## 15 Nov 2012

### Are models* useful?

SW asked if he should take a course with a heavy emphasis on modelling as a means of understanding the economy and environment. Here's my response...

LOTS of people use models, but the big question is whether they are USEFUL.

Academics love them (modelling in your room, alone). Policy makers love them (move the levers of society and the future). Do businesses love them? That's not so clear since THEY have real money on the line.

Modelling is often abused and misleading, so the main idea is to NOT take your results as conclusive but as a demonstration of your thoughts on how the world works -- or demonstration of how little you know of how it works!

Alfred Marshall (famous economist from 100 yrs ago) said this:
1. Use mathematics as shorthand language, rather than as an engine of inquiry.
2. Keep to them till you have done.
3. Translate into English.
4. Then illustrate by examples that are important in real life.
5. Burn the mathematics.
6. If you can’t succeed in 4, burn 3. This I do often.
If you want an example of using models to "understand," then look at my most intensive modelling paper, where you will see the pre-ordained result of my model specification and parameterization. The big point, then, is that I used it to clarify my OWN feelings/intuition. Does it prove anything? No. Will it change someone's mind? Not unless they agree with my pre-conditions. Will it be useful? Yes, if they reconsider how THEY model the world and those issues.

Bottom Line: Modelling is only as good as your starting conditions: garbage in, garbage out.

* There's a blog called "economists do it with models" but she's a bit too clever for my taste.

#### 1 comment:

Jay said...

The statistician George E. P. Box said, "All models are wrong, but some are useful."

You can learn a lot by building models. One of the most important things you can learn is how sensitive your results are to your assumptions and inputs.

Chaotic systems show a sensitive dependence on inital conditions. A change in the fourth decimal place in the input can lead to large differences in the output of things like weather systems (and economic systems).

Models make you state your assumptions. They should also make you humble about how precise and how accurate your predictions are. The track record of people making predictions should be disclosed with each new prediction.