Outbound Investment Controls: A Deep Dive
“How indispensable is US venture capital to Chinese firms in these sensitive sectors? That’s the key question.”
At long last! The US’s new outbound investment screening rules are here … sort of.
ChinaTalk sat down with Emily Benson (CSIS) and Martin Chorzempa (PIIE) to discuss the White House’s new executive order and the US Treasury’s Advance Notice of Proposed Rulemaking (ANPRM) on novel outbound investment screening rules targeting China’s AI, quantum, and semiconductor industries.
Why it took so long to get new controls after the shock and awe of October 7;
How the controls became so narrow, and what the public comment process achieves;
The dilemma US policymakers face regarding access to China’s leading tech sectors;
The allied dimension of America’s investment screening play;
What comes next for the regulations and firms impacted.
The Long March to Outbound Controls
Jordan Schneider: So, what’s happening with these new outbound investment controls?
Emily Benson: For the first time ever, the administration this week released an executive order asking the Treasury to close a perceived gap in trade and investment authority.
This is under the assumption that we control outbound items, technology, and software with export controls and that we also screen inbound capital through the Committee on Foreign Investment in the United States for national security purposes.
That begs a sort of simple and obvious question: why don’t we already control outbound capital to countries that pose a national security risk?
This is not the first time we’ve had this discussion. This is not unique to the Biden administration. It surfaced initially in 2017–2018, during the ECRA and FIRRMA debates that resulted in a comprehensive reform to these trade and investment authorities in the United States.
In the last year or so, the administration has become increasingly keen on getting this thing through, and talks have really accelerated in the past couple of months.
What’s interesting is that this executive order was imminent since last October, when the Bureau of Industry and Security at the US Department of Commerce levied these sweeping export controls on advanced AI chips to China.
The October 2022 export controls had a much greater effect throughout the foreign policy community. It had the effect of slowing down this impending executive order.
When it slowed down, that created a greater opportunity in the interagency process for additional stakeholders — including people like US Treasury Secretary Janet Yellen — to intervene and say we really want this tool to be sufficiently narrow. We don’t want the scope to be too wide. We want to make sure that it achieves this end goal of protecting US national security while not overly ensnaring our allies.
So talks have been very much under wraps the past couple of months. What’s interesting is that in the month of July, things changed in Congress. There was recently a fairly resounding vote in the Senate that signals a broad, bipartisan desire to get something like this finished.
That added increased pressure on the administration to actually get this order out of the door.
That’s what happened two days ago when Biden signed this EO directing Treasury to stand up this new program. What it’s doing is opening up for public comment this new program to screen outbound investments and capital in semiconductors, microelectronics, artificial intelligence, and quantum technology.
Martin Chorzempa: You can’t export an advanced semiconductor to China, send chip design software, or send equipment that makes the most advanced chips — but you could invest in Chinese companies trying to make advanced chips or chip-making software.
The idea is there was a loophole where US capital could do what US exports couldn’t. Ultimately, the effect of that would be that US capital would help China become so good at making its own stuff that it wouldn’t need US Stuff anymore. US export controls would be neutered.
That’s the core concern this was trying to address. It’s very much related to October 7.
What Does Congress Think about All This?
Martin Chorzempa: On the politics, it’s been really fascinating to watch because the initial proposals from Congress were really broad and did not get traction. There was an interesting partisan divide here — very different from the usual.
Some Republicans were very much for this and were driving factors behind the idea of having this outbound regime. Other Republicans said, “Absolutely no way. This expands the power of government and gives it the ability to ask questions about all sorts of investments that we haven’t done before. We’re an open, free-flow-of-capital kind of country.” That led to a lot of opposition.
The Biden administration was in a tough position where the legislative branch couldn’t really get a law passed that everyone could agree to. They kind of said, “Okay, Biden administration, do something yourself. And by the way, there’s no political consensus on exactly what you should cover.”
That makes it really challenging to get something.
Having an excessively broad or poorly written version of this would be enormous and could bog down US investment in allied countries that have a subsidiary in China. It could lead to a huge financial decoupling between the two countries that was not intended and could really damage the US’s ability to access Chinese technology and know-how. It could bog down US firms that have operations in China. It could damage the US position as a global financial center.
