Robots, slaves, horses, and Malthus

I think this is the model that Karl Smith has in mind.

Assume robots are the same as humans. Robots can do all the work that humans can do. Robots need the same amount of energy/food to stay functioning as humans do, but robots themselves can produce that energy/food just like humans can. Robots will need maintenance and training as humans do, but robots themselves can produce that maintenance and training just like humans can. Robots can produce other robots, just as humans can produce other humans.

The only difference between robots and humans is that robots are owned by humans. Robots are just like human slaves.

Robots will earn the same wages as humans, but those wages, minus the costs of the robots' subsistence, will go to the robots' owners. Just like slaves.

Robots' owners will program their robots to build more robots, as long as the surplus of wages over subsistence is positive. Just like slave-owners will want their slaves to reproduce. And just like human workers reproduce in a Malthusian model. Given diminishing marginal product of human plus robot labour, the total population of humans plus robots will expand until the Malthusian limit is reached. At that Malthusian limit, wages are driven down to subsistence.

But unlike Malthus, who allowed that subsistence wages might be determined in part by custom and culture, the level of subsistence wages for robots would be a strict biological constant. It's the bare minimum that robots need to survive and reproduce.

At that strict Malthusian steady state, robots, like slaves, are no longer scarce. There's no point in owning either if the marginal product of their labour equals the cost of their subsistence.

In the Malthus/Ricardo model, land was the only scarce factor of production in the steady state. Land rents increased as population expanded and wages were driven down to subsistence. Land rents were the only surplus.

Time preference - a scarcity of capital in the strict (Austrian) sense - would slow down the transition to the Malthusian steady state. Robots breed robots, but doing so means future benefits are compared to present costs. The high rate of growth of the consumption of robot owners will mean those future benefits are discounted at a high rate of interest. So robot owners will limit the time that robots spend investing in new robots. But as growth in robots, and hence growth in the owners' consumption, slows as the steady state is approached, that rate of interest will decline too. Malthus/Ricardo, plus a consumption-Euler equation for the rate of interest.

Horses were once like robots. Horses could do a lot of the same work that humans could do. Humans and horses can pull things, if you feed them. But then mechanical horses, called tractors, were invented, that could pull heavier things with cheaper food. Tractors pushed horses' wages below subsistence, so the horse population declined.

The robot horse displaced horses, just as horses displaced humans from all the jobs where humans pulled things. But humans, unlike horses, can do lots of other jobs beside pulling things. Humans are very versatile. Horses can't really do anything except pull things. So humans switched to doing other jobs, while horses couldn't. And the marginal product of labour, and hence wages in those other jobs, increased. Horses and tractors were complementary factors to human labour in those other jobs.

But that won't happen if robots are invented that really are just like humans, and can do all the jobs that humans can do. Robots that are just like humans would be just like slaves, rather than like tractors and horses.

I own land.

Comments

[edeast]

Isn't there only 2 jurisdictions in the world where you can own land. A municipality in Texas, and an island off Scotland? Everywhere else is like a license from the crown, not sure what the Americans call it to square it with the revolution. So I guess we need to make sure that the robot slaves can't vote. Otherwise eminent domain over all those pesky farms.

[Nick Rowe]

edeast: dunno. My view has always been that if the government really wants something it will take it. Or just tax it, like Lloyd George and the unobtanium in Avatar.

[adjacent / q]

but i don't see why land is valuable, unless it is in short supply and useful for making more robots.
for instance if robots were easy to make, it would be very easy for someone to restrict your access to the land.
robots would presumably be very small and hard to keep off your land.
it seems like the only safe place to be would be inside a robot-controlling monopoly, and those don't exist yet.

[Frances Woolley]

Nick, terrific post. You don't own the mineral rights, b.t.w., but you can get them quite cheaply - it's been a bit of an issue recently in some areas of W. Quebec.

