How to Buy Stuff Like DARPA Does

"He was going to find his head chopped off"

Government agency procurement processes are legendarily clunky. If an agency wants to acquire something, or ink a contract, it has to follow the Federal Acquisitions Regulation, or FAR. As a result, government contracting is far slower than private sector contracting. Federal contracts routinely go to the same narrow set of contractors, and run well over budget and schedule.

But a few agencies manage to avoid this rigmarole entirely. Instead, they’ve pioneered the use of Other Transaction Authority, or OTA: which lets them sign flexible contracts. Among the regular users of OTA: the Defense Advanced Research Projects Agency (DARPA), largely responsible for the invention of the Internet, GPS, drones, and the personal computer. And SpaceX’s first deal with NASA was through OTA.

Today, we talked to Richard “Rick” Dunn, the man who brought Other Transaction Authority to DARPA. Before serving as DARPA’s first general counsel, he served at NASA, negotiating the Space Act Agreement involving the first launch of a privately owned space launch vehicle in 1982.

This is an arcane topic, but I don’t think this is an overstatement: the best work the federal government has done with the private sector over the last 30 years has happened through OTA. Without Rick Dunn, much of that work might not have happened.

As you’ll see, Dunn believes far more agencies could use OTA far more often. So why don’t they?

---

When you were general counsel, you proposed DARPA receive Other Transaction Authority, this freedom to make flexible contracts. How’d that come about?

I had been in the role for about a year or so, and on my desk came a tasking from Congress. The House Appropriations Committee had put into the Authorization Act, a requirement for DARPA to write up a report on an alternative management system. The concern was really compensation for scientists and engineers: can you get world-class talent if you’re paying civil-service salaries? We were going to do a study of that. I ended up being the primary government author, and we brought in a consulting firm to actually do the compensation part. By the way, I have a background in this: my LLM [Master of Laws] is in labor and personnel law. 

In any event, we had this tasking from Congress, and so in addition to looking at the issue of compensation, I took that opportunity to plug in a section that said, “DARPA needs flexible contracting authority.” Because we don’t do business, or shouldn’t do business, the way everybody else does. That report went through the Pentagon legislative affairs team and so forth and went to Congress.

To clarify: Congress hadn’t specifically asked for information on this, but it hadn’t ruled it out in the report? You seized your opportunity.

Yes. My role was unique. I had a direct reporting line to the General Counsel of the Department of Defense, and only a dotted reporting line to the Director of DARPA. I almost never saw the general counsel of the department – it speaks to the dysfunctionality of the way the Pentagon organizes things.

I may be getting into too much detail about this sort of stuff.

This is the perfect amount of detail.

As DARPA’s first General Counsel, what was the organization like when you arrived in 1987?

It was in a strange state. DARPA had been around since 1958. For a long time, it relied on outside counsel as well as doing its own contracting. That’s in part because DARPA wanted to stay a small, flat organization, and didn’t want more administration and overhead. But that was starting to change. DARPA planned to establish a Contracts Management Office and was in search of a General Counsel.

I applied for the position, a senior executive position, and then the position notice was canceled. A few weeks later, it came out again, and I just decided not to apply, because it seemed like the fix was in for somebody. 

But then the general counsel of NASA called me in and said, “Are you interested in executive advancement? DARPA called, and they’re looking for a general counsel, and they may be calling you.” That’s how my journey with DARPA began. 

When I went over to the interview, things got even stranger. I spoke to the director of DARPA, and as I was leaving, he said, “I hope they hire you.” I was confused, because I thought he was the one making the decision: who’s “they?” It turned out he was dual hatted. He was also appointed as the director of defense, research, and engineering at the DOD; he still had his DARPA hat on, but he spent most of his time at the Pentagon.

The director had two deputies, one for technology and one for research, who appeared to hate one another. And the reason the personnel announcement had been canceled was that the deputy for technology wanted to bring in his fair-haired boy. But because it was an SES (senior executive service) position, the candidate had to go through an SES-quality review. They rewrote the position description to take some of the requirements out, but he couldn’t pass that. So eventually they decided to downgrade the role below senior executive service. I was lateraled in as a GS-15 and eventually became a senior executive a year or so later. That’s probably more detail than you needed to know.

