Launch Hazards: Problems Getting Insurance
SHOWSTOPPER: 3rd Party Liability Insurance
Shortly after we became the lead agency, the large companies and said there was an insurance problem. Both NASA and DOD were telling the companies that as a requirement of using their launch range facilities, they had to have insurance. NASA and the Air Force were requiring the companies to purchase the “$500 million 3rd party liability insurance or the maximum amount of insurance available on the market at a commercially reasonable price”. 3rd party liability insurance covers the cost of damage to 3rd parties, e.g. innocent bystanders. It does not cover the cost of loss of satellites or launch vehicle.
The problem was that the first launch company seeking insurance was able to purchase about $500 million dollars for a launch. But the second company was not able to come close to this amount and the price of what little insurance was available was unaffordable.
Without going into the actual details of how the space insurance industry worked at the time, the problem could be summarized this way. 3rd party liability insurance was not covered by a pre-existing fund. It was covered by underwriters who were essentially gambling that there would be no 3rd party claim. If the rocket launched successfully with no outside fatalities, they pocketed the money they received from the premium. If the rocket failed, killed people, and the claims totaled more than the premium they collected, that remaining balance came out of their pocket. If one misunderstood the risks, it was possible to lose huge amounts of money underwriting any space launch activity, much less 3rd party liability insurance. For a couple years in a row, underwriters actually paid out for first party insurance losses (i.e. satellite owners), than they took in. Most underwriters were very leery of insuring private rocket launches because of the risk. Even if they were willing to assume the risk of 3rd party liability claims for one launch, they believed that they could not afford the risk of underwriting a second launch except for an exceptionally high premium.
As a result, the first launch company that purchased 3rd party liability insurance essentially bought up all the capacity available in the market. The next company trying to buy 3rd party liability insurance would find that there was no remaining “low price” insurance, and that underwriters would provide insurance only at an exceptionally high and unaffordable rate.
There were a number of factors at play. NASA and the US Air Force had done a pretty good job of advertising to the Nation that space was risky business. By requiring $500 million insurance, the government was telling the insurance industry and underwriters that the space business was extremely risk. In 1986 $500M was comparable to more than $1 billion today. Further, the way that NASA and USAF portrayed the risk, there was no way that insurers could determine whether, in the event of an accident, would the entire $500 million insurance amount be used? “Why was so much insurance required? “ the underwriters would ask the space community. The prevailing answer in that community (note: I am not ascribing this answer to either NASA or USAF because I do not know where this viewpoint originated.) the $500M would be what was needed for this scenario: a rocket goes off course and lands in a stadium holding a convention of trial lawyers. So the problem amounted to a risk adverse government culture, a misperception in the space community, and an insurance industry that relied on both for information. Because insurers only go by what NASA, DOD and the space community believed. insurers assumed it was so dangerous that it was uninsurable.
Thus, when the second company sought insurance, the insurance companies found very few underwriters willing to bid at any rate of return.
It Didn’t Make Sense
The idea that these launches were so risky that insurance should not be available seemed extremely odd to me. My experience was in the transportation world, and there were many modes of transportation, e.g. rail, large commercial aircraft, or tanker shipping vessels that seemingly posed much greater risks to the public than space launches. In commercial transportation we often do not keep the dangers away from the public, but instead have them interspersed with the public.
To illustrate, my office was on the 10th floor and Northeast corner of the Nassif Building. When I looked out the window I could see the Nation’s capitol. In addition, I looked over four railroad tracks part of the Northeast Corridor railroad system and every day I saw many trains pass by, many hauling large numbers of tanker cars, some filled with explosive and/or toxic materials. One day, I watched an entire train derail. The same thing happened the very next day. But DOT didn’t stop trains from moving through the city. Similarly, when planes crash on take off, they often crash into public areas. Only two years before we became lead agency, I had watched from the Deputy Transportation Secretary’s office the flashing lights of emergency and rescue vehicles on the 14th street bridge after Air Florida Flight 90 crashed into the bridge due to icing. The next day, airplanes were taking off from National Airport.
How could the 3rd party risks of a space launch be greater than what I had personally seen? Whereas, in transportation it was easy to deliver the risk in the midst of the public as illustrated above, the opposite was true with respect to launch vehicles. Rockets at the time were so strictly controlled, they could never reach the public.
The other question I had was how much explosive power did the launch vehicles have versus a commercial transportation vehicle. The General Dynamics rocket used RP-1, essentially the same as used by a 747. Which carried more fuel? I knew that rail and truck tankers routine carried highly explosive materials, such as fertilizer, gasoline, and propane and liquefied natural gas. These were hugely dangerous and yet they were often found in populated areas.
