Exponential Economist Meets Finite Physicist
physics.ucsd.edu
physics.ucsd.edu
https://news.ycombinator.com/item?id=5814769
https://news.ycombinator.com/item?id=6346156
This is from the same site, from the title looks to be prior work on the same topic, also submitted previously (6 years ago) with no comments:
http://physics.ucsd.edu/do-the-math/2011/07/can-economic-gro...
How about saving the existential dread for your therapist? Let economists focus on economics.
I mean, how many physical theories can be argued to work into the infinite future? Does quantum mechanics or GR hold past the end of the universe? It's not even a meaningful question. Why hold economic models to the equivalent standard, when all that is needed is whether they hold in the relevant domain?
I feel that this question and argument is often used to try to draw the conclusion that we must stop trying for economic growth NOW, because infinite growth is impossible. Maybe it is impossible, but I don't think something happening or not happening at infinity or even in a thousand years has any bearing on what we should be doing now.
Looking at total energy and net heat production just hammers home that within an exceptionally short timeframe by historical measure growth of human activity on Earth must come to an end. And if you look at more complex systems models, there's strong argument that sustainable carrying capacity has been exceeded. When and by how much provides a wide range of estimates -- I've seen good arguments for from 50m to 10 billion, though lower-end estimates in the 500m - 2 billion range strike me as most plausible.
The first-order argument is to demonstrate that infite growth is impossible that's what Murphy demonstrates here.
Then there's the question of assessing what actual physical resource needs do exist. Looking at total population, relative biomass of humans vs. other terrestrial vertebrates (http://i1176.photobucket.com/albums/x330/chefurka/Zoomass_zp...), land area per person, human appropriation of net primary production (a fancy term for "how much plant growth to humans consume directly or indirectly), total arable land, available critical and limiting resources (particularly freshwater, topsoil, phosphorus, copper, and of course, fossil fuels, see Leibig's Law: https://en.wikipedia.org/wiki/Leibig%27s_law), incident solar flux per person, the situation for effluent sinks (sewerage, CO₂, industrial pollution, heavy metals, radioactive waste), systemic risks (see David Korowicz, "Financial System Supply-Chain Cross-Contagion": http://www.feasta.org/wp-content/uploads/2012/06/Trade-Off1....).
On land area and solar flux per person, if you assume that the only long-term energy option is solar (meaning: solar, wind, hydroelectric, wave, and biomass, all of which derive from it, but excluding tidal, and geothermal energy, as well as nuclear), then we're looking at, with 8 billion people on the planet, less than 0.5 MW of total available recoverable solar, at present conversion rates, per person via photovoltaics. This is with blanket coverage of the planet in PV panels. Yes, that's in excess of how much energy an American ses on an ongoing basis (about 10 kW continuous), by about 45x, but that excludes conversions to other factors, including liquid, solid, or gas fuels for later or mobile consumption. It's also a total land-area usage: 3.18 million km² without storage and fuels allowances. On a large-scale engineering basis, I'd really like a far larger margin of error. In 1850 it would have been 250x, in 1AD, 20,000x.
https://ello.co/dredmorbius/post/_bi5UhywbDyukHFY-EaYJw https://d324imu86q1bqn.cloudfront.net/uploads/asset/attachme...
At some point, the prospects of 1) Business as Usual, 2) "developing world" advancement to first-world standards (even EU vs. US, at about 50% of per-capita resource consumption), demographic transition, and surviving even a "modest" 8-9 billion total humans on the planet ... looks exceptionally dicey.
An exceptionally common response is wishful thinking, expressed variously as "there's no problem", techno-optimism, "they're" working on it (whoever "they" might be), etc. A significant problem with the techno-optimist viewpoint is that there's a compelling argument that technological level (complexity) is dependent on, not independent of, net energy throughput.
