The above graph is an amusing example of futurologists falling in love with the idea of exponential growth. It was reprinted by historian Sean Manning in a long essay in the form of a blog post, on the difference between how historians and economists think about economic history.
You’ll have to read the whole thing–also it’s written as a review of a book, “Slouching Toward Utopia,” by economist Brad DeLong, that I haven’t read–but here’s Manning:
Slouching Towards Utopia is . . . a big ideas book, or as DeLong prefers to call it a grand narrative. . . . Almost any paragraph of this book poses two or three fun subjects for debate. But this week, I [Manning] will put most of them aside and show how this book tells one story while the author wants to tell a different one.
A Grand Narrative
Slouching speaks of things which were, and things which are, and things which will be. . . . DeLong is an economist, and economists love to model things as exponential growth. So DeLong makes up numbers for world population and GDP in 6000 BCE, 1000 BCE, 1 BCE, and 1500 CE . . .
Slouching ends in the recent past, but it is optimistic about the future. Incomes are no longer increasing as fast in the rich countries, but much of the world is still poor, and prosperity in the rich countries which grows at 0.5% a year instead of 2% a year is still growth. “There is every reason to believe prosperity will continue to grow at an exponential rate in the centuries to come” (p. 11)
This is a grand narrative about incomes and population and exponential growth, three of economists’ favourite things. It is a narrative where before 1870 population grows slowly and prosperity barely grows at all, then from 1870 to 2010 both grow rapidly, and after 2010 prosperity is likely to continue to grow while populations stabilize or shrink. Economists love to tell grand narratives about trends, even though they know why historians have trouble believing in them. I have not written a book with a grand narrative. But I can summon the old magic of Max Weber and use the facts in this book to tell a different story.
A Counter-Narrative of What Was
I [Manning] don’t see much value in estimating the population of the world in 6000 BCE when we can’t agree on the population of the Americas in 1492 within a factor of 5 [Manning’s original post said “a factor of 20” here, but after looking back at the literature he changed this to a factor of 5. — AG], and took decades to agree on the population of the Roman empire under Augustus within a factor of two. But we do have data on population and prices in specific places such as England since Domesday Book or Achaemenid Babylonia. These show cycles of rising and falling, with rising population tending to correspond to falling wages and vice versa. In the best times, an ordinary worker or peasant might consume twice as much as the worst times.
I think most people could agree that populations tended to slowly increase, particularly when new crops or methods of farming were introduced. . . . But its not clear that standards of living trended upwards. . . . data before the 19th century shows standards of living rising and falling in cycles, not stable with a gradual upwards trend. . . .
In past societies with great material wealth, such as the Roman empire, its not clear to me that the health benefits from tiled roofs or plastered aqueducts outpaced the health dangers of crowding people into cities and moving them back and forth. Height is a good proxy for health and nutrition, and height in Europe also goes up and down before the 20th century, rather than having a clear trend. Notably, it is higher in the material-poor early middle ages than the material-rich Roman empire or nineteenth century.
A Counter-Narrative of What is and What Will Be
Beginning in the 19th century, population and welfare begin to rise at the same time across much of the world. But countries which tried to imitate the richest countries often failed, and growth in any one country often stalled. . . . Economists can calculate, but physicists can too. Colleagues like University of California physics professor Tom Murphy have done the math . . . Economic growth that is decoupled from energy consumption won’t look much like GDP, and there are only so many hours in the day to enjoy our futuristic entertainment centers or teleport to our beach huts in Maui and stroll on the beach. Even if we avoid ecological catastrophe, a devastating plague, exhaustion of resources, and the replacement of humans-as-we-know them with something else, growth in prosperity is likely to end within the next century or two because of physical limits.
Manning concludes:
As a historian, I don’t think that exponential growth is the natural state of things. I think in terms of cycles and tradeoffs and phase changes. The simultaneous rise in health, wealth, and population in the long 20th century was a new thing in world history, and much of the world is transitioning into a new state. I think that looking at the past with a mind trained on the long 20th century is a mistake, and so is imagining the future that way. . . .
I think that [De Long’s] book would have been better if it presented exponential growth as typical of the long 20th century, not as the way economies usually behave.
Interesting. Slouching Toward Utopia also has a blurb from Paul Krugman, so maybe he and De Long can respond in comments.
