What Happens When Space Projects Go Over Budget? The Curious Case of MSL’s Overrun

In a previous blog post, “Recovery. Peak. Collapse. Planetary Science from 1990-2014,” I included a plot of NASA’s planetary science budget. It showed that, between FY 2003 and FY 2005, the Planetary Science Division received its highest levels of funding historically, and those years were at the top of a period of strong funding extending from FY 2000 to FY 2010. Some readers questioned whether the peak was in fact a result of the Mars Science Laboratory, now known as Curiosity, going over budget.

The answer is no, the peak had nothing to do with covering Curiosity’s budget woes. In fact, we can attribute the peak to two factors. The first is that NASA was funding a robust portfolio of planetary missions, resulting in the high number of spacecraft we now have studying the solar system, while also pursuing strong research and technology development programs at facilities here on Earth.

The second reason is that NASA, between 2001 and 2006, moved the funds for the Deep Space Network from the Space Operations Mission Directorate budget line to the Planetary Science Division budget line. In 2007, DSN went back to Space Operations. The DSN program cost roughly $290M a year in 2013 dollars, which temporarily inflated the planetary science bottom line.

But if MSL’s budget woes did not contribute to the 2003-2005 peak, what were the effects of Curiosity’s cost overruns?

Development of the Mars Science Laboratory

In 2003, the National Research Council recommended in its Decadal Survey for planetary science, New Frontiers in the Solar System: An Integrated Exploration Strategy, that NASA construct a Mars Science Laboratory. This moderate-cost rover with an undetermined science package would serve primarily as a “technology-demonstration precursor” to the planned Mars Sample Return mission.

As a medium-cost mission, MSL would have been in the $410–$820 million range (all figures here are adjusted to 2013 dollars). By comparison, the Mars Exploration Rover missions (Spirit and Opportunity) cost a combined total of roughly $800 million to launch, with an additional $40 million a year in approximate operating costs.

In 2007, NASA leadership confirmed MSL with a baseline development cost of $1.1 billion. The mission was no longer a mid-cost solar-powered technology demonstration for a future mission; it had become a flagship mission. NASA released the official confirmation notice to Congress—along with baseline cost and schedule commitments and development cost projections for the project—in the FY 2008 budget request.

The budget problems for MSL began early. Budget issues in the Mars Exploration Program forced deferment of some work on MSL to FY 2008. The Jet Propulsion Laboratory requested additional funds in 2007 and 2008 to meet the 2009 launch date. The Critical Design Review in June revealed additional cost overruns. Materials NASA planned to use on the heat shield didn’t pass tests, resulting in more costs for redesign efforts. Mission designers planned to use dry, titanium-based actuators (motors that move the wheels, robotic arm, and camera mast) that would allow MSL to operate in a wider range of conditions on Mars, but the technology didn’t pan out, requiring more funds for replacement and redesign. Many other issues plagued the MSL team, and the 2009 launch window kept getting closer.

In late 2008, with a nearly $220 million shortfall looming and no margin left in the schedule, NASA Administrator Mike Griffin, Associate Administrator of the Science Mission Directorate Ed Weiler, Director of Planetary Science Jim Green, and Director of the Mars Exploration Program Doug McCuistion announced that MSL would not launch in 2009. A two-year delay would add an additional $430 million in “standing army” costs to the price tag just to keep all the necessary people on the team for an additional 26 months. Added to previous overruns, this brought the total cost of the mission to more than $2.4 billion. But launching in 2011 would allow the additional costs to be spread over several years, and gave the team plenty of time to correct mistakes or re-work and re-test systems they had rushed in an effort to meet the 2009 launch window.

Section 103 of the NASA Authorization Act of 2005 requires Congressional approval for continuation of any project exceeding its baseline development costs by 30 percent or its schedule by six months.  NASA leadership sought this approval in FY 2010 through a re-baseline for the project, which means that NASA changed the commitments it made to Congress and Congress agreed to the change. So technically, it is inaccurate to compare MSL’s cost and schedule performance to the 2007 baseline numbers, since NASA and Congress changed the agreement. The project performed well against the new baseline.

However, it is fair to say that the cost overruns affected NASA’s plans for the Mars Exploration Program, and for the Planetary Science Division. The question is, what were those effects? It’s difficult to pinpoint exact impacts, since priorities among NASA’s programs and projects often change, and for many reasons. NASA postpones new starts or shifts money from one program to another frequently for perfectly legitimate reasons that have nothing to do with cost overruns. That said, let’s take a look at some budget numbers and see what they suggest.

