Complex problems don't have final solutions
Almost every strategy works at first and eventually fails
One of the curious facts about weight-loss diets is that they all “seem to work to some extent – even those that are based on diametrically opposed principles, such as low-fat, low-carbohydrate, paleo, and vegan diets”1. The same is also true for strength training in which “for six weeks or so, everything works”2.
This principle is true for control projects in most complex systems - almost every strategy works at first and eventually fails if carried too far. In his book ‘Seeing Like A State’, James Scott describes how “scientific” forestry in 18th and 19th century Germany transformed German forests into monocultures created with the sole purpose of maximising the volume of wood that could be extracted. Although the exercise ended in catastrophic failure, the extent of the failure was only clear after a century and what preceded it was, in Scott’s own words, “a resounding success”. Similarly, the ‘Greenspan Put’ era of monetary policy started with a long period of success, the ‘Great Moderation’, but we may now be living through its failure mode.
As Holling and Meffe have observed,
The initial phase of command and control is nearly always quite successful: insect pests are reduced by pesticide use; fishing and hunting are enhanced by stocking or predator removal; forest fires are suppressed for years; floods are minimized by levees.
So what explains this phenomenon? There are a couple of fundamental reasons for this pattern of initial success and eventual failure.
System Adaptation
There is a time lag between the time of the introduction of the control strategy and the time by which the system and its components/agents have adapted to the strategy. The length of this lag depends upon the nature of this adaptation. For example, the “short run” of success in Scott’s example of German forestry was rather long by human standards given that “a single crop rotation of trees might take eighty years to mature” and the eventual failure was due to the emergence of pests specialised to the monocultures planted to maximise yield. Similarly, new diets can temporarily change your adiposity set point and new exercise regimens can “shock” your muscles, but this does not last.
If we are dealing with a system where the control strategy has to deal with human adaptation rather than natural selection, the timescales can be much shorter. Nevertheless, there is almost always a significant time lag before performance deteriorates.
Built-up Capital and its Depletion
Most systems have a built-up reservoir of capital that can be drawn upon for a long period before system performance deteriorates. This is especially true of the human body. Even nutritionally deficient diets can be maintained for long periods before the deficiency leads to health problems as the built-up nutritional base is slowly consumed and run down. Similarly, an imbalanced exercise regimen focused on one body part can be maintained for weeks before it leads to negative consequences.
The first generation of trees planted in modern German forestry succeeded “because it was living off (or mining) the long-accumulated soil capital of the diverse old-growth forest that it had replaced”.
No final solutions
The failure modes described in the examples above do not imply that the control strategies were wrong. After all, the entire history of agriculture is littered with this pattern of initial successes and subsequent failures, but our species has typically found a way to recover from the failure. What it does mean is that there are no final solutions to any complex problem. Sooner or later, the strategy must be changed, and a new approach must be adopted. For example, any strength training regimen only works for so long, after which a change is needed to maintain progress.
Failure is merely success taken too far
We must also not allow the initial success to blind us to the eventual failure. A failed strategy is merely a successful strategy taken too far and held on for too long. We often learn too much from our success. For example, the successes in the Dano-Prussian War in 1864, the Austro-Prussian War in 1866 and the Franco-Prussian War in 1870-71 embedded the flaws of the Prussian army “ever deeper into the country’s institutions, only to be repeated with ever more disastrous results in the subsequent two world wars”3.
Often, we mistake the success of the strategy to be an absolute scientific “truth”. Central bankers today view the characteristics of the monetary policy regime of the past few decades - independent central banking, inflation targeting, the Greenspan/Bernanke put doctrine etc. - as components of a scientific macroeconomic policy whereas they may simply be tools that happened to prove useful in fighting battles of the past and are now past their sell-by date.
As William Langewiesche once argued, everything that can go wrong does not go wrong. In fact, “everything that can go wrong usually goes right, and then we draw the wrong conclusion”.
Stephan Guyenet in ‘The Hungry Brain’.
Dan John, Dave Draper, and Pavel Tsatsouline in ‘Never Let Go’.
Peter Wilson in ‘Iron and Blood’.