinese economists) the minimum pension is modest and could be increased further, the system of unemployment insurance remains thin and fragmented along provincial lines, and out of pocket health care costs are high

2/

Second, China has expanded its social spending primarily by increasing relatively regressive taxes -- and the structure of the payments into the system of social contributions is (per the IMF among others) very regressive

3/

That has limited its impact on national savings -- the hefty tax on low income working households in the formal economy reduces their purchasing power

4/

Third, I at least have also advocated increasing China's personal income tax (which collects very little -- ~  1 pp of GDP) and funding the social insurance system more progressively.  Taxing high propensity to save households to fund benefits for high propensity to consume

Finally, I at least (unlike the IMF ...) have never insisted that the increase in funding for social spending be budget neutral.  In fact, to drive rebalancing, I have often suggested that it be financed by a bigger central government fiscal deficit

6/

And the main channel through which such a shift impacts national savings is through more government dissavings, not through any change in household behavior (tho of course a fall in household savings would be welcome)

7/

The change how China taxes (more taxes more progressive) and how it spends (more on household income support, less on subsidies for new factories in "involuted" sectors) agenda was always a rather hopeful agenda -- 

8/

If that positive agenda won't in fact reduce household savings/ raise household consumption help address China's demand deficit, my challenge to Nick and Andrew is to identify an agenda that will ...

9/

Otherise, a China that needs 1.5 pp of growth from net exports is on a collision course with the rest of the global economy ...

10/

One last suggestion: The IMF produced a series of useful papers on these issues in the late teens -- it would be useful to update them as part of the Fund's new focus on imbalances (the IMF's effort to be a fiscal policeman for China proved to be a dead end imo)

11/11

so there is still a chance that a fiscal deficit run to expand social expenditure could be a two-fer (traditional macro channels + change in household incentives to save :) ...) tho you put low odds on the second channel. fair enough ...

The Economist this highlights the dollar's recent drop &amp; the risk of a future fall.

It was striking that they made the argument essentially on fiscal grounds (tax cuts + a desire for easy money spook the bond market).   The balance of payments enters only tangentially 

1/ 

External factors weren't entirely ignored, but they weren't the focus either -- even though conceptually a discal crisis is a bit different than a dollar crisis (folks can sell bonds for cash w-o exiting the dollar)

2/ 

The Economist thus joins the IMF in, well, thinking almost exclusively in fiscal terms (see the IMF's market access debt sustainability framework) even though a true dollar crisis requires, well a flight out of the dollar into foreign assets ... 

3/

https://www.economist.com/leaders/2025/04/16/how-a-dollar-crisis-would-unfold

What the Economist left out is the clear deterioration in what I would call the dollar's external fundamentals (and unlike some, I do think they matter even for the dollar) ...

4/ 

The non-oil current account is back above 4% of GDP (and will be 5-6% of GDP in q1 with all the tariff distortions), the petrol surplus will fade away at current oil prices and the investment income surplus has disappeared

5/ 

The current account deficit is about equal to US manufacturing and ag exports (using the SITC definition)

6/ 

Adding in the "intangibles" export base (net FDI income , which essentially is all in the world's tax havens and IP related services exports) doesn't change the story much, as those sources of foreign earnings have also shrunk as a share of GDP

 7/ 

So the current account deficit has gotten to be large absolutely, and large relative to the United States sources of external earnings - and that deficit is now financed by large private inflows.   Reserve growth has dried up

8/ 

the net result, as James Aitken highlights, is that the financing of the US external deficit was a function of the "American exceptionalism" trade 

9/

https://www.aitkenadvisors.com/blog/the-end-of-the-beginning

It isn't enough though for the world to just hold on to their existing exposure -- to fund the US external deficit they needed to keep adding to it ...

US external debt wasn't rising v US GDP (which was inflated by the strong dollar) but it was rising relative to WGDP

10/ 

The dollar doesn't need to lose its central position in the global financial position for it to fall -- it just has to start to be a bit less exceptional (a self induced recession will do that ... ) relative to other financial assets globally

11/

Another way to think about the risks is that the world has been running a big carry trade, holding say US corporate bonds against yen liabilities or as part of a European pension portfolio ... 

12/ 

And when carry trades unwind, the "destination" currency tends to fall -- which would explain how US bond yields could rise/ bond prices fall while the dollar also fell ...

13/

https://www.nytimes.com/2025/04/14/business/us-dollar-trump-tariffs.html

Put differently -- the risks around the US aren't limited to a loss of confidence in US fiscal policy. 

The underlying BoP position has started to weaken + there are a lot of "stale" dollar/ US asset longs that were premised on a world and US that is no more

14/14

Lots of rumors (best I can tell without strong supporting evidence) of China selling the 10y (and more implausibly, the 30y).   A few observations.

1/ China entered Trump's second term with a lot of deposits and bills

thread 

Data is lagged, but China doesn't need to sell longer dated notes to get liquidity.   The state banks also have in the past uses the swaps market to borrow dollars to fund spot sales.

2/

The exact composition of China's treasury portfolio isn't known, but my sense is that it historically has had a bit more duration than the average central bank.  Even so China has a ton of maturing 10ys bought in 2015, 5ys bought in 2020 &amp; 2ys bought in 2023.

3/

SAFE also has a big equity portfolio and the TIC data clearly shows that China has been rebalancing its portfolio as the US stock market soared over the last 2 years (trying to keep its equity and USD share from passively rising)

4/

I doubt SAFE is doing this (it responds with a lag to market moves in my judgement) but a true portfolio rebalancer would actually be selling bonds and buying equities as the equity market sold off ... 

