October 20, 2016
On 11 October, the US Air Force received prospective contractors’ proposals its its Ground-Based Strategic Deterrence (GBSD) program, its effort to begin replacing its Boeing LGM-30 Minuteman III intercontinental ballistic missiles (ICBMs) by 2030. The USAF is sufficiently concerned about its choices for cost-effective replacement that the service’s request for proposals restricted its contractors from exclusive dealing. At least one of two subcontractors thus lost an opportunity for a one-time payout from a prime contractor, and it’s possible that both did. The idea is certainly at odds with the Total Package Procurement concept of the 1960s, and the equally unlamented Total System Program Responsibility concept of the 1990s. In contrast, using the government’s buying power in this way in the 2010s makes for quite a smart acquisition strategy.

To begin, some background. The current ICBM force consists of 440 Minuteman III missiles, each now with a single warhead of (probably) 300 kilotons. The missiles are spread amongst 450 silos in three fields, in Montana, Wyoming, and North Dakota. By February 2018, a further 40 missiles will be taken out of service, to comply with warhead limits in the New START Treaty. (Stew Magnuson offered a good summary of the specifics in the August issue of National Defense magazine.)  Even with those 40 additional spares, the missile force will run out of missiles for annual testing a few years after replacement is due to start. At that point, the comfortable feeling about the 20th Air Force’s inventory starts to degrade. Indeed, General Robin Rand of Global Strike Command said at September’s Air Force Association conference that the reliability of those missiles is already failing. Steven Koonin, a physicist who has been advising the Energy Department on maintaining its warheads, rather agrees. In last weekend's Wall Street Journal, he offered a long essay on whether America can trust its aging nuclear arsenal even now.

Not everyone is equally enthusiastic about replacement, of course. There are some usual suspects. In War On The Rocks, Geoff Wilson and Noah Williams of the Ploughshares Fund recently argued that land-based ballistic missiles are obsolete. On the other hand, as Peter Huessy of the Mitchell Institute recently wrote for Real Clear Defense—citing authorities from our very own Brent Scowcroft to the disarmament group Global Zero—eliminating land-based missiles hugely simplifies an enemy’s targeting problem. The location of each land-based American missile is precisely known by potential enemies, but destroying any single missile probably requires expending a matching nuclear warhead. Destroying the force thus starts a massive nuclear war with the United States, a prospect no opponent takes lightly. Perhaps for this reason, as Byron Callan of Capital Alpha recently wrote to investors, “there appears to be no debate in the DoD or Congress about ICBM modernization.” So let us suppose that this program is proceeding.

Full replacement calls for at least 22 missiles for developmental test and evaluation, followed by 642 operational weapons. Presumably 400 will go directly into silos, and the remaining 242 will be used as spares and ongoing test articles for decades. Three companies have bid: Boeing, Lockheed Martin, and Northrop Grumman. None of these firms today makes solid rockets, and liquid-fueled ICBMs have long been obsolete. Those Musk and Bezos guys, building liquid-fueled launchers at SpaceX and Blue Origin, are totally consumed with getting to Mars, and were probably never much interested in building nuclear missiles anyway. So it was to be one of the three, just with another company’s motor.

However, as prospective subcontractors, there are just two domestic solid rocket motor manufacturers: Aerojet Rocketdyne and Orbital ATK. The USAF probably never considered soliciting input from Airbus Safran Launchers, which builds the solid rocket motors for its own M51 ballistic missiles, used in the France’s Triomphant-class submarines. Like France, Britain builds its own nuclear warheads. Unlike France, however, Britain has long gotten its ballistic missiles in the United States, so why the United States couldn’t in theory buy French is an interesting question. It’s just a whole other talk show.

What’s more notable is how the USAF's solicitation barred any of the prospective primes from exclusive dealing with either of the prospective rocket motor suppliers. That is, none of the missile-makers could demand that either of the rocket motor builders work only with one firm. Why would the government ever tolerate exclusive dealing? As occasional rocket failures remind us, going exoatmospheric remains tricky business. An exclusive, long-term commitment ensures that management at both a rocket motor manufacturer and a missile maker will be keep their attention on a single solution. On the other hand, exclusivity can then preclude better or cheaper approaches, particularly should a technically challenging program bog down in development.

The USAF’s solicitation seemingly did not require that prospective primes remain open to either rocket motor manufacturer, but at least one has. As Vivienne Machi reported for National Defense, Lockheed Martin announced its team when it submitted its proposal, and the lineup notably included both Aerojet and Orbital as alternative suppliers. Lockheed program manager John Karas explained that decisions about rockets will be made later. For reasons I will provide below, if Lockheed is open to either now without a mandate, it’s probable that Lockheed would not have dealt exclusively if the solicitation had so allowed. Doing so would have cost the company considerably. In contrast, as Aaron Mehta observed for Defense News, Boeing and Northrop Grumman are saying nothing so far. They may have foreclosed some options, but they’re not saying so just yet.

