This week some Pentagon officials morphed into street cleaners as the Defense Department's F-35 "Joint Strike Fighter" left yet another load of unpleasantness on the street for all to see. It came in the form of major new revelations from Jason Sherman at InsideDefense.com with an article titled "DOD Warns Congress JSF Costs Could Skyrocket To $388 Billion." The new, higher cost estimate intensified the sticker shock for the already unaffordable F-35. The word went out from the "E" ring of the Pentagon; reporters and others - including myself - were told it was all "shaky math," "garbage," "totally wrong."
It was also directly from a DOD report, sent to Congress, obtained by InsideDefense.com and other reporters, and released to subscribers at the InsideDefense.com website.
The "Selected Acquisition Report (SAR) F-35 As of December 31, 2009" repeated some earlier information about F-35 costs, but it also dropped a new load of unwelcome new data. Previously, Congress had been told that the average price for each F-35 would be $79 million to $95 million, with primary emphasis on the lower figure.
But the F-35 SAR report made painfully obvious two devices used to understate the cost.
First, as some aware observers - none of them present at recent congressional hearings - already knew, the $79 to $95 million band of unit cost did not include all expenses, specifically, about $50 billion in research, development, test, and evaluation (RDT&E) expenditures.
Second, that band of cost was stated in "base year" dollars. "Base year" dollars are normalized to an early date in the program's history and, therefore, understate the amount of dollars that actually must be appropriated. "Then Year" dollars are the ones you want to know. They are the amounts that Congress actually appropriates.
The Pentagon likes both "base year" dollars and cost estimates that ignore huge portions of program outlays; they help to get the camel's nose under the tent with low ball price estimates.
Careful reading of the SAR report (for example, on pp. 36) reveals that including the RDT&E expenses yields a total program unit cost of $97.1 million for each F-35. But, it's still in those worse than useless "base year" dollars, specifically fiscal year 2002 dollars. (If you can find BMW car dealership that sells cars in 2002 dollars, buy one; it won't cost you more than a Volkswagen.)
The F-35 SAR reveals that $97.1 million in base year dollars translates into $133.6 million in actual future appropriations ("then year" dollars).
But, there's more in a well buried passage, on page 37. There, we are told that a new "complete Independent Cost Estimate is in process. The Department expects this analysis will result in increases to the stated [unit cost] estimates. The projected range of estimates are $97 - $115 million ... in Base Year 2002 dollars."
So it looks like the $97 million, or rather the more honest $134 million, unit cost estimate is about to be overtaken by events. We've got the new, higher estimate ($115 million), but it is only in those worse than useless "base year" 2002 dollars. Nowhere do we find in the report the more straightforward "then year" dollar cost for the $115 million. Luckily, however, eight grade math and ever-helpful inside the Pentagon sources both provide the same answer: $158 million per aircraft. So, it's not $79 million per aircraft; it's not $134 million; it's $158 million. That's twice what the Pentagon was pretending last month.
Those same sources, and the same math, enable us to convert the old - soon to be over taken by events - cost for the entire program to what insiders in the Pentagon now expect: $388 billion, or what Sherman reported in his April 6 article. It was this figure that caused the major E ring eruption. However, on April 8, officials in the Pentagon admitted to Sherman, and the public, he was basically right - "in the ballpark."
Even more unfortunate for the after-parade Pentagon street cleaners, there's a lot more in the F-35 SAR report.
As Sherman also reported in another article on April 7 (See "DOD: JSF Combat Radius Shrinks, Logistic Footprint Grows As Design Matures"), the F-35 is beginning to go south on some of its performance specs. Specifically, the projected performance for range, payload, logistics requirements, and sortie rate of the various F-35 models are beginning to deteriorate from the originally stated "baseline" performance estimates and program objectives. It is a process that will continue as the original F-35 performance promises meet reality in the form of flight testing, now only 3 percent complete.
Finally, pages 25 and 28 of the SAR report show what Pentagon mythmakers are pretending they will pay for future F-35 production. For the next Pentagon budget proposal for fiscal year 2012, the unit cost - counting only production, not RDT&E, costs - appears painfully reasonable: the Navy will produce 21 copies of its F-35 versions for $5.1 billion. That calculates to $243 million each. That is what Navy F-35s are actually costing these days.
But the so-called "learning curve" promises to make Navy F-35s much more affordable.
For 2013, the Navy plans to pay $188 million for each of its F-35s.
In 2014, it plans to pay $173 million.
In 2015, it will be $156 million
$118 million in 2016.
Tables on Air Force production show the same trends.
Nirvana! The learning curve permits the advocates to pretend that the average unit price can be lowered to rescue the program from even higher costs than those now projected: As we get deeper into production, optimization of production processes will result in cheaper and cheaper aircraft. That's what happens in mass production, they argue.
There is only one problem: modern combat aircraft are not Chevrolets. For them, the learning curve barely exists. As Franklin "Chuck" Spinney has shown us for aircraft like the F-14, F-15, F-18, Apache Helicopter, V-22 Tilt Rotor, B-1 bomber, and even the less complex F-16 and relatively simple A-10 in the 1980s and 1990s, the projected "learning curves" were always vastly overstated. In fact, Spinney's analysis, which was based on official data in the Pentagon's budget planning and execution documents, showed that the Pentagon could not even use learning curves to predict the costs of runway cleaners and pickup trucks bought on the commercial market.
Once you look at the details, this denial of the prevailing conventional wisdom becomes perfectly obvious. A major reason is that there never exists a stable design to mass produce. Engineering change proposals, upgrades in the form of new production blocks, product improvements, and new requirements from the user never end. Modern tactical aircraft procurement programs never really allow a design to stabilize to enable "mass production" or any meaningful optimization of production and of the resulting cost. For this and several other reasons, the learning curve is mostly illusion.
In the case of the F-22, the closest design and fabrication cousin to the F-35, the learning curve actually went backwards late in production; the unit costs went up. That is by no means new -- the same thing happened with the F-14 and the F-15. For the late models of the "mass produced" F-16, the block 50 model, unit costs were about twice the amount of early production models.
When the non-learning curve starts to become obvious in the F-35, the latest total program cost, $388 billion, will be a thing of the past, a long gone whisper of what never could have been.
Costing more and doing less is the standard profile for failing Pentagon programs. The F-35 is becoming today's poster for what in the 1980s Washington Post cartoonist Herblock depicted as a $600 toilet seat hanging around the neck of then Secretary of Defense Casper Weinberger. The modern day cartoonist for the Post, Tom Toles, would be entirely correct to give us a depiction of new millstone for Secretary of Defense Robert Gates, the F-35.
Perhaps, if we are lucky at some future date, when still more F-35 cost growth is first extracted from a user-unfriendly DOD document by an enterprising journalist, the E ring of the Pentagon will not lunge for invective, but will appreciate that it is being told something it didn't know and that others in the building didn't want them to know.
Alas, perhaps someday.