The more I've looked at the situation, the less credible I find any of these options. And "solving" any one issue (e.g., energy, which YC have addressed with several recent startups), still leaves multiple other challenges.
https://www.reddit.com/r/dredmorbius
https://www.reddit.com/r/dredmorbius/wiki/faq
It's complicated:
(not an economist, but broad-minded enough to recognize someone who has spend so much time in his own field to be completely rigid-minded about it. An affliction rather common to CS types too, I'm sad to say, since as a software engineer in research I spend quite some time around those)
That said, I'm not sure it's a real exchange. It feels somewhat manufactured to me. But that's somewhat irrelevant to the point, as long as the arguments each side presents are conceivable for their professions and the current prevailing beliefs within them (which I'm not qualified to judge).
In the end though, that's all irrelevant. The real question is whether you find the facts of the argument compelling and whether it changes your view on anything.
I am tempted to save this and put it under a file titled something like "why scientists should not be economists" but I'm afraid it's a little too involved for that purpose.
I mean no disrespect to the author: these all seem like good and reasonable questions. To understand why this form of analysis does not work is to cut to the heart of what the difference between economics and hard empirical sciences. I will give a couple of examples, but I fear that folks who missed it the first time around will not get it this time.
First, note that no matter what period in human history this conversation occurred, it would follow the same pattern. "Here is a chart of the rate of fishing from the pond near our Roman village. Certainly catastrophe awaits us in just a few years", or one that actually occurred, "Population growth will stagnate and we will have famine when the 1900s come. Why? Because we will have ran out of grazing land for all the horses we will need"
Growth lines can't continue forever. Seems obvious. Yet somehow it always turns out the thing you were measuring wasn't the important thing all along. Sure, you can only fish so much -- but there are other ponds. Yep, can't have that many meadows -- yet meadows aren't a constraining factor.
The reason why such analysis fails is that science is always seeking a constrained and known system. Given controlled conditions, the falsifiability and reproducibility of science says that given these inputs, these outputs are likely to occur. Economics, because it deals with what people want, is under no such constraints. Let's say you give everyone on the planet all the food, shelter, and healthcare they need. End of money, right? Nope. People would trade Justin Bieber concert tickets, or autographs of Elvis Presley. People always want stuff they don't have, and they always like trading for it. That's economics.
If economic systems were the same as physical systems, wow, we'd have a lot of physicists who have made fortunes playing the stock market. We do not.
Poorly written code might not be very valuable, because it's slow and crashes often. Well written code would arguably be more valuable. It would also waste less electrical energy, and produce less heat.
Indeed, code could do entirely new things, using less material and energy inputs, that are worth much more. Maybe I'm missing something, but that seems distinct from increasing efficiency. There's a limit to efficiency, but no limit to value.
The most serious objection to the propositions in the article is refuted at the beginning, when they agree that our future will be tied to one planet - "I assume you’re happy to confine our conversation to Earth". There are good reasons to expect that this will be invalidated somewhen in the next 400 years.
I keep hearing this but I'm very doubtful we could or even should do it.
Since the use of economic theory is in guiding our long-term decision making, I'd rather that theory be focused on getting it right with what we have, and be pleasantly surprised if we get a bit more growth from a distant space rock, rather than being reliant on the never ending (and growing) series of space rocks.
[1] Few space rocks besides Mars are particularly habitable, each subsequent one we chase after will be less so, and after that we're going outside the solar system where timescales are massive. Our best knowledge of physics/engineering today is that we can colonize these things, but our best knowledge is not that we can reach them at anywhere near FTL.
But the whole context of this discussion is a society that has thousands of times more energy available than we do today. That's why it's not at all unreasonable to include space in the discussion. It's hardly fair to posit advanced technology that decreases the cost of energy by a factor of a many thousands while claiming that dropping the cost of space transport by a mere factor of 100 is unreasonable.
No new physics is required. We already know how to build lasers that could lift cargo into orbit -- it's just a question of energy cost.
Habitability is irrelevant: space is for robots. We already know that population growth will level off in only a few decades, so "more people" is not the problem -- it's "more industry". The industry doesn't need to be where the people are -- not when our robots are already good and getting better very fast.
What's it tied to?
How does it grow without some real, resource-based, basis?
What evidence do you have for such a thing existing?
How about evidence that it cannot? What would falsify your premise?
Citation required.
There are awfully good reasons to believe it won't.