DARPA was located in Rosslyn, Virginia, next to the Potomac, in the Architects Building, which had actually won an award from an architectural magazine years before. It was pretty run down when I was there. DARPA had originally been in the Pentagon, and when it moved out of the Pentagon, many people thought, “they've really distanced us, we've been downgraded.”

Later, not being in the Pentagon came to be seen as a great benefit. It was a physical removal, but it also encouraged this idea that, “We're DARPA, we do things in a different way.” Before I had gotten there, there was a question: “This lawyer, is he going to be just a contracts review lawyer, or is he going to be a general full-service counsel?” My background was such that I was capable of handling a range of issues.

Can you tell me about your experience with Other Transaction Authority at NASA? What wisdom did you bring over to DARPA from there?

My summer job during my first year of law school was in the Apollo procurement office at the Houston Manned Spacecraft Center. That was my first exposure to government contracts, and I won’t go into detail, but the head of that office was a really slimy character. The government shouldn’t treat its citizens, including its contractors, the way I saw the government doing business. 

In any event, I went on active duty in the Air Force for about 10 years, and then a year and a half in private practice before joining NASA. I was originally there as head of labor law, doing personnel work. For some reason, soon after I joined, I began to receive special assignments. I don’t know exactly how that happened, but I went to my boss and said, “Boss, they’re not coming through you: they’re coming directly to me.” 

But one of the first assignments I got was related to Deke Slayton, one of the original seven astronauts, but he was medically disqualified and became Chief of the Astronaut Office. He had retired, and was associated with a group of people that wanted to pioneer private space launch capability. Slayton was the kind of guy who could make a phone call to the administrator of NASA.

When was this?

1981, early 1982. So the NASA administrator then asked his General Counsel if NASA could legally assist a private company in demonstrating commercial launch capabilities. This was a private company that was not contracting with NASA, that didn’t have anything to do with NASA. 

I was assigned to look into this, and wrote a 12-page legal opinion. I concluded that not only did NASA have the legal authority to assist, but it was also part of our mission to aid the industry in the utilization of space. As a result, I got involved in negotiating the Space Act Other Transactions Agreement launch demonstration. And what NASA contributed was rocket motors. In 1982, they launched a vehicle from Matagorda Island, Texas. It did not go into orbit, but it demonstrated all of the phases of putting a satellite up. There was a mock satellite on the vehicle, there was a platform, there were communications involved. They basically demonstrated that they could put a satellite into orbit. 

That was the first agreement that I was involved in, and I wrote the underlying legal opinion and then actually negotiated the agreement. It's worth mentioning why lawyers were so involved in this. At NASA, our practice for Space Act agreements was unique, especially for those involving “other transactions.” The procurement team didn't handle these. The lawyers did.

We had a couple of lawyers who specialized in launch service agreements. These are a type of Space Act agreement where NASA provides launch services and gets reimbursed. Usually, when a government agency receives outside funds, it goes straight to the treasury. But NASA's General Counsel, who actually wrote the provisions of the Space Act, took the position that NASA had the authority to reimburse its appropriations.

So you took that experience with OTA to DARPA.

That started my work in Space Act agreements, and I eventually reached the conclusion that NASA was way underutilizing its authority to do this sort of thing. The Space Act gives NASA a wide-open authority. It allows NASA to enter into “contracts, leases, cooperative agreements, and such other transactions as may be necessary” under terms that the administrator shall determine. It's just an open-ended authority.

Why did NASA have that broad authority before anyone else? 

Because Paul Dembling, the author of the Space Act, intentionally wrote it into the statute, and his view was that most government contracts are highly regulated compared to the world of procurement contracts. Contracts can be almost anything that two or more parties agree to. And yet the government finds itself limited. If you’re the Department of Interior, you sell timber, but you have a special set of rules for selling timber and so forth. 