The issue of exponential growth in the economy came up last year in our post, How to think about the claim by Justin Wolfers that “the income of the average American will double approximately every 39 years”?, where economist Wolfers shared this graph:
And here are some earlier posts regarding historians’ efforts to acknowledge historical uncertainty in social science and journalism:
History, historians, and causality
Uncertainties regarding historical facts
Decorative statistics and historical records
Uncertainties regarding historical facts
P.S. Manning’s counter-narratives can be given a “left” or “right” political spin. From the left, Manning’s arguments push against the right-leaning view that, whatever happens, the existing system of capitalism will bail us out: maybe a steady state of economic expansion will not be expected even in the absence of government interference. From the right, Manning’s arguments push against the left-leaning view in which gradually increasing prosperity is a given, and so our society can afford economic redistribution.
To put it another way, both the left and the right have contradictions in their views of the economy. Economic leftists oscillate between saying that the capitalist system is failing and needs to be replaced before we all choke on our own filth, and saying that, as the richest society in the world, we should be able to afford to massively reduce poverty and inequality. On the other side, economic rightists are torn between viewing capitalism as an unstoppable force (the resonant phrase is “there is no alternative”) and the idea that it is a fragile golden goose whose supply of eggs is endangered by any reforms.
In saying this, I’m not trying to “both-sides” the debate or imply that leftist and rightist views on economics are equally correct–it would be more accurate to say that I have various policy preferences on specific economic issues but I don’t have any coherent view of macroeconomics myself. As a political scientist, I’m interested in the arguments used by different sides in these debates, and I’m interested in how various ideas and talking points can implictly depend on the understanding of historical claims that themselves are in legitimate dispute.
P.P.S. More from Manning here, where he lists some factual errors or dubious claims in Slouching Towards Utopia, a book which Manning seems to like overall.
I think this is a mischaracterization: “economists love to model things as exponential growth.” I think a more accurate statement is that economists are inconsistent in their views about exponential growth, and it is a serious inconsistency. One thing economists frequently complain about is that non-economists ignore principles such as diminishing marginal returns which argue against continued linear, not to mention exponential, growth. The old book “The Limits to Growth” was roundly criticized by economists for its failure to properly account for market mechanisms which would make future trends deviate from the past (for example, as fossil fuels become scarce their prices rise, and this brings about a number of market responses that make continued growth in demand unlikely.
I am not trying to defend economists here – but characterizing them as believing in economic growth misses the more important point, in my view, that they rely on exponential growth in some matters and not in others. I have long found the underlying belief system of economists either lacking or distasteful. Income can continue to grow exponentially, but environmental damage somehow will not. To me, these inconsistencies are a more accurate critique.
I’m not sure I see the inconsistency.
The raw material of growth is the stock of stuff. Thus, exponential growth is more natural to think about that arithmetic growth, in which the stock is immaterial. The raw material of pollution and resource exhaustion is also the stock of stuff, and its effects on growth are negatively exponential. These two thngs alone will allow all sorts of observed behavior, including cycles of growth and shrinkage in material mellbeing. But anything that depends on the stock of stuff is, one way or another, inherently exponential (if not in the long run).
Innovative leaps are not exponential (this is controversial among economists, but I think it is correct) although their diffusion may be exponential early on and negative exponential late, with inflection points which depend on all sorts of things. Some sort of logistic-like diffusion curve is in common use here by economists (and others) and modifies the “everything is exponential” narrative.
So I confess to a fondness for exponentials of various sorts, with the understanding that uniform exponentials are sometimes countenanced (though not by me) as an exception to one of the most cherished of all economics dictums: Stein’s Law — “If a thing can’t go on forever, it won’t” Stein’s Law has no exceptions.
Moore’s law is exponential, and performed spectacularly well for a long period of time. To put it another way: “All exponential growth rules are wrong. Some are useful.”
As Tom Murphy pointed out, if we extrapolate energy growth rates from Shakespeare to now and then now forward the same amount, we’ve done something ridiculous. Im not sure which thing would have to look back at the blog posts but maybe we’ve made the surface of the earth above the boiling point of water, or are using the entire output of the sun. If I remember, a 1000 years or so and we are using the entire output of all the galaxies stars. We have documents from 1000 years ago, its well within human cultural memory.
https://ourworldindata.org/energy-production-consumption
Suggests since 1990 we’ve had global rates of growth around 2% dropping a lot from the 6% ish of 1960. I’d be surprised if this 2% continues 50 years. With climate issues more and more of it will be used just to maintain the stock of physical capital as flooding and fire and ice storms and things become more common. So energy growth doesn’t even mean improved physical conditions necessarily. Within my lifetime I expect global GDP to reach peak and decline.
Dale, I am an ancient historian who never took coursework in modern economics so I don’t want to go too far into criticizing mainstream economics or unweaving the differences between an economics conference, Economics 101, and the business section of my local paper (economist John Quiggin has a book that does the later). But I see this lust to model populations and economies as exponential growth all over writing on economic and demographic history, and the difference between those models and specific case studies where sometimes population is roughly static then someone introduces paddy farming and it rapidly grows and then there is a pandemic and it falls.