[edeast]

Found what I was talking about. Allodial title. http://en.wikipedia.org/wiki/Allodial_title
I first came across it a couple of years ago, following the presidential campaign of Ron Paul. After his failed nomination bid, people on the forums were casting about in despair for a libertarian utopia. I think most of the sane people left after that but the derivative movement has continued. eg The first tea parties, were publicity for the nominations. I'm not sure when it became the 'tea party'. But that's a whole lot of tangent. What we have is fee simple.

[Andrew F]

"Isn't there only 2 jurisdictions in the world where you can own land. A municipality in Texas, and an island off Scotland? Everywhere else is like a license from the crown,"

Sounds like that could end up being an important distinction. If human wages were driven below subsistence en masse, it's the difference between 'redundancy' and a Star Trek-esque utopia.

[JDUB]

Would adding a small probability (as a functiion of the stock of robots) of an Asimov-esque robot take over affect your rates of robot accumulation?

[Nick Rowe]

adjacent/q: it's not so much that land (i.e natural resources) is in short supply, but that labour (human+robot) is so abundant. So land is the only thing that's scarce.

This is just like the very old Ricardo/Malthus model, with output a function of two factors of production: land (in fixed supply) and labour (in perfectly elastic supply in the long run). Land and labour are each paid their marginal product. Given constant returns to scale, land rents plus labour wages exhaust total output. So as the labour/land ratio increases, the marginal product of labour falls and the marginal product of land rises.

Ricardo squeezed capital into the model via a wage fund, and I've squeezed it in via procreation of robots plus an Euler Equation savings/investment function. Then I added robot labour to human labour.

Thanks Frances!

I had better check out the mineral rights.

[Nick Rowe]

JDUB: My (very cursory) reading of the history of slavery says no. Each individual robot owner has an incentive to acquire more robots, because the risk of a robot revolt depends on the total number of robots, not on the number he himself owns.

[Determinant]

That "Island off Scotland" was South Uist. Unlike England and the colonies, Scotland had a true Feudal land system until 2000. The vassal in possession was the true owner (actual control of the property), but feudal dues were often payable and many modern housing estates used feudal structures as a sort of neighbourhood association for building standards, as a way to enforce a positive duty to do something. England has banned subinfeudation (a vassal having their own vassals, thus forming a chain) for centuries, Scotland didn't.

Anyway, the residents of Uist formed a community company and bought out the feudal landlord of their island.

As a further point of trivia, unceded native land in Canada is held allodially. The Crown has no right to it. In Ontario there are still two unceded reserves, the eastern end of Manitoulin Island and Walpole Island near Sarnia.

[RSJ]

Hmm, but do you really *believe* all this teleology? Wasn't it just inserted ad hoc as a type of mathematical convenience? Sort of like a physicist worrying that we'll never be able to stay in one place as humans are spheres.

For example, when land is expensive, we can build up, or we can invent transportation to take us to where land is cheaper. Is the factor of production "land" or is it "office space", or is it just "means to coordinate workers"?

Here's a study arguing that there was basically no commercial real estate price appreciation in Manhattan from 1899 to 1999, and they also cite another study arguing for no real price appreciation from 1835 to 1900. But there are long (e.g. multi-decade) boom/bust periods.

//web.mit.edu/CRE/research/papers/WP90wheatonbaranski.pdf

[Nick Rowe]

RSJ: "Hmm, but do you really *believe* all this teleology? Wasn't it just inserted ad hoc as a type of mathematical convenience? Sort of like a physicist worrying that we'll never be able to stay in one place as humans are spheres."

I'm not at all sure what you mean by that. But Malthus and Ricardo certainly did believe that land was the ultimate scarce factor. They didn't just bung it into their models as an ad hoc mathematical convenience. Population pressures on natural resources were absolutely central.

Now, whether technology to use land more efficiently can continue to improve quickly enough to offset those population pressures has been the $64 trillion question these last 200 years. The Greenies say "no", and even though they've been wrong so far, they might be right or wrong in future. But the Hansonian "singularity" and self-replicating robots would sure throw some more fat into that old fire.

