Infactorium

I am a Systems Engineer. That’s what I do for a living, and I’m at least reasonably good at it. I specialize in healthcare, so that’s where my general expertise lies from a subject-matter perspective, but at heart I am a methodologist: I have a group of tools I use to study how Complex Systems behave. Hospitals, from a mathematical perspective, are not different from factories, or airports. And they share remarkable similarities with other systems like traffic, economies, ecosystems, and climate.

All of these systems fall into the broad category we call “hybrid dynamic systems (HDS)” (except perhaps climate, which is pretty exclusively a continuous dynamic system – which is a subset of HDSs where the discrete portion is empty… more later). Sufficiently complex HDSs, as I’m sure I’ve written before, are provably intractable. What that means is, there’s no way to write a closed-form equation that will allow you to solve for the system state at any time in the future. This is true even if the system has no random phenomena. But they almost all do, making it even harder.

Most HDSs can be conceptualized as a series of tanks and flows between them. In the case of a hospital, the things flowing may be patients, and the tanks may be inpatient units. Flow comes from admissions, the emergency department, etc.. Patients consume some resources (which flow from a reserve like inventory) and then are eventually discharged. Depending on the granularity of our model, we may be interested in the volume of patients in a unit, or in the whole hospital, or in all the hospitals in San Francisco, or the United States.

Tanks and flows are also good tools to use for modeling climate. Heat and water flow through various tanks like oceans and lakes and the atmosphere. Heat moves like water does (though generally more slowly). Tanks and flows work well for ecosystems too; consider a simple predator-prey model. Two tanks, one of predator population and another of prey, which are related by rates of reproduction and predation.

And tanks and flows work reasonably well for economies too. Volumes of money in various governmental and private holdings which flow according to spending, regulation, taxation, supply and demand, etc.. Modeling economies as HDSs is really the only way to do it, and has been the standard approach under various names and approaches for many decades.

But having a good model of a system is not the same as being able to predict how that system will behave, or even knowing how to influence the system. Without having and using a good model, it’s entirely impossible. This is because despite the fact that all these systems superficially resemble one another, they are have fundamentally distinct underlying dynamics. And as you should recall from about the only thing Jurassic Park got right: the outcome of a system is dramatically influenced by its initial conditions. Small changes in how a system is constructed or managed can cause totally different outcomes, far afield from one another.

With large scale systems, there are a couple of basic tenets that few people appropriately consider when imagining how the system will respond. The most obvious is simply inertia. When it comes to things like economies and climate, the tank part of the system is a lot bigger than the flow part of the system. That’s exactly the same situation as having a massive cargo ship and a very small engine. It takes a long time to speed it up or slow it down. With a small rudder, it takes a long time to get the vessel to turn.

To an observer on the outside, it is not at all obvious when the vessel begins to make a shift. The wheel is turned, but the ship proceeds lazily along its original trajectory. The engines are put in reverse, but the ship keeps moving forward for miles. This is true with economies too: what we call the “control surfaces” are far too small to make rapid changes in the system trajectory. A new tax, a new spending program, a deregulation, these take years to show their effects.

This is true for climate too: the massive tanks of carbon in the atmosphere, and especially in things like fossil fuels, take enormous amounts of time to be released, and once released to influence the behavior of the overall system. For more than a century we’ve been dumping carbon into the atmosphere on an industrial scale and the system is only now swerving into critical territory. And because our control surface to manage it back is so small, I have very little hope that we can do anything about planetary warming within my lifetime. Even if the entire human race was on board.

Trump is almost certainly going to shift the levers toward a worsening economy. Depressing governmental resources that allow money to flow from larger tanks to smaller ones. Deregulating checks that prevent the accumulation and hoarding of wealth into subsystems designed to extract it. Kicking supports out from education and social systems. But the effects of these policies will take a long time to have an effect on the system trajectory. The system has a lot of inertia.

So far, Trump’s worst ideas have been reasonably well-thwarted. He hasn’t implemented his repeal of the healthcare laws yet, nor has he implemented disastrous tax cuts in an already very low-tax country. I think those things are coming, but they’re not here yet. And even if they do come, they’ll take years to have their effect on the economy. Trump doesn’t get credit for the job growth or stock market yet, and he won’t have earned the blame yet if they sour in the next few months. Things just don’t move that fast.

What Trump gets blame for is being an anti-constitutional traitor, full of stupid and dangerous ideas. For being a racist and a hatemonger. For being a thief and a liar. And for being an incompetent, bewildered narcissist. But not yet for ruining (or having any other observable effect on) the economy though policy implementation.