Even people whose job it is to be optimistic about space and science aren't.
Charlie Stross:
http://www.antipope.org/charlie/blog-static/2007/06/the-high...
"I write SF for a living. Possibly because of this, folks seem to think I ought to be an enthusiastic proponent of space exploration and space colonization. Space exploration? Yep, that's a fair cop — I'm all in favour of advancing the scientific enterprise. But actual space colonisation is another matter entirely, and those of a sensitive (or optimistic) disposition might want to stop reading right now ..."
http://www.space.com/29862-kim-stanley-robinson-aurora-inter...
Kim Stanley Robinson on interstellar / interplanetary colonisation:
"Q: What about interstellar colonization, in particular?"
"A: There are a lot of people, even powerful, influential people, who seem to think that the goal of humanity is to spread itself. I want this book to make people think really hard about — maybe there's only one planet where humanity can do well, and we're already on it."
The author seems to be (accidentally or intentionally) part of an ultra-heterodox economic school that believes that all economic value, across all times and places, comes from the consumption of energy.
They base this mostly on that graph near the top of the article that shows that energy consumption was growing exponentially during the industrial revolution.
He then goes on to imagine that 3% economic growth means that in 200 years we'll all have 400 times as many cars and refrigerators (or maybe just one car the size of a luxury yacht), eat 400 times as much food, have houses 400 times as big, etc.
Ultimately all life is based on exploiting energy gradients. Economic value without using more energy is ultimately an efficiency improvement, an investment of past energy which is proving beneficial.
https://en.wikipedia.org/wiki/Jevons_paradox
Basically that improvements in efficiency leads paradoxically to more usage, as new uses becomes feasible.
> Note: This conversation is my contribution to a series at www.growthbusters.org honoring the 40th anniversary of the Limits to Growth study. You can explore the series here. Also see my previous reflection on the Limits to Growth work. You may also be interested in checking out and signing the Pledge to Think Small and consider organizing an Earth Day weekend house party screening of the GrowthBusters movie.
I do agree that exponential growth in energy use won't end well. I'm reminded of the polar heat radiators on Trantor in Foundation by Isaac Asimov ;)
I also find the linear fit in log energy use vs time[0] somewhat dubious. It's confusing to just look at energy use. What matters is the relationship between economic growth and energy use. I'm sure that's been done. I'll look into it and report back :)
[0] http://physics.ucsd.edu/do-the-math/wp-content/uploads/2011/...
http://ourfiniteworld.com/2012/08/29/the-long-term-tie-betwe...
There are very good thermodynamic reasons to suspect the relationship will continue to hold.
Although tbh I find the exercise of imagining an economy with abundant (yet limited) energy quite difficult.
I read it as an explanation of why in 200 years we won't have 400 times as many cars and refrigerators. Looked at another way, freed from material and energy constraints, why wouldn't we have a fleet of drones and other appliances to fulfill our needs? Why wouldn't we have, if not a giant house or property, at least properties spread around in all the geographic locations we cared about? I think the answer is constraints on space, material and energy.
Alternatively, you can read that growth curve not as the individual's increasing utilization of resources, but as the increase of individuals themselves and the increase of resources this entails.
Some years ago, I found myself in a very small art gallery in a very expensive place. In front of my face was a weathered pine board, with 96 (as I recall) small red rubber serum caps nailed to it, in a reasonably regular array.
The price was ~$1000 ;)
Anyway, maybe the software is just charming.
"If the flow of energy is fixed, but we posit continued economic growth, then GDP continues to grow while energy remains at a fixed scale. This means that energy—a physically-constrained resource, mind—must become arbitrarily cheap as GDP continues to grow and leave energy in the dust."
It goes on to describe the paradox that if we can generate arbitrary wealth in such a system (through software, for example), it becomes possible to just buy all the energy[1] and break the system, which means that this doesn't constitute a really valid economic model.
And yes, it goes on to say that you can still have growth, perhaps even something that can be called "economic growth," but we've placed a fundamental limit on the growth of the GDP- the natural way economists would consider economic growth today.
[1] "Corner the energy market," if you're an Alpha Centauri fan.