Dembling was interviewed for NASA’s 50th anniversary, and he laid out how people would come up and say, “Can you do a transaction like this? Would this break the rules?” Eventually he wrote in the words “such other transactions as may be necessary” into the Space Act, and that’s how OTA started. 

NASA’ first use of that agreement is also very instructive. NASA and the Department of Defense were interested in communication satellites, they had various programs going. One of them was basically to bounce radio signals off a balloon in orbit. But they also had another project to forward the message to the satellite, and when it comes around the next time, it downloads the message. Then a very sophisticated project, a communications satellite in geosynchronous orbit: nobody knew whether you could actually put a satellite in geosynchronous orbit.

Then AT&T, Bell Labs, comes along and says, “We're working on this, but we have no way to get it into orbit, and therefore we need you, NASA, we need your launch capability.” This raised a couple of questions.

One, the government already has its own communication satellite programs. The balloon was actually in orbit. They'd actually bounce signals off the balloon. And the other two programs were in some stage of development. So, “does it make sense for us to have a competing commercial version of it?” And the policy question was, “do we have legal authority to do this?” Because of the broad language of the Space Act, they were able to engage AT&T, and AT&T designed, built, and owned the satellite, Telstar 1. 

It was the first privately owned satellite to operate in space, the first active communication satellite in service. It had sensors on it, and as it passed through the Van Allen belt, it gave us an initial understanding of radiation damage to communications equipment. 

It was launched in 1962. So less than four years from the time the Space Act had been enacted, there was use of this authority in this rather unique way, doing business differently. We were getting paid to launch things. 

So you'd come from NASA to DARPA with a sense of the power of this transaction authority. And when you got the chance to tell Congress “We want to have this at DARPA too,” you jumped at it.

Absolutely.

And how difficult was it to convince Congress? Did you do active work on the Hill?

In addition to the report that I wrote, one other thing that ended up on my desk was a letter from some distinguished former defense officials, the "graybeards," most of whom had contributed to the Packard Commission Report of 1986, highlighting inefficiencies in defense procurement. This report and their input helped me push for adopting flexible transaction authority at DARPA, similar to what we had at NASA.

The Packard Commission pointed out that many defense projects were too expensive and took too long. This perspective came from people deeply involved in defense during the 1950s. Think about that era – we developed Intercontinental Ballistic Nuclear Missiles and deployed them in just a few years. The Von Neumann Commission report in 1954 was a turning point. It noted that because hydrogen bombs were smaller than the older fission bombs, it was now feasible to deploy ICBMs.

In just five years, we were able to develop a fully functional ICBM. We developed entirely new systems in five years. So these fellows were looking at the state of defense acquisition in the late 1980s, and said “you guys are screwed up.”

They had written a letter that basically exhorted officials from the Department of Defense to try to be more commercial in nature, to be more innovative. That letter had bounced around the Pentagon, it was sent to DARPA, and ended up on my desk. So I drafted a reply which basically said, “You guys have great ideas, but we can’t do them under our current legal authority. But if we had more flexible legal authority, we could do them.”

You enlisted them into your project.

I’d say we had a mutually beneficial relationship. 

The head of that group was retired 4-star General Bernard Schriever, also known as the father of America's ICBM program and Defense Space Program. He was the kind of guy who could call up Sam Nunn, then chairman of the Senate Armed Services Committee. Schriever briefed what he was interested in, and there was this guy at DARPA attending. About a week later, the deputy director who had been my advocate in the hiring process had a meeting with Sam Nunn, and Nunn said, “Hey, these graybeards came in and they started talking to me about this stuff. What’s the deal?” And the deputy director said, “my general counsel knows.” Senator Nunn replied, “I’ll have my guy call your guy.”

Nine months later, Other Transaction Authority was law. But it wasn't smooth sailing. It wasn't just the Senate. We had to get through the House and there were people that raised objections.

Ultimately, I backed off and said, “We're only going to do this for research, it's limited to DARPA,” and so forth. And that was later expanded. 

