1/11
A new paper by the NBER on the McKinley tariffs of the late 1890s claims that the US economy did not benefit from the tariffs, mainly because they "may have reduced labor productivity in manufacturing."
2/11
Tyler Cowen (along with a number of other economists and journalists) argues that this paper is evidence that if the US were to impose tariffs today (or other trade intervention policies, presumably), they too would hurt the economy.
3/11
But this argument makes the same mistake as claims about the similar lessons of the Smoot Hawley tariffs of 1930. It treat tariffs a little hysterically, either as inherently and always bad for the economy, or as inherently and always good for the economy.
4/11
But tariffs are neither. They are simply one of a huge range of industrial and trade policies that work (much like currency devaluation) by shifting income from households (as net importers) to producers (as net exporters).
5/11
To put it another way, tariffs work in large part by forcing up the domestic savings share of GDP. For that reason their impacts on the economy must depend in large part on whether investment in the economy is constrained by scarce savings or by weak demand.
6/11
In economies running persistent trade surpluses, saving exceeds investment by definition, with the very purpose of trade surpluses being to resolve weak domestic demand. In that case policies that further weaken domestic demand and boost savings are not likely to help.
7/11
On the contrary, they need the opposite policies. That is why most economists, for example, call on China to implement policies that increase the consumption share of GDP (i.e. reduce the savings share). China should, in other words, reduce tariffs and strengthen the RMB.
8/11
But the impact of tariffs on deficit economies will be radically different. In that case by pushing up the savings share, these economies can either enjoy more investment and growth, or the same amount of investment and growth driven by less debt.
9/11
The US had been running large surpluses for over 20 years in 1900 and for over 60 years in 1930. It is not at all surprising that increasing tariffs was unlikely to benefit the economy. Surplus countries should implement the opposite transfers.
10/11
Today, however, the US has been running massive deficits for roughly five decades. It should surprise no one that policies that benefit the economy under one set of imbalances are unlikely to do the same under a set of diametrically opposed imbalances.
11/11
That's why instead of pounding the table about whether tariffs are inherently good or inherently bad, we should instead discuss what the conditions are under which tariffs (and other trade and industrial policies) will or won't benefit the economy.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
1/7 Good Martin Wolf piece on the global return of mercantilism. What is new about this piece is that it seems part of a growing recognition among global opinion makers that mercantilism and trade war didn't start when deficit economies with... ft.com/content/cd68b3…
2/7 open external accounts began to implement trade restrictions and otherwise control their external accounts. It started earlier, when economies that controlled their external accounts implemented trade and industrial policies that led to beggar-thy-neighbor trade surpluses.
3/7 We are returning, in other words, to Joan Robinson and her 1937 explanation of how trade conflict emerges. What I would add is that in a hyperglobalized trading system (i.e one in which transportation costs, communications costs, and the costs of... ia802806.us.archive.org/16/items/essay…
1/6 Wall Street bankers and owners of movable capital would hate it, but if the rest of the world were to reduce its dependence on the US dollar, this would be good for the US economy, good for US manufacturing, and good for US farmers and workers. wsj.com/finance/curren…
2/6 The claim that the US benefits from the global use of the dollar is one of those things that people believe even though they can't explain why – except perhaps in terms of sanctions. None of the world's fastest-growing economies (including... foreignaffairs.com/united-states/…
3/6 advanced economies like the US in the 100 years before the 1970s, Germany in the same time period, or post-war Japan, Taiwan and South Korea) had major reserve currency status, and yet they all had rapidly growing economies driven by even more rapid growth in manufacturing.
1/4 Bloomberg: "“Even with strong determination and sufficient resources, transforming China’s economy into one driven by consumption and services will take years,” Goldman said. “With a more reluctant, measured approach, it could take decades.”" bloomberg.com/news/articles/…
2/4 Goldman is right, of course, unless a debt crisis, or a serious acceleration of trade war, forces a much faster, disruptive adjustment. While the latter might happen, the former is, for now at least, pretty unlikely.
3/4 A long adjustment, however, means a Japanese-style adjustment over two or three decades, in which consumption growth continues at more or less the same pace it had in the past while GDP growth drops sharply, and investment growth goes negative.
1/10
SCMP: "Kenya has reached a preliminary trade deal with China for duty-free exports of key products including coffee, tea and cut flowers – a major step towards narrowing the East African nation’s long-standing trade gap with Beijing."
via @scmpnewssc.mp/gg0zg?utm_sour…
2/10
This kind of incrementalist thinking is one of the reasons why global trade is so unbalanced and so poorly understood. China does not run a trade surplus with Kenya because of tariffs on coffee, tea and cut flowers.
3/10
It runs a massive trade surplus with the world because of equally-massive domestic imbalances. Reducing tariffs on Kenyan coffee, tea and cut flowers will have almost no effect at all on China's domestic imbalances, and so no affect on China's need for a trade surplus.
1/4 Aggregate financing in China, the most widely-used proxy for total debt, ended 2025 at RMB 442.12 trillion, an 8.3% increase over last year's outstanding amount. This is a relatively small increase in total debt compared to earlier years. english.news.cn/20260115/3e5af…
2/4 But of course nominal GDP growth is also much lower, so the RMB 35.6 trillion increase in aggregate financing in 2025 represents a 12 percentage-point increase in China's debt-to-GDP ratio. This is higher than the 11 percentage-point increases in 2024 and 2023.
3/4 China's debt data isn't always comparable over time, but I think only the COVID year of 2020 saw a higher increase in China's debt-to-GDP ratio, and because this was partially reversed in 2021, the average annual increase over the two years was only ten percentage points.
1/5 NYT: "The U.S. trade deficit in goods and services shrank to $29.4 billion in October, down from $48.1 billion the prior month. The figure was the lowest monthly trade deficit recorded since June 2009." nytimes.com/2026/01/08/bus…
2/5 If this persists, it may be the most important factor for those thinking about what is likely to happen in 2026. In a three-month period during which the Chinese trade surplus has surged, the US trade deficit has declined.
3/5 Simple arithmetic tells us that the difference must be reflected in the trade balances of other countries. Some of this will have showed up initially as rising trade deficits among developing countries, but this will ultimately be limited by their abilities to finance them.