[FRIAM] agent-based macroeconomic model

Sun May 4 04:22:00 EDT 2025

The *World3 Limits to Growth* model, developed in the early 1970s and made
famous through the Club of Rome’s report, is a well-known example of how
internal conceptual frameworks are externalized and explored through
computer modeling. In this case, the authors constructed a systems dynamics
model to simulate long-term global trends in population, resource
consumption, industrial output, and pollution. Their primary aim was to
communicate a particular vision of the future: one in which exponential
growth in a finite system inevitably leads to overshoot and collapse unless
proactive changes are made.

What is often overlooked, however, is the degree to which the conclusions
of such models are shaped by the assumptions built into them—assumptions
that may not always be made explicit. In the case of the *Limits to Growth*
model, one of the most critical assumptions underpinning its projections is
that technological development during the period from 1900 to 2100 will not
proceed rapidly enough to fundamentally change the dynamics of resource
availability and consumption. In other words, the model implicitly assumes
that no breakthrough technologies will emerge in time to mitigate the
depletion of key resources or to radically improve the efficiency and
sustainability of industrial activity.

This is a crucial point, because the validity of the model’s most dramatic
conclusions—such as widespread societal collapse by the late 21st
century—hinges heavily on that assumption. If, in contrast, technological
innovation does accelerate and lead to the discovery or creation of new
resources, energy sources, or modes of production, then the trajectory of
global development may look very different from the model’s projections.

I freely admit that I do not know whether this key assumption is ultimately
correct. No one can say with certainty what the pace or direction of
technological progress will be over the next 75 years. However, based on
historical trends and current trajectories, I am inclined to believe that
the assumption is too pessimistic. In my view, it underestimates humanity’s
capacity for innovation and the accelerating feedback loop between
knowledge, technology, and problem-solving.

Of course, we must acknowledge that the Earth's natural resources are
finite. At some point—whether decades or centuries from now—this finitude
will impose constraints on economic and population growth. That said, the
timing and severity of these constraints depend heavily on how we define
and access resources. Technological advancements can dramatically alter
both. For instance, materials once considered scarce or inaccessible can
become viable through improved extraction techniques, recycling, or even
synthesis. Similarly, previously unusable energy sources may become
dominant through innovation, as was the case with oil in the early 20th
century.

To illustrate this point more concretely, consider the domain of energy.
Energy is foundational to almost every aspect of economic growth and
societal development. If we are able to develop clean, scalable, and
abundant sources of energy, many other constraints—such as water scarcity,
food production, and even material shortages—can potentially be addressed.
Sam Altman and others have argued persuasively that the combination of
abundant energy and advanced intelligence systems (even if narrow and
artificial) could usher in an era of material abundance.

It is important to clarify what is and isn’t meant by this. I am not
referring to fantastical technologies that violate the known laws of
physics. I am speaking of plausible, science-based advances that are
already in development or on the horizon. Nor am I invoking some vague
notion of sentient or “real” artificial intelligence. The kind of AI I have
in mind is the narrow, task-specific form we see today—tools that, while
limited, are increasingly capable of solving complex problems, optimizing
systems, and accelerating scientific discovery when combined with human
ingenuity.

Nor should this vision be interpreted as a license for ecological
destruction. On the contrary, I argue that the path to abundance must be
rooted in sustainability. Technological progress should enable us to reduce
our environmental footprint while increasing our capacity to meet human
needs. Clean energy technologies, circular economies, and efficient
material use are essential components of this future. One promising example
is the development of thorium-based nuclear reactors, such as the one
currently being built in China. These reactors offer the potential for
abundant, safe, and low-waste energy—possibly serving as a bridge until
nuclear fusion becomes a practical reality.

In sum, the *Limits to Growth* model is a valuable intellectual exercise
and a cautionary tale. It highlights the risks of unchecked growth in a
finite system. However, its conclusions are not inevitable. They are based
on a set of assumptions, the most consequential of which concerns the pace
of technological development. If one believes that innovation will
stagnate, then the model presents a sobering and perhaps accurate warning.
But if one believes—as I do—that human creativity and technological
capacity will continue to grow, then a more optimistic future is plausible.

Ultimately, the debate over such models is less about data than it is about
belief—about how we weigh uncertainty and how we envision the future. I do
not claim certainty in my outlook. Rather, I argue that we should be
cautious in drawing fatalistic conclusions from models that may
underestimate the transformative power of innovation. No one can guarantee
that a future of abundance awaits us. But likewise, no one can confidently
assert that it does not.

In that uncertainty lies both the risk and the opportunity of our time.

On Sat, 3 May 2025 at 07:46, Jochen Fromm <jofr at cas-group.net> wrote:

> Nice model! Not bad. One aspect we could try to model is the distribution
> of supply chains in a globalized world.
>
> In 2021 the average gross income in the US was about 70,430, in Taiwan
> 21,689, in China 11,890, in India 2170, and in Myanmar 1,140. Supply chains
> of companies in a world where income differs so much will obviously end
> sooner or later in low income countries. Typical supply chain lines for
> Apple are for instance Apple (California) > TSMC (Taiwan) > Factory
> (Guangdong Province, China) or Apple (California) > Foxconn (Taiwan) >
> Factory (Henan Province, China).
>
> How are supply chains affected if transportation costs rise or tariffs are
> imposed?
>
> What would cause a world-wide economic crisis in such a model and how
> would it look like?
>
> One of the most famous models on a global scale is the world3 model from
> the Club of Rome. Brian Hayes decided to rewrite the world3 model in
> Javascript
>
> http://bit-player.org/limits
>
> The article is here
>
>
> https://www.americanscientist.org/article/computation-and-the-human-predicament
>
>
> -J.
>
>
>
> -------- Original message --------
> From: Pieter Steenekamp <pieters at randcontrols.co.za>
> Date: 5/1/25 7:48 PM (GMT+01:00)
> To: The Friday Morning Applied Complexity Coffee Group <friam at redfish.com>
>
> Subject: [FRIAM] agent-based macroeconomic model
>
> I made a very "quick and dirty" start on developing an agent-based
> macroeconomic model.
>
> Posted about it on X:
> https://x.com/pietersteenekam/status/1917995128678986170 , the post reads
> as follows:
>
> Me: “Let’s understand global macroeconomic policy better.”
> Also me: Builds a crude ABM with AI because economists can’t agree on
> tariffs.
>
> Is it useful? Not yet.
> Is it cool? Heck yes.
>
> 🔗 https://github.com/pieterSteenekamp/bottom-up-macroeconomics"
> .- .-.. .-.. / ..-. --- --- - . .-. ... / .- .-. . / .-- .-. --- -. --. /
> ... --- -- . / .- .-. . / ..- ... . ..-. ..- .-..
> FRIAM Applied Complexity Group listserv
> Fridays 9a-12p Friday St. Johns Cafe   /   Thursdays 9a-12p Zoom
> https://bit.ly/virtualfriam
> to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com
> FRIAM-COMIC http://friam-comic.blogspot.com/
> archives:  5/2017 thru present
> https://redfish.com/pipermail/friam_redfish.com/
>   1/2003 thru 6/2021  http://friam.383.s1.nabble.com/
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20250504/d568f05a/attachment.html>