For and Against Degrowth

1st of April, 2026

Degrowth is one of those magic words that make surrounding people temporarily less intelligent once mentioned. Soon after, you can expect to receive a bouquet of keywords from your conversation partner signifying which camp they belong to (the cool one, usually the same as yours), all without treading near a substantive argument. I'm exactly like this as well, and feel an urge to start a rant right about now. I find the concept of degrowth intuitively wrong. But there is reason behind it, and this reason recalls another problem I find more interesting, so it's worth examining charitably. The basic argument is as follows.

Earth has finite resources. The current world economy is using a significant proportion of some of the easily depletable resources. As economic activity grows, more resources are used. Therefore, if the economy continues to grow, important resources will be depleted fast. At lower economic activity per person, the same amount of resources can support more happy human lives in total. Therefore, we should lower per capita economic activity.

This argument is very straightforward, and in many ways true. Let's go through every step.

Earth has finite resources.

This is trivially true, but it does hide one more assumption not usually discussed seriously in the degrowth literature, namely that Earth is the only meaningful source of resources. Most importantly, if we could harness the energy of the sun beyond what reaches Earth, the energy available to humanity would be so large that we could sustain much higher per capita energy consumption. The moon provides more mass and surface area, and so do other celestial bodies in our solar system. That said, moving a significant amount of solar panels (or mass in general) into space might take too long to be practically helpful for issues like climate change. Global energy consumption was at about 20 terawatts in 2023From Enerdata. A good solar panel at scale might get about 25% efficiency and weigh 3 kg/m², and solar irradiance near Earth is 1.36 W/m², meaning 10 g/W, so to get to replace the Earth's energy consumption without any loss when moving energy from space to Earth, we need about 200 million tons of solar panels in space. While that's not impossible at all, a comparison with the current annual launch capacity of around 3,000 tons per year tells us that it will take a while. The broader point of the degrowth argument still stands. It will be very hard to move industry off the Earth in time. Even if we do, the solar system is still finite, and other stars aren't relevant for this conversation.

The current world economy is using a significant proportion of some of the easily depletable resources.

This is also straightforwardly true. The resources that are commonly raised as easily depletable are capacity of the Earth to deal with increased greenhouse gas levels, and surface area housing natural beauty, ecologically diverse life et cetera. The IPCC AR6 Working Group III calculates 1150 GtCO₂ until 2 degrees global warming (after 2020)From IPCC AR6 WGIII Technical Summary, with global emissions up to 2019 totaling 2400 GtCO₂, and annual global emissions in 2024 were 37.8 GtCO₂From IEA Global Energy Review 2025. As a representative example for an important ecosystem, about 26% of the area of the Amazon rainforest has been semi-permanently destroyed as of 2022From Amazon Watch. I personally come from a country where 100% of the original forest area has been harvested at least once, and I wouldn't wish that on anybody else.

As economic activity grows, more resources are used.

Again, this is empirically true. On a per country basis, GDP is highly correlated (linearly, in fact) with things like energy consumption (ρ=0.7), CO₂ emissions (ρ=0.89) and material consumption (ρ=0.50)Calculated with data from Our World in Data, sourced from World Bank, Global Carbon Budget, the Energy Institute, and the UN Environment Programme (GDP, CO₂, Energy, Materials). A priori this makes sense as well; economically active societies do a lot of things that need physical materials and surface area to facilitate.

Therefore, if the economy continues to grow, important resources will be depleted fast.

This follows from the previous three statements.

At lower economic activity per person, the same amount of resources can support more happy human lives in total.

The basic premise is that marginal returns to wealth, and therefore resource usage, diminish as wealth increases. There have been many investigations on wealth or GDP per capita vs. happiness, on the societal and individual level, with various definitions of happiness. Not one of them, though, has concluded that happiness increases without diminishing returns. So this is true as well.

Therefore, we should lower per capita economic activity.

This follows from the previous two statementsPut this way, it might seem like the statements about the urgency of resource depletion is not even needed to complete the argument, but in practice most people heavily discount future generations, so this only goes through if resource depletion happens soon..