Was that your way of just getting your foot in the door? 

Absolutely. My idea from the beginning was a full-scale alternative to the acquisition system. So we would start with research, prototyping, and then the whole shebang. 

And why did bounding it just to research initially make it an easier sell for Congress? 

There are fewer budget implications. It's not everything, it’s just one place, and it was DARPA-specific. It didn't look like it was goring anybody's ox. 

Were you involved in the push later to expand OTA authority to more parts of DoD?

I was consulted. I don’t think I pushed that per se. I drafted the prototype authority language.

How did getting OTA change DARPA’s operations?

I’m not sure where to start… It got the director of DARPA fired.

Well, say more on that.

The first OT was with a venture-capital supported firm that was doing work in gallium arsenide. The authority was enacted in 1989, and the first agreement was signed in April of 1990, with a startup that had about three rounds of venture capital support, about $12 million, and they were developing chips with gallium arsenide instead of silicon, and there was some defense interest in that. 

Politically, what was going on at that time, of course, was the Cold War was ending, and there was this question of what the government's involvement would be in directing technology, in industrial policy. Some people were adamant that the government should not be involved in industrial policy, picking winners and losers. 

Were those people made more powerful by the end of the Cold War? Was there less support for the industrial policy DARPA was doing as the Cold War ended?

I don't think DARPA conceived of what it was doing as industrial policy at all. DARPA was interested in enhancing defense capabilities. And one of the techniques for doing that was to have a strong industrial base in whatever the technology was that defense needed. And it wasn't just a technology baked with the big contractors, but a vibrant national industry that we were interested in, and getting beyond the narrow, kept cadre of contractors that DOD normally does business with. In any event we had this first agreement, which I've written about, was managed by Arati Prabhakar, who later became the director of DARPA and is now science advisor to the president.

She managed the technical aspect of it, and I partnered to get the agreement negotiated. The agreement looked nothing like a procurement contract. It was about five pages long. The way she conducted oversight was she went out to California and sat in on the company's board of directors. 

It was just completely different from the way the government normally does it. Because we were supporting a VC firm, we wrote into the agreement that, if the project is successful and goes to market, the government will receive a payback, and the payback can either be a royalty on product sales, or the company's increase in value can be prorated to the government investment and the government will be paid. 

So there were options, but the point was the government could make money out of the deal.

Because you were putting in money before this was a mature company, right?

It had no products, it was entirely VC-supported, and we were basically acting as your next stage of VC, so we were going to get compensated in that way. When the company publicized it, it looked like the government had taken an equity interest in the company, and that was enough to cause a firestorm. There was a paper published in the Harvard Business Review that quoted Vice President Cheney [then the Secretary of Defense] saying if Dr. Craig Fields, the director of DARPA, stuck his neck out one more time, he was going to find his head chopped off. This was the occasion to do that. Of course, he was a senior executive and the government could move him where they wanted to – he wasn't fired, but he was moved to a closet in the Pentagon.

So days after that, the Deputy Secretary of Defense [Donald Atwood Jr.] was before the Senate Armed Services Committee, and they said, “What did you do with Dr. Craig Fields, and what's going on with this agreement with this little company?” 

Apparently the people in the Pentagon believed that we at DARPA had gone off the deep end and done this all on our own. Actually, we had gone over and briefed the undersecretary of defense for R&D ahead of time [Charles M. Herzfeld, who had previously been Director of DARPA and authorized the creation of ARPANET, the Internet’s predecessor]. I said we were thinking about starting a small OTA experiment and working up from there, and he said, “Congress gave me this authority and they meant for you to use it, so you go ahead with this agreement.” 

But Herzfeld was out of the country when this hit the newspapers. There was no one available to say, “Hey, a senior person signed off on this.” So when the Deputy Secretary got before Congress, he of course denied that Fields had been fired. They said he’d been reassigned, he’d been given a new opportunity. Fine. 

