We’ve been interested in how the federal government decides to buy things for a while: see past installments on biotech, oil, Afghan infrastructure contracting, NASA, challenge prizes, DARPA, and so on. But none of these topics has been quite as salient in recent years as how the government procures military equipment. There’s a growing Washington consensus that we simply can’t buy the weapons we need, in the quantities we need, on the timelines we need.
To better understand what’s going wrong, we talked to Dr. Arun Seraphin. Dr. Seraphin (the doctorate is from MIT, in Electronic Materials) has served in many roles, including as a Congressional staffer, an Assistant Director at the White House Office of Science and Technology Policy, and (currently) Executive Director of The Emerging Technologies Institute at the National Defense Industrial Association.
But we spoke to him about something else: Seraphin just finished serving as a commissioner on a 14-person “blue ribbon commission” to investigate reforms to the way Congress and the military coordinate to buy things. We got into:
How to design a commission to matter
Why the Pentagon’s IT doesn’t work
The value of pork
Directed energy weapons
Killed DoD programs
Is the Asian pivot happening?
You were appointed to the Commission on Planning Programming Budgeting and Execution (PPBE) Reform, which recently put out its final report on federal defense procurement.
How do you make sure a commission is useful to policymakers’ decisions downstream?
From an executive or legislative branch perspective, you've got these thorny policy issues, whether it's to move money, change the law, write regulations, change the way you manage a program. Some problems are so politically fraught and complicated that they just last forever. You can chip away at issues yourself, or you can ask an independent group for feedback. There are places you can go for that, like the think tank community in the government, Federally Funded Research and Development Centers, RAND, MITRE, and IDA in the DoD world, or the National Academies.
But sometimes, these problems are so high profile or politically fraught that you want to “turn it up to 11.” You go to what's called a blue ribbon commission. That gets you bigger budgets, more staff, and the ability to go deeper. It usually gives you more time and a higher profile of person participating. The commissioners can be former congressmen or former secretaries.
Defense procurement is an example of a perpetual thorny issue. But no one ever took a big, holistic, deep-dive look at it at a very high level. Why is that helpful? One, the resources let you turn over the problem and understand it. Second, with those higher-level people participating, in theory they can then walk around with the results and try to make change happen.
I helped launch a couple of these over the years, like the AI Commission; there's a thing called the 809 Panel, which was an Acquisition Reform Commission. Those are very high-level people. They did a lot of work. Then they came around and bothered the Pentagon, the White House, and the Hill. That high-level push from former secretaries and senators is helpful in trying to get this stuff done.
It sounds like the importance of a blue ribbon commission is not so much that it describes a fact pattern as that it points important people to specific problems. Is that a fair read?
Yeah. And depending on how the commission is structured, you can have active or in-name-only participation by senior people. Plenty of blue ribbon commissions have been accused of being staff products. You can tell the difference, because those members have slightly less enthusiasm to push hard on what the staff produced.
The ones that were most successful in my time were the 9/11 Commission and the AI Commission. There, the members of the commission worked on the recommendations. They internalized them, felt bought in, and were much better advocates. It won’t happen every time, but those commissioners can be very good advocates when done well.
Is another function of a commission to outsource political risk?
Yeah, that's exactly right. In fact, we built this group to be Noah's Ark style: two by two, one Republican, one Democrat. The commission I served on, the PPBE Reform Commission, was built that way. I was nominated by a Democratic member of Congress, Senator Schumer. The executive branch also had a few people. That gave it this bipartisan feeling, even on an issue as obscure and wonky as how the budget is built. That's helpful.
Let's get into the commission's report. How would you summarize the findings?
The commission was stood up because of a growing realization: it is very difficult for the Pentagon to put money in appropriate places in a timely fashion. Technology and threats resulting from it are accelerating in their pace of development and deployment — bureaucratic processes aren't keeping up anymore. There's been a ton of reform attempts on how the Pentagon contracts for good ideas, new technologies, and hiring people. There hasn't been a lot of discussion about how it handles money.
The high-level recommendations focused on building a budget based on the US’s understanding of the global threat environment. The highest-level reforms focused on how the Pentagon and Congress work together to build budget requests and final budgets, called appropriations bills.
After that money reaches the Pentagon, it has some ability to spend that money in a way that makes the most sense. You'll see recommendations on how the Pentagon and Congress can communicate better, IT systems to handle that information, and ideas for the rules by which you can move money for new things.
The goal is to create more speed, but in no way lose any transparency. The thing to remember is this is not private money. The taxpayer deserves some visibility into how this is all spent.
