Some Launch Economics

A commenter on my post about Soyuz launches from Kourou raised the question of whether launch prices would really drop with an increase in flight rate. This question may be due to discussions about a report prepared for NASA by a Dr Hertzfeld of GWU that was discussed on the Space Show last year. I’ve been meaning to discuss the topic of launch economics for a while, since it was raised during my Space Show, and also because its a rather important question to answer for the emerging space industry as a whole.

There are a few pieces of anecdotal evidence that leads me to believe that increased flight rates typically will lead to lower prices. This doesn’t mean Dr H’s research was wrong, or that I’m more knowledgeable about economics than he is (because I’m not), but might mean that some people may be drawing incorrect conclusions from his paper.

I remember at one point while developing our current line of rocket engines, we had found an analytical model to help us estimate a couple of design parameters. The model (which is a fairly well-accepted one) was suggesting that our design flat out wasn’t going to work, and that we were going to have to try some drastic measures to get it to work. One of my coworkers though had been looking over actual data from our system, and questioned the model. It turns out that he was right, and as far as I can tell my understanding of this model (or possibly the model itself) was off by over a factor of 3! Ever since this situation, I’ve tried to be a little more humble about my capacity for correctly understanding things the first time, and have been a little less willing to brush-off inconvenient facts that ruin my otherwise perfect theories.

Two of the pieces of evidence that lead me to believe that there is something wrong with the theory that increases in flight rate won’t decrease price are:

  1. There is a lot of evidence that when you sign an order for multiple flights, you can usually negotiate a better per-flight price than if you order them one at a time. One publicly available piece of evidence corroborating this common knowledge is shown on the SpaceX site:

    Pricing and Performance
    SpaceX offers open and fixed pricing that is the same for all customers, including a best price guarantee. Modest discounts are available for contractually committed, multi-launch purchases.

    I’m sure that others that have dealt with launch providers (such as Dennis Wingo) could provide additional support to this piece of evidence–that “bulk buys” of launches will almost always net you a lower price per launch. The price reduction is probably modest mostly because most such bulk buy contracts are for a relatively small amount of launches spread out over several years.

  2. As part of their study to try and determine if Atlas V could be used for servicing Bigelow Aerospace’s planned Sundancer station, the ULA Atlas V team investigated the business case as well as the technical requirements. According to several sources, the total per-person price for a visit to Bigelow’s station (including launch price, the price of the capsule flight, and the price of spending time there at the station) is actually substantially below what the current prices of Atlas V would suggest. One online article (whose numbers jibe with several other sources I’ve seen) states that:

    So how much will a space cruise cost the Average Joe? One week on the Sundancer alone will cost $7.9 million per person, predicts Bigelow. While pricey, it’s a steal compared to the $20 million that Space Adventures, another orbital flight company, charges for a week-long whirl on the International Space Station.

    Even assuming the $7.9M/seat only covered the Atlas V launch costs (and all the info I’ve seen says that includes Bigelow’s fee and the capsule providers fee as well as the Atlas V launch price) , you’re talking about less than $60M for an Atlas V 401 flight. That’s more than a factor of 2 drop from the current Atlas V 401 price, and is only possible because Bigelow is talking about providing for a much higher flight rate than the measly 1-2 flights per year Atlas V has been bringing in so far. Now obviously whether they will charge that price or not depends very strongly on if they think they can get enough demand to justify it. If Bigelow only buys 2-4 flights per year from them, the price will likely be a lot higher. But if the flight rate ever reaches the 16/year that Bigelow was talking about, their case can close on the $8-10M/person price point that Bigelow has been talking about.

While those two examples don’t conclusively prove that in all situations increases in flight rate will drop price, they do show that that is in fact the case at least some of the time. Another commenter, Habitat Hermit, I think gave a much better explanation about the fundamental idea that governs whether prices will drop or rise–profit maximization:

[T]here isn’t a conflict between demand & supply elastics (what you’re reacting to) and profit maximization (what you’re alluding to even if you manage to contradict yourself), quote contrary the two depend upon each other (which is sort of beautiful if you think about it and its consequences).

The launch industry’s publicly traded companies will always seek to maximize their profits, they’re even bound by the law to do this to protect the interests of their shareholders (privately owned companies without shareholders is a different story but I don’t know of any currently available launch company that is).

Now as to whether the price will increase or decrease can be simplistically illustrated by where on the supply & demand curve the company presently is and where it is moving to. You are absolutely right that increased demand does not always lead to decreased prices and both scarcity and oversupply are important factors.

