The Cracked Crystal Ball: The Problem with Predicting Defense Jobs and Sequestration

As the threat of sequestration has grown closer, the level of noise surrounding the looming budget cuts has only grown louder. Last week, the volume reached a new level with the issue of a report prepared on the projected economic costs. 

The report was prepared on behalf of the Aerospace Industries Association (AIA) by Stephen Fuller, a professor at George Mason University, and Chmura Economics & Analytics, a consulting company. It painted a particularly dismal picture of the jobs impact of the defense portion of the potential cuts.  “A total of 1,090,359 jobs with a total labor income of $46.5 billion would be lost due to DOD budget cuts in FY 2012-FY 2013.” It then broke down the figures on a state by state basis, noting that such key election battleground state like Virginia and Florida would experience severe job losses (136,191 and 41,905) at a time of already worrisome unemployment.

The report landed like a bombshell. A range of newspapers, including the Los Angeles Times and Washington Post, directly quoted the report’s jobs findings, while cable and online networks such as CNN and Fox News also devoted significant coverage to the mind-boggling numbers.  Similarly, policy leaders on both left and right have cited it, such as both New Hampshire Senators Kelly Ayotte (R) and Jeanne Shaheen (D). And in testimony before the House Armed Services Committee, timed the day following the report’s release, senior defense industry executives echoed the report’s apocalyptic numbers, which furthered its spread and assumption as fact.

There is only one problem: none of the people citing the numbers have looked under the hood. Such amazingly specific predictions are actually based on very crude and sometime erroneous assumptions and calculations. To be very kind, while they may make for good headlines, they are not useful for actual policymaking.

#FunnyNumbers:  The Multiple Flaws of Sequestration Projections on Defense Jobs

Setting aside the absurd specificity of the numbers used in such projections of the future (1,090,359? Really? A report projecting into the future across the entire economy couldn’t round off even to ten?), the first flawed notion is the disconnect between the scale of the industry and the scale of the potential loss.

While many still have a vision of Eisenhower’s defense industrial complex, several recent studies by experts in the field have explained this is not the case in the 21st century. As Nate Fick, CEO of the Center for New American Security, and Tom Davis, Vice President of Strategic Planning at General Dynamics, explain in a 2011 report, “In 1950, 1 in 3 jobs in the U.S. economy involved manufacturing. Now that number is 1 in 10, with the national security industrial sector accounting for only 15 percent of total manufacturing jobs. In other words, only about 1 out of every 70 workers in the United States is now involved in aerospace and defense.”

Indeed, just four months ago, Deloitte issued a report on the financial and economic impact of the industry, notably sponsored by the very same organization, the AIA,that sponsored the jobs report. That report found that there were approximately 3.53 million jobs – direct, indirect and induced – sustained by the aerospace and defense industry. This is for the entire workforce, not just the part sustained by the Pentagon budget (such as workers making passenger jets, etc). So if 1.09 million direct, indirect, and induced jobs were lost from the 10 percent cut of sequestration, that would mean nearly a third of the overall jobs sustained by the industry would evaporate, an extreme projection to say the least. 

These kinds of projections grow equally questionable when you look at them from the bottom up, rather than top down. For instance, in recent testimony to Congress, Bob Stevens, the CEO of Lockheed, estimated that if sequestration occurred, his company might have to lay off as many as 10,000 workers. Stevens’ estimate was a far more sensible projection, as that is approximately 8.3% of his workforce, not a third, far more in line with the scale of the potential cuts.

Thus, when you start to add up the numbers that would be required to scale the losses at Lockheed to the overall industry, things also fall apart for the projection. Of the top 100 defense firms last year, Lockheed is ranked at the top, doing just over $33 billion in contracts, or 8.9 percent of the total funds awarded by the DoD. The slide is steep in size, with number 100, Groupe CGI, at $404 million and 0.11 percent of total contracts. Of course, beneath these top 100, there are a host of smaller firms. But if the biggest company in the industry is losing around 10,000 jobs, and it dwarfs the next firms literally by orders of magnitude, you have a hard time reaching the projected numbers, even with an assumed ripple effect.

It’s equally problematic if one deconstructs the jobs projection not from above or below, but from another angle, with a look at the supposed ripple effect job losses that make up much of the supposed numbers. The report found that for every $1 million cut to DoD procurement (from the $45.1 billion), roughly 24 jobs would be lost (7 direct workers, and then ripple effects of 6 indirectly employed, and nearly 11 induced jobs). But such numbers are well beyond what we’ve seen play out in reality or frankly even in the claims made by defense firms in their other lobbying efforts.