You mentioned the Pentagon's financial systems are not really designed to handle things that move quickly. What are they designed to handle?
The Pentagon's IT was built in the 1990s and 2000s, maybe even before that. It's built to help the Pentagon absorb all funds — this year it’s $850 billion — and distribute them across the Army, Navy, Air Force, Space Force, installations, and program offices all over the world to spend in the time frame that Congress deems appropriate. You can spend the money on a research account in two years, but if you're buying something instead of just researching, you get three years in a procurement account.
It's not designed to reflect the strategy, right? And it's certainly not designed for speed. When I say the system, that means both the IT, the people and their training to operate all of these, this enterprise, and the culture that wraps around.
Go a little deeper on the reforms the commission proposes.
One of the reforms tries to improve transparency by creating different ways for the Pentagon and Congress to share information.
The Pentagon has a process called Planning, Programming, Budgeting, and Execution (PPBE). It's a process that the Pentagon feels like it owns, but it turns out its ownership is shared between the Pentagon and Congress. The White House also plays a big role in building the Pentagon's and all federal agency budgets.
Also, the E in PPBE is execution. All that execution occurs in the private sector. This is a shared system between multiple parts of the national security enterprise, executive branch, legislative branch, and private sector. Unfortunately, communication does not exist.
What should exist are shared IT enclaves, places to quickly share useful information between the Pentagon and the Congress. Why? Every year, the Pentagon requests a budget. Congress is going to absorb that request and turn it into an appropriations bill. That data is shared in incredibly antiquated systems. Let's create some shared IT systems to know what the Pentagon is asking for.
An example of cultural reform is that when the executive branch produces its budget request, delivered around the State of the Union, it is culturally programmed to defend the request exactly as it was printed. The problem is that the Pentagon's budget is immense. It takes multiple years to build it. Some of those decisions were made well before changes in technology, threats, or requirements in places like Ukraine or Israel popped up.
Can you give me an example of that?
The burn rate of munitions in a place like Ukraine could not have been predicted two years ago. Ideally, you might have asked for more money to improve the manufacturing, not just the buying.
New threats have also appeared: more use of drones, for example. If you locked your budget decision process a year ago just because of its immense nature, it's probably months or years out of date.
There's a phrase that's used in Washington by executive branch employees: "I support the president's budget." I used to say that when I was a DARPA employee. Everybody knows you don't really support the president's budget. While that budget sits on Capitol Hill before becoming a final appropriations bill, can we have a more open dialogue?
That's a culturally hard thing to do.
Is there a tension between more open dialogue and getting the yearly budget process in on time?
No, I would actually argue the other way. Congress’s difficulty getting the information it wants slows down the process.
The volume of information produced annually, posted on public websites, delivered in briefings and hearings, and classified rooms is huge. How much of that is useful information? Not a huge amount.
What slows things down at the end are also the politics of passing a bipartisan budget. One party in the White House creates the budget request for the Pentagon, Health and Human Services, Social Security, and taxes. It's not shocking that whoever the opposite party is doesn't just accept everything the White House sends over. Why would they?
In this report, the authors, including yourself, call for “improvements in analytic methodology that enable DoD resourcing decisions."
What does that mean?
Two years before the Pentagon budget is delivered to the Hill, analysts look at the threat environment and the DoD's forces to see how well they align. What new trends are we worried about in counter-terrorism? Are we worried about Ukraine? Are we worried about China?
Then all of that strategizing is turned into programs and budgets. That is a huge analytical process. There's not enough workforce to do all this. IT tools aren't as good as they could be, certainly not as good as the commercial sector. There's also no way of even measuring your strategy, so the link between the strategy, budget, and outcome doesn't exist. We recommended a bunch of reforms, both at the people and technology level, but also even at the process level, like merging some of these steps and eliminating roles of certain sub-offices.
If these reforms are taken up by Congress, what vehicle will that be in?
You'll see these reforms taken up in three or four places, primarily by the executive branch itself. Many of the reforms are processes the executive branch can do itself, but bureaucratic cultural resistance keeps you from doing them.
Deputy Secretary Hicks has already adopted a lot of our interim report recommendations. The services below her have also started up these task forces to look at some of these recommendations and are starting to implement them.
On the Hill, you might see these things happen legally through the National Defense Authorization Act. For example, establishing a group to execute some of these recommendations and see their progress. That's a classic authorization kind of thing to do.