A little bit of quick research, and I knew how the answers would turn out. This was an area where we, as DOT as regulators, could show the space community how we could bring value added to the space program. The first area involved beginning to define the actual risks posed, not only by commercial space launches, but all dimensions of space transportation. This was important not only for setting insurance, but for all future aspects of commercial space transportation, including airborne launches, re-entry vehicles, new launch sites and spaceports, and even part of commercial human flight. (Note: again, while I mention human space flight, the goal was to occupy the space so that other agencies didn’t perceive a vacuum. To claim authority and the capability prevented other agencies from seeing a potential future opportunity. If Congress excluded OCST from regulating human space flight, the door is now again open. I’d wager if members of the House and Senate were polled, each agency would have proponents for handling commercial human space flight.)
From the very beginning of the space launch program, for their own reasons DOD and NASA had taken essentially an “absolute no risk based approach”. They chose launch sites away from populated areas, and essentially implemented a system that would ensure no rocket ever could threaten 3rd parties. They developed a sophisticated safety system and models that could allow them to protect the public from malfunctioning vehicles, and safeguards to keep their officers with the fingers on the destruct button from inadvertently destroying a good launch vehicle. Boiled down to the essence, their system was simple. Launch over the ocean and blow up rockets that get close to people. But their reason for doing this made sense. One mistake, and their budgets and public image are hurt. Congress and the public tend to be harsh on government agency mistakes.
That wouldn’t work for commercial space transportation—not for what I envisioned. This is a small illustration of commercial transportation hazards we accept every day:
- cars and gas tankers and trucks filled with hazardous materials to travel highways next to each other at 65 mph;
- LNG railroad tank cars traveling through cities with tens of thousands of people nearby;
- Aircraft flying over rush hour full bridges ¼ mile from the end of the runway;
- Natural gas pipelines located under densely populated city blocks,
- 82 people being killed each day in automobiles and trucks, etc.
There is absolutely no area of life, whether it is transportation, energy, medicine, etc., that Americans don’t experience and accept risk, whether voluntary or involuntary. For this industry to grow and advance, it meant that commercial space transportation was going to have to be treated in the same context. Commercial launch vehicles might one day launch from the US Midwest, space vehicles might fly over Miami, or everything needed to ensure public safety might be handled by the vehicle itself, not a human being. Unlike NASA or DOD, for DOT risk was going to be managed, not eliminated.
(At this point, someone is going to say, “But the Outer Space Treaties…. They make the US absolutely liable….” It’s true that there are these treaties that to space experts make this all different. Technically, that’s true. However, the “space experts” are addressing the concept of treaties. But let’s examine that. In real monetary terms from the US government perspective: the cost impact of those treaties are nil. In the most extreme case, say a rocket hits an airliner with 400 people on it resulting in $1 billion of suits (which by the way shouldn’t happen if FAA ATC does its job.) One billion dollars is what the US government routinely pays just in fraudulent Medicare claims every 6 days. DOD’s inspector general paid $720M in late fees on shipping containers, and GAO found that DOD had stockpiled $9.2 billion in excess parts. In 2017 the Trump administration is submitting a budget request of $29B to cover just part of the recent hurricane damage; that huge amount is simply to compensate for something the US experiences on an involuntary basis which it accepts without arguing. This is not a criticism of DOD or HHS nor a weather report, but instead this illustrates that United States government incurs “lost costs” on many things that it deems important. The “absolute liability” risks and potential monetary costs posed by the Outer Space Treaties are insignificant compared to the benefit the US derives by having Elon Musk, Jeff Bezos, or Richard Branson assume the cost risk of developing new space technologies. Thus, while the treaties are real, and impose responsibilities on the United States, in the greater scheme of things, they pose no greater regulatory issues than most other industrial activities. Please note once again: I acknowledge the importance of treaties.)
Tackling the Problem
The insurance issue arose very early in OCST’s history, and had extremely limited funds. Our first priority was to start documenting the real and relative risks that commercial space transportation posed. The key to setting a proper insurance level was associated with real, not perceived or anecdotal risks. Beyond insurance, however, knowing the real 3rd party risks that commercial space transportation presented also was key to evaluating commercial space launch applications.