Maybe this isn't relevant to your point, but I think there might be parallels between Moore's law, which the author argues will end, and OP's argument against exponential energy growth.
Take for example solar: right now, pretty much all of that energy (sans the work performed for propelling ocean currents and production of future fossil fuels via photosynthesis of plants) is irradiated back into space anyway. Using more of that energy to perform work on surface of Earth won't change a thing from global thermodynamics perspective. That energy is already on the books.
Or consider hydro (a consequence of solar itself): the moving water already carries energy and performs work. Re-purposing that energy in the form of electricity would have no impact on the overall energy budget.
Granted the energy growth is theoretically limited, but the limit might not be as near if we consider that much of energy generation could be essentially energy conversion.
My point was that the timeframe would likely be quite different, e.g. millennia instead of centuries. From theoretical perspective doesn't change anything, but perhaps makes the argument a bit less dramatic.
To "convert" energy, you need an entropic gradient.
Those are fundamentally limited.
Here's your inventory:
Total human annual energy consumption presently is about 560 million terajoules (TJ), 156,000 TWh, or 530 quad (quadrillion BTU), or 13.3 billion tons of oil equivalent (btoe), or 91.5 billion barrels of oil, or 19.1 billion tons of coal, or 4 million tonnes of processed uranium nuclear fuel.
Oil, gas, and coal combined are 11.0 btoe, or 83%, of all energy consumption. Nuclear and hydro, both possibly close to tapped out, account for another 1.4 btoe.
● Solar. About 1kW/m² at Earth's surface, modulo conversion efficiency -- 37% max for single-layer PV, 85% max for infinite layer, or concentrated solar power at Carnot efficiencies determined by input/output thermal differentials, 40-45% likely upper bound. And spacing factor (dependent on latitude, about 55% if you're in the continental US, lower in Europe, higher in the tropics), capacity factor, 20-30%, the amount of time you can capture sunlight.
● Solar-derived: biomass, wind, hydroelectricity, and wave energy are all based on solar flux. Biomass is largely tapped out and competes with both human food and natural ecosystem uses.
● Biomass is stubbornly persistent as a part of human energy consumption. It's also heavily dependent on fossil-based inputs: fertilisers, pesticides, equiipment, transport, processing, refrigeration and storage. About 6% of present net energy, unlikely to change. Recognise that all wastestream sources (food waste, sileage, sewerage, animal waste, forestry waste, and garbage) either derive directly from biomass or fossil fuels. I've penciled these out to at best 1-2% of present fuel consumption. Britain might be able to run its bus fleet on sewerage-derived biomethane, but that's about it. For the US, most biofuel crop yields would require multiples of total national land area to provide the present level of fuel consumption.
● Hydro's similarly largely tapped out, though some third-world potential remains. My suspicion is that hydro will increasingly fill a storage / dispatchable load role. Mind that it is among the highest return on energy investment options available. It's just not very available.
● Wind actually has good metrics on per plant return, though at roughly 3-4 MW per turbine, you'd need a lot to address global energy requirements (about 4.4 million nameplate, 12-21 million at 21-35% capacity factors). Oh, and the best sites get picked first, so output/capacity will likely fall with build-out.
● Geothermal. Absent a few global hotspots, not likely a large player. Already surprisingly well developed in Iceland, the US, Japan, New Zealand and the Philippines. Possible potential in Kenya, Indonesia and a few other regions. Proven, but limited. Enhanced geothermal, using boreholes and consumptively extracting heat has proven both expensive and disappointing to date. Total global1potential of 35 GW to 2 TW (306 TWh/yr - 18,000 WTh/yr). At the high end, that's 12% of present total energy consumption, which isn't to be sneezed at. But it's also not "unlimited potential underneath our feet".
● Tidal. This is far more limited, difficult to extract, and distributed than people seem to think. It also has massive local environmental impacts. Locally, plants with a few tens of MW of continuous capacity might be created. As storage systems, they're likely more useful than energy. Potential of 0.26% of global use per Tom Murphy.
