Michèle Flournoy on The US-China Strategic Calculus
Michèle Flournoy recently joined me on an episode of the ChinaTalk podcast. Eric Lofgren of the fantastic Acquisition Talk blog and podcast cohosted.
Jordan: Today, I am co-hosting with Eric Lofgren of Acquisition Talk. We are joined by Michèle Flournoy. Michèle Flournoy was Under Secretary of Defense for Policy during the first Obama administration and is the founder of both WestExec Advisors and the Center for a New American Security, where I'm an adjunct fellow. Welcome to ChinaAcquisitionTalk.
Michèle: Glad to be with you.
Jordan: How would you apply a social-psychological approach to US-China relations?
Michèle: That's a great question. We have a lot of work to do to understand how Chinese leaders think: what their strategic calculus is, what they value, what costs they fear, the fundamentals that would actually deter their problematic action.
There was a whole industry around this during the Cold War vis-à-vis the Soviet Union. There were people whose only job was to really understand how the Soviet leadership thought. I am not suggesting we're in a “Cold War” with China but I am suggesting them as a rising competitor. We have to do a much better job of really understanding their calculus and how to influence it.
Jordan: How would you rate the performance of that cottage industry during the cold war?
Michèle: It's a mixed record as in any of these things. I think there were some really good insights that helped influence U.S. policy, and then there were times when we came dangerously close to getting it wrong as in the Cuban Missile Crisis.
The lesson I take from that is that you need a pretty diverse range of views around the table. Not only the people who are narrowly focused on the question but people that are going to bring orthogonal perspectives to the table so that you can debate the issues, so that you can hear dissent and think about managing the risks associated with getting it wrong.
The Challenge of Peace
John Paul II in Japan, 1981. [Source]
Jordan: So another riff from your archives… I have two sentences for you, Michèle. “In the words of the Holy Father, we need a ‘moral about-face.’”
Michèle: What? That doesn't sound like me.
Jordan: Oh, no, it’s not you.
Michèle: Oh! Oh, okay. I thought you were saying that was a quote from me and I was like, “I don't think so.” Nothing against the Holy Father, I'm just not in the habit of quoting him.
Jordan:[Quoting from a pastoral letter written by the National Conference of Catholic Bishops, “The Challenge of Peace: God's Promise and Our Response”:] “The whole world must summon the moral courage and technical means to say ‘no’ to nuclear conflict; ‘no’ to weapons of mass destruction; ‘no’ to an arms race which robs the poor and the vulnerable; and ‘no’ to the moral danger of a nuclear age which places before humankind indefensible choices of constant terror or surrender.”
And McGeorge Bundy's response: “Nations have never come near to any agreement that would reliably eliminate nuclear weapons. Fear of the bomb itself has always been less powerful than fear of an adversary's bomb.”
Michèle: I think the question today is: How do we continue to reduce the number and importance of nuclear weapons in strategy, placing them on the “back burner” in the background of nuclear deterrence, but prevent rising powers like China and revanchist powers like Russia from ever thinking about using them?
I also think the real name of the game is opening up the aperture of strategic stability.
It's not just about nuclear weapons anymore. It's about how countries will use cyber tools. It's about how they will behave in the domain of space. It's about a lot of other things that can affect strategic stability between the nuclear powers. So I think that is a great area for dialogue with both China and with Russia. It's tough to get them to the table to honestly discuss these things but I think it's imperative to try.
Jordan: What do you think the debate around nuclear policy in the U.S. led by the Catholic Church can teach us? What lessons does that episode have today?
Michèle: Are you talking about the Bishop’s letter way back when? [Linked to above]
Jordan: Yeah. That's what I was quoting.
Michèle: You may be digging into the archives to discover that I wrote my Master's thesis on the ethics of nuclear deterrence. That may be where this is coming from.
I think it is worth trying to apply the principles of Just War Theory to how we think about the use of force generally, including the use or threat to use nuclear weapons.
We’ve integrated some key principles of Just War Theory into domestic and international law. Discrimination: we discriminate between military targets and civilians. We try to avoid targeting civilians directly but also avoid collateral damage indirectly. Proportionality: you want to make sure that any response to aggression is proportionate. I think there are some very important principles that came out of Just War Theory that are pretty well woven into how we think about these issues in places like the Situation Room today.
Ending Legacies: It’s Not Just The Ivies
Eric: In recent weeks, the discussion around divesting from “legacy weapon systems” in order to reinvest in a future of great power competition with China has been picking up steam. Some people seem to find legacy to mean systems like the F-15. We recently had the first flight of the F-15EX, but that's based on a model designed in the ’60s and first built in the ’70s. Some say things like the F-35 and the Ford- class aircraft carriers are also in the legacy bucket. So… there's a lot of definitions running around. Where do you stand on this legacy debate?
