Dr. Metzger's model of economics of starting a city on Mars

[su27k]

There're some other threads discussing similar topics but I think it might be better to have a thread focusing on this specific model.

https://twitter.com/DrPhiltill/status/1468772477333557248

Hey Mars settlement fans: I modeled the economics of starting a city on Mars, including the use of Mars resources to build over a few decades, and including income to offset costs. The model indicates much cheaper than @elonmusk estimated here:
https://twitter.com/elonmusk/status/1159964499975135232?s=20
.../1


2/ For this model, I used actual data from the US Census bureau, the bureau of labor statistics, etc etc, assuming most segments of the economy will be built on Mars. A portion of the residents will work in the services sector (NAICS 51) just as in the US economy, so...


3/ ...exports from those workers will be "massless", easy to transport back to Earth. Including only those exports, the settlement can completely pay for ongoing imports of materials from Earth within a couple decades, thus keeping total cost low...


4/ I was surprised. Using Elon's transportation cost estimate and actual all-sectors economics data, the total Elon would need to outlay looks quite low. This is a zero-interest J-curve (i.e., the loan covered by Elon). It goes no lower than $2.5B at the deepest point.


5/ Certain segments of our current economy would be totally omitted (e.g., leather and wood) and the new costs for making air & water plus doing indoor agriculture are not that large compared to the total economy, thus easily absorbed.


6/ The biggest challenge was to estimate the number of unique items manufactured in a closed (self-sustaining) supply chain. Do we make 100,000 unique electronic components? What capital assets are needed to make them? Can we redesign everything to use fewer unique parts?


7/ Some researchers I know are working on reducing the number of unique parts for an off-planet economy. The extreme limit is a "general self-replicator" where *one* machine is able to make all of itself. This new model avoids fanciful extremes, so...


8/ ...it only assumes a modest reduction in supply chain breadth. Instead of 1000 types of bolts, Mars might start out by making only 100 types of bolts and design everything for just that. But this question is IMO the biggest source of uncertainty in the model.


9/ Everybody who knows me knows I am more of a cislunar person: put industry in cislunar space using lunar and asteroid resources to save the Earth (and get revenue servicing Mars transports). This modeling result surprised me. I no longer think settling Mars will be that hard.


10/10 I plan to present this model on Friday at the ESA's Space Based Solar Power workshop as part of a discussion on how we can use space resources to build clean energy for Earth.

[su27k]

https://twitter.com/DrPhiltill/status/1468787591889793027

The label was bad on the graph. That was net imports/exports but did not include initial hardware development costs or human immigration costs. Those costs can never be recouped (per this model) so Elon & immigrants will need to just pay them. A rough estimate says… 1/2


2/2...only $10B for the initial habitat extrapolating from NASA's design reference mission + this cost estimator: https://globalsecurity.org/military/intro/reference/calc/AMCM.htm#Block%20Number with economies of scale and cheap transport reducing further. The model includes ongoing habitat expansion via self-funded in-situ industry.

[colbourne]

"My question is, what’s your current estimate on the total cost of having a self-sustaining civilization on Mars?"

If we assume that there will be no inputs from Earth to be classed as self sustaining (Maybe Earth civilisation has been destroyed. This is the eventuality that Elon wants to prepare for to allow human civilisation to be more robust ). This requires the seed technologies and resources to allow a civilisation to grow. I expect it would start small as too many mouths to feed and provide living space for would be harder.
They would need the ability to make sealed  tunnels, produce enough food, process minerals (meteorites) , manufacture suits and airlocks, and some kind of power source (probably solar or wind). Solar could be PV panels or reflectors (powering a turbine which might be simpler to construct than PV  ).
I think we need to work out how to sustainably fulfill these few critical tasks and then we could have a chance at a sustainable growing civilisation on Mars. Big might not be better initially. Ten people with the right skills might have the best chance of being the seed.

[JohnFornaro]

For this model, I used actual data from the US Census bureau...

This seems to be a faulty assumption.

[JohnFornaro]

The extreme limit is a "general self-replicator" where *one* machine is able to make all of itself. This new model avoids fanciful extremes, so...