5/

SAFE's holdings aren't as big relative to the US market as they once were -- even counting euroclear/ belgium, US bond holdings are only around 4% of US GDP, and only a slightly higher fraction of the stock of outstanding marketable Treasuries (ex the Fed).  6/ 

And China simply hasn't been a big net buyer or net seller of Treasuries over the last 10 years.   The US BoP data shows a roughly flat overall portfolio, with a modest shift toward bills recently

7/ 

The baseline from the last few years is actually a bit of a passive rolloff (as maturing Treasuries aren't all reinvested in the notes market) and likely a bit of maturity shortening (why buy 10ys when 2s yield more)

8/ 

SAFE could of course be an active shorter of the 10y if it wanted to -- borrowing 10ys to sell them.   But if it did so the Fed would know pretty quickly ... as the GSIBS who would lend the Treasuries are are regulated  by the Fed/ US friends

9/

The data I use is lagged, but I think I have a reasonably good read on what SAFE has been doing over the last 10 and lat 5ys (read some of my blog posts).   Am all ears though if anyone has real market color ...

10/10 

Little Belgium is in the news right now, and not entirely for the best of reasons.* But judging from the US TIC data, Belgium doesn't have much to worry about 

*Belgian politics seem to have been captured by a custodian, which is a bit strange -- 

1/

For what it is worth, nothing much showed up in the OCtober TIC data, foreign private and foreign official holdings were more or less flat

2/

I am sympathetic to the Belgian argument that Euroclear is part of Europe's financial infrastructure, and that Euroclear specific risks should not be born by Belgium alone (even tho the Euroclear tax windfall was initially enjoyed by Belgium alone)

3/

At the same time, Euroclear shouldn't overstate its argument.   The initial freeze of Russia's reserves in early 2022 was WAY more consequential than the current debate over the use of the frozen assets (it was just done quickly, so there was less debate)

4/

And there is no evidence that the freezing of Russian assets adversely impacted Euroclear's role -- notably as a custodian for US Treasuries. Euroclear holdings have more or less doubled since early 2022 ... 

5/

Now China did shift out of US custodians (and more recently into shorter duration debt) back in 2022 ... but that was a boon to Euroclear and the broader European custodial ecosystem

6/

Bottom line -- not that many people understand the mechanics of the reparations loan (it works within the context of a freeze, it isn't an asset seizure, all the rhetoric aside), and Euroclear's concerns are wildly overstated.  It has done just fine after 2022

7/7

Japan is an interesting case in a lot of ways.  It has a ton of domestic debt (and significant domestic financial assets) which generates heated concerns about its solvency/ ability to manage higher rates.   But it is also a massive global creditor --

1/

Japan's net holdings of bonds (net of foreign holdings of JGBs) is close to 50% of its GDP (a creditor position as big v GDP as the US net det position).  That includes $1 trillion in bonds held in Japan's $1.175 trillion in reserves, + over $2 trillion in other holdings

2/

That translates into big holdings of US debt -- the MoF's Treasuries all show up in the US TIC data, but the corporate bonds held by the lifers, postbank and the GPIF are only partially captured in the US data b/c of third party management/ the use of EU custodians

3/

Japan has a big net FDI position as well -- so the net international investment position is much bigger than just the net position in bonds

4/

This translates into a 4.8 pp of GDP current account surplus even with the trade deficit -- all from investment income

5/

Coupon income from the bonds is now close to 2.5% of GDP - a real sum, even if it is smaller that the global profit of Japanese firms (Toyota etc)

6/

The other key thing about Japan is that a large share of the country's foreign assets are held by the public sector:

MoF has $1.175 trillion in reserves 

GPIF has over $900b in foreign assets, including over $450b in bonds

7/

Post bank has another $600b in foreign bonds (mostly corporate bonds, and mostly hedged) -- tho is it finally starting to raise its holdings of JGBS (its financial statement also shows how it became overweight in long-dated JGBs)

8/

Sum that all up and it is a huge amount of foreign assets -- over $2.5 trillion, mostly unhedged ... and with massive mark to market gains that the public sector could realize to reduce its gross debt at any point in time!

9/

right now the interest income on Japan's foreign assets is more or less offsetting domestic interest payments, so net interest is tiny -- despite massive debts (the rate on yen liabilities tho is poised to rise a bit)

10/

And capital gains on all of the foreign assets at MoF and the GPIF + smaller far fiscal deficits than in the US have brought net debt down to be within shooting distance of much of the rest of the G-7

11/

So to my way of thinking, heavily influenced by the strength of Japan's external fx balance sheet, some of the concerns about Japan's proposed fiscal loosening are overdone. Takaichi isn't proposing half of what Trump has tweeted out ... and her starting point is better

12/

And Japan has options that most countries with lots of fiscal debt don't have, as most countries with a large stock of domestic fiscal debt aren't also massive external creditors with big flow earnings on their offshore assets.

13/13

p.s. The flow out of Japan and into global bonds in recent years has been modest, a net flow of around $70 billion 0.2 pp of US GDP).  That is much lower than pre-COVID.

le to support government pension liabilities, money is fungible, tho my understanding is that Japan hasn't drawn on it yet.   Agree re Post bank tho with post bank Japan could reverse the flow into foreign assets and get a guaranteed bid on its own bonds (PB could go back to its

don't follow -- is there a link to the BIS set of concerns?

yes, interest costs will rise, that's clear from the JGB curve -- gamble is that real rates won't rise much.   seems like a good time to take profits on fx assets and reduce the debt stock to me.

spell out the long term risks and i might agree.

MoF can take profits easiliy.  GPIF can hedge/ shift back to JGS.  lots of options for reducing funding stress