What I’m trying to point out is that you don’t see exponential growth in environmental damage all over economics writing. Income can somehow rise exponentially, but carbon and habitat loss will not. There is a hidden set of values displayed in these models.
Dale, to me it seems more like Economics 101 tries not to think about some aspects of production and consumption (eg. health or topsoil). They get covered in passing, and eccentric economists like Diane Coyle give them close attention, but are not central to discussions of economics like buying and selling are. Back in 1987 the paper “Sex and Death in the Rational World of Defense Intellectuals” criticized what aspects of nuclear war a “nuclear strategist” was allowed to talk about.
Tom Murphys blog just absolutely devastates the continued exponential growth view of economists https://dothemath.ucsd.edu/2012/04/economist-meets-physicist/
Here’s the thing, you can take the view that while exponential growth can’t continue forever, its still got maybe 100 to 200 years… Or you can take the view that measured by more reasonable measures the US has been in degrowth for 15-30 years already and the world is facing flat or degrowth at any moment. For example Chinas population has declined several years.
Blair Fix shows a pattern of energy usage that he says is indicative of a life cycle of empires https://economicsfromthetopdown.com/2025/05/04/the-half-life-of-empire/
By his measure the British empire hit peak in 1900 or so and the US empire hit peak in 1950 or so. His half-life concept ended for Britain in 1950 and for the US in about 2000.
Here’s the thing. From a physical perspective we dont have too many fossil fuel filled empires left in our world. We haven’t run out of gas and oil, but we’ve pretty well run out of ways to handle the pollution they cause. Wildfires in CA in Jan nearly burned my home to the ground and did burn 10,000 structures for tens of Billions in losses. Just this week massive flooding in NJ and NY. Massive fires across Canada a few years ago. Crop failures are more common. A heat dome over Portland OR started killing people a couple years ago. Also a cold storm a year or so ago had my sister huddling in sleeping bags with power out and indoor temperatures at 35F for multiple days a few years ago. Texas had similar situations.
We simply aren’t gonna get 200 more years of fossil fuel growth. The real question is what will we get? The Limits To Growth people got mostly sharp collapse in their model. If they actively tried to stabilize many things they could get a soft landing, but if anything we are destabilizing things more and more with Russia invading Europe, resparking wars in the Middle East, and saber rattling surrounding Taiwan.
Fossil fuel use exponential growth will end. Likely in maybe 50 years or less. Solar can maybe slow the decline, but is no panacea for a world addicted to 3%/yr increase. The absence of that growth will end many other types of growth. The wildfires and flooding and things will wipe out much physical capital of society. Food production will suffer. Population will decline probably rather more rapidly than people would like. All of this is essentially physically inevitable in some form. That a guy writing today from an economist perspective can write a book about centuries of exponential growth with a straight face is evidence that we aren’t likely to experience the controlled soft landing of the ideal Limits To Growth scenario.
Its all rather sad honestly. The Pollyanna view of Economics is definitely not helping.
Murphy’s book, “Energy and Human Ambitions on a Finite Planet,” is interesting, and I use parts of it for my Renewable Energy course, which coincidentally I just posted a recap of (https://eighteenthelephant.com/2025/07/18/course-recap-physics-of-solar-and-renewable-energies-spring-2025/). However, though the book overall is great, I’m not fond of the first third, Murphy’s railing against rosy pictures of growth, which are as simplistic as the “exponential forever” views but in the opposite way. There’s a good review of the book here: https://pubs.aip.org/aapt/ajp/article/89/9/897/593796/Energy-and-Human-Ambitions-on-a-Finite-Planet .
Thanks for the link to the review. I’ve been reading Murphy’s blog a lot lately and I really like it. He is smart and his writing is clear and compelling, so good to know that other physicists disagree. The points raised in the review seem good ones.
I haven’t read the book, only the blog. So I don’t know about the first third. I can say that the review seems to miss the point as far as my memory of what the blog says. From the review:
“The result is an apparent reductio ad absurdum: that a worldwide demographic transition cannot be completed because it would require roughly a “factor of ten” increase in energy use.”
But that’s exactly the point of those calculations as I remember them on the blog at least. when the reviewer then says “If we assume that the world’s per-capita energy use in 2100 will be the same as that of the U.K. today (less than half that of the U.S. and typical for much of Europe), we obtain an estimated total world energy demand in 2100 of a little over double—not ten times—what it is now.”
