The Space Shuttle Was Not Great, Part II: Cost
This is part two of a three-part series by Daniel Koenig.
Continued from Part I: Failed Goals.
By the time of the Challenger accident in 1986, just five years after the first launch of the space shuttle, it was clear that the program would never be able to achieve its goals. Challenger showed that the shuttle could not deliver safe and routine access to space, and the loss of the seven crew-members on a satellite delivery mission demonstrated just how disastrous it was to send a crew on missions which could be done by unmanned rockets.
One of the program goals not discussed in depth in the last article was that of reducing the cost of access to space. The difficulty and cost associated with getting to orbit is mostly a result of the speed which something must reach in order to enter a stable orbit around the earth – more than 7 kilometers per second. To reach this speed, every pound of payload requires 20 to 40 pounds of rocket underneath it – almost all of it fuel. Rocket stages and their engines are typically used only once and then discarded as they stage away from the vehicle on ascent. This contributes to the high cost – imagine how expensive an airplane ticket would be if the airplane was tossed into the Atlantic Ocean after every flight. The space shuttle was supposed to reduce the cost of access to space by making it routine, and by reusing several parts of the vehicle. The orange external tank was the only part not recovered and reused. But this partial reuse had the opposite of the intended effects, as is common with government programs, and the cost of the shuttle was more expensive, not less, than other vehicles with similar capabilities. And not just a little more expensive either.
The space shuttle was outrageously expensive.
In 1969, at a conference dedicated to the proposed space shuttle, NASA Deputy Director Robert Mueller said,
The goal we have set for ourselves is the reduction of the present costs of operating in space from the current figure of $1,000 a pound for a payload delivered in orbit by the Saturn V, down to a level of somewhere between $20 and $50 a pound. By so doing we can open up a whole new era of space exploration. [source]
Studies by NASA in the 1970s estimated that the low operational costs of the proposed shuttle could bring the cost of space access down to as low as $118 per pound within a few years.
A 1972 NASA review stated that the space shuttle could produce,
[a launch cost savings] of 50% of the historical baseline. The cost of developing payloads (the major portion of space shuttle program costs) can be reduced primarily as a result of the shuttle’s ability to retrieve payloads for reuse, refurbishment, and updating. [source]
The cost savings were intended to come from several areas – the reusability of the shuttle meant that a new launch vehicle didn’t have to be built for each launch, the fast turnaround time meant that the manpower necessary for each launch would remain low, and the ability of the shuttle to retrieve, repair, and redeploy satellites already in orbit would reduce the need for expensive new satellites.
But, like the estimates of flight rate, NASA’s cost estimates were disconnected from reality. In 2003, Shuttle Director Robert F. Thomas explained that when the shuttle was being developed,
the people in Washington got a company called Mathematica to come in and do an analysis of operating costs. Mathematica discovered that the more you flew, the cheaper it got per flight. Fabulous. So they added as many flights as they could… [source]
Despite knowing that the shuttle would not be able to fly as frequently as was needed to keep costs low, NASA published the numbers anyway. Thompson added that the cost estimates of operating the shuttle didn’t include costs associated with but not directly related to operations.
So we didn’t try to throw the cost of ownership into that. It would have made it look much bigger. So that’s where those very low cost-per-flight numbers came from. They were never real. [source]
Even as the shuttle was being developed, it was known that NASA was lying about the savings. Brian O’Leary wrote in a 1973 article (a full 8 years before the first shuttle flight) criticizing the program,
Even when compared with current boosters the shuttle fares poorly. Taking $8.1 billion as the initial development cost and assuming NASA’s model of 514 shuttle flights during the 1980s, an independent scientist, Ralph Lapp… obtained a launch cost of at least $5,250 per pound of payload. This is more than five times the rate of Titan III-C and other present launches and 70 times the rate claimed for the shuttle by NASA. [source]
Shuttle critic Gregg Easterbrook wrote, one year before the first shuttle launch,
“The shuttle will be able to carry three Delta-class payloads,” says Chet Lee, the shuttle pricing director. That means, for instance, three communications satellites and the extra boosters needed to push them to high orbits. To launch three Delta-class payloads on Deltas would cost three times $23 million–$69 million.
Suppose the shuttles fly 500 sorties, as predicted, and cost $13 billion to build. That works out to an investment cost of $26 million per flight. Add that to the $40 million “true cost” of a launch. Suddenly a shuttle launch costs $66 million just about the same as three Deltas. Now, suppose the shuttles fly only the pessimistic 200 flights. The investment cost leaps to $65 million per flight. Suddenly the total cost of a shuttle flight becomes $105 million–almost twice the cost of three of those wasteful Delta rockets. [source]
His estimates of the cost of the shuttle program turned out to be too optimistic by a factor of more than 10. The space shuttle program cost $200 billion with a yield of 135 flights, which works out to a staggeringly massive $1.5 billion per launch to do what could be done on traditional rockets for $70 million.
NASA expected a large market for satellite repair and redeployment, but there was no basis for that expectation. Most of the satellites targeted for repair and redeployment were communications satellites operating out of reach of the space shuttle. Easterbrook wrote,
More than two thirds of the satellites being launched are sent out to geosynchronous orbits – 22,000 miles up, where, relative to a spot on the rotating earth, they hang in the same place all the time. Others are sent to nearly-as-high sun-synchronous orbits, where they follow the movements of the sun.