As our final report was coming out, we saw that last year's appropriations bill was also finished. They've already taken some steps to follow our recommendations. For example, when the Pentagon gets a budget, money is allocated, but sometimes reality changes. We said the ability to reprogram money between those accounts should be improved. We asked for a high level of reprogramming, and the appropriations committee has ratcheted up a little bit.
I think what you'll see in this cycle is changes in behavior in the executive branch and requests for more flexibility. Where all of that then becomes reality, does industry feel like the dollars are moving faster and more rationally? That'll take multiple years to see whether it sorts itself out.
You've mentioned that hundreds of provisions have been written to fix the defense acquisition process.
Why is it so hard to solve?
Truthfully, most of the provisions have been about contracting and different, clever ways of getting access to companies.
For example, years ago, we wrote Prize Authority. It took multiple years for anyone to pay attention to us, but it eventually took root. Prizes are a thing now, everyone does it, and it's wonderful, right? Other transactions have similar stories. It just takes a while.
We had a great interview with Rick Dunn on Other Transactions Authority.
[laughing] 25 years talking to Rick Dunn, I've never made that man happy. I've tried, right? On the other hand, we've never really gone after the money piece.
Why is that?
Because it's the money piece. It's so near and dear to people's hearts. It's also controlled with a financial management culture, which is worried about living up to the transparency requirements, following the letter of the appropriations law, and whose motivations, as discussed earlier, are just to absorb and spend a budget without measuring outcome. We're trying to inject some ideas that follow what we did on contracting and personnel reform, but this time in financial management reform.
And why are you hopeful about the current round?
The group was bipartisan and bicameral. I had a lot of Hill experience. Other people had a lot of Pentagon experience. In fact, there were also people with a lot of industry experience. It was a very well-informed group, arguably better informed than other commissions I may have started. For example, I love the AI commission. Some of its stuff was super clever, but that clever stuff will never survive in Washington. This was very informative. We're getting yelled out on both sides. It's not ambitious enough. It's crazy and too ambitious. But it's very practical so I am hopeful.
The other thing that gives me hope is that the executive branch has already adopted recommendations. If you look at the appropriations bill from last year, they made statements about speed being important in the system. They positioned themselves to say, “We're not the slow part of the system.” They poked at the Hill and the Pentagon. Now, both sides are saying speed is important and trying to show they're not the slow ones.
People say the DoD needs to be more ruthless in its priority setting, that it can't do everything at once.
What do you make of that idea?
It's always been true. $850 billion doesn't go as far as it used to. So you have to make choices. Your roof is always leaking, but you're always saving for college. You are always trying to balance those things, and the taxpayer is not going to give you blank checks, so you have to be as ruthless as possible in shutting things down. That gets back to this analysis piece. If you could inform the analysis on what's actually happening in the programs and the threat community, you probably can lay out your budget in a more informed way.
There's a political element to all of this. Every program has a set of people in the government that love it and a set of people in the private sector that love it. Those people in the private sector have representatives in the legislative branch that love it. It's a zero-sum game, a battle between where the resources come from and go.
As an outsider, my impression is that many of our political institutions are not especially good at this kind of priority setting. The priorities are set for us reactively.
If I want to be encouraged by our ability to ruthlessly prioritize, where should I look?
There are examples of Congress and the executive branch working together to get the DoD out of things. Programs have been terminated.
Can you give some examples?
A big army helicopter program recently had billions of dollars at stake. The Army killed it, and killed some big radio programs, joint tactical radio systems, and big command and control programs. A famous old army program, Future Combat Systems, was also killed. Despite political pressure to have a 500- or 600-ship navy, it still stays at this size, at three hundred. People are grumpy.
There are always examples of fighting through compromise and not spending enough. What do the people who would have gotten the money want? Is it as much as they wanted? No, but remember, one of the things to think about is that everyone who's selling a product always thinks their product is better. It's true if it’s Lockheed Martin. It's true if it's a cool Silicon Valley company. I have no reason to believe one over the other because they're both in it for profit and serving the taxpayer. There's no halo on any of them. So why should I assume there is?
Let me ask you about a recent Marathon Initiative white paper from Elbridge Colby and Robert Delfeld. They talk about how, in the aftermath of the Cold War, we saw a lot of geographic consolidation in the defense industrial base: “Three states receive over a third of DoD spending outlays.”
They argue that this consolidation limits the appeal of increasing defense investment within Congress, from a pure horse-trading perspective: legislators want to send money back to their districts, and most proposals for fixing America's defense industrial base ignore this political problem.
What do you make of that argument?