It’s just proving Murphy’s point, which is more or less that growth at current rates stops, it stops below the level of the most developed world today, and that even that level applied across the board is unlikely to be achievable in a sustainable way.
How long can we handle a constant population at 2x current energy per capita without a catastrophe? Not long. Esp. since the catastrophe is already underway. My own neighborhood burned to the ground in a disaster larger than any in CA history even after inflation (at least that’s what I read somewhere). A few years ago there was another wildfire disaster of epic proportions in CA as well. People in Texas got like $10k monthly electric bills a few years back. Without that electricity they’d have frozen to death. Flooding right now all over the eastern seaboard. People still recovering from the flooding in the winter. My sister huddled in the dark at 35F inside her house for days a few years ago in Portland OR while electricity trucks were prevented from fixing the power due to slick ice conditions. Heatwaves, crop failures, all of it are already here and likely will get worse.
The big problem isn’t physics though, it’s politics. Physics just shows us that there are limits and we’ve come close to hitting them already, and the upper limit likely declines with time and so we need to come to live in a condition below what those of us in the US are accustomed to.
It could still be better than the condition of Kenyan dirt farmers or Bangladeshi click-farmers chained to a board full of iPhones for 12 hours a day or whatever, but in addition to the physical constraints, we also have the political wealth-inequality imposed constraints. The richest 1% own 35% of all assets worldwide according to OurWorldInData. Even within developed countries like US, it’s like a fractal, the richest 1% own 36% of US assets.
That makes them able to say how the rest of us experience life. And it also means the share going to the rest of us declines faster than the share going to those few rich. If there’s overall growth, then we can say this means the rest of us have to wait longer for the growth to come to us. Maybe we experience 2% growth and the rich experience 6% but we still have that 2% and get doubly better off in 35 years! But if there’s constant or declining consumption, then we deteriorate to serfs unless the system for distribution changes.
Tom Murphys’ ultimate thesis, in my opinion, is that physics tells us the economic system we have today can’t continue. Hopefully we will invent something new without running a giant world war and burning up the last possibility we have to invest in more renewables etc. If we’d done that investment in 2001 instead of running 20 years of middle eastern wars… we’d be a lot better off.
Raghu, thanks for the pointer to the review, I would add that Murphy seems to have gone in a “green doomer” direction like John Michael Greer since he wrote his classic essays. I was really struck that DeLong does not take seriously any future for earth except more everything forever (or more prosperity for several centuries), and that seems to be emotional rather than his professional judgement. He is worried about climate change and curious about LLMs but does not seem like a transhumanist and certainly not a degrowther.
My uncle, an economist for the Pentagon, used to say that if trends had continued everyone should now own five Chevrolet cars.
As a retired geographer who thinks that people are much better at rationalizing than at being rational, I take a pessimistic view. I learned the basic physics of human-caused climate change in 1971, and since at time I thought that people were more rational than I do now, I thought governments would take effective action to address it. Now, we have an administration that claims climate change is a hoax, and is undoing the efforts of the Biden Administration to address it. Climate change is already driving human migration across boarders (from droughts and just being too damn hot), and that is giving a boost to xenophobic dictators. That is just going to get worse, and, some of those people are going to have nuclear weapons. And, for all our technology, we still depend on plants for food, plants depend on water, and climate change is making fields effectively drier (more evaporation). And don’t get me started about disease. Who knows when we are going to get a pandemic with really high mortality rates (could be H5N1)? Etc., etc., grumble, grumble.
I agree that exponential growth isn’t going to be sustained over the long term, but I am an optimist about the future. Specifically, I imagine that in a thousand years there will be a billion humans on this planet living in harmony with nature and each other. They will have technologies that we would mistake for magic and their songs and legends will celebrate our struggles and achievements.
The main reason for this is that I see population peaking this century (at around 10 billion) and then declining by about a billion every century until it stabilizes.
I am always curious to see how other people draw the population curve from AD 1000 to 3000, given their worst and best case policy scenarios. What do they worry will happen? What do they hope could happen instead?
P.S. I’m reading Henry George.
I thought the thesis of “Slouching to Utopia” was that exponential growth specifically did not kick in until around the 1870s, with previous periods alternating between growth, stagnation, and decline.
From Noah Smith’s review:
“Why did the world become rich? DeLong argues that only after 1870 did technological progress accelerate to the point where it managed to outstrip human population growth, thus freeing humanity from Malthusian constraints.”
So this is entirely consistent with “data before the 19th century shows standards of living rising and falling in cycles, not stable with a gradual upwards trend”.