An extraordinary claim, if not an extreme one.

[Twark_Main]

For this model, I used actual data from the US Census bureau...

This seems to be a faulty assumption.

It's an *imperfect* assumption to be sure (like all analogies), but do you have a better suggestion about what assumption should have been used instead?

[JohnFornaro]

For this model, I used actual data from the US Census bureau...

This seems to be a faulty assumption.

It's an *imperfect* assumption to be sure (like all analogies), but do you have a better suggestion about what assumption should have been used instead?

[Edit, 12-31-21. 

I am watching his presentation, linked below.  Although I still hold that the modeling assumption of the entirety of the US economy is "faulty", I would like to walk back the apparent severity of my terse summary above.]

The salaries of the astros involved are immaterial to the costing, as I see it.  In a competitive world, fairly administered, plenty of qualified volunteers would be available for no cost whatsoever.  NSoV.  No Shortage of Volunteers.

Also, Musk will throw out a twit and a hundred thousand people will salute it.  This doesn't mean that Musk cannot accurately estimate costs and profits.  When a contractor gives a homeowner a price that spans two orders of magnitude, the homeowner knows that it is not a guaranteed price in the least.

[lamontagne]

The element that worries me the most would be the evaluation of the cost of food on Mars, that may not have been properly evaluated from the brief mention in the tweets. Enclosing everything in a pressure vessel must add to the cost of living :-)

Looking forward to seeing the model though.

[Twark_Main]

For this model, I used actual data from the US Census bureau...

This seems to be a faulty assumption.

It's an *imperfect* assumption to be sure (like all analogies), but do you have a better suggestion about what assumption should have been used instead?

The salaries of the astros involved are immaterial to the costing, as I see it.

I see nothing in those tweets to suggest that this is how the model works.

[su27k]

Dr. Metzger's presentation starts at 2:43:15:

[JohnFornaro]

Dr. Metzger's presentation starts at 2:43:15:

Thanks for this.  Metzler presents Economics of Space Derived propellant here:

[Edit:  The time stamp feature didn't work properly on this post]

See the index for various time stamped conference sections.

[JohnFornaro]

For this model, I used actual data from the US Census bureau...

This seems to be a faulty assumption.

It's an *imperfect* assumption to be sure (like all analogies), but do you have a better suggestion about what assumption should have been used instead?

The salaries of the astros involved are immaterial to the costing, as I see it.

I see nothing in those tweets to suggest that this is how the model works.

His model is based on economics, which includes the salaries of all participants in the model.  My point, poorly worded, is that there would be a large number of *colonists* who would make the trip and work without pay for some time.  If true, this suggests that the *economic* costs might be lower.   However, I think the total costs will be far higher than suggests, which is not an argument not to make the attempt, but rather an argument to tighten up the estimate.

[Robotbeat]

Yeah, kind of like people going on the Mars Analogue missions for The Mars Society. They aren't paid. If anything, I think they might have to pay for them.

[su27k]

https://twitter.com/DrPhiltill/status/1479235313394462724

I’ve done a lot of work on this question and it is surprisingly affordable (relatively speaking). Making air etc. is negligible cost compared to iPhones and everything else so is easily absorbed. But the problem is the scale needed to be self-sufficient. It requires O(1M) ppl /1


2/…or possibly O(100k) ppl.

https://twitter.com/DrPhiltill/status/1479257235591245833

The modeling was based on extensive US economic data so was realistic except the key assumption is that we can “skinny down” the supply chain by redesigning everything to reduce the number of unique parts and materials. E.g., instead of using 1000 types of bolts, use just 100./1


2/ i had to make a crude guess at the amount we can narrow the supply chain in each sector. The mode then assumes that we only increase population as we build the factory machines that need people to operate them (or whatever other kind of job is needed) so no extra ppl to feed.


3/ The model assumes we import everything from Earth but evolve toward sufficiency by building capital equipment & buildings over time. We retire first those sectors that produce the most mass per capital mass to reduce import mass ASAP. Move through the sectors over time.


4/ The model found that it takes about 40 years by this lowest-expense method before the economy can be self-sustaining, but that figure depends heavily on my guesses at how much we can narrow the supply chain.