Michèle: I think it's really important to define your terms. The way I think of it is, we have a Program of Record. That includes everything from the F-15EX to JSF to the carriers and so forth.
What concerns the Pentagon leadership and others is that if you project that Program of Record, that planned force, into the future and you war game that over the course of the next decade, the United States loses its competitive edge. It loses its ability and we are at much greater risk of being on the losing end of any kind of conflict with China.
If you just continue with the plan you’re eventually going to be in a pretty bad situation. The question I think we should be asking is, “Do we want to trade-off some capacity, some quantity, in order to move that money into the emerging technologies and capabilities: AI, cybersecurity, compute at the edge, etc.?
Those are exactly the capabilities that, when they're integrated into the program force, will make it survivable, combat effective, while also having the reach and the lethality that will be needed to deter China in the future. So it's not an either-or . Nobody's talking about throwing away all the legacy systems and starting over. It's a question of how do you pair or integrate new capabilities and emerging technologies onto the force that you've got in place? What trade-offs do you make at the margin in terms of quantity versus investing in the new capabilities?
Rendering of F-15EX. [Source]
Funding, Funding, Funding
Jordan: We've seen debates on the Hill about funding. It seems Congress is pretty gung-ho to spend tens, maybe hundreds of millions, of dollars to reinvest in certain domestic strategic industries. Do you have any thoughts though, on what the trade-offs are there and how to best structure these sorts of incentives in a way that is economical and doesn't lead to the regulatory capture and suboptimal performance that you see in other countries’ industrial policies?
Michèle: I don't think we want to take a reshoring approach across the board, as controversial as that may sound. We do have to pay attention to the economics behind these supply chains and what will keep our own companies competitive globally.
There are key areas where we need to apply a national security lens and say, “where is it really important to have a home-based supply chain or a supply chain that is inclusive of our allies, but not necessarily reliant on China?” Obviously semiconductors. There's a lot of discussion about investing in a U.S. foundry. People talk about rare earths as another area of vulnerability.
Then we need to use Federal policy: whether it's leading with federal R&D money or whether it's tax policy to create incentives for private sector investment. There are lots of different levers the government can pull to try to attract investment in those areas and to make the economics work because of the strategic imperative.
Eric: How does the DOD work with Congress to free up more flexibility to bring in emerging technologies to complement the major programs that we've already invested a lot of time and effort into?
Michèle: Instead of throwing a budget over the transom to Congress each year, the DOD has to bring in Congresspeople and brief them on the wargaming, brief them on what happens if we just stay the course, brief them on the stakes. We need members of Congress sharing in that sense of urgency.
Congress should get credit. They've actually done a very good job of giving the Department authority to do things more flexibly, to prototype technologies, to use other transactional authorities, to acquire things quickly, to break out of the traditional requirements-driven process of acquisition. It is very sequential. It takes many years and it's totally antithetical to the way that software-driven systems and emerging technologies are developed, which is by using an agile, iterative approach.
The problem is the Department has not trained or incented the acquisition core to really adopt those at scale and to use them regularly. You still have this situation where promising technologies are prototyped and demonstrated and might be the greatest thing since sliced bread but then there's this Valley of Death between the successful prototype and actually getting a spot in the Program of Record. So that is where I think the DoD needs to do a better job.
Both DoD and Congress need to accept a little bit more risk in the development cycle of critical technologies. You have to accept a little bit more risk, not into the war fighter downstream, but in the development cycle of the technology because in an agile process you're going to have some failures. You're gonna learn from those failures. You're going to address them and build better capability. You have to tolerate learning from a mistake rather than beating people up on the Hill every time such a failure in the development process occurs.
“Green Berets”... For Tech Acquisition?
Eric: I want to get your view on two kinds of solutions that can help reinvigorate the Department. First is the idea of a “bridge fund.” In 2019, you brought this up to Congress because you need some kind of flexible pot of money to get emerging technologies over the Valley of Death. Another view is that bridge funding might just push the Valley of Death a little bit to the right. That it's not really solving the institutional problems of portfolio management.
Michèle: You need both. I do think occasionally you'll need a bridge fund. When you have a successful prototype with an end-user customer who wants it and is willing to put money against it in the Program of Record, but just because of the budget cycle that point in time is 12 to 18 months away. So how do you use that time to continue the development and the refinement of the capability so that you're that much farther down the road when you actually go into production? That's the bridge fund case.
