Thursday, September 11, 2008


"What's going to happen to the economy over the next six months?"

That's a question often asked to economists, and one which they are ill-equipped to answer.

An economy is a complicated, moving thing comprised of a myriad interacting forces, each of which affects the others, causing repercussions by which it is affected in return.

By and large (pace, time series analysis) economists gave up long ago on trying to figure out what happens to the economy as it moves. Instead, they focus their efforts on calculating what the economy will look like once it stops moving.

Wait - didn't I just say an economy is always in motion?

Quite so.

This is why shocks - stuff well out of the ordinary - are the darlings of economics. "What's going to happen in this complicated but smoothly-run economy that hasn't had a crisis in a hundred years?" is a very difficult question. "What's going to happen to the economy now that every forest in the country has burned down?" is much easier to answer.

A shock is something that happens suddenly, and has ripples that disappear in finite time. By looking at a long enough time scale, an economist can find a 'resting-point', and describe the characteristics of the economy once the shock has run its course. This can be the basis of useful advice, allowing governments and individuals to prepare and plan for the future.

The resting-points are called 'equilibria', after the Latin for 'equal weights'.

When equal weights are placed on either side of a scale, it is balanced. Eventually it will reach a state in which the two sides of the scale are at an equal height. All forces cancel each other out. The weights have no reason to rise, or to fall. The dust has cleared, and all influences have played their course.

To see how economic inquiry works, consider a kitten.

A kitten is in the living room. If you're lucky, the living room is sealed off from the rest of the house. A friend asks you, "What's the kitten going to do over the next six hours?"

It is a hopeless task to try to calculate the exact motion of a kitten in advance. There are too many uncertainties and random factors at play to be able to make anything other than a wild guess.

A reasonable non-economist would answer the question with, "I dunno. Walk around? Swat at things? Meow? Hopefully not make a mess..."

An economist would answer, "The kitten will be neither asleep nor awake, but on the cusp of both states."

(There's a reason many physicists become economists.)

"Buh?" the startled friend may answer.

This is where the economist would pull out a graph.

Assume that the kitten's behaviour can be entirely described by the changes in its exertion over the six-hour period. Everything else is, by this assumption, constant.

The two forces driving the kitten are curiosity and exhaustion. The kitten exerts itself in order to satisfy its curiosity - running over to a promising shiny thing on the other side of the room, for example. Exertion, therefore, lowers curiosity. If we plot Curiosity on a graph with Exertion as the horizontal axis, it should look like a downward-sloping... well, something.

We have no idea how it is that the kitten's curiosity varies with exertion, exactly, so we'll just assume the relationship is linear, and draw a downward-sloping straight line.

What's that you say? There could be trouble with our results if it turns out the true relationship is more of a curve, or an M-shape, or the outline of a gazebo? Or if we're out to lunch entirely and pulling our assumptions from somewhere unpleasnt? Okay, sure, but that's something for the econometricians (economists who work with real-world numbers) to figure out later. This is theory.

Now for exhaustion. Clearly, as the kitten exerts itself, its level of exhaustion rises. So we draw an upward-sloping line on our graph to represent exhaustion.

We have an upward-sloping line, and a downward-sloping line. If they cross at all, they cross exactly once - and they MUST cross.

This is actually easy to prove. Suppose the kitten has not yet exerted itself. Exertion is zero. Its curiosity must be very high, and its exhaustion must be very low. In particular, curiosity must be higher than exhaustion - otherwise the kitten would just sleep forever. Hunger? Oh, we're assuming that's constant. Yes, forever. No, the kitten won't die from starvation.

If curiosity starts out higher than exhaustion, and curiosity is a downward-sloping line while exhaustion is an upward-sloping line, eventually they MUST cross - after which they'll never cross again.

Before they cross, curiosity is higher than exhaustion. This means that the kitten will keep exerting itself, in order to satisfy its curiosity. We'll keep moving to the right on our graph, in the direction of increasing exertion.

After they cross, exhaustion is higher than curiosity. The kitten is too tired to explore, and falls asleep. Sleep can be modeled as a state of negative exertion - resting allows the kitten to recover energy and renews its curiosity, possibly due to a very short attention span and lousy short-term memory. We move left-ward on our graph, in the direction of decreasing exertion.

The end result? There is only one equilibrium in this model: the point at which exhaustion and curiosity are exactly equal to each other. The kitten is not exploring, since curiosity is not higher than exhaustion, but neither is it sleeping, because exhaustion is not higher than curiosity. The two forces are completely balanced. Should the kitten by some random accident wander away from this equilibrium, circumstances would conspire to move it back to this odd state, as we have argued above.

Hence, the answer to the question "What will the kitten do for the next six hours?": "The kitten will be neither asleep nor awake, but on the cusp of both states."

I DID mention economists were lousy at talking about the motion of a smoothly-functioning economy, right?

Introduce a shock, though...

Suppose that the living room is flooded with poison gas while the kitten is still inside. If the economist is again asked the same question, she will answer correctly: "The kitten will eventually be dead, and will remain in that steady state for the rest of time, assuming that there are no (presumably electrical) shocks to its system."

I leave the diagram for that situation as an exercise to the reader. It's distressingly similar to the one I walked you through.

I've been reading Naomi Klein's Shock Doctrine at the request of a student. It's an excellent book, and well worth reading by anyone who feels comfortable with the English language. The discussion is more about psychology, power and politics than economics, but it does talk a lot about economists and the phenomenally tragic consequences that can follow their mistakes.

One thing that struck me while reading the book is that many of the worst economic mistakes it refers to were made by people who appeared to have taken economics's focus on equilibrium literally. That is, these people acted as if they believed that because equilibria are what economists solve for, they must therefore be the only relevant thing, and everything else should be swept aside. The results are predictably tragic.

Consider our kitten example, and a well-meaning but much too enthusiastic economist. After modeling the kitten's situation, she comes to the following conclustion:

"No matter how you draw the two lines of curiosity and exhaustion, the result is always the same: the only equilibrium, and the point the kitten will move toward, is the twilight state of 'barely awake'. The only thing that distinguishes one equilibrium from another is the amount of exertion at which this state is reached."

The policy implications? That's clear enough. The economist's advice would be as follows:

"Clearly, it's better to be at an equilibrium with low exertion than one with high exertion. It is more restful, and on the rare occasions when due to random chance the kitten stumbles into momentary wakefulness, there is a higher level of curiosity, which is good for the mind. I will therefore drug the kitten with morphine, so that it is always in this twilight state, even when its body tells it that it is fully rested."

What about possible side effects from giving morphine to a kitten? Addiction? Shouldn't the kitten be fed once in a while? Wasn't it going to be in the living room for only six hours?

None of those considerations are in the model. And if they were, well, those are matters for other specialties to consider. The economist has done her job, and besides, the reasoning is mathematical and ironclad.

It's not terribly difficult to see how an economist who is too enthusiastic about models can wreck an economy.

At the same time, an economist can be of great help to an economy. That's why I'm in the profession, after all. I truly believe that educating others about economic thinking can help the world - more than that, I believe that economic education is absolutely necessary if the world's situation is to improve. At least as much harm has been done through ignorance of economics as by its earnest misapplication (see Zimbabwe).

The economist's profession is one of the few that can actually help to end world hunger. This is a great opportunity, and a terrible responsibility - the same force that can end starvation may also inflict it, if misapplied.

1 comment:

Rebecca said...

You know the only thing this post is missing is a link to That Darn Video of the cute kitten falling alseep on its paws.
Pweeze? ^^