When people suggest that human ingenuity will find a way out of the various predicaments we have wandered into, they base this optimism in part on extrapolation. History shows what the British call “muddling through”, adding a bunch of ad hoc rules and policies to adapt gradually to changes. The trouble with this is familiar to any software developer. You end up with a system that is overly complex, poorly understood, fragile, and increasingly difficult to muddle with. This is the usual dynamic of failed software.
The other basis for such confidence is the history of ingenuity itself, which is actually an impulse toward elegance and informed design. If you delegate well-defined technical problems to appropriate experts, they generally find a way out. The delegation process is fraught, of course. So where possible, you let anyone willing and able to declare themselves expert, and set up a competition between them. The competition can in many cases solve the delegation process – whoever succeeds in solving the problem is rewarded.
Feedback appears twice in this sort of solution. The first is obvious – you solve the delegation problem by mapping it onto a marketplace problem, and reward the winner, that is, the outfit or outfits that best solve the problem, handsomely. Demand for the product measures the fitness of the product for the problem at hand. Then perhaps, you may choose to muddle things. You modify policies to encourage the further progress of the new solution.
But there’s another place feedback enters – within the individual solutions at hand: say, mining systems, power plants, bridges, pharmaceutical plants, etc. When you look at the actual technical solutions that constitute the real basis of modern civilization, you find explicit, mathematically elegant feedback loops wherever you look. My prior brief essay was intended to introduce people to the rudiments of feedback and control theory. It’s not that I expect conversation here to proceed in the Laplace Transform Domain. However, I would like it if people understood that there is a whole domain of expertise about feedback systems that touches on predictability, stability, and communication, key concepts for coping with the Very Big Problem.
Norbert Wiener, a pioneer in this domain who coined the term “cybernetics” to identify it, failed to get very far beyond the engineering problems which so advanced the practice of engineering in the mid-twentieth century. But he did note that many problems of governance had analogous structures. In addition to his well-known early dabbling in Artificial Intelligence (which gave us the vernacular meaning of “cybernetic” in use today) Wiener also was very concerned with the relationship of employment to the larger structure of society. He was deeply concerned as to what would happen as a consequence of the” automation” of manufacturing processes leading to a decrease in demand for low to medium skilled labor. (How this was avoided for a half century, and why it is finally rising to bite us now, is something on which I have some speculative hypotheses, but I’ll stick here to what I’m pretty certain of).
All the talk of “feedback” in the climate system also relates to these concepts, even though the climate is not (as yet) an engineered system.
So we can start with natural homeostasis. A “homeostatic” system is one which has a tendency to return to a given state when preserved. The short term temperature (assuming fixed atmospheric concentration) of the earth’s surface is roughly homeostatic. If a perturbation occurs that warms the surface, more infrared radiation is emitted from the surface, allowing the surface to cool. Cloud albedo, notoriously, makes this process more complex but not so much as to run it backwards – that would make the process unstable as opposed to homeostatic. To fully characterize the system leads to mathematics far more complex and less tractable than that in an engineering situation, but the presence of the key features of homeostasis are there.
One key feature of such systems is the dominant time constant – how long it takes a perturbation to go away. One of the key design goals of engineered systems is often to minimize this time (though with systems like shock absorbers, it is to extend it). It can be of the order of seconds or even much shorter. In a natural homeostasis, a key objective is to discover that time. (For technical reasons the “time constant” is typically about a third of the time it takes a perturbed system to settle back into the noise – it’s the inverse of the rate of the decaying exponential, sometimes called the “e-folding time”.)
Economic systems are artifacts, but they are not properly considered engineered systems. What is or isn’t privately owned (surface water, highways, airports, schools, land, air, music, pictures of mice with big round black ears) is contextual, and is decided by the political process of “muddling through.” Economic theory, to my limited understanding, tends to divorce itself from all this messiness and focus on theoretical aggregations of exchanges among free agents. If my understanding is right, this is one of the points where it suffers badly from its lack of overlap with other sciences. It’s as if a climate model were to ignore the surface boundary level altogether, for instance. You’d end up with a much simpler abstraction, and one not entirely without utility (see “quasigeostrophic”) but you would not have a complete theory of climate or a model that could be integrated indefinitely to replicate the gross features of climate.