So what is the current situation?
– almost all launch companies are looking for more customers (could be interpreted as oversupply but it isn’t, the correct term for this is overcapacity: the business as a whole has spare unused capacity it would like to fill to maximize its profits further)
– however it’s not like the launch companies are constantly producing more launch vehicles than they have any use for (that would be real oversupply as well as the opposite of profit maximizing and if they did that the leadership would be quickly thrown out by the shareholders)
– there are plenty of available launch windows, especially for LEO (no scarcity)
– there is no lack of material or factory capacity for increasing the production of launch vehicles (no scarcity)
– most launch complexes can be easily expanded (and NASA is trying to get rid of a rather large one in an excellent location…)
the main cost to the launch company is in maintaining the infrastructure and workforce required (fixed capital costs) and not in the launch vehicle production and launch itself (variable cost depending on units produced). Actually the fixed costs dwarf the variable costs, the two aren’t remotely close.

Regarding the last point there are at least two things worth mentioning:
1. This is the explanation for why the barrier to entry into the industry is rather high unless you can evade or mitigate the huge fixed costs. A NewSpace company like SpaceX is able to attempt it because it is lead by an “angel” investor (Elon Musk) willing to spend a lot of money for little in return in order to establish the company. Another NewSpace company is attempting to evade the fixed costs by using air-launches (that would be t/Space) and thus has less of a need for money (but they’re still struggling getting enough investors).
2. The second thing is that this is a very strong indication that increased production even if at a lower price would be more profitable. In other words if a somewhat lower price would bring in significantly more sales your total profit would rise (thus profit maximization).

What would cause the price of a launch for the customer to go up even though there was high demand? Scarcity in the launch industry. If the launch industry has little additional capacity and/or the fixed costs are no longer such a major cost relative to the product production and/or when investing in additional infrastructure (industry expansion) would not lead to increased profits through increased sales, then profit maximization would result in price increases to the customer (taking advantage of the oversupply of customers).

However that simply isn’t the case in the immediate future and even an additional 40 Atlas launches a year would not make it so for the industry as a whole, likely not even for LM. It’s also important to understand that guaranteed sales as would be the case with Bigelow Aerospace is somewhat different than one-by-one sales and makes any need for additional capital investments much more tenable (if theres should be such a need at all that is).

Might as well stop now or this is going to become very long and cumbersome. I guess any introductory textbook on microeconomics should do a better job at this than I have.

I could probably go on as well, but I think that Habitat Hermit put it very well. The ULA team will do whatever they think will lead to the most profit for their investors, as will any other publicly held company (and even most privately held ones unless they are owned and run by the sole investor–like SpaceX or Bigelow), and in many cases the best way to maximize profits will involve dropping their launch prices.

The following two tabs change content below.
Jonathan Goff

Jonathan Goff

President/CEO at Altius Space Machines
Jonathan Goff is a space technologist, inventor, and serial space entrepreneur who created the Selenian Boondocks blog. Jon was a co-founder of Masten Space Systems, and is the founder and CEO of Altius Space Machines, a space robotics startup in Broomfield, CO. His family includes his wife, Tiffany, and five boys: Jarom (deceased), Jonathan, James, Peter, and Andrew. Jon has a BS in Manufacturing Engineering (1999) and an MS in Mechanical Engineering (2007) from Brigham Young University, and served an LDS proselytizing mission in Olongapo, Philippines from 2000-2002.
This entry was posted in Business, Economics, Launch Vehicles. Bookmark the permalink.

10 Responses to Some Launch Economics

  1. Monte Davis says:

    Hertzfeld’s argument has never been that costs/prices don’t go down with increasing flight rate (first derivative), but that there’s scant evidence so far of “knees” in the curve (second derivative) where a dollar invested in ramping up the rate more than pays for itself.

    What makes suborbital tourism such an interesting potential lever is that it brings a psychological multiplier into the mix. A commsat or remote sensor (or propellant tank for an orbital depot 🙂 sitting on the ground doesn’t “know” or “care” that its peers are being launched. But it may turn out that lots of people watching the first flights of SS2 will say to themselves “I gotta try that.”

  2. mz says:

    You get a (shallow) virtuous cycle though even if the price at flight rate derivative is low and the second derivative is zero.

