Hysteresis is a Hiatus Based on Hardware History

Hysteresis is a phenomenon that has been identified in many places in engineering, science and even economics. The idea behind hysteresis is that a system's future outputs will depend on past and present inputs. In other words, the way something behaves in the future depends on how it was treated in the past.

Let's take a rubber band as a simple but accurate example. You are putting weights onto a rubber band and measuring how much it stretches. Imagine this is a brand new rubber band, the kind that's really stiff until you've stretched it out a couple of times. As you put weights on your new rubber band, it stretches, but not by very much. If you drew a graph of the load you applied versus the stretch distance, it might look like this:

Now let's say you've stretched the rubber band almost until it breaks and you start taking weights off again because you're going to need this rubber band to be intact later so you can shoot it at your friend/coworker/arch nemesis. The rubber band has stretched out and it's elastic properties have changed a little since you started, so the rubber band doesn't contract to exactly the same position it was in when it last had a particular amount of weight on it. If we drew this situation on the same graph as before, it might look like this:

This effect - where the rubber band doesn't follow the same path in unloading as it did during loading - is hysteresis. We call this kind of a graph a "hysteresis loop." And we can draw this same loop for many different kinds of phenomenon, like magnetism. Below is a curve of how a magnet's level of magnetism varies with a changing magnetic field.

Here there is hysteresis because not all of the magnetic domains in a magnet go back to zero/neutral when the magnetic field is removed. Both the rubber band and the magnet "remember" the loads that they saw in the past and it affects how they respond now.

But how can you apply hysteresis to something completely different from science, like economics? It turns out you can think of unemployment in an economy over time as a hysteresis loop. Let's say we have very low unemployment and the economy is flourishing - we're at a corner of the loop. Then something bad happens to the economy and we see a fast rise in unemployment. Eventually the economy slowly starts to recover and unemployment decreases again. However, unemployment doesn't drop as quickly as it rises - remember that hysteresis says a system responds in the future based on past events. In this case, some workers who are unemployed stay unemployed for longer because their skills atrophy, or they struggle with the job hunt and they have a hard time finding a job. So while many workers can quickly lose their jobs in an economic downturn, there are a variety of factors that slow down the rate at which workers regain their jobs. This is an example of hysteresis because these workers don't get their jobs back as quickly as they lose them.