I don’t have my copy of Slouching Towards Utopia with me to check, but pages 32 and elsewhere postulate tiny exponential growth in global income and population until 1500 CE or so, then less tiny growth until 1870 or so, then rapid growth. Technically that first figure is not income but a different construct out of an Introduction to Economics textbook but it sort of corresponds to “the amount of stuff a thousand people can get done in a year.” The book is more interested in estimating a global trend than in cases like the rise and fall of the Roman empire, it gets more interested in specific places and times once its Long Twentieth Century begins.
Came here to say this. Delong absolutely talks about how slow and cyclical economic growth per capita was pre 1870 and how Malthusian logic turned any technological advance into population growth which reduced living standards.
“ Economic leftists oscillate between saying that the capitalist system is failing and needs to be replaced before we all choke on our own filth, and saying that, as the richest society in the world, we should be able to afford to massively reduce poverty and inequality.”
But I find this to be a false dichotomy. The way to reduce poverty and inequality is to re-localize economies.
Economic consolidation is constructed from hierarchy and hierarchies are associated with greater energy intensity. Factor in scale, and a “mystery” like Jevons paradox disappears.
To give a brief example: when city finances are audited over the long-term, 15 to 20 years, where the revenue of the city receives versus the expenses it’s committed to we find that big box stores (Walmart etc) leave the city in the red. The primary reason for this is all the debt and all the maintenance obligations the city undertakes to support big box operation far exceeds any revenues created by that operation. Cities that invest in local business building out their business supporting their operation benefit greatly they see positive revenue. This is because supporting the infrastructure for local business is less expensive in part because it is more compact.
A quoted figure I saw was for every hundred dollars spent on local businesses 47 recirculate. The number drops down to 14% or 17% if it goes into a big box store.
One of the chief problems is that capitalism has been confounded with market economies. This is on purpose the two are not equivalent. It is one more false equivalent used by mainstream economics to persuade people to defer to economic consolidation.
+1
“when city finances are audited over the long-term, 15 to 20 years, where the revenue of the city receives versus the expenses it’s committed to we find that big box stores (Walmart etc) leave the city in the red…”
What sources do you have for this claim? Seems interesting.
There’s a lot here, much of which I’m completely unqualified to comment on. But I have to disagree with the following statement from Manning: “Economic growth that is decoupled from energy consumption won’t look much like GDP.” A few centuries ago, GDP mostly consisted of agricultural production. Today, the percentage contribution from agriculture is rather small. In that sense, GDP has been decoupled from agriculture.
By analogy, it should be perfectly possible to move to a post-energy economy in which economic growth continues while energy production remains more or less stable. New goods and services can be developed that require less and less energy to deliver. Prosperity can continue to increase.
I have no idea whether that optimistic scenario will actually come to pass. I’m only saying it’s a possibility. If prosperity hits a wall, it will be for reasons other than an absolute physical limit on energy production.
Murphy is certainly correct that energy production can’t continue on an exponential rise forever. But the ratio of energy production to prosperity (however defined) can become arbitrarily close to zero.
As a political scientist, you should be deeply worried about a zero sum world, where it is no longer true that a rising tide floats all boats, and it is indeed the case that allocating resources to one group requires that they be taken from another.
Just got back from a trip and saw this. I’m always annoyed when someone rails against “endless growth”. This is simply meaningless, since the fact that everything ends sooner or later has nothing to do with whether it should end, say, tomorrow. And people who are so sure about this generally have no idea of what “growth” means. A little under a fifth of the US economy is health care. This includes lots of shiny machines, but mostly it’s people and medications that are supposedly expensive because of their research costs. That’s GDP. Education? That’s GDP. I suppose Trump could say he wants to shrink higher ed and axe the Dept of Education because he’s against “endless growth”. In my climate book I mention the music education program in Finland which has made that country an unlikely mecca for music of all sorts — and that’s GDP. GDP can be stupid, wasteful and polluting. It can also be low footprint and elevating. We get to choose, or we should get to.
As for backcasting exponential growth, I’ve done a bit of that myself when teaching development. We don’t have much data to go on except for occasional inventories like Domesday or sporadic credit records. And of course the index number problem associated with changing composition becomes overwhelming on a scale of centuries or longer. So any concept of economic growth over that timespan is really only notional, and the main purpose for backcasting is to demonstrate how unprecedented the sustained growth of the past 150 or so years has been. You can do the same for population. I haven’t read DeLong’s book, but if he attributes much more significance to such estimates than that, he’s probably reading in.