5/ The cost of establishing this industry (a full supply chain on Mars) is offset be “massless” exports (data, services, tech licenses, etc.) back to Earth. Using US economic data for these sectors the exports can be very substantial and keep net costs quite low.


6/ By the time about $150k ppl live in Mars the debt per capita on Mars is comparable to the cost of an Orlando home. By about 20 years or so the settlement is profitable so debt is decreasing. (But this assumes Elon covers the cost of initial settlement on a zero interest loan.)


7/7 These results are preliminary and I plan to make the model more rigorous before publishing. But I was surprised that it came out as optimistic as it did.

[Slarty1080]

There are some trades that are really difficult to unpack at the moment. They are also subject to change over time:
How expensive will it be to shift mass to Mars?
How much money will be available and for how long to support the effort?
What proportion of the Mars cargos can be replaced with locally produced materials instead and how quickly?
What is the correct balance point between importing people and materials?

[Genial Precis]

The model appears to depend on Mars being able to sell services (as opposed to goods) to Earth. I don't understand how/why that would happen.

[deltaV]

The model appears to depend on Mars being able to sell services (as opposed to goods) to Earth. I don't understand how/why that would happen.

My guess is that the Mars exports that will be best able to compete with Earth versions will be goods and services where an authentic Mars version is a lot more valuable or marketable than Earth equivalents. For anything else the Mars version will be too expensive to compete with Earth-built equivalents due to Mars's inhospitable environment making everything expensive and/or the cost of shipping the goods or services to Earth.

Examples of possible exports:
* Mars soil and rocks can be sold to scientists and collectors
* Tourists will need lodging, food, air, tour guides, and so on
* Film movies set on Mars.
* Books on life on Mars written by Mars residents (with ghost writing and editing done on Earth)
* Televised sports with athletes beating many Earth records due to the low gravity and/or thin atmosphere.
* Can't afford to visit Mars in person? Instead rent a Mars robot for a day and walk around the settlement, write your initials in the Mars soil, or whatever. (You'd need a lot of artificial intelligence so the customer can give high-level instructions that work with the many-minute speed of light delay. You could also do this with a human with video cameras following instructions instead of a robot.)

[Genial Precis]

In the model services are broken out specifically as things not requiring transport of material. Tourism not a part of it.

Selling pet rocks on Earth has been proposed before. To make that work you need a cult like NFTs, MLM, Trump merch etc.

A lot of other stuff is probably killed by the 10-40 min latency in communications, especially as Earth stuff is increasingly habituated to more automation.

So, as I said, I don't understand how it could work. If the author understands, that is not clear to me.

[high road]

For this model, I used actual data from the US Census bureau...

This seems to be a faulty assumption.

It's an *imperfect* assumption to be sure (like all analogies), but do you have a better suggestion about what assumption should have been used instead?

At least acknowledge that to make it work as a model, Mars needs to import as much as the US does in quantity, at Mars-import cost.

edit: ah, having now seen the presentation, the model is: if we ignore the setup costs (which is Elon's 100 billion - 1 trillion dollars), the cost to transport the infrastructure that works on Earth (but would not on Mars at that price due to environment and supply chain dependencies) and handwave the part that explains how imports would evolve over time, and transport that at an admittedly low estimate of less than 1/3 of the current best price to LEO, and ignore that Mars would actually be paying for transporting exports (by having to compete with Earth prices to sell their stuff), that all amounts to a much smaller amount than the setup costs themselves. Which I assume he forgets that this amount is either part of or in addition to the setup costs.

So yeah, not really sure what kind of information we can get from this. Maybe it would be better if we would reduce the exports to mass-less exports and see how long it would take then to cover the total of the imports? Assuming we're making abstraction of what is being imported. That still assumes those kinds of exports would be comparable to the US economy for some reason, but as a thought expirement I see the value. You might indeed want to invest more in those industries and postpone ISRU with lower import mass savings.

[high road]

Dr. Metzger's presentation starts at 2:43:15:

What does SBSP have to do with an economic model for Mars?

Edit: the Mars part starts at 2:57:35 ish.