Thursday, April 30, 2020

The most accurate model-based prediction of all times

The "base case" scenario from the 1972 edition of "The Limits to Growth." This scenario described the trajectory of the world's economy on the basis of the data and assumptions that were judged to be the most reliable ones. This run might turn out to have been amazingly accurate some fifty years after it was proposed.


One of the most remarkable features of the story of the "Limits to Growth" study of 1972 is how effectively it was possible to convince almost everyone that it was completely wrong. Amazingly, though, the most vituperated model-based prediction in history may turn out to have been perhaps the most accurate one.

Note how the scenario above, the "base case" scenario, saw the start of the decline around 2010 and the start of the collapse maybe a decade afterward, that is now. If the oil collapse generated by the coronavirus takes the whole economy with it, as it may well happen, then this scenario turns out to have been unbelievably accurate. And that for a prediction made 50 years ago. Truly amazing!

Now, of course, this story has to be taken with some caution, predictions can be right even by mere chance. But, in this case, there is a certain logic in this result: the base case scenario had been already noted by Graham Turner to have been following the real-world data. But that was true for the growth side of the diagram: even standard economic models had been predicting economic growth. The crucial test for the model was to be the sharp change in slope expected to take place around 2010-2020.

Of course, no model could have predicted that the turning point would have been triggered by a word pandemic -- as it happened. But "something" had to give and the virus is not a cause of anything, it is just the straw that breaks the camel's back. The little push that sent the system in a direction where it had to go.

So, it IS possible to use models to predict the future. Another example of a good prediction is the famous one by Marion King Hubbert of the peak of oil production in the US. In 1956, he had proposed 1970 as the likely date and he had been right. On the other hand, predictions are not always so good. In 1970, Hubbert himself had predicted the global "peak oil" for the year 2000. Later on, ASPO (association for the study of peak oil) had estimated the peak for 2010. Both predictions were not so bad, but a little pessimistic if the peak arrived in 2020.

Perhaps the most surprising discovery, here, is how the most vituperated predictions turned out to be the most accurate. Conversely, many economic models that predicted infinite growth were much praised, but they seem to have badly missed the ongoing collapse. Maybe vituperation is a good yardstick to judge whether a prediction is good or bad. In any case, always remember that the future always takes you by surprise. You can't really predict it, but you may always be prepared for it.


Who

Ugo Bardi is a member of the Club of Rome, faculty member of the University of Florence, and the author of "Extracted" (Chelsea Green 2014), "The Seneca Effect" (Springer 2017), and Before the Collapse (Springer 2019)