The Cost Estimator's Problem: Network Effects

This will provide a brief introduction to defense cost estimating and the broader economic problem of production networks that is not often addressed. 

The government, of course, is not a vertically integrated supplier of defense. It procures weapon and information systems almost exclusively from the private sector creating a relatively thin market. Government agents want to attain advanced capabilities from at a “fair” price. Thus, while expecting efficient and innovative production, they prefer suppliers to earn low and stable economic profits. Due to the scale and complexity of many defense acquisitions, government agents also prefer transparent program oversight because the direction of resources cannot often be precisely specified during contract negotiations.

The DOD’s problems are then two-fold: 1) monitoring performance on long and variable contracts; and 2) return value to the taxpayer by alleviating the asymmetric information problem for future procurement decisions. EVM reporting intends to address the former issue while CSDRs the latter. When considering the second case, it is clear that the government requires insight into both the intensity and mix of resources used to develop certain system functionality (i.e. WBS element). An important concern to the estimator, regardless of data quality or accessibility, is that future cost outcomes are contingent on the performing contractor’s current and future workflows. This is caused by joint production of multiple goods with large common costs, making it difficult to forecast the marginal cost of any one good because prices depend on production across all goods.

Industrial Base Considerations in a Defense System Cost Estimate

Since a firm will seek maximize efficiency at the level of the business unit, it is difficult to understand their net position by observing indirect rates for any given contract. Figure 1 below depicts some business base considerations in an estimate. The total facility workflow has been between theoretical thresholds represented by two horizontal dash lines. The top line represents the amount of workflow which necessitates a scaling up through major fixed investments. A workflow closer to, but not exceeding, this top line represent lower indirect rates because many of the same indirect costs are spread among more direct work. The bottom line represents the point which a downsizing of fixed investments is required, and workflows closer to this threshold represent higher rates as the same indirect costs are spread across less direct work. It would be difficult to say what “fair” indirect rates would be when observing Program A and historical analogies in a vacuum. Luckily, DCMA does much of this analysis and cost estimators often strip the historical indirect costs to apply their approved rates.

More on this later as there is much to unpack and extend.

F-35 and Risk Management

F-35A Assembly at Lockheed Ft. Worth

It has been no secret that the F-35 Joint Strike Fighter, the air-superiority aircraft intended replace numerous legacy counterparts, has had its fair share of difficulties. Acquisition costs have ballooned to $400bn and another $1tn (yes, trillion) in operating and support costs are expected. The Air Force even had to repaint their refueling trucks from green to white because the former color absorbed too much heat and the resulting fuel temperature was inoperable with the F-35B short-take off, vertical landing (STOVL) variant. A fascinating article from David Axe sheds light on what we are really buying through one of the government’s largest and most important programs. An unnamed test pilot wrote a scathing unclassified briefing after a dogfight test with an F-16D.

"The stealth fighter proved too sluggish to reliably defeat the F-16, even with the F-16 lugging extra fuel tanks. ‘Even with the limited F-16 target configuration, the F-35A remained at a distinct energy disadvantage for every engagement,’ the pilot reported.

‘Insufficient pitch rate.’ ‘Energy deficit to the bandit would increase over time.’ ‘The flying qualities in the blended region (20–26 degrees AoA) were not intuitive or favorable.’

The F-35 jockey tried to target the F-16 with the stealth jet’s 25-millimeter cannon, but the smaller F-16 easily dodged. ‘Instead of catching the bandit off-guard by rapidly pull aft to achieve lead, the nose rate was slow, allowing him to easily time his jink prior to a gun solution,’ the JSF pilot complained.

And when the pilot of the F-16 turned the tables on the F-35, maneuvering to put the stealth plane in his own gunsight, the JSF jockey found he couldn’t maneuver out of the way, owing to a ‘lack of nose rate.’

The F-35 pilot came right out and said it — if you’re flying a JSF, there’s no point in trying to get into a sustained, close turning battle with another fighter. ‘There were not compelling reasons to fight in this region.’ God help you if the enemy surprises you and you have no choice but to turn.

[….]In the end, the F-35 — the only new fighter jet that America and most of its allies are developing — is demonstrably inferior in a dogfight with the F-16, which the U.S. Air Force first acquired in the late 1970s.”

This process of testing  your system against handicapped opponents or unrealistic situations is rife. The Ballistics Missile Defense Organization (BMDO) would put a infrared beacon or radar reflector on the target missile in order to make the test intercept easier.

As Axe explained in another article, the F-35’s design was compromised to fulfill the needs of the Marine Corps who insisted on the STOVL. In order to get the lift-fan mechanism to work, it was decided to dump 11 pounds of worth of valves and fuses making the plane 25% more likely to get destroyed by enemy fire. The Chinese, who likely hacked and stole engineering data from Lockheed, have quickly built an F-35 replica without the complexity stemming from the STOVL requirements.

Here is my take on the subject, which comes down to risk management. Even if there is a slight chance of total ruin, the precautionary principle applies. Climate change is an often used case of this logic. Even if the chances that humans are causing climate change are miniscule, the potential effects are so great that the burden of proof must be on those who wish to do nothing. Nissim Talebhas applied the precautionary principle to GMOs because an unintended blight could destroy the entire ecosystem.

The precautionary principle can be applied in defense acquisitions as well. The F-35 has monopoly in its sphere and we are led to believe that there is at least a reasonable chance that it will be completely outclassed by competitors. If air superiority turns out to be a decisive factor in the next war, I would argue that diversification is a sound choice. 

Interestingly enough, enacting the precautionary principle usually comes at a cost of "higher returns" (e.g. emissions regulations dampen economic growth), but the case here is that we are paying more for accepting this risk because of the nature of massive acquisition programs (more on this discussed later). 

The strengths of the F-16 were born in a reaction to the complexities of the F-15, and has proved itself simple, effective, and enduring. The F-35 is also intended to replace the A-10, whose strengths come from features that stand in stark contrast to the F-16. 

I don’t see an easy resolution to the F-35 problem. The primary impediment to change is how insular the program is from the public eye (aided and abetted by poor media coverage).