● Nuclear. Conventional uranium resources are finite, with roughly 60 years' supply at present rates of use. Ramped to full human energy consumption, they'd be good for about six years. Suggestions are that lower-grade ores, or recovery from seawater might be possible, but this is generally unproven at scale. Likewise thorium, despite much cheerleading. Nuclear might offer potential, but it comes with numerous extreme challenges and risks, not all well or fairly discussed by proponents. .
And that's pretty much it. There are some extremely long shots:
● Fusion: still not working after all these years. First successful sustained nuclear fission was achieved within four years of theoretical understanding. Applied plants were operating within a decade, and significant commercialisation within two. Though those plants haven't been without their problems. Now, 70 years after first demonstration, nuclear proponents are assuring us they've got all the kinks worked out. I'm dubious.
● Space-based solar. Factor in several kg per kWh capacity, and launch costs of $200-$10,000 per kg. Your benefits are:
More sunlight per panel. No day/night factor, no latitudinal losses, no weather losses. Greater total area potential. Net gain is about 3 over ground-based desert siting. Half that is lost in transmission to Earth. Launch costs to geosync orbit are $20k/kg. Murphy follows a NASA study which assumes $100-200/kg costs (presently a few thousand to LEO is still quite good). At 1kW/kg, that's 17.7 megatonne of material, and $354 trillion in launch costs -- a transport bill alone, on top of engineering, fabrication, manufacture, etc. Murphy sees this as at least 4x the cost of ground-based installations.
Black holes or antimatter might offer more extreme options, but again those would be exceptionally difficult to engineer and are now entirely theoretical.
To the author: you scoot off to /r/badeconomics. Sure. OK. Economic growth cannot continue to look like it currently does forever. Gasp, some economists may not have ever pondered that particular limit case, given the importance to their everyday lives. I know that it was super thrilling experience for you ("That should not have happened!" as if the economist were a debate club champion and superhuman?) but trapping a man in a word game over dinner is does not mean you've shattered his profession - not even if he falls into your clever trap. The whole discussion has no bearing on what economics is or what economists do.
To be more concrete: what economists do involves more data, accepting that there is a use to building mathematically tractable models with limited domains of explanation even if it means some implausible assumptions in limit cases, and it certainly involves less worrying about what the world will look like 2500 years from now. I'm not saying we shouldn't care about the ultra-long-horizon future of humanity. I'm sure there are some smart people working on those issues, but econ departments are not where people do it. There are no data, no way to establish causal links that far out, no arguments that could make it into a respectable economics journal.
It doesn't have to be a main research area for all of them, but it _should_ be for some. And all economists should know the basics (physical limitations, ...) about this area of economics.
P.S. Always interested in some decent reading material about this one!
There's far too much that we don't know about the economy right now to include wild speculation, even if it's mysterious and thrilling, in the category of economic knowledge.
As a familiar example in physics, take Hooke's law: a spring will try to retract with a force that is proportional to its extension. I think everyone will agree that this is only an approximation, and as soon as you start pulling too much you'll get out of the linear regime, or if you pull long and hard enough you'll start seeing plastic effects and the spring will change shape permanently.
This is to say that having a linear model of springs or an exponential model of economic growth is very useful, but it is important to train scholars to know the limits of their approximations so that they can recognize the need to use a more complex theory.
As an admitted guy-in-the-middle, I'll add this deliberate provocation to the physicist/hard science guy: if the soft science guy dealt with simple systems as the ones you study [1], then maybe he would care about 100% abstract/infinitely-proof theories.
[1] Simple, as in, without complex organisms and interactions. Simple, as in, with less (or a finite set of) variables to account for.
As I recall, in Niven's books an alien race moved their planet away from its sun in order to deal with the problem, but I assume that wouldn't even buy us an order of magnitude in planetary energy dissipation capacity.
If not for that constraint i guess economies in western countries would have already shrunken quite a bit due to automatisation. For me the cause seems to be to a fundamental bug in our intertemporal trade system.
So this whole "discussion" is totally wrong-headed.
This whole dialogue is utterly wrong-headed, and I don't for a second believe that that economist is a real economist.
The outcome would be that people start migrating into space, where heat can be rejected more freely.