Did Congress Increase NASA’s Budget to Pay for MSL Overruns?

First, let’s compare NASA’s original spending plans for MSL development to actual expenditures:

The figure above clearly shows the cost overruns in 2008 leading up to the decision to delay MSL’s launch, and then the extra costs from 2010 to 2012 associated with the new baseline. So where did all the “extra” money come from?

Next, we should see if Congress provided the Planetary Science Division any additional funds:

The chart above shows that in 2008 and 2011, NASA’s Planetary Science Division received less funding than requested. In 2010, PSD received just under $20 million more than requested, and in 2012 nearly $13 million more. However, in 2010, MSL ran nearly $200 million over the original cost plan, and in 2012 it was $136 million over.

And while funding levels for the Planetary Science Division were higher in 2011 and 2012, they were just $90 million and $140 million higher than the 2010 budget, respectively. The MSL costs for those years were $191 million and $136 million over projections. So, even though Congress was supplying some additional funds, they still did not cover all of the MSL costs. This means the “extra” money came from within NASA’s budget. But as we will see, determining precisely where the money came from is extremely difficult.

Did MSL Impact Other Planetary Science Programs?

When Administrator Griffin first announced the MSL launch delay, he said that any additional funding for MSL would come initially out of the Mars Exploration Program budget, but might eventually affect other planetary science programs. The plan was to avoid cuts in any existing programs, and instead shift resources to MSL by delaying the start of new Mars projects, such as a Mars Sample Return mission or the Mars Atmosphere and Volatile Evolution (MAVEN) mission.

Interestingly, if we look at the Planetary Science Division’s spending projections in the FY 2008 NASA budget (the year NASA confirmed MSL) and compare them to actual expenditures, we see a startling shift in priorities for the Planetary Science Research Program:

These numbers shouldn’t be taken entirely at face value, as they only indicate what NASA leadership was expecting for the program in 2008, with assumptions about funding levels, other project and program needs, and instructions from the White House and Congress that may have changed in following years. Some of the cuts can probably be explained by the later addition of the Lunar Quest budget line and an increase in the Technology budget. Additionally, the Planetary Science Research program doesn’t remain static, as some research is completed and new research projects begun, while other projects may be shifted to another budget line (such as the Cassini extended mission, which NASA moved to the new Outer Planets program line in FY 2009).

Even so, when we look at NASA’s requests for each fiscal year from 2008 to 2012 compared to actual expenditures, the shift in prioritization is evident, particularly as MSL engineers were rushing to complete the spacecraft before the 2009 launch date:

Again, we cannot attribute the reductions in the Planetary Science Research budget directly to MSL. In times of budget shortfalls for aerospace programs, leaders typically reduce funding for research and technology development efforts to spare flight projects from catastrophic cuts or cancellation. We can say that MSL overruns were a significant pressure on the planetary science budget, and likely contributed to NASA leadership’s decisions regarding annual funding allocations within the Planetary Science Division, but we cannot say that NASA took dollars from Planetary Science Research and gave them to MSL.

Did MSL Affect New Project Starts?

Administrator Griffin stated that if possible, he wanted the cost overruns on MSL to be paid out of the Mars Exploration Program budget line by delaying new project starts. At the time, the Mars Exploration Program ran a series of competitively selected small missions called Mars Scouts. Based on the successful Discovery program, the Mars Scout program began in 2001, launching the Mars Phoenix lander in 2007. The program office released an Announcement of Opportunity (AO) soliciting mission proposals for the second Mars Scout mission in 2006, with a planned launch in 2011.

In December, 2007 (two months before MSL’s overruns became a threat to other planetary projects), NASA announced that it was delaying the Mars Scout launch until 2013 because a member of the selection board in charge of choosing the next mission had a previously unrecognized conflict of interest. The board would have to start the selection process over with new membership, and the anticipated four-month delay did not leave the proposed projects enough margin to build and launch their spacecraft by the 2011 launch window.

The following September, NASA selected MAVEN as the second Mars Scout mission, which would launch in 2013. Thus, the timing of all these events coincides with the cost overruns on MSL, but in this case the delay of MAVEN’s launch seems to have had nothing to do with MSL’s issues.