[Vladimir]

Nick,

An excellent post! I've seen this issue discussed on many occasions, including on economics blogs, but I've never read a treatment that gets to the heart of the matter as accurately and poignantly as the conclusion of your post.

One fallacy I frequently see in discussions of this sort is the invocation of comparative advantage. Supposedly, even if robots are more productive than humans across the board, humans will still be able to practice whatever their comparative advantage will be. Trouble is, a correct refutation of this argument is also applicable in a similar form to some of the popular arguments often heard in the context of contemporary political economy. From what I've seen, the unwillingness of many people to accept critiques of comparative advantage theories in this latter context also makes them fallaciously optimistic when it comes to the robot scenario.

[Blackadder]

My problem with this line of argument is that as we have gotten closer to the robot world real wages have gone up, not down.

[Steve T]

I think this discussion misses the point.

Ownership of robots used in production will be highly concentrated. Most robots will be owned by corporations and a wealthy elite. The result will be substantial structural unemployment and low wages for the vast majority of humans that now rely on jobs.

And the result of that will be a massive drop in consumer spending and consumer confidence. The point is that robots take jobs and incomes but do not act as "end consumers" in the economy. Robots do consume energy and resources but these are inputs to production. Without end consumers to drive the economy, the result will be a deflationary spiral. Many of the robots would be shut down. Imagine production robots used by General Motors. If there are no humans with income and confidence to buy cars, the robots will be shut down.

Only people and governments act as end consumers in the economy. This drives ALL economic activity. Businesses purchase inputs to production and they also invest, but business investment is based on anticipated consumer spending (ie. Apple invests to create an iPad because it expects millions of consumers to want and be able to pay for an iPad in the future.) Businesses are NOT end consumers.

You cannot get around this problem unless you turn the robots into real consumers. To do that they would have to be sentient and have needs and desires and the ability to purchase products and services -- and they would have to be free. That implies a very high level of technology, and before we get there, we would have less advanced robots capable of doing all the work, but not capable of acting as end consumers. Hence the problem.

For an in depth discussion of this issue, please check out this book:

The Lights in the Tunnel: Automation, Accelerating Technology and the Economy of the Future.

http://www.amazon.com/Lights-Tunnel-Automation-Accelerating-Technology/dp/1448659817

there is also a free PDF at http://www.thelightsinthetunnel.com
and the author has a blog at http://econfuture.wordpress.com

[Lord]

There could be other scenarios though. Prohibiting slavery, leaving them to themselves, where they would support some level of custom and culture, or individuals reproducing themselves as robots towards the end of their lives to live on, willing their property to them, displacing children whether in wills or in fact, or retreat into the virtual world where most would cognitate and earn their incomes without ever instantiating into a robot or only doing so for a specific task or period. It would also matter whether they were longer lived robots (i, robot) or shorter lived synthetics (blade runner).

[Vladimir]

Steve T,

Ford's argument was addressed by Robin Hanson a while ago:
http://www.overcomingbias.com/2009/10/take-both-econ-tech-seriously.html

As Hanson says, "[H]igh income inequality might be ethically bad, and threaten political instability, but it does not at all threaten economic collapse – producers can focus on giving the rich what they want, and innovation and growth is just as feasible for elite products as for mass products."

[Steve T]

Vladimir,

I can't think of too many major industries that are not mass market industries: autos, computers, retail banking, media/advertising, Google (adwords), Facebook/Social Media....they all depend on millions (or tens or even hundreds of millions) of potential customers. Destroying all that and replacing it with an entirely new set of industries that focus entirely on the rich seems a stretch to me. Also everyone would default on their mortage and that would likely bring down the finanical system.

See Ford's response to Hanson here:

http://econfuture.wordpress.com/2009/10/29/why-i-think-robin-hanson-is-wrong-about-%E2%80%9Ceconomic-growth-given-machine-intelligence%E2%80%9D/

and a related post here:

http://econfuture.wordpress.com/2009/10/26/the-economic-implications-of-intelligent-machines/

[Jim Rootham]

It's a little nastier (or better, depending on how you take it). There is evidence that equality makes everyone better off.