Let’s use the KC-X air tanker as an example, as it is one of the most important defense jobs programs of the last generation. During the tanker competition, for instance, both Boeing and Northrop Grumman/EADS promised to create approximately 48,000 jobs for a contract worth $35 billion. This meant rather than a ratio of 24 jobs per $1 million, they claimed 1.37 jobs would be created for every $1 million spent.  Of note here, the numbers that the firms used are actually almost double what the Department of Commerce uses (whose formula comes out to 0.71 per $1 million, finding that the $35 billion contract would create 25,000 jobs directly and indirectly).

There’s no question that job creation and elimination take place at different rates, but such a discrepancy is pretty staggering. It could be solved if there were exact data, but that is the very point: There is no data, and whatever the prediction of the would-be ripple effects depends greatly on the assumptions built into the model. And this is not even to go into all the issues that arise when trying to project into second and third order ripple effects into a dynamic economy. Such analysis treats each job created or lost as some isolated chain of events. But the job prospects of that teacher or steelworker who draws from the defense worker’s economic impact is also shaped by all sorts of other forces. That teacher could lose their job because of the cutback in a defense firm in their town. But that teacher could equally lose their job because of tradeoff budget cutbacks in other parts in the non-defense budget. Or they could keep their job because the budget axe fell in defense and not on the Education Department. That steelworker could lose their job because of lesser demand for steel to build ships. But they could also keep it because of the impact that stabilizing the debt may or may not have on the broader economy, bond rates, investment climate, etc. To project any exact number for these multiple degrees of separation from the original spending implies a level of certainty that just doesn’t exist in the real world.

A related problem of treating the jobs process in isolation is how such projections also ignore ongoing changes in the defense industry itself. The analysis directly equates increased Pentagon spending to increased defense jobs, which seems reasonable enough. Except go back to the state of the overall industry. The amount of spending on defense has gone up since 9-11, but the number of defense jobs has continued to shrink. The reason for this, which these formulations just ignore, is the impact on jobs of massively increased productivity as well as changes in the structure of the industry.. For instance, the 1960s-era F-4 fighter jet was built with zero parts using direct digital manufacturing (also known as “3-D printing,” where the part is digitally designed and made, shrinking the assembly line process, and thus reducing the number of workers involved). The 1980s F-18 has some 90 parts made this way. The current F-35 has over 900 parts made this way. This is a huge gain for the process, but it comes with a loser: the number of workers involved. Similarly, industry is going through a wave of rationalization, with all sorts of division cuts and potential mergers and acquisitions that would have happened regardless of sequestration.

There is yet another problem with such incredibly specific jobs projections about the future: how incredibly dependent they are on another current uncertainty: what slice of the budget is cut? For instance, the report assumes 82% of the reduction would be placed on procurement and R&D.  But this is yet to be decided. Top experts in the field don’t, for example, yet know whether the cuts would or wouldn’t fall on the personnel side, whether Pentagon civilians and or those uniform, which could have a hugely variable impact. Moreover, it ignores common practice. In the real world, policymakers have often turned to more fungible pots like construction or O&M to buffer cuts. Finally, the certainty of a number like 82% assumes that the world stays static. In a post-sequestration political environment, there might be changes in anything from personnel costs structures to Tricare and other benefits systems, as all sorts of things not on the table might be pushed into play.

But where the certainty of such reports should be most questioned is actually in the part that makes the local news headlines and creates the greatest alarm in Congress, the state-by-state job loss projections. The idea that we can predict future job losses drilled down to their zip code, without knowing where the actual budget cuts fall, is deeply misleading. 

Such projections are based on the cuts falling in a generalized distribution of payroll and procurement spending. There’s only one problem with this model: cuts don’t actually fall in a generalized manner. Depending on where and how the cuts are applied – to personnel, programs, new contracts, etc. – each state and the factories and facilities within it will be affected in dramatically different ways.  If, for instance, the Pentagon slashes orders of the F-35, then California, Texas and Connecticut would take the hardest hits.  But if the Pentagon decides to scale back future shipbuilding, Virginia’s and Maine’s shipyards would bear the brunt of losses in jobs.  Were federal workers forced out in large numbers, states with the largest military and DOD footprints, such as Colorado, would shoulder a greater share of the losses.