On the Air Force’s side, a similar strategy seems to be at work in the JSTARS recapitalization program. Radar-makers Raytheon and Northrop Grumman have recently each received sole-source contracts to build candidate radars for the new ground-surveillance plane. For the actual aircraft, three prospective prime contractors have a variety of solutions: Boeing with its own 737, Lockheed Martin with a Bombardier, and (again) Northrop Grumman with a General Dynamics’ Gulfstream. By contracting separately for the radars, the USAF has effectively precluded any vertical restraint by Northrop. That’s an economist’s technical term for a company's refusing to sell its own upstream product to a downstream competitor who wants it. The difference between this and exclusive dealing basically lies in whether the supply chain is vertically integrated. The economic effect, though, is broadly similar.

The key concern with GBSD is probably cost, and not quality. The USAF’s first nuclear-armed ballistic missile, the liquid-fueled SM-65 Atlas, was fielded in 1959, so the question likely isn’t whether this can be done, but for how much. While buying new land-based ballistic missiles will be less expensive than buying new sea-based ballistic missiles (which includes their Columbia-class nuclear-powered submarines), or air-breathing B-21 bombers, the program will still be expensive. As Sydney Freedberg wrote for Breaking Defense, the USAF thinks that the full development and procurement program will cost about $62 billion, or $93 million per missile, fully loaded. The Pentagon’s Cost Assessment and Program Evaluation (CAPE) office thinks that this will all cost at least $85 billion, which is more like $128 million per missile, and maybe a lot more.

Why the stark difference? The Air Force Department based its estimate partly on the cost of the development and production of the Minuteman III missile, which entered service in 1970. CAPE used information from the Navy’s program for UGM-133 Trident II missiles, which entered service in 1990. For some reason, the Air Force never got the information it needed from the Navy. Fairly, though, it’s unclear whether accounts from 26 years ago are remarkably more useful analogs than accounts from 46 years ago. In any case, after the first nine months of a “technology maturation and risk reduction” (TMRR) effort with the chosen contractor, the Air Force will update its cost estimate with the necessarily more detailed information then available. For all this angst, it is thus probable that the service has restricted exclusive dealing to increase the likelihood that the low-cost missile maker can team with the low-cost rocket maker, without another missile-maker’s economic intervention.

How might that have happened? Consider the counterfactual: if the USAF had not precluded exclusivity, would a defense contractor have sought such a deal? Lockheed has already signaled that it would not have, but Boeing or Northrop might have. Take the field somewhat abstractly. Assuming that each prospective prime plans a single missile design, and each prospective rocket subcontractor plans a single rocket motor design, there would be six possibilities. For convenience, I’ll represent the primes with single-letter abbreviations, and the subcontractors with two-letter abbreviations. The winner of the competition will come from the following set:

{B + AR, B + OA, L + AR, L + OA, N + AR, N + OA}

Not all of these combinations would necessarily have been offered, but no competitor could presume that any one of them would not be offered. Then, if one of the three prospective prime contractors perceived that one of the two prospective subcontractors had a clear cost advantage, it might offer good terms to secure that cheaper motor on its team alone. If that prime conservatively assumed no similar movement by other prospective primes, it might have taken that money out of its expected profits. If it gambled, it might have increased its bid, effectively transferring that value from the government to the subcontractor.

In either case, the government’s set of possibilities would shrink, as securing any one exclusive deal would cut the number of possibilities from six to three. To see how, consider how Rocketdyne—less Aerojet, with which it was merged in 2013—was actually owned by Boeing from 1996 through 2005. Had Boeing retained that company, and had the government wanted to prevent vertical restraint, the Air Force may have needed to pursue separate contractors, as with the JSTARS recapitalization, and with possible knock-on integration problems. Suppose now, though, that the two independent companies, Boeing and Aerojet Rocketdyne, were allowed to pursue an exclusive arrangement, because each saw something in the other that it liked. The possibilities for contracting teams available to the government would reduce to

{B + AR, L + OA, N + OA}

So much for the incentives of the primes. Without inducement, the subcontractors would be less likely to agree, because ceteris paribus, exclusivity reduces from one-half to one-third the likelihood that either will win a spot on the program. As such, because there are only two rocket options, those rocket-makers have greater market power than the missile-makers, even though they do not control the relationship with the customer. However, that lowered probability of a win could be traded for some other consideration, perhaps an investment by the prime in its design or tooling, that larger share of the profits, or just a cash payment.

Any further deal for exclusivity would reduce the number of possible bids to two, completely removing one of the prospective primes. If, for example, Northrop Grumman reached such an accord with Orbital ATK, after Boeing did a deal with Rocketdyne, Lockheed Martin would be completely out:

{B + AR, N + OA}

Note that this confers an unusual second-mover advantage in the rocket motor business. The first solid rocket firm to enter into an exclusive arrangement could have extracted some value from its chosen prospective prime, but the second firm to move extracts more. Whichever remaining missile-maker doesn’t win its affections would have been out of the competition, unless its management wanted to try to enter the solid rocket motor business at that scale. The USAF would not likely be impressed by de novo upstream vertical integration with a program so important. That alone would have driven up the bidding for the final deal on the dance card. In the process, even more value could have been effectively ex ante transferred from the government to the two subcontractors. The rocket-builder on the winning team would earn much more, of course, but both would extract some value.

And for avoiding that outcome, Air Force Acquisition deserves a little acclaim.

James Hasík is a senior fellow at the Brent Scowcroft Center on International Security.