meichen zhang @MeichenZhang

@cesifoti complexity.world/program2025#sl… great job!😃😃😃😊😊

@cesifoti It's great chance! Will there be an online channel to participate? Visa is too late😭

Is the US trade deficit as bad as Trump says? Or do trade numbers fail to account for America's amazing digital trade surplus? In 2024 the United States exported about $2 trillion in physical goods and imported about $3.27 trillion. At face value, that would translate into a trade deficit of about $1 trillion, or 3 to 4 percent of America's GDP. But for some time now the US has been exporting through routers and importing through ports. A prime example of this is the state of California, which in 2024 exported $183B worth of physical goods and imported about $490B. How did we pay for the difference? Using their amazing trade surplus in bits. Every time a foreigner streams a movie on Netflix or buys an ad on Facebook, California is exporting. Digital trade, however, is hard to capture because much it falls into what is known as "GATS mode 3 trade," which is trade delivered by having a physical presence in a foreign country (think of a subsidiary). A European buying cloud computing services from Amazon doesn't count as exports from the US because in Europe Amazon "esports" from Luxembourg. This means that there is a lot of hidden trade, like the invisible part of an iceberg. This is particularly true for digital products, that can enjoy a minimal physical presence in a foreign country. So how big is the US digital trade surplus? A few years ago, we began exploring this question by building a machine learning model with @ViktorStojkoski, Eva Coll, and @philippmkoch. The model transformed corporate revenue data and digital consumption patterns into digital trade estimates that we published in Nature Communications (nature.com/articles/s4146…) and that we have continued to update since. Our estimates are far from perfect, and intentionally conservative (we prioritize domestic consumption over exports as much as we can). Still, they reveal a fabulous digital trade surplus in favor of the US, with billions of dollars of surplus in categories that are not captured by physical trade data, such as digital advertising, cloud computing, video on demand, and payment services. We estimate that the US enjoys a trade surplus of at least $600B in digital products. This is comparable to the total exports of France, which is the 7th largest exporter in the world. When we dig into these categories, the statistics are even more mind-blowing. The top physical products exported by the US are crude petroleum and refined petroleum, with about $100B in exports per year each. We estimate US exports in digital advertising and cloud computing alone to be, respectively, about $260B and $184B. That means the US top digital product export categories are way larger than the US top physical product exports. So, is the US trade deficit as dire as Trump claims? Not quite—at least not when accounting for America’s digital trade. While traditional metrics highlight a trillion dollar deficit, the products generated by Silicon Valley’s "bits"—from cloud computing to streaming, generate an enormous surplus. The real story isn’t only about ports and shipping containers; it’s also about routers and data centers. It is about digital exports that are untouched by tariffs. If we measured trade in the 21st century the way we conduct it, the narrative may flip. The deficit isn’t disappearing—it’s just hiding in plain sight. You are looking at it as you read this message.

keyboard_arrow_left Previous keyboard_arrow_right Next

At the moment, the semiconductor supply chain is probably the most important global supply chain from a national security perspective. Here's how the supply chain works 👇

Finding it difficult to install Python packages into @Stata? This step-by-step guide, written by Chuck Huber, helps you understand how to install Python packages into Stata using pip. Find out more: bit.ly/Python-Stata #Stata #Python #pip bit.ly/Python-Stata

Baltagi’s “Econometric Analysis of Panel Data, Sixth Edition” is a standard reference for performing estimation and inference on panel datasets from an econometric standpoint and makes for a handy desk reference for seasoned researchers. buff.ly/3BhnCqI

Another econ postdoc on "R&D and ag productivity in a changing climate" to work w/ me & Bob Chambers (Maryland) in collaboration w/ economists @USDA_ERS. Position is approved & link will be coming soon when grant money arrives. Keep tabs on this if interested.

A great resource on all sorts of #AcademicWriting oacommunity.org/resources (cover letters, CV, articles, grants, elevator pitch, peer review, ...) Thank you, @OpenAcademics, for this collection of free resources! #phdchat #AcademicChatter

keyboard_arrow_left Previous keyboard_arrow_right Next

How to write a PhD introduction: 10 questions to help you get started + tips on length & timing buff.ly/3lyItP1 #phdchat #phdadvice #phdforum #phdlife #ecrchat #acwri

How to write a literature review: the purpose, where to place it in your PhD & what the scope & focus should be buff.ly/3cS5GaR #phdchat #phdadvice #phdforum #phdlife #ecrchat #acwri

Check out the freshly published book "Networks of International Trade and Investment" edited by @SaraGorgoni @aaamigh @SmithMatt90 with our (Massimo Riccaboni & me) chapter on World Input-Output Network vernonpress.com/book/96

Real footage of an economist cleaning data in R

Thomas Chaney and I are looking for summer research assistants (and beyond as FT pre-docs/research associates) to work on exciting empirical projects at the intersection of trade/econ history/growth. 🌍 1/3

Mark

NBER @nberpubs

5 years ago

A survey of the economics literature on global value chains, from @pol_antras and Davin Chor nber.org/papers/w28549

0 54 135 0 19

@tblrj Singapore has basically the same population as Denmark, and is on the top of the list. In general, country complexity estimated using exports data doesn't correlate with population (but complexity estimated with employment data for cities does). Fig 4: centerforcollectivelearning.org/s/s42254-020-0…

keyboard_arrow_left Previous keyboard_arrow_right Next

@TicketySplitUK 🤣🤣

R Views on rstudio::conf 2020 Workshops. @RStudioJoe reviews the 18 courses on offer, and where to find materials if you weren’t able to attend. buff.ly/37xTZ5n #rstudioconf

@YourPaperSucks Haha🤣🤣

Stay with you.❤️