The shuttle, on the other hand, orbits at 200 miles. The highest it could reach, unloaded, would be 600 miles. Routinely, the shuttle will fly to 200 miles and release a satellite mounted on the LUS booster, or a smaller booster called a SUS (for Spinning Upper Stage). The booster then blasts the satellite out to its final destination. Once there, it is beyond the shuttle’s reach. Period. When something is 22,000 miles away, getting 200 miles closer isn’t much of a help. [source]
The few satellites operating at altitudes low enough to be reached by the shuttle were not worth repairing.
A communications satellite has absolutely no salvage value… They almost never fail, and by the time they wear out, after seven to ten years, they’re obsolete. It’s much cheaper to build a new one. [source]
Satellites, like most computer equipment, are subject to the pressure of advancing technology. Repair is often more costly than replacement, and the equation skews hard with age. In fact, beside the Hubble repairs, the shuttle only repaired three satellites over the course of its lifetime, worth no more than $200 million each when they were brand new. In every case, launching a new satellite would have been more cost effective than repair.
Joseph McGolrick of NASA’s Launch Vehicle Office said,
The users that were contacted indicated no interest in [repair]. Usually, what you were talking about was a satellite that was at the end of its life or was partway through its life, and they really didn’t want it back. It was, effectively, garbage. [source]
And what about capturing a dead satellite and bringing it back to earth for repair?
Refurbishment on the ground was even less promising: you’re bringing back junk and relaunching, and you’ve got an extra launch in there to be paid for. [source]
Realizing that their early estimates were wildly off-base, NASA suggested instead that satellite operators could perform low-orbit tests of their satellites before sending them to the unreachable geosynchronous altitude. The idea was that perhaps if problems were discovered on a brand new satellite while it was still in low orbit, repair would be possible and worth doing. But that idea was just as unsuccessful.
We asked the communications satellite people if they expected to check their payloads out in low earth orbit. And the answer came back that they would not anticipate doing an extensive test of the satellites, if for no other reason than that would require deploying solar arrays and then retracting them and putting them back together again. They felt that the benefit from that was outweighed by the additional risks that they would go through in going through that additional deployment and retraction. [source]
The one success that shuttle-advocates hold on to is the Hubble Space Telescope, which was delivered to orbit and subsequently serviced five times by shuttle crews.
But this success doesn’t hold up to scrutiny – most of the cost of the telescope was due to the truly gargantuan shuttle overhead. Hubble was originally scheduled to launch in 1986, but the launch date was postponed when all shuttles were grounded following the Challenger accident. Storing the telescope for the next 4 years cost NASA $6 million per month – adding up to several hundred million dollars – money that would not have been spent were it not for the shuttle.
After Hubble was launched, it was discovered that a flaw in the primary mirror severely diminished the quality of the images it was returning. NASA sent the first servicing mission in 1993, by which time they could have, for a similar cost, built an entirely new telescope complete with technological upgrades made available in the years since the original telescope was designed. That repair was followed up with 4 more servicing missions, each requiring a full shuttle launch.
Later in the shuttle program, the International Space Station was picked up as the justification du jour for the shuttle program. The ISS required a few dozen shuttle flights to assemble, and proponents argue that the shuttle was necessary to build the station.
But the ISS is only one of the 10 manned space stations that have been put into orbit, and it is the only one that was constructed using the space shuttle. That reliance on the shuttle isn’t a feature of stations in general, rather it’s like the old saying goes: when all you have is hammer, every problem looks like a nail. It was built with the shuttle because the shuttle is what was available.
The Soviet station Mir was assembled by self-docking modules using a robotic manipulator arm to place modules correctly. Mir was a third of the mass of the ISS, and nothing about the way it was assembled would have prevented an ISS-sized station from being assembled using similar methods. The assembly probably would have been much cheaper if done using traditional rockets rather than the space shuttle – the overhead of the massive orbiter and large crew to place modules weighing a few tons at a time was a colossal waste of resources.
In fact, there are several additional modules scheduled to be added to the station in the coming years, all of them to be delivered by traditional rockets.
Shuttle proponents point to the products created by the shuttle program, products developed as part of the shuttle program and later repurposed for other uses, as a defense of the costs of the program. NASA’s Spinoff series claims,
The Space Shuttle Program alone has generated more than 100 technology spinoffs. [source]
Spinoff names automotive insulation, infrared cameras, and jeweler’s equipment, among their successes.
This entire line of reasoning is completely absurd when the actual numbers are considered – a $200 billion program producing 100 technological products puts the cost of each advance at $2 billion. To produce what? Here is one example of a $2 billion technological advance, to quote from Spinoff,
Responding to a request from the orthopedic appliance industry, NASA recommended that the foam insulation used to protect the shuttle’s external tank replace the heavy, fragile plaster used to produce master molds for prosthetics. [source]
And another example,
NASA engineers also discovered that adding small tabs to the back of a truck smoothed air flow around the vehicle, reducing drag—a technology made commercially available. [source]
That’s $4 billion between those technological marvels.
Wouldn’t it be cheaper and faster to develop the commercial products directly without the massive overhead of the world’s most expensive space program?
Not only was the shuttle incredibly expensive, but it was also incredibly dangerous. The unusual layout of the vehicle as well as the requirements of the program led to the deaths of 14 people – more than all other space vehicles combined.
Continue to Part III: Safety.