The allocation of taxpayer resources is going to be a political process. If you can tie the interests of local communities and multiple members of Congress to your project, you're more likely to get broader political support from the Executive Branch side.
It was always helpful for us to tie in multiple agencies. The Department of Energy gets some money, and the Department of Defense gets some money. They were all happy, right? The DoD, Army, Navy, and Air Force all have to be treated equally to build up the political will on the executive and constituent sides, which then feeds to the congressional side. Spreading your bets to as geographically wide a range as you can is good.
The challenge is the Executive Branch is charged with finding the best vendors, wherever they should be. For example, in a more free system, like how VCs spend their money, why do we see a clustering of VC money? There's nothing that prevents them from doing 50 states’ worth of money, but it must be economically efficient to cluster for other reasons. There's a political versus an economically efficient trade-off here.
Are there ways to grease the skids politically without this horse-trading or cutting everybody a piece of the pie? If you want to improve American manufacturing capability or defensive readiness, it sounds like you have to accept this inefficiency by spreading the wealth across all the agency branches and states.
Is there any way around that? Or is that just a brute political fact?
No. If you could just keep drifting up and up, you could probably define a politically optimal, efficient allocation of resources. The problem is the way Congress produces appropriations bills. In theory, 12 individual appropriations bills are wrapped into a single giant package. The executive branch produces budget requests by having individual agencies and sub-organizations optimize for themselves instead of what's good for the nation. There is organizational inefficiency and regional inefficiency on both sides.
You could use a giant supercomputer sitting above, trying to optimize across sectors, but horse-trading is what actually makes the system work. I guess if you drift too far into that command-and-control economy, you sound a lot like the nation we're competing against.
What's your technological priority list?
I would say the 14 critical technology areas, ranging from hypersonics to directed energy lasers to biotech. In my opinion, the thing that will change warfighting most is the application of AI.
With regard to the budget, we should think about the world we currently live in, where more money is spent by the private sector in some research and development areas than the public sector. The public sector will likely be outspent in software, bio, and AI. But the Pentagon can position its budgets to be an early adopter.
Other areas are still super important: space, hypersonics, directed energy, energetic materials for explosives, nuclear weapons, etc. Those unique military technologies really drive deterrence. There are these core military technologies that might be a little “yesterday” to the cool guys, but gosh, without nuclear weapons deterrence, without hypersonic strike capabilities, without submarines, you don't really have deterrents.
Are hypersonic weapons no longer cool to the cool guys? It sounds pretty cool to me.
Everything's AI and software, right?
Sure.
It's still hard to convince a set of innovative people that hardware is a thing.
Who are you referring to?
So the Pentagon has been flirting with Silicon Valley for 20 years, right?
Longer, right?
Yeah. A long time, forever. Federal investment in the semiconductor industry created the phrase Silicon Valley in the first place. But we’re yet to see Silicon Valley develop a combat vehicle that's being used as a hypersonic weapon that's being flown, stuff that actually might be used in a high-operational tempo conflict, right? Are they interested in that? I don't know. It's a slightly different style of investment and return on investment than that community is used to.
Does Anduril count as a success story here?
Yeah, Anduril's starting to break in.
How much should we expect Silicon Valley to play a role in this, given that the return profiles don't look like typical software investments?
I think it'll grow over time. The commission talks about this also. The Pentagon's money can now be used to signal and attract private sector investment. You can attract the private sector to take down technical risks, produce some prototypes, and then swoop in to buy them. The Pentagon can better position itself and become an addressable market.
We already do this with the medical and automotive communities. The DoD using this stuff then turns into commercial sales. Dollars going to “nontraditional contractors” increase investment from nontraditional sources into defense markets.
Let me return to the Pentagon's financial process. Why should readers see that as a major bottleneck compared to, say, procurement?
Why should we see it as an essential piece of this reform process?
When you say the word procurement, I think what people are talking about is contracting, or “How can I pass money to the private sector to develop products or IP for the military?” The tools that exist today in 2024 are so different from those that existed the first day I showed up in 1996. If anything, what's missing and what frustrates everyone is a lack of use of the tools.
On the other hand, this PPBE process was developed by Secretary McNamara in the 50s and 60s because it was commercial best practice to do things this way back then, and for the most part, has not changed since.
When you say it was commercial best practice, it’s borrowed from...
The automotive industry. It was a very structured way of saying, “This is what the CEO wants to do.” But it was not intended to be fast. It doesn't work well in a completely federated system. A CEO driving a company is closer to a one-party system than what we've got going.
Let me change gears a little bit.