A final note. There are two contradictory facts about the Roman Empire. One is that it vastly improved the material culture of those who benefited from it. Just walk through the Roman-Germanic museum in Cologne, and see the sudden deterioration of the artifacts housed there corresponding to the Germanic conquest of Rome’s outposts. The other fact is that every tribe that had a ghost of a chance, or thought they did, fought Roman conquest because the Empire was brutal and ruthlessly exploitative. It’s both and.
Peter, the Tom Murphy essay that I link in the original blog goes through “what if the non-material part of the economy grows while consumption of material resources and energy flattens?” Maybe its wrong but the book I reviewed does not really address this sort of argument. AFAIK its generally agreed that GDP is ~ energy consumption and that it has trouble measuring free information goods like Wikipedia and Project Gutenberg. https://dothemath.ucsd.edu/2011/07/can-economic-growth-last/
Thanks, Sean. The simplest way to think about GDP is that it measures the fungible portion of the economy, stuff that in some sense trades for money. (This includes the public sector, although its purchases are indirect.) It leaves out lots of the services we depend on and is not a measure of anything, really, except what makes up the income-expenditure identity. As you go back in time, less and less of what people produce and consume is traded for money, so estimating historic GDP becomes increasingly meaningless. I don’t know why that isn’t obvious.
As for the future — green growth and all that — the challenge of climate change is unprecedented, and the measures we need to take to contain it are unprecedented as well. History is again no guide. There has never been an energy transition in which the energy source being superseded just disappeared within the space of a decade or two. And previous transitions were not purposely engineered, they just responded to changes in cost, availability and suitability to changing end uses. I can’t think of anything in history that corresponds to a climate program that rapidly phases out fossil fuels.
Peter, Tom Murphy has spent now more than a decade “doing the math” on the “Do the Math” blog. He even wrote a whole textbook on the topic: https://escholarship.org/uc/energy_ambitions
His point is basically this, not only can the strongly material part of the economy (furniture, cars, electronics, clothing etc) not grow forever, but also the less material stuff like education or music production or whatever can’t grow forever either. I mean, this is obvious to you and me, nothing grows forever. But he takes it farther to get general estimates of just how much longer things can “look like” the period between say 1950-now. Pretty much the answer is none, none more longer (a Spinal Tap reference).
We’ve already seen it, energy usage per capita has held fairly constant in the world for 20 years. It’s declined for most of the major developed countries like US, UK, France etc, basically by shifting that usage to China and Brazil etc. But we should expect China to flatten out or even decline in 20 years.
He points out that even if stuff just holds constant instead of growing at 3% per year that’s already a major “breakage” of the current world system. Include increased climate based destruction from flooding, fires, desertification, etc and essentially we are already at the end of modernity people just haven’t really figured it out. What the future looks like isn’t clear, but it won’t be skyscrapers and gigabit fiber to everyone in Ghana 40 years from now. Hell, it likely won’t even be cars and highways and airplane travel and etc at the level we have in the US today 40 years from now (they’ll exist but at some lower level). How are we supposed to have all that while per-capita energy consumption has been falling since 2000, and is likely to fall farther? That’s true even if we don’t want it to be, because the *energy* cost of extracting energy has been increasing. So it’s just *harder to get* energy even if you discount the climate issue. And the climate issue isn’t gonna be discounted. Climate will increase the energy cost of staying still in terms of physical capital (ie. destructive events will become more common and bigger). So even if energy per capita held constant somehow, capital stock would decline or stay constant at the expense of other consumption.
There are some things we could do about all this. We could target deployment of solar, wind, and a new kind of nuclear with better waste handling perhaps. But it still doesn’t let us grow a lot longer, because at the growth rates of the last 50-100 years the *waste heat* output becomes a bigger problem than the climate change gasses not that far out… https://tmurphy.physics.ucsd.edu/papers/limits-econ-final.pdf he calculates that the waste heat would equal the solar heating in 400 years. Since we can’t get anywhere close to this number and remain safe on earth. This puts an upper bound on growth on the order of 100 years max (or say within a multiplicative factor of 2 of that). So somewhere between 50 and 200 years no matter what tech we use. Given climate change is a real thing, and transition entirely away from fossil fuels is likely impossible on the 50 year timescale, we really are well and truly done in terms of growth at least.
Per-capita growth might still occur if population declines rather a large amount. But since the lifetime of humans is well more than 50 years, we aren’t gonna get fast enough decline without a lot of people dying prematurely.
To put it kind of poetically, the modern world ended more or less on 9/11/2001 and just hasn’t realized it yet.
Well, I haven’t read Murphy, so I can’t evaluate what he says, although the waste heat thing is new to me, and I’ve read a lot of energy analysis.
