When a rich man's software predicted the end of civilization

Fifty years ago, David Rockefeller formed the Club of Rome with the mission to "promote understanding of the global challenges facing humanity and to propose solutions through scientific analysis, communication, and advocacy."  That is, the Club presented itself as able to predict and warn of future global catastrophes and worthy to guide world leaders in navigating them.  With these lofty aspirations, the Club enlisted as members many prominent people from many vocations, often connected with the United Nations.

In short order, the Club discovered Dr. Jay Forrester, inventor of a new scientific discipline called "system dynamics," a prominent feature of which was to simulate via computer the essential behavior of the system under investigation, including (very importantly) the positive and negative feedback loops inherent in that behavior.  Dr. Forrester had previously applied this approach to the fields of industrial dynamics and urban dynamics and had received prestigious awards for his books on those subjects.

The Club enlisted Dr. Forrester to apply his approach to "world dynamics" – i.e., to simulate the essential behavior of the world as a single system, including commerce, population, natural resource and energy consumption, pollution, and other major variables that appeared important. 

In 1971, this work was documented in Forrester's book, World Dynamics, which included not only his conclusions, but also the DYNAMO source code he used for the simulation. 

This work and its major conclusions were also captured in a 1974 Australian TV documentary, which is available here.  Also included is an interview with two Club of Rome officials (hat tip: WND).

I was once enthusiastic about Forrester's work, and I bought a copy of World Dynamics when it first hit bookshelves.  I translated Forrester's DYNAMO source code to APL and faithfully reproduced his results.  Then I modified his initial input values to create a variety of simulations.

It was immediately apparent that one could produce world catastrophe, as he did (with civilization collapsing in the year 2040), or a relatively benign or even idyllic result, with a few simple changes in initial input conditions. 

The kicker is that no one can possibly know which input conditions are even approximately correct.  For example, there is no distinction among the various contributors to "natural resources," but there is a single value that represents how much (all) of them is assumed to be recoverable at the beginning of the simulation (the year 1900).  How is this representation anything but meaningless? 

An additional flaw is that Forrester's model takes account of human learning only in a gradual sense across the total time of the simulation.  (The one exception is that he allows for potentially different global birth rates before and after 1970.)

- In reality, the fracking revolution was a sudden development that dramatically increased the "known reserves" of oil. 

- In reality, in all countries, the birth rate has been increasingly responsive to changes in affluence and education. 

- In reality, short-term results from a single recent election have suddenly changed much of the world's expectation of economic growth.

The clear conclusion I drew from my experimentation was that Forrester's choice of input conditions was arbitrary and accidental, or else (more likely, I decided) the conditions were selected to produce the catastrophic results that conformed to the Club of Rome's desires or expectations.  In either case, garbage in, garbage out.

Forrester's seemingly superior approach had beguiled me into what seemed later to be a large waste of time.  On the other hand, I learned that science is easily corrupted by politics and that liberal groups like the Club of Rome tend to have Malthusian misconceptions about how the world actually works.

This year, the Club of Rome is contemplating how to manage a world with 10 billion humans.  In 1974, the Club didn't think such a world was possible – at least without its members' guiding hands.

Fifty years ago, David Rockefeller formed the Club of Rome with the mission to "promote understanding of the global challenges facing humanity and to propose solutions through scientific analysis, communication, and advocacy."  That is, the Club presented itself as able to predict and warn of future global catastrophes and worthy to guide world leaders in navigating them.  With these lofty aspirations, the Club enlisted as members many prominent people from many vocations, often connected with the United Nations.

In short order, the Club discovered Dr. Jay Forrester, inventor of a new scientific discipline called "system dynamics," a prominent feature of which was to simulate via computer the essential behavior of the system under investigation, including (very importantly) the positive and negative feedback loops inherent in that behavior.  Dr. Forrester had previously applied this approach to the fields of industrial dynamics and urban dynamics and had received prestigious awards for his books on those subjects.

The Club enlisted Dr. Forrester to apply his approach to "world dynamics" – i.e., to simulate the essential behavior of the world as a single system, including commerce, population, natural resource and energy consumption, pollution, and other major variables that appeared important. 

In 1971, this work was documented in Forrester's book, World Dynamics, which included not only his conclusions, but also the DYNAMO source code he used for the simulation. 

This work and its major conclusions were also captured in a 1974 Australian TV documentary, which is available here.  Also included is an interview with two Club of Rome officials (hat tip: WND).

I was once enthusiastic about Forrester's work, and I bought a copy of World Dynamics when it first hit bookshelves.  I translated Forrester's DYNAMO source code to APL and faithfully reproduced his results.  Then I modified his initial input values to create a variety of simulations.

It was immediately apparent that one could produce world catastrophe, as he did (with civilization collapsing in the year 2040), or a relatively benign or even idyllic result, with a few simple changes in initial input conditions. 

The kicker is that no one can possibly know which input conditions are even approximately correct.  For example, there is no distinction among the various contributors to "natural resources," but there is a single value that represents how much (all) of them is assumed to be recoverable at the beginning of the simulation (the year 1900).  How is this representation anything but meaningless? 

An additional flaw is that Forrester's model takes account of human learning only in a gradual sense across the total time of the simulation.  (The one exception is that he allows for potentially different global birth rates before and after 1970.)

- In reality, the fracking revolution was a sudden development that dramatically increased the "known reserves" of oil. 

- In reality, in all countries, the birth rate has been increasingly responsive to changes in affluence and education. 

- In reality, short-term results from a single recent election have suddenly changed much of the world's expectation of economic growth.

The clear conclusion I drew from my experimentation was that Forrester's choice of input conditions was arbitrary and accidental, or else (more likely, I decided) the conditions were selected to produce the catastrophic results that conformed to the Club of Rome's desires or expectations.  In either case, garbage in, garbage out.

Forrester's seemingly superior approach had beguiled me into what seemed later to be a large waste of time.  On the other hand, I learned that science is easily corrupted by politics and that liberal groups like the Club of Rome tend to have Malthusian misconceptions about how the world actually works.

This year, the Club of Rome is contemplating how to manage a world with 10 billion humans.  In 1974, the Club didn't think such a world was possible – at least without its members' guiding hands.