Understanding the actual risks posed by commercial launches was important, but that was not sufficient basis for us to set insurance requirements credibly. Nor was it going to be enough to pressure NASA or USAF to change their requirements; although the CSLA gave DOT responsibility to set insurance requirements, DOD and NASA were going to have to agree that DOT could do it better than they did, and have confidence in the levels we set.
We tasked a team at the Transportation System Center in Cambridge, Massachusetts to start identifying, analyzing and documenting risks. I wanted to not only the risks that launch vehicles caused, but how many people got killed at the end of the runway in airport crashes. I wanted to know how many people had gotten killed by meteorites, falling satellites, and also the blue ice that that the media reports falls from airplanes and crashes through people’s roofs. The more that we could put all aspects of space transportation activity, including rockets and satellites, falling back to earth, the better we could explain to a the “space focused” experts why our approaches were sound.
Maximum Probable Loss
I was looking for the strategy and approach to better compute the proper insurance level. There were a lot of approaches we could use. We could look at government insurance requirements in other fields, or set it based on risk. However, Don Trilling suggest I talk to Dick Will, an individual who Don had hired for projects like this in the past. Dick suggested we consider using “maximum probably loss” as a starting strategy.
There are three key aspects of setting insurance using maximum probably loss that are key: the probability of certain events happening at various probability thresholds, the selection of a level of risk one accepts (e.g. 1 in a million [10 -6], or 1 in 10 million 10 -7) and a value of life. Only the first is achieved by computation and analysis. The second two are subjective and manipulatable.
The second aspect, i.e. selection of a level of risk, involves simply this: against which frequency of event are we protecting the government ? Note: because of the outer space treaties, the US government was liable, so the insurance is to protect the government.) For example, 1 in a million (10 -6) is a rare event; if a commercial rocket launched every single day, this a 10 -6 event would be expected to occur once every 2,700 years. What should we choose? 10-6? 10 -7? The more remote the risk of the event, the larger. Hence if a 10 -6 launch accident would kill 1 person, the 10 -7 event would certainly kill more. If I chose the latter, I would be protecting the government for an event I expected to occur every 27,000 years. In most cases, regulators would not do anything to protect a person for an event that would occur every one in a million events. How remote is one in a million (10 -6)? Your chances of being killed by lightning in any one year is 1 in 700,000. So, despite the fact that at current launch rates was pretty rare, electing a risk level of 10 -6 would make the most sense.
The third aspect, i.e. value of life, is even more subjective. Value of life is often used by regulators to assess the cost benefit impact of regulations. One simply multiplies the value of life times the number of lives saved and compares that total against the total cost impact of the regulation. Dividing the total cost of the regulation into the lives saved gives the cost:benefit ratio for the regulation.
There is no standard for value of life. An old person has less future income potential than a child. A doctor has a different income potential than a musician. There are so many different factors to be considered, that the value can be almost anything providing one comes up with a rationale for the number. Different regulators use different values because it assists their cause. Because so many different agencies use so many different numbers, we had our pick. At the time, most agencies were using $1M-$2M as the value of life. That would work for us too, we thought.
We modeled this approach quite a bit using large launch vehicles and small ones. We played around with different scenarios. We really needed to test the efficacy of this approach. At the same time, we were looking at other approaches to setting insurance.
We told members of the space community, such as launch companies, congressional staff, NASA and DOD, what we were doing. There was a lot of excitement about our efforts because DOT was talking about an analytical approach to setting insurance. The more we talked, the more DOD and NASA started talking about letting OCST set the insurance rates for their facilities. They had been burned by their method, and would have input into our approach.
Unbeknownst to the outside world, there was a problem with maximum probable loss. The fact of the matter was that at the one in a million event there usually would be no or just a few deaths. We found that we had to go up to the one in 1 in 10,000,000 level to get any 3rd party life losses. Moreover, at the lower value of life levels, we would be setting a very low insurance level–too low to be acceptable to the established space community, and perhaps others.
We were in a bind. By this time, our approach of “maximum probable loss” was being treated as the answer, although no one had seen or really understood how it would work. The conventional wisdom about space transportation being the most dangerous activity in existence still persisted and companies were talking to the Administration and Congress about indemnification. As they described it, they didn’t want to “bet the company” on a space launch accident. The Administration (along with OCST) and Congress were indicating they would go along (everyone was worried about losing the commercial launch industry.)
We were walking a tightrope. On the one hand, we could not have insurance requirements so high that they would hurt the companies, either by an inability to get insurance or due to unaffordable premiums. On the other hand, Congress, the Administration, NASA and the USAF, and opponents of giving industry unwarranted subsidies were watching to make sure we didn’t set the insurance rates too low.