    Crude example:
    If you buy ten comsat launches but there is a decrease in price because you just bought so many, you can perhaps launch one more, making it eleven. (Not another ten!) Say with a rough model, if the price drop is 1% per ordered launch, then 10 launches would save you about ten percent over the price, thus one launch. If you used fixed money, you could in fact launch eleven.
    (If the satellites are not so “elastic”, it wouldn’t be plausible etc)

    And since the launches now would be cheaper to other instances too (without any effort by them!), they could potentially buy more of them.

    Now, there are instances like NASA that could do this buying of launch services to kinda “jumpstart” the industry. They could multiply US launch demand. Or increase the world’s significantly.
    But that’s not explicitly in their charter, so it’s understandable.

    Btw notable in this “fixed percentage” model is that the gains get smaller. In absolute dollars, as the price gets smaller, the savings by ordering many are less.

    With 1% percentage cheapening rate it’d take 70 launches to drop the price per launch to half. (0.99^70 = about 0.5)

    Anyway, these models have not that much to do with the real world when extrapolated so much. I could have choosen any other formula just as well.

    We’re not even in the fixed vs marginal costs discussion yet…

  3. mz says:

    Damn I was tired yesterday when I wrote that. I’m sorry about the low quality of the comment, the crypticity and the typos. By the derivatives I meant that the rate of cheapening wouldn’t get any better with time in the model. No knees indeed, but an exponential decay: current_price*0.99^n.
    (I don’t believe this model reflects reality much.)

    Anyway, if you assume the number of clients is fixed and if spaceflight gets cheaper it is possible they will not spend more money on it. If they spend the same amount or not much less, it is still possible to get a “virtuous cycle”, since a cheaper launch service gets a greater portion of the market.

    Of course, it’s debatable that at current spending situation, the point becomes very quickly where even if you have all of the current commercial market you still launch too rarely to make an RLV profitable.

    In 2005 there were 52 total launches worldwide. (Some of them contained multiple satellites.) A big portion were commercially competed.

    http://planet4589.org/space/log/launchlog.txt

    Market research would be fun.

    With COTS, Bigelow, VSE, and even some suborbital tourism, some changes and synergies might happen.

  4. Paul Dietz says:

    Correct me if I’m wrong, but don’t federal government contracts typically require that the feds receive a price that is no higher than is offered to private consumers of the same product/service?

    If this is the case, then having the feds as an anchor customer for a launcher could be a huge two-edged sword. The supplier would not be able to offer discounted launch services (for example, at the marginal cost of adding another launch) to attract new customers without damaging the revenue stream from the government.

  5. Bill White says:

    Paul, this is exactly why I oppose selecting an EELV to carry CEV:

    If this is the case, then having the feds as an anchor customer for a launcher could be a huge two-edged sword. The supplier would not be able to offer discounted launch services (for example, at the marginal cost of adding another launch) to attract new customers without damaging the revenue stream from the government.

    Boeing voluntarily withdrew its Delta line from the com-sat launch business for very similar reasons.

    It was more lucrative to sell fewer launches to NASA & DoD than chase the commercial market and compete on price.

  6. Jardinero1 says:

    Here is what I said earlier in the comments that Jon alludes to:

    “There is a glut of launch providers and little demand. If and when launch demand increases; for on orbit refueling or whatever, launch prices won’t go down, they will go up. Increased demand does not reduce prices, it increases them. Even if increased launch rates drive down marginal costs it doesn’t follow that a launch provider is going to charge less. In a market with rising demand he has every incentive to charge more and keep the difference as profit.”

    I respect the technical expertise that Jon and other posters bring to this blog. It baffles me why the same posters are going through contortions to refute an axiom of economics that was pretty well settled in the nineteenth century. The only way you can get the outcome you argue for is to switch the supply and demand curves with one another. From what I am reading, all I can figure is that somehow you are conflating the cost dynamic of a particular launch provider with the dynamic of what same will charge a launch buyer. A launch provider is going to charge as much as the market will bear, no matter how low his own costs are.

    The issue of bulk pricing is raised as an instance where more launches lower the price. While a valid example, it isn’t used appropriately. Bulk pricing exists in many industries; autos, airplanes, pencils, advertising, insurance. It does not demonstrate a reduction in pricing to the end buyer unless and only if you compare it to a competitor’s bulk price for the same package. Comparing the price of 10 Atlas launches to one Ariane launch proves nothing. You have to compare 10 Atlas launches to 10 Ariane launches to have a valid comparison. My guess is that a launch provider is going to charge nearly as much as his competitor in a bulk pricing arrangement. He would be stupid not to.