But what about the other flight programs in the Planetary Science Division, specifically Discovery and New Frontiers?

When NASA began the Discovery program in 1992, it planned to launch a small-class project every 12-24 months. The medium-class New Frontiers program, started in 2002, would launch a mission every 24-36 months. NASA included a schedule for future Announcements of Opportunity(AO) and mission selection dates in the FY 2005 budget request.

As we see from the graph, NASA missed the target release date of the Discovery 11 AO by a year, and the mission selection (GRAIL) by another six months. More interesting in the context of MSL, though, is the span between GRAIL’s selection in late 2007 and the release of the Discovery 12 AO a year and a half later. Again, we cannot directly attribute the delay of the Discovery 12 AO release (and the resulting delay of the InSight project start) to MSL overruns, but the timing of the two events coincides.

The fact that the Discovery 13 AO came out just last month, more than two years after NASA selected InSight as the Discovery 12 mission, indicates that the Planetary Science Division is still operating under tremendous budgetary pressure.

The story for New Frontiers is similar. The AO release and mission selection for New Frontiers 2 (Juno) occurred on schedule, but the New Frontiers 3 AO release missed the target date by nearly two and a half years. NASA schedules two years between the AO and mission selection on New Frontiers missions, which is roughly how long it took for the selection of OSIRIS Rex as the NF-3 pick, so the downselect was not delayed beyond the gap resulting from the AO release. The gap between selecting Juno and the AO release for NF-3 spanned from June 2005 to February 2009, which effectively delayed the OSIRIS Rex mission start by nearly two and a half years. We can’t say this gap was a direct result of MSL, but it does align with the MSL overruns.

And as further evidence of the current budget pressures on the Planetary Science Division, NASA has not announced any plans to release an AO for a New Frontiers 4 mission.

Did MSL Result in Any Project Cancellations?

In 2009, NASA and the European Space Agency signed the Mars Exploration Joint Initiative, which led to plans for two joint Mars missions collectively known as ExoMars. The first would have been an orbiting spacecraft called the Trace Gas Orbiter (TGO) set to launch in 2016, followed by a 2018 dual-rover mission featuring a U.S.-built robot called MAX-C and a European rover. NASA agreed to provide the launch vehicles for both missions.

In early 2011, NASA said it could no longer afford the rocket to launch the Mars 2016 mission, so the Europeans purchased a Proton rocket from the Russian space agency, Roscosmos. NASA further re-scoped its commitment in late 2011, negotiating with ESA to reduce the payload on the 2018 mission from two rovers to one, MAX-C, while also reducing NASA’s—and increasing ESA’s—proposed overall budgetary commitment. Finally, and in spite of ESA’s increased commitment, NASA’s FY 2013 budget request eliminated all NASA funds for both missions.

MAX-C would have been the first of three missions recommended as the highest priority flagships for planetary science in the National Research Council’s 2012 Decadal Survey, Visions and Voyages for Planetary Science in the Decade 2013-2023, part of the long-anticipated Mars Sample Return program. When White House officials rejected the mission on the grounds that they did not want to commit to three concurrent flagship missions, regardless of the timeline, they did not cite MSL overruns (or the James Webb Space Telescope overruns) for their reluctance to approve a new flagship, but it’s hard to imagine they didn’t factor into the decision.

So, What’s the Verdict?

Funding NASA development and operations is an incredibly complex enterprise with many moving parts at any given time.Trying to determine direct impacts of one project’s overruns on other projects and programs borders on futile. Clearly, cost overruns have consequences, but since they usually affect projects that have not yet started, effects are very difficult to measure. I have tried here to illustrate some of the feasible impacts of the cost overruns on MSL, but this just isolates MSL as one of many factors.

In later posts, I will delve a bit deeper into why projects exceed cost estimates and look at possible systemic causes for some of the rising costs (or inaccurate estimates). In the meantime, I think the largest impact of cost overruns on NASA projects is to the confidence in the agency’s ability to complete missions at cost and on schedule. While federal and public confidence in NASA’s technical capabilities remains high, the days of regularly exceeding space project budgets and continuing to receive adequate funding for a balanced portfolio of small, medium, and large space missions are likely gone for good.

Fortunately, NASA has improved its record of remaining within baselines on planetary projects in recent years, MSL notwithstanding. But the larger the project, the greater potential impact an overrun will have on other projects and programs. And the events of MSL only add to the difficulty in advocating for future flagship missions.