Some thoughts on this thought-provoking piece on what to do about technology export controls by @MartijnRasser. He argues US needs to refocus use of the Entity List (EL) framework on preserving US tech competitiveness, to target competitors’ attempts at tech indigenisation.../1 https://twitter.com/cnastech/status/1336459795965091841

... & to work with allies. They need to be integrated into broader strategy. He says the Trump administration’s use of EL was ‘blunt’ & it needs to be used more sparingly & strategically. He also outlines the downsides of EL inappropriate/over-use:../2

... US tech getting designed out, costs & volatility for US tech exporters, helping foreign competitors, etc. /3

As first principles go, this is pretty good. He’s right – this stuff needs a rethink, not least because it was designed for a different world. And its clear going back to the way things were before doesn’t work..../4

...The old “Run Faster” strategy is not going to work with China’s market scale/subsidies/hacking/own tech abilities. And rethinking this hugely complex area is really complicated so props for taking it on. In the spirit of pushing the debate fwd, I have three broad questions./5

First, policy is only policy when it hits the road. I’d love to know how these principles would work in practice. I might have missed it, but in the piece there were no egs of a Trump admin EL action which met or did not meet these criteria. /

Second, I was left asking what the alternatives are for issues which do not meet the criteria? Third, what do we do when our allies don’t play ball? Those three in turn. /7

First, let’s get specific. The EL case which caused the most pain for US firms was Huawei. What would this framework have meant in that case? On the one hand, there’s hardly a firm in the world today which was more adapt at tech indigenization or,.../8

...with its global 5G infra roll-out, undermining western firms’ edge over an important platform. So the EL was correct? Well maybe not because at least before the revised rule which shut down access to TSM for Huawei, US firms like Qualcomm were losing sales, Huawei was.../9

... designing US tech out of its supply chain, and other countries did not seem that excited about working multilaterally on 5G. So, the Huawei action: right or wrong? There are other examples. There’s SMIC, China’s wanabee TSMC, which is not (yet) under EL but has been.../10

...classified as a PLA supplier, thus limiting sales of semi manufacturing equipment from US firms like Applied etc. Again, they’re central to China’s semi-indigenization program, so presumably they’d be OK to go on the EL in this framework? /11

But they’re still reliant on US semi-equipment and cutting them off some might argue will only encourage China to design-out Applied etc. Then, there’s YMTC, the 3D NAND firm in Wuhan which is slowly but surely ramping up to take on Micron, Samsung, SK Hynix. Again, it’s.../12

... definitely part of the indigenizing program. And Fujian Jinhua – who were EL-ed ostensibly because they stole US IP -  was that a legitimate use of the tool? Tough questions - not easy to determine. /13

Second issue; what’s the alternative? We can agree that use of EL on PRC firms involved in Xinjiang human rights abuses or building out the South China Seas did not meet the tech indigenization criteria. But what alternative framework do we have for raising those firms’.../14

...costs/dissuading others from doing similar? Should we use sanctions on the firms/Magnitsky-like moves on their corporate officers? Or do we rely on more classic diplomatic options for these issues? /15

Third, those pesky allies. Ideally we want a strengthened Wassenaar system, where the democratic allies together expand the list of what tech should not be exported to China. But the current system is only voluntary; everyone has to agree for a technology to go on the list,../16

...it’s a bureaucratic process, & often the Europeans & Asian governments have different views/priorities. What does the US do then? Trump admin pressure on the Dutch appears to have succeeded in stopping a license to ASML to ship leading-edge EUV machine to SMIC. /17

If SMIC are indeed supplying chips to China’s space/missile systems, then was that the right move? In sum, these are complex issues, hard to balance, they’ll always be losers. great we’re thinking afresh about them. Love to hear your thoughts @MartijnRasser if you’re game. //18