How Not to Buy Weapons

Defense acquisition programs are FUBAR’d. I don’t think that’s a big surprise to anyone here. Heck, when your programs are so bad they get the Star Wars treatment, you’re in the hurt locker:

[scribd id=65528746 key=key-15o79c12e2b47i0lvqrl mode=list]

Sure, the .pdf is tongue in cheek. But it does raise some good questions about what we’re doing when it comes to major weapon system programs. What is the last major weapon system you can think of that came in on time and under budget? I’ve been following acquisition programs since the early 80s, and I can’t think of a single one. Oh, sure, several were ultimately quite successful. The Army’s “Big Five” programs of the 1970s (M1 Abrams, M2/3 Bradley, UH-60 Blackhawk, AH-64 Apache and Patriot missile system) were all undoubtedly successful programs, but even they ran over budget, and in fact, each program was the successor to a program that had been cancelled for cost overruns and technical overreach. The Navy’s very successful Spruance/Ticonderoga/Burke class ships were “austere” versions of ships that had been developed for over 20 years. And the Navy’s F-14 and the Air Force F-15 fighters were both born out of the disastrous F-111 fighter program. Current procurement disasters include the Navy’s LPD-17 class and LCS ship programs, the Army’s aborted Future Combat System, and the Air Force/Navy/Marine Corps F-35 Joint Strike Fighter.

So why is it that buying weapons is so hard? First, weapons are incredibly complex. From a mechanical and even more from a software viewpoint, a weapon system such as the F-35 is a nightmare just to design. On the software front, if Bill Gates screws up and your Windows crashes, you lose the fantasy football stats you were trying to save. When Lockheed Martin screws up, a $200 million fighter gets turned into a smoking hole in the ground. Debugging the software before market release takes time, and that means money.

Even worse, the process for simply defining what a weapon system should do is a nightmare. It used to be, when the Army decided it needed a new tank, it looked at what the last batch could do, then padded the numbers a bit, and asked the contractors to build a new one. When the Navy decided it needed a new ship, it got together the people that used ships and asked them what they’d like. Same thing with the Air Force. The Marines were screwed. Nobody asked them what they wanted. They got whatever the Navy and the Army were buying. All in all, it worked pretty well.

One of the great joys of Robert McNamara’s revolution in the Department of Defense was his “operational analysis” applied to roles and missions. Instead of just letting the Navy buy a ship to do a mission that the Navy had always done, OA insisted that the mission had to be examined to see if one of the other services could better perform that role. This added a layer of complexity to buying weapons. It also had the knock on effect of adding to the complexity of the weapons themselves. You couldn’t risk losing a fight over whether a plane was the right tool for a job. You had to make that plane so outstandingly capable that no other possible platform could even hold a candle to it. That drove up complexity, strung out development time, and in turn drove up costs. Rising costs leads to the “procurement death spiral.” When the overall cost of a system rises, the end numbers of units bought gets cut. That in turn drives up the unit cost, which exerts great pressure to lower the cost of the program, usually by reducing the total number bought. Which drives up the unit cost…

Then there’s the fact that we’ve made developing weapon systems as profitable for defense contractors as producing them. It used to be that the design and development team of a contractor was a cost center, an investment made in hopes of receiving a production contract. Now, the cost and risk of having a union production team means actually producing a system is often the cost center, and the juicy development contracts pay to support large engineering staffs. We’ve perversely incentivized contractors to keep weapon systems in development as long as possible. Now, they aren’t dragging their feet on purpose, but as long as they get development contracts, they aren’t about to complain, either.

Finally, politics gets to play a role as well. Lots of money gets spent, and Congress controls the purse strings. That means many a Congressman wants to see defense contracts coming to his district. Whenever you hear a congressman mention the importance of maintaining our industrial base, it is a good bet they’ve go a contractor in their district that they think is critical. Maintaining the industrial base IS important. But when Congress exerts influence to pick and choose winners and losers, the results are about what you’d expect.