I think portfolio management is really important because it would be hard to find a CEO that's in the world of manufacturing that's not doing some form of portfolio management. You have to be able to regularly assess how different efforts are performing against an objective and move resources around to accelerate the ones that are working and cut off the ones that are failing or put them on the back burner. The way our system works doesn't really allow that.
We should get Congress to give the DoD some kind of authority for portfolio managers to make those trade-offs. You can require reporting or testifying or engaging Congress, and so forth. At the moment our lack of flexibility is really hurting us because even when we know we should make a trade-off, it's very hard to actually make the change.
Eric: We have the legacy of an industrial-era system. How do we modify that in order to address the problems that we're facing today? Because we can't have 20, 30, 40, 50 year plans and presume that commercial technology won't change in that time.
Michèle: I came into DoD in the ‘90s as a reformer thinking we needed to completely change the system. We do… but over time I’ve really come to believe that you really can't boil the ocean. If you try to change the whole acquisition system, it's a noble cause but you’re tilting at a windmill. So I think the more important thing to do is to take a sub cadre of folks and really train them well on how to manage technology and emerging technology programs and how to accelerate development and adoption. Eric Schmidt said, “DoD does not have an innovation problem; it has an innovation adoption problem.”
We haven't trained people and we haven't incented them. We haven't made it a great thing for their career.
We need a sort of cadre of “Green Berets” who are emerging technology acquisition experts. It doesn't have to be a huge number of people. If we were to train and incent and reward them for it, I think we’d get faster development and adoption of some of these critical capabilities.
We’ve also got to help DoD to become a better, smarter customer. This means fostering a better understanding of tech and its development and acquisition cycle, as well as understanding the business models. What to negotiate in a contract. How to drive price down or costs down when you're dealing with an industry partner. We don't invest enough in the tech or business acumen of our professionals and that would make a big difference.
You see work that's being done in the Navy, where they really are driving down the ONM costs associated with key platforms, like the F-18. What's happening there is the Navy is becoming a smarter customer in terms of negotiating those contracts and dealing with that supply chain. So there's a lot of goodness there's a lot that we could get out of investing in that human capital.
Eric Do we know how China approaches that human capital aspect in relation with its industry?
Michèle: China is a totally different system. The sort of civil-military fusion, which is: if something interesting is happening in the commercial world, they step in and direct it to be shared with the PLA. That is obviously not a system that would work for us but we need to find our own answer to that in terms of accelerating innovation and the adoption of innovation.
Eric: The bipartisan Future of Defense Task Force called for a [Manhattan Project for artificial intelligence (AI). Your colleague Bob Work called for something slightly different, a Rickover-style reform for AI in the military. So what's your take on what is needed in the AI+ML for scaling the various solutions, as well as enterprise tools and architectures, we need.
Michèle: I think machine learning and AI applications are going to be critical not only for the U.S. military and for speeding the quality of decision-making for the military by enabling human-machine teaming with operators and unmanned systems. It's also going to be really critical to our own economic competitiveness commercially around the world.
I’ve just been reading through this National Security Commission on Artificial Intelligence report chaired by Eric Schmidt and Bob Work. I think this is one of the most important permissions since the 9/11 Commission Report.
They argue for putting us on a national footing with investments in science and technology, research and development, human capital, the whole nine yards, to really be competitive in AI across the board. I would like to see the vast majority of those recommendations implemented.
The point about Rickover is a good one too. One of the big challenges in DoD for both civilian and military leaders is the short tours that they do. Military leaders, general officers, flag officers typically turn over every two to three years, maybe four years at the most. It's very hard to do change management over three or four years.
When a CEO takes on change management at a company, they plan for a decade of stability and leadership to do that. So we're making it very hard on ourselves with all the churn. Then on the civilian side, the average life of a political appointee is what, 18 months? 24 months? That doesn't help either.
I think in areas where we really want to double down by making a big bet on technology to accelerate our advancement, we've got to stabilize some of that leadership to make sure we can drive the change over time.
Eric: I love Rickover because he backed up his plans with workforce. He spent so much time recruiting people. He interviewed tens of thousands of people coming into the nuclear division. Then he spent all this time making sure they're trained up. I think he started on nuclear reactors in the late ‘40s and he was there until the ‘70s or early ‘80s, even. It seems like we need that kind of stability and leadership.
Michèle: He understood that human capital is the long pole in the tent. Not only recruiting the right talent, but developing them, creating career paths for them, incentivizing them, managing that pipeline of talent over time and really creating that whole cadre. That was what was going to ultimately put nuclear power on the map for the Navy.
