
"The Trade War Was a Problem Before Iran—Congress Can End It" via @KyleLHandley cato.org/blog/trade-war-was-…

Professor of Economics at @UCSanDiego @GPS_UCSD. Trail runner, gravel biker, news reader, international economist, tweets on all of the above.

"The Trade War Was a Problem Before Iran—Congress Can End It" via @KyleLHandley cato.org/blog/trade-war-was-…

Twilight is killing it tonight in SD. If we were still on daylight savings I would be making dinner and miss it all. These photos are 19 minutes apart

Reading the comments to this made me realize these Hitler Downfall memes are new to some folks. But this basically used to be the internet, even before Facebook was widespread. There was a whole Hitler Learns Topology meme shared among Econ grad students (and surely math majors as well) 20 years ago that is still out there reddit.com/r/mathmemes/comme…
Lmao been waiting for this….absolute cinema….no notes!

"Yet the finding that [the China Shock] caused one in four manufacturing jobs lost from 1990 to 2007 has not aged well. Neither has the populist turn against free trade that began around the time of the original research" 🔥 economist.com/finance-and-ec…

Multilateral controls on dual-use goods reduce leakage and adversaries’ real military spending. Acting alone, even with an embargo, cannot match the security impact, from Jingting Fan, @kylelhandley, and Nuno Limão nber.org/papers/w35898

So the question isn't just "how much trade gets blocked?" That's not a measure of success on its own, especially if leakage rates are very high. We need to know what relative security gain export controls buy, and at what civilian economic cost. Our "price of security" can be extended to compares those trade-offs across coalitions (see Fig 7) and sectors (see Fig 8). Thanks for reading. 10/10

One might be tempted to do better by going it alone, but US-only controls can backfire. Beyond a point, lost export revenue and production scale REDUCE the relative security gain even as civilian costs rise. Even a US embargo in the 10 controlled sectors can't match coordination. Using the model we can sketch out the civilian cost, to security benefit trade-off (the red-line below is a unilateral U.S. counterfactual, it bends back on itself because the own US losses outweight the effects on adversaries.) 9/

We are economists, so we don't try to game out nuclear, drone or cyber warfare scenarios. We compare costs on for allies and adversaries for military and civilians instead... Coordinated controls lower US real military expenditure too. But adversaries' falls more: US real military expenditure relative to adversaries rises 0.99 log points vs 0.14 under US-only unilateral controls. US civilian costs are small and similar. 8/

But there are TWO channels. Controls raise the military price index of adversaries relative to the US. Yet shifting production toward adversaries also raises their relative income, i.e. if they can’t buy the dual-use goods from allied countries, countries can make it themselves. This offsets about half the relative military-price effect. And the flip side is that U.S. and allied partners operate at lower scale, which imposes costs on them. Is it worth it? 7/

In our 2024 model counterfactuals, the “leakage rate” whereby other suppliers replace 28% of the reduction in controlled exports when the US acts alone, vs 7% with coordination. Adversaries' controlled imports fall 5.7 times as much under coordinated controls. (See Table 14) 6/

We then put those estimates into a global trade model with supply chain linkages, entry, civilian vs dual-use production, and separate civilian and military final demand. The last distinction matters: controls don't affect military and civilian prices equally. 5/

If the controls are not bans, then what are they? They operate like non-tariff export barriers. We estimate control frictions using exports, shipment size, and licensing outcomes. For CCL goods, the export-cost equivalent is about 29 log points higher for non-members than for multilateral-regime members. Export entry is affected too. This is not necessarily a bad idea though, we don't want a lot of things in the hands of certain end-users or for certain end-uses 4/

Controls aren't uniform bans. We digitized available BIS data for 2016–22, which shows average license approval rates of 93% for multilateral-regime members vs 72% for adversary destinations (technically, the EAR D-lists). Waiting times: 23 vs 43 days. In recent years approvals rates have fallen and waiting times increased. 3/

Export controls touch way more trade than you think. In our US export data (exc. Canada), 62% of goods exports in 2024 included dual-use items listed on the U.S. Commerce Control List or a broader catch-all control code known as EAR99, up from 47% in 2016. Depending on the destination and the good, these exports may have required a license, a license exception, or trade with no license required (but still reported as controlled). 2/

🚨New Paper Alert🚨: Export controls aim to limit adversaries' access to military-relevant inputs. But firms, supply chains, and incomes adjust. The countries imposing controls pay a price too. Using new data, we quantify when and how they can be effective and thus whether the trade-offs are worth it. 🧵 1/
New @nberpubs: "Export Controls and the Economics of National Security" nber.org/papers/w35898#fromr… viia @KyleLHandley

New @nberpubs: "Export Controls and the Economics of National Security" nber.org/papers/w35898#fromr… viia @KyleLHandley