Almost all major defense programs are competitively bidded. Now, you’d think that would lead to the best, most cost effective system being chosen. You’d be wrong. The ground rules of the competition aren’t so simple as a head to head comparison of capabilities. Development costs, production costs, life-cycle maintenance costs and a myriad of other similar factors play a role in the competitions. Unfortunately, history has shown that the estimates of these future costs is mostly wishcasting. There just really is no way to accurately predict out-year costs for programs with life cycles of upwards of 50 years. Nonetheless, Congress, both directly and indirectly, has a huge influence of the terms of the competition, and can use these and hundreds of other parameters to shade the rules so as to often almost predetermine the outcome for a favored client. Or worse, they end up setting up a competition that is so rife with loopholes that even when the service has a clear preference for one platform over another, the losing bidder can sue to demand a recompetition, delaying the development and procurement of any platforms. This happened so many times in the Air Force’s program for a new Combat Search and Rescue helicopter that eventually the Air Force just gave up. The Air Force’s next generation tanker program took over a decade and three complete competitions to decide on the Boeing KC-46 program. Dude, it’s a tanker. We’ve been buying tankers since the B-29 was still in service, and they haven’t gotten much more technically complex since then. Take an airliner, add a boom, add some extra gas tanks, you’re done. It ain’t rocket science.

So what are some successful procurement programs, and what do they have in common?

Two programs spring to mind. First, the MRAP, or Mine Resistant Ambush Protected vehicle program. As the insurgency in Iraq heated up, and the low levels of protection on early Humvees became apparent, the Army (and to some extent, the Marines) were in dire need of vehicles that were less vulnerable to IEDs than early Humvees. Up-armored Humvees were quickly produced and fielded, but there was limited room from growth on the basic vehicle. Building on designs pioneered in South Africa, several companies offered began building large purpose built trucks that were far more capable of withstanding IED attacks, and gave good protection from small arms fire. Indeed, the main complaint in the press was that the Army wasn’t buying enough vehicles fast enough (never mind the shortages of armor quality steel and lack of production facilities and workers to build them any faster).

The other program which comes to mind was the Army’s Stryker family of wheeled personnel carriers. The eight-wheeled Strykers were intended as an interim vehicle to provide units with some level of armor protection, but to be more easily transported and less expensive than fully armored units such as tank and Bradley equipped brigades. The Stryker was basically an off the shelf Canadian design with the Army’s digital battlefield internet system grafted on. The concept was that the Strykers would serve short term until the definitive Future Combat System family of vehicles were available.

In the event, the Stryker program was actually a huge procurement program, and yet it was budgeted with very little controversy or public scrutiny. While the vehicle is hardly perfect, it was a very successful procurement program, mostly on time and on budget. The “definitive” Future Combat System vehicles, on the other hand, never even made it off the drawing board. The program was so bloated and behind schedule, costly, and so far from delivering any capability that it was scrapped.

Why were the MRAP and Stryker programs successful? They both were narrowly focused programs, designed to fit a very specific need. Further, they were evolutionary designs, building upon mature technologies. Neither program was “transformational” or “cutting edge” technically. The Stryker program did introduce new levels of digital integration to units, but the Army had been developing such systems for 15 years, and the systems are now common in all combat vehicle fleets.

Lessons learned?

  1. Set realistic program goals. Not every program can be a 30 year leap into the future. That’s not to say research and development aren’t important. But trying to develop critical technologies concurrent with the development of the weapon system programs that will use them is foolish. Inevitably, there are delays and difficulties in developing the underlying technologies, which delays the parent program. Changes in the developing technologies also means that baseline configurations can’t be set, leading to constant changes in the parent design. This stretches out the development timeline, increasing costs.
  2. Settle for 80%. The MRAP, Stryker, M1, Ticonderoga/Burke, and F-14/F-15 programs all accepted that they were getting most of what the developers wanted, but not all. By settling for 80% of the capabilities they desired, the developers were able to freeze the designs, get them into production, and still achieve a great net increase in capability. The last 20% of desired capability ends up being gold-plating, and an outsized cost driver. Further, if that capability is later determined to be critical, it can usually be retrofitted. One example is the Commander’s Independent Thermal Viewer on the M1A2 series tanks. It was omitted from earlier production for cost reasons, but provisions were made in the original design to fit one if later needed. Similarly, Vertical Launch Systems for Tomahawk missiles were added to later production Virginia class submarines, because space and weight was reserved for them (the same thing was done with the Los Angeles class subs twenty years earlier).
  3. Be ruthless in controlling costs. The easiest way to do that is to be ruthless in controlling the growth of a program. When the services propose a system, they have to get buy-in from various factions within that service. It takes a forceful leader to ensure the minimal amount of bloat occurs. The sooner in the definition phase a program manager can freeze the design, the quicker development and fielding of a platform. That alone works to keep costs down more than almost any other factor.
  4. Leverage existing technologies. Virtually every successful procurement program was based on existing technologies. The Ticonderoga cruisers used the hull and machinery developed for the Spruance class destroyers, and mated them with the Aegis weapons system that had been in development (in various guises) for almost 30 years. The F-14 program used the AWG-9/Phoenix missile system developed during the TFX/F-111 program as well as the TF30 engines.