But they required all future agreements to be run through the director of procurement and acquisition policy, who knew nothing about non-procurement agreements. We were really slowed down for a year and a half until the administration changed. The last senior “advice and consent” official who was in place was another former director of DARPA, and he told the procurement folks, “You’re no longer reviewing DARPA procurement.” Sorry to get into all this detail.

No, this is exactly what we’re looking for! 

OTA at DARPA is clearly seen as a success, right? In that OTA has been rolled out to other agencies, parts of DOE, the CHIPS Act deployment. How long did it take for other transactions to become widely recognized as a useful tool? 

As I said, we went into a hiatus in early 1990. We got to 1992 with very little action going on. There were only a few agreements, and people started bailing out of the department after the election. And Congress got involved: again, we’re at the end of the Cold War, defense is downsizing, acquisition budget is shrinking.

Defense firms are under stress. People start thinking that although DOD is really technically advanced, the nation is going to lose that technical advantage, it’s going to atrophy.

So Congress enacted a series of eight programs to address this issue of defense companies under stress. They need to commercialize. We need to have defense technologies flowing into the private sector. We need to have private sector technologies flowing into defense. We need a workforce that can handle new technologies. Anyway, a mumbo jumbo of issues. And the results of that can still be seen in law today. I think Chapter 381 and following in the current version in the US Code, Defense Technology and Reinvestment Act is what it was called. 

In addition to these programs that were enacted, half a billion dollars came to DARPA. That was at a time when DARPA's budget was only a billion and a half, so a very substantial increase. and the statute said, the secretary of the defense may use other transactions in carrying out these programs.

It was such a big addition to our program, and some of the elements of this defense reinvestment statute really smacked of NIST (the National Institute of Standards and Technology) and DOE (Department of Energy) and NSF (the National Science Foundation). We actually brought people from them onto the governing board, and used their manpower as program managers to execute some of these programs, because we ended up with hundreds of awards. So that was the big impact of other transactions.

Now, some of the other agencies, when they operated on our behalf, used cooperative agreements rather than other transactions, because that's what they thought their legal authority was. Actually a lot of these agencies had OT authority, if they’d actually read their enabling statutes. They didn't know what the language meant. 

You're talking about DOE: DOE had other transactions authority in its enabling statute in 1977, and yet Congress had to force it down DOE’s throat in 2005. Even then, they implemented it in the most backward way.

That's something I wanted to ask you about. It seems that agencies tend not to use OTA to the extent that they could by statute, and you've written about this. Help me understand why an agency wouldn’t use all the tools available to it. 

At NASA, they didn't let the procurement office touch space act agreements. At DARPA, my idea was, we were going to get the Contracts Management Office involved. OTA was going to become an integral part of how we operated.

Other agencies, instead of saying, “this is really special, it gives us lots of flexibility and authority,” what they say is, “this is just another tool, another contracting tool. We'll either give it to the procurement people or to the assistance people.” And these people are trained as rule followers. Part 1 of the FAR says the first duty of a contracting officer is to make sure that all laws, regulations, executive orders, policies are followed. That's what they're trained to do, that's what their mindset is. And it's absolutely inimical to executing the flexibility of other transactions. 

At ARPA, after building up experience, we eventually got the Contracts Management Office involved. They were all experienced contracting officers.

Half of them had the time of their lives doing OTs. Half of them were utterly incapable of operating in a flexible environment like that. 

Because it ran afoul of everything they'd learned?

Right. And I think that's happened in many agencies. We had a guy from the National Science Foundation call us, and he really wanted to get into this. He got pushback from his employees, rather than the employee saying, “Yes, sir, boss, we'll do what you want.” They were saying, “What’s this all about? Why are we doing this?” The Federal Aviation Administration was the same way. 

OTA is defined in the negative: it's not a loan and it's not a grant. How capacious is that in practice? What can OTA actually do? 

“As may be necessary in the conduct of its work and on such terms as it may deem appropriate.” As long as it doesn't require people to commit murder or fraud or something that's clearly illegal. The laws relating to contracting per se don't apply, whether it's sales of forest products or purchasing goods and services. They don't apply to transactions themselves, as long as they pass a basic test of, “Are we supporting the mission of the agency?”