[Vladimir]

Steve T,

I agree that a dramatic fall in employment and wages due to automation could mean political instability. (Hanson doesn't dispute that either.) But it seems to me that Ford makes a much stronger argument, namely that the existence of an expanding high-tech economy somehow inherently depends on great masses of affluent consumers. That surely isn't true. A shift from mass-market to elite-market consumerism would represent a huge episode of "creative destruction," and like any other huge social transition it might have unpredictable political consequences, but there is nothing impossible or contradictory about the ultimate prospect of elite-oriented production.

In my opinion, even the prospect of serious political instability is uncertain. History offers many examples of societies in which great masses of people were at some point impoverished to the point of mass starvation -- just look at the Malthusian famines that regularly struck all parts of the world until modern times, or at the historically recent famines caused by communism and/or war. What these examples show is that as long as the elites have some minimal level of self-confidence and competence, ordinary people won't rebel and will just starve to death peacefully. Even in those rare cases when a mass rebellion does happen, it never poses any real threat to the established order, unless the elites are disastrously incompetent or significant parts of them decide to side with it for some reason.

[Steve T]

Vladimir:

"What these examples show is that as long as the elites have some minimal level of self-confidence and competence, ordinary people won't rebel and will just starve to death peacefully."

Wow, ok. I'll stop worrying then... However, I don't think I want to be the one who has to explain to the Tea Party folkes that the "liberty" they are fighting for is the liberty to starve.

[Vladimir]

Steve T,

I'm not saying that we should not worry -- on the contrary! I'm just arguing that a transition from the present high-wage, mass-consumerism economy to an unemployable-masses, elite-consumerism economy would not mean an end to economic expansion and innovation, and it would not necessarily mean cataclysmic political instability either. I'm definitely not claiming that things would look pretty!

Now, during the past Malthusian famines, even the most benevolent elites had no alternative but to watch the masses starve; a bad harvest in a Malthusian economy, let alone several successive ones, meant that there was simply not enough food to feed everyone. In contrast, the elites in a robot-based economy would have the option to keep the unemployable masses alive out of charity. Feeding and clothing them certainly won't be a problem; even nowadays, the cost of non-fancy but adequate food and clothing is minuscule. The greatest problem will be that humans require an awful lot of space to subsist tolerably, so that the opportunity cost of providing living space the unemployable poor will be very large. (There is also the problem that people living off handouts and lacking meaningful employment tend to develop all sorts of social pathologies, but that's a whole complex topic in its own right.)

[K]

I made this point about automation and land value in a comment here a few months ago. As then, I will take the opportunity to point out that the solution is to shift taxation away from income and towards land (as proposed by Smith, Ricardo, George and countless others). It is truly bizarre that when we have known about this perfectly efficient revenue source for ages we persist in taxing income and consumption when land is incompletely taxed. The other part of the solution is the citizen's dividend, I.e. a shift towards recognition of a natural claim of all citizens on the fruits of the earth rather than just on the fruits of their labour.

[RSJ]

"But Malthus and Ricardo certainly did believe that land was the ultimate scarce factor."

Yes! But they had no real interest in understanding production or trying to model it. They were in that tradition of economics that cares more about understanding the process by which the produced good is assigned a market price, rather than the process by which the good is produced -- i.e. auction-centered economics, rather than technique centered-economics. This was during the industrial revolution, in which England, with a poor endowment of land, was exploding in increasing returns industries that were capital intensive. But to me, the ultimate scarce factor is human innovation, and this will drive everything else. Human capital is more valuable than land.