We hear a lot about greedy defense contractors. And to be sure, they haven’t covered themselves in glory. But the fact remains, they only sell what the DoD is buying. And they only develop what DoD pays for.

The acquisition framework is so complicated that simply defining what a system should do takes years. This stretched out procurement cycle drives costs up, which is ironic, since the stated purpose of the entire DoD acquisition architecture is to ensure that money is wisely spent. It is a classic case of paralysis by analysis. That’s not to say strong oversight and justification isn’t needed in procurement. But we need to strike a balance between oversight and actually developing weapons.

Finally, procurement in the last generation has been far too willing to accept unreasonable technological risk instead of reasonable risk to troops. It’s impossible to fully remove risks to our troops, and attempts to do so by overreaching for technical solutions instead leads to troops using obsolescent existing weapons. The prime example of this would be the Marines and the MV-22. How many Marines died in CH-46 crashes waiting more than 20 years for the MV-22 to be developed, when a “good enough” replacement could have been bought in as little as a quarter of the time?

We owe it to the taxpayer and the troops to get our acquisition strategy under control.

6 thoughts on “How Not to Buy Weapons”

  1. Much of the problem can be boiled down to “Revolution vs Evolution.” The best programs have been evolutionary rather than revolutionary.

    In the case of fighter aircraft, we used to have competitions which produced “Y” aircraft. The last I know about is the YF-16 vs YF-17. The USAF bought the 16, but the 17 got warmed over to the F/A-18 of Lex’s warm memories.

    As I recall, the F-14 later was re-engined with the same engine used in the F-16.

  2. Fielding the M1 with the available 105mm main gun, while designing the tank and building it with the intent of putting the 120mm in later is another example of building flexibility into the program. Separate but related is the willingness or unwillingness to buy foreign weapons. While our 120mm is US -built, and has a different breech, it is a German design, fielded in their Leos.

  3. Well, the Army has gotten over it’s don’t by foreign attitude. MRAP is South African, Stryker is Swiss through Canada, the M252 mortar, M119 howitzer and M777 howitzer are British systems and both the M249 and M240 series are Belgian. We may be license building them but they are not home grown. My own program I am working in is based on SAP software a German company.

    1. It was too bad that didn’t obtain in the late 50s – early 60s. I would rather have seen us adopt the FN FAL or the H&K G3 instead of the M-14 then the M-16. The M-16 was OK after they got cleaning kits to everyone (they issued the piece without supplying cleaning kits to all and sundry. Ordnance Corps hated the thing and refused to support in the early going). Kids were literally sending cleaning rods home to get more made because the Army didn’t have nearly enough to go around.

  4. It was hard to go against the grain in the late 1950s with the M-14…it was the result of the John Garand school and it was hard to argue against him. Also we were not too much into buy Euro at the time for a lot of reasons…soem good, some bad. The M-16 had its issues but a lot have been overblown and it turned out to be a fine rifle which fit our doctrine. And I personnaly found it more difficult to shoot an FN/SLR than an M-16.

    All of that said we try to work with the NATO allies with weapons development…the MBT-70 comes to mind first. It may have failed but it lead to the M1. There were also several aircraft which were jointly developed. Also mso tof the 1/4 tons used in USAREUR through the 1950s into the mid-1960s were made in French factories…a result of the Marshall Plan.

    Finally, there was a real issue of the amount the Communist parties were involved with the European labor unions and we did not want weapon systems tied up in teh possible problems those could pose.

  5. How soon we forget.

    Gatling guns and electric battleships. And the Jeep, my God the Jeep, and its successor (?) the HMMWV. Defense procurement has always been a mess, and it always will be. MacNamara’s Band of the Best and Brightest just put the icing on the cake.

Comments are closed.