As a practical matter, I think reducing global CO2 emissions by 4-5% annually, indefinitely, which is what we need to do to stay within the 2 degree remaining carbon budget, is going to cut substantially into economic growth and probably cause at least some temporary reversal. Of course, not doing this will have an even more devastating effect, which is why we need to do it anyway. This is not because of some claim about “endless growth”, but just the immediate constraints imposed by the suppression of fossil fuels at a rate that exceeds their potential replacement. Whether economic growth resumes further down the road, I don’t know. We’ll see. (Those of us who are young enough to be around then.)
I think it’s a mistake to extrapolate current technologies or make hard predictions regarding what new technologies will become feasible at scale. We’ll invent some new stuff we didn’t predict, and we’ll also encounter areas where we hit a wall. The one thing that seems clear to me is that we should begin shifting our end uses (consumable goods and services) in the direction of less energy consumption. That’s why Finland’s music program is such a good example.
I like the theory that the “purpose” of life is to maximize the rate of entropy production, ie take a few high energy photons from the sun and ultimately use them generate many lower energy IR photons. And do this at an ever faster rate until the earth, then entire solar system (then galaxy) looks like a black body to the rest of the universe.
Eg, we consider photosynthesis to be something like 1% “efficient”. Based on the idea that the purpose of something is what it does, that is a misnomer. Generating “inefficient waste heat” is actually the purpose of photosynthesis. The other stuff like metabolism, reproduction, etc is more like a “side effect” or positive feedback. It helps to even further increase rate of entropy production. Animals are originally only secondary contributors whose role is to spread seeds, minerals, and nutrients around for the plants. Of course, tech can change this.
The universe selects for net contributors to this process. Failure of a phenomenon to do so essentially marks it (eg, species) for extinction.
And, if you think about it, the mark of a successful species/civilization is how much “waste heat” it generates: https://en.wikipedia.org/wiki/Kardashev_scale
Finally, like “junk DNA”, it is the height of hubris to call the primary thing all lifeforms do “waste heat”. It should be given a more neutral term at least.
Re: I [Manning] don’t see much value in estimating the population of the world in 6000 BCE when we can’t agree on the population of the Americas in 1492 within a factor of 5 [Manning’s original post said “a factor of 20” here, but after looking back at the literature he changed this to a factor of 5. — AG],”
I think that’s still too high, factor of 2 seems more like it from recent research.
There’s been much analysis since Caldwell & Schindlmayr(2002) and Henige(2008),
and Koch, Brierly, Maslin, Lewis wrote “Earth system impacts of the European arrival and Great Dying in the Americas after 1492” (2018)
https://www.sciencedirect.com/science/article/pii/S0277379118307261
24p, including ~17 pages of analysis of the various population estimates, concluding:
“Our estimate of the number of people living in the Americas in 1492 CE is 60.5 million, with an interquartile range (IQR) of 44.8-78.2 million, using the FFT approach (Fig. 4A). An alternate Monte-Carlo approach to combining the data results in a similar best estimate of 64.1 million (IQR 48.4-82.3 million; Table S2).”
(For context, much research was inspired by Bill Ruddiman in papers and book “Plows, Plagues and Petroleum”, which included hypothesis that the unique-within-millennia CO2 drop in the 1500s was partly due ot massive die-off in the Americas. People pushed back by saying “not enough people”, but as more people looked at issue they found evidence of more people (and also larger land footprints, so more trees cut).
Of course, expontential growth is another issue.
https://en.wikipedia.org/wiki/Total_factor_productivity
“It has been shown that there is a historical correlation between TFP and energy use[5] as well as corresponding energy conversion efficiency into useful work.[6] ”
In particular Robert Ayres & Benjamin Warr showed that TFP ~ energy*efficiency, so that much of the growth rates in 20th century was due to increasing use of fossil fuels, and many forecasts use that history. Needless to say, 21st century = hard work to convert fossil fuel use to renewables.
https://en.wikipedia.org/wiki/Robert_Ayres_(scientist) physicist turned economist, actully worried about real-world energy issues.
John:
Assuming for simplicity or approximation a normal distribution, the interquartile range is +/- 0.67 sd. The 95% range is +/- 2 sd. If we consider the upper and lower 95% points as a reasonable range of uncertainty, this would give a factor of (78.2/44.8)^3 = 5.3. So “a factor of 5” seems about right to me, if you trust that source.
If you take as your best guess the geometric median, your guess is expected to be accurate to within sqrt(5.3)=2.3 so we’re back to a factor of 2 again.
John:
I think that “within a factor of X” implies that the quantity is estimated roughly in the range (a, b), and X = b/a.