[RSJ]

P.S. This chart may be of interest:

http://3.bp.blogspot.com/_fevQMK7kLEI/S62M-TiWAyI/AAAAAAAAAL0/94Ap5ZKl-Ws/s1600/Land_Share.png

I was once more interested in r.e. prices, and wanted to see whether land in constrained areas (e.g. san francisco) was a greater proportion of the house price than land in unconstrained areas (e.g. phoenix). It is, but over time, that proportion doesn't trend up. You do not see share of land --> 1 in any market. If the share of land is a measure of the inelasticity of demand for that particular location, so you would not expect it to go up over time unless the location was becoming more and more desirable relative to the unconstrained areas. This is another reason why house prices in desirable areas should not be infinite. And there is plenty of land -- you could put the whole world into Texas, if it was willing to move there, and population growth is slowing.

[Nick Rowe]

Vladimir: I've been giving a lot of thought to what you said about comparative and absolute advantage. Here's what I've come up with:

Let there be two countries: "Land" and "Labour", which differ in factor endowments. They trade for standard comparative advantage reasons. Then a third country, "Robots" appears out of the sea, which is exactly like Labour, except it keeps growing and growing. Robots will also trade with Land, for standard comparative advantage reasons. But the appearance and growth of Robots will worsen the terms of trade between Labour and Land.

[Nick Rowe]

Why was Malthus/Ricardo wrong? Why has land not become generally more scarce over the last 200 years? I think there are two reasons:

1. Technology improved, so we could use existing land more efficiently. (It was "land-augmenting" technical change.)

2. The population of workers did not continue to grow as Malthus predicted it would. Real wages increased, but we don't all have 20 surviving kids.

Robots would knock out the second reason. Because robots *would* all have 20 surviving kids. Their owners would tell them to.

[Nick Rowe]

I agree generally with all of Vladimir's comments above. In particular:

1. There is no reason to believe that Robots will cause a deficiency of aggregate demand. The owners of robots will want to spend the income earned by robots just as much as anyone will want to spend income.

2. Robots will increase output and hence income per person. That is not the problem. But they will reduce the marginal product of labour, and hence real wages. Which is a problem, from the point of view of the distribution of income. In that respect, as Vladimir notes, the robot-Malthusian problem is very different from the old Malthusian problem. The robot-Malthusian problem can, in principle, be handled by redistribution. The old Malthusian problem was low output per human. The robot-Malthusian problem is high output per human, but low marginal product of labour.

[Jim Rootham]

Don't we have a current example of this running now? I am referring to the petro states. How do their economies behave?

I know they have social pathologies, and I don't see what I would consider a sound economy from my observations from a great distance.

[Vladimir]

Nick,

Regarding the classic Malthusian vs. robot-Malthusian scenario, the critical assumption is that robots are treated as passive and obedient non-persons. If you care for the robots too, then the robot-Malthusian outcome is every bit as bad -- robots would toil for bare subsistence, and a recession would instantly push masses of them below subsistence, just like a bad harvest would mean mass starvation centuries ago. Moreover, it is almost certain that robots matching human intelligence would technologically advance far beyond in a fairly short time -- and while human elites normally have little difficulty keeping human masses obedient even in awful misery, lording over much smarter creatures would be a harder challenge.

[Vladimir]

Nick,

Regarding comparative advantage, I agree with your example. The trouble is that this scenario also has unpleasant implications for the contemporary discussions of free trade, since it clearly illustrates that despite comparative advantage, the emergence of trade with a new party can leave an arbitrarily high percentage of actors worse off to an arbitrary extent in the new equilibrium. (In an extreme case such as the robot scenario, perhaps even pushing them below subsistence.)

This effectively refutes the often seen knee-jerk invocations of comparative advantage as a supposedly universal and conclusive argument in favor of trade liberalization. Now of course, it may well be that such pessimistic scenarios are not realistic in the present situation, but the very logical possibility of such scenarios means that simply brushing off protectionist arguments by invoking comparative advantage, as is often done, is not a valid argument, and that instead a more complex analysis is necessary to prove anything in each particular case. From what I've observed in various internet discussions, many economically literate people find this ideologically troublesome and are thus unwilling to really think these things through, which in turn makes it hard for them to understand the gloomy robot-Malthusian scenario. That's why I've really enjoyed your no-nonsense analysis of the issue.