I think there is a huge bunch of lessons to be learned there. One of my favorite things to ask nuclear Navy folks is, “Tell me your Rickover story.” Every single one of them had to sit down with this very tough Admiral. It was a classic stress interview model. And so there's some great Rickover interview stories out there.
Hyman G. Rickover, Admiral in the U.S. Navy. [Source]
Eric: He would ask people to delay their engagement and marriage until after they've completed. He’d put them into closets. He’d cut off part of the leg chair so that you're rocking back and forth. It's definitely an interesting model. You can't just boil the ocean, as you said, we have to pick areas to build out the human capital. Then with that human capital, then you can leverage the flexibility and take things off the rails, a little, potentially.
Michèle: I agree. Human capital and the right incentive structure. I don't mean financial. Other kinds of rewards: promotions, or opportunities that really align behavior with your objective. Those are two critical things.
I will just note that a lot of people talk about “how do we get outside tech talent into the DoD?” That is very important and we should be doing better there, but we have an enormous amount of tech talent in the military, graduating from the academies every year.
Two-thirds of each class are in some kind of STEM major, but they come into the Officer Corps and then we put them in something that has nothing to do with their tech background.
Even if they want to be a technologist for their service, they're told that's career-ending, a blind alley. You can't make it to flag officer if you're a technologist. That has to change, we just got to start making better use of the tech talent that we actually have in our hands, but we're not managing them well.
Jordan: How did we get to this point?
Michèle: I think a lot of it is the outsourcing of most of the tech and engineering jobs to the private sector. The folks inside the department are not given as much of a role. I think given the era that we're entering and the criticality of people who understand how to integrate commercial technology into military systems, we have to change that balance.
Eric: I have some nostalgia for the Bureau in Arsenal days of old where the government would actually help develop at least components or some pieces in-house. But then a lot of people also saw them as like backwards, right? They were fixed to their own things and they loved their own legacy solutions. So there's ossification, if you let it go too long. What do you think about a system of churn that keeps the field dynamic?
Michèle: I’m not suggesting we go back to the Arsenal Model but I am suggesting we need people who understand technology peppered throughout the system as operators, as experimenters, as concept developers, as acquisition program managers, all of that. There has to be more tech savvy throughout the ranks for all of this to work. We have to really make it exciting for people.
We don't have the concepts and systems that we need to be effective with China long-term. If you tell the current generation of young officers, “We want you to be part of the solution. So you're going to do competitive concept development for your next chair, and you are going to take new technologies coming out of Silicon Valley and experiment with them for a tour, or you're going to be the one to figure out how to to accelerate getting this really cool new thing that's going to save lives into the fleet faster or into the force faster,” it will become more attractive
But we're not making those the sexy jobs for young people coming in.
I think there are pockets of the force that are doing that. I think one of the advantages of SOCOM and the special ops world is they're so small that it’s much easier to gain that agility. They have such urgent requirements from guys who are in harm's way every day, that they've actually done one of the better jobs of figuring out how to bring in new technologies and adapt them for military purposes. They get them into the field very quickly, albeit at a small scale because they are small. But that's the kind of muscle that we've got to develop Department-wide.
The other thing I'll mention is coding. Right now the department is dependent on industry for everything. There's some coding that we should be able to do in-house, simple coding for different applications. In some corners, there's now a move towards low-code or no-code building of AI applications where it's literally putting building blocks of existing algorithms together to build new applications. That's the world of the future. We've got to prepare the tech talent to be able to do some of that in-house.
Jordan: There's also an aspect to this where once you spin up this kind of work working for the department of defense starts to sound more appealing than being the three thousandth project manager at Facebook.
The Indo-Pacific’s New Reality: The End of Sustained Superiority
Eric: Jordan and I like to debate about the relative importance of basic research and applied research. Where do you land?
Michèle: I don't think it's either/or. I do think that we need to be investing more in basic science and technology research but… Given the urgency of what we're facing with China and the fact that we don't have enough good answers to some of the asymmetric approaches that they're going to confront us with, I think we need to put a particular emphasis on applied R&D and moving that into development and rapid fielding.
I think in two time frames. There's a sort of 5-7 year timeframe, which is really not about new systems. It’s about how do we take what we have and use it differently to get a new capability?
For example, if you want to have options for a President other than striking the Chinese homeland, the homeland of a nuclear power (the President better have other options in that), but you want to be able to hold at risk forces that are projecting power to take Taiwan or to intimidate a partner in the South China Sea.
The first question is how could we do that better with what we have today? Maybe you pair a Navy long-range precision munition with an existing air force bomber. We’ve got to think out of the box across services to say: “We have everything on the table. It's the Apollo exercise.” How do we mix and match these pieces that we have to solve this hard problem.