Agreed 100%, IF this were just a a statistical distribution with no external constraints/information.
I probably should have explaine more.
Over the last few decades, there has been a lot of work on land-use impacts of early agriculture, as well as people flying sensor-laden planes over the Amazon and finding signs of pre-1492 civilization now buried in jungles. (I’ve talked to some of the researchers.) These tend to constrain the outer edges of the distribution, often overriding interpretations of some of the earlier results, i.e., I think the likely distribution has high kurtosis … or I wouldn’t have suggested 2X rather than 5X.
Some of this comes from climate scientists like Bill Ruddiman happening to meet land-use ecologists like Erle Ellis, plus much input from other agricultural historians.
https://en.wikipedia.org/wiki/William_Ruddiman
https://en.wikipedia.org/wiki/Erle_Ellis
John, thanks for the citation! I will add it to the original blog post. Chapter 2.2 of Koch et al. 2019 shows that estimates of the population of the Americas vary in ranges such as 100,000 to 8,000,000 on Hispanola in 1492 (factor of 80), 2.2 to 52 million in Central Mexico (factor of 24), 2.3 million to 13 million in Yucatan (factor of 6), 1 to 20 million in Amazonia, 0.9 to 18 million in the USA and Canada (factor of 20). I would be wary of performing arithmetic on all those and assuming that the truth must be in the middle, because often these different estimates come from different assumptions, eg. “did Augustus’ census figures include just adult men or women and children?” or “were the Amazonians all small-scale hunters and gardeners like their descendants, or intensive farmers through the terra preta zone?” And as David Henige and Timothy W. Guinnane have explained, often there is no solid basis for any number except the first census and feelings about how much higher or lower figures must have been in the past. Its well known that when you ask people to guess a number for something (say the population of Haida Gwaii in 1774), and there is a minimum but no maximum (say a Canadian Government count from 1921 after massive losses to disease), the average guess will be skewed high.
Do we even know the population today to within 50%? 25%?
https://www.nature.com/articles/s41467-025-56906-7
Also search for China in particular, which has many families with illegal (unreported) female births, and large issues with politically fudged numbers.
Even in the US, the census data was used to round up Japanese into camps during WW2. No doubt there are many people off the books due to similar fears. Maybe 10%?
Anoneuoid, Statistics Canada estimates that the 2021 census was a 3% undercount (about 5% of the population missed, and 2% counted twice). That is the sort of error I would expect in a rich country. I have not looked closely at the aerial study of rural China which argued that the Chinese underestimate their population, or the arguments that the UK has lost track. https://www12.statcan.gc.ca/census-recensement/2021/ref/98-303/2021001/chap1-eng.cfm
It seems like they mention homeless (similar issues as very rural) but never provide any method to address it:
https://www12.statcan.gc.ca/census-recensement/2021/ref/98-303/2021001/chap3-eng.cfm
Ie, they got their error ranges from calling people they found in other types of records. People who aren’t going to answer/cooperate will be invisible.
Well, it is a contentious issue, as people seek to narrow the bounds of uncertainty.
I posted an answer to Andrew, I think there is a lot of land-use resaerch by folks like Erle Ellis and the size of rthe CO2 drop that constrain the outer edges of the distributions. But this is from a fairly complex mass of research, much in last decade or two, and very multidisciplinary.
I recommend BilL Ruddiman’s 2013 AGU Tyndall lecture, https://www.youtube.com/watch?v=1TOTsmqgmL8
or his wonderful book, Earth Transformed(2013).
https://www.amazon.com/Earth-Transformed-William-F-Ruddiman/dp/1464107769
This puts me in mind of how when I took demography they had basically these “time to double population” estimates, but they were never accurate because they are just models with assumptions. Meanwhile people stop smoking, get vaccines, face pandemics and suffer from war. And now we have all this talk about “below replacement level” birth rates that don’t really have much to do with population size (not age distribution) because e.g. my parents have 2 living children and 4 living grand children and are still alive, so even though they had “replacement level children” they increase the population size in 2025 by 6. Likewise the meaning of being in your 70s (or in their cases your 90s) in the US and much of the rest of the world has really changed.
In fairness to neoclassical economics, the basic Ramsey and Solow models of economic growth predict that the economy would reach a steady state even in the absence of environmental constraints, unless pushed by exogenous factors (e.g., population growth, technological progress). It is true that the economics profession did not like these results and there have been several attempts to develop models of “endogenous growth” (Lucas, Romer and others in the ’80s and ’90s), but I don’t think these models are taken too seriously nowadays.
Disclaimer: I basically stopped following the macroeconomics literature thirty years ago, so take the above with the customary grain of salt.