World3 : the first computational model to incorporate natural physical limits in the economy.
The book ”Limits To Growth” (LTG) [5] was commissioned by the Club of Rome and has been published for the first time in 1972. It became a major reference as one of the first analyses to describe the impact of an economic system promoting growth on a planet with limited resources. To this end, LTG based its analysis on a systemic model World3. This model simulates the world under a bunch of scenarios with real data and various assumptions. In this article, we aim at summarizing the key contributions and reviews of LTG.
The model. The model World3 describes the world at a high level. It focuses on most important variables (e.g. population, pollution, industrial capital …etc) and models their interactions with linear and non-linear relations (see Fig. 1). An important feature of the model is the exponential growth or depletion emerging from the existence of positive and negative feedback loops in the model. These simple but realistic interactions may lead however to surprising results for lay public. A case in point of exponential growth is a sheet of paper folded 42 times is as thick as the distance Earth-Moon. If you are curious about the details, you can check out all the variables and interactions on this webpage and run different simulations :)
The all World3 system is fed by an input of resources. We can distinguish between renewable resources (Land, water, forest, ecosystem) and non-renewable resources (Fossil fuels, materials). These sources describe a first set of unchangeable physical constraints of the Earth. On the output side, World3 spits out wastes and pollution (CO2, toxic pollution…). Then, those have to be absorbed by natural sinks which describe a second set of physical constraints. Regeneration rate of the sources and absorption rate of the sinks are of primary importance for a sustainable system. Overuse of sources or overflow of sinks are indeed not sustainable.
The scenarios. The purpose of the model is subject to a lot of uncertainty, and thus cannot be considered as a predictive model according to the authors. The goal of World3 is to describe different possible trends assuming different undertaken actions. Among thousands of tested models, LTG presents 8 selected scenarios with different assumptions relative to the desire for a constant growth, the existing physical limits of sources and sinks, the delay and the quality of distorted signals received from the economy and from nature. For each selected scenario, LTG reports computes the trajectories until 2100 of most important indicators reflecting the state of the world (e.g. food, population, industrial output, pollution, resources), the material standard of living (e.g. live expectancy, consumer goods/person, food/person, services/person) and the human welfare and footprint (e.g. human welfare index, human ecological footprint).
Most scenarios end up in either overshoot and oscillations or overshoot and collapse. In this case, overshoot means a rapid change after crossing the physical resources limits. Scenarios without overshoot are very rare or usually rely on unrealistic assumptions (e.g. scenario ”Infinity In/Infinity Out” assumes unlimited resources). Scenarios without quick actions and relying on self regulating market and technologies lead to collapse. Scenarios with quick and deep modifications of the problematic forward loops avoid a collapse after the overshoot. In particular, the scenarios using the historic data (scenario ”standard run”) or with optimistic technology development (scenario ”comprehensive technology”) lead to a collapse in World3. Another important scenario assuming a reduction of children and an enhancement of the concept of sufficient standard of living (scenario ”Stabilized world”) manage however to avoid a strong collapse. We show examples of scenarios on the in figure 2.
The problems. An infinite exponential growth in a finite work is obviously problematic: the sources will eventually reach their production exhaustion and the sinks will reach their absorption limit. A first major issue is the general desire of constant growth which inevitably leads to approaching the limits with at a high speed.
Second, neither the limit signals nor the answers to these signals are instan-taneous. Combined together, the first and second problems imply a delay before the corrective actions start to take effect, and thus create an overshoot.
Moreover, natural resources are not only finite but also erodable. In otherwords, the degradation of the natural resources by their over exploitation is not easily reversible. Hence, it is not guaranteed that society will be again capable to recover the same natural resource capacity in a short future.
Eventually, the pursued goal of a consumer society is unlikely to reduce poverty. We observed an increase of the gap between rich and poor at country and individual levels. As an example, it leads to mostly two outcomes rich countries with small population growth and poor countries with large population growth. Both cases are unsustainable as they have an increasing ecological footprint. The authors relate this phenomenon to the ”success to successful” positive feedback loop which reward the wealthiest.
The solutions. First, LTG’s authors mention that it is feasible to come back from beyond the limit and take the ozone story as an example. The international will to take actions jointly with an international research collaboration helped to solve the depletion of the ozone layer. This will probably be required to face the LTG challenges.
The transition to a sustainable system requires to gain time with more efficient solutions and change the feedback loops at the root cause. Hence, the society changes should deals with the number of children per family, giving with the notion of sufficient standard of living and with leveraging technology for more efficient resources use. In other words, it should enhance concepts like long term planning, sufficiency and efficiency. We should also question ourselves on real need for humans.
Authors insist however that relying on self regulating market and technology development alone is not a reliable solution. The two main reasons are that market and technologies serve the goal of a consumer society and are subject to delayed signals. In particular, money is not an indicator of real physical quantities but it only conveys information about relative costs and values.
To conclude, LTG suggests five powerful non-scientific tools to start the sustainability revolution: (1) Visioning what we really want and need, (2) Networking to have global and fair connections, (3) Truth-telling to avoid biases and simplifications, (4) Learning by going slowly, collecting information and trying solutions, and (5) Taking care of people.
The reviews. LTG got both negative and positive reviews. On one hand, most famous criticism by the university of Sussex [6], Bjorn Lomborg et al. [4] or Julian Simon [8] stated that the model rely on non-valid assumptions, poor data and achieved wrong catastrophic predictions. As an example, cross-validating World3 on the past would have lead to an early collapse which did not happen. Later, another competing model DICE (”Dynamic Integrated Climate-Economy model”) with neoclassical economics assumptions rather than physical assumptions has been developed by William Nordhaus. This model allowed him to win the Nobel Prize in 2018.
On the other hand, LTG and its updates versions repeated that the goal of the model is not purely predictive and rather aims at a global trend description. Furthermore, other defenders [2] mentioned that critics came mainly from groups seeing a threat in their activities (e.g BP, economists, catholic church) and focused on isolated equations which might not be relevant for a global model. Besides, additional studies [9, 10] observed that the scenario ”standard run” of World3 achieved much more accurate predictions than the ”stabilized world” or ”comprehensive technology” scenarios given 30 or 40 more years of data (See Fig. 2). It suggests that the standard run is the best representation of the reality among the 8 key scenarios of LTG.
To conclude, World3 turned out to be a seminal and groundbreaking model to estimate the impact of the human activity on the planet. While it achieved good predictions so far, warnings about the model are legitimate. Indeed, a model relying on poor data and having a poor generalization capacity cannot be desired. Recently, other models and reports (e.g. GIEC [3]) presented updated estimations for world resource consumption which turns out not to be more optimistic. Moreover, Machine learning techniques which appeared to generalize to various tasks might be a promising direction to improve hand-crafted models (examples of programs: [7], [1]…). In particular, it might be interesting to improve the reliability of economic models by estimating the uncertainty of their predictions.
References
[1] Using ai to better understand and model the earth system.
[2] Thomas Burch. The limits to growth revisited.
[4] B. Lomborg and Olivier Rubin. The dustbin of history: Limits to growth.
[5] D. Meadows, J. Randers, and D. Meadows. Limits to Growth: The 30-Year Update.
[6] Felix Pirani. Thinking about the future: A critique of the limits to growth.
[8] Julian L. Simon. The ultimate resource.
[9] Graham Turner. A comparison of the limits to growth with 30 years of reality.