The first page of the commission's report talks about the People's Republic of China as a “large technologically advanced strategic adversary with corresponding global reach that has profoundly threatened the rules-based order advocated by the US and its allies.”
The report later goes on to call China a “pacing threat.” Within the defense industrial establishment, how consensus is this view of China?
Nearly everyone believes that China is the long-range pacing threat. China is a very interesting combination of what we used to deal with the Soviet Union on a military global competition, and Japan in an economic global competition. It's both simultaneously, but with a very different system.
How long has that been the settled view?
The last 10 to 15 years. I started in this business in the 1990s. It was absolutely not true then. People were talking about it then, but it was a minority view. I think it became more real with China's economic growth, which — I'm no economist — took off in the 2000s. The distraction from the national security community was Iraq and Afghanistan during that decade.
What's going on in Europe?
The things you buy and the forces you structure to deal with a hot war in the Pacific may not be exactly the things you buy and the forces you structure to deal with the Russian invasion of Ukraine.
Let’s compare the Hill and the Pentagon on whether China is a threat. What does each institution broadly get right that the other gets wrong?
That's a good one. I think it's difficult for the Pentagon to see the competition with China as anything but a military competition. For the most part, they're not going to see the broader economic competition or even the people flows that are going on. They're going to have difficulty understanding how that economic competition, played out incorrectly, in turn affects their own ability to get what they want from their own industry.
If you view everything as a military question, it can drive you down the path of export controls on every last thing that might be useful to a foreign military. But if you do that with no economic context, you've basically told your own companies you can't sell overseas, in a world where you can't stay afloat unless you're selling overseas, and you're driving them to just disappear to failure, which then means you can't even get it yourself. That said, the Pentagon is way better at this than they once were.
From a congressional side, “Congress” doesn't think anything, right? There's nobody in charge of Congress: it’s a set of tribes. Depending on which tribe you're on, you can paint China in two stark pictures. You can assume there's nothing good we can do with them either on an economic or military front, losing sight of the fact that there is still good technology cooperation or scientific cooperation that can go on. On the other hand, a set of people in Congress and the scientific community do not fully understand that China is a competitor. China’s views on the rule of law, IP, and academic integrity are very different from anything we've dealt with before.
You can't paint either institution with a single broad brush, but there's a timeframe, culturally, that comes with dealing with China. They're very comfortable thinking about how long they will take to achieve their goals, which I think is very different than an American time frame. I don't know if we can with our quick return on investment on the VC side, quick political return on the political side, those styles, whether they need to be modified at all to deal with someone who's got way more patience.
Let me throw a couple more unrelated questions your way, and we'll wrap up.
You mentioned directed energy weapons. How close are we to seeing them as a prominent feature of modern warfare?
I think you're going to see routinely within this decade directed energy weapons used in dealing with smaller-level threats, drones, fast boats, those kinds of things. I think within the next 15 years, we’re gonna see stuff that starts to resemble what you might see in a science fiction movie. But the thing to remember is that it's not going to look like a science fiction movie, because you actually can't see anything. A directed energy war is pretty much a silent movie.
Disappointing for us.
Stuff just lights on fire, occasionally explodes. It's more like a kid with a magnifying glass burning ants than anything else.
If you had your full suite of reforms, whether from the commission report or generally from your time on the Hill, what are three that would meaningfully improve our defense industrial capacity?
Some of these are pretty in the weeds.
Please.
I think that Congress has provided the authority for the Pentagon to make much more use of its procurement dollars rather than its R&D dollars to drive technology and scientific innovation. Those reforms have not been taken up as much as possible.
There are lots of different clever contracting mechanisms that allow the private sector to spend money to bring technical risk down, and then when they finish and the robot jumps, the Pentagon steps in and buys a thousand. There are things like prizes for procurement, which are hardly ever used.
The budget reform committee talks about using your budget to signal where the market will be. In the tech space, making the Pentagon understand there are certain areas where you still have to drive R&D. More often, you just should be a buyer.
Another thing I think would be good is if we created mechanisms to create more flow between different sectors — government, industry, across agencies, and back to universities. We need more of that flow, the ability for people to move across sectors temporarily and share their experience so you have this confluence of interest. Because of all of their industry experience, they are smarter customers and smarter regulators.
The last one is oversight of the budget. Oversight exists to either protect taxpayers or prevent bad things from happening. Reform in the oversight process, which we've pushed out through data sharing, streamlines both reporting and testing processes. I would love for people to go after that.
How to Fix Defense Procurement