I’m going to get a little esoteric, but it’s actually a fundamental principle. This goes back to an 1831 Supreme Court case, and in 1831, the Supreme Court ruled that an agency did not have to have specific statutory authority to enter into a contractual relationship. As long as the agency existed and it had a mission, the agency itself was a corporate entity.

Corporate entities have the power to enter into contracts. It isn't that you have to look for authority to do this, it is authority unless it's been bounded. 

I think I know what you're going to say, but does that authority apply to any appropriations? Can any appropriations that are not specifically set aside for a particular loan form or grant be used in an OTA?

As the DOD OTA guide says, the choice of award instrument and the choice of appropriation are separate.

Got it. You describe training as one of the big reasons procurement officers are wary of using OTA.

How would you go about changing that training? 

I wouldn't even try to use procurement officers. I was part of a task force of the defense science board that put out a report in 2009, “Fulfillment of Urgent Operational Needs.” That report says speed is counter-cultural to the normal acquisition process. If you want to do something fast, you want to do something different? You need a different organization, you need different ways to get different people. And you can create a different organization. 

But instead, DOD has just doubled down on DAU [Defense Acquisition University, which publishes guides on OTAs, among other things] and they just haven't tried to do this at all. You don't need a contracting officer to sign. The statute says the Director of DARPA or the Secretary of Defense has the authority, and they can delegate to whomever they want. But in the DOD, they just flow it down to contracting. And for those people, it's not just a matter of education. They have to unlearn things.

What's the best argument against OTA that critics have?

They have all kinds of arguments, but the best argument I think was made by Frank Kendall in an article for Forbes when he was between offices, and he says the workforce isn't trained to use these. He also said they're not very well trained in using procurement contracts, both of which are true, but they're certainly not trained for using OTs.

That statement is ironic, however, considering that he had just been the Undersecretary of Defense for Acquisition and Sustainment, and it was his job to see that he had a trained workforce.

I spoke to president of DAU Jim Woolsey in 2018 and I made the statement, “An OT agreements officer does not necessarily have to be a contracting officer.” He jumped back in shock. “You mean you don't need a contracting office?” “Yeah. Not only do you not need one, it's probably counterproductive to have one.”

As you look around at different government projects or aims of different agencies, which ones stand out to you as places that would benefit from more use of OTA? 

I generally don't make that judgment. I don't want to make that judgment. I want to unleash people who are called program managers or program directors, who have a mission and have a budget to accomplish it. 

I want them to know what other transactions are, and that their flexibility is almost unlimited. I want the ideas to come from that kind of a person. 

This program that I told you about earlier where Congress added some money to DARPA's budget, we called it the Technology Reinvestment Project (TRP). And that's when we brought in people from other agencies. It showed that agencies can collaborate, and government missions can be accomplished with a mix of government and outsourced funds. We actually did this with the Commercial Operations and Support Savings Initiative.

This began at DARPA as a follow-on to the TRP. When we started, people said, “Sustainment is eating our lunch. What do we do about that? We need to focus on saving money on sustainment.” So we ran a program that, rather than targeting this platform or that platform, saved money in sustainment. 

The program was the government provided some R&D money, some non-recurring engineering money for a firm that thought that it had a product that could substitute for a government product. That product was then put into a kit that could plug into the government system in its form, fit, and functionality. It could perform its function at least as well as the legacy part in the existing system, and it cost less. If you could do at least as well and cost less, the government would commit to buying your substitute product.

We ran that program out of DARPA for two years. We executed a hundred million dollars of R&D funds. Over ten years, it saved three billion dollars in procurement. It did this successfully despite the fact that back then we did not have follow-on production authority for OTs, and the production had to be procured under Part 12, Commercial Items Acquisition.

Will you describe that briefly? 

Part 12 is commercial acquisition. Fewer rules apply. If you can show me that you're selling this product on the market competitively, we're not going to take your intellectual property. We're not going to require accounting data, etc. And traditionalists hate this. 