There were tons of awful scenarios that could have been the case if they rushed something through. It’s really positive they took it slow.
We’ve been really careful and asking lots of questions to make sure we don’t get into these nightmare scenarios.
Jordan Schneider: How genuine was the GOP opposition?
Martin Chorzempa: There’s real sincerity there.
There’s a long-standing tradition in the Republican Party of being suspicious of the expansion of government, red tape regulations, and oversight of firms. That clashes with the idea that many firms are helping China in ways that hurt US national security.
House Financial Services Committee Chair Patrick McHenry argues we are actually protecting China from our own national security threat.
CFIUS reviews inbound investment into the United States and prevents Chinese citizens fromom sitting on the boards of sensitive US firms. We don’t want their VC investment. We don’t want anything because we’re worried they’re going to learn all sorts of secrets that could hurt our national security.
At the same time, we’re saying we don’t want any Americans on the boards of Chinese AI companies or any sensitive Chinese companies. It’s like, well, wait a second … if we think it’s going to hurt US national security to have the Chinese on our boards, wouldn’t it harm Chinese national security to have our people on their boards? Wouldn’t they learn interesting things about what’s happening in these sensitive markets?
There’s this assumption that the US is ahead of China on everything and that any interaction with China benefits China more than the US. If you have a more nuanced view — which I think Pat McHenry had to some extent — you can acknowledge there are a lot of things we can learn from China.
They are doing really interesting things in these sensitive technologies — not necessarily things that we like or agree with. But involving US Investors in these firms could benefit US national security as well.
Emily Benson: The more targeted derisking or decoupling there is in certain high-tech sectors, the less visibility you have over time into these foreign sectors. That can have adverse effects when we’re trying to figure out where we are maintaining a leading edge or where China might be pulling ahead.
Those data points are extremely important in formulating policies that can combat the current threat environment. With reduced visibility, we just naturally have less knowledge about what’s happening in those sectors. That’s one downside of additional controls.
Cutting Off Access
Jordan Schneider: The US is not going to be the dominant player in every technology either now or in the future. This executive order seems to argue that there’s a net-negative trade-off that comes with boosting an adversary’s techno-industrial capability even as you benefit from the knowledge transfer.
Martin Chorzempa: There’s a historical counterargument to that completely valid argument.
The NSA was against China having digital telecoms and wanted to keep them in analog because it was easier to snoop on them. When they figured out that China would get digital telecoms from someone else, they said American companies should absolutely sell them the digital stuff.
We knew their vulnerabilities and capabilities because we knew exactly what equipment was going in there. If we banned it, they would get the equipment anyway. We would lose that visibility. The current situation is analogous.
But how indispensable is US venture capital to Chinese firms in these sensitive sectors? That’s the key question.
It’s not clear to me just how much of a benefit they get from this versus, say, a Chinese venture capitalist who spent many years and cut their teeth at a US venture capital firm.
I think we agree that the capital is pretty interchangeable. But on the know-how and network side, I just haven’t seen a rigorous argument for how they’re getting a huge boost here from having US venture capital in their funding rounds. I just don’t know.
Jordan Schneider: It’s a revealed preference, though, is it not? These funding rounds keep happening with the hottest firms, which are ostensibly closely aligned with China’s five-year plan and push toward self-reliance. It seems like an odd play for the US run over the coming decades.
Emily Benson: The US Treasury partly intends to set up a mandatory notification regime so that they can collect more data. The Advance Notice of Proposed Rulemaking from Treasury does not require prior notice of a transaction. It says to notify the US Treasury within 30 days of having closed the transaction. That’s a very concrete signal that this is to collect more information.
Of course, there are national security considerations in these very targeted prohibitions.
Will the data reveal if any of this will actually bear fruit in the long run? Will it be worth it? Will the nature and volume of transactions essentially substantiate the entire standing up of this new program at Treasury?
I’m not sure. But there may also be a gap in current US Statutory authority. I think there probably is. So let’s get more information on that. If there is a viable risk, let’s deal with it. It’s a little bit early.
The public comments will reveal exactly the contours of what this needs to look like. They will reveal the companies working in the three covered sectors and show which ones are particularly exposed.
What to Expect When You’re Expecting … Great Power Competition!