The analyses showed we had a serious problem. If we used reasonable risk and value of life levels, the insurance requirements would be adequate to protect the US government, but after a world in which was accustomed to NASA’s and the USAF $500M requirement we would face a credibility crisis. For example, the US was safe requiring Conatec to have NO insurance. At the 10 -6 event level there were no fatalities. Only by going to a 10 -7 event did a fatality appear. Even then, using a $1M or $2M value of life only resulted in a $2M insurance requirement at most. No one in would be willing to accept a $0 insurance requirement. In fact, they wouldn’t believe the results of the analyses because they were pre-conditioned to believe this was super dangerous.
So we didn’t use a “pure” maximum probable loss approach. I directed that we start using other scenarios based on sound assumptions that would boost the low end of the numbers. We increased the value of life to $3M and upped the risk threshold to 10 -7. Even then, we used a marriage of the other approaches we had researched to justify a minimum of $10M. Using this modified “blended” strategy we were able to arrive at insurance requirements which industry found to be entirely reasonable, and which at the same time we could defend against critics who were arguing that we were putting the US government at risk because of the government policy to indemnify companies against all damages above the insurance requirement that OCST set.
It Almost Didn’t Work
As with everything where we broke new ground (which was virtually everything), I required we document all our analyses and issue them in the form of reports. The entire range of options is outlined in a report with the cumbersome title: Alternative Methodologies for Setting Government Third-Party Insurance Requirements For Commercial Space Launches with Focus on the Conatec Application # CON/003.
We coordinated this report with NASA, USAF, congressional staff and relevant offices within the Administration. The reception was extremely favorable. We had demonstrated that it was possible to set reasonable insurance rates based on sound analyses. Moreover, this went a long way to establish our technical bona-fides in the space community.
However, as I mentioned, we were walking a tightrope. Industry had to agree that it was not too high. The Administration had to agree it was not too low. NASA and USAF were okay with our process. Unfortunately, the Department of Justice stepped in.
The Department of Justice responded with a 7 page letter to OCST’s Program Counsel, Gerald Musarra, that raised questions about our approach, the potential of collusion within the insurance industry, and whether we were subsidizing the industry, among many other issues. They didn’t challenge the technical analyses, but were instead uncomfortable with the low insurance requirement. DOJ thought we were exposing the US government to risk because of the indemnification feature offered by the US. In fact, DOJ thought we were being governed by our “promotional” role. Gerald Musarra and I met with the Justice Deputy Assistant Attorney General in a long meeting to address DOJ’s concerns. It was an unsatisfactory meeting. DOJ did not have a different number to offer, but it also would not concur with our approach.
We were stuck and at an impasse. Without DOJ concurrence, we couldn’t issue the insurance requirements. I figured, this cannot continue, and I was ready to play “hardball”. After several days of no change, I sent Gerald Musarra a memorandum outlining how we should deal with this, and attached a draft memorandum for him to send to DOJ. One paragraph of my memorandum to Gerald Musarra stated”
“The reason we should take the next step is to prevent (DOJ official’s name withheld for privacy reasons) opinion from becoming both his and Justices’ final position. I think we will achieve this if we stand firmly behind our product and our position that we have adequately protected the government. Rather, we achieve our goal if Justice acknowledges DOT’s analysis, and passes the buck back to DOT. The strategy for Justice on this issue should be the same that we used to get the State (Department) to agree to transfer ITAR to us. Simply, we should make it safer for them to be quiet than to venture into an area of uncertainty beyond their technical expertise.”
Gerald Musarra sent a memorandum, based on my draft though slightly gentler, stating our confidence in our analysis and approach. It laid out our next steps which included issuing the insurance requirements to Conatec. It closed with a message to the effect that if DOJ wanted a different number, they needed to provide a number backed up with their analysis.
That ended the issue. We had walked the tightrope successfully. I issued the first commercial launch insurance orders requiring $10M 3rd party liability insurance to Conatec on May 9, 1988.
I didn’t feel good about the floor of $10M for 3rd party liability insurance that we had to set, even though we barely made it past the Department of Justice with that figure. Five years later, a company complained to the DOT Inspector General that our “minimum” cost them several hundred thousand dollars more in premiums. For small rocket launching companies, that would have been a huge economic blow considering the price they could procure for their rocket launch. This is a great example where no matter how much a government official is in tune with the needs of the public or an industry, his/her hands can be tied by other public policy issues. To this day, I believe that a $2M insurance requirement would have amply protected the US government.