Before weighing in on the other side, one caveat. Some degrowth thinkers stress non-human conscious life, the inherent value of ecological systems, and especially the connection with the kind of economy that generates economic activity (i.e. capitalism vs. something else). These are in my view all sideshows that weaken the central argument, which as we saw is strong in its simplest form. To resolve the question at hand, as with all important questions, it's unnecessary to first settle which of the grand ideologies is the correct one.

Now, if I agree with all the premises, and buy the structure of the argument in general, then why do I still disagree with the conclusion?

The first part of the answer lies in the solution proposed by degrowth. If we can impose a constraint that completely flies in the face of the forces shaping the global economy today, then surely we can instead impose a constraint that only slightly flies in the face of the global economy. One only slightly face-flying constraint might be: Don't touch the remaining 74% of the Amazon. However hard that might be to enforce, and it certainly is extremely hard, it is still easier than disallowing the entire world from getting richer. The classic case is the depletion of the ozone layer by atmospheric emissions of various substances that catalyze O₃ into O₂ and O, which was successfully fixed through international regulation starting with the Montreal Protocol in 1987.

Nonetheless resources are finite, and as the economy grows, we have to impose more constraints to not destroy the Earth. So one could object that granular regulation only buys us a little bit more GDP before we hit a hard limit. Here I see the fundamental flaw of the degrowth argument. The limit on economic activity doesn't look like a budget to be depleted. Instead it looks more like the Landauer limit, the lowest possible amount of energy expended to perform one (non-reversable) computation. Or perhaps the Bremermann limit, the highest number of computations doable with a computer of fixed weight in a fixed time.

Though the principle applies to all things, not just integrated chips, they are the best example to illustrate: The Intel 4004, released in 1971, has a max. clock speed of f = 750 kHz at P = 0.5 W, a modern CPU like the Intel i5-14600K has a maximum clock speed of f = 5.3 GHz with n = 20 threads at P = 181 W. Energy consumption per clock cycle is P/(n × f), so 6.7 × 10⁻⁷ J for the old CPU and 1.7 × 10⁻⁹ J for the new CPU. This means the new CPU is about 400 times more efficient. The new model is cheaper, easier to use; it is simply better. If we move the same calculation to the transistor level, we can compare to the Landauer limit E = k_b · T · ln(2) = 3.4 × 10⁻²¹ J. Multiplying the entire CPU's energy consumption by the number of transistors typically active in the CPU at a given time (divided by 4, the number of transistor to make an elementary non-reversible gate, the NAND gate), we get an approximation of the energy per transistor operation. Even today's CPUs are still a factor of 1000 away from the theoretical limit. The below figure shows the same calculation for a number of Intel CPUs along with a horizontal line at the Bremermann limit.

We see that even the most advanced, aggressively optimized technologies still have room for improvement. Other technologies, say a brick, do not have as large a potential to improve, but given that "human happiness" or "economic activity" are such broad goals, I believe that chip-sized efficiency gains are plentiful overall across the economy. That is, if we try hard to optimize, the resource usage of most subsystems of the economy can be drastically reduced. Another illustrative example: There is no law of nature that says that a coal-fired power station can't be carbon-negative. Burning a given mass of coal releases much more energy than what is needed in theory to bind all the released CO₂, and Herzog (2024) calculates that even with the inefficiencies of today's immature carbon capture technology, removing >85% of the produced CO₂ only needs <25% of the produced electricity.

In comparison to the possible multiplier obtainable from technology, the fraction of Earth's resources we want to use (e.g. 0.1% vs. 10%) is a tiny factor.

If you buy my counterargument, I'd still encourage you to be aware of its structure. I did not propose a technological solution to solve the problems pointed at by the degrowth argument. Instead I proposed the abstract ideal of a technological solution. To actually implement anything, we need many highly skilled people actively developing technologies that allow the economy to grow further without causing environmental destruction. These are incredibly hard problems, and the people making progress on them are heroic. Generically regulating or deregulating the economy does not get us anywhere, though narrow regulation (e.g. carbon credits) and deregulation (e.g. fast permitting for carbon capture) can help a lot. Least helpful of all is rejecting degrowth out of cultural osmosis and reasoning backwards that the problems it raised must not have mattered. Admittedly, another example of not helping is this article. A better use of time than writing this might be to actually work on technology that solves things.

Managing environment-related externalities in the global economy is not the hardest problem we are facing. It is however a formidable lesson on leaving the default path in favor of more beautiful life on our special planet.