[K]

I like your analogy, Nick. But the citizens of Land had better start building fences cause the citizens of Labour will be gunning for them - if the citizens of Robot don't get them first.  If we want to avoid that, we will need to reconsider the moral foundations of economics (as well as Asimov's Laws of Robotics), but as that doesn't seem to be in the cards for now, I too will be staying long land, betting that governments will maintain my fences for me and keep out the riffraff.

[K]

Vladimir:  I'm not sure I follow the consequences of your analogy.  Are you saying that

1) Developed world workers are disadvantaged by trade in greater numbers than developing world workers are advantaged; or

2) That in a deep recession, the developing world workers (the "Robots") would suffer in enormous numbers - more so than they would have been suffering in a world without liberalized trade; or

3) Despite the enormous gains from trade with the "robots", the gains aren't redistributed to the citizens of Labour, and therefore only benefit a tiny minority of actors?

I think the first two are wrong, and the third is not new.  It is also a problem of redistribution rather than a problem of trade.

[Vladimir]

K,

Basically, (3) is the closest to what I had in mind, with the caveat that I agree with the standard free-trader argument that "Robots" (i.e. poor country workers) benefit a lot, since unlike the hypothetical future robots, in the present-day world they're not doomed to a Malthusian catastrophe. The problem is that commonly seen comparative advantage arguments are often framed with an implicit (and sometimes even explicit) claim that *everyone* necessarily benefits by removing trade barriers except for nefarious rent-seeking special interests -- whereas in fact, there is at least a theoretical possibility that the losers will include a very large percentage of "Labor."

I'm not claiming that this is an original insight, but it does seem to me that this argument is often not given due consideration for ideological reasons. This is of course a politically charged issue, which I'm brining up only because I think that people often carry over their prejudices from it into discussions of the robot issue. Basically, the over-simplified "more trade is always good" principle leads them to fallaciously conclude that we can only be all better off if we can also trade with a bunch of robots.

(As for the first two items, I think (1) is definitely wrong, and (2) implausible, though not outside the realm of the possible. Also, while comparative advantage doesn't imply that everyone is a winner, it does imply that the total gains of the winners exceed the total losses of the losers, so yes, you can frame it as a redistribution problem.)

[K]

I understand your point and I see how Nick's analogy makes it more explicit:  by placing Labour and Land in separate countries it becomes explicit that what benefits one has zero benefit for the other.  It makes it impossible to conflate capital and labour using what, you rightfully point out, is a standard neoliberal sleight of hand.

[RSJ]

I don't understand the trade happening in this example.

Robots in robot-land export products (e.g. consumption goods) to Labour land.

What does labor-land send to Robot-land in return? At first, I thought it was the energy/maintenance needed to keep the robots operational, but then I realized that the robots can produce this themselves. What goes in the other direction?

[K]

RSJ:  I don't think that trade axis is very important in the model.  I think Nick's approximation is that neither Robot nor Labour have any comparative advantage over the other so they wouldn't need to trade.  But they are both subject to the same Malthusian catastrophe as the number of robots grows to infinity.

In reality, of course, they each have some comparative advantage (and so have significant trade where labour and automation are complementary - e.g. factories need to be operated) but Robot is constantly gaining on Labour. At the point of the singularity, as you point out, Labour will cease to have any comparative advantage and so will starve.

So Nick's model is really at the threshold of the singularity where human and robot skills are exactly equal, at which point the number of robots would quite suddenly explode, limited only by the resources of Land.

[Nick Rowe]

Yep. My model, which assumes that robots are identical to labour, is a very extreme case. There are no gains from trade between labour and robots. This case is a worst case scenario for people who only have labour to sell.

Of course, you could argue that it's an unrealistic case as well. We tend to try to build robots that are good at doing the things humans are very bad at doing, or really don't like doing.