Why? 

Because it doesn't follow all the usual rules. Auditors don't get to do all the things that the auditors do. And we don't, we have less paperwork involved. Not all of them hate it: there's many of them that just don't know how to do it, they only know how to follow one set of rules. 

I did some reading and you've criticized the most recent DOD guide to OTA usage. Can you flesh out for our readers what's wrong with it?

My biggest objection is that section where I point out they have stuck requirements back into the guide, whereas the previous guide said OTs are about problem solving: “What problem do we have? What are the potential solutions?” 

In the paradigm of the procurement world, the requirement pre-exists. Somebody has a requirement, and they hand the requirement to the acquisition system, which then does what it does. In OTs, we're saying, that paradigm doesn't exist. We could send a DARPA guy out to an army base and actually talk to people who were in the fight. And they might say, “My car has this feature, but I don’t have it in my airplane. Why can’t I have it?”

In one of our podcasts, we had Raj Shah on, the second director of Defense Innovation Unit (DIU), talking about his time as an F-16 pilot, flying along the Iraq-Iranian border. He didn't have a moving map display in his aircraft: if he accidentally crossed the border he could be shot down. It would be an international incident. So he bought a commercial one and strapped it to his thigh.

I'm not a big advocate of just doing the same thing over and over again, but you at least need to know what happened in the past. Open your mind up to say, “My problem today is this. If they could do that back then, why can't I do this today?” That's the kind of thinking I would like to see in government agencies and what I've been trying to push, and nobody really wants to listen to me. 

If you were the President's advisor today, aiming to encourage more risk-taking and using all available tools, where would you start? 

I would make it mandatory for the government agencies to follow all of the FAR, specifically FAR 35. 002, Part 35 is the R&D section of the Federal Acquisition Regulation. The primary purpose of contracted R&D programs is “to advance scientific and technical knowledge and apply that knowledge to the extent necessary to achieve agency and national goals.” 

FAR 35. 003, contract, meaning procurement: “Contracts shall be used only when the principal purpose is the acquisition of supplies or services for the direct benefit or use of the federal government.” Acquisition of supplies and services. All R&D should be conducted outside the federal acquisition process in the FAR.

Apart from that, many agencies have legal authority beyond just R&D, prototyping and follow-on production for DOD. DOE has open-ended authority. ARPA-E's authority is even broader than their basic enabling statute.

Are there levers that Congress can use to force agencies into better procurement practices?

They've already created a policy for this. The U.S. Code, 10 USC 4811, talks about the civil-military integration policy already. It says we should move away from relying on companies that depend on us. We want our defense contractors to be diversified companies.

How do you do that? You don't do it through traditional procurement. So there's a policy that's not being implemented by senior leadership in the DOD now, but what can Congress do? It can say, “of these R&D funds, not less than X amount is available only for the conduct of a program similar to the commercial operations and support savings initiatives.” It can also mandate that the Department of Defense comply with FAR Parts 35 when conducting development and prototyping programs. 

If you had to make a short list of the most obvious OTA successes to flog to policymakers, what's on that list?

Telstar 1 basically created a whole new industry in telecom, satellite telecommunications. Then there's Global Hawk. This was the first OT agreement I talked about, with Gazelle Microcircuits for gallium arsenide. Gazelle got bought out by TriQuint — eventually Corvo became the recipient of that technology. Now basically every smartphone and base station has a gallium arsenide component.

That trail was blazed with the very first OT back in April 1990. Sure, there's still a lot of silicon out there, but DOD is still finding ways to use gallium arsenide and gallium nitride for special purpose stuff, but OTs were the contracting vehicle that led the way to actually get that knowledge into you.

So original OT agreement: Global Hawk, NASA's first satellite launch of a private satellite NASA's Falcon 9 with SpaceX, an entirely new launch capability [here’s the launch services agreement]. We used to just have a couple of big companies. The defense space alliance or whatever it's called, and now here we have private money coming in for an entirely new launch capability.