This makes economics hard, because if viewed as a science, it includes politics as a key system. Climate projection can factor this out into “emissions scenarios”, based on a few large decisions that we will eventually make, even if only by default. The relationship of economics to policy is far more tightly coupled. And predicting politics is surely beyond foreseeable human capacity, Asimov’s Foundation notwithstanding. (I mean, even though I read the trilogy at 14, I knew he was desperately handwaving.)
But there is no question that there are homeostatic subsystems within economics. These have, perhaps, emerged to some extent in a competition among social structures. Some of them, perhaps, were built in to the very idea of a market. If people won’t buy my product at my price, I have to change my product, or my price, or the way I market it. My failure in the marketplace feeds back on my behavior. Then, when people do start buying, I get feedback that I am doing something right, and can start to scale up.
There is homeostasis at a macroeconomic level as well. A particularly important one was discovered by Jane Jacobs, and explained in her book “Cities and the Wealth of Nations”. Her thesis in that book is “one city, one currency”. She illustrates how currency forms the feedback mechanism in a healthy homeostasis in world trade. In short, if an economic zone is doing well, it can export, causing the value of its currency to rise, and allowing increased value of imports. Conversely, if a zone is in decline, its currency will fall, and it is obliged to rely more on local resources, building up local skills and services, and eventually allowing it a position in export again. She points out that most nations are larger than a single economic zone, and this is what encourages the formation of economically dominant regions in a currency zone, whose interests are best represented by that currency: London, Toronto, Madrid, Milano, the Ruhr.
Just as physics is both the greatest of the sciences and a poor paradigm for the rest, the USA is the greatest of the economies and also a poor paradigm. On the whole, the American people are immensely mobile and willing to move for employment, far more than people in other countries who tend to be rooted to their ancestral region. The country is so vast, so naturally wealthy, and so young, that a dominant region has not emerged since the relative decline of New York. Both conurbations in California, Chicago, the Texas triangle, and even swaths of the southeast and northwest are competitive. Many people are willing to move between them. So in Jacobs’ sense, America manages to be more like an enormous city than like a large nation.
This fact was an enormous distraction for the Europeans. Nowhere else on earth are national boundaries as important as in Europe, especially western Europe, which has long been sorted into linguistic and cultural affinity zones. The idea of creating a common currency was meant largely to be competitive with American industrial might but also with America’s geopolitical influence. The trouble is, Europeans are not fungible. While there may be important exceptions, on the whole a Greek fellow cannot function very well in Dusseldorf. And the Greeks maintain their own parliament, their own electoral rules, their own laws, and their own political traditions.
What this does is break the cybernetic loop that Jacobs defines (without actually getting into cybernetics jargon). The performance of the Greek system does not affect the Greek currency, because the Greek currency is dominated by what happens in Dusseldorf. So the process of import-substitution does not take place, because what money does end up in the hands of a Greek can equally purchase goods and services from Paris or Dusseldorf as from Athens. But Athens does not become a provincial city ancillary to Dusseldorf because the Athenians have not conceded control of their governance along with control of their currency. Consequently, the system becomes unstable.
(When the Euro idea first came along, I thought it would fail along these lines. I should say what I think even if it’s unorthodox, I guess. Jane Jacobs lived until 2006. I wonder if she said anything about the Euro.)
The lesson here goes beyond the Euro crisis. There is a key generalization here, one which is fairly obvious but somehow not widely perceived. A homeostatic system requires feedback. Break the feedback loop and you lose control.
You cannot control what you don’t measure. This is so obvious in an engineering worldview that it isn’t even mentioned, yet the whole of Europe somehow missed the point, (Admittedly, Jacobs’ import-substitution model is not mainstream economics. More’s the pity, apparently.)
And this is the context in which I want to consider how the economic systems we have deal with sustainability (badly). The reason that the system is off the rails is only crudely because “we don’t account for externalities”. That would, of course, be better than what we have. But the question is how the accounting should take place.