  7. mz says:

    Paul, this is interesting in relation to Ariane and the new EELV fixed subsidies. Can they sell to commercial entities with profit if the price is just above the marginal cost, if the government pays for the fixed costs anyway. (btw that brings money to the launch company and then the government needs to pay less to sustain it.)

    The approach of Ariane is, as far as I know, to try to make as much money with the commercial sector and make the difference (since you operate at some loss) up with government subsidies.

    With EELV:s it’s trying to be “compete the companies and sell a launch to the government at a price that includes the profit” approach, except the companies are doing terrible and threaten to withraw so subsidies are needed.

    In both cases it’s the fundamental political decision that some kind of space access is needed that makes it so different from many other areas.

  8. Habitat Hermit says:

    I noticed George Sowers (vice president for business development and advanced programs at ULA) talking about launch prices in this Space Review article

    Here’s a choice paragraph from said article:
    “Sowers was optimistic that additional demand for the Atlas 5 from commercial orbital manned missions could benefit all users of the vehicle. “The launch vehicle industry is very highly invested in fixed costs,” he said. “If there’s a new big market we think we can get factors of two to four, nearly, in cost reduction by increasing launch rates by factors of two to four.””

    That’s even better than I would have expected and gives LM/ULA much more room to manouver. Great to see them talking about this out in the open.

  9. kert says:

    bytheway
    have you taken a look at http://www.lunartransportationsystems.com/technologies.aspx

    The architectures proposed there seem to align very well with what has been advocated on this blog and elsewhere.

    Using a Delta II as a baseline vehicle for lunar sorties, no less.

    The bonus, they have illustrations 🙂

  10. Habitat Hermit says:

    Jardinero1 I’ve been thinking about how to resolve the misunderstanding between us as I see it and how to do it clearly and thoroughly.

    First we have the most fundamental example of a supply & demand curve. I’m going to borrow illustrations from wikipedia; here’s the illustration). The accompanying paragraph does a fair job of pointing out exactly what this graph illustrates.

    If I have understood you correctly I think one of the misunderstandings you have is that you see the upwards angle of the supply curve in that illustration as proof that prices will always go up. But look closer at the graph and its meaning; it is the intersection between the supply and the demand curve that matters and unless that shifts then any further supply from the producer will only lead to oversupply (the orange area) that customers aren’t willing to buy – i.e. it will not lead to increased prices only a decrease in profit to the seller as they’re left with surplus stock (incurring costs but no profit) no-one wants at the current price.

    Side-note: when this happens the producer will try a host of different approaches to reach the equilibrium price and quantity including but not limited to artificially raising demand (for example advertising or discounts) to evade further costs of storage, or simply destroy the surplus supply, or stop or reduce output for a limited period etc. – whichever practical solution costs the least to the producer.

    Second we have shifts in the supply and/or demand curve which result in the movement of the intersection of the two (the price and quantity equilibrium). Wikipedia has several sections with different illustrations introducing shifts in the demand curve, shifts in the supply curve and so on but I won’t be using all of it. I’ll focus on the illustration that most closely resembles what I’m arguing, this one from the section on Supply curve shifts.

    Now, remember the first illustration and the orange field representing oversupply. The launch industry does not produce such an oversupply because it would reduce profits (goes for every business) however the launch industry has over-capacity. They would like to use that over-capacity if it ends up increasing their overall profits.

    So in the second illustration let us say that we are currently on the intersection P1Q1 with a launch industry (supply) that has
    – very high fixed costs
    – few sales compared to what they could produce (over-capacity)

    At P2Q2 you have
    – a lower price than P1Q1
    – more sales than P1Q1 (less over-capacity than at P1Q1)

    Now here’s the deal, if a company in the launch industry would have a greater total profit for a product at P2Q2 than at P1Q1 (through the increased sales) and if they see an opportunity to move to P2Q2 (because of customer interest at that price) then they will of course try very hard to make that possible to maximize their profits.

    On the other hand if they don’t see greater total profit at P2Q2 they obviously won’t try to shift to it.

    If you have very high fixed costs (costs that are there no matter how many or few products you supply) then producing only a few units means that the high costs have to be paid for by those few units and they are going to be very expensive. If you produce a lot of units you have more units to pay for the very high fixed costs and each of them can be cheaper.

    A lot more detail could be added to this discussion but I’ve tried to get at the root miconceptions with this post instead of dissecting your post sentence by sentence. I hope it at least established some common ground upon which to base further disagreements ^_^

Leave a Reply

Your email address will not be published. Required fields are marked *