[MyName]

I don't think this approach makes very much sense for a number of reasons. First, you assume that the thing that robots need to recreate themselves is cheap/plentiful but this doesn't necessarily hold true. Humans can be made essentially out of land (using solar powered crops) and time. It requires much more energy to create even the simplest electronic components. And, even assuming that AI is feasible, there's no reason to believe that it may not take as much time, relatively speaking, for a robot to learn what it needs to learn to develop as it does for the human brain to fully develop. If 15 or so years is the best nature can do after millions of years, maybe there's a lower limit on this sort of thing.

You also assume that robots have the same static goals as humans and that they won't evolve into something approaching a natural ecology. Robots are more likely to compete with other robots than they are with humans as they both need the same things, while humans don't. Even if, given enough replacement parts, you are functionally immortal, there will still be a competitive advantage for newer/faster robots to develop that can reproduce by predation and scrapping of other robots, so there will still be competition over resources needed by other robots. In which case, I think even the robot future is more interesting and different than you're talking about. And the future for humans is dependent upon which resources we need and how much overlap there is with what the robots need. Not too different from how humans deal with the billions of other humans on the planet.

[edeast]

Thanks Determinant.

[Bill Woolsey]

When reading this, I wanted to know what is happening to consumption per human.

I guess it is increasing all the time, it is just that the only way to get any consumption is by owning land. If you owned only robots, then your return goes to zero and you get now consumer goods. If you don't own anything and just work for a living, you get nothing (or just repair and depreciation on robots, which is assumed to be equal to enough for a person to stay alive and raise two children?)

Anyway, isn't the reason why things don't really turn out this way is improving technology? We get more consumer goods per worker (and robot?) Or does this amount to new ways to save on land?

For example, suppose cheap intersteller travel makes land cheap. What happens to consumer goods per worker?

Ignoring that, suppose I have one child and use the savings to invest in land? My children do the same? What happens to our consumption per person?

[Nick Rowe]

Bill: If the supply of land were to increase in proportion to the supply of humans+robots, and assuming constant returns to scale, real wages per worker (or per robot) would not fall. But if robots are exactly like workers, and not at all like land, then the invention of robots increases the effective labour/land ratio, and reduces the MP of labour while increasing the MP of land.

Total output and income per human rises, but all the extra income, and then some (as income is redistributed) goes to owners of land.

"Ignoring that, suppose I have one child and use the savings to invest in land? My children do the same? What happens to our consumption per person?"

Unless you can forecast better than the market, or save more than the representative agent, you won't be able to do any better by saving and investing in land. The price of land will already have risen so much you won't want to do this. It's already too late!

[Peter T]

Nice thought experiment. Has a few actual parallels.

I assume by "land" you mean all limited resources (land, nitrogen turnover, CO2, water etc)? If so, robots and humans are in direct competition at the species level, and the final balance between the two depends on their relative degrees of fitness - a matter of endless details.

At the individual level, more robots is clearly good. But this advantage disappears at the group level (a human group with fewer robots will be less rich than a competing group with more, but will have more human resources). Same within the group - the issues are how many robots at the group level, and who owns them within the group. If the robots are autonomous, then they are for all practical purposes humans, and the competition is at the species level.

For parallels, have a look a the Mamluks in Egypt, or similar "slave soldier" societies. The Mamluk elite are the humans, the quiescent local populations the robots. Keeping the elite small (a high robot/human ratio) worked well for the elites, UNTIL these societies were challenged by groups who could field many more human resources.

For the within-group dynamics, see the anthropologist Marvin Harris on pigs and people in the Middle East - his thesis is that the individual advantage of pig-keeping (cheap protein) was balanced against the group disadvantage (fewer people) by religious prohibition.

[Tim Tyler]

Re: "Assume robots are the same as humans."

A situation which is very unlikely in the first place - and extremely unlikely to last for very long if it somehow magically happened. Machines already vastly exceed human capabilities in many areas.

This essay stops before things get interesting. What will politicians do if they have a big mountain of unemployed human voters to feed?