Jordan Schneider: What can policymakers do when Chinese firms outpace the US technologically? One day, Beijing very well may flip the dynamic with its own export controls. Should we expect America it some point to conduct its own industrial espionage?
Emily Benson: The realization that China might be pulling ahead right now is that meme with the panicking head. “Oh my gosh. We’re not the leader in every single technology ever. That is uncomfortable.”
Stephen Walt argues in Foreign Policy that Americans are fundamentally uncomfortable with a multipolar world. That’s sadly true. Now we’re acknowledging finally that we actually are not the leader in all of these technologies. That means we need an affirmative agenda for pulling ahead where we can.
This is where a solid trade policy comes in.
It really plays into the promote side. Going back to the European Economic Security Strategy, the US must protect, promote, and partner with allies over a long time horizon, maintaining efficiencies and open trade and investment wherever possible.
But at the end of the day, our national security could be better protected than it currently is. Squaring that circle is very tricky, especially when you have competing constituencies.
Jordan Schneider: I’ve been spending a fair amount of time thinking about the US’s competition with Japan in the 1980s. The irony, from today’s historical vantage, is that Japan was a treaty ally. It didn’t have a military and had a deeply cooperative posture toward the US.
Now fast-forward to today. “America Plus Friends” will make up about two-thirds of the global economy whereas “China Plus Friends” will likely top out at no more than a third. Keeping that in mind makes these challenges less daunting.
Martin Chorzempa: I think that’s a crucial point.
If you take these restrictions too far and get too extraterritorial, you’ll start to alienate the friends and will have a much harder time building a coalition that really matters.
You have to really sell the national security justification, and they have to buy it. This is tricky because sometimes the US thinks, rightly, that it needs to lead by example and push along other countries that don’t want to incur any costs.
At other times, however, the US has to restrain itself and not overuse its tools.
One of the things that makes us so nervous about this is that China may use the same kind of tools on the US. We have shown that we are absolutely willing — when we think it affects our national security or commercial interests — that we will weaponize key nodes in the supply chain. We have good reason to think China would do the same thing were it to dominate in these areas. There’s an irony in that. We’re a democracy. We talk about upholding liberal order. They don’t really have that. It’s not the equivalent thing.
But China is looking at how we have used these tools, and it might follow us in ways we don’t like.
Jordan Schneider: All right, so let’s take dive and talk about what dropped on Wednesday. What’s in there? And what’s notable?
The Solace of Quantum
Jordan Schneider: There’s two ways this can go wrong, right? It could be too big and burdensome, or it could be too small and not get at the problem.
But if all we’re going to do is notify and collect data, why does that have to be only three technologies?
Last year the White House put out its list of fifteen critical emerging technologies. We’ve got engineering materials, gas turbine engine tech, directed energy, renewable energy generation and storage, and all this other stuff.
But these new outbound controls ended up focusing on only three critical technologies — semiconductors, quantum, and AI.
Semiconductors were probably the most predictable. They kind of copy-and-pasted a lot of what was in the October 7 export controls and said, “Look, if you’re doing anything over these thresholds or in these types of industries that we’re trying to control, that’s not cool. We’re not going to let you do it.”
We haven’t yet had big export controls on quantum. They asked a lot of questions in the comment period — “Did we get it right? Is there any quantum stuff we missed?”
Then AI was the most curious. There’s a big, open discussion within the broader AI commercial research community about whether foundational models or specialized models will be the ones to rule the world. Yet the TAC the Treasury used for framing focused on defining this stuff as specialized models.
Then we have military mass surveillance, digital forensic tools, penetration testing, control of robotic systems, and facial recognition. But those models are very different from GPT-4 which is multimodal. But if AI ends up developing in a way that sidelines specialized models, then the key is just to have the biggest, baddest foundational model you can apply to any nefarious thing.
Is AI the wrong focus for the US Government or is it the all-important, key dual-use technology of the future?
In the second part of this conversation, we discuss:
How the new controls differ from CFIUS, and why that’s good;
The scope and loopholes inherent in the restrictions;
A deep dive into the specifics of the regulations around semis, quantum and AI, and what they entail for future industries like biotech which may come on the chopping block in the future
How US allies are (sometimes) getting on board with these new tools;
How these regulations may end up providing a playbook for a “larger yard” going forward.