Then looking beyond to 10 years and beyond is where accelerating the development and adoption of emerging technologies can help you. That you're going to get something fundamentally new that you can integrate into the force in the mid-to longer-term. You gotta think in both of those time frames to be successful, I think.
Eric: You mentioned that China is approaching us in an asymmetric way. They're focusing on systems destruction warfare as well, going after our concepts of how we operate. What are you seeing there and what does the DoD need to do to read?
Michèle: Systems destruction warfare is the idea that China will try early on to disrupt, defeat, and degrade our networks.
Their preferred scenario is using cyber attacks, using maneuvers in space that mess with our ability to see, to navigate, to move, to target before we even get out of the East Coast bases or the West Coast bases.
The key thing is that the environment that we will encounter in the Indo-Pacific is going to be highly contested and it's going to stay contested. The traditional approach from the Gulf War and since is, the U.S. comes in, establishes domain superiority, gains air superiority, gains maritime superiority, gains superiority in various areas and then has the freedom of action to go in and prosecute campaigns.
That moment of sustained superiority is never going to happen. We will get it. We'll lose it. We'll have to regain it. It's going to be episodic.
And so what we have to do is think asymmetrically about China: meaning how do we fight in that more contested domain? How do we build a resilient network of networks to be able to continue to operate effectively and communicate and command and control in that contested environment?
But also where are the points of vulnerability for China? They've focused on cyber attacks on critical infrastructure, attacks on space-based systems that could blind or mobilize us. What's the equivalent insight for us? How do we take an asymmetric approach to them?
I do think the network of networks, the resilient, what's being called join-all-domain command and control or advanced battle management system (ABMS) by the Air Force. That's a really key piece. I think leveraging AI for decision support so that we can cull through the data and get the insight and make the decision faster than they can. That's critical human-machine teaming.
The Air Force’s advanced battle management system (ABMS) has come under a lot of fire recently from defense insiders as well as congress. One problem is that there are a lot of interesting exercises, but critics point to exercises being all over the place (e.g., a robotic dog patrolling Nellis AFB). This seems to translate into a messaging problem.
I think Michele put ABMS in the right context. China’s strategy is to degrade our networks that allow the US to deploy its exquisite assets. Since the Korean War, the US has basically had free rein in terms of its C4ISR infrastructure. Even in that permissive environment, weapon systems are developed in stovepipes that cannot easily interoperate without a significant time penalty associated with manual processes.
A satellite that detects and tracks a target cannot relay that information directly to an available aircraft. That’s one sensor-to-shoot possibility, but in a degraded battlefield the US will need many ways of sensing and shooting. Importantly, every potential pair of systems that needs to talk to each other cannot wait years for the acquisition system to churn out a solution.
Perhaps that’s why ABMS seems to be running in so many directions. Ultimately, the Air Force will have to decide what user stories are particularly important when thinking about resilient command and control and focus on getting some high-value wins. But the program will continue to come under fire because defense analysts love to talk about the kinetic effects of weapon systems, such as payload, range, accuracy, and so forth. ABMS doesn’t have that kind of well-defined military requirement.
One potential analytical lens is to say: Suppose China takes THIS capability off the table. That means we cannot deliver THESE kinetic effects. A survey of the types of at-risk capabilities, their overlaps, and how they impact military effects will start to uncover the highest valued priorities for the ABMS roadmap. But it also can translate into dollars — or the opportunity cost of military weapons that can no longer use their capabilities in a degraded environment (when they are needed the most!).
We're always going to be at a quantitative disadvantage in their backyard in the Indo-Pacific given just the time-distance equation. How do we marry unmanned systems with manned systems to get some of that capacity back and regain some of that advantage and be able to operate in a very lethal environment where we may not want to send 5,000 people on an aircraft carrier? It's asymmetric thinking. As important as the technology is, the mind-share, the intellectual conceptual development piece is really critical as well.
Xi Jinping sends out his first Weibo, a New Years message to the People’s Liberation Army. [Video]
Michèle: Jordan, you get the award for being the first person who's ever asked me about my Oxford thesis.
Jordan: I actually got a scanned copy of your undergraduate one as well.
Michèle: Oh God. Did you see all the different typefaces and the pages didn't match? My typist for my thesis…He was accepting chapters for weeks and the night before it was due, he came and gave it all back and said, “I’m so sorry. I've overcommitted. I can't do this.”
So I had 20 friends divide up the pages of my thesis all night and then slap it together with a mea culpa letter from the typist at the front. So that it wouldn't be downgraded because of the way it looked. You must've seen that and said, what's her problem? Why can't she type properly? But now I know where the psycho-social question came from…