The notion of a perfectly even distribution, from the 136,191 workers supposedly to be let go in Virginia to the exactitude of predicting that 65 unlucky workers would be fired in Delaware (not, 64, and not 66, but exactly 65) may make it seem real to those in Congress worried about votes in their home districts.  But it bears no actual link to reality. As Bob Stevens put it, “We don’t know with much precision yet which lines of business, which sites, which contracts, which programs, or which technologies would be affected.”

Here again, we see the good sense of the actual defense business leader working in the field. His focus on “lines of business,” “contracts,” and “programs,” is important, as it also points out the final flaw in these overly specified jobs projections. Economists using generalities ignore the interplay that actually goes on in the real world between defense firms, the systems they make, and the workers they employ. Cuts do not hit by the same exact percentage across all states and all programs, even if one were to enact them as a perfectly aligned cut (which the interplay between OMB and OSD wouldn’t let happen). A loss in one program can actually be a gain to another, not just in the immediate budget, but also to the life cycle of the system.

For example, one of the most likely ways the Pentagon would deal with the scale of cuts in sequestration would be to further shrink and delay Navy, Air Force, and Marine purchases of the new F-35 fighter jet, which is one of the biggest programs of record. This would be very bad news for the workers at Lockheed’s fighter plane plant in Fort Worth, Texas. It would mean fewer jets bought in the short term, which would further escalate the per unit costs of the F-35. This would then be more bad news for those workers in the long-term, as the Pentagon (and foreign allies on the export market) would likely buy even fewer of the now more expensive jets. That same bad news scenario for Texas, however, would actually be good news to the workers at Boeing’s facilities in places like St. Louis, Missouri. In this scenario, their lines for the F-18 and F-15 fighter jets would likely stay viable and active for a longer period of time as the US keeps them in service longer and foreign allies buy more of them as stopgaps (as Australia is already doing).

Similarly, sequestration would likely doom or at least delay many of the new ground systems that the Army wants like the Joint Light Tactical Vehicle and the Ground Combat Vehicle. But no real projection can weigh the locale or size of these potential job losses to the workers behind the JLTV or GCS with any kind of certitude, as these contracts have not been awarded yet. But the very same bad news for some number of future workers in some unidentified locales, who wouldn’t have their vehicles bought, would be celebrated by actual workers currently making parts for the older HUMVEEs and the Bradley Fighting Vehicles in places like South Bend, Indiana and York, Pennsylvania. A rising tide may lift all boats, but a sinking tide puts some people under water while leaving others on dry land.  


We do not mean this analysis to be an attack on the defense firms that supply the arsenal of democracy. Any cut to the defense budget would certainly reverberate onto these firms and workers that draw revenue from it. Jobs would be lost, which would hit companies, communities, and, most importantly, individual workers and their families in a tragic way. That is certain.

Rather, our point is to show what happens when you actually look a bit deeper at these kinds of reports that make such news. The notion of projecting specifically what would happen all the way down to the exact job numbers lost, and in the exact location, is deeply flawed, all the more so when based on uncertain or even erroneous assumptions. Serious people working on the serious issues of defense should not give them serious credence.

A good analysis (that would better serve the firms that sponsored the study) would instead engage the uncertainty. It would lay out the complex nature of the system; explore the multiple scenarios that might occur; and identify the key variables and assumptions that would have the most impact. Moreover, as all of these companies that supply the national defense would agree, it would not ignore the biggest most important elephant in the room. The budget should not be viewed as just a jobs program.  Indeed, if one actually only cared about job creation or loss, studies going back nearly 25 years have found that spending on defense is not the optimal way to spend federal discretionary dollars. But that, of course, is not how we should look at the defense spending in the first or even last place. We spend on national defense to protect our nation, something completely set aside in these kinds of analyses.

In closing, we want to be utterly clear here: We do not intend this deconstruction of the far too common practice of highly specified jobs projections as a defense of sequestration. We believe sequestration to be an un-strategic, largely ineffective manner to go about dealing with America’s debt dilemma. To allow sequestration to happen, rather than implementing a compromise that focuses on both entitlement and revenue reform, would be irresponsible.

But bad ideas should not be fought with bad analysis. Some may believe that hype and hysteria, even if based on false reports, are the only way to force action in the broken political climate of today. We simply note that hype, hysteria, and relying on dubious assumptions as if they were fact are exactly the kind of behavior that originally got us into this predicament.

Throwing misleading figures about as if they were certain truths muddies an important debate, both to the nation’s future and the future health of the defense industry. It undermines any sensible understanding of the stakes, radicalizes the involved parties, and makes the very needed compromise less likely in the end.

Cracked crystal balls should not be our guides to the future.