A Question about economic models?

Question

Can the usefulness of a model be deduced from the realism of its assumption?

Answer 1

Arguably not. Let's use a (ridiculous, but pertinent) example:

The idea that your computer will work better if you scoop ice cream into it is patently absurd. It is a totally unrealistic assumption. But if some wacko out there DID scoop a bunch of ice cream into a bunch of computers and made one work better because of it, do you ignore the better-working computer, or take the ice cream out? Of course not.

So to get back to your question, if someone made a whole bunch of unrealistic assumptions in an economic model but it works anyway, then you can hardly say the model isn't useful - it produced useful results, even if you don't understand how it can possibly do so! In some senses it might be MORE useful, because the fact that it works may help economists adjust theories on what is actually important and what is not.

All the same, I wouldn't START by making unrealistic assumptions in a theory.

In the short short version - it's not what goes into a model that determines its usefulness, but what comes out!

Answer 2

There a classical economist joke about this:

three scientists are the sole survivors of an aircrash in the ocean, they've reached a desert island. By searching the wrecks, they found a crate of tincans, but they're wondering how they could open them.

The first scientist is a biologist, so he says:
"We should leave the cans into a pool of seawater, so they will rust and we can open them easily."
The others disagree, they say it would take too long and that they're gonna starve.

The second one is a physicist, and he says:
"We should drop the cans from this cliff over there, so the impact will open the cans."
The others disagree, they say the content is gonna be splattered and spoiled if they do this.

The third one, an economist, starts then:
"Let's assume we have a can opener..."

The thing is, the best "models", those which give the most accurate predictions, are usually statiscal models, with very little theory behind them, but they're not useful at all when there are disturbances.

Most mainstream economists consider unrealistic assumptions are ok if the results are satisfying, and some even consider that the assumptions should be the norms to change reality...

I would consider that economics also serves to understand a certain aspect of the world, hence the necessity of realistic assumptions.

Answer 3

No. The assumptions need not be realistic for the model to work well. If the model predicts behavior adequately then the assumptions can be false or nearly false. The test for a model is its predictive power. Economists often make simplifying assumptions particularly in the segments of the model that they are not studying. So, for example, you may make simplifying assumptions about the money market if you are studying employment but not make simplifying assumptions about employment because you are studying that.

Answer 4

realism of assumption is required as well as the realism of its out come or deductions,,,,, this means that the more people can relate their problem and their solutions to ur model ... the more useful it is..... although some far from realism models can be proven true but not useful as every one cant relate to them and hence cant be generalized leaving them useless.... got it?


lets reffer to the example above... ie it may be possible to get the pc work better with icecream ... . but how many ppl will b able to adapt to it will show